Kristina Lewis, MD, MPH

Study Overview

Objective. To evaluate whether higher BMI alone contributes to risk of cardiovascular disease (CVD) and death.

Study design. Cohort study of weight-discordant monozygotic twin pairs

Setting and participants. This study took place in Sweden, using a subset of data from the Swedish Twin Registry and the Screening Across Lifespan Twin (SALT) study, which aimed to screen Swedish twins born prior to 1958 for the development of “common complex diseases.” From a total of 44,820 individuals, the current study limited to a subset of 4046 monozygotic twin pairs where both twins had self-reported height and weight data, and where calculated body mass index (BMI) was discordant between the twins, defined as a difference > 0.01 kg/m². No other inclusion or exclusion criteria are mentioned. Data for the study were collected from several different sources, including telephone interviews (eg, height and weight, behaviors such as physical activity and smoking), national registries on health conditions (eg, myocardial infarction [MI], stroke, diabetes) or prescriptions (eg, diabetes medications), the national causes of death register, and a nationwide database containing socioeconomic variables (eg, income and education). The primary exposure of interest for this study was weight status, categorized as “leaner” or “heavier,” depending on the relative BMI of each twin in a given pair. “Leaner” twins were assumed to have lower adiposity than their “heavier” counterparts, and yet to have identical genetic makeup, thereby allowing the authors to eliminate the contribution of genetic confounding in evaluating the relationship between weight status and CVD risk. The classification system could mean that one person with a BMI of 26 kg/m² would be placed in the “leaner” category if their twin had a BMI of 28, while someone else in another twin pair but also with a BMI of 26 kg/m² might be classified in the “heavier” category if their twin had a BMI of 22. Twin pairs were followed for up to 15 years to assess for incident outcomes of interest, with baseline data collected between 1998 and 2002, and follow-up through 2013.

Main outcome measures. The primary outcome of interest was the occurrence of incident MI or death from any cause. As above, these outcomes were assessed using national disease and death registries spanning 1987-2013, and ICD-9 or -10 codes of interest. A secondary outcome of incident diabetes was also specified, presumably limited to development of type 2 diabetes mellitus, and identified using the same datasets, as well as the national prescription registry. Kaplan-Meier curves for incident MI and death were constructed comparing all “leaner” twins against all “heavier” twins, and Cox proportional hazards modeling was used to compare the hazard of the primary composite outcome between groups. Logistic regression was used to evaluate the odds of each outcome including diabetes incidence, and several models were built, ranging from an unadjusted model to one adjusting for a number of lifestyle factors (eg, smoking status, physical activity), baseline health conditions, and sociodemographic factors.

The authors separately examined risk of MI/death in the subgroup of twins where the “heavier” twin had a BMI ≤ 24.9 kg/m² at baseline (ie, despite being labeled “heavier” they still had a technically normal BMI), and examined the impact of weight trajectory prior to the defined baseline (eg, they were able to incorporate into models whether someone had been actively gaining or losing weight over time prior to the baseline exposure categorization). The authors also conducted several sensitivity analyses, including running models excluding twins with < 1 year of follow-up in an effort to insure that results of the main analysis were not biased due to differential loss to follow-up between exposure categories.

Results. Of the 4046 twin pairs in this study, 56% (2283 pairs) were female, and mean (SD) age at baseline was 57.6 (9.5) years. Race/ethnicity was not reported but presumably the vast majority, if not all, are non-Hispanic white, based on the country of origin. In comparing the group of “heavier” twins to “leaner” twins, several important baseline differences were found. By design, the “heavier” twins had significantly higher mean (SD) BMI at study baseline (25.9 [3.6] kg/m² vs. 23.9 [3.1] kg/m²) and reported greater increases in BMI over the 15–20 years preceding baseline (change since 1973 was +4.3 [2.9] BMI units for “heavier” twins, vs. +2.6 [2.6] for “leaner” twins). Smoking status differed significantly between groups, with 15% of “heavier” twins reporting they were current smokers versus ~21% of “leaner” twins. “Leaner” twins were also slightly more active than their “heavier” counterparts (50.4% reported getting “rather much or very much” exercise versus 46.5%). The groups were otherwise very similar with respect to marital status, educational level, income, and baseline diagnoses of MI, stroke, diabetes, cancer or alcohol abuse.

In fully adjusted models over a mean (SD) 12.4 (2.5)-year follow-up, “heavier” twins had a significantly lower odds of MI or death (combined) than “leaner” twins (odds ratio [OR] 0.75, 95% CI 0.63–0.91). Because the “heavier” vs. “leaner” dichotomy did not map to clinical definitions of overweight or obesity, the investigators also examined this primary outcome among subgroups with more clinical relevance. Being “heavier” actually had the greatest protective effect against MI/death (OR 0.61, 95% CI 0.46–0.80) among pairs where the so-called “heavier” twin had a normal BMI (< 25.0 kg/m²), and this subgroup appeared to be driving the overall finding of lower odds of MI/death in the “heavier” group as a whole. This pattern was underscored when examining the subgroup of twin pairs where the “heavier” twin had a BMI ≥ 30 kg/m² at baseline – in this group the protective effect of being “heavier” disappeared (OR 0.92, 95% CI 0.60 to 1.42). Besides not always reflecting clinically relevant weight categories, the “heavier” vs. “leaner” twin dichotomy could, in some cases, amount to a very small difference in BMI between twins (anything > 0.01 unit counted as discordant). As such, the investigators sought to examine whether their results held up when looking at pairs with a higher threshold for BMI discordance (1.0 to 7.0 units or more difference between twins), finding that risk of MI or death did not increase among the “heavier” group in these more widely split twin pairs, even when adjusting for smoking status and physical activity.

In contrast to the MI/mortality analyses, “heavier” twins did have significantly greater odds of developing diabetes during follow-up compared to their “leaner” counterparts (OR 1.94, 95% CI 1.51 to 2.48, adjusted for smoking and physical activity). Also unlike the MI/death analyses, this relationship of increased diabetes risk among “heavier” twins was enhanced by increasing BMI dissimilarity between twins, and among twins who had been gaining weight prior to baseline BMI measurement.

Sensitivity analyses excluding twins with less than 1 year of follow-up did not result in changes to the main findings—“heavier” twins still had similar odds of MI/death as “leaner” twins.

Conclusion. The authors conclude that among monozygotic twin pairs, where the possibility for genetic confounding has been eliminated, obesity is not causally associated with increased risk of MI or death, although the results do support an increased risk of developing incident diabetes among individuals with higher BMI.

Commentary

Obesity is a known risk factor for many chronic conditions, including diabetes, osteoarthritis, sleep apnea, and hypertension [1]. However, the relationship between obesity and cardiovascular outcomes, particularly coronary artery disease and death from heart disease, has been more controversial. Some epidemiologic studies have demonstrated reduced mortality risk among patients with obesity and heart failure, and even among those with established coronary artery disease—the so-called “obesity paradox” [2]. Others have observed that overweight older adults may have lower overall mortality compared to their normal weight counterparts [3]. On the other hand, it is known that obesity increases risk for diabetes, which is itself a clear and proven risk factor for CVD and death.

As the authors of the current study point out, genetic confounding may be a potential reason for the conflicting results produced in studies of the obesity–CVD risk relationship. In other words, patients who have genes that promote weight gain may also have genes that promote CVD, through pathways independent of excess adipose tissue, with these hidden pathways acting as confounders of the obesity–CVD relationship. By studying monozygotic twin pairs, who have identical genetic makeup but have developed differential weight status due to different environmental exposures, the investigators designed a study that would eliminate any genetic confounding and allow them to better isolate the relationship between higher BMI and CVD. This is an important topic area because, at a population level, we are faced with an immense number of adults who have obesity. Treatment of this condition is resource intense and it is critical that patients and health care systems understand the potential risk reduction that will be achieved with sustained weight loss.

The strengths of this study include the use of a very unique dataset with longitudinal measures on a large number of monozygotic twin pairs, and the authors’ ability to link this dataset with nationwide comprehensive datasets on health conditions, health care use (pharmacy), sociodemographics, and death. Sweden’s national registries are quite impressive and permit these types of studies in a way that would be very difficult to achieve in the United States, with its innumerable separate health care systems and few data sources that contain information on all citizens. Because of these multiple data sources, the authors were able to adjust for some important lifestyle factors that could easily confound the weight status-MI/death relationship, such as smoking and physical activity. Additionally, their models were able to factor in trajectory of weight on some individuals prior to baseline, rather than viewing baseline weight only as a “snapshot” which could risk missing an important trend of weight gain or loss over time, with important health implications.

There are several limitations of the study that are worth reviewing. First, and most importantly, as pointed out in a commentary associated with the article, the categorization of “leaner” and “heavier” can be somewhat misleading if the true question is whether or not excess adiposity is an independent driver of cardiovascular risk [4]. BMI, at the individual level, is not an ideal measure of adiposity and it does not speak to distribution of fat tissue, which is critically important in evaluating CVD risk [5]. For example, 2 siblings could have identical BMIs, but one might have significantly more lean mass in their legs and buttocks, and the other could have more central adipose tissue, translating to a much higher cardiovascular risk. Measures such as waist circumference are critical factors in addition to BMI to better understand an individual’s adipose tissue volume and distribution.

Although the authors did adjust for some self-reported behaviors that are important predictors of CVD (smoking, exercise), there is still potential for confounding due to unscreened or unreported exposures that differ systematically between “leaner” and “heavier” twins. Of note, smoking status—probably the single most important risk factor for CVD—was missing in 13% of the cohort, and no imputation techniques were used for missing data. Another limitation of this study is that its generalizability to more racial/ethnically diverse populations may be limited. Presumably, the patients in this study were non-Hispanic white Swedes, and whether or not these findings would be replicated in other groups, such as those of African or Asian ancestry, is not known.

Finally, the finding that “heavier” twins had greater odds of developing diabetes during follow-up is certainly consistent with existing literature. However, it is also known that diabetes is a strong risk factor for the development of CVD, including MI, and for death [6]. This raises the question of why the authors observed an increased diabetes risk yet no change in MI/death rates among heavier twins. Most likely the discrepancy is due to inadequate follow-up time of incident diabetes cases. Complications of diabetes can take a number of years to materialize, and, with an average of 12 years’ total follow-up in this study, there simply may not have been time to observe an increased risk of MI/death in heavier twins.

Applications for Clinical Practice

For patients interested in weight loss as a way of reducing CVD risk, this paper does not support the notion that lower body weight alone exerts direct influence on this endpoint. However, it reinforces the link between higher body weight and diabetes, which is a clear risk factor for CVD. Therefore, it still seems reasonable to advise patients who are at risk of diabetes that improving dietary quality, increasing cardiorespiratory fitness, and losing weight can reduce their long-term risk of CVD, even if indirectly so.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To evaluate whether higher BMI alone contributes to risk of cardiovascular disease (CVD) and death.

Study design. Cohort study of weight-discordant monozygotic twin pairs

Setting and participants. This study took place in Sweden, using a subset of data from the Swedish Twin Registry and the Screening Across Lifespan Twin (SALT) study, which aimed to screen Swedish twins born prior to 1958 for the development of “common complex diseases.” From a total of 44,820 individuals, the current study limited to a subset of 4046 monozygotic twin pairs where both twins had self-reported height and weight data, and where calculated body mass index (BMI) was discordant between the twins, defined as a difference > 0.01 kg/m². No other inclusion or exclusion criteria are mentioned. Data for the study were collected from several different sources, including telephone interviews (eg, height and weight, behaviors such as physical activity and smoking), national registries on health conditions (eg, myocardial infarction [MI], stroke, diabetes) or prescriptions (eg, diabetes medications), the national causes of death register, and a nationwide database containing socioeconomic variables (eg, income and education). The primary exposure of interest for this study was weight status, categorized as “leaner” or “heavier,” depending on the relative BMI of each twin in a given pair. “Leaner” twins were assumed to have lower adiposity than their “heavier” counterparts, and yet to have identical genetic makeup, thereby allowing the authors to eliminate the contribution of genetic confounding in evaluating the relationship between weight status and CVD risk. The classification system could mean that one person with a BMI of 26 kg/m² would be placed in the “leaner” category if their twin had a BMI of 28, while someone else in another twin pair but also with a BMI of 26 kg/m² might be classified in the “heavier” category if their twin had a BMI of 22. Twin pairs were followed for up to 15 years to assess for incident outcomes of interest, with baseline data collected between 1998 and 2002, and follow-up through 2013.

Main outcome measures. The primary outcome of interest was the occurrence of incident MI or death from any cause. As above, these outcomes were assessed using national disease and death registries spanning 1987-2013, and ICD-9 or -10 codes of interest. A secondary outcome of incident diabetes was also specified, presumably limited to development of type 2 diabetes mellitus, and identified using the same datasets, as well as the national prescription registry. Kaplan-Meier curves for incident MI and death were constructed comparing all “leaner” twins against all “heavier” twins, and Cox proportional hazards modeling was used to compare the hazard of the primary composite outcome between groups. Logistic regression was used to evaluate the odds of each outcome including diabetes incidence, and several models were built, ranging from an unadjusted model to one adjusting for a number of lifestyle factors (eg, smoking status, physical activity), baseline health conditions, and sociodemographic factors.

The authors separately examined risk of MI/death in the subgroup of twins where the “heavier” twin had a BMI ≤ 24.9 kg/m² at baseline (ie, despite being labeled “heavier” they still had a technically normal BMI), and examined the impact of weight trajectory prior to the defined baseline (eg, they were able to incorporate into models whether someone had been actively gaining or losing weight over time prior to the baseline exposure categorization). The authors also conducted several sensitivity analyses, including running models excluding twins with < 1 year of follow-up in an effort to insure that results of the main analysis were not biased due to differential loss to follow-up between exposure categories.

Results. Of the 4046 twin pairs in this study, 56% (2283 pairs) were female, and mean (SD) age at baseline was 57.6 (9.5) years. Race/ethnicity was not reported but presumably the vast majority, if not all, are non-Hispanic white, based on the country of origin. In comparing the group of “heavier” twins to “leaner” twins, several important baseline differences were found. By design, the “heavier” twins had significantly higher mean (SD) BMI at study baseline (25.9 [3.6] kg/m² vs. 23.9 [3.1] kg/m²) and reported greater increases in BMI over the 15–20 years preceding baseline (change since 1973 was +4.3 [2.9] BMI units for “heavier” twins, vs. +2.6 [2.6] for “leaner” twins). Smoking status differed significantly between groups, with 15% of “heavier” twins reporting they were current smokers versus ~21% of “leaner” twins. “Leaner” twins were also slightly more active than their “heavier” counterparts (50.4% reported getting “rather much or very much” exercise versus 46.5%). The groups were otherwise very similar with respect to marital status, educational level, income, and baseline diagnoses of MI, stroke, diabetes, cancer or alcohol abuse.

In fully adjusted models over a mean (SD) 12.4 (2.5)-year follow-up, “heavier” twins had a significantly lower odds of MI or death (combined) than “leaner” twins (odds ratio [OR] 0.75, 95% CI 0.63–0.91). Because the “heavier” vs. “leaner” dichotomy did not map to clinical definitions of overweight or obesity, the investigators also examined this primary outcome among subgroups with more clinical relevance. Being “heavier” actually had the greatest protective effect against MI/death (OR 0.61, 95% CI 0.46–0.80) among pairs where the so-called “heavier” twin had a normal BMI (< 25.0 kg/m²), and this subgroup appeared to be driving the overall finding of lower odds of MI/death in the “heavier” group as a whole. This pattern was underscored when examining the subgroup of twin pairs where the “heavier” twin had a BMI ≥ 30 kg/m² at baseline – in this group the protective effect of being “heavier” disappeared (OR 0.92, 95% CI 0.60 to 1.42). Besides not always reflecting clinically relevant weight categories, the “heavier” vs. “leaner” twin dichotomy could, in some cases, amount to a very small difference in BMI between twins (anything > 0.01 unit counted as discordant). As such, the investigators sought to examine whether their results held up when looking at pairs with a higher threshold for BMI discordance (1.0 to 7.0 units or more difference between twins), finding that risk of MI or death did not increase among the “heavier” group in these more widely split twin pairs, even when adjusting for smoking status and physical activity.

In contrast to the MI/mortality analyses, “heavier” twins did have significantly greater odds of developing diabetes during follow-up compared to their “leaner” counterparts (OR 1.94, 95% CI 1.51 to 2.48, adjusted for smoking and physical activity). Also unlike the MI/death analyses, this relationship of increased diabetes risk among “heavier” twins was enhanced by increasing BMI dissimilarity between twins, and among twins who had been gaining weight prior to baseline BMI measurement.

Sensitivity analyses excluding twins with less than 1 year of follow-up did not result in changes to the main findings—“heavier” twins still had similar odds of MI/death as “leaner” twins.

Conclusion. The authors conclude that among monozygotic twin pairs, where the possibility for genetic confounding has been eliminated, obesity is not causally associated with increased risk of MI or death, although the results do support an increased risk of developing incident diabetes among individuals with higher BMI.

Commentary

Obesity is a known risk factor for many chronic conditions, including diabetes, osteoarthritis, sleep apnea, and hypertension [1]. However, the relationship between obesity and cardiovascular outcomes, particularly coronary artery disease and death from heart disease, has been more controversial. Some epidemiologic studies have demonstrated reduced mortality risk among patients with obesity and heart failure, and even among those with established coronary artery disease—the so-called “obesity paradox” [2]. Others have observed that overweight older adults may have lower overall mortality compared to their normal weight counterparts [3]. On the other hand, it is known that obesity increases risk for diabetes, which is itself a clear and proven risk factor for CVD and death.

As the authors of the current study point out, genetic confounding may be a potential reason for the conflicting results produced in studies of the obesity–CVD risk relationship. In other words, patients who have genes that promote weight gain may also have genes that promote CVD, through pathways independent of excess adipose tissue, with these hidden pathways acting as confounders of the obesity–CVD relationship. By studying monozygotic twin pairs, who have identical genetic makeup but have developed differential weight status due to different environmental exposures, the investigators designed a study that would eliminate any genetic confounding and allow them to better isolate the relationship between higher BMI and CVD. This is an important topic area because, at a population level, we are faced with an immense number of adults who have obesity. Treatment of this condition is resource intense and it is critical that patients and health care systems understand the potential risk reduction that will be achieved with sustained weight loss.

The strengths of this study include the use of a very unique dataset with longitudinal measures on a large number of monozygotic twin pairs, and the authors’ ability to link this dataset with nationwide comprehensive datasets on health conditions, health care use (pharmacy), sociodemographics, and death. Sweden’s national registries are quite impressive and permit these types of studies in a way that would be very difficult to achieve in the United States, with its innumerable separate health care systems and few data sources that contain information on all citizens. Because of these multiple data sources, the authors were able to adjust for some important lifestyle factors that could easily confound the weight status-MI/death relationship, such as smoking and physical activity. Additionally, their models were able to factor in trajectory of weight on some individuals prior to baseline, rather than viewing baseline weight only as a “snapshot” which could risk missing an important trend of weight gain or loss over time, with important health implications.

There are several limitations of the study that are worth reviewing. First, and most importantly, as pointed out in a commentary associated with the article, the categorization of “leaner” and “heavier” can be somewhat misleading if the true question is whether or not excess adiposity is an independent driver of cardiovascular risk [4]. BMI, at the individual level, is not an ideal measure of adiposity and it does not speak to distribution of fat tissue, which is critically important in evaluating CVD risk [5]. For example, 2 siblings could have identical BMIs, but one might have significantly more lean mass in their legs and buttocks, and the other could have more central adipose tissue, translating to a much higher cardiovascular risk. Measures such as waist circumference are critical factors in addition to BMI to better understand an individual’s adipose tissue volume and distribution.

Although the authors did adjust for some self-reported behaviors that are important predictors of CVD (smoking, exercise), there is still potential for confounding due to unscreened or unreported exposures that differ systematically between “leaner” and “heavier” twins. Of note, smoking status—probably the single most important risk factor for CVD—was missing in 13% of the cohort, and no imputation techniques were used for missing data. Another limitation of this study is that its generalizability to more racial/ethnically diverse populations may be limited. Presumably, the patients in this study were non-Hispanic white Swedes, and whether or not these findings would be replicated in other groups, such as those of African or Asian ancestry, is not known.

Finally, the finding that “heavier” twins had greater odds of developing diabetes during follow-up is certainly consistent with existing literature. However, it is also known that diabetes is a strong risk factor for the development of CVD, including MI, and for death [6]. This raises the question of why the authors observed an increased diabetes risk yet no change in MI/death rates among heavier twins. Most likely the discrepancy is due to inadequate follow-up time of incident diabetes cases. Complications of diabetes can take a number of years to materialize, and, with an average of 12 years’ total follow-up in this study, there simply may not have been time to observe an increased risk of MI/death in heavier twins.

Applications for Clinical Practice

For patients interested in weight loss as a way of reducing CVD risk, this paper does not support the notion that lower body weight alone exerts direct influence on this endpoint. However, it reinforces the link between higher body weight and diabetes, which is a clear risk factor for CVD. Therefore, it still seems reasonable to advise patients who are at risk of diabetes that improving dietary quality, increasing cardiorespiratory fitness, and losing weight can reduce their long-term risk of CVD, even if indirectly so.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To evaluate whether higher BMI alone contributes to risk of cardiovascular disease (CVD) and death.

Study design. Cohort study of weight-discordant monozygotic twin pairs

Setting and participants. This study took place in Sweden, using a subset of data from the Swedish Twin Registry and the Screening Across Lifespan Twin (SALT) study, which aimed to screen Swedish twins born prior to 1958 for the development of “common complex diseases.” From a total of 44,820 individuals, the current study limited to a subset of 4046 monozygotic twin pairs where both twins had self-reported height and weight data, and where calculated body mass index (BMI) was discordant between the twins, defined as a difference > 0.01 kg/m². No other inclusion or exclusion criteria are mentioned. Data for the study were collected from several different sources, including telephone interviews (eg, height and weight, behaviors such as physical activity and smoking), national registries on health conditions (eg, myocardial infarction [MI], stroke, diabetes) or prescriptions (eg, diabetes medications), the national causes of death register, and a nationwide database containing socioeconomic variables (eg, income and education). The primary exposure of interest for this study was weight status, categorized as “leaner” or “heavier,” depending on the relative BMI of each twin in a given pair. “Leaner” twins were assumed to have lower adiposity than their “heavier” counterparts, and yet to have identical genetic makeup, thereby allowing the authors to eliminate the contribution of genetic confounding in evaluating the relationship between weight status and CVD risk. The classification system could mean that one person with a BMI of 26 kg/m² would be placed in the “leaner” category if their twin had a BMI of 28, while someone else in another twin pair but also with a BMI of 26 kg/m² might be classified in the “heavier” category if their twin had a BMI of 22. Twin pairs were followed for up to 15 years to assess for incident outcomes of interest, with baseline data collected between 1998 and 2002, and follow-up through 2013.

Main outcome measures. The primary outcome of interest was the occurrence of incident MI or death from any cause. As above, these outcomes were assessed using national disease and death registries spanning 1987-2013, and ICD-9 or -10 codes of interest. A secondary outcome of incident diabetes was also specified, presumably limited to development of type 2 diabetes mellitus, and identified using the same datasets, as well as the national prescription registry. Kaplan-Meier curves for incident MI and death were constructed comparing all “leaner” twins against all “heavier” twins, and Cox proportional hazards modeling was used to compare the hazard of the primary composite outcome between groups. Logistic regression was used to evaluate the odds of each outcome including diabetes incidence, and several models were built, ranging from an unadjusted model to one adjusting for a number of lifestyle factors (eg, smoking status, physical activity), baseline health conditions, and sociodemographic factors.

The authors separately examined risk of MI/death in the subgroup of twins where the “heavier” twin had a BMI ≤ 24.9 kg/m² at baseline (ie, despite being labeled “heavier” they still had a technically normal BMI), and examined the impact of weight trajectory prior to the defined baseline (eg, they were able to incorporate into models whether someone had been actively gaining or losing weight over time prior to the baseline exposure categorization). The authors also conducted several sensitivity analyses, including running models excluding twins with < 1 year of follow-up in an effort to insure that results of the main analysis were not biased due to differential loss to follow-up between exposure categories.

Results. Of the 4046 twin pairs in this study, 56% (2283 pairs) were female, and mean (SD) age at baseline was 57.6 (9.5) years. Race/ethnicity was not reported but presumably the vast majority, if not all, are non-Hispanic white, based on the country of origin. In comparing the group of “heavier” twins to “leaner” twins, several important baseline differences were found. By design, the “heavier” twins had significantly higher mean (SD) BMI at study baseline (25.9 [3.6] kg/m² vs. 23.9 [3.1] kg/m²) and reported greater increases in BMI over the 15–20 years preceding baseline (change since 1973 was +4.3 [2.9] BMI units for “heavier” twins, vs. +2.6 [2.6] for “leaner” twins). Smoking status differed significantly between groups, with 15% of “heavier” twins reporting they were current smokers versus ~21% of “leaner” twins. “Leaner” twins were also slightly more active than their “heavier” counterparts (50.4% reported getting “rather much or very much” exercise versus 46.5%). The groups were otherwise very similar with respect to marital status, educational level, income, and baseline diagnoses of MI, stroke, diabetes, cancer or alcohol abuse.

In fully adjusted models over a mean (SD) 12.4 (2.5)-year follow-up, “heavier” twins had a significantly lower odds of MI or death (combined) than “leaner” twins (odds ratio [OR] 0.75, 95% CI 0.63–0.91). Because the “heavier” vs. “leaner” dichotomy did not map to clinical definitions of overweight or obesity, the investigators also examined this primary outcome among subgroups with more clinical relevance. Being “heavier” actually had the greatest protective effect against MI/death (OR 0.61, 95% CI 0.46–0.80) among pairs where the so-called “heavier” twin had a normal BMI (< 25.0 kg/m²), and this subgroup appeared to be driving the overall finding of lower odds of MI/death in the “heavier” group as a whole. This pattern was underscored when examining the subgroup of twin pairs where the “heavier” twin had a BMI ≥ 30 kg/m² at baseline – in this group the protective effect of being “heavier” disappeared (OR 0.92, 95% CI 0.60 to 1.42). Besides not always reflecting clinically relevant weight categories, the “heavier” vs. “leaner” twin dichotomy could, in some cases, amount to a very small difference in BMI between twins (anything > 0.01 unit counted as discordant). As such, the investigators sought to examine whether their results held up when looking at pairs with a higher threshold for BMI discordance (1.0 to 7.0 units or more difference between twins), finding that risk of MI or death did not increase among the “heavier” group in these more widely split twin pairs, even when adjusting for smoking status and physical activity.

In contrast to the MI/mortality analyses, “heavier” twins did have significantly greater odds of developing diabetes during follow-up compared to their “leaner” counterparts (OR 1.94, 95% CI 1.51 to 2.48, adjusted for smoking and physical activity). Also unlike the MI/death analyses, this relationship of increased diabetes risk among “heavier” twins was enhanced by increasing BMI dissimilarity between twins, and among twins who had been gaining weight prior to baseline BMI measurement.

Sensitivity analyses excluding twins with less than 1 year of follow-up did not result in changes to the main findings—“heavier” twins still had similar odds of MI/death as “leaner” twins.

Conclusion. The authors conclude that among monozygotic twin pairs, where the possibility for genetic confounding has been eliminated, obesity is not causally associated with increased risk of MI or death, although the results do support an increased risk of developing incident diabetes among individuals with higher BMI.

Commentary

Obesity is a known risk factor for many chronic conditions, including diabetes, osteoarthritis, sleep apnea, and hypertension [1]. However, the relationship between obesity and cardiovascular outcomes, particularly coronary artery disease and death from heart disease, has been more controversial. Some epidemiologic studies have demonstrated reduced mortality risk among patients with obesity and heart failure, and even among those with established coronary artery disease—the so-called “obesity paradox” [2]. Others have observed that overweight older adults may have lower overall mortality compared to their normal weight counterparts [3]. On the other hand, it is known that obesity increases risk for diabetes, which is itself a clear and proven risk factor for CVD and death.

As the authors of the current study point out, genetic confounding may be a potential reason for the conflicting results produced in studies of the obesity–CVD risk relationship. In other words, patients who have genes that promote weight gain may also have genes that promote CVD, through pathways independent of excess adipose tissue, with these hidden pathways acting as confounders of the obesity–CVD relationship. By studying monozygotic twin pairs, who have identical genetic makeup but have developed differential weight status due to different environmental exposures, the investigators designed a study that would eliminate any genetic confounding and allow them to better isolate the relationship between higher BMI and CVD. This is an important topic area because, at a population level, we are faced with an immense number of adults who have obesity. Treatment of this condition is resource intense and it is critical that patients and health care systems understand the potential risk reduction that will be achieved with sustained weight loss.

The strengths of this study include the use of a very unique dataset with longitudinal measures on a large number of monozygotic twin pairs, and the authors’ ability to link this dataset with nationwide comprehensive datasets on health conditions, health care use (pharmacy), sociodemographics, and death. Sweden’s national registries are quite impressive and permit these types of studies in a way that would be very difficult to achieve in the United States, with its innumerable separate health care systems and few data sources that contain information on all citizens. Because of these multiple data sources, the authors were able to adjust for some important lifestyle factors that could easily confound the weight status-MI/death relationship, such as smoking and physical activity. Additionally, their models were able to factor in trajectory of weight on some individuals prior to baseline, rather than viewing baseline weight only as a “snapshot” which could risk missing an important trend of weight gain or loss over time, with important health implications.

There are several limitations of the study that are worth reviewing. First, and most importantly, as pointed out in a commentary associated with the article, the categorization of “leaner” and “heavier” can be somewhat misleading if the true question is whether or not excess adiposity is an independent driver of cardiovascular risk [4]. BMI, at the individual level, is not an ideal measure of adiposity and it does not speak to distribution of fat tissue, which is critically important in evaluating CVD risk [5]. For example, 2 siblings could have identical BMIs, but one might have significantly more lean mass in their legs and buttocks, and the other could have more central adipose tissue, translating to a much higher cardiovascular risk. Measures such as waist circumference are critical factors in addition to BMI to better understand an individual’s adipose tissue volume and distribution.

Although the authors did adjust for some self-reported behaviors that are important predictors of CVD (smoking, exercise), there is still potential for confounding due to unscreened or unreported exposures that differ systematically between “leaner” and “heavier” twins. Of note, smoking status—probably the single most important risk factor for CVD—was missing in 13% of the cohort, and no imputation techniques were used for missing data. Another limitation of this study is that its generalizability to more racial/ethnically diverse populations may be limited. Presumably, the patients in this study were non-Hispanic white Swedes, and whether or not these findings would be replicated in other groups, such as those of African or Asian ancestry, is not known.

Finally, the finding that “heavier” twins had greater odds of developing diabetes during follow-up is certainly consistent with existing literature. However, it is also known that diabetes is a strong risk factor for the development of CVD, including MI, and for death [6]. This raises the question of why the authors observed an increased diabetes risk yet no change in MI/death rates among heavier twins. Most likely the discrepancy is due to inadequate follow-up time of incident diabetes cases. Complications of diabetes can take a number of years to materialize, and, with an average of 12 years’ total follow-up in this study, there simply may not have been time to observe an increased risk of MI/death in heavier twins.

Applications for Clinical Practice

For patients interested in weight loss as a way of reducing CVD risk, this paper does not support the notion that lower body weight alone exerts direct influence on this endpoint. However, it reinforces the link between higher body weight and diabetes, which is a clear risk factor for CVD. Therefore, it still seems reasonable to advise patients who are at risk of diabetes that improving dietary quality, increasing cardiorespiratory fitness, and losing weight can reduce their long-term risk of CVD, even if indirectly so.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To understand the risk of diabetes, coronary heart disease (CHD), stroke, and death over time by comparing adults who were “metabolically healthy” but obese at baseline and those with baseline cardiometabolic abnormalities, and to understand the stability of metabolic health over time.

Design. Secondary analysis of data from 2 large prospective epidemiologic cohort studies of cardiovascular risk and outcomes.

Setting and participants.  This study relied on data from the Atherosclerosis Risk in Communities (ARIC) study and the Coronary Artery Risk Development in Young Adults (CARDIA) study. ARIC includes adults who were recruited in late middle-age (45–64 years at baseline) from several clinical sites in the Southeastern and Midwestern United States and who have been followed over time with multiple examinations to assess cardiovascular risk factors and outcomes. CARDIA similarly assessed cardiovascular risk behaviors and events over time in a large cohort of American men and women, however, it recruited younger participants at baseline (18–30 years old). For the present study, the authors used data from all available ARIC and CARDIA participants who had complete information on body mass index (BMI) and cardiometabolic health status and who had not already developed one of the outcomes of interest at baseline, which led to a final sample of 4990 individuals from CARDIA and 14,685 from ARIC.

The independent variable of interest in this study was twofold, describing baseline status in terms of cardiometabolic risk markers and weight. Cardiometabolic risk was categorized as either “healthy,” “suboptimal,” or “unhealthy” based on the presence or absence of 3 risk factors: (1) elevated blood pressure (with untreated threshold of < 130/85 mm Hg considered negative); (2) elevated blood glucose (with untreated threshold of fasting < 100 mg/dL or hemoglobin A1c < 5.7% considered negative); and (3) dyslipidemia (with untreated total cholesterol < 240 mg/dL and HDL cholesterol > 40 for men or > 50 for women considered negative). Participants who were negative for all 3 risk factors were deemed metabolically healthy, those with 1 or 2 risk factors “suboptimal,” and those with all 3 risk factors “unhealthy.” Participants were then further characterized by baseline weight status as “lean” (BMI < 25), “overweight” (BMI 25–29.9), or “obese” (BMI ≥ 30). Combining a participant’s metabolic and weight status therefore yielded 9 possible exposure categories, ranging from  healthy-lean to unhealthy-obesity. The group of greatest interest for this study was participants in the metabolic-ally healthy-obesity (MHO) category—those with a BMI ≥ 30 but with zero of the 3 cardiometabolic risk factors.

Main outcome measures. The investigators assessed  both the stability of MHO over time, and the relative contribution of BMI status vs. cardiometabolic abnormalities to key health outcomes.

To assess the stability of MHO over time, the investigators used follow-up data from both studies to create descriptive statistics on the frequency with which patients (1) changed weight status or (2) changed metabolic status, over up to 10 years of follow-up in ARIC and up to 20 years of follow-up in CARDIA.

Among only ARIC participants (the older adults at baseline), risk of several health outcomes over up to 10 years of follow-up was compared across groups. The health outcomes of interest were incident diabetes, incident CHD (myocardial infarction or coronary death), incident stroke, or all-cause mortality. To visually represent incidence of these outcomes over time, the investigators constructed Kaplan-Meier survival curves. To determine whether the risk of an outcome differed significantly according to baseline exposure category, they used Cox proportional hazards modeling, adjusting for covariates including age, sex, race, income, education, and tobacco and alcohol use. All available follow-up time was used from baseline until an outcome of interest developed or a participant was censored. Reasons for censoring a participant were not outlined. Multivariable Cox models were separately conducted for the 4 outcomes of diabetes, CHD, stroke, and mortality across the 9 main exposure categories and across 15 categories in an additional analysis where “suboptimal cardiometabolic health” was split according to whether participants had 1 or 2 baseline risk factors. The reference group for all analyses was patients with MHO. A P value of < 0.05 was specified as statistically significant.

Results. Baseline characteristics were presented only for ARIC patients. Among that group (n = 14,685), just 2% (n = 260) were characterized as MHO at baseline. Just over one-quarter (27%) were obese at baseline, and the vast majority of patients with obesity at baseline (94%) had either suboptimal cardiometabolic risk (SO) or metabolically unhealthy obesity (MUO—all 3 risk factors present). Mean follow-up time for ARIC participants was 18.7 years.

Just under half of the ARIC sample were women (45%), 25% were black (the remaining were white), mean age was 54.3 years, and mean BMI was 27.7. Covariates such as education and income were not reported in the table of patient characteristics. No statistical testing was reported comparing exposure categories at baseline, however, within the “healthy”, “suboptimal,” and “unhealthy” categories, increasing weight status appeared to track with increasing blood pressure, fasting glucose, insulin resistance, and waist circumference.

With MHO participants as the reference group, there were no significant differences between baseline weight categories (lean, overweight, obese) of “healthy” (zero risk factors) participants for CHD, stroke, or mortality during follow-up. In other words, baseline weight status did not significantly impact the risk of these 3 outcomes, assuming someone started out metabolically healthy. However, significant differences did emerge among participants with 1 or more risk factor at baseline. For those in the “suboptimal” (SO) category (1 or 2 risk factors), all 3 weight subgroups (lean, overweight, and obese) had significantly higher risk of CHD and stroke during follow-up relative to the MHO participants (hazard ratios [HRs] for CHD: lean 2.3, overweight 2.5, obese 3.0; HRs for stroke: lean 2.6, overweight 2.7, obese 3.0), and mortality risk was higher among lean-SOs (HR 1.4) and obese-SOs (HR 1.7). For those in the unhealthy category at baseline, there was significantly higher risk of CHD, stroke, and mortality across all 3 baseline weight categories relative to participants who were MHO at baseline – that is, even “lean” at baseline patients who had 3 risk factors had significantly higher risk of all of these outcomes than “healthy obese” patients (CHD HR 3.6; stroke HR 2.9; mortality HR 1.9).

Diabetes results differed slightly. For this outcome, participants in the lean-healthy baseline category had about half the risk of developing diabetes during follow-up (HR 0.47) compared to the MHO participants. Those in the overweight-healthy and lean-suboptimal health categories had no difference in risk of diabetes compared to MHO. All other subgroups had higher risk of developing diabetes over time (eg, lean-unhealthy diabetes HR 2.3; obese-unhealthy diabetes HR 5.4) relative to MHO.

Data from both ARIC and CARDIA were used to evaluate the stability of weight and metabolic health during follow-up. Consistent with nationally observed trends in the U.S., many participants gained weight during follow-up, with 17.5% of initially lean and 67.3% of overweight CARDIA patients transitioning to obesity over time. For patients initially categorized as metabolically healthy, a large fraction developed 1 or 2 metabolic abnormalities in follow-up (52% in ARIC and 35% in CARDIA). Very few participants from either study transitioned from a healthy state of 0 risk factors to the unhealthy state of 3 risk factors during follow-up.

Conclusion. The authors conclude that patients with MHO have lower risk for diabetes, CHD, stroke, and mortality than unhealthy subjects regardless of weight status. They did note that obesity increased diabetes risk, even in the absence of detectable baseline abnormalities, relative to lean healthy individuals.

Commentary

Metabolically healthy obesity describes a state where a patient’s body mass index is above 30 kg/m² yet the individual lacks traditional measures of cardiometabolic derangement often associated with excess adiposity. The definition of MHO varies between studies but often requires that a patient with obesity display less than 3 metabolic syndrome criteria, sometimes allowing for even fewer abnormalities (eg, 0 or 1) [1]. It is estimated that anywhere between 10% to one-third of adults with obesity may fall into this category of relative metabolic health despite an elevated BMI [1,2]. Some controversy surrounds how the MHO state should be viewed and its practical implications for clinical management. It is unclear whether patients with MHO are simply in a transient state (ie, “pre-metabolic,” akin to “pre-diabetic”) that will later convert to metabolically unhealthy obesity (MUO), or whether they are truly somehow genetically able to handle excess weight without ever developing the sequelae that are so commonly observed in most patients with obesity. This is an important distinction for clinicians, as it may have implications for how aggressively weight loss is pursued and how long-term risks of excess weight are framed for these individuals. Consider a 40-year-old female patient with a BMI of 32 who is otherwise healthy and active, on no BP medications, with an optimal lipid profile and no signs of insulin resistance. Should this patient be encouraged to lose weight? What health risks does she face in the next few decades if she does not?

In this secondary data analysis from 2 large cohort studies of cardiovascular risk, Guo and Garvey conclude that it is the cardiometabolic risk markers of elevated blood pressure, dyslipidemia, and elevated blood glucose that confer far more risk in terms of long-term cardiovascular outcomes than excess weight in and of itself. Their analyses make use of data from 2 large and rigorous cohort studies of cardiovascular outcomes, which lend credibility to the outcomes they aimed to study (ie, we can be confident that if someone is listed as having had a myocardial infarction in ARIC or CARDIA, they probably did) and provided them with the unique ability to study long-term outcomes on a large sample size. In short, this study would have been difficult to do with many other sources of data or methodologies. Their statistical methods for comparing the risk of the outcomes over long-term follow-up appear robust, particularly given the high event rates in some of the groups and therefore inevitable high levels of censoring over time. Importantly, they control for a number of potential confounders in their study, including tobacco use. On the other hand, they perform quite a large number of statistical comparisons, therefore it is possible they may have found fewer significant differences between groups with a more stringent cutoff for their P value (eg, with a Bonferroni correction).

Regarding the stability of the metabolically healthy state over time, it appears there was significant crossover of participants from “healthy” to “suboptimal,” and significant weight gain occurred during follow-up. It is not clear whether an individual’s baseline exposure category was permitted to change over time in the statistical models, which could have impacted their results. Clinically, it is not surprising that there was a lot of movement between categories over up to 20 years of follow-up. It underscores the notion that even if a patient is obese and metabolically healthy cross-sectionally, many of these individuals will not remain metabolically healthy over time. Additionally, although the study abstract describes using data from both ARIC and CARDIA, the health outcomes component relied solely on ARIC participants. These were a group of relatively older adults at baseline who had already made it through much of their adult lives without developing any of the outcomes of interest (diabetes, CHD, stroke or mortality), therefore could have represented a sample that is somehow more metabolically resilient than the general population. As stated by the authors in their limitation section, the assertion that MHO patients are not at increased risk of cardiovascular disease (CVD) outcomes should not be extrapolated to younger patients based on this cohort. This is particularly true because, again, the stability of metabolic health appears relatively low—over half of baseline “healthy” participants in ARIC and over one-third in CARDIA developed 1 or more risk factors in follow-up, and therefore presumably also developed greater risk of CVD than if they had remained “metabolically healthy.” The likelihood that young adults with MHO will go on to develop new risk factors over time is underscored by how rare the MHO state was in the ARIC sample—it represented only 2% of the overall population, and 6% of those with obesity.

Additionally, as the authors noted, while CVD appeared to be much more influenced by the risk factor trio than by obesity alone, obesity did increase diabetes risk even in the “metabolically healthy” group. This finding aligns with prior work suggesting that patients with MHO are at increased risk of diabetes but not CVD, compared with their normal weight metabolically healthy counterparts [3].

Applications for Clinical Practice

Regardless of weight status, patients with risk factors such as elevated blood pressure, glucose, or lipids would benefit from interventions to reduce their long-term cardiovascular risk and mortality. On the other hand, patients with obesity who lack traditional cardiometabolic risk factors represent a clinical population where it is more difficult to advise on some of the potential benefits of weight loss. Adults with MHO can be advised with confidence that weight loss may reduce their risk of developing diabetes, and they may have other important motivations for weight loss that can be supported as well. Importantly, young adults with MHO who are not interested in weight loss should not be assumed to be “in the clear” for cardiovascular risk; they should be monitored for development of new risk factors over time and for the ensuing need for increased intensity of weight loss recommendations and interventions.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To understand the risk of diabetes, coronary heart disease (CHD), stroke, and death over time by comparing adults who were “metabolically healthy” but obese at baseline and those with baseline cardiometabolic abnormalities, and to understand the stability of metabolic health over time.

Design. Secondary analysis of data from 2 large prospective epidemiologic cohort studies of cardiovascular risk and outcomes.

Setting and participants.  This study relied on data from the Atherosclerosis Risk in Communities (ARIC) study and the Coronary Artery Risk Development in Young Adults (CARDIA) study. ARIC includes adults who were recruited in late middle-age (45–64 years at baseline) from several clinical sites in the Southeastern and Midwestern United States and who have been followed over time with multiple examinations to assess cardiovascular risk factors and outcomes. CARDIA similarly assessed cardiovascular risk behaviors and events over time in a large cohort of American men and women, however, it recruited younger participants at baseline (18–30 years old). For the present study, the authors used data from all available ARIC and CARDIA participants who had complete information on body mass index (BMI) and cardiometabolic health status and who had not already developed one of the outcomes of interest at baseline, which led to a final sample of 4990 individuals from CARDIA and 14,685 from ARIC.

The independent variable of interest in this study was twofold, describing baseline status in terms of cardiometabolic risk markers and weight. Cardiometabolic risk was categorized as either “healthy,” “suboptimal,” or “unhealthy” based on the presence or absence of 3 risk factors: (1) elevated blood pressure (with untreated threshold of < 130/85 mm Hg considered negative); (2) elevated blood glucose (with untreated threshold of fasting < 100 mg/dL or hemoglobin A1c < 5.7% considered negative); and (3) dyslipidemia (with untreated total cholesterol < 240 mg/dL and HDL cholesterol > 40 for men or > 50 for women considered negative). Participants who were negative for all 3 risk factors were deemed metabolically healthy, those with 1 or 2 risk factors “suboptimal,” and those with all 3 risk factors “unhealthy.” Participants were then further characterized by baseline weight status as “lean” (BMI < 25), “overweight” (BMI 25–29.9), or “obese” (BMI ≥ 30). Combining a participant’s metabolic and weight status therefore yielded 9 possible exposure categories, ranging from  healthy-lean to unhealthy-obesity. The group of greatest interest for this study was participants in the metabolic-ally healthy-obesity (MHO) category—those with a BMI ≥ 30 but with zero of the 3 cardiometabolic risk factors.

Main outcome measures. The investigators assessed  both the stability of MHO over time, and the relative contribution of BMI status vs. cardiometabolic abnormalities to key health outcomes.

To assess the stability of MHO over time, the investigators used follow-up data from both studies to create descriptive statistics on the frequency with which patients (1) changed weight status or (2) changed metabolic status, over up to 10 years of follow-up in ARIC and up to 20 years of follow-up in CARDIA.

Among only ARIC participants (the older adults at baseline), risk of several health outcomes over up to 10 years of follow-up was compared across groups. The health outcomes of interest were incident diabetes, incident CHD (myocardial infarction or coronary death), incident stroke, or all-cause mortality. To visually represent incidence of these outcomes over time, the investigators constructed Kaplan-Meier survival curves. To determine whether the risk of an outcome differed significantly according to baseline exposure category, they used Cox proportional hazards modeling, adjusting for covariates including age, sex, race, income, education, and tobacco and alcohol use. All available follow-up time was used from baseline until an outcome of interest developed or a participant was censored. Reasons for censoring a participant were not outlined. Multivariable Cox models were separately conducted for the 4 outcomes of diabetes, CHD, stroke, and mortality across the 9 main exposure categories and across 15 categories in an additional analysis where “suboptimal cardiometabolic health” was split according to whether participants had 1 or 2 baseline risk factors. The reference group for all analyses was patients with MHO. A P value of < 0.05 was specified as statistically significant.

Results. Baseline characteristics were presented only for ARIC patients. Among that group (n = 14,685), just 2% (n = 260) were characterized as MHO at baseline. Just over one-quarter (27%) were obese at baseline, and the vast majority of patients with obesity at baseline (94%) had either suboptimal cardiometabolic risk (SO) or metabolically unhealthy obesity (MUO—all 3 risk factors present). Mean follow-up time for ARIC participants was 18.7 years.

Just under half of the ARIC sample were women (45%), 25% were black (the remaining were white), mean age was 54.3 years, and mean BMI was 27.7. Covariates such as education and income were not reported in the table of patient characteristics. No statistical testing was reported comparing exposure categories at baseline, however, within the “healthy”, “suboptimal,” and “unhealthy” categories, increasing weight status appeared to track with increasing blood pressure, fasting glucose, insulin resistance, and waist circumference.

With MHO participants as the reference group, there were no significant differences between baseline weight categories (lean, overweight, obese) of “healthy” (zero risk factors) participants for CHD, stroke, or mortality during follow-up. In other words, baseline weight status did not significantly impact the risk of these 3 outcomes, assuming someone started out metabolically healthy. However, significant differences did emerge among participants with 1 or more risk factor at baseline. For those in the “suboptimal” (SO) category (1 or 2 risk factors), all 3 weight subgroups (lean, overweight, and obese) had significantly higher risk of CHD and stroke during follow-up relative to the MHO participants (hazard ratios [HRs] for CHD: lean 2.3, overweight 2.5, obese 3.0; HRs for stroke: lean 2.6, overweight 2.7, obese 3.0), and mortality risk was higher among lean-SOs (HR 1.4) and obese-SOs (HR 1.7). For those in the unhealthy category at baseline, there was significantly higher risk of CHD, stroke, and mortality across all 3 baseline weight categories relative to participants who were MHO at baseline – that is, even “lean” at baseline patients who had 3 risk factors had significantly higher risk of all of these outcomes than “healthy obese” patients (CHD HR 3.6; stroke HR 2.9; mortality HR 1.9).

Diabetes results differed slightly. For this outcome, participants in the lean-healthy baseline category had about half the risk of developing diabetes during follow-up (HR 0.47) compared to the MHO participants. Those in the overweight-healthy and lean-suboptimal health categories had no difference in risk of diabetes compared to MHO. All other subgroups had higher risk of developing diabetes over time (eg, lean-unhealthy diabetes HR 2.3; obese-unhealthy diabetes HR 5.4) relative to MHO.

Data from both ARIC and CARDIA were used to evaluate the stability of weight and metabolic health during follow-up. Consistent with nationally observed trends in the U.S., many participants gained weight during follow-up, with 17.5% of initially lean and 67.3% of overweight CARDIA patients transitioning to obesity over time. For patients initially categorized as metabolically healthy, a large fraction developed 1 or 2 metabolic abnormalities in follow-up (52% in ARIC and 35% in CARDIA). Very few participants from either study transitioned from a healthy state of 0 risk factors to the unhealthy state of 3 risk factors during follow-up.

Conclusion. The authors conclude that patients with MHO have lower risk for diabetes, CHD, stroke, and mortality than unhealthy subjects regardless of weight status. They did note that obesity increased diabetes risk, even in the absence of detectable baseline abnormalities, relative to lean healthy individuals.

Commentary

Metabolically healthy obesity describes a state where a patient’s body mass index is above 30 kg/m² yet the individual lacks traditional measures of cardiometabolic derangement often associated with excess adiposity. The definition of MHO varies between studies but often requires that a patient with obesity display less than 3 metabolic syndrome criteria, sometimes allowing for even fewer abnormalities (eg, 0 or 1) [1]. It is estimated that anywhere between 10% to one-third of adults with obesity may fall into this category of relative metabolic health despite an elevated BMI [1,2]. Some controversy surrounds how the MHO state should be viewed and its practical implications for clinical management. It is unclear whether patients with MHO are simply in a transient state (ie, “pre-metabolic,” akin to “pre-diabetic”) that will later convert to metabolically unhealthy obesity (MUO), or whether they are truly somehow genetically able to handle excess weight without ever developing the sequelae that are so commonly observed in most patients with obesity. This is an important distinction for clinicians, as it may have implications for how aggressively weight loss is pursued and how long-term risks of excess weight are framed for these individuals. Consider a 40-year-old female patient with a BMI of 32 who is otherwise healthy and active, on no BP medications, with an optimal lipid profile and no signs of insulin resistance. Should this patient be encouraged to lose weight? What health risks does she face in the next few decades if she does not?

In this secondary data analysis from 2 large cohort studies of cardiovascular risk, Guo and Garvey conclude that it is the cardiometabolic risk markers of elevated blood pressure, dyslipidemia, and elevated blood glucose that confer far more risk in terms of long-term cardiovascular outcomes than excess weight in and of itself. Their analyses make use of data from 2 large and rigorous cohort studies of cardiovascular outcomes, which lend credibility to the outcomes they aimed to study (ie, we can be confident that if someone is listed as having had a myocardial infarction in ARIC or CARDIA, they probably did) and provided them with the unique ability to study long-term outcomes on a large sample size. In short, this study would have been difficult to do with many other sources of data or methodologies. Their statistical methods for comparing the risk of the outcomes over long-term follow-up appear robust, particularly given the high event rates in some of the groups and therefore inevitable high levels of censoring over time. Importantly, they control for a number of potential confounders in their study, including tobacco use. On the other hand, they perform quite a large number of statistical comparisons, therefore it is possible they may have found fewer significant differences between groups with a more stringent cutoff for their P value (eg, with a Bonferroni correction).

Regarding the stability of the metabolically healthy state over time, it appears there was significant crossover of participants from “healthy” to “suboptimal,” and significant weight gain occurred during follow-up. It is not clear whether an individual’s baseline exposure category was permitted to change over time in the statistical models, which could have impacted their results. Clinically, it is not surprising that there was a lot of movement between categories over up to 20 years of follow-up. It underscores the notion that even if a patient is obese and metabolically healthy cross-sectionally, many of these individuals will not remain metabolically healthy over time. Additionally, although the study abstract describes using data from both ARIC and CARDIA, the health outcomes component relied solely on ARIC participants. These were a group of relatively older adults at baseline who had already made it through much of their adult lives without developing any of the outcomes of interest (diabetes, CHD, stroke or mortality), therefore could have represented a sample that is somehow more metabolically resilient than the general population. As stated by the authors in their limitation section, the assertion that MHO patients are not at increased risk of cardiovascular disease (CVD) outcomes should not be extrapolated to younger patients based on this cohort. This is particularly true because, again, the stability of metabolic health appears relatively low—over half of baseline “healthy” participants in ARIC and over one-third in CARDIA developed 1 or more risk factors in follow-up, and therefore presumably also developed greater risk of CVD than if they had remained “metabolically healthy.” The likelihood that young adults with MHO will go on to develop new risk factors over time is underscored by how rare the MHO state was in the ARIC sample—it represented only 2% of the overall population, and 6% of those with obesity.

Additionally, as the authors noted, while CVD appeared to be much more influenced by the risk factor trio than by obesity alone, obesity did increase diabetes risk even in the “metabolically healthy” group. This finding aligns with prior work suggesting that patients with MHO are at increased risk of diabetes but not CVD, compared with their normal weight metabolically healthy counterparts [3].

Applications for Clinical Practice

Regardless of weight status, patients with risk factors such as elevated blood pressure, glucose, or lipids would benefit from interventions to reduce their long-term cardiovascular risk and mortality. On the other hand, patients with obesity who lack traditional cardiometabolic risk factors represent a clinical population where it is more difficult to advise on some of the potential benefits of weight loss. Adults with MHO can be advised with confidence that weight loss may reduce their risk of developing diabetes, and they may have other important motivations for weight loss that can be supported as well. Importantly, young adults with MHO who are not interested in weight loss should not be assumed to be “in the clear” for cardiovascular risk; they should be monitored for development of new risk factors over time and for the ensuing need for increased intensity of weight loss recommendations and interventions.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To understand the risk of diabetes, coronary heart disease (CHD), stroke, and death over time by comparing adults who were “metabolically healthy” but obese at baseline and those with baseline cardiometabolic abnormalities, and to understand the stability of metabolic health over time.

Design. Secondary analysis of data from 2 large prospective epidemiologic cohort studies of cardiovascular risk and outcomes.

Setting and participants.  This study relied on data from the Atherosclerosis Risk in Communities (ARIC) study and the Coronary Artery Risk Development in Young Adults (CARDIA) study. ARIC includes adults who were recruited in late middle-age (45–64 years at baseline) from several clinical sites in the Southeastern and Midwestern United States and who have been followed over time with multiple examinations to assess cardiovascular risk factors and outcomes. CARDIA similarly assessed cardiovascular risk behaviors and events over time in a large cohort of American men and women, however, it recruited younger participants at baseline (18–30 years old). For the present study, the authors used data from all available ARIC and CARDIA participants who had complete information on body mass index (BMI) and cardiometabolic health status and who had not already developed one of the outcomes of interest at baseline, which led to a final sample of 4990 individuals from CARDIA and 14,685 from ARIC.

The independent variable of interest in this study was twofold, describing baseline status in terms of cardiometabolic risk markers and weight. Cardiometabolic risk was categorized as either “healthy,” “suboptimal,” or “unhealthy” based on the presence or absence of 3 risk factors: (1) elevated blood pressure (with untreated threshold of < 130/85 mm Hg considered negative); (2) elevated blood glucose (with untreated threshold of fasting < 100 mg/dL or hemoglobin A1c < 5.7% considered negative); and (3) dyslipidemia (with untreated total cholesterol < 240 mg/dL and HDL cholesterol > 40 for men or > 50 for women considered negative). Participants who were negative for all 3 risk factors were deemed metabolically healthy, those with 1 or 2 risk factors “suboptimal,” and those with all 3 risk factors “unhealthy.” Participants were then further characterized by baseline weight status as “lean” (BMI < 25), “overweight” (BMI 25–29.9), or “obese” (BMI ≥ 30). Combining a participant’s metabolic and weight status therefore yielded 9 possible exposure categories, ranging from  healthy-lean to unhealthy-obesity. The group of greatest interest for this study was participants in the metabolic-ally healthy-obesity (MHO) category—those with a BMI ≥ 30 but with zero of the 3 cardiometabolic risk factors.

Main outcome measures. The investigators assessed  both the stability of MHO over time, and the relative contribution of BMI status vs. cardiometabolic abnormalities to key health outcomes.

To assess the stability of MHO over time, the investigators used follow-up data from both studies to create descriptive statistics on the frequency with which patients (1) changed weight status or (2) changed metabolic status, over up to 10 years of follow-up in ARIC and up to 20 years of follow-up in CARDIA.

Among only ARIC participants (the older adults at baseline), risk of several health outcomes over up to 10 years of follow-up was compared across groups. The health outcomes of interest were incident diabetes, incident CHD (myocardial infarction or coronary death), incident stroke, or all-cause mortality. To visually represent incidence of these outcomes over time, the investigators constructed Kaplan-Meier survival curves. To determine whether the risk of an outcome differed significantly according to baseline exposure category, they used Cox proportional hazards modeling, adjusting for covariates including age, sex, race, income, education, and tobacco and alcohol use. All available follow-up time was used from baseline until an outcome of interest developed or a participant was censored. Reasons for censoring a participant were not outlined. Multivariable Cox models were separately conducted for the 4 outcomes of diabetes, CHD, stroke, and mortality across the 9 main exposure categories and across 15 categories in an additional analysis where “suboptimal cardiometabolic health” was split according to whether participants had 1 or 2 baseline risk factors. The reference group for all analyses was patients with MHO. A P value of < 0.05 was specified as statistically significant.

Results. Baseline characteristics were presented only for ARIC patients. Among that group (n = 14,685), just 2% (n = 260) were characterized as MHO at baseline. Just over one-quarter (27%) were obese at baseline, and the vast majority of patients with obesity at baseline (94%) had either suboptimal cardiometabolic risk (SO) or metabolically unhealthy obesity (MUO—all 3 risk factors present). Mean follow-up time for ARIC participants was 18.7 years.

Just under half of the ARIC sample were women (45%), 25% were black (the remaining were white), mean age was 54.3 years, and mean BMI was 27.7. Covariates such as education and income were not reported in the table of patient characteristics. No statistical testing was reported comparing exposure categories at baseline, however, within the “healthy”, “suboptimal,” and “unhealthy” categories, increasing weight status appeared to track with increasing blood pressure, fasting glucose, insulin resistance, and waist circumference.

With MHO participants as the reference group, there were no significant differences between baseline weight categories (lean, overweight, obese) of “healthy” (zero risk factors) participants for CHD, stroke, or mortality during follow-up. In other words, baseline weight status did not significantly impact the risk of these 3 outcomes, assuming someone started out metabolically healthy. However, significant differences did emerge among participants with 1 or more risk factor at baseline. For those in the “suboptimal” (SO) category (1 or 2 risk factors), all 3 weight subgroups (lean, overweight, and obese) had significantly higher risk of CHD and stroke during follow-up relative to the MHO participants (hazard ratios [HRs] for CHD: lean 2.3, overweight 2.5, obese 3.0; HRs for stroke: lean 2.6, overweight 2.7, obese 3.0), and mortality risk was higher among lean-SOs (HR 1.4) and obese-SOs (HR 1.7). For those in the unhealthy category at baseline, there was significantly higher risk of CHD, stroke, and mortality across all 3 baseline weight categories relative to participants who were MHO at baseline – that is, even “lean” at baseline patients who had 3 risk factors had significantly higher risk of all of these outcomes than “healthy obese” patients (CHD HR 3.6; stroke HR 2.9; mortality HR 1.9).

Diabetes results differed slightly. For this outcome, participants in the lean-healthy baseline category had about half the risk of developing diabetes during follow-up (HR 0.47) compared to the MHO participants. Those in the overweight-healthy and lean-suboptimal health categories had no difference in risk of diabetes compared to MHO. All other subgroups had higher risk of developing diabetes over time (eg, lean-unhealthy diabetes HR 2.3; obese-unhealthy diabetes HR 5.4) relative to MHO.

Data from both ARIC and CARDIA were used to evaluate the stability of weight and metabolic health during follow-up. Consistent with nationally observed trends in the U.S., many participants gained weight during follow-up, with 17.5% of initially lean and 67.3% of overweight CARDIA patients transitioning to obesity over time. For patients initially categorized as metabolically healthy, a large fraction developed 1 or 2 metabolic abnormalities in follow-up (52% in ARIC and 35% in CARDIA). Very few participants from either study transitioned from a healthy state of 0 risk factors to the unhealthy state of 3 risk factors during follow-up.

Conclusion. The authors conclude that patients with MHO have lower risk for diabetes, CHD, stroke, and mortality than unhealthy subjects regardless of weight status. They did note that obesity increased diabetes risk, even in the absence of detectable baseline abnormalities, relative to lean healthy individuals.

Commentary

Metabolically healthy obesity describes a state where a patient’s body mass index is above 30 kg/m² yet the individual lacks traditional measures of cardiometabolic derangement often associated with excess adiposity. The definition of MHO varies between studies but often requires that a patient with obesity display less than 3 metabolic syndrome criteria, sometimes allowing for even fewer abnormalities (eg, 0 or 1) [1]. It is estimated that anywhere between 10% to one-third of adults with obesity may fall into this category of relative metabolic health despite an elevated BMI [1,2]. Some controversy surrounds how the MHO state should be viewed and its practical implications for clinical management. It is unclear whether patients with MHO are simply in a transient state (ie, “pre-metabolic,” akin to “pre-diabetic”) that will later convert to metabolically unhealthy obesity (MUO), or whether they are truly somehow genetically able to handle excess weight without ever developing the sequelae that are so commonly observed in most patients with obesity. This is an important distinction for clinicians, as it may have implications for how aggressively weight loss is pursued and how long-term risks of excess weight are framed for these individuals. Consider a 40-year-old female patient with a BMI of 32 who is otherwise healthy and active, on no BP medications, with an optimal lipid profile and no signs of insulin resistance. Should this patient be encouraged to lose weight? What health risks does she face in the next few decades if she does not?

In this secondary data analysis from 2 large cohort studies of cardiovascular risk, Guo and Garvey conclude that it is the cardiometabolic risk markers of elevated blood pressure, dyslipidemia, and elevated blood glucose that confer far more risk in terms of long-term cardiovascular outcomes than excess weight in and of itself. Their analyses make use of data from 2 large and rigorous cohort studies of cardiovascular outcomes, which lend credibility to the outcomes they aimed to study (ie, we can be confident that if someone is listed as having had a myocardial infarction in ARIC or CARDIA, they probably did) and provided them with the unique ability to study long-term outcomes on a large sample size. In short, this study would have been difficult to do with many other sources of data or methodologies. Their statistical methods for comparing the risk of the outcomes over long-term follow-up appear robust, particularly given the high event rates in some of the groups and therefore inevitable high levels of censoring over time. Importantly, they control for a number of potential confounders in their study, including tobacco use. On the other hand, they perform quite a large number of statistical comparisons, therefore it is possible they may have found fewer significant differences between groups with a more stringent cutoff for their P value (eg, with a Bonferroni correction).

Regarding the stability of the metabolically healthy state over time, it appears there was significant crossover of participants from “healthy” to “suboptimal,” and significant weight gain occurred during follow-up. It is not clear whether an individual’s baseline exposure category was permitted to change over time in the statistical models, which could have impacted their results. Clinically, it is not surprising that there was a lot of movement between categories over up to 20 years of follow-up. It underscores the notion that even if a patient is obese and metabolically healthy cross-sectionally, many of these individuals will not remain metabolically healthy over time. Additionally, although the study abstract describes using data from both ARIC and CARDIA, the health outcomes component relied solely on ARIC participants. These were a group of relatively older adults at baseline who had already made it through much of their adult lives without developing any of the outcomes of interest (diabetes, CHD, stroke or mortality), therefore could have represented a sample that is somehow more metabolically resilient than the general population. As stated by the authors in their limitation section, the assertion that MHO patients are not at increased risk of cardiovascular disease (CVD) outcomes should not be extrapolated to younger patients based on this cohort. This is particularly true because, again, the stability of metabolic health appears relatively low—over half of baseline “healthy” participants in ARIC and over one-third in CARDIA developed 1 or more risk factors in follow-up, and therefore presumably also developed greater risk of CVD than if they had remained “metabolically healthy.” The likelihood that young adults with MHO will go on to develop new risk factors over time is underscored by how rare the MHO state was in the ARIC sample—it represented only 2% of the overall population, and 6% of those with obesity.

Additionally, as the authors noted, while CVD appeared to be much more influenced by the risk factor trio than by obesity alone, obesity did increase diabetes risk even in the “metabolically healthy” group. This finding aligns with prior work suggesting that patients with MHO are at increased risk of diabetes but not CVD, compared with their normal weight metabolically healthy counterparts [3].

Applications for Clinical Practice

Regardless of weight status, patients with risk factors such as elevated blood pressure, glucose, or lipids would benefit from interventions to reduce their long-term cardiovascular risk and mortality. On the other hand, patients with obesity who lack traditional cardiometabolic risk factors represent a clinical population where it is more difficult to advise on some of the potential benefits of weight loss. Adults with MHO can be advised with confidence that weight loss may reduce their risk of developing diabetes, and they may have other important motivations for weight loss that can be supported as well. Importantly, young adults with MHO who are not interested in weight loss should not be assumed to be “in the clear” for cardiovascular risk; they should be monitored for development of new risk factors over time and for the ensuing need for increased intensity of weight loss recommendations and interventions.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To compare several behavioral strategies for weight gain prevention in young adults.

Study design. Randomized clinical trial.

Setting and participants. The study took place at 2 U.S. academic centers between 2010 and 2016. Participants were recruited using email and postal mailings if they were 18–35 years old, had a body mass index (BMI) between 21 and 30.9 (ie, they ranged from normal body weight to class I obesity), spoke English, had internet access, and did not have contraindications to participating in a behavioral weight management intervention (eg, eating disorders). Once recruited, participants were block randomized, stratified by site, sex, and ethnic group, in to 1 of 3 study arms. The control arm of the study consisted of a single in-person meeting where behavioral strategies to prevent weight gain were discussed, as well as quarterly newsletters and personalized reports on interim weight data during follow-up.

Intervention. There were 2 intervention arms in the study. Both intervention groups had 10 in-person group-based visits over the initial 4 months of the intervention, at which strategies to prevent weight gain were discussed. Additionally they received annual invitations to participate in online refresher courses and the same newsletter frequency and content as the control group. Advice to the 2 intervention groups differed, however. Those in the “small changes” group were advised to decrease caloric intake by about 100 kcal per day in order to prevent weight gain. Additionally they were given pedometers, with a goal of increasing their daily step counts by about 2000. In the “large changes” group, participants were given lower calorie targets and more aggressive physical activity goals, with a goal of producing weight loss over the first 4 months of follow-up (2.3 kg for those with normal baseline BMI, and 4.5 kg if overweight or obese at baseline). Participants in all groups were encouraged to engage in self-monitoring behaviors such as daily weighing, and to report these weights to study staff by email, text, or on the web. Aside from pre-specified study follow-up assessments, most follow-up beyond the initial 4 month “small” or “large” changes phase was done using email or web-based intervention.

Main outcome measures. All participants were scheduled for follow-up assessments at 4 months, 1 year, and 2 years, with some early participants having additional follow-ups at 3 and 4 years. The primary outcome of interest was change in weight from baseline through follow-up, with additional outcome measures including the proportion in each group who gained at least 0.45 kg, or developed obesity. Additionally, the investigators did a thorough evaluation of intervention implementation and delivery. Weight change was modeled using mixed effects linear models, adjusting for clinic site. They corrected for multiple measures using Bonferroni adjustment to minimize the risk of type I error and used multiple imputation to examine the impact of missing data on their results. Pre-specified subgroup comparisons between several groups of patients were conducted—those in the normal weight vs. overweight category at baseline, those younger vs. older than age 25 at baseline, and men vs. women.

Results. 599 participants were randomized to the control (n = 202), small changes (n = 200), or large changes groups (n = 197), with no significant differences between groups in terms of measured baseline characteristics. The majority of participants were women (78%) and non-Hispanic white (73%). Mean (SD) baseline age was 28.2 (4.4) years and BMI was 25.4 (2.6) kg/m². The group as a whole was highly educated—between 77% and 82% had college degrees. The series of 10 intervention sessions in the first 4 months was very well-attended (87% attendance on average for large changes group, 86% for small changes group), and by 4 months of follow-up, a majority of participants in both intervention groups endorsed the behavior of daily self-weighing (75% in large changes, 72% in small changes).

Both intervention groups had statistically significant weight losses compared to control (average weight change in control +0.3 kg, in small change –0.6 kg, and in large change –2.4 kg, over an average of 3 years), with large change participants also having significantly greater average weight loss in follow-up than small change participants. Significantly fewer participants in the intervention groups went on to develop obesity than in the control group (16.9% incidence in control, vs. 7.9% incidence in small changes [P = 0.002] and 8.6% in large changes [P = 0.02]). Importantly, the trajectories of weight gain (or regain) after the initial 4-month intervention differed between the small and large change groups, with small change participants experiencing a more gradual rate of gain throughout follow-up, versus a steeper rate of gain in the large changes group, such that the groups were at very similar weights by the final time point. The investigators did not observe any differences in effect between subgroups according to participant baseline BMI, sex, age, or race.

Conclusion. The authors conclude that these scalable small- and large-change interventions reduced longer-term weight gain and even promoted weight loss in a group of young adults, with the large-change intervention having a greater impact on weight than the small-change intervention.

Commentary

Treatment of obesity is difficult, leading to frustration for many patients and clinicians. Although it is often possible to help patients lose weight with tools such as low-calorie diets and increased physical activity, the long-term maintenance of weight loss is quite challenging. There is a growing awareness that the difficulty in maintaining weight loss has strong physiologic underpinnings. The human body has complex energy regulatory systems that may oppose weight loss by lowering metabolic rate, increasing hunger cues, and limiting satiety cues, when faced with energy restriction or weight loss [1,2].

In order to decrease the number of patients who ultimately require treatment for obesity, an alternative approach may be to try to prevent weight gain in the first place. Young adults in the U.S. tend to gain weight steadily over time, yet this insidious pattern is unlikely to be addressed by physicians [3]. Given that gradual weight gain seems to be the norm for most young adults, it may be beneficial for primary care providers to advise all young adult patients to make small behavioral changes in order to prevent the onset of overweight or obesity. Preventing weight gain is an attractive approach for broad application because it may require lower intensity programs, and less behavioral commitment from patients, compared to what is required for weight loss [4].

In this randomized trial, Wing et al investigated several relatively low-intensity approaches for weight gain prevention. Strengths of the study include aspects of the design and analysis, including its randomized nature, the relatively long follow-up period, the use of multiple imputation to address missing data, and the use of statistical methods to account for the large number of comparisons made between groups over time (Bonferroni correction). More importantly, however, this study represents an important innovation in how physicians might think about obesity, with a shift toward prevention rather than treatment. Historically, many obesity prevention efforts have fallen in the domain of public health or population-level interventions, and it may be the case that physicians have felt they did not really have a role in prevention. On the other hand, doctors who have engaged in obesity treatment—trying to help patients lose weight—may have felt that they lacked the resources or training needed to implement successful programs to promote long-term weight loss. By testing several lower-intensity strategies for weight gain prevention, this study sheds light on what could possibly be a new role for primary care providers or health care systems who care for otherwise healthy young adults. As the authors point out, the methods they employed could also be easily scaled or disseminated using public health approaches and community organizations.

In addition to addressing an important topic, this study relied on intervention methods that would be relatively easy to replicate in clinical practice or in community settings. Aside from the initial 4-month intervention, which involved 10 face-to-face group sessions (which were very well attended by participants), the remainder of the ~3 year follow-up consisted mostly of contact that took place electronically using email and/or text messaging. These modes of communication align well with the move toward electronic health records (eg, e-visits) and are probably ideally suited for young adults, who as a group rely heavily on these methods of communication.

The study has several limitations, most of which are addressed by the authors in the discussion section of the paper. As with most studies of behavioral weight interventions, the majority of participants in this study were women, with relatively few racial and ethnic minorities. Furthermore this was a highly educated group of participants and it is unclear whether these results would generalize to a more diverse clinical population with fewer resources or lower health literacy. Given that the control arm of the study experienced less weight gain over time than would be expected based on population averages, it could be that the participants in this study were a select group of individuals who were more motivated around preventing long-term health problems than a general clinical population. One additional point of possible concern is that, while participants in the “large changes” group did, as per the design, lose weight at the beginning of the trial, they also went on to regain much of that weight and experienced a steeper trajectory of overall gain during follow-up compared to the “small changes” group, so that the 2 intervention groups were not statistically different from each other in terms of overall weight change from baseline by 2 years. Therefore, whether the “large changes” approach is truly more beneficial for long-term obesity prevention than the more modest “small changes” approach is not entirely clear from this study.

Applications for Clinical Practice

The identification of young adults who are gaining weight, but who are not yet obese, represents an opportunity for providers and health care systems. Efforts to promote modest dietary and physical activity changes in this population may prevent obesity, and may be achievable even in busy clinical practice settings. Whether weight-gain prevention programs should include an attempt to first foster a small amount of weight loss as a “buffer” against later gains is still not entirely clear.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To compare several behavioral strategies for weight gain prevention in young adults.

Study design. Randomized clinical trial.

Setting and participants. The study took place at 2 U.S. academic centers between 2010 and 2016. Participants were recruited using email and postal mailings if they were 18–35 years old, had a body mass index (BMI) between 21 and 30.9 (ie, they ranged from normal body weight to class I obesity), spoke English, had internet access, and did not have contraindications to participating in a behavioral weight management intervention (eg, eating disorders). Once recruited, participants were block randomized, stratified by site, sex, and ethnic group, in to 1 of 3 study arms. The control arm of the study consisted of a single in-person meeting where behavioral strategies to prevent weight gain were discussed, as well as quarterly newsletters and personalized reports on interim weight data during follow-up.

Intervention. There were 2 intervention arms in the study. Both intervention groups had 10 in-person group-based visits over the initial 4 months of the intervention, at which strategies to prevent weight gain were discussed. Additionally they received annual invitations to participate in online refresher courses and the same newsletter frequency and content as the control group. Advice to the 2 intervention groups differed, however. Those in the “small changes” group were advised to decrease caloric intake by about 100 kcal per day in order to prevent weight gain. Additionally they were given pedometers, with a goal of increasing their daily step counts by about 2000. In the “large changes” group, participants were given lower calorie targets and more aggressive physical activity goals, with a goal of producing weight loss over the first 4 months of follow-up (2.3 kg for those with normal baseline BMI, and 4.5 kg if overweight or obese at baseline). Participants in all groups were encouraged to engage in self-monitoring behaviors such as daily weighing, and to report these weights to study staff by email, text, or on the web. Aside from pre-specified study follow-up assessments, most follow-up beyond the initial 4 month “small” or “large” changes phase was done using email or web-based intervention.

Main outcome measures. All participants were scheduled for follow-up assessments at 4 months, 1 year, and 2 years, with some early participants having additional follow-ups at 3 and 4 years. The primary outcome of interest was change in weight from baseline through follow-up, with additional outcome measures including the proportion in each group who gained at least 0.45 kg, or developed obesity. Additionally, the investigators did a thorough evaluation of intervention implementation and delivery. Weight change was modeled using mixed effects linear models, adjusting for clinic site. They corrected for multiple measures using Bonferroni adjustment to minimize the risk of type I error and used multiple imputation to examine the impact of missing data on their results. Pre-specified subgroup comparisons between several groups of patients were conducted—those in the normal weight vs. overweight category at baseline, those younger vs. older than age 25 at baseline, and men vs. women.

Results. 599 participants were randomized to the control (n = 202), small changes (n = 200), or large changes groups (n = 197), with no significant differences between groups in terms of measured baseline characteristics. The majority of participants were women (78%) and non-Hispanic white (73%). Mean (SD) baseline age was 28.2 (4.4) years and BMI was 25.4 (2.6) kg/m². The group as a whole was highly educated—between 77% and 82% had college degrees. The series of 10 intervention sessions in the first 4 months was very well-attended (87% attendance on average for large changes group, 86% for small changes group), and by 4 months of follow-up, a majority of participants in both intervention groups endorsed the behavior of daily self-weighing (75% in large changes, 72% in small changes).

Both intervention groups had statistically significant weight losses compared to control (average weight change in control +0.3 kg, in small change –0.6 kg, and in large change –2.4 kg, over an average of 3 years), with large change participants also having significantly greater average weight loss in follow-up than small change participants. Significantly fewer participants in the intervention groups went on to develop obesity than in the control group (16.9% incidence in control, vs. 7.9% incidence in small changes [P = 0.002] and 8.6% in large changes [P = 0.02]). Importantly, the trajectories of weight gain (or regain) after the initial 4-month intervention differed between the small and large change groups, with small change participants experiencing a more gradual rate of gain throughout follow-up, versus a steeper rate of gain in the large changes group, such that the groups were at very similar weights by the final time point. The investigators did not observe any differences in effect between subgroups according to participant baseline BMI, sex, age, or race.

Conclusion. The authors conclude that these scalable small- and large-change interventions reduced longer-term weight gain and even promoted weight loss in a group of young adults, with the large-change intervention having a greater impact on weight than the small-change intervention.

Commentary

Treatment of obesity is difficult, leading to frustration for many patients and clinicians. Although it is often possible to help patients lose weight with tools such as low-calorie diets and increased physical activity, the long-term maintenance of weight loss is quite challenging. There is a growing awareness that the difficulty in maintaining weight loss has strong physiologic underpinnings. The human body has complex energy regulatory systems that may oppose weight loss by lowering metabolic rate, increasing hunger cues, and limiting satiety cues, when faced with energy restriction or weight loss [1,2].

In order to decrease the number of patients who ultimately require treatment for obesity, an alternative approach may be to try to prevent weight gain in the first place. Young adults in the U.S. tend to gain weight steadily over time, yet this insidious pattern is unlikely to be addressed by physicians [3]. Given that gradual weight gain seems to be the norm for most young adults, it may be beneficial for primary care providers to advise all young adult patients to make small behavioral changes in order to prevent the onset of overweight or obesity. Preventing weight gain is an attractive approach for broad application because it may require lower intensity programs, and less behavioral commitment from patients, compared to what is required for weight loss [4].

In this randomized trial, Wing et al investigated several relatively low-intensity approaches for weight gain prevention. Strengths of the study include aspects of the design and analysis, including its randomized nature, the relatively long follow-up period, the use of multiple imputation to address missing data, and the use of statistical methods to account for the large number of comparisons made between groups over time (Bonferroni correction). More importantly, however, this study represents an important innovation in how physicians might think about obesity, with a shift toward prevention rather than treatment. Historically, many obesity prevention efforts have fallen in the domain of public health or population-level interventions, and it may be the case that physicians have felt they did not really have a role in prevention. On the other hand, doctors who have engaged in obesity treatment—trying to help patients lose weight—may have felt that they lacked the resources or training needed to implement successful programs to promote long-term weight loss. By testing several lower-intensity strategies for weight gain prevention, this study sheds light on what could possibly be a new role for primary care providers or health care systems who care for otherwise healthy young adults. As the authors point out, the methods they employed could also be easily scaled or disseminated using public health approaches and community organizations.

In addition to addressing an important topic, this study relied on intervention methods that would be relatively easy to replicate in clinical practice or in community settings. Aside from the initial 4-month intervention, which involved 10 face-to-face group sessions (which were very well attended by participants), the remainder of the ~3 year follow-up consisted mostly of contact that took place electronically using email and/or text messaging. These modes of communication align well with the move toward electronic health records (eg, e-visits) and are probably ideally suited for young adults, who as a group rely heavily on these methods of communication.

The study has several limitations, most of which are addressed by the authors in the discussion section of the paper. As with most studies of behavioral weight interventions, the majority of participants in this study were women, with relatively few racial and ethnic minorities. Furthermore this was a highly educated group of participants and it is unclear whether these results would generalize to a more diverse clinical population with fewer resources or lower health literacy. Given that the control arm of the study experienced less weight gain over time than would be expected based on population averages, it could be that the participants in this study were a select group of individuals who were more motivated around preventing long-term health problems than a general clinical population. One additional point of possible concern is that, while participants in the “large changes” group did, as per the design, lose weight at the beginning of the trial, they also went on to regain much of that weight and experienced a steeper trajectory of overall gain during follow-up compared to the “small changes” group, so that the 2 intervention groups were not statistically different from each other in terms of overall weight change from baseline by 2 years. Therefore, whether the “large changes” approach is truly more beneficial for long-term obesity prevention than the more modest “small changes” approach is not entirely clear from this study.

Applications for Clinical Practice

The identification of young adults who are gaining weight, but who are not yet obese, represents an opportunity for providers and health care systems. Efforts to promote modest dietary and physical activity changes in this population may prevent obesity, and may be achievable even in busy clinical practice settings. Whether weight-gain prevention programs should include an attempt to first foster a small amount of weight loss as a “buffer” against later gains is still not entirely clear.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To compare several behavioral strategies for weight gain prevention in young adults.

Study design. Randomized clinical trial.

Setting and participants. The study took place at 2 U.S. academic centers between 2010 and 2016. Participants were recruited using email and postal mailings if they were 18–35 years old, had a body mass index (BMI) between 21 and 30.9 (ie, they ranged from normal body weight to class I obesity), spoke English, had internet access, and did not have contraindications to participating in a behavioral weight management intervention (eg, eating disorders). Once recruited, participants were block randomized, stratified by site, sex, and ethnic group, in to 1 of 3 study arms. The control arm of the study consisted of a single in-person meeting where behavioral strategies to prevent weight gain were discussed, as well as quarterly newsletters and personalized reports on interim weight data during follow-up.

Intervention. There were 2 intervention arms in the study. Both intervention groups had 10 in-person group-based visits over the initial 4 months of the intervention, at which strategies to prevent weight gain were discussed. Additionally they received annual invitations to participate in online refresher courses and the same newsletter frequency and content as the control group. Advice to the 2 intervention groups differed, however. Those in the “small changes” group were advised to decrease caloric intake by about 100 kcal per day in order to prevent weight gain. Additionally they were given pedometers, with a goal of increasing their daily step counts by about 2000. In the “large changes” group, participants were given lower calorie targets and more aggressive physical activity goals, with a goal of producing weight loss over the first 4 months of follow-up (2.3 kg for those with normal baseline BMI, and 4.5 kg if overweight or obese at baseline). Participants in all groups were encouraged to engage in self-monitoring behaviors such as daily weighing, and to report these weights to study staff by email, text, or on the web. Aside from pre-specified study follow-up assessments, most follow-up beyond the initial 4 month “small” or “large” changes phase was done using email or web-based intervention.

Main outcome measures. All participants were scheduled for follow-up assessments at 4 months, 1 year, and 2 years, with some early participants having additional follow-ups at 3 and 4 years. The primary outcome of interest was change in weight from baseline through follow-up, with additional outcome measures including the proportion in each group who gained at least 0.45 kg, or developed obesity. Additionally, the investigators did a thorough evaluation of intervention implementation and delivery. Weight change was modeled using mixed effects linear models, adjusting for clinic site. They corrected for multiple measures using Bonferroni adjustment to minimize the risk of type I error and used multiple imputation to examine the impact of missing data on their results. Pre-specified subgroup comparisons between several groups of patients were conducted—those in the normal weight vs. overweight category at baseline, those younger vs. older than age 25 at baseline, and men vs. women.

Results. 599 participants were randomized to the control (n = 202), small changes (n = 200), or large changes groups (n = 197), with no significant differences between groups in terms of measured baseline characteristics. The majority of participants were women (78%) and non-Hispanic white (73%). Mean (SD) baseline age was 28.2 (4.4) years and BMI was 25.4 (2.6) kg/m². The group as a whole was highly educated—between 77% and 82% had college degrees. The series of 10 intervention sessions in the first 4 months was very well-attended (87% attendance on average for large changes group, 86% for small changes group), and by 4 months of follow-up, a majority of participants in both intervention groups endorsed the behavior of daily self-weighing (75% in large changes, 72% in small changes).

Both intervention groups had statistically significant weight losses compared to control (average weight change in control +0.3 kg, in small change –0.6 kg, and in large change –2.4 kg, over an average of 3 years), with large change participants also having significantly greater average weight loss in follow-up than small change participants. Significantly fewer participants in the intervention groups went on to develop obesity than in the control group (16.9% incidence in control, vs. 7.9% incidence in small changes [P = 0.002] and 8.6% in large changes [P = 0.02]). Importantly, the trajectories of weight gain (or regain) after the initial 4-month intervention differed between the small and large change groups, with small change participants experiencing a more gradual rate of gain throughout follow-up, versus a steeper rate of gain in the large changes group, such that the groups were at very similar weights by the final time point. The investigators did not observe any differences in effect between subgroups according to participant baseline BMI, sex, age, or race.

Conclusion. The authors conclude that these scalable small- and large-change interventions reduced longer-term weight gain and even promoted weight loss in a group of young adults, with the large-change intervention having a greater impact on weight than the small-change intervention.

Commentary

Treatment of obesity is difficult, leading to frustration for many patients and clinicians. Although it is often possible to help patients lose weight with tools such as low-calorie diets and increased physical activity, the long-term maintenance of weight loss is quite challenging. There is a growing awareness that the difficulty in maintaining weight loss has strong physiologic underpinnings. The human body has complex energy regulatory systems that may oppose weight loss by lowering metabolic rate, increasing hunger cues, and limiting satiety cues, when faced with energy restriction or weight loss [1,2].

In order to decrease the number of patients who ultimately require treatment for obesity, an alternative approach may be to try to prevent weight gain in the first place. Young adults in the U.S. tend to gain weight steadily over time, yet this insidious pattern is unlikely to be addressed by physicians [3]. Given that gradual weight gain seems to be the norm for most young adults, it may be beneficial for primary care providers to advise all young adult patients to make small behavioral changes in order to prevent the onset of overweight or obesity. Preventing weight gain is an attractive approach for broad application because it may require lower intensity programs, and less behavioral commitment from patients, compared to what is required for weight loss [4].

In this randomized trial, Wing et al investigated several relatively low-intensity approaches for weight gain prevention. Strengths of the study include aspects of the design and analysis, including its randomized nature, the relatively long follow-up period, the use of multiple imputation to address missing data, and the use of statistical methods to account for the large number of comparisons made between groups over time (Bonferroni correction). More importantly, however, this study represents an important innovation in how physicians might think about obesity, with a shift toward prevention rather than treatment. Historically, many obesity prevention efforts have fallen in the domain of public health or population-level interventions, and it may be the case that physicians have felt they did not really have a role in prevention. On the other hand, doctors who have engaged in obesity treatment—trying to help patients lose weight—may have felt that they lacked the resources or training needed to implement successful programs to promote long-term weight loss. By testing several lower-intensity strategies for weight gain prevention, this study sheds light on what could possibly be a new role for primary care providers or health care systems who care for otherwise healthy young adults. As the authors point out, the methods they employed could also be easily scaled or disseminated using public health approaches and community organizations.

In addition to addressing an important topic, this study relied on intervention methods that would be relatively easy to replicate in clinical practice or in community settings. Aside from the initial 4-month intervention, which involved 10 face-to-face group sessions (which were very well attended by participants), the remainder of the ~3 year follow-up consisted mostly of contact that took place electronically using email and/or text messaging. These modes of communication align well with the move toward electronic health records (eg, e-visits) and are probably ideally suited for young adults, who as a group rely heavily on these methods of communication.

The study has several limitations, most of which are addressed by the authors in the discussion section of the paper. As with most studies of behavioral weight interventions, the majority of participants in this study were women, with relatively few racial and ethnic minorities. Furthermore this was a highly educated group of participants and it is unclear whether these results would generalize to a more diverse clinical population with fewer resources or lower health literacy. Given that the control arm of the study experienced less weight gain over time than would be expected based on population averages, it could be that the participants in this study were a select group of individuals who were more motivated around preventing long-term health problems than a general clinical population. One additional point of possible concern is that, while participants in the “large changes” group did, as per the design, lose weight at the beginning of the trial, they also went on to regain much of that weight and experienced a steeper trajectory of overall gain during follow-up compared to the “small changes” group, so that the 2 intervention groups were not statistically different from each other in terms of overall weight change from baseline by 2 years. Therefore, whether the “large changes” approach is truly more beneficial for long-term obesity prevention than the more modest “small changes” approach is not entirely clear from this study.

Applications for Clinical Practice

The identification of young adults who are gaining weight, but who are not yet obese, represents an opportunity for providers and health care systems. Efforts to promote modest dietary and physical activity changes in this population may prevent obesity, and may be achievable even in busy clinical practice settings. Whether weight-gain prevention programs should include an attempt to first foster a small amount of weight loss as a “buffer” against later gains is still not entirely clear.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To compare weight regain after rapid versus slower loss of an equivalent amount of weight.

Study design. Randomized clinical trial.

Setting and participants. This study took place in a single medical center in the Netherlands. Investigators recruited 61 adults (no age range provided) with body mass index (BMI) between 28–35 kg/m² and at a stable weight (no change of > 3 kg for the past 2 months) to participate in a weight loss study. Individuals with type 2 diabetes, dyslipidemia, uncontrolled hypertension, or liver, heart or kidney disease were excluded, as were those who were currently pregnant or reported consuming more than moderate amounts of alcohol.

Once consented, participants were randomized into one of 2 study arms. The rapid weight loss arm was prescribed a very-low-calorie diet (VLCD) with just 500 kcal/day (43% protein/43% carb/14% fat) for 5 weeks, after which they transitioned to a 4-week “weight stable” period, and then a 9-month follow-up period (overall follow-up time of ~11 months; 10 months after weight loss). In contrast, the slower weight loss arm was prescribed a low-calorie diet (LCD) with 1250 kcal/day (29% protein/48% carb/23% fat) for 12 weeks, after which they also transitioned to a 4-week weight stable period and 9 months of follow-up (overall follow-up time of ~13 months; 10 months after weight loss). VLCD (rapid weight loss) participants received 3 meal replacement shakes per day (totaling 500 kcal) during the weight loss period and were also told they could consume unlimited amounts of low-calorie vegetables. The LCD (slower weight loss) participants received 1 meal replacement shake per day during their 12 weeks of weight loss and were responsible for providing the remainder of their own meals and snacks according to guidelines from a study dietitian. Following active weight loss, both groups then shifted to higher-calorie, food-based diets during a “weight stable” 4-week period and were responsible during this time for providing all of their own food. The researchers do not specify the details of the diet composition for this weight stable period. Exposure to the registered dietitian was the same in both groups, with 5 consultations during weight loss (weekly for VLCD, presumably more spaced out for LCD) and 4 during weight stable period. No further diet advice or meal replacement support was given during the 9-month follow-up period, but participants came in for monthly weigh-ins.

Main outcome measure. The primary outcome measure was change in weight (ie, amount of weight regained) during the 9-month follow-up period, compared between groups using an independent samples t test. Additional biometric measures included change in waist circumference and changes in body composition. For the latter, the researchers used a “Bod Pod” to conduct air-displacement plethysmography and determine what percentage of an individual’s weight was fat mass (FM) versus lean mass/water (FFM [fat-free mass]). They then compared the amount of FFM lost between groups, again using the independent samples t test.

The researchers also collected information on self-reported physical activity (questionnaire) and self-reported history of weight cycling (number of times a participant had previously lost and regained at least 5 kg) prior to this study. These were not outcomes per-se, but were collected so that they could be examined as correlates of the biometric outcomes above, using Pearson and Spearman’s correlation coefficients.

Results. The LCD (n = 29) and VLCD (n = 28) groups were similar at baseline with no significant differences reported. Of the 61 individuals initially enrolled, 57 (93%) completed the study. Summary statistics are reported only for these 57 individuals. No imputation or other methods for handling missing data were used. There were slightly more women than men in the study (53% women); the average (SD) age was 51.8 (1.9) years in the LCD group and 50.7 (1.5) years in the VLCD group. Mean starting BMI was 31 kg/m² (31.3 [0.5] in LCD, 31.0 [0.4] in VLCD) and both groups had just under 40% body fat at baseline (39.9% [1.8] in LCD, 39.7% [1.5] in VLCD).

After 12 weeks of weight loss for LCD, or 5 weeks of weight loss for VLCD, both groups lost a similar amount of total weight (8.2 [0.5] kg in LCD vs. 9.0 [0.4] kg in VLCD), then had no significant changes in weight during the subsequent 4-week “weight stable” period. However, during the weight stable period VLCD patients had an average 0.8 (0.6) cm increase in waist circumference (a rebounding after a decrease of 7.7 cm during weight loss), while LCD patients on average had a continued decrease of 1.0 (0.5 cm) in waist circumference (P = 0.003).

There was no significant difference between groups for the primary outcome of weight regain during 9-months of follow-up (4.2 [0.6] kg regained for LCD, 4.5 [0.7] for VLCD; P = 0.73). The only significant correlates of weight regain were amount of FFM lost (more lean mass lost predicted more weight regain), and amount of physical activity reported during follow-up (more activity predicted less regain). Participant sex, age, starting BMI, history of weight cycling, and amount of weight lost did not correlate with rate of re-gain.

One area where there was a significant between-group difference, both after initial weight loss and persisting after the weight stable period, was in the amount of FFM lost (a rough approximation of lost lean mass, eg, muscle mass). VLCD participants had more FFM loss (1.6 [0.2] kg) than LCD participants (0.6 [0.2] kg) (P < 0.01) after active weight loss, and continued to have significantly more FFM loss (0.8 [0.2] kg vs. 0.2 [0.2] kg) after the 4-week weight stable period.

There were no between-group differences at the end of weight loss or at the end of follow-up for hip or waist circumference or for blood pressure.

Conclusion. The authors conclude that rate of weight loss does not affect one’s risk of weight regain after a diet, after a similar amount of weight has been lost.

Commentary

The failure of most diets to produce durable weight loss is a frustration for patients, clinicians, and researchers. In general, regardless of the composition of a diet, the majority of patients will regain some or all of their lost weight within several years after completing the diet. The reasons for weight regain are complex, and include reversion to old eating or physical activity behaviors but also a strong physiologic drive by the body to reverse weight loss that it perceives as a threat to health [1].

One area in diet research that has recently generated some controversy is whether or not rate of initial weight loss might impact a patient’s ability to maintain that weight loss, with the conventional wisdom (and national guidelines, in some cases), suggesting that slower weight loss is preferable to rapid weight loss for this reason [2]. A handful of studies have challenged this notion, however, and suggested that rapid weight loss does not necessarily lead to greater weight regain [3,4]. Previous such studies, however, have not generally been designed to compare regain after equal amounts of weight loss, which may make their results more difficult to interpret.

The present study contributes another piece of evidence to the argument that rapid initial weight loss may not increase a patient’s risk of regain. This small randomized trial is timely and has several features that make it a unique contribution. First, the design of the study allowed for both groups, despite losing weight at very different rates, to reach the same amount of total weight loss before being followed forward in time. This made weight regain much easier to compare between groups during follow-up. Second, the study included measurement of changing body composition—ie, what kind of weight was being lost (fat vs. fat-free mass)—rather than just the total amount of weight. This allowed the researchers to present data for an outcome that is mechanistically related to metabolic rate (and therefore weight regain), and one that might have implications for longer-term health after rapid versus more moderate-pace weight loss.

Several aspects of the study design, however, may limit the impact of the findings. For example, in both arms, while a certain type of diet was “prescribed,” there is no comment about assessment of participant fidelity to the prescribed diet, and there is potential for very different levels of adherence between groups, especially in active weight loss, when basically all meals were provided to the VLCD arm, but LCD subjects were responsible for about 90% of their own meals. This could have led to larger discrepancies between prescribed and actual diet in the LCD arm relative to VLCD. Granted, the rate of weight loss was the exposure of interest, and that clearly varied between groups as expected, implying at least moderate fidelity to prescribed caloric content of each diet, but how much protein vs. fat vs. carb was actually consumed by each group is not clear. Additionally, while 9 months of post weight-loss follow-up is certainly a good start in terms of follow-up duration, it may not have been sufficient to observe differences that would later emerge between the groups for weight regain. Other long-term weight loss maintenance studies have followed patients for several years or longer after initial weight loss [5].

Using data from all participants, the researchers reported that the amount of FFM an individual lost was a predictor of weight regain during follow-up. This finding is in keeping with the idea that more lean mass loss leads to lower metabolic rate and predisposes to weight regain (hence the conventional wisdom to avoid rapid weight loss with low-protein diets). In keeping with this theme, VLCD patients, whose protein intakes and activity levels were lower, did lose more FFM (ie, lean mass) than LCD patients. It was therefore surprising that in between-group analyses there was no statistical difference in weight regain. On some level, this raises concerns about the robustness of the overall finding. Perhaps with a larger sample, more precise measures of FFM lost (eg, with DEXA scanning instead of the “bod pod” or longer follow-up, this difference in lost lean mass between groups actually would have predicted greater weight regain for VLCD patients. The researchers attribute some of the FFM loss after the caloric restriction phase to decreased water and glycogen stores, rather than muscle mass, and speculate that this is why no impact on weight regain was seen between groups.

From a generalizability standpoint, there are important safety concerns with the use of VLCDs, aside from subsequent risk of weight regain, that are not addressed with this study. Many patients simply cannot tolerate a 500 kcal per day diet, including those with more severe obesity (who have higher daily energy requirements) or those with complicated chronic medical conditions who might be at higher risk of complications from such low energy intake. Accordingly, these kinds of patients were not included in this study, so it is not clear whether results might generalize to them.

Applications for Clinical Practice

Despite the conventional wisdom that slower weight loss may be more sustainable over time, several recent trials have suggested otherwise. Nonetheless, rapid weight loss produced with the use of VLCDs is not appropriate for every patient and must be carefully overseen by a weight management professional. Furthermore, rapid weight loss may place patients at increased risk of preferentially losing lean mass, which does correlate with risk of weight regain and could set them up for other health problems in the long-term. More studies are needed in this area before a definitive judgment can be made regarding the long term risks and benefits of rapid versus moderate-pace weight loss.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To compare weight regain after rapid versus slower loss of an equivalent amount of weight.

Study design. Randomized clinical trial.

Setting and participants. This study took place in a single medical center in the Netherlands. Investigators recruited 61 adults (no age range provided) with body mass index (BMI) between 28–35 kg/m² and at a stable weight (no change of > 3 kg for the past 2 months) to participate in a weight loss study. Individuals with type 2 diabetes, dyslipidemia, uncontrolled hypertension, or liver, heart or kidney disease were excluded, as were those who were currently pregnant or reported consuming more than moderate amounts of alcohol.

Once consented, participants were randomized into one of 2 study arms. The rapid weight loss arm was prescribed a very-low-calorie diet (VLCD) with just 500 kcal/day (43% protein/43% carb/14% fat) for 5 weeks, after which they transitioned to a 4-week “weight stable” period, and then a 9-month follow-up period (overall follow-up time of ~11 months; 10 months after weight loss). In contrast, the slower weight loss arm was prescribed a low-calorie diet (LCD) with 1250 kcal/day (29% protein/48% carb/23% fat) for 12 weeks, after which they also transitioned to a 4-week weight stable period and 9 months of follow-up (overall follow-up time of ~13 months; 10 months after weight loss). VLCD (rapid weight loss) participants received 3 meal replacement shakes per day (totaling 500 kcal) during the weight loss period and were also told they could consume unlimited amounts of low-calorie vegetables. The LCD (slower weight loss) participants received 1 meal replacement shake per day during their 12 weeks of weight loss and were responsible for providing the remainder of their own meals and snacks according to guidelines from a study dietitian. Following active weight loss, both groups then shifted to higher-calorie, food-based diets during a “weight stable” 4-week period and were responsible during this time for providing all of their own food. The researchers do not specify the details of the diet composition for this weight stable period. Exposure to the registered dietitian was the same in both groups, with 5 consultations during weight loss (weekly for VLCD, presumably more spaced out for LCD) and 4 during weight stable period. No further diet advice or meal replacement support was given during the 9-month follow-up period, but participants came in for monthly weigh-ins.

Main outcome measure. The primary outcome measure was change in weight (ie, amount of weight regained) during the 9-month follow-up period, compared between groups using an independent samples t test. Additional biometric measures included change in waist circumference and changes in body composition. For the latter, the researchers used a “Bod Pod” to conduct air-displacement plethysmography and determine what percentage of an individual’s weight was fat mass (FM) versus lean mass/water (FFM [fat-free mass]). They then compared the amount of FFM lost between groups, again using the independent samples t test.

The researchers also collected information on self-reported physical activity (questionnaire) and self-reported history of weight cycling (number of times a participant had previously lost and regained at least 5 kg) prior to this study. These were not outcomes per-se, but were collected so that they could be examined as correlates of the biometric outcomes above, using Pearson and Spearman’s correlation coefficients.

Results. The LCD (n = 29) and VLCD (n = 28) groups were similar at baseline with no significant differences reported. Of the 61 individuals initially enrolled, 57 (93%) completed the study. Summary statistics are reported only for these 57 individuals. No imputation or other methods for handling missing data were used. There were slightly more women than men in the study (53% women); the average (SD) age was 51.8 (1.9) years in the LCD group and 50.7 (1.5) years in the VLCD group. Mean starting BMI was 31 kg/m² (31.3 [0.5] in LCD, 31.0 [0.4] in VLCD) and both groups had just under 40% body fat at baseline (39.9% [1.8] in LCD, 39.7% [1.5] in VLCD).

After 12 weeks of weight loss for LCD, or 5 weeks of weight loss for VLCD, both groups lost a similar amount of total weight (8.2 [0.5] kg in LCD vs. 9.0 [0.4] kg in VLCD), then had no significant changes in weight during the subsequent 4-week “weight stable” period. However, during the weight stable period VLCD patients had an average 0.8 (0.6) cm increase in waist circumference (a rebounding after a decrease of 7.7 cm during weight loss), while LCD patients on average had a continued decrease of 1.0 (0.5 cm) in waist circumference (P = 0.003).

There was no significant difference between groups for the primary outcome of weight regain during 9-months of follow-up (4.2 [0.6] kg regained for LCD, 4.5 [0.7] for VLCD; P = 0.73). The only significant correlates of weight regain were amount of FFM lost (more lean mass lost predicted more weight regain), and amount of physical activity reported during follow-up (more activity predicted less regain). Participant sex, age, starting BMI, history of weight cycling, and amount of weight lost did not correlate with rate of re-gain.

One area where there was a significant between-group difference, both after initial weight loss and persisting after the weight stable period, was in the amount of FFM lost (a rough approximation of lost lean mass, eg, muscle mass). VLCD participants had more FFM loss (1.6 [0.2] kg) than LCD participants (0.6 [0.2] kg) (P < 0.01) after active weight loss, and continued to have significantly more FFM loss (0.8 [0.2] kg vs. 0.2 [0.2] kg) after the 4-week weight stable period.

There were no between-group differences at the end of weight loss or at the end of follow-up for hip or waist circumference or for blood pressure.

Conclusion. The authors conclude that rate of weight loss does not affect one’s risk of weight regain after a diet, after a similar amount of weight has been lost.

Commentary

The failure of most diets to produce durable weight loss is a frustration for patients, clinicians, and researchers. In general, regardless of the composition of a diet, the majority of patients will regain some or all of their lost weight within several years after completing the diet. The reasons for weight regain are complex, and include reversion to old eating or physical activity behaviors but also a strong physiologic drive by the body to reverse weight loss that it perceives as a threat to health [1].

One area in diet research that has recently generated some controversy is whether or not rate of initial weight loss might impact a patient’s ability to maintain that weight loss, with the conventional wisdom (and national guidelines, in some cases), suggesting that slower weight loss is preferable to rapid weight loss for this reason [2]. A handful of studies have challenged this notion, however, and suggested that rapid weight loss does not necessarily lead to greater weight regain [3,4]. Previous such studies, however, have not generally been designed to compare regain after equal amounts of weight loss, which may make their results more difficult to interpret.

The present study contributes another piece of evidence to the argument that rapid initial weight loss may not increase a patient’s risk of regain. This small randomized trial is timely and has several features that make it a unique contribution. First, the design of the study allowed for both groups, despite losing weight at very different rates, to reach the same amount of total weight loss before being followed forward in time. This made weight regain much easier to compare between groups during follow-up. Second, the study included measurement of changing body composition—ie, what kind of weight was being lost (fat vs. fat-free mass)—rather than just the total amount of weight. This allowed the researchers to present data for an outcome that is mechanistically related to metabolic rate (and therefore weight regain), and one that might have implications for longer-term health after rapid versus more moderate-pace weight loss.

Several aspects of the study design, however, may limit the impact of the findings. For example, in both arms, while a certain type of diet was “prescribed,” there is no comment about assessment of participant fidelity to the prescribed diet, and there is potential for very different levels of adherence between groups, especially in active weight loss, when basically all meals were provided to the VLCD arm, but LCD subjects were responsible for about 90% of their own meals. This could have led to larger discrepancies between prescribed and actual diet in the LCD arm relative to VLCD. Granted, the rate of weight loss was the exposure of interest, and that clearly varied between groups as expected, implying at least moderate fidelity to prescribed caloric content of each diet, but how much protein vs. fat vs. carb was actually consumed by each group is not clear. Additionally, while 9 months of post weight-loss follow-up is certainly a good start in terms of follow-up duration, it may not have been sufficient to observe differences that would later emerge between the groups for weight regain. Other long-term weight loss maintenance studies have followed patients for several years or longer after initial weight loss [5].

Using data from all participants, the researchers reported that the amount of FFM an individual lost was a predictor of weight regain during follow-up. This finding is in keeping with the idea that more lean mass loss leads to lower metabolic rate and predisposes to weight regain (hence the conventional wisdom to avoid rapid weight loss with low-protein diets). In keeping with this theme, VLCD patients, whose protein intakes and activity levels were lower, did lose more FFM (ie, lean mass) than LCD patients. It was therefore surprising that in between-group analyses there was no statistical difference in weight regain. On some level, this raises concerns about the robustness of the overall finding. Perhaps with a larger sample, more precise measures of FFM lost (eg, with DEXA scanning instead of the “bod pod” or longer follow-up, this difference in lost lean mass between groups actually would have predicted greater weight regain for VLCD patients. The researchers attribute some of the FFM loss after the caloric restriction phase to decreased water and glycogen stores, rather than muscle mass, and speculate that this is why no impact on weight regain was seen between groups.

From a generalizability standpoint, there are important safety concerns with the use of VLCDs, aside from subsequent risk of weight regain, that are not addressed with this study. Many patients simply cannot tolerate a 500 kcal per day diet, including those with more severe obesity (who have higher daily energy requirements) or those with complicated chronic medical conditions who might be at higher risk of complications from such low energy intake. Accordingly, these kinds of patients were not included in this study, so it is not clear whether results might generalize to them.

Applications for Clinical Practice

Despite the conventional wisdom that slower weight loss may be more sustainable over time, several recent trials have suggested otherwise. Nonetheless, rapid weight loss produced with the use of VLCDs is not appropriate for every patient and must be carefully overseen by a weight management professional. Furthermore, rapid weight loss may place patients at increased risk of preferentially losing lean mass, which does correlate with risk of weight regain and could set them up for other health problems in the long-term. More studies are needed in this area before a definitive judgment can be made regarding the long term risks and benefits of rapid versus moderate-pace weight loss.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To compare weight regain after rapid versus slower loss of an equivalent amount of weight.

Study design. Randomized clinical trial.

Setting and participants. This study took place in a single medical center in the Netherlands. Investigators recruited 61 adults (no age range provided) with body mass index (BMI) between 28–35 kg/m² and at a stable weight (no change of > 3 kg for the past 2 months) to participate in a weight loss study. Individuals with type 2 diabetes, dyslipidemia, uncontrolled hypertension, or liver, heart or kidney disease were excluded, as were those who were currently pregnant or reported consuming more than moderate amounts of alcohol.

Once consented, participants were randomized into one of 2 study arms. The rapid weight loss arm was prescribed a very-low-calorie diet (VLCD) with just 500 kcal/day (43% protein/43% carb/14% fat) for 5 weeks, after which they transitioned to a 4-week “weight stable” period, and then a 9-month follow-up period (overall follow-up time of ~11 months; 10 months after weight loss). In contrast, the slower weight loss arm was prescribed a low-calorie diet (LCD) with 1250 kcal/day (29% protein/48% carb/23% fat) for 12 weeks, after which they also transitioned to a 4-week weight stable period and 9 months of follow-up (overall follow-up time of ~13 months; 10 months after weight loss). VLCD (rapid weight loss) participants received 3 meal replacement shakes per day (totaling 500 kcal) during the weight loss period and were also told they could consume unlimited amounts of low-calorie vegetables. The LCD (slower weight loss) participants received 1 meal replacement shake per day during their 12 weeks of weight loss and were responsible for providing the remainder of their own meals and snacks according to guidelines from a study dietitian. Following active weight loss, both groups then shifted to higher-calorie, food-based diets during a “weight stable” 4-week period and were responsible during this time for providing all of their own food. The researchers do not specify the details of the diet composition for this weight stable period. Exposure to the registered dietitian was the same in both groups, with 5 consultations during weight loss (weekly for VLCD, presumably more spaced out for LCD) and 4 during weight stable period. No further diet advice or meal replacement support was given during the 9-month follow-up period, but participants came in for monthly weigh-ins.

Main outcome measure. The primary outcome measure was change in weight (ie, amount of weight regained) during the 9-month follow-up period, compared between groups using an independent samples t test. Additional biometric measures included change in waist circumference and changes in body composition. For the latter, the researchers used a “Bod Pod” to conduct air-displacement plethysmography and determine what percentage of an individual’s weight was fat mass (FM) versus lean mass/water (FFM [fat-free mass]). They then compared the amount of FFM lost between groups, again using the independent samples t test.

The researchers also collected information on self-reported physical activity (questionnaire) and self-reported history of weight cycling (number of times a participant had previously lost and regained at least 5 kg) prior to this study. These were not outcomes per-se, but were collected so that they could be examined as correlates of the biometric outcomes above, using Pearson and Spearman’s correlation coefficients.

Results. The LCD (n = 29) and VLCD (n = 28) groups were similar at baseline with no significant differences reported. Of the 61 individuals initially enrolled, 57 (93%) completed the study. Summary statistics are reported only for these 57 individuals. No imputation or other methods for handling missing data were used. There were slightly more women than men in the study (53% women); the average (SD) age was 51.8 (1.9) years in the LCD group and 50.7 (1.5) years in the VLCD group. Mean starting BMI was 31 kg/m² (31.3 [0.5] in LCD, 31.0 [0.4] in VLCD) and both groups had just under 40% body fat at baseline (39.9% [1.8] in LCD, 39.7% [1.5] in VLCD).

After 12 weeks of weight loss for LCD, or 5 weeks of weight loss for VLCD, both groups lost a similar amount of total weight (8.2 [0.5] kg in LCD vs. 9.0 [0.4] kg in VLCD), then had no significant changes in weight during the subsequent 4-week “weight stable” period. However, during the weight stable period VLCD patients had an average 0.8 (0.6) cm increase in waist circumference (a rebounding after a decrease of 7.7 cm during weight loss), while LCD patients on average had a continued decrease of 1.0 (0.5 cm) in waist circumference (P = 0.003).

There was no significant difference between groups for the primary outcome of weight regain during 9-months of follow-up (4.2 [0.6] kg regained for LCD, 4.5 [0.7] for VLCD; P = 0.73). The only significant correlates of weight regain were amount of FFM lost (more lean mass lost predicted more weight regain), and amount of physical activity reported during follow-up (more activity predicted less regain). Participant sex, age, starting BMI, history of weight cycling, and amount of weight lost did not correlate with rate of re-gain.

One area where there was a significant between-group difference, both after initial weight loss and persisting after the weight stable period, was in the amount of FFM lost (a rough approximation of lost lean mass, eg, muscle mass). VLCD participants had more FFM loss (1.6 [0.2] kg) than LCD participants (0.6 [0.2] kg) (P < 0.01) after active weight loss, and continued to have significantly more FFM loss (0.8 [0.2] kg vs. 0.2 [0.2] kg) after the 4-week weight stable period.

There were no between-group differences at the end of weight loss or at the end of follow-up for hip or waist circumference or for blood pressure.

Conclusion. The authors conclude that rate of weight loss does not affect one’s risk of weight regain after a diet, after a similar amount of weight has been lost.

Commentary

The failure of most diets to produce durable weight loss is a frustration for patients, clinicians, and researchers. In general, regardless of the composition of a diet, the majority of patients will regain some or all of their lost weight within several years after completing the diet. The reasons for weight regain are complex, and include reversion to old eating or physical activity behaviors but also a strong physiologic drive by the body to reverse weight loss that it perceives as a threat to health [1].

One area in diet research that has recently generated some controversy is whether or not rate of initial weight loss might impact a patient’s ability to maintain that weight loss, with the conventional wisdom (and national guidelines, in some cases), suggesting that slower weight loss is preferable to rapid weight loss for this reason [2]. A handful of studies have challenged this notion, however, and suggested that rapid weight loss does not necessarily lead to greater weight regain [3,4]. Previous such studies, however, have not generally been designed to compare regain after equal amounts of weight loss, which may make their results more difficult to interpret.

The present study contributes another piece of evidence to the argument that rapid initial weight loss may not increase a patient’s risk of regain. This small randomized trial is timely and has several features that make it a unique contribution. First, the design of the study allowed for both groups, despite losing weight at very different rates, to reach the same amount of total weight loss before being followed forward in time. This made weight regain much easier to compare between groups during follow-up. Second, the study included measurement of changing body composition—ie, what kind of weight was being lost (fat vs. fat-free mass)—rather than just the total amount of weight. This allowed the researchers to present data for an outcome that is mechanistically related to metabolic rate (and therefore weight regain), and one that might have implications for longer-term health after rapid versus more moderate-pace weight loss.

Several aspects of the study design, however, may limit the impact of the findings. For example, in both arms, while a certain type of diet was “prescribed,” there is no comment about assessment of participant fidelity to the prescribed diet, and there is potential for very different levels of adherence between groups, especially in active weight loss, when basically all meals were provided to the VLCD arm, but LCD subjects were responsible for about 90% of their own meals. This could have led to larger discrepancies between prescribed and actual diet in the LCD arm relative to VLCD. Granted, the rate of weight loss was the exposure of interest, and that clearly varied between groups as expected, implying at least moderate fidelity to prescribed caloric content of each diet, but how much protein vs. fat vs. carb was actually consumed by each group is not clear. Additionally, while 9 months of post weight-loss follow-up is certainly a good start in terms of follow-up duration, it may not have been sufficient to observe differences that would later emerge between the groups for weight regain. Other long-term weight loss maintenance studies have followed patients for several years or longer after initial weight loss [5].

Using data from all participants, the researchers reported that the amount of FFM an individual lost was a predictor of weight regain during follow-up. This finding is in keeping with the idea that more lean mass loss leads to lower metabolic rate and predisposes to weight regain (hence the conventional wisdom to avoid rapid weight loss with low-protein diets). In keeping with this theme, VLCD patients, whose protein intakes and activity levels were lower, did lose more FFM (ie, lean mass) than LCD patients. It was therefore surprising that in between-group analyses there was no statistical difference in weight regain. On some level, this raises concerns about the robustness of the overall finding. Perhaps with a larger sample, more precise measures of FFM lost (eg, with DEXA scanning instead of the “bod pod” or longer follow-up, this difference in lost lean mass between groups actually would have predicted greater weight regain for VLCD patients. The researchers attribute some of the FFM loss after the caloric restriction phase to decreased water and glycogen stores, rather than muscle mass, and speculate that this is why no impact on weight regain was seen between groups.

From a generalizability standpoint, there are important safety concerns with the use of VLCDs, aside from subsequent risk of weight regain, that are not addressed with this study. Many patients simply cannot tolerate a 500 kcal per day diet, including those with more severe obesity (who have higher daily energy requirements) or those with complicated chronic medical conditions who might be at higher risk of complications from such low energy intake. Accordingly, these kinds of patients were not included in this study, so it is not clear whether results might generalize to them.

Applications for Clinical Practice

Despite the conventional wisdom that slower weight loss may be more sustainable over time, several recent trials have suggested otherwise. Nonetheless, rapid weight loss produced with the use of VLCDs is not appropriate for every patient and must be carefully overseen by a weight management professional. Furthermore, rapid weight loss may place patients at increased risk of preferentially losing lean mass, which does correlate with risk of weight regain and could set them up for other health problems in the long-term. More studies are needed in this area before a definitive judgment can be made regarding the long term risks and benefits of rapid versus moderate-pace weight loss.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine whether in-person visits for primary care patients resulted in improved weight loss maintenance relative to monthly mailings, with both groups receiving access to portion-controlled meals.

Design. Randomized clinical trial.

Setting and participants. This study took place within 2 university-affiliated primary care clinics in Colorado. For the first phase of the study, investigators enrolled 104 obese adult patients (18–79 years; BMI 30–49.9 kg/m²) who had been diagnosed with at least one of the following: type 2 diabetes, sleep apnea, hypertension, or hyperlipidemia. Patients who had independently lost weight prior to trial entry (> 5% in 6 months), were on weight-gain–promoting medications such as steroids, or had previously undergone bariatric surgery were excluded. The trial started with a 6-month run-in phase where active weight loss was promoted using a high-intensity behavioral intervention based on the Diabetes Prevention Program as well as access to subsidized portion-controlled foods (Nutrisystem). At the end of the 6-month run-in, the remaining participants (n = 84, 79.3%) were then randomized, stratified by gender and whether or not they achieved 5% weight loss, into the 2 main study arms.

Intervention. The experimental study arm (n = 41, “intensified maintenance”) relied on monthly in-person visits and monthly phone calls to prevent weight regain (thus, these participants had twice monthly contact during maintenance). Both visit types in this arm were conducted by a graduate-level research assistant and included some structured educational content as well as problem-solving around diet and lifestyle issues. In contrast, the control arm (n = 43, “standard maintenance”) relied just on monthly mailings (or emails) of educational and support materials to promote weight loss maintenance. Participants in both groups had the opportunity to purchase subsidized portion-controlled foods/meals from Nutrisystem in order to facilitate continued adherence to the caloric restriction required for weight loss maintenance.

Main outcome measures. The primary outcome for this trial was change in weight, measured in kgs, during the 12-month maintenance period. Other biometric outcomes included changes in blood pressure, serum glucose, lipid levels, and the inflammatory marker hs-CRP. Patient-reported outcomes included changes in medication use. The investigators used intention-to-treat analysis, with mixed linear models adjusted for age and gender. No imputation techniques for missing data are reported, although complete follow-up data was obtained on 94% of patients.

Results. Participants in the standard and intensified weight maintenance arms of the trial were similar with respect to measured baseline characteristics. The average age of participants was 56 years, and three-quarters (75%) were female. The majority in both groups were white (77% in standard arm; 88% in intense), and over half had either a college or advanced degree (58.1% in standard arm, 51.2% in intense). Approximately one- third had diabetes (32.6% in standard arm, 34.1% in intense) and over half had hypertension (67.4% in standard arm, 63.4% in intense). Of the 84 participants who were randomized in the weight maintenance phase of the study, 79 completed the 12-month follow up (94%; no difference in attrition between groups).

After 12 months of maintenance, participants in the intensified maintenance arm regained just 1.6 (± 1.3) kg of lost weight, while those in the standard arm regained 5.0 (± 0.8) kg, a statistically significant difference (P = 0.01). The investigators also examined the subgroup of participants who, after the 6-month run-in, had lost at least 5% of their initial body weight. For these individuals, almost three-quarters in the intensified maintenance arm (71.9%) maintained that > 5% loss by 18 months, compared to 51.7% in the standard group. This difference between groups was not statistically significant. There was a significant difference between groups for change in hs-CRP over the 12-month maintenance period, with the intensive group’s hs-CRP ending up an average of 1.46 mg/L lower than that of the standard group (P = 0.03). Although there was a similar trend favoring the intensive intervention for other biometric measures (change in waist circumference, glucose, blood pressure, and lipids were all more favorable in this arm), the between-group differences for these measures did not reach statistical significance. No significant differences between groups were observed with respect to changes in medication use over the 12-month maintenance intervention.

Conclusion. After 5 months of active weight loss, twice-monthly contact (using one in-person and one phone visit) plus portion-controlled foods during a 12-month weight maintenance phase resulted in significantly less weight regain than monthly mail or email-based counseling plus portion-controlled foods.

Commentary

Behavioral weight loss interventions, which typically require high-intensity in-person counseling over several months to a year, may be difficult to accomplish in the average primary care practice [1]. On the other hand, it may be the case that primary care practices are well-suited to assist patients who have already lost weight, as they enter weight-loss maintenance. While numerous studies have shown that patients who adhere to calorie-restricted diets (almost regardless of diet composition) are able to achieve clinically significant weight loss, less is known about effective methods of preventing weight regain. Several large trials have suggested that, as is the case with behavioral weight loss interventions, maintenance interventions are also more successful if they include regular contact, at least some of which is face-to-face [2,3]. These visits, along with other practices such as self-weighing and food diaries, may help patients maintain the energy balance necessary to stay at their new, lower body weight. There remains a gap, however, in terms of knowing whether the maintenance interventions from large randomized trials can be translated into the sometimes messy real world of clinical practice, where clinicians and patients are typically overburdened and busy.

The current study by Tsai et al does address some aspects of this important question. By recruiting “real-world” chronically ill patients from a primary care practice to participate in the trial, the results of this study may be more likely to generalize to the patient populations seen by practicing clinicians than the typically healthier, younger, community-recruited volunteers in large trials. Additionally, although the interventions in this study were not delivered by the primary care practice per se, they were low enough in intensity that they could theoretically be translated into most clinical practice settings, assuming reimbursement is not an issue. Monthly in-person visits certainly could be done by a physician (as under current CMS reimbursement guidelines), but would not have to be (the visits in this study were done by a graduate student with no formal training in behavioral interventions), and telephone visits could easily be done by clinical support staff. Even with this low level of visit intensity, patients had significantly less weight regain than those who were receiving monthly email or postal mail support (which, realistically, would still require some work on the part of primary care practices). Furthermore, there were suggestions of numerous parallel cardiometabolic benefits that might have been statistically significant with a larger sample size. This study benefited from several strengths in addition to its highly practical point of view. It was a randomized trial with a strong control group and long follow-up duration (18 months total). It used a run-in period for weight loss so that all who entered maintenance were doing so based on exposure to the same weight loss intervention. Happily, though, the investigators did not require successful weight loss (> 5%) for entry into the maintenance phase, which likely further contributed to the generalizability of their results. Another area where the run-in likely helped was with retention of subjects—94% of those randomized for maintenance contributed complete data at the end of the 12-month study period.

As acknowledged by the authors, this study also has some important limitations. As with most weight loss/diet interventions, the participants in this study were mostly female, and mostly non-Hispanic white, and thus a substantially less diverse population than is represented by patients with obesity in the US. Furthermore, although some aspects of the patient population did promote generalizability (recruitment from primary care, chronic illness burden), these patients were fairly highly educated, which may have impacted their adherence and results.

The use of subsidized portion-controlled meals in this study, while evidence-based, may have clouded the results somewhat. Perhaps the effect of both interventions would have been less pronounced had patients not been provided with subsidies to access these foods. In their discussion, the investigators acknowledge that the study lacked a comparison group with no access to portion-controlled foods and that, in a post-hoc analysis, greater use of these foods corresponded with better weight loss and weight loss maintenance among all participants.

Finally, although it was beyond the scope of this study, this trial does not provide any information about how weight loss medications in either the weight loss or maintenance phases might impact these types of interventions. Now that the FDA has approved a number of such medications for long-term use, it would be very helpful to have more information about how medications might be integrated into these types of strategies, for interested patients, as physicians could clearly play an integral role in the pharmacologic management of weight, alongside effective behavioral interventions.

Applications for Clinical Practice

Low-to-moderate intensity in-person and telephone-based visits during weight maintenance may help to protect against weight regain, and could realistically be an option for many primary care practices and their patients. However, aside from Medicare patients, for whom monthly primary care–based weight maintenance visits are now covered, physicians would need to understand how to code and bill such visits appropriately in order to avoid having patients face unexpected charges.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine whether in-person visits for primary care patients resulted in improved weight loss maintenance relative to monthly mailings, with both groups receiving access to portion-controlled meals.

Design. Randomized clinical trial.

Setting and participants. This study took place within 2 university-affiliated primary care clinics in Colorado. For the first phase of the study, investigators enrolled 104 obese adult patients (18–79 years; BMI 30–49.9 kg/m²) who had been diagnosed with at least one of the following: type 2 diabetes, sleep apnea, hypertension, or hyperlipidemia. Patients who had independently lost weight prior to trial entry (> 5% in 6 months), were on weight-gain–promoting medications such as steroids, or had previously undergone bariatric surgery were excluded. The trial started with a 6-month run-in phase where active weight loss was promoted using a high-intensity behavioral intervention based on the Diabetes Prevention Program as well as access to subsidized portion-controlled foods (Nutrisystem). At the end of the 6-month run-in, the remaining participants (n = 84, 79.3%) were then randomized, stratified by gender and whether or not they achieved 5% weight loss, into the 2 main study arms.

Intervention. The experimental study arm (n = 41, “intensified maintenance”) relied on monthly in-person visits and monthly phone calls to prevent weight regain (thus, these participants had twice monthly contact during maintenance). Both visit types in this arm were conducted by a graduate-level research assistant and included some structured educational content as well as problem-solving around diet and lifestyle issues. In contrast, the control arm (n = 43, “standard maintenance”) relied just on monthly mailings (or emails) of educational and support materials to promote weight loss maintenance. Participants in both groups had the opportunity to purchase subsidized portion-controlled foods/meals from Nutrisystem in order to facilitate continued adherence to the caloric restriction required for weight loss maintenance.

Main outcome measures. The primary outcome for this trial was change in weight, measured in kgs, during the 12-month maintenance period. Other biometric outcomes included changes in blood pressure, serum glucose, lipid levels, and the inflammatory marker hs-CRP. Patient-reported outcomes included changes in medication use. The investigators used intention-to-treat analysis, with mixed linear models adjusted for age and gender. No imputation techniques for missing data are reported, although complete follow-up data was obtained on 94% of patients.

Results. Participants in the standard and intensified weight maintenance arms of the trial were similar with respect to measured baseline characteristics. The average age of participants was 56 years, and three-quarters (75%) were female. The majority in both groups were white (77% in standard arm; 88% in intense), and over half had either a college or advanced degree (58.1% in standard arm, 51.2% in intense). Approximately one- third had diabetes (32.6% in standard arm, 34.1% in intense) and over half had hypertension (67.4% in standard arm, 63.4% in intense). Of the 84 participants who were randomized in the weight maintenance phase of the study, 79 completed the 12-month follow up (94%; no difference in attrition between groups).

After 12 months of maintenance, participants in the intensified maintenance arm regained just 1.6 (± 1.3) kg of lost weight, while those in the standard arm regained 5.0 (± 0.8) kg, a statistically significant difference (P = 0.01). The investigators also examined the subgroup of participants who, after the 6-month run-in, had lost at least 5% of their initial body weight. For these individuals, almost three-quarters in the intensified maintenance arm (71.9%) maintained that > 5% loss by 18 months, compared to 51.7% in the standard group. This difference between groups was not statistically significant. There was a significant difference between groups for change in hs-CRP over the 12-month maintenance period, with the intensive group’s hs-CRP ending up an average of 1.46 mg/L lower than that of the standard group (P = 0.03). Although there was a similar trend favoring the intensive intervention for other biometric measures (change in waist circumference, glucose, blood pressure, and lipids were all more favorable in this arm), the between-group differences for these measures did not reach statistical significance. No significant differences between groups were observed with respect to changes in medication use over the 12-month maintenance intervention.

Conclusion. After 5 months of active weight loss, twice-monthly contact (using one in-person and one phone visit) plus portion-controlled foods during a 12-month weight maintenance phase resulted in significantly less weight regain than monthly mail or email-based counseling plus portion-controlled foods.

Commentary

Behavioral weight loss interventions, which typically require high-intensity in-person counseling over several months to a year, may be difficult to accomplish in the average primary care practice [1]. On the other hand, it may be the case that primary care practices are well-suited to assist patients who have already lost weight, as they enter weight-loss maintenance. While numerous studies have shown that patients who adhere to calorie-restricted diets (almost regardless of diet composition) are able to achieve clinically significant weight loss, less is known about effective methods of preventing weight regain. Several large trials have suggested that, as is the case with behavioral weight loss interventions, maintenance interventions are also more successful if they include regular contact, at least some of which is face-to-face [2,3]. These visits, along with other practices such as self-weighing and food diaries, may help patients maintain the energy balance necessary to stay at their new, lower body weight. There remains a gap, however, in terms of knowing whether the maintenance interventions from large randomized trials can be translated into the sometimes messy real world of clinical practice, where clinicians and patients are typically overburdened and busy.

The current study by Tsai et al does address some aspects of this important question. By recruiting “real-world” chronically ill patients from a primary care practice to participate in the trial, the results of this study may be more likely to generalize to the patient populations seen by practicing clinicians than the typically healthier, younger, community-recruited volunteers in large trials. Additionally, although the interventions in this study were not delivered by the primary care practice per se, they were low enough in intensity that they could theoretically be translated into most clinical practice settings, assuming reimbursement is not an issue. Monthly in-person visits certainly could be done by a physician (as under current CMS reimbursement guidelines), but would not have to be (the visits in this study were done by a graduate student with no formal training in behavioral interventions), and telephone visits could easily be done by clinical support staff. Even with this low level of visit intensity, patients had significantly less weight regain than those who were receiving monthly email or postal mail support (which, realistically, would still require some work on the part of primary care practices). Furthermore, there were suggestions of numerous parallel cardiometabolic benefits that might have been statistically significant with a larger sample size. This study benefited from several strengths in addition to its highly practical point of view. It was a randomized trial with a strong control group and long follow-up duration (18 months total). It used a run-in period for weight loss so that all who entered maintenance were doing so based on exposure to the same weight loss intervention. Happily, though, the investigators did not require successful weight loss (> 5%) for entry into the maintenance phase, which likely further contributed to the generalizability of their results. Another area where the run-in likely helped was with retention of subjects—94% of those randomized for maintenance contributed complete data at the end of the 12-month study period.

As acknowledged by the authors, this study also has some important limitations. As with most weight loss/diet interventions, the participants in this study were mostly female, and mostly non-Hispanic white, and thus a substantially less diverse population than is represented by patients with obesity in the US. Furthermore, although some aspects of the patient population did promote generalizability (recruitment from primary care, chronic illness burden), these patients were fairly highly educated, which may have impacted their adherence and results.

The use of subsidized portion-controlled meals in this study, while evidence-based, may have clouded the results somewhat. Perhaps the effect of both interventions would have been less pronounced had patients not been provided with subsidies to access these foods. In their discussion, the investigators acknowledge that the study lacked a comparison group with no access to portion-controlled foods and that, in a post-hoc analysis, greater use of these foods corresponded with better weight loss and weight loss maintenance among all participants.

Finally, although it was beyond the scope of this study, this trial does not provide any information about how weight loss medications in either the weight loss or maintenance phases might impact these types of interventions. Now that the FDA has approved a number of such medications for long-term use, it would be very helpful to have more information about how medications might be integrated into these types of strategies, for interested patients, as physicians could clearly play an integral role in the pharmacologic management of weight, alongside effective behavioral interventions.

Applications for Clinical Practice

Low-to-moderate intensity in-person and telephone-based visits during weight maintenance may help to protect against weight regain, and could realistically be an option for many primary care practices and their patients. However, aside from Medicare patients, for whom monthly primary care–based weight maintenance visits are now covered, physicians would need to understand how to code and bill such visits appropriately in order to avoid having patients face unexpected charges.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine whether in-person visits for primary care patients resulted in improved weight loss maintenance relative to monthly mailings, with both groups receiving access to portion-controlled meals.

Design. Randomized clinical trial.

Setting and participants. This study took place within 2 university-affiliated primary care clinics in Colorado. For the first phase of the study, investigators enrolled 104 obese adult patients (18–79 years; BMI 30–49.9 kg/m²) who had been diagnosed with at least one of the following: type 2 diabetes, sleep apnea, hypertension, or hyperlipidemia. Patients who had independently lost weight prior to trial entry (> 5% in 6 months), were on weight-gain–promoting medications such as steroids, or had previously undergone bariatric surgery were excluded. The trial started with a 6-month run-in phase where active weight loss was promoted using a high-intensity behavioral intervention based on the Diabetes Prevention Program as well as access to subsidized portion-controlled foods (Nutrisystem). At the end of the 6-month run-in, the remaining participants (n = 84, 79.3%) were then randomized, stratified by gender and whether or not they achieved 5% weight loss, into the 2 main study arms.

Intervention. The experimental study arm (n = 41, “intensified maintenance”) relied on monthly in-person visits and monthly phone calls to prevent weight regain (thus, these participants had twice monthly contact during maintenance). Both visit types in this arm were conducted by a graduate-level research assistant and included some structured educational content as well as problem-solving around diet and lifestyle issues. In contrast, the control arm (n = 43, “standard maintenance”) relied just on monthly mailings (or emails) of educational and support materials to promote weight loss maintenance. Participants in both groups had the opportunity to purchase subsidized portion-controlled foods/meals from Nutrisystem in order to facilitate continued adherence to the caloric restriction required for weight loss maintenance.

Main outcome measures. The primary outcome for this trial was change in weight, measured in kgs, during the 12-month maintenance period. Other biometric outcomes included changes in blood pressure, serum glucose, lipid levels, and the inflammatory marker hs-CRP. Patient-reported outcomes included changes in medication use. The investigators used intention-to-treat analysis, with mixed linear models adjusted for age and gender. No imputation techniques for missing data are reported, although complete follow-up data was obtained on 94% of patients.

Results. Participants in the standard and intensified weight maintenance arms of the trial were similar with respect to measured baseline characteristics. The average age of participants was 56 years, and three-quarters (75%) were female. The majority in both groups were white (77% in standard arm; 88% in intense), and over half had either a college or advanced degree (58.1% in standard arm, 51.2% in intense). Approximately one- third had diabetes (32.6% in standard arm, 34.1% in intense) and over half had hypertension (67.4% in standard arm, 63.4% in intense). Of the 84 participants who were randomized in the weight maintenance phase of the study, 79 completed the 12-month follow up (94%; no difference in attrition between groups).

After 12 months of maintenance, participants in the intensified maintenance arm regained just 1.6 (± 1.3) kg of lost weight, while those in the standard arm regained 5.0 (± 0.8) kg, a statistically significant difference (P = 0.01). The investigators also examined the subgroup of participants who, after the 6-month run-in, had lost at least 5% of their initial body weight. For these individuals, almost three-quarters in the intensified maintenance arm (71.9%) maintained that > 5% loss by 18 months, compared to 51.7% in the standard group. This difference between groups was not statistically significant. There was a significant difference between groups for change in hs-CRP over the 12-month maintenance period, with the intensive group’s hs-CRP ending up an average of 1.46 mg/L lower than that of the standard group (P = 0.03). Although there was a similar trend favoring the intensive intervention for other biometric measures (change in waist circumference, glucose, blood pressure, and lipids were all more favorable in this arm), the between-group differences for these measures did not reach statistical significance. No significant differences between groups were observed with respect to changes in medication use over the 12-month maintenance intervention.

Conclusion. After 5 months of active weight loss, twice-monthly contact (using one in-person and one phone visit) plus portion-controlled foods during a 12-month weight maintenance phase resulted in significantly less weight regain than monthly mail or email-based counseling plus portion-controlled foods.

Commentary

Behavioral weight loss interventions, which typically require high-intensity in-person counseling over several months to a year, may be difficult to accomplish in the average primary care practice [1]. On the other hand, it may be the case that primary care practices are well-suited to assist patients who have already lost weight, as they enter weight-loss maintenance. While numerous studies have shown that patients who adhere to calorie-restricted diets (almost regardless of diet composition) are able to achieve clinically significant weight loss, less is known about effective methods of preventing weight regain. Several large trials have suggested that, as is the case with behavioral weight loss interventions, maintenance interventions are also more successful if they include regular contact, at least some of which is face-to-face [2,3]. These visits, along with other practices such as self-weighing and food diaries, may help patients maintain the energy balance necessary to stay at their new, lower body weight. There remains a gap, however, in terms of knowing whether the maintenance interventions from large randomized trials can be translated into the sometimes messy real world of clinical practice, where clinicians and patients are typically overburdened and busy.

The current study by Tsai et al does address some aspects of this important question. By recruiting “real-world” chronically ill patients from a primary care practice to participate in the trial, the results of this study may be more likely to generalize to the patient populations seen by practicing clinicians than the typically healthier, younger, community-recruited volunteers in large trials. Additionally, although the interventions in this study were not delivered by the primary care practice per se, they were low enough in intensity that they could theoretically be translated into most clinical practice settings, assuming reimbursement is not an issue. Monthly in-person visits certainly could be done by a physician (as under current CMS reimbursement guidelines), but would not have to be (the visits in this study were done by a graduate student with no formal training in behavioral interventions), and telephone visits could easily be done by clinical support staff. Even with this low level of visit intensity, patients had significantly less weight regain than those who were receiving monthly email or postal mail support (which, realistically, would still require some work on the part of primary care practices). Furthermore, there were suggestions of numerous parallel cardiometabolic benefits that might have been statistically significant with a larger sample size. This study benefited from several strengths in addition to its highly practical point of view. It was a randomized trial with a strong control group and long follow-up duration (18 months total). It used a run-in period for weight loss so that all who entered maintenance were doing so based on exposure to the same weight loss intervention. Happily, though, the investigators did not require successful weight loss (> 5%) for entry into the maintenance phase, which likely further contributed to the generalizability of their results. Another area where the run-in likely helped was with retention of subjects—94% of those randomized for maintenance contributed complete data at the end of the 12-month study period.

As acknowledged by the authors, this study also has some important limitations. As with most weight loss/diet interventions, the participants in this study were mostly female, and mostly non-Hispanic white, and thus a substantially less diverse population than is represented by patients with obesity in the US. Furthermore, although some aspects of the patient population did promote generalizability (recruitment from primary care, chronic illness burden), these patients were fairly highly educated, which may have impacted their adherence and results.

The use of subsidized portion-controlled meals in this study, while evidence-based, may have clouded the results somewhat. Perhaps the effect of both interventions would have been less pronounced had patients not been provided with subsidies to access these foods. In their discussion, the investigators acknowledge that the study lacked a comparison group with no access to portion-controlled foods and that, in a post-hoc analysis, greater use of these foods corresponded with better weight loss and weight loss maintenance among all participants.

Finally, although it was beyond the scope of this study, this trial does not provide any information about how weight loss medications in either the weight loss or maintenance phases might impact these types of interventions. Now that the FDA has approved a number of such medications for long-term use, it would be very helpful to have more information about how medications might be integrated into these types of strategies, for interested patients, as physicians could clearly play an integral role in the pharmacologic management of weight, alongside effective behavioral interventions.

Applications for Clinical Practice

Low-to-moderate intensity in-person and telephone-based visits during weight maintenance may help to protect against weight regain, and could realistically be an option for many primary care practices and their patients. However, aside from Medicare patients, for whom monthly primary care–based weight maintenance visits are now covered, physicians would need to understand how to code and bill such visits appropriately in order to avoid having patients face unexpected charges.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To evaluate the efficacy of liraglutide for weight loss in a group of nondiabetic patients with obesity.

Design. Randomized double-blind placebo-controlled trial.

Setting and participants. This trial took place across 27 countries in Europe, North America, South America, Asia, Africa and Australia. It was funded by NovoNordisk, the pharmaceutical company that manufactures liraglutide. Participants were 18 years or older, with a BMI of 30 kg/m² (or 27 kg/m² with hypertension or dyslipidemia). Patients with diabetes, those on medications known to induce weight gain (or loss), those with history of bariatric surgery, and those with psychiatric illness were excluded from participating. Patients with prediabetes were not excluded.

Intervention. Participants were randomized (2:1 in favor of study drug) to liraglutide or placebo, stratified according to BMI category and pre-diabetes status. They were started at a 0.6–mg dose of medication and up-titrated as tolerated to a dose of 3.0 mg over several weeks. All received counseling on behavioral changes to promote weight loss. Participants were then followed for 56 weeks. A small subgroup in the liraglutide arm was randomly assigned to switch to placebo after 12 weeks on medication to examine for durability of effect of medication, and to evaluate for safety issues that might occur on drug discontinuation.

Main outcome measures. This study focused on 3 primary outcomes: individual-level weight change from baseline, group-level percentage of participants achieving at least 5% weight loss, and percentage of participants with at least 10% weight loss, all assessed at 56 weeks.

Secondary outcomes included change in BMI, waist circumference, markers of glycemia (hemoglobin A1c, insulin level), markers of cardiometabolic health (blood pressure, lipids, CRP), and health-related quality of life (using several validated survey measures). Adverse events were also assessed.

The investigators used an intention-to-treat analysis, comparing outcomes among all patients who were randomized and received at least 1 dose of liraglutide or placebo. For patients with missing values (eg, due to dropout), outcome values were imputed using the last-observation-carried-forward method. A multivariable analysis of covariance model was used to analyze changes in the primary outcomes and included a covariate for the baseline measure of the outcome in question. Sensitivity analyses were conducted in which the investigators used different imputation techniques (multiple imputation, repeated measures) to account for missing data.

Results. The trial enrolled 3731 participants, 2487 of whom were randomized to receive liraglutide and 1244 of whom received placebo. The groups were similar on measured baseline characteristics, with a mean age of 45 years, mostly female participants (78.7% in liraglutide arm, 78.1% in placebo), and the vast majority of participants identified as “white” race/ethnicity (84.7% in liraglutide, 85.3% in placebo). Mean baseline BMI was 38.3 kg/m² in both groups. Although overweight patients with BMI 27 kg/m² or greater were included, they represented a small fraction of all participants (2.7% in liraglutide group and 3.5% in placebo group). Furthermore, although patients with overt diabetes were excluded from participating, over half of the participants qualified as having prediabetes (61.4% in liraglutide group, 60.9% in placebo group). Just over one-third (34.2% of liraglutide group, 35.9% placebo) had hypertension diagnosed at baseline. Study withdrawal was relatively substantial in both groups – 71.9% remained enrolled at 56 weeks in the liraglutide group, and 64.4% remained in the placebo arm. The investigators note that withdrawal due to adverse events was more common in the liraglutide group (9.9% of withdrawals vs. 3.8% in placebo), while other reasons for withdrawing (ineffective therapy, withdrawal of consent) were more common among placebo participants.

Liraglutide participants lost significantly more weight than placebo participants at 56 weeks (mean [SD]  8.0 [6.7] kg vs. 2.6 [5.7] kg). Similarly, more patients in the liraglutide group achieved at least 5% weight loss (63% vs. 27%), and 10% weight loss (33.1% vs. 10.6%) than those taking placebo. When subgroups of patients were examined according to baseline BMI, the investigators suggested that liraglutide appeared to be more effective at promoting weight loss among patients starting below 40 kg/m².

Hemoglobin A1c dropped significantly more (–0.23 points, P < 0.001) among liraglutide participants than among placebo participants. Similarly, fasting insulin levels dropped by 8% more (P < 0.001) in the liraglutide group at 56 weeks. In keeping with the greater weight loss, markers of cardiometabolic health also improved to a greater extent among liraglutide participants, with larger decreases in blood pressure (SBP –2.8 mm Hg lower in liraglutide, P < 0.001), and LDL (–2.4% difference, P = 0.002), and a larger increase in HDL (1.9% difference, P = 0.001). By week 56, 14% of prediabetic patients in the placebo arm had received a new diagnosis of diabetes, compared to just 4% in the liraglutide group (P < 0.001).

Quality of life scores were higher for liraglutide participants on all included measures except those related to side effects of treatment, where placebo participants reported lower levels of side effects. The most common side effects reported by liraglutide participants related to GI upset, including nausea (40%), diarrhea (21%), and vomiting (16%). More serious events, including cholelithiasis (0.8%), cholecystitis (0.5%), and pancreatitis (0.2%), were also reported. Somewhat surprisingly, although liraglutide is also used to improve glycemic control in diabetics, rates of reported spontaneous hypoglycemia were fairly low in the liraglutide group (1.3% vs. 1.0% in placebo).

Conclusion. Liraglutide given at a dose of 3.0 mg daily, along with lifestyle advice, produces clinically significant weight loss and improvement in glycemic and cardiometabolic parameters that is sustained after 1 full year of treatment.

Commentary

Over the past few years, the FDA has approved a growing list of medications for the treatment of obesity [1,2]. Unlike the prior mainstay for prescription weight management, phentermine, which can only be used for a few months at a time due to concerns about abuse, many of these newer medications are approved for long-term use, aligning well with the growing recognition of obesity as a chronic illness. Interestingly, most of the drugs that have emerged onto the market do not represent novel compounds, but rather are existing drugs that have been repurposed and repackaged for the indication of weight management. These “recycled” medications include Qsymia (a mix of phentermine and topiramate) [1], Contrave (naltrexone and buproprion) [2], and now, Saxenda (liraglutide, also marketed as Victoza for treatment of type 2 diabetes). Liraglutide is a glucagon-like-peptide 1 (GLP-1) analogue, meaning it has an effect similar to that of GLP-1, a gut hormone that stimulates insulin secretion, inhibits pancreatic beta cell apoptosis, inhibits gastric emptying, and decreases appetite by acting on the brain’s satiety centers [3]. For several years, endocrinologists and some internists have been using liraglutide (Victoza) to help with glycemic control in diabetics, with the known benefit that, unlike some other diabetes medications, it tends to promote modest weight loss [4].

In this large multicenter trial, Pi-Sunyer et al evaluated the efficacy of liraglutide at a 3.0 mg daily dose (almost twice the dose used for diabetes) for weight management. The trial utilized a strong study design, with double blinding, randomization of a subgroup for early discontinuation (to evaluate for weight regain and stopping-related side effects), and, importantly, the intervention for both groups also included a behavior change component (albeit one of relatively low intensity, based on the limited description). Patients were followed for 56 weeks on the medication, making the “intervention” phase of the study longer than what has been done in many diet trials. Testing for a long-lasting impact on weight, and at the same time attempting to quantify risks associated with longer-term use of a medication, was an important contribution for this study given that liraglutide is being marketed for long-term use.

After a year on liraglutide, participants in that group had lost around 12 lb more, on average, than those using placebo, and had achieved greater improvements cardiometabolic risk markers, with a much lower risk of developing diabetes. While these findings are promising from a clinical standpoint, it is not clear whether the moderate health impacts of this drug will be sufficient to outweigh several issues that may impede its widespread use in practice. The rate of GI side effects (nausea, vomiting, diarrhea) in liraglutide participants was fairly high, and it is worth considering whether the side effects themselves could have been driving some of the weight loss observed in that group. Furthermore, the out of pocket cost of this medication, when used for weight loss in nondiabetics, is likely to be around $1000 per month. For most patients, this high price will prohibit longer-term use of liraglutide. Even in the setting of a trial where participants faced no out of pocket costs, almost one-third in the liraglutide arm did not complete a year of treatment. On a related note, the primary analysis for this trial used a “last observation carried forward” approach—somewhat concerning given that patients are likely to regain weight after stopping any weight loss intervention, pharmaceutical or otherwise. The authors do report that a range of sensitivity analyses with varying imputation techniques were conducted and did not change the main conclusions of the trial.

Despite the promising findings from this trial, several important clinical questions remain. What is the durability of health effects for patients who discontinue the medication after a year? What safety concerns may arise in those who can afford to continue using liraglutide at this higher dose for several years? A 2-year follow-up study on participants from the current trial has been completed and those results are expected soon, which may help to shed light on some of these issues [5].  Cost-effectiveness evaluations, and head-to-head comparisons of liraglutide with lower cost weight management options would also be very helpful for clinicians presenting a range of treatment options to patients with obesity.

Applications for Clinical Practice

Liraglutide at a daily dose of 3.0 mg represents a new option for treatment of patients with obesity. It should be used in conjunction with behavioral interventions that promote a more healthful diet and increased physical activity, and may result in clinically meaningful weight loss and decreased risk of diabetes. On the other hand, the medication is costly and associated with some unpleasant GI side effects, both important factors that may limit patients’ ability to use it in the long-term. More studies are needed to establish durability of effects and safety beyond a year and that offer direct comparisons with other evidence-based weight loss tools, pharmaceutical and otherwise.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To evaluate the efficacy of liraglutide for weight loss in a group of nondiabetic patients with obesity.

Design. Randomized double-blind placebo-controlled trial.

Setting and participants. This trial took place across 27 countries in Europe, North America, South America, Asia, Africa and Australia. It was funded by NovoNordisk, the pharmaceutical company that manufactures liraglutide. Participants were 18 years or older, with a BMI of 30 kg/m² (or 27 kg/m² with hypertension or dyslipidemia). Patients with diabetes, those on medications known to induce weight gain (or loss), those with history of bariatric surgery, and those with psychiatric illness were excluded from participating. Patients with prediabetes were not excluded.

Intervention. Participants were randomized (2:1 in favor of study drug) to liraglutide or placebo, stratified according to BMI category and pre-diabetes status. They were started at a 0.6–mg dose of medication and up-titrated as tolerated to a dose of 3.0 mg over several weeks. All received counseling on behavioral changes to promote weight loss. Participants were then followed for 56 weeks. A small subgroup in the liraglutide arm was randomly assigned to switch to placebo after 12 weeks on medication to examine for durability of effect of medication, and to evaluate for safety issues that might occur on drug discontinuation.

Main outcome measures. This study focused on 3 primary outcomes: individual-level weight change from baseline, group-level percentage of participants achieving at least 5% weight loss, and percentage of participants with at least 10% weight loss, all assessed at 56 weeks.

Secondary outcomes included change in BMI, waist circumference, markers of glycemia (hemoglobin A1c, insulin level), markers of cardiometabolic health (blood pressure, lipids, CRP), and health-related quality of life (using several validated survey measures). Adverse events were also assessed.

The investigators used an intention-to-treat analysis, comparing outcomes among all patients who were randomized and received at least 1 dose of liraglutide or placebo. For patients with missing values (eg, due to dropout), outcome values were imputed using the last-observation-carried-forward method. A multivariable analysis of covariance model was used to analyze changes in the primary outcomes and included a covariate for the baseline measure of the outcome in question. Sensitivity analyses were conducted in which the investigators used different imputation techniques (multiple imputation, repeated measures) to account for missing data.

Results. The trial enrolled 3731 participants, 2487 of whom were randomized to receive liraglutide and 1244 of whom received placebo. The groups were similar on measured baseline characteristics, with a mean age of 45 years, mostly female participants (78.7% in liraglutide arm, 78.1% in placebo), and the vast majority of participants identified as “white” race/ethnicity (84.7% in liraglutide, 85.3% in placebo). Mean baseline BMI was 38.3 kg/m² in both groups. Although overweight patients with BMI 27 kg/m² or greater were included, they represented a small fraction of all participants (2.7% in liraglutide group and 3.5% in placebo group). Furthermore, although patients with overt diabetes were excluded from participating, over half of the participants qualified as having prediabetes (61.4% in liraglutide group, 60.9% in placebo group). Just over one-third (34.2% of liraglutide group, 35.9% placebo) had hypertension diagnosed at baseline. Study withdrawal was relatively substantial in both groups – 71.9% remained enrolled at 56 weeks in the liraglutide group, and 64.4% remained in the placebo arm. The investigators note that withdrawal due to adverse events was more common in the liraglutide group (9.9% of withdrawals vs. 3.8% in placebo), while other reasons for withdrawing (ineffective therapy, withdrawal of consent) were more common among placebo participants.

Liraglutide participants lost significantly more weight than placebo participants at 56 weeks (mean [SD]  8.0 [6.7] kg vs. 2.6 [5.7] kg). Similarly, more patients in the liraglutide group achieved at least 5% weight loss (63% vs. 27%), and 10% weight loss (33.1% vs. 10.6%) than those taking placebo. When subgroups of patients were examined according to baseline BMI, the investigators suggested that liraglutide appeared to be more effective at promoting weight loss among patients starting below 40 kg/m².

Hemoglobin A1c dropped significantly more (–0.23 points, P < 0.001) among liraglutide participants than among placebo participants. Similarly, fasting insulin levels dropped by 8% more (P < 0.001) in the liraglutide group at 56 weeks. In keeping with the greater weight loss, markers of cardiometabolic health also improved to a greater extent among liraglutide participants, with larger decreases in blood pressure (SBP –2.8 mm Hg lower in liraglutide, P < 0.001), and LDL (–2.4% difference, P = 0.002), and a larger increase in HDL (1.9% difference, P = 0.001). By week 56, 14% of prediabetic patients in the placebo arm had received a new diagnosis of diabetes, compared to just 4% in the liraglutide group (P < 0.001).

Quality of life scores were higher for liraglutide participants on all included measures except those related to side effects of treatment, where placebo participants reported lower levels of side effects. The most common side effects reported by liraglutide participants related to GI upset, including nausea (40%), diarrhea (21%), and vomiting (16%). More serious events, including cholelithiasis (0.8%), cholecystitis (0.5%), and pancreatitis (0.2%), were also reported. Somewhat surprisingly, although liraglutide is also used to improve glycemic control in diabetics, rates of reported spontaneous hypoglycemia were fairly low in the liraglutide group (1.3% vs. 1.0% in placebo).

Conclusion. Liraglutide given at a dose of 3.0 mg daily, along with lifestyle advice, produces clinically significant weight loss and improvement in glycemic and cardiometabolic parameters that is sustained after 1 full year of treatment.

Commentary

Over the past few years, the FDA has approved a growing list of medications for the treatment of obesity [1,2]. Unlike the prior mainstay for prescription weight management, phentermine, which can only be used for a few months at a time due to concerns about abuse, many of these newer medications are approved for long-term use, aligning well with the growing recognition of obesity as a chronic illness. Interestingly, most of the drugs that have emerged onto the market do not represent novel compounds, but rather are existing drugs that have been repurposed and repackaged for the indication of weight management. These “recycled” medications include Qsymia (a mix of phentermine and topiramate) [1], Contrave (naltrexone and buproprion) [2], and now, Saxenda (liraglutide, also marketed as Victoza for treatment of type 2 diabetes). Liraglutide is a glucagon-like-peptide 1 (GLP-1) analogue, meaning it has an effect similar to that of GLP-1, a gut hormone that stimulates insulin secretion, inhibits pancreatic beta cell apoptosis, inhibits gastric emptying, and decreases appetite by acting on the brain’s satiety centers [3]. For several years, endocrinologists and some internists have been using liraglutide (Victoza) to help with glycemic control in diabetics, with the known benefit that, unlike some other diabetes medications, it tends to promote modest weight loss [4].

In this large multicenter trial, Pi-Sunyer et al evaluated the efficacy of liraglutide at a 3.0 mg daily dose (almost twice the dose used for diabetes) for weight management. The trial utilized a strong study design, with double blinding, randomization of a subgroup for early discontinuation (to evaluate for weight regain and stopping-related side effects), and, importantly, the intervention for both groups also included a behavior change component (albeit one of relatively low intensity, based on the limited description). Patients were followed for 56 weeks on the medication, making the “intervention” phase of the study longer than what has been done in many diet trials. Testing for a long-lasting impact on weight, and at the same time attempting to quantify risks associated with longer-term use of a medication, was an important contribution for this study given that liraglutide is being marketed for long-term use.

After a year on liraglutide, participants in that group had lost around 12 lb more, on average, than those using placebo, and had achieved greater improvements cardiometabolic risk markers, with a much lower risk of developing diabetes. While these findings are promising from a clinical standpoint, it is not clear whether the moderate health impacts of this drug will be sufficient to outweigh several issues that may impede its widespread use in practice. The rate of GI side effects (nausea, vomiting, diarrhea) in liraglutide participants was fairly high, and it is worth considering whether the side effects themselves could have been driving some of the weight loss observed in that group. Furthermore, the out of pocket cost of this medication, when used for weight loss in nondiabetics, is likely to be around $1000 per month. For most patients, this high price will prohibit longer-term use of liraglutide. Even in the setting of a trial where participants faced no out of pocket costs, almost one-third in the liraglutide arm did not complete a year of treatment. On a related note, the primary analysis for this trial used a “last observation carried forward” approach—somewhat concerning given that patients are likely to regain weight after stopping any weight loss intervention, pharmaceutical or otherwise. The authors do report that a range of sensitivity analyses with varying imputation techniques were conducted and did not change the main conclusions of the trial.

Despite the promising findings from this trial, several important clinical questions remain. What is the durability of health effects for patients who discontinue the medication after a year? What safety concerns may arise in those who can afford to continue using liraglutide at this higher dose for several years? A 2-year follow-up study on participants from the current trial has been completed and those results are expected soon, which may help to shed light on some of these issues [5].  Cost-effectiveness evaluations, and head-to-head comparisons of liraglutide with lower cost weight management options would also be very helpful for clinicians presenting a range of treatment options to patients with obesity.

Applications for Clinical Practice

Liraglutide at a daily dose of 3.0 mg represents a new option for treatment of patients with obesity. It should be used in conjunction with behavioral interventions that promote a more healthful diet and increased physical activity, and may result in clinically meaningful weight loss and decreased risk of diabetes. On the other hand, the medication is costly and associated with some unpleasant GI side effects, both important factors that may limit patients’ ability to use it in the long-term. More studies are needed to establish durability of effects and safety beyond a year and that offer direct comparisons with other evidence-based weight loss tools, pharmaceutical and otherwise.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To evaluate the efficacy of liraglutide for weight loss in a group of nondiabetic patients with obesity.

Design. Randomized double-blind placebo-controlled trial.

Setting and participants. This trial took place across 27 countries in Europe, North America, South America, Asia, Africa and Australia. It was funded by NovoNordisk, the pharmaceutical company that manufactures liraglutide. Participants were 18 years or older, with a BMI of 30 kg/m² (or 27 kg/m² with hypertension or dyslipidemia). Patients with diabetes, those on medications known to induce weight gain (or loss), those with history of bariatric surgery, and those with psychiatric illness were excluded from participating. Patients with prediabetes were not excluded.

Intervention. Participants were randomized (2:1 in favor of study drug) to liraglutide or placebo, stratified according to BMI category and pre-diabetes status. They were started at a 0.6–mg dose of medication and up-titrated as tolerated to a dose of 3.0 mg over several weeks. All received counseling on behavioral changes to promote weight loss. Participants were then followed for 56 weeks. A small subgroup in the liraglutide arm was randomly assigned to switch to placebo after 12 weeks on medication to examine for durability of effect of medication, and to evaluate for safety issues that might occur on drug discontinuation.

Main outcome measures. This study focused on 3 primary outcomes: individual-level weight change from baseline, group-level percentage of participants achieving at least 5% weight loss, and percentage of participants with at least 10% weight loss, all assessed at 56 weeks.

Secondary outcomes included change in BMI, waist circumference, markers of glycemia (hemoglobin A1c, insulin level), markers of cardiometabolic health (blood pressure, lipids, CRP), and health-related quality of life (using several validated survey measures). Adverse events were also assessed.

The investigators used an intention-to-treat analysis, comparing outcomes among all patients who were randomized and received at least 1 dose of liraglutide or placebo. For patients with missing values (eg, due to dropout), outcome values were imputed using the last-observation-carried-forward method. A multivariable analysis of covariance model was used to analyze changes in the primary outcomes and included a covariate for the baseline measure of the outcome in question. Sensitivity analyses were conducted in which the investigators used different imputation techniques (multiple imputation, repeated measures) to account for missing data.

Results. The trial enrolled 3731 participants, 2487 of whom were randomized to receive liraglutide and 1244 of whom received placebo. The groups were similar on measured baseline characteristics, with a mean age of 45 years, mostly female participants (78.7% in liraglutide arm, 78.1% in placebo), and the vast majority of participants identified as “white” race/ethnicity (84.7% in liraglutide, 85.3% in placebo). Mean baseline BMI was 38.3 kg/m² in both groups. Although overweight patients with BMI 27 kg/m² or greater were included, they represented a small fraction of all participants (2.7% in liraglutide group and 3.5% in placebo group). Furthermore, although patients with overt diabetes were excluded from participating, over half of the participants qualified as having prediabetes (61.4% in liraglutide group, 60.9% in placebo group). Just over one-third (34.2% of liraglutide group, 35.9% placebo) had hypertension diagnosed at baseline. Study withdrawal was relatively substantial in both groups – 71.9% remained enrolled at 56 weeks in the liraglutide group, and 64.4% remained in the placebo arm. The investigators note that withdrawal due to adverse events was more common in the liraglutide group (9.9% of withdrawals vs. 3.8% in placebo), while other reasons for withdrawing (ineffective therapy, withdrawal of consent) were more common among placebo participants.

Liraglutide participants lost significantly more weight than placebo participants at 56 weeks (mean [SD]  8.0 [6.7] kg vs. 2.6 [5.7] kg). Similarly, more patients in the liraglutide group achieved at least 5% weight loss (63% vs. 27%), and 10% weight loss (33.1% vs. 10.6%) than those taking placebo. When subgroups of patients were examined according to baseline BMI, the investigators suggested that liraglutide appeared to be more effective at promoting weight loss among patients starting below 40 kg/m².

Hemoglobin A1c dropped significantly more (–0.23 points, P < 0.001) among liraglutide participants than among placebo participants. Similarly, fasting insulin levels dropped by 8% more (P < 0.001) in the liraglutide group at 56 weeks. In keeping with the greater weight loss, markers of cardiometabolic health also improved to a greater extent among liraglutide participants, with larger decreases in blood pressure (SBP –2.8 mm Hg lower in liraglutide, P < 0.001), and LDL (–2.4% difference, P = 0.002), and a larger increase in HDL (1.9% difference, P = 0.001). By week 56, 14% of prediabetic patients in the placebo arm had received a new diagnosis of diabetes, compared to just 4% in the liraglutide group (P < 0.001).

Quality of life scores were higher for liraglutide participants on all included measures except those related to side effects of treatment, where placebo participants reported lower levels of side effects. The most common side effects reported by liraglutide participants related to GI upset, including nausea (40%), diarrhea (21%), and vomiting (16%). More serious events, including cholelithiasis (0.8%), cholecystitis (0.5%), and pancreatitis (0.2%), were also reported. Somewhat surprisingly, although liraglutide is also used to improve glycemic control in diabetics, rates of reported spontaneous hypoglycemia were fairly low in the liraglutide group (1.3% vs. 1.0% in placebo).

Conclusion. Liraglutide given at a dose of 3.0 mg daily, along with lifestyle advice, produces clinically significant weight loss and improvement in glycemic and cardiometabolic parameters that is sustained after 1 full year of treatment.

Commentary

Over the past few years, the FDA has approved a growing list of medications for the treatment of obesity [1,2]. Unlike the prior mainstay for prescription weight management, phentermine, which can only be used for a few months at a time due to concerns about abuse, many of these newer medications are approved for long-term use, aligning well with the growing recognition of obesity as a chronic illness. Interestingly, most of the drugs that have emerged onto the market do not represent novel compounds, but rather are existing drugs that have been repurposed and repackaged for the indication of weight management. These “recycled” medications include Qsymia (a mix of phentermine and topiramate) [1], Contrave (naltrexone and buproprion) [2], and now, Saxenda (liraglutide, also marketed as Victoza for treatment of type 2 diabetes). Liraglutide is a glucagon-like-peptide 1 (GLP-1) analogue, meaning it has an effect similar to that of GLP-1, a gut hormone that stimulates insulin secretion, inhibits pancreatic beta cell apoptosis, inhibits gastric emptying, and decreases appetite by acting on the brain’s satiety centers [3]. For several years, endocrinologists and some internists have been using liraglutide (Victoza) to help with glycemic control in diabetics, with the known benefit that, unlike some other diabetes medications, it tends to promote modest weight loss [4].

In this large multicenter trial, Pi-Sunyer et al evaluated the efficacy of liraglutide at a 3.0 mg daily dose (almost twice the dose used for diabetes) for weight management. The trial utilized a strong study design, with double blinding, randomization of a subgroup for early discontinuation (to evaluate for weight regain and stopping-related side effects), and, importantly, the intervention for both groups also included a behavior change component (albeit one of relatively low intensity, based on the limited description). Patients were followed for 56 weeks on the medication, making the “intervention” phase of the study longer than what has been done in many diet trials. Testing for a long-lasting impact on weight, and at the same time attempting to quantify risks associated with longer-term use of a medication, was an important contribution for this study given that liraglutide is being marketed for long-term use.

After a year on liraglutide, participants in that group had lost around 12 lb more, on average, than those using placebo, and had achieved greater improvements cardiometabolic risk markers, with a much lower risk of developing diabetes. While these findings are promising from a clinical standpoint, it is not clear whether the moderate health impacts of this drug will be sufficient to outweigh several issues that may impede its widespread use in practice. The rate of GI side effects (nausea, vomiting, diarrhea) in liraglutide participants was fairly high, and it is worth considering whether the side effects themselves could have been driving some of the weight loss observed in that group. Furthermore, the out of pocket cost of this medication, when used for weight loss in nondiabetics, is likely to be around $1000 per month. For most patients, this high price will prohibit longer-term use of liraglutide. Even in the setting of a trial where participants faced no out of pocket costs, almost one-third in the liraglutide arm did not complete a year of treatment. On a related note, the primary analysis for this trial used a “last observation carried forward” approach—somewhat concerning given that patients are likely to regain weight after stopping any weight loss intervention, pharmaceutical or otherwise. The authors do report that a range of sensitivity analyses with varying imputation techniques were conducted and did not change the main conclusions of the trial.

Despite the promising findings from this trial, several important clinical questions remain. What is the durability of health effects for patients who discontinue the medication after a year? What safety concerns may arise in those who can afford to continue using liraglutide at this higher dose for several years? A 2-year follow-up study on participants from the current trial has been completed and those results are expected soon, which may help to shed light on some of these issues [5].  Cost-effectiveness evaluations, and head-to-head comparisons of liraglutide with lower cost weight management options would also be very helpful for clinicians presenting a range of treatment options to patients with obesity.

Applications for Clinical Practice

Liraglutide at a daily dose of 3.0 mg represents a new option for treatment of patients with obesity. It should be used in conjunction with behavioral interventions that promote a more healthful diet and increased physical activity, and may result in clinically meaningful weight loss and decreased risk of diabetes. On the other hand, the medication is costly and associated with some unpleasant GI side effects, both important factors that may limit patients’ ability to use it in the long-term. More studies are needed to establish durability of effects and safety beyond a year and that offer direct comparisons with other evidence-based weight loss tools, pharmaceutical and otherwise.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine the association between bariatric surgery and long-term mortality rates among patients with severe obesity.

Design. Retrospective cohort study.

Setting and participants. This analysis relied upon data from Veteran’s Administration (VA) patients undergoing bariatric surgery between 2000 and 2011 and a group of matched controls. For this data-only study, a waiver of informed consent was obtained. Investigators first used the VA Surgical Quality Improvement Program (SQIP) dataset to identify all bariatric surgical procedures performed at VA hospitals between 2000 and the end of 2011, excluding patients who had any evidence of body mass index (BMI) less than 35 kg/m² and those with certain baseline diagnoses that would be considered contraindications for surgery, as well as those who had prolonged inpatient stays immediately prior to their surgical date. No upper or lower age limits appear to have been specified, and no upper BMI limit appeared to have been set.

Once all surgical patients were identified, the investigators attempted to find a group of similar control patients who had not undergone surgery. Initially they pulled candidate matches for each surgical patient based on having the same sex, age-group (within 5 years), BMI category (35-40, 40-50, >50), diabetes status (present or absent), racial category, and VA region. From these candidates, they selected up to 3 of the closest matches on age, BMI, and a composite comorbidity score based on inpatient and outpatient claims in the year prior to surgery. The authors specified that controls could convert to surgical patients during the follow-up period, in which case their data was censored beginning with the surgical procedure. However, if a control patient underwent surgery during 2012 or 2013, censoring was not possible given that the dataset for identifying surgeries contained only procedures performed through the end of 2011.

Main outcome measures. The primary outcome of interest was time to death (any cause) beginning at the date of surgery (or baseline date for nonsurgical controls) through the end of 2013. The investigators built Cox proportional hazards models to evaluate survival using multivariable models to adjust for baseline characteristics, including those involved in the matching process, as well as others that might have differentially impacted both likelihood of undergoing surgery and mortality risk. These included marital status, insurance markers of low income or disability, and a number of comorbid medical and psychiatric diagnoses.

In addition to the main analyses, the investigators also looked for effect modification of the surgery-mortality relationship by a patient’s sex and presence or absence of diabetes at the time of surgery, as well as the time period in which their surgery was conducted, dichotomized around the year 2006. This year was selected for several reasons, including that it was the year in which a VA-wide comprehensive weight management and surgical selection program was instituted.

Results. The surgical cohort was made up of 2500 patients, and there were 7462 matched controls. The surgical and control groups were similar with respect to matched baseline characteristics, tested using standardized differences (as opposed to t test or chi-square). Mean (SD) age was 52 (8.8) years for surgical patients versus 53 (8.7) years for controls. 74% of patients in both the surgical and control groups were men, and 81% in both groups were white (ethnicity not specified). Mean (SD) baseline BMI was 47 (7.9) kg/m² in the surgical group and 46 (7.3) kg/m² for controls.

Some between-group differences were present for baseline characteristics that had not been included in the matching protocol. More surgical patients than controls had diagnoses of hypertension (80% surgical vs. 70% control), dyslipidemia (61% vs. 52%), arthritis (27% vs. 15%), depression (44% vs. 32%), GERD (35% vs.19%), and fatty liver disease (6.6% vs. 0.6%). In contrast, more control patients than surgical patients had diagnoses of alcohol abuse (6.2% in controls vs. 3.9% in surgical) and schizophrenia (4.9% vs. 1.8%). Also, although a number of different surgical types were represented in the cohort, the vast majority of procedures were classified as Roux-en-Y gastric bypasses (RYGB). 53% of the procedures were open RYGB, 21% were laparoscopic RYGB, 10% were adjustable gastric bands (AGB), and 15% were vertical sleeve gastrectomies (VSG).

Mortality was lower among surgical patients than among matched controls during a mean follow-up time of 6.9 years for surgical patients and 6.6 years for controls. Namely, the 1-, 5- and 10-year cumulative mortality rates for surgical patients were: 2.4%, 6.4%, and 13.8%. Unadjusted mortality rates for nonsurgical controls were lower initially (1.7% at 1 year), but then much higher at years 5 (10.4%), and 10 (23.9%). In multivariable Cox models, the hazard ratio (HR) for mortality in bariatric patients versus controls was nonsignificant at 1 year of follow-up. However, between 1 and 5 years after surgery (or after baseline), multivariable models showed an HR (95% CI) of 0.45 (0.36–0.56) for mortality among surgical patients versus controls. For those with more than 5 years of follow up, the HR was similar (0.47, 95% CI 0.39–0.58) for death among surgical versus control patients. The investigators found that the year during which a patient underwent surgery (before or after 2006) did impact mortality during the first postoperative year, with those who had earlier procedures (2000-2005) exhibiting a significantly higher risk of death in that year relative to non-operative controls (HR 1.66, 95% CI 1.19–2.33). No significant sex or diabetes interactions were observed for the surgery-mortality relationship in multivariable Cox models. There was no information provided as to the breakdown of cause of death within the larger “all-cause mortality” outcome.

Conclusion. Bariatric surgery was associated with significantly lower all-cause mortality among surgical patients in the VA over a 5- to 14-year follow-up period compared with a group of severely obese VA patients who did not undergo surgery.

Commentary

Rates of severe obesity (BMI ≥ 35 kg/m²) have risen at a faster pace than those of obesity in the United States over the past decade [1], driving clinicians, patients and payers to search for effective methods of treating this condition. Bariatric surgery has emerged as the most effective treatment for severe obesity; however, the existing surgical literature is predominated by studies with short- or medium-term postoperative follow-up and homogenous participant populations containing large numbers of younger non-Hispanic white women. Research from the Swedish Obesity Study (SOS), as well as smaller US-based studies, has suggested that severely obese patients who undergo bariatric surgery have better long-term survival than their nonsurgical counterparts [2,3].Counter to this finding, a previous medium-term study utilizing data from VA hospitals did not find that surgery conferred a mortality benefit among this largely male, older, and sicker patient population [4].The current study, by the same group of investigators, attempts to update the previous finding by including more recent surgical data and a longer follow-up period, to see whether or not a survival benefit appears to emerge for VA patients undergoing bariatric surgery.

A major strength of this study was the use of a large and comprehensive clinical dataset, a strength of many studies utilizing data from the VA. The availability of clinical data such as BMI, as well as diagnostic codes and sociodemographic variables, allowed the authors to match and adjust for a number of potential confounders of the surgery-mortality relationship. Another unique feature of VA data is that members of this health care system can often be followed regardless of their location, as the unified medical record transfers between states. This is in contrast to many claims-based or single-center studies of surgery, where patients are lost to follow-up if they move or transfer insurance providers. This study clearly benefited from this aspect of VA data, with a mean postoperative follow-up period of over 5 years in both study groups, much longer than is typically observed in bariatric surgical studies, and probably a necessary feature for examining more of a rare outcome such as mortality (as opposed to comparing weight loss or diabetes remission). Another clear contribution of this study is that it focused on a group of patients not typical of bariatric cohorts—this group was slightly older and sicker, with far more men than women, and therefore at a much higher risk of mortality than the typically younger females that are part of most studies.

Although the authors did adjust for many factors when comparing the surgical and nonsurgical groups, it is possible, as with any observational study, that unmeasured confounders may have been present. Psychosocial and behavioral features that may be linked both to a person’s likelihood of undergoing surgery, and to their mortality risk are of particular concern. It is worth noting, for example, that far more patients in the nonsurgical group were identified as schizophrenic, and that the rate of schizophrenia in that severely obese group was much higher than that of the general population. This pattern may have some relationship to the weight-gain promoting effect of antipsychotic medications and the unfortunate reality that patients with severe obesity and severe mental illness may not be as well equipped to seek out surgery (or viewed as acceptable candidates) as those without severe mental illness. One possible limitation mentioned by the authors was that control group patients who underwent surgery in 2012 or 2013 would not have been recognized (and thus had their data censored in this study), possibly leading to incorrect categorization of exposure category for some amount of person-time during follow-up. In general, though, there is a low likelihood of this phenomenon impacting the findings, given both the relative infrequency of crossover observed in the cohort prior to 2011, and the relatively short amount of person-time any later crossovers would have contributed in the later years of the study.

Although codes for baseline disease states were adjusted for in multivariable analyses, the surgical patients were in general a medically sicker group at baseline than control patients. As the authors point out, if anything, this should have biased the findings in favor of seeing higher mortality rate in the surgical group, the opposite of what was found. Further strengthening the finding of a correlation between survival and surgery is the mix of procedure types included in this study. Over half of the procedures were open RYGB surgeries, with far fewer of the more modern and lower risk procedures (eg, laparoscopic RYGB) represented. Again, this mix of procedures would be expected to result in an overestimation of mortality in surgical patients relative to what might be observed if all patients had been drawn from later years of the cohort, as surgical technique evolved.

Applications for Clinical Practice

This study adds to the evidence that patients with severe obesity who undergo bariatric surgery have a lower risk of death up to 10 years after their surgery compared with patients who do not have these procedures. The findings of this work should provide encouragement, particularly for managaing older adults with more longstanding comorbidities. Those who are strongly motivated to pursue weight loss surgery, and who are deemed good candidates by bariatric teams, may add years to their lives by undergoing one of these procedures. As always, however, the quality of life experienced by patients after surgery, and a realistic expectation of the ways in which surgery will fundamentally change their lifestyle, must be a critical part of the discussion.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine the association between bariatric surgery and long-term mortality rates among patients with severe obesity.

Design. Retrospective cohort study.

Setting and participants. This analysis relied upon data from Veteran’s Administration (VA) patients undergoing bariatric surgery between 2000 and 2011 and a group of matched controls. For this data-only study, a waiver of informed consent was obtained. Investigators first used the VA Surgical Quality Improvement Program (SQIP) dataset to identify all bariatric surgical procedures performed at VA hospitals between 2000 and the end of 2011, excluding patients who had any evidence of body mass index (BMI) less than 35 kg/m² and those with certain baseline diagnoses that would be considered contraindications for surgery, as well as those who had prolonged inpatient stays immediately prior to their surgical date. No upper or lower age limits appear to have been specified, and no upper BMI limit appeared to have been set.

Once all surgical patients were identified, the investigators attempted to find a group of similar control patients who had not undergone surgery. Initially they pulled candidate matches for each surgical patient based on having the same sex, age-group (within 5 years), BMI category (35-40, 40-50, >50), diabetes status (present or absent), racial category, and VA region. From these candidates, they selected up to 3 of the closest matches on age, BMI, and a composite comorbidity score based on inpatient and outpatient claims in the year prior to surgery. The authors specified that controls could convert to surgical patients during the follow-up period, in which case their data was censored beginning with the surgical procedure. However, if a control patient underwent surgery during 2012 or 2013, censoring was not possible given that the dataset for identifying surgeries contained only procedures performed through the end of 2011.

Main outcome measures. The primary outcome of interest was time to death (any cause) beginning at the date of surgery (or baseline date for nonsurgical controls) through the end of 2013. The investigators built Cox proportional hazards models to evaluate survival using multivariable models to adjust for baseline characteristics, including those involved in the matching process, as well as others that might have differentially impacted both likelihood of undergoing surgery and mortality risk. These included marital status, insurance markers of low income or disability, and a number of comorbid medical and psychiatric diagnoses.

In addition to the main analyses, the investigators also looked for effect modification of the surgery-mortality relationship by a patient’s sex and presence or absence of diabetes at the time of surgery, as well as the time period in which their surgery was conducted, dichotomized around the year 2006. This year was selected for several reasons, including that it was the year in which a VA-wide comprehensive weight management and surgical selection program was instituted.

Results. The surgical cohort was made up of 2500 patients, and there were 7462 matched controls. The surgical and control groups were similar with respect to matched baseline characteristics, tested using standardized differences (as opposed to t test or chi-square). Mean (SD) age was 52 (8.8) years for surgical patients versus 53 (8.7) years for controls. 74% of patients in both the surgical and control groups were men, and 81% in both groups were white (ethnicity not specified). Mean (SD) baseline BMI was 47 (7.9) kg/m² in the surgical group and 46 (7.3) kg/m² for controls.

Some between-group differences were present for baseline characteristics that had not been included in the matching protocol. More surgical patients than controls had diagnoses of hypertension (80% surgical vs. 70% control), dyslipidemia (61% vs. 52%), arthritis (27% vs. 15%), depression (44% vs. 32%), GERD (35% vs.19%), and fatty liver disease (6.6% vs. 0.6%). In contrast, more control patients than surgical patients had diagnoses of alcohol abuse (6.2% in controls vs. 3.9% in surgical) and schizophrenia (4.9% vs. 1.8%). Also, although a number of different surgical types were represented in the cohort, the vast majority of procedures were classified as Roux-en-Y gastric bypasses (RYGB). 53% of the procedures were open RYGB, 21% were laparoscopic RYGB, 10% were adjustable gastric bands (AGB), and 15% were vertical sleeve gastrectomies (VSG).

Mortality was lower among surgical patients than among matched controls during a mean follow-up time of 6.9 years for surgical patients and 6.6 years for controls. Namely, the 1-, 5- and 10-year cumulative mortality rates for surgical patients were: 2.4%, 6.4%, and 13.8%. Unadjusted mortality rates for nonsurgical controls were lower initially (1.7% at 1 year), but then much higher at years 5 (10.4%), and 10 (23.9%). In multivariable Cox models, the hazard ratio (HR) for mortality in bariatric patients versus controls was nonsignificant at 1 year of follow-up. However, between 1 and 5 years after surgery (or after baseline), multivariable models showed an HR (95% CI) of 0.45 (0.36–0.56) for mortality among surgical patients versus controls. For those with more than 5 years of follow up, the HR was similar (0.47, 95% CI 0.39–0.58) for death among surgical versus control patients. The investigators found that the year during which a patient underwent surgery (before or after 2006) did impact mortality during the first postoperative year, with those who had earlier procedures (2000-2005) exhibiting a significantly higher risk of death in that year relative to non-operative controls (HR 1.66, 95% CI 1.19–2.33). No significant sex or diabetes interactions were observed for the surgery-mortality relationship in multivariable Cox models. There was no information provided as to the breakdown of cause of death within the larger “all-cause mortality” outcome.

Conclusion. Bariatric surgery was associated with significantly lower all-cause mortality among surgical patients in the VA over a 5- to 14-year follow-up period compared with a group of severely obese VA patients who did not undergo surgery.

Commentary

Rates of severe obesity (BMI ≥ 35 kg/m²) have risen at a faster pace than those of obesity in the United States over the past decade [1], driving clinicians, patients and payers to search for effective methods of treating this condition. Bariatric surgery has emerged as the most effective treatment for severe obesity; however, the existing surgical literature is predominated by studies with short- or medium-term postoperative follow-up and homogenous participant populations containing large numbers of younger non-Hispanic white women. Research from the Swedish Obesity Study (SOS), as well as smaller US-based studies, has suggested that severely obese patients who undergo bariatric surgery have better long-term survival than their nonsurgical counterparts [2,3].Counter to this finding, a previous medium-term study utilizing data from VA hospitals did not find that surgery conferred a mortality benefit among this largely male, older, and sicker patient population [4].The current study, by the same group of investigators, attempts to update the previous finding by including more recent surgical data and a longer follow-up period, to see whether or not a survival benefit appears to emerge for VA patients undergoing bariatric surgery.

A major strength of this study was the use of a large and comprehensive clinical dataset, a strength of many studies utilizing data from the VA. The availability of clinical data such as BMI, as well as diagnostic codes and sociodemographic variables, allowed the authors to match and adjust for a number of potential confounders of the surgery-mortality relationship. Another unique feature of VA data is that members of this health care system can often be followed regardless of their location, as the unified medical record transfers between states. This is in contrast to many claims-based or single-center studies of surgery, where patients are lost to follow-up if they move or transfer insurance providers. This study clearly benefited from this aspect of VA data, with a mean postoperative follow-up period of over 5 years in both study groups, much longer than is typically observed in bariatric surgical studies, and probably a necessary feature for examining more of a rare outcome such as mortality (as opposed to comparing weight loss or diabetes remission). Another clear contribution of this study is that it focused on a group of patients not typical of bariatric cohorts—this group was slightly older and sicker, with far more men than women, and therefore at a much higher risk of mortality than the typically younger females that are part of most studies.

Although the authors did adjust for many factors when comparing the surgical and nonsurgical groups, it is possible, as with any observational study, that unmeasured confounders may have been present. Psychosocial and behavioral features that may be linked both to a person’s likelihood of undergoing surgery, and to their mortality risk are of particular concern. It is worth noting, for example, that far more patients in the nonsurgical group were identified as schizophrenic, and that the rate of schizophrenia in that severely obese group was much higher than that of the general population. This pattern may have some relationship to the weight-gain promoting effect of antipsychotic medications and the unfortunate reality that patients with severe obesity and severe mental illness may not be as well equipped to seek out surgery (or viewed as acceptable candidates) as those without severe mental illness. One possible limitation mentioned by the authors was that control group patients who underwent surgery in 2012 or 2013 would not have been recognized (and thus had their data censored in this study), possibly leading to incorrect categorization of exposure category for some amount of person-time during follow-up. In general, though, there is a low likelihood of this phenomenon impacting the findings, given both the relative infrequency of crossover observed in the cohort prior to 2011, and the relatively short amount of person-time any later crossovers would have contributed in the later years of the study.

Although codes for baseline disease states were adjusted for in multivariable analyses, the surgical patients were in general a medically sicker group at baseline than control patients. As the authors point out, if anything, this should have biased the findings in favor of seeing higher mortality rate in the surgical group, the opposite of what was found. Further strengthening the finding of a correlation between survival and surgery is the mix of procedure types included in this study. Over half of the procedures were open RYGB surgeries, with far fewer of the more modern and lower risk procedures (eg, laparoscopic RYGB) represented. Again, this mix of procedures would be expected to result in an overestimation of mortality in surgical patients relative to what might be observed if all patients had been drawn from later years of the cohort, as surgical technique evolved.

Applications for Clinical Practice

This study adds to the evidence that patients with severe obesity who undergo bariatric surgery have a lower risk of death up to 10 years after their surgery compared with patients who do not have these procedures. The findings of this work should provide encouragement, particularly for managaing older adults with more longstanding comorbidities. Those who are strongly motivated to pursue weight loss surgery, and who are deemed good candidates by bariatric teams, may add years to their lives by undergoing one of these procedures. As always, however, the quality of life experienced by patients after surgery, and a realistic expectation of the ways in which surgery will fundamentally change their lifestyle, must be a critical part of the discussion.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine the association between bariatric surgery and long-term mortality rates among patients with severe obesity.

Design. Retrospective cohort study.

Setting and participants. This analysis relied upon data from Veteran’s Administration (VA) patients undergoing bariatric surgery between 2000 and 2011 and a group of matched controls. For this data-only study, a waiver of informed consent was obtained. Investigators first used the VA Surgical Quality Improvement Program (SQIP) dataset to identify all bariatric surgical procedures performed at VA hospitals between 2000 and the end of 2011, excluding patients who had any evidence of body mass index (BMI) less than 35 kg/m² and those with certain baseline diagnoses that would be considered contraindications for surgery, as well as those who had prolonged inpatient stays immediately prior to their surgical date. No upper or lower age limits appear to have been specified, and no upper BMI limit appeared to have been set.

Once all surgical patients were identified, the investigators attempted to find a group of similar control patients who had not undergone surgery. Initially they pulled candidate matches for each surgical patient based on having the same sex, age-group (within 5 years), BMI category (35-40, 40-50, >50), diabetes status (present or absent), racial category, and VA region. From these candidates, they selected up to 3 of the closest matches on age, BMI, and a composite comorbidity score based on inpatient and outpatient claims in the year prior to surgery. The authors specified that controls could convert to surgical patients during the follow-up period, in which case their data was censored beginning with the surgical procedure. However, if a control patient underwent surgery during 2012 or 2013, censoring was not possible given that the dataset for identifying surgeries contained only procedures performed through the end of 2011.

Main outcome measures. The primary outcome of interest was time to death (any cause) beginning at the date of surgery (or baseline date for nonsurgical controls) through the end of 2013. The investigators built Cox proportional hazards models to evaluate survival using multivariable models to adjust for baseline characteristics, including those involved in the matching process, as well as others that might have differentially impacted both likelihood of undergoing surgery and mortality risk. These included marital status, insurance markers of low income or disability, and a number of comorbid medical and psychiatric diagnoses.

In addition to the main analyses, the investigators also looked for effect modification of the surgery-mortality relationship by a patient’s sex and presence or absence of diabetes at the time of surgery, as well as the time period in which their surgery was conducted, dichotomized around the year 2006. This year was selected for several reasons, including that it was the year in which a VA-wide comprehensive weight management and surgical selection program was instituted.

Results. The surgical cohort was made up of 2500 patients, and there were 7462 matched controls. The surgical and control groups were similar with respect to matched baseline characteristics, tested using standardized differences (as opposed to t test or chi-square). Mean (SD) age was 52 (8.8) years for surgical patients versus 53 (8.7) years for controls. 74% of patients in both the surgical and control groups were men, and 81% in both groups were white (ethnicity not specified). Mean (SD) baseline BMI was 47 (7.9) kg/m² in the surgical group and 46 (7.3) kg/m² for controls.

Some between-group differences were present for baseline characteristics that had not been included in the matching protocol. More surgical patients than controls had diagnoses of hypertension (80% surgical vs. 70% control), dyslipidemia (61% vs. 52%), arthritis (27% vs. 15%), depression (44% vs. 32%), GERD (35% vs.19%), and fatty liver disease (6.6% vs. 0.6%). In contrast, more control patients than surgical patients had diagnoses of alcohol abuse (6.2% in controls vs. 3.9% in surgical) and schizophrenia (4.9% vs. 1.8%). Also, although a number of different surgical types were represented in the cohort, the vast majority of procedures were classified as Roux-en-Y gastric bypasses (RYGB). 53% of the procedures were open RYGB, 21% were laparoscopic RYGB, 10% were adjustable gastric bands (AGB), and 15% were vertical sleeve gastrectomies (VSG).

Mortality was lower among surgical patients than among matched controls during a mean follow-up time of 6.9 years for surgical patients and 6.6 years for controls. Namely, the 1-, 5- and 10-year cumulative mortality rates for surgical patients were: 2.4%, 6.4%, and 13.8%. Unadjusted mortality rates for nonsurgical controls were lower initially (1.7% at 1 year), but then much higher at years 5 (10.4%), and 10 (23.9%). In multivariable Cox models, the hazard ratio (HR) for mortality in bariatric patients versus controls was nonsignificant at 1 year of follow-up. However, between 1 and 5 years after surgery (or after baseline), multivariable models showed an HR (95% CI) of 0.45 (0.36–0.56) for mortality among surgical patients versus controls. For those with more than 5 years of follow up, the HR was similar (0.47, 95% CI 0.39–0.58) for death among surgical versus control patients. The investigators found that the year during which a patient underwent surgery (before or after 2006) did impact mortality during the first postoperative year, with those who had earlier procedures (2000-2005) exhibiting a significantly higher risk of death in that year relative to non-operative controls (HR 1.66, 95% CI 1.19–2.33). No significant sex or diabetes interactions were observed for the surgery-mortality relationship in multivariable Cox models. There was no information provided as to the breakdown of cause of death within the larger “all-cause mortality” outcome.

Conclusion. Bariatric surgery was associated with significantly lower all-cause mortality among surgical patients in the VA over a 5- to 14-year follow-up period compared with a group of severely obese VA patients who did not undergo surgery.

Commentary

Rates of severe obesity (BMI ≥ 35 kg/m²) have risen at a faster pace than those of obesity in the United States over the past decade [1], driving clinicians, patients and payers to search for effective methods of treating this condition. Bariatric surgery has emerged as the most effective treatment for severe obesity; however, the existing surgical literature is predominated by studies with short- or medium-term postoperative follow-up and homogenous participant populations containing large numbers of younger non-Hispanic white women. Research from the Swedish Obesity Study (SOS), as well as smaller US-based studies, has suggested that severely obese patients who undergo bariatric surgery have better long-term survival than their nonsurgical counterparts [2,3].Counter to this finding, a previous medium-term study utilizing data from VA hospitals did not find that surgery conferred a mortality benefit among this largely male, older, and sicker patient population [4].The current study, by the same group of investigators, attempts to update the previous finding by including more recent surgical data and a longer follow-up period, to see whether or not a survival benefit appears to emerge for VA patients undergoing bariatric surgery.

A major strength of this study was the use of a large and comprehensive clinical dataset, a strength of many studies utilizing data from the VA. The availability of clinical data such as BMI, as well as diagnostic codes and sociodemographic variables, allowed the authors to match and adjust for a number of potential confounders of the surgery-mortality relationship. Another unique feature of VA data is that members of this health care system can often be followed regardless of their location, as the unified medical record transfers between states. This is in contrast to many claims-based or single-center studies of surgery, where patients are lost to follow-up if they move or transfer insurance providers. This study clearly benefited from this aspect of VA data, with a mean postoperative follow-up period of over 5 years in both study groups, much longer than is typically observed in bariatric surgical studies, and probably a necessary feature for examining more of a rare outcome such as mortality (as opposed to comparing weight loss or diabetes remission). Another clear contribution of this study is that it focused on a group of patients not typical of bariatric cohorts—this group was slightly older and sicker, with far more men than women, and therefore at a much higher risk of mortality than the typically younger females that are part of most studies.

Although the authors did adjust for many factors when comparing the surgical and nonsurgical groups, it is possible, as with any observational study, that unmeasured confounders may have been present. Psychosocial and behavioral features that may be linked both to a person’s likelihood of undergoing surgery, and to their mortality risk are of particular concern. It is worth noting, for example, that far more patients in the nonsurgical group were identified as schizophrenic, and that the rate of schizophrenia in that severely obese group was much higher than that of the general population. This pattern may have some relationship to the weight-gain promoting effect of antipsychotic medications and the unfortunate reality that patients with severe obesity and severe mental illness may not be as well equipped to seek out surgery (or viewed as acceptable candidates) as those without severe mental illness. One possible limitation mentioned by the authors was that control group patients who underwent surgery in 2012 or 2013 would not have been recognized (and thus had their data censored in this study), possibly leading to incorrect categorization of exposure category for some amount of person-time during follow-up. In general, though, there is a low likelihood of this phenomenon impacting the findings, given both the relative infrequency of crossover observed in the cohort prior to 2011, and the relatively short amount of person-time any later crossovers would have contributed in the later years of the study.

Although codes for baseline disease states were adjusted for in multivariable analyses, the surgical patients were in general a medically sicker group at baseline than control patients. As the authors point out, if anything, this should have biased the findings in favor of seeing higher mortality rate in the surgical group, the opposite of what was found. Further strengthening the finding of a correlation between survival and surgery is the mix of procedure types included in this study. Over half of the procedures were open RYGB surgeries, with far fewer of the more modern and lower risk procedures (eg, laparoscopic RYGB) represented. Again, this mix of procedures would be expected to result in an overestimation of mortality in surgical patients relative to what might be observed if all patients had been drawn from later years of the cohort, as surgical technique evolved.

Applications for Clinical Practice

This study adds to the evidence that patients with severe obesity who undergo bariatric surgery have a lower risk of death up to 10 years after their surgery compared with patients who do not have these procedures. The findings of this work should provide encouragement, particularly for managaing older adults with more longstanding comorbidities. Those who are strongly motivated to pursue weight loss surgery, and who are deemed good candidates by bariatric teams, may add years to their lives by undergoing one of these procedures. As always, however, the quality of life experienced by patients after surgery, and a realistic expectation of the ways in which surgery will fundamentally change their lifestyle, must be a critical part of the discussion.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine the effect of phentermine and topiramate extended release (PHEN/TPM ER) treatment on progression to type 2 diabetes and/or cardiometabolic disease in subjects with prediabetes and/or metabolic syndrome (MetS) at baseline.

Design. Sub-group analysis of a larger double-blind, randomized, placebo-controlled trial of PHEN/TPM ER in overweight and obese adults.

Setting and participants. The larger study had 2 phases —a 56-week weight loss trial (CONQUER, n = 866), followed by an extension of the drug trial out to 108 weeks (SEQUEL, n = 675) in a sub-group of CONQUER participants. The CONQUER trial, based at 93 U.S. centers, enrolled overweight or obese patients with at least 2 obesity-related comorbidities and randomly assigned them to receive either placebo or PHEN/TPM ER at a lower (7.5 mg/46 mg) or higher (15 mg/92 mg) daily dose. All 3 groups also received lifestyle modification counseling that included an evidence-based diet and exercise curriculum. Participants received study drug and lifestyle counseling in the setting of monthly visits during the 60- (CONQUER) or 108-week (SEQUEL) follow-up period.

The analyses presented in this paper focus on the 475 participants who completed both CONQUER and SEQUEL and who were characterized as pre-diabetic or as having the metabolic syndrome (MetS) at baseline. Pre-diabetes was defined as having a blood glucose level of 100–125 mg/dL or higher while fasting, or 140–199 mg/dL after an oral glucose tolerance test (GTT). MetS was characterized in participants who displayed 3 or more of the following at baseline: waist circumference ≥ 102 cm in men or 88 cm in women; triglycerides ≥ 150 mg/dL or on a lipid-lowering medication; HDL < 40 mg/dL in men or < 50 mg/dL in women; systolic BP ≥ 130 mm Hg or diastolic BP ≥ 85 mm Hg (or on antihypertensive); and fasting glucose ≥ 100 mg/dL or on treatment for elevated glucose.

Main outcome measures. The primary outcome for this study was percent weight loss at 108 weeks of follow-up (or early termination). Secondary outcomes included cardiometabolic changes, such as development of type 2 diabetes and changes in lipid measures, blood pressure, and waist circumference. These were assessed at baseline, week 56, and week 108 (or at early termination). Rates of progression to type 2 diabetes were compared between the treatment groups using chi-square testing. Intention-to-treat (ITT) ANCOVA analysis was performed with multiple imputation techniques to address missing data, as well as with an alternative analysis using last observation carried forward.

Results. The study arms were similar with respect to baseline characteristics. Average age was 51 years in the high dose PHEN/TPM ER arm and 52 in the other arms. Over half (65%) of participants were women and 86% were Caucasian. Mean BMI was 36 kg/m² (class II obesity). Over half of participants were on antihypertensive medications at baseline but with well-controlled blood pressure (mean 128/80 mm Hg). Of the 475 people in this analysis, 316 met criteria for prediabetes, 451 for MetS, and 292 for both prediabetes and MetS.

Weight loss at 2 years was significantly greater in subjects taking PHEN/TPM ER (10.9% in the lower dose group, 12.1% in the higher dose group) compared to those taking placebo (2.5%) (P < 0.001). Mirroring weight loss results, type 2 diabetes incidence was also significantly lower in the drug treatment arms than in the placebo arm at 2 years after randomization—annualized incidence was 6.1% for placebo vs. 1.8% for lower-dose drug and 1.3% for higher-dose drug (P < 0.05). Greater weight loss was associated with greater decrease in diabetes incidence across all 3 arms of the study. Those persons who did not achieve at least a 5% weight loss at 2 years had the highest annualized risk of developing diabetes (6.3%), compared with a 0.9% risk among those who lost at least 15% of their weight. Improvements in other cardiometabolic parameters, including HDL, triglycerides, waist circumference, and insulin sensitivity index, was more common among the PHEN/TPM ER participants compared with placebo. Blood pressure decreased slightly for all 3 groups and there was no significant difference between the drug arms and the placebo arm.

Discontinuation of study medication occurred in all 3 groups (3.1% in placebo, 6.1% in lower-dose medication, and 5.5% in higher-dose medication), with serious adverse events in 5%, 7%, and 8.5%, respectively. There were no deaths.

Conclusion. PHEN/TPM ER administered over a 2-year period significantly improved weight loss and decreased progression to type 2 diabetes relative to placebo in a group of at-risk participants.

Commentary

Diabetes and related cardiometabolic disease are major contributors to morbidity and mortality in adults. With the exception of invasive treatments such as bariatric surgery, reversal of diabetes once it is established has proven quite difficult [1,2], and thus there is an increased emphasis from the public health and medical communities on preventing the development of this disease in the first place. Complicating the picture, recently broadened criteria for pre-diabetes will likely result in a very large number of these at-risk individuals being identified [3,4]. Although intensive lifestyle interventions resulting in a 5% to 7% weight loss among pre-diabetics have been shown to delay progression to diabetes [5], the translation of these programs into real-world settings has, so far, shown less promise than the original randomized trials might have indicated [4]. Although there is ongoing work to try to improve results and uptake in community-based lifestyle intervention programs, for many patients and clinicians these resource-intensive programs currently prove difficult to do well on a large scale.

Alternative methods of helping patients achieve and maintain that critical > 5% weight loss are desperately needed, not only for preventing diabetes, but also for impacting the numerous other risks associated with obesity. This particular trial capitalized on the notion that it is probably successful weight loss, not the intervention format used to achieve that weight loss, which drives decreased diabetes risk. This study was a sub-analysis of a larger randomized trial, and many of the strengths of that larger study are therefore reflected in this paper. Participants and study staff were blinded to treatment arm with the use of placebo, a very important strength when adverse reactions and drug intolerances need to be measured. Furthermore, this likely equalized motivation to comply with the lifestyle recommendations across the treatment arms—this might not have been the case if patients were aware that they were or were not receiving study drug. Another key strength of the study is its duration. PHEN/TPM ER is unique in that it is approved by the FDA for long-term use. Whereas many studies of weight loss show maximum intervention effect at about 6 months followed by weight regain, this study showed sustained weight loss up to 2 years after starting therapy, presumably because participants could actually continue the therapy for the full 2 years. Most importantly, the intervention itself (medication plus low-intensity lifestyle counseling) is likely highly replicable in clinical practice.

There are some important limitations to consider when interpreting the results from this study. First, the participants analyzed in this paper were comprised entirely of people who had already participated in a full year of the parent study and therefore probably represent a sub-group that might have been experiencing greater success as a result of their participation, potentially generating an overly optimistic estimate of weight loss and health effect for all of the groups relative to what might be seen in a general population. This feature of the design also limits this study’s ability to comment on drug intolerance or early adverse reactions—those who didn’t stick with the pills for at least a year would not have been included in these analyses. In terms of generalizability, although the infrastructure required from a clinical standpoint is much lower for an intervention like this (prescribing a medication) compared to an intensive lifestyle intervention, these drugs are still costly, and many insurers/providers may not offer them on formulary. Thus, to realize the long-term benefits of sustained weight loss, patients may need to face significant out-of-pocket costs, which may limit uptake of this therapy to those with financial means. For this and other reasons, it will be important to do future studies looking at how quickly weight is re-gained once people stop taking the medication. Another threat to generalizability is the racial makeup of the participants—the vast majority of them were non-Hispanic white. Furthermore, although a majority of the participants had hypertension, it was well-controlled in all (a prerequisite for taking the medication), and it is unclear whether in a real-world patient population hypertension would be adequately controlled in a large number of patients.

Another issue to consider when looking at the use of weight loss medications for prevention of diabetes is the relative risk of prolonged medication use compared with the risk for developing diabetes. Clearly, for obese patients who are interested in losing weight for other reasons, prevention of diabetes is a wonderful side effect of achieving that goal. However, it is worth noting that even in the highest-risk group of participants in this study (those who lost < 5% of weight), the annualized risk of developing diabetes was about 6% (< 20% cumulative risk projected over 3 years). Compare this to the 7% to 8% serious adverse event rate observed in those on drug therapy. Although the medication did reduce annualized diabetes risk significantly, the vast majority of people in all the arms did not develop diabetes during follow-up. This drives home the point that our current categorization of pre-diabetes is far from perfect in identifying people who are at imminent risk of becoming diabetic, and reinforces the notion that any treatment we provide to them in the name of diabetes prevention should be free from risk of harm. Rather than applying a long-term medication with potentially harmful side effects to a large group of at-risk patients, more research is needed to provide tools for clinicians to think carefully about which of their patients are truly at highest risk of going on to develop diabetes in the near future.

Applications for Clinical Practice

Although clinicians ought not use PHEN/TPM ER exclusively for diabetes prevention based on the results from this trial, delay of diabetes onset is a possible and important benefit of the use of PHEN/TPM ER in obese patients, provided that they are willing to also make and sustain lifestyle changes in order to lose a clinically significant amount of weight.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine the effect of phentermine and topiramate extended release (PHEN/TPM ER) treatment on progression to type 2 diabetes and/or cardiometabolic disease in subjects with prediabetes and/or metabolic syndrome (MetS) at baseline.

Design. Sub-group analysis of a larger double-blind, randomized, placebo-controlled trial of PHEN/TPM ER in overweight and obese adults.

Setting and participants. The larger study had 2 phases —a 56-week weight loss trial (CONQUER, n = 866), followed by an extension of the drug trial out to 108 weeks (SEQUEL, n = 675) in a sub-group of CONQUER participants. The CONQUER trial, based at 93 U.S. centers, enrolled overweight or obese patients with at least 2 obesity-related comorbidities and randomly assigned them to receive either placebo or PHEN/TPM ER at a lower (7.5 mg/46 mg) or higher (15 mg/92 mg) daily dose. All 3 groups also received lifestyle modification counseling that included an evidence-based diet and exercise curriculum. Participants received study drug and lifestyle counseling in the setting of monthly visits during the 60- (CONQUER) or 108-week (SEQUEL) follow-up period.

The analyses presented in this paper focus on the 475 participants who completed both CONQUER and SEQUEL and who were characterized as pre-diabetic or as having the metabolic syndrome (MetS) at baseline. Pre-diabetes was defined as having a blood glucose level of 100–125 mg/dL or higher while fasting, or 140–199 mg/dL after an oral glucose tolerance test (GTT). MetS was characterized in participants who displayed 3 or more of the following at baseline: waist circumference ≥ 102 cm in men or 88 cm in women; triglycerides ≥ 150 mg/dL or on a lipid-lowering medication; HDL < 40 mg/dL in men or < 50 mg/dL in women; systolic BP ≥ 130 mm Hg or diastolic BP ≥ 85 mm Hg (or on antihypertensive); and fasting glucose ≥ 100 mg/dL or on treatment for elevated glucose.

Main outcome measures. The primary outcome for this study was percent weight loss at 108 weeks of follow-up (or early termination). Secondary outcomes included cardiometabolic changes, such as development of type 2 diabetes and changes in lipid measures, blood pressure, and waist circumference. These were assessed at baseline, week 56, and week 108 (or at early termination). Rates of progression to type 2 diabetes were compared between the treatment groups using chi-square testing. Intention-to-treat (ITT) ANCOVA analysis was performed with multiple imputation techniques to address missing data, as well as with an alternative analysis using last observation carried forward.

Results. The study arms were similar with respect to baseline characteristics. Average age was 51 years in the high dose PHEN/TPM ER arm and 52 in the other arms. Over half (65%) of participants were women and 86% were Caucasian. Mean BMI was 36 kg/m² (class II obesity). Over half of participants were on antihypertensive medications at baseline but with well-controlled blood pressure (mean 128/80 mm Hg). Of the 475 people in this analysis, 316 met criteria for prediabetes, 451 for MetS, and 292 for both prediabetes and MetS.

Weight loss at 2 years was significantly greater in subjects taking PHEN/TPM ER (10.9% in the lower dose group, 12.1% in the higher dose group) compared to those taking placebo (2.5%) (P < 0.001). Mirroring weight loss results, type 2 diabetes incidence was also significantly lower in the drug treatment arms than in the placebo arm at 2 years after randomization—annualized incidence was 6.1% for placebo vs. 1.8% for lower-dose drug and 1.3% for higher-dose drug (P < 0.05). Greater weight loss was associated with greater decrease in diabetes incidence across all 3 arms of the study. Those persons who did not achieve at least a 5% weight loss at 2 years had the highest annualized risk of developing diabetes (6.3%), compared with a 0.9% risk among those who lost at least 15% of their weight. Improvements in other cardiometabolic parameters, including HDL, triglycerides, waist circumference, and insulin sensitivity index, was more common among the PHEN/TPM ER participants compared with placebo. Blood pressure decreased slightly for all 3 groups and there was no significant difference between the drug arms and the placebo arm.

Discontinuation of study medication occurred in all 3 groups (3.1% in placebo, 6.1% in lower-dose medication, and 5.5% in higher-dose medication), with serious adverse events in 5%, 7%, and 8.5%, respectively. There were no deaths.

Conclusion. PHEN/TPM ER administered over a 2-year period significantly improved weight loss and decreased progression to type 2 diabetes relative to placebo in a group of at-risk participants.

Commentary

Diabetes and related cardiometabolic disease are major contributors to morbidity and mortality in adults. With the exception of invasive treatments such as bariatric surgery, reversal of diabetes once it is established has proven quite difficult [1,2], and thus there is an increased emphasis from the public health and medical communities on preventing the development of this disease in the first place. Complicating the picture, recently broadened criteria for pre-diabetes will likely result in a very large number of these at-risk individuals being identified [3,4]. Although intensive lifestyle interventions resulting in a 5% to 7% weight loss among pre-diabetics have been shown to delay progression to diabetes [5], the translation of these programs into real-world settings has, so far, shown less promise than the original randomized trials might have indicated [4]. Although there is ongoing work to try to improve results and uptake in community-based lifestyle intervention programs, for many patients and clinicians these resource-intensive programs currently prove difficult to do well on a large scale.

Alternative methods of helping patients achieve and maintain that critical > 5% weight loss are desperately needed, not only for preventing diabetes, but also for impacting the numerous other risks associated with obesity. This particular trial capitalized on the notion that it is probably successful weight loss, not the intervention format used to achieve that weight loss, which drives decreased diabetes risk. This study was a sub-analysis of a larger randomized trial, and many of the strengths of that larger study are therefore reflected in this paper. Participants and study staff were blinded to treatment arm with the use of placebo, a very important strength when adverse reactions and drug intolerances need to be measured. Furthermore, this likely equalized motivation to comply with the lifestyle recommendations across the treatment arms—this might not have been the case if patients were aware that they were or were not receiving study drug. Another key strength of the study is its duration. PHEN/TPM ER is unique in that it is approved by the FDA for long-term use. Whereas many studies of weight loss show maximum intervention effect at about 6 months followed by weight regain, this study showed sustained weight loss up to 2 years after starting therapy, presumably because participants could actually continue the therapy for the full 2 years. Most importantly, the intervention itself (medication plus low-intensity lifestyle counseling) is likely highly replicable in clinical practice.

There are some important limitations to consider when interpreting the results from this study. First, the participants analyzed in this paper were comprised entirely of people who had already participated in a full year of the parent study and therefore probably represent a sub-group that might have been experiencing greater success as a result of their participation, potentially generating an overly optimistic estimate of weight loss and health effect for all of the groups relative to what might be seen in a general population. This feature of the design also limits this study’s ability to comment on drug intolerance or early adverse reactions—those who didn’t stick with the pills for at least a year would not have been included in these analyses. In terms of generalizability, although the infrastructure required from a clinical standpoint is much lower for an intervention like this (prescribing a medication) compared to an intensive lifestyle intervention, these drugs are still costly, and many insurers/providers may not offer them on formulary. Thus, to realize the long-term benefits of sustained weight loss, patients may need to face significant out-of-pocket costs, which may limit uptake of this therapy to those with financial means. For this and other reasons, it will be important to do future studies looking at how quickly weight is re-gained once people stop taking the medication. Another threat to generalizability is the racial makeup of the participants—the vast majority of them were non-Hispanic white. Furthermore, although a majority of the participants had hypertension, it was well-controlled in all (a prerequisite for taking the medication), and it is unclear whether in a real-world patient population hypertension would be adequately controlled in a large number of patients.

Another issue to consider when looking at the use of weight loss medications for prevention of diabetes is the relative risk of prolonged medication use compared with the risk for developing diabetes. Clearly, for obese patients who are interested in losing weight for other reasons, prevention of diabetes is a wonderful side effect of achieving that goal. However, it is worth noting that even in the highest-risk group of participants in this study (those who lost < 5% of weight), the annualized risk of developing diabetes was about 6% (< 20% cumulative risk projected over 3 years). Compare this to the 7% to 8% serious adverse event rate observed in those on drug therapy. Although the medication did reduce annualized diabetes risk significantly, the vast majority of people in all the arms did not develop diabetes during follow-up. This drives home the point that our current categorization of pre-diabetes is far from perfect in identifying people who are at imminent risk of becoming diabetic, and reinforces the notion that any treatment we provide to them in the name of diabetes prevention should be free from risk of harm. Rather than applying a long-term medication with potentially harmful side effects to a large group of at-risk patients, more research is needed to provide tools for clinicians to think carefully about which of their patients are truly at highest risk of going on to develop diabetes in the near future.

Applications for Clinical Practice

Although clinicians ought not use PHEN/TPM ER exclusively for diabetes prevention based on the results from this trial, delay of diabetes onset is a possible and important benefit of the use of PHEN/TPM ER in obese patients, provided that they are willing to also make and sustain lifestyle changes in order to lose a clinically significant amount of weight.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine the effect of phentermine and topiramate extended release (PHEN/TPM ER) treatment on progression to type 2 diabetes and/or cardiometabolic disease in subjects with prediabetes and/or metabolic syndrome (MetS) at baseline.

Design. Sub-group analysis of a larger double-blind, randomized, placebo-controlled trial of PHEN/TPM ER in overweight and obese adults.

Setting and participants. The larger study had 2 phases —a 56-week weight loss trial (CONQUER, n = 866), followed by an extension of the drug trial out to 108 weeks (SEQUEL, n = 675) in a sub-group of CONQUER participants. The CONQUER trial, based at 93 U.S. centers, enrolled overweight or obese patients with at least 2 obesity-related comorbidities and randomly assigned them to receive either placebo or PHEN/TPM ER at a lower (7.5 mg/46 mg) or higher (15 mg/92 mg) daily dose. All 3 groups also received lifestyle modification counseling that included an evidence-based diet and exercise curriculum. Participants received study drug and lifestyle counseling in the setting of monthly visits during the 60- (CONQUER) or 108-week (SEQUEL) follow-up period.

The analyses presented in this paper focus on the 475 participants who completed both CONQUER and SEQUEL and who were characterized as pre-diabetic or as having the metabolic syndrome (MetS) at baseline. Pre-diabetes was defined as having a blood glucose level of 100–125 mg/dL or higher while fasting, or 140–199 mg/dL after an oral glucose tolerance test (GTT). MetS was characterized in participants who displayed 3 or more of the following at baseline: waist circumference ≥ 102 cm in men or 88 cm in women; triglycerides ≥ 150 mg/dL or on a lipid-lowering medication; HDL < 40 mg/dL in men or < 50 mg/dL in women; systolic BP ≥ 130 mm Hg or diastolic BP ≥ 85 mm Hg (or on antihypertensive); and fasting glucose ≥ 100 mg/dL or on treatment for elevated glucose.

Main outcome measures. The primary outcome for this study was percent weight loss at 108 weeks of follow-up (or early termination). Secondary outcomes included cardiometabolic changes, such as development of type 2 diabetes and changes in lipid measures, blood pressure, and waist circumference. These were assessed at baseline, week 56, and week 108 (or at early termination). Rates of progression to type 2 diabetes were compared between the treatment groups using chi-square testing. Intention-to-treat (ITT) ANCOVA analysis was performed with multiple imputation techniques to address missing data, as well as with an alternative analysis using last observation carried forward.

Results. The study arms were similar with respect to baseline characteristics. Average age was 51 years in the high dose PHEN/TPM ER arm and 52 in the other arms. Over half (65%) of participants were women and 86% were Caucasian. Mean BMI was 36 kg/m² (class II obesity). Over half of participants were on antihypertensive medications at baseline but with well-controlled blood pressure (mean 128/80 mm Hg). Of the 475 people in this analysis, 316 met criteria for prediabetes, 451 for MetS, and 292 for both prediabetes and MetS.

Weight loss at 2 years was significantly greater in subjects taking PHEN/TPM ER (10.9% in the lower dose group, 12.1% in the higher dose group) compared to those taking placebo (2.5%) (P < 0.001). Mirroring weight loss results, type 2 diabetes incidence was also significantly lower in the drug treatment arms than in the placebo arm at 2 years after randomization—annualized incidence was 6.1% for placebo vs. 1.8% for lower-dose drug and 1.3% for higher-dose drug (P < 0.05). Greater weight loss was associated with greater decrease in diabetes incidence across all 3 arms of the study. Those persons who did not achieve at least a 5% weight loss at 2 years had the highest annualized risk of developing diabetes (6.3%), compared with a 0.9% risk among those who lost at least 15% of their weight. Improvements in other cardiometabolic parameters, including HDL, triglycerides, waist circumference, and insulin sensitivity index, was more common among the PHEN/TPM ER participants compared with placebo. Blood pressure decreased slightly for all 3 groups and there was no significant difference between the drug arms and the placebo arm.

Discontinuation of study medication occurred in all 3 groups (3.1% in placebo, 6.1% in lower-dose medication, and 5.5% in higher-dose medication), with serious adverse events in 5%, 7%, and 8.5%, respectively. There were no deaths.

Conclusion. PHEN/TPM ER administered over a 2-year period significantly improved weight loss and decreased progression to type 2 diabetes relative to placebo in a group of at-risk participants.

Commentary

Diabetes and related cardiometabolic disease are major contributors to morbidity and mortality in adults. With the exception of invasive treatments such as bariatric surgery, reversal of diabetes once it is established has proven quite difficult [1,2], and thus there is an increased emphasis from the public health and medical communities on preventing the development of this disease in the first place. Complicating the picture, recently broadened criteria for pre-diabetes will likely result in a very large number of these at-risk individuals being identified [3,4]. Although intensive lifestyle interventions resulting in a 5% to 7% weight loss among pre-diabetics have been shown to delay progression to diabetes [5], the translation of these programs into real-world settings has, so far, shown less promise than the original randomized trials might have indicated [4]. Although there is ongoing work to try to improve results and uptake in community-based lifestyle intervention programs, for many patients and clinicians these resource-intensive programs currently prove difficult to do well on a large scale.

Alternative methods of helping patients achieve and maintain that critical > 5% weight loss are desperately needed, not only for preventing diabetes, but also for impacting the numerous other risks associated with obesity. This particular trial capitalized on the notion that it is probably successful weight loss, not the intervention format used to achieve that weight loss, which drives decreased diabetes risk. This study was a sub-analysis of a larger randomized trial, and many of the strengths of that larger study are therefore reflected in this paper. Participants and study staff were blinded to treatment arm with the use of placebo, a very important strength when adverse reactions and drug intolerances need to be measured. Furthermore, this likely equalized motivation to comply with the lifestyle recommendations across the treatment arms—this might not have been the case if patients were aware that they were or were not receiving study drug. Another key strength of the study is its duration. PHEN/TPM ER is unique in that it is approved by the FDA for long-term use. Whereas many studies of weight loss show maximum intervention effect at about 6 months followed by weight regain, this study showed sustained weight loss up to 2 years after starting therapy, presumably because participants could actually continue the therapy for the full 2 years. Most importantly, the intervention itself (medication plus low-intensity lifestyle counseling) is likely highly replicable in clinical practice.

There are some important limitations to consider when interpreting the results from this study. First, the participants analyzed in this paper were comprised entirely of people who had already participated in a full year of the parent study and therefore probably represent a sub-group that might have been experiencing greater success as a result of their participation, potentially generating an overly optimistic estimate of weight loss and health effect for all of the groups relative to what might be seen in a general population. This feature of the design also limits this study’s ability to comment on drug intolerance or early adverse reactions—those who didn’t stick with the pills for at least a year would not have been included in these analyses. In terms of generalizability, although the infrastructure required from a clinical standpoint is much lower for an intervention like this (prescribing a medication) compared to an intensive lifestyle intervention, these drugs are still costly, and many insurers/providers may not offer them on formulary. Thus, to realize the long-term benefits of sustained weight loss, patients may need to face significant out-of-pocket costs, which may limit uptake of this therapy to those with financial means. For this and other reasons, it will be important to do future studies looking at how quickly weight is re-gained once people stop taking the medication. Another threat to generalizability is the racial makeup of the participants—the vast majority of them were non-Hispanic white. Furthermore, although a majority of the participants had hypertension, it was well-controlled in all (a prerequisite for taking the medication), and it is unclear whether in a real-world patient population hypertension would be adequately controlled in a large number of patients.

Another issue to consider when looking at the use of weight loss medications for prevention of diabetes is the relative risk of prolonged medication use compared with the risk for developing diabetes. Clearly, for obese patients who are interested in losing weight for other reasons, prevention of diabetes is a wonderful side effect of achieving that goal. However, it is worth noting that even in the highest-risk group of participants in this study (those who lost < 5% of weight), the annualized risk of developing diabetes was about 6% (< 20% cumulative risk projected over 3 years). Compare this to the 7% to 8% serious adverse event rate observed in those on drug therapy. Although the medication did reduce annualized diabetes risk significantly, the vast majority of people in all the arms did not develop diabetes during follow-up. This drives home the point that our current categorization of pre-diabetes is far from perfect in identifying people who are at imminent risk of becoming diabetic, and reinforces the notion that any treatment we provide to them in the name of diabetes prevention should be free from risk of harm. Rather than applying a long-term medication with potentially harmful side effects to a large group of at-risk patients, more research is needed to provide tools for clinicians to think carefully about which of their patients are truly at highest risk of going on to develop diabetes in the near future.

Applications for Clinical Practice

Although clinicians ought not use PHEN/TPM ER exclusively for diabetes prevention based on the results from this trial, delay of diabetes onset is a possible and important benefit of the use of PHEN/TPM ER in obese patients, provided that they are willing to also make and sustain lifestyle changes in order to lose a clinically significant amount of weight.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To examine whether an early response (or non-response) to lorcaserin therapy predicts ≥ 5% weight loss achieved at 1 year.

Study design. Secondary analysis of data collected in 3 placebo-controlled blinded randomized trials.

Setting and participants. This study relied upon data collected as part of 3 separate phase 3 clinical trials of lorcaserin, a weight loss drug and selective serotonin 2c (5-HT2c) agonist. The first study, “Behavioral Modification and Lorcaserin for Overweight and Obesity Management” (BLOOM; n = 3182) enrolled overweight (with at least 1 comorbidity) or obese (no comorbidity needed) adult patients (18–65 yr) without diabetes, to determine the safety and efficacy of lorcaserin. The second trial, “Behavioral Modification and Lorcaserin Second Study of Obesity Management” (BLOSSOM; n = 4008) enrolled a similar population as BLOOM. For both BLOOM and BLOSSOM, patients were randomly assigned to receive either lorcaserin (10 mg po bid) or placebo for a 1-year period, and all patients received advice and instruction in exercise goals (at least 30 min/day) and caloric intake (600 kcal less than recommended for weight maintenance for that individual) necessary to promote weight loss. The third trial, BLOOM-DM (n = 604) focused on overweight or obese diabetic patients, but otherwise was similar in methodology to BLOOM and BLOSSOM. All studies took place in multiple US academic and private medical centers and were funded by Arena Pharmaceuticals. For the current analysis, the investigators used data from these trials and classified participants as either “responders” or “non-responders” based on each participant’s early weight loss response to either lorcaserin or placebo.

Main outcome measures. The investigators used area under the curve for the receiver operating characteristic (AUC for ROC) analysis to determine whether an early weight loss response to lorcaserin or placebo predicted a patient’s longer-term (52-week) weight loss. Several steps were used to conduct these analyses.

First, the investigators needed to determine what amount of weight loss at which of several early time-points would qualify a participant as a “responder” to either drug or placebo. They compared weight lost at weeks 2, 4, 8 and 12, using AUC for ROC analysis to identify the appropriate “responder” or “non-responder” cut-points, and classified all participants with data points in these early weeks as such. Second, all of the early responder and non-responder participants with 52-week weight data were then classified as to whether or not they had achieved at least a 5% weight loss at the end of the study. AUC for ROC analysis was again used to determine whether this early categorization was predictive of final study response. In addition to looking at early response as predictive of final weight loss, the investigators also examined the response/non-response variable’s ability to predict other health outcomes, including changes in lipid levels, blood pressure, and, for type 2 diabetic participants, changes in glycemic control (fasting plasma glucose [FPG] and HgbA1c). Finally, the investigators examined the incidence of adverse events in the different groups as well.

Results. The investigators identified a 4.6% weight loss by week 12 on lorcaserin or placebo as the optimal cut-point for determining whether a participant was a “responder” or “non-responder” (“W12R” or “W12NR”). This cut-point had an AUC (95% CI) of 0.849 (0.828–0.870) for predicting ≥ 5% weight loss at 52 weeks, with a positive predictive value (PPV) of 0.855 and negative predictive value (NPV) of 0.740, thus optimizing specificity and sensitivity of the time/weight cut-point compared to those at weeks 2, 4, or 8. Given the need for practical clinical recommendations, however, the investigators used a cut-point of 5% weight loss by week 12 to determine response/non-response for the health outcome analyses. The breakdown of responders vs. non-responders was as follows: For the pooled BLOOM/BLOSSOM participants, there were 1251 lorcaserin-recipient responders and 1286 lorcaserin-recipient non-responders (about 40% of those randomized to lorcaserin were “responders”). Among placebo recipients, there were 541 early responders and 1852 non-responders (about 17% of those randomized to placebo were “responders”). For the diabetic BLOOM-DM participants, the ratios were similar although slightly less favorable, with only about 30% (n = 78) of lorcaserin patients classified W12 responders (139 non-responders), and 10% (n = 25) of placebo patients as W12 responders (192 non-responders).

The lorcaserin and placebo groups in BLOOM and BLOSSOM were similar to one another, with overall mean (SD) age of 43.8 (11.6) years for lorcaserin and 44.0 (11.4) years for placebo. The vast majority of participants in these 2 trials were female (81.7% in lorcaserin arms, 81.0% in placebo), and the majority were non-Hispanic white (67.6% in lorcaserin and 66.2% in placebo). The mean (SD) baseline body mass index (BMI, kg/m²) was 36.1 (4.3) for lorcaserin and 36.1 (4.2) for placebo. The BLOOM-DM participants were also similar in the lorcaserin and placebo arms, although they were older (mean age, 53.2 years lorcaserin, 52 years placebo), and more likely to be female (53.5% lorcaserin, 54.4% placebo). Otherwise, the BLOOM-DM participants were similar on reported demographic characteristics to those in the other 2 trials.

Importantly, however, for all 3 trials there were differences in demographic characteristics between those participants characterized as responders and those characterized as non-responders. Amongst the nondiabetic participants in the BLOOM and BLOSSOM studies, responders (to both lorcaserin and placebo) were more likely to be non-Hispanic white (as opposed to African American or Hispanic participants, who were more likely to be non-responders), and responders were older than non-responders. Interestingly, for the diabetics in the BLOOM-DM trial, the responder/non-responder differences were less pronounced, although the responders were still slightly more likely to be non-Hispanic white and older, particularly for placebo.

Among BLOOM and BLOSSOM participants who received lorcaserin, mean weight loss at 52 weeks was 10.8% among W12Rs and only 2.7% among W12NRs. A similar pattern was observed in the BLOOM and BLOSSOM placebo participants; W12Rs averaged 9.5% weight loss at 52 weeks, versus just 1.1% in W12NRs.  Among diabetics receiving lorcaserin in the BLOOM-DM study, weight loss at 1 year was 9.1% in W12Rs versus 3.1% in W12NRs. Similarly, in placebo-recipients in BLOOM-DM, weight loss at 1 year was 7% for W12Rs and 1.3% for W12NRs. When the weight loss at 1 year was categorized in terms of whether or not participants achieved at least 5% or 10% weight loss, once again early responders to either lorcaserin or placebo had higher rates of achieving both thresholds. Namely, 85.5% of nondiabetic W12Rs had achieved or maintained 5% weight loss at week 52, while only 26% of the W12NRs ultimately did so. Seventy percent of diabetic W12Rs to lorcaserin had ≥ 5% weight loss at week 52 and 25.2% of W12NRs did. The pattern of prediction for achieving 10% weight loss at week 52 was even more pronounced, with, for example, 49.8% of nondiabetic W12Rs having lost at least 10% of their starting weight at 1 year, versus just 4.7% of W12NRs.

When cardiometabolic outcomes were examined, the differences between W12 lorcaserin responders and non-responders appeared to be somewhat attenuated. For example, among diabetic patients, W12 lorcaserin responders had a mean decrease of 1.2% in their A1c level by study end, compared to a nearly 1% decrease in W12NRs. For fasting plasma glucose, the improvement at week 52 was pronounced (about 30 mg/dL lower than baseline) and very similar in W12 responders and non-responders.

Among nondiabetics, average blood pressure lowering (SBP and DBP) at week 52 was greater among lorcaserin W12 responders (SBP dropped 4 mm Hg on average, DBP 3 mm Hg) than it was among non-responders (SBP and DBP dropped by about 1 mm Hg). Other than triglycerides, which decreased substantially among W12 responders (whether on placebo or lorcaserin), changes to lipid profile were relatively small for nondiabetics. Among diabetics, however, LDL and HDL both increased on average in all 4 groups (W12 responders/non-responders to placebo/lorcaserin) by week 52.

Common adverse events for lorcaserin-treated patients included headache (15%–17%), upper respiratory infections (9%–14%), nausea (8%–9%), and dizziness (8% among nondiabetics). Among diabetics, hypoglycemia occurred in 29.3% of those treated with lorcaserin (vs. 21% on placebo). Week 12 responders and non-responders appeared to have a similar adverse event profile, and, in general, adverse events were more common among lorcaserin than placebo participants.

Conclusion. The authors of this study concluded that a week-12 weight loss of ≥ 5% on lorcaserin was a strong predictor of achieving at least that same amount of weight loss, as well as improvements in some cardiometabolic parameters, at 1 year.

Commentary

In 2013, the American Medical Association officially recognized obesity as a disease. This shift in terminology, coupled with a movement towards reimbursing primary care providers for obesity-related interventions, has created a growing awareness among providers that better treatment options for this chronic condition are sorely needed. Just as we treat patients with hypertension and type 2 diabetes by titrating medications, discontinuing those that aren’t effective and continuing those that are, so should we approach the management of our patients with obesity. Although behavioral interventions centered around lifestyle changes (diet/exercise) remain first-line therapies for the treatment of obesity [1], many patients will seek additional tools, such as meal replacement, medication, or even bariatric surgery, to help achieve and maintain weight loss.

In the past 2 to 3 years, there has been a flurry of activity by the FDA to approve new medications for weight loss. In keeping with the view of obesity as a chronic condition, some of these medications, including lorcaserin and phentermine-topiramate ER, have even been approved for patient long-term use [2]. While the addition of new options to the weight loss toolkit is exciting, it may also be daunting for clinicians who have witnessed a bevy of weight loss drugs come on, and then off, the market over the years due to serious adverse events experienced by patients. For physicians and patients considering the use of a new weight loss medication, there is therefore a clear need to minimize risk for adverse effects related to the drug, while maximizing the patient’s chances of losing weight.

Growing evidence from trials of behavioral interventions as well as weight loss medications suggests that the individuals who will ultimately achieve weight loss success with a given intervention/medication, tend to indicate that success relatively early on in the course of therapy [3–5]. For clinicians, this fact is extremely useful, because it may allow the physician and patient to more rapidly make a decision to discontinue a likely ineffective option in favor of another that has not yet been tried, thus minimizing risks for adverse events while maximizing chances of weight loss outcomes.

In this paper, Smith and colleagues addressed this very important issue for one of the more recently FDA-approved medications, lorcaserin. This 5-HT2c agonist is a useful addition to the list of weight loss medications, as it has relatively few contraindications, other than that it cannot be used in pregnancy/lactation and should be avoided in those with a history of heart failure. However, lorcaserin is still relatively costly (eg, compared to phentermine) and, if it is going to be used for long-term weight loss/maintenance, the financial outlay faced by patients might be considerable. In addition to answering an important question, this paper also examined not only weight loss outcomes but also cardiometabolic impacts of the medication. Furthermore, the authors separately examined outcomes for diabetic and nondiabetic patients, as the risk/benefit ratio of remaining on this medication could be quite different between the 2 groups.

Importantly, the study represented a group of secondary analyses of data aggregated from several trials—trials that were not originally designed to answer this question. Although the majority of original trial participants did have data at weeks 12 and 52 (requirement for inclusion in this analysis), up to a quarter of patients in some groups were missing one or the other measure. Whether or not those analyzed represented a biased subsample, and therefore do not have generalizable results, cannot be ascertained.

In reviewing the outcomes achieved by early responders and non-responders, it was very interesting to note that so-called “responders” to placebo followed a nearly identical weight loss trajectory as those on lorcaserin. This fact should not be taken to indicate that lorcaserin is no different from placebo, as the overall chances of achieving weight loss were significantly greater among the lorcaserin participants. However, it is interesting that, for those placebo patients who clearly followed the recommended lifestyle changes, they did just as well as patients receiving active study drug. This underscores the need to educate patients and encourage them, first and foremost, to make a real effort to diet and exercise regardless of what other tools are employed to achieve weight loss.

Another issue to consider for this study is that there are clear differences in the racial/ethnic makeup of responders versus non-responders. This finding is not unexpected, as in many prior weight loss trials, particularly for behavioral interventions, African-American women have experienced less weight loss than their non-Hispanic white counterparts [6]. These differences were observed both for lorcaserin and placebo patients, raising a concern that the lifestyle intervention component of the study was not equally successful for minorities compared to the non-Hispanic white participants. More research is needed on behavioral interventions that work well in diverse populations.

One finding of interest is that among diabetic participants (BLOOM-DM), glycemic control parameters improved nearly equally between lorcaserin early responders and non-responders, despite the differences between those groups for year-end weight loss. The reasons for this are not clear but could merit further investigation.

Ultimately however, even among this large group of randomized trial participants, who were likely highly motivated, only about 40% of nondiabetics and 30% of diabetics were classified as week 12 responders to lorcaserin. That means that likely well over half of the real-world patients who initiate the drug may not achieve their desired weight loss goals with it. Given the cost of the medication, this must be considered before prescribing it, and it reinforces the importance of being willing to reassess a patient’s weight loss progress early and often so that the medication can be discontinued in favor of other therapies as needed.

Applications for Clinical Practice

For providers interested in prescribing lorcaserin to their patients, a clear plan should be made to have regular and early follow-up to assess the patient’s response to the medication. Patients should understand that if they are not responding to the medication within 3 months, or perhaps sooner if they are experiencing any negative side effects, their physician may elect to discontinue it. Importantly, they should only be given lorcaserin if they are also willing to undertake the behavioral changes necessary to promote weight loss, and it should be underscored that their chances of successful weight loss with or without the medication will be greatly enhanced by doing so.

 —Kristina Lewis, MD, MPH

Study Overview

Objective. To examine whether an early response (or non-response) to lorcaserin therapy predicts ≥ 5% weight loss achieved at 1 year.

Study design. Secondary analysis of data collected in 3 placebo-controlled blinded randomized trials.

Setting and participants. This study relied upon data collected as part of 3 separate phase 3 clinical trials of lorcaserin, a weight loss drug and selective serotonin 2c (5-HT2c) agonist. The first study, “Behavioral Modification and Lorcaserin for Overweight and Obesity Management” (BLOOM; n = 3182) enrolled overweight (with at least 1 comorbidity) or obese (no comorbidity needed) adult patients (18–65 yr) without diabetes, to determine the safety and efficacy of lorcaserin. The second trial, “Behavioral Modification and Lorcaserin Second Study of Obesity Management” (BLOSSOM; n = 4008) enrolled a similar population as BLOOM. For both BLOOM and BLOSSOM, patients were randomly assigned to receive either lorcaserin (10 mg po bid) or placebo for a 1-year period, and all patients received advice and instruction in exercise goals (at least 30 min/day) and caloric intake (600 kcal less than recommended for weight maintenance for that individual) necessary to promote weight loss. The third trial, BLOOM-DM (n = 604) focused on overweight or obese diabetic patients, but otherwise was similar in methodology to BLOOM and BLOSSOM. All studies took place in multiple US academic and private medical centers and were funded by Arena Pharmaceuticals. For the current analysis, the investigators used data from these trials and classified participants as either “responders” or “non-responders” based on each participant’s early weight loss response to either lorcaserin or placebo.

Main outcome measures. The investigators used area under the curve for the receiver operating characteristic (AUC for ROC) analysis to determine whether an early weight loss response to lorcaserin or placebo predicted a patient’s longer-term (52-week) weight loss. Several steps were used to conduct these analyses.

First, the investigators needed to determine what amount of weight loss at which of several early time-points would qualify a participant as a “responder” to either drug or placebo. They compared weight lost at weeks 2, 4, 8 and 12, using AUC for ROC analysis to identify the appropriate “responder” or “non-responder” cut-points, and classified all participants with data points in these early weeks as such. Second, all of the early responder and non-responder participants with 52-week weight data were then classified as to whether or not they had achieved at least a 5% weight loss at the end of the study. AUC for ROC analysis was again used to determine whether this early categorization was predictive of final study response. In addition to looking at early response as predictive of final weight loss, the investigators also examined the response/non-response variable’s ability to predict other health outcomes, including changes in lipid levels, blood pressure, and, for type 2 diabetic participants, changes in glycemic control (fasting plasma glucose [FPG] and HgbA1c). Finally, the investigators examined the incidence of adverse events in the different groups as well.

Results. The investigators identified a 4.6% weight loss by week 12 on lorcaserin or placebo as the optimal cut-point for determining whether a participant was a “responder” or “non-responder” (“W12R” or “W12NR”). This cut-point had an AUC (95% CI) of 0.849 (0.828–0.870) for predicting ≥ 5% weight loss at 52 weeks, with a positive predictive value (PPV) of 0.855 and negative predictive value (NPV) of 0.740, thus optimizing specificity and sensitivity of the time/weight cut-point compared to those at weeks 2, 4, or 8. Given the need for practical clinical recommendations, however, the investigators used a cut-point of 5% weight loss by week 12 to determine response/non-response for the health outcome analyses. The breakdown of responders vs. non-responders was as follows: For the pooled BLOOM/BLOSSOM participants, there were 1251 lorcaserin-recipient responders and 1286 lorcaserin-recipient non-responders (about 40% of those randomized to lorcaserin were “responders”). Among placebo recipients, there were 541 early responders and 1852 non-responders (about 17% of those randomized to placebo were “responders”). For the diabetic BLOOM-DM participants, the ratios were similar although slightly less favorable, with only about 30% (n = 78) of lorcaserin patients classified W12 responders (139 non-responders), and 10% (n = 25) of placebo patients as W12 responders (192 non-responders).

The lorcaserin and placebo groups in BLOOM and BLOSSOM were similar to one another, with overall mean (SD) age of 43.8 (11.6) years for lorcaserin and 44.0 (11.4) years for placebo. The vast majority of participants in these 2 trials were female (81.7% in lorcaserin arms, 81.0% in placebo), and the majority were non-Hispanic white (67.6% in lorcaserin and 66.2% in placebo). The mean (SD) baseline body mass index (BMI, kg/m²) was 36.1 (4.3) for lorcaserin and 36.1 (4.2) for placebo. The BLOOM-DM participants were also similar in the lorcaserin and placebo arms, although they were older (mean age, 53.2 years lorcaserin, 52 years placebo), and more likely to be female (53.5% lorcaserin, 54.4% placebo). Otherwise, the BLOOM-DM participants were similar on reported demographic characteristics to those in the other 2 trials.

Importantly, however, for all 3 trials there were differences in demographic characteristics between those participants characterized as responders and those characterized as non-responders. Amongst the nondiabetic participants in the BLOOM and BLOSSOM studies, responders (to both lorcaserin and placebo) were more likely to be non-Hispanic white (as opposed to African American or Hispanic participants, who were more likely to be non-responders), and responders were older than non-responders. Interestingly, for the diabetics in the BLOOM-DM trial, the responder/non-responder differences were less pronounced, although the responders were still slightly more likely to be non-Hispanic white and older, particularly for placebo.

Among BLOOM and BLOSSOM participants who received lorcaserin, mean weight loss at 52 weeks was 10.8% among W12Rs and only 2.7% among W12NRs. A similar pattern was observed in the BLOOM and BLOSSOM placebo participants; W12Rs averaged 9.5% weight loss at 52 weeks, versus just 1.1% in W12NRs.  Among diabetics receiving lorcaserin in the BLOOM-DM study, weight loss at 1 year was 9.1% in W12Rs versus 3.1% in W12NRs. Similarly, in placebo-recipients in BLOOM-DM, weight loss at 1 year was 7% for W12Rs and 1.3% for W12NRs. When the weight loss at 1 year was categorized in terms of whether or not participants achieved at least 5% or 10% weight loss, once again early responders to either lorcaserin or placebo had higher rates of achieving both thresholds. Namely, 85.5% of nondiabetic W12Rs had achieved or maintained 5% weight loss at week 52, while only 26% of the W12NRs ultimately did so. Seventy percent of diabetic W12Rs to lorcaserin had ≥ 5% weight loss at week 52 and 25.2% of W12NRs did. The pattern of prediction for achieving 10% weight loss at week 52 was even more pronounced, with, for example, 49.8% of nondiabetic W12Rs having lost at least 10% of their starting weight at 1 year, versus just 4.7% of W12NRs.

When cardiometabolic outcomes were examined, the differences between W12 lorcaserin responders and non-responders appeared to be somewhat attenuated. For example, among diabetic patients, W12 lorcaserin responders had a mean decrease of 1.2% in their A1c level by study end, compared to a nearly 1% decrease in W12NRs. For fasting plasma glucose, the improvement at week 52 was pronounced (about 30 mg/dL lower than baseline) and very similar in W12 responders and non-responders.

Among nondiabetics, average blood pressure lowering (SBP and DBP) at week 52 was greater among lorcaserin W12 responders (SBP dropped 4 mm Hg on average, DBP 3 mm Hg) than it was among non-responders (SBP and DBP dropped by about 1 mm Hg). Other than triglycerides, which decreased substantially among W12 responders (whether on placebo or lorcaserin), changes to lipid profile were relatively small for nondiabetics. Among diabetics, however, LDL and HDL both increased on average in all 4 groups (W12 responders/non-responders to placebo/lorcaserin) by week 52.

Common adverse events for lorcaserin-treated patients included headache (15%–17%), upper respiratory infections (9%–14%), nausea (8%–9%), and dizziness (8% among nondiabetics). Among diabetics, hypoglycemia occurred in 29.3% of those treated with lorcaserin (vs. 21% on placebo). Week 12 responders and non-responders appeared to have a similar adverse event profile, and, in general, adverse events were more common among lorcaserin than placebo participants.

Conclusion. The authors of this study concluded that a week-12 weight loss of ≥ 5% on lorcaserin was a strong predictor of achieving at least that same amount of weight loss, as well as improvements in some cardiometabolic parameters, at 1 year.

Commentary

In 2013, the American Medical Association officially recognized obesity as a disease. This shift in terminology, coupled with a movement towards reimbursing primary care providers for obesity-related interventions, has created a growing awareness among providers that better treatment options for this chronic condition are sorely needed. Just as we treat patients with hypertension and type 2 diabetes by titrating medications, discontinuing those that aren’t effective and continuing those that are, so should we approach the management of our patients with obesity. Although behavioral interventions centered around lifestyle changes (diet/exercise) remain first-line therapies for the treatment of obesity [1], many patients will seek additional tools, such as meal replacement, medication, or even bariatric surgery, to help achieve and maintain weight loss.

In the past 2 to 3 years, there has been a flurry of activity by the FDA to approve new medications for weight loss. In keeping with the view of obesity as a chronic condition, some of these medications, including lorcaserin and phentermine-topiramate ER, have even been approved for patient long-term use [2]. While the addition of new options to the weight loss toolkit is exciting, it may also be daunting for clinicians who have witnessed a bevy of weight loss drugs come on, and then off, the market over the years due to serious adverse events experienced by patients. For physicians and patients considering the use of a new weight loss medication, there is therefore a clear need to minimize risk for adverse effects related to the drug, while maximizing the patient’s chances of losing weight.

Growing evidence from trials of behavioral interventions as well as weight loss medications suggests that the individuals who will ultimately achieve weight loss success with a given intervention/medication, tend to indicate that success relatively early on in the course of therapy [3–5]. For clinicians, this fact is extremely useful, because it may allow the physician and patient to more rapidly make a decision to discontinue a likely ineffective option in favor of another that has not yet been tried, thus minimizing risks for adverse events while maximizing chances of weight loss outcomes.

In this paper, Smith and colleagues addressed this very important issue for one of the more recently FDA-approved medications, lorcaserin. This 5-HT2c agonist is a useful addition to the list of weight loss medications, as it has relatively few contraindications, other than that it cannot be used in pregnancy/lactation and should be avoided in those with a history of heart failure. However, lorcaserin is still relatively costly (eg, compared to phentermine) and, if it is going to be used for long-term weight loss/maintenance, the financial outlay faced by patients might be considerable. In addition to answering an important question, this paper also examined not only weight loss outcomes but also cardiometabolic impacts of the medication. Furthermore, the authors separately examined outcomes for diabetic and nondiabetic patients, as the risk/benefit ratio of remaining on this medication could be quite different between the 2 groups.

Importantly, the study represented a group of secondary analyses of data aggregated from several trials—trials that were not originally designed to answer this question. Although the majority of original trial participants did have data at weeks 12 and 52 (requirement for inclusion in this analysis), up to a quarter of patients in some groups were missing one or the other measure. Whether or not those analyzed represented a biased subsample, and therefore do not have generalizable results, cannot be ascertained.

In reviewing the outcomes achieved by early responders and non-responders, it was very interesting to note that so-called “responders” to placebo followed a nearly identical weight loss trajectory as those on lorcaserin. This fact should not be taken to indicate that lorcaserin is no different from placebo, as the overall chances of achieving weight loss were significantly greater among the lorcaserin participants. However, it is interesting that, for those placebo patients who clearly followed the recommended lifestyle changes, they did just as well as patients receiving active study drug. This underscores the need to educate patients and encourage them, first and foremost, to make a real effort to diet and exercise regardless of what other tools are employed to achieve weight loss.

Another issue to consider for this study is that there are clear differences in the racial/ethnic makeup of responders versus non-responders. This finding is not unexpected, as in many prior weight loss trials, particularly for behavioral interventions, African-American women have experienced less weight loss than their non-Hispanic white counterparts [6]. These differences were observed both for lorcaserin and placebo patients, raising a concern that the lifestyle intervention component of the study was not equally successful for minorities compared to the non-Hispanic white participants. More research is needed on behavioral interventions that work well in diverse populations.

One finding of interest is that among diabetic participants (BLOOM-DM), glycemic control parameters improved nearly equally between lorcaserin early responders and non-responders, despite the differences between those groups for year-end weight loss. The reasons for this are not clear but could merit further investigation.

Ultimately however, even among this large group of randomized trial participants, who were likely highly motivated, only about 40% of nondiabetics and 30% of diabetics were classified as week 12 responders to lorcaserin. That means that likely well over half of the real-world patients who initiate the drug may not achieve their desired weight loss goals with it. Given the cost of the medication, this must be considered before prescribing it, and it reinforces the importance of being willing to reassess a patient’s weight loss progress early and often so that the medication can be discontinued in favor of other therapies as needed.

Applications for Clinical Practice

For providers interested in prescribing lorcaserin to their patients, a clear plan should be made to have regular and early follow-up to assess the patient’s response to the medication. Patients should understand that if they are not responding to the medication within 3 months, or perhaps sooner if they are experiencing any negative side effects, their physician may elect to discontinue it. Importantly, they should only be given lorcaserin if they are also willing to undertake the behavioral changes necessary to promote weight loss, and it should be underscored that their chances of successful weight loss with or without the medication will be greatly enhanced by doing so.

 —Kristina Lewis, MD, MPH

Study Overview

Objective. To examine whether an early response (or non-response) to lorcaserin therapy predicts ≥ 5% weight loss achieved at 1 year.

Study design. Secondary analysis of data collected in 3 placebo-controlled blinded randomized trials.

Setting and participants. This study relied upon data collected as part of 3 separate phase 3 clinical trials of lorcaserin, a weight loss drug and selective serotonin 2c (5-HT2c) agonist. The first study, “Behavioral Modification and Lorcaserin for Overweight and Obesity Management” (BLOOM; n = 3182) enrolled overweight (with at least 1 comorbidity) or obese (no comorbidity needed) adult patients (18–65 yr) without diabetes, to determine the safety and efficacy of lorcaserin. The second trial, “Behavioral Modification and Lorcaserin Second Study of Obesity Management” (BLOSSOM; n = 4008) enrolled a similar population as BLOOM. For both BLOOM and BLOSSOM, patients were randomly assigned to receive either lorcaserin (10 mg po bid) or placebo for a 1-year period, and all patients received advice and instruction in exercise goals (at least 30 min/day) and caloric intake (600 kcal less than recommended for weight maintenance for that individual) necessary to promote weight loss. The third trial, BLOOM-DM (n = 604) focused on overweight or obese diabetic patients, but otherwise was similar in methodology to BLOOM and BLOSSOM. All studies took place in multiple US academic and private medical centers and were funded by Arena Pharmaceuticals. For the current analysis, the investigators used data from these trials and classified participants as either “responders” or “non-responders” based on each participant’s early weight loss response to either lorcaserin or placebo.

Main outcome measures. The investigators used area under the curve for the receiver operating characteristic (AUC for ROC) analysis to determine whether an early weight loss response to lorcaserin or placebo predicted a patient’s longer-term (52-week) weight loss. Several steps were used to conduct these analyses.

First, the investigators needed to determine what amount of weight loss at which of several early time-points would qualify a participant as a “responder” to either drug or placebo. They compared weight lost at weeks 2, 4, 8 and 12, using AUC for ROC analysis to identify the appropriate “responder” or “non-responder” cut-points, and classified all participants with data points in these early weeks as such. Second, all of the early responder and non-responder participants with 52-week weight data were then classified as to whether or not they had achieved at least a 5% weight loss at the end of the study. AUC for ROC analysis was again used to determine whether this early categorization was predictive of final study response. In addition to looking at early response as predictive of final weight loss, the investigators also examined the response/non-response variable’s ability to predict other health outcomes, including changes in lipid levels, blood pressure, and, for type 2 diabetic participants, changes in glycemic control (fasting plasma glucose [FPG] and HgbA1c). Finally, the investigators examined the incidence of adverse events in the different groups as well.

Results. The investigators identified a 4.6% weight loss by week 12 on lorcaserin or placebo as the optimal cut-point for determining whether a participant was a “responder” or “non-responder” (“W12R” or “W12NR”). This cut-point had an AUC (95% CI) of 0.849 (0.828–0.870) for predicting ≥ 5% weight loss at 52 weeks, with a positive predictive value (PPV) of 0.855 and negative predictive value (NPV) of 0.740, thus optimizing specificity and sensitivity of the time/weight cut-point compared to those at weeks 2, 4, or 8. Given the need for practical clinical recommendations, however, the investigators used a cut-point of 5% weight loss by week 12 to determine response/non-response for the health outcome analyses. The breakdown of responders vs. non-responders was as follows: For the pooled BLOOM/BLOSSOM participants, there were 1251 lorcaserin-recipient responders and 1286 lorcaserin-recipient non-responders (about 40% of those randomized to lorcaserin were “responders”). Among placebo recipients, there were 541 early responders and 1852 non-responders (about 17% of those randomized to placebo were “responders”). For the diabetic BLOOM-DM participants, the ratios were similar although slightly less favorable, with only about 30% (n = 78) of lorcaserin patients classified W12 responders (139 non-responders), and 10% (n = 25) of placebo patients as W12 responders (192 non-responders).

The lorcaserin and placebo groups in BLOOM and BLOSSOM were similar to one another, with overall mean (SD) age of 43.8 (11.6) years for lorcaserin and 44.0 (11.4) years for placebo. The vast majority of participants in these 2 trials were female (81.7% in lorcaserin arms, 81.0% in placebo), and the majority were non-Hispanic white (67.6% in lorcaserin and 66.2% in placebo). The mean (SD) baseline body mass index (BMI, kg/m²) was 36.1 (4.3) for lorcaserin and 36.1 (4.2) for placebo. The BLOOM-DM participants were also similar in the lorcaserin and placebo arms, although they were older (mean age, 53.2 years lorcaserin, 52 years placebo), and more likely to be female (53.5% lorcaserin, 54.4% placebo). Otherwise, the BLOOM-DM participants were similar on reported demographic characteristics to those in the other 2 trials.

Importantly, however, for all 3 trials there were differences in demographic characteristics between those participants characterized as responders and those characterized as non-responders. Amongst the nondiabetic participants in the BLOOM and BLOSSOM studies, responders (to both lorcaserin and placebo) were more likely to be non-Hispanic white (as opposed to African American or Hispanic participants, who were more likely to be non-responders), and responders were older than non-responders. Interestingly, for the diabetics in the BLOOM-DM trial, the responder/non-responder differences were less pronounced, although the responders were still slightly more likely to be non-Hispanic white and older, particularly for placebo.

Among BLOOM and BLOSSOM participants who received lorcaserin, mean weight loss at 52 weeks was 10.8% among W12Rs and only 2.7% among W12NRs. A similar pattern was observed in the BLOOM and BLOSSOM placebo participants; W12Rs averaged 9.5% weight loss at 52 weeks, versus just 1.1% in W12NRs.  Among diabetics receiving lorcaserin in the BLOOM-DM study, weight loss at 1 year was 9.1% in W12Rs versus 3.1% in W12NRs. Similarly, in placebo-recipients in BLOOM-DM, weight loss at 1 year was 7% for W12Rs and 1.3% for W12NRs. When the weight loss at 1 year was categorized in terms of whether or not participants achieved at least 5% or 10% weight loss, once again early responders to either lorcaserin or placebo had higher rates of achieving both thresholds. Namely, 85.5% of nondiabetic W12Rs had achieved or maintained 5% weight loss at week 52, while only 26% of the W12NRs ultimately did so. Seventy percent of diabetic W12Rs to lorcaserin had ≥ 5% weight loss at week 52 and 25.2% of W12NRs did. The pattern of prediction for achieving 10% weight loss at week 52 was even more pronounced, with, for example, 49.8% of nondiabetic W12Rs having lost at least 10% of their starting weight at 1 year, versus just 4.7% of W12NRs.

When cardiometabolic outcomes were examined, the differences between W12 lorcaserin responders and non-responders appeared to be somewhat attenuated. For example, among diabetic patients, W12 lorcaserin responders had a mean decrease of 1.2% in their A1c level by study end, compared to a nearly 1% decrease in W12NRs. For fasting plasma glucose, the improvement at week 52 was pronounced (about 30 mg/dL lower than baseline) and very similar in W12 responders and non-responders.

Among nondiabetics, average blood pressure lowering (SBP and DBP) at week 52 was greater among lorcaserin W12 responders (SBP dropped 4 mm Hg on average, DBP 3 mm Hg) than it was among non-responders (SBP and DBP dropped by about 1 mm Hg). Other than triglycerides, which decreased substantially among W12 responders (whether on placebo or lorcaserin), changes to lipid profile were relatively small for nondiabetics. Among diabetics, however, LDL and HDL both increased on average in all 4 groups (W12 responders/non-responders to placebo/lorcaserin) by week 52.

Common adverse events for lorcaserin-treated patients included headache (15%–17%), upper respiratory infections (9%–14%), nausea (8%–9%), and dizziness (8% among nondiabetics). Among diabetics, hypoglycemia occurred in 29.3% of those treated with lorcaserin (vs. 21% on placebo). Week 12 responders and non-responders appeared to have a similar adverse event profile, and, in general, adverse events were more common among lorcaserin than placebo participants.

Conclusion. The authors of this study concluded that a week-12 weight loss of ≥ 5% on lorcaserin was a strong predictor of achieving at least that same amount of weight loss, as well as improvements in some cardiometabolic parameters, at 1 year.

Commentary

In 2013, the American Medical Association officially recognized obesity as a disease. This shift in terminology, coupled with a movement towards reimbursing primary care providers for obesity-related interventions, has created a growing awareness among providers that better treatment options for this chronic condition are sorely needed. Just as we treat patients with hypertension and type 2 diabetes by titrating medications, discontinuing those that aren’t effective and continuing those that are, so should we approach the management of our patients with obesity. Although behavioral interventions centered around lifestyle changes (diet/exercise) remain first-line therapies for the treatment of obesity [1], many patients will seek additional tools, such as meal replacement, medication, or even bariatric surgery, to help achieve and maintain weight loss.

In the past 2 to 3 years, there has been a flurry of activity by the FDA to approve new medications for weight loss. In keeping with the view of obesity as a chronic condition, some of these medications, including lorcaserin and phentermine-topiramate ER, have even been approved for patient long-term use [2]. While the addition of new options to the weight loss toolkit is exciting, it may also be daunting for clinicians who have witnessed a bevy of weight loss drugs come on, and then off, the market over the years due to serious adverse events experienced by patients. For physicians and patients considering the use of a new weight loss medication, there is therefore a clear need to minimize risk for adverse effects related to the drug, while maximizing the patient’s chances of losing weight.

Growing evidence from trials of behavioral interventions as well as weight loss medications suggests that the individuals who will ultimately achieve weight loss success with a given intervention/medication, tend to indicate that success relatively early on in the course of therapy [3–5]. For clinicians, this fact is extremely useful, because it may allow the physician and patient to more rapidly make a decision to discontinue a likely ineffective option in favor of another that has not yet been tried, thus minimizing risks for adverse events while maximizing chances of weight loss outcomes.

In this paper, Smith and colleagues addressed this very important issue for one of the more recently FDA-approved medications, lorcaserin. This 5-HT2c agonist is a useful addition to the list of weight loss medications, as it has relatively few contraindications, other than that it cannot be used in pregnancy/lactation and should be avoided in those with a history of heart failure. However, lorcaserin is still relatively costly (eg, compared to phentermine) and, if it is going to be used for long-term weight loss/maintenance, the financial outlay faced by patients might be considerable. In addition to answering an important question, this paper also examined not only weight loss outcomes but also cardiometabolic impacts of the medication. Furthermore, the authors separately examined outcomes for diabetic and nondiabetic patients, as the risk/benefit ratio of remaining on this medication could be quite different between the 2 groups.

Importantly, the study represented a group of secondary analyses of data aggregated from several trials—trials that were not originally designed to answer this question. Although the majority of original trial participants did have data at weeks 12 and 52 (requirement for inclusion in this analysis), up to a quarter of patients in some groups were missing one or the other measure. Whether or not those analyzed represented a biased subsample, and therefore do not have generalizable results, cannot be ascertained.

In reviewing the outcomes achieved by early responders and non-responders, it was very interesting to note that so-called “responders” to placebo followed a nearly identical weight loss trajectory as those on lorcaserin. This fact should not be taken to indicate that lorcaserin is no different from placebo, as the overall chances of achieving weight loss were significantly greater among the lorcaserin participants. However, it is interesting that, for those placebo patients who clearly followed the recommended lifestyle changes, they did just as well as patients receiving active study drug. This underscores the need to educate patients and encourage them, first and foremost, to make a real effort to diet and exercise regardless of what other tools are employed to achieve weight loss.

Another issue to consider for this study is that there are clear differences in the racial/ethnic makeup of responders versus non-responders. This finding is not unexpected, as in many prior weight loss trials, particularly for behavioral interventions, African-American women have experienced less weight loss than their non-Hispanic white counterparts [6]. These differences were observed both for lorcaserin and placebo patients, raising a concern that the lifestyle intervention component of the study was not equally successful for minorities compared to the non-Hispanic white participants. More research is needed on behavioral interventions that work well in diverse populations.

One finding of interest is that among diabetic participants (BLOOM-DM), glycemic control parameters improved nearly equally between lorcaserin early responders and non-responders, despite the differences between those groups for year-end weight loss. The reasons for this are not clear but could merit further investigation.

Ultimately however, even among this large group of randomized trial participants, who were likely highly motivated, only about 40% of nondiabetics and 30% of diabetics were classified as week 12 responders to lorcaserin. That means that likely well over half of the real-world patients who initiate the drug may not achieve their desired weight loss goals with it. Given the cost of the medication, this must be considered before prescribing it, and it reinforces the importance of being willing to reassess a patient’s weight loss progress early and often so that the medication can be discontinued in favor of other therapies as needed.

Applications for Clinical Practice

For providers interested in prescribing lorcaserin to their patients, a clear plan should be made to have regular and early follow-up to assess the patient’s response to the medication. Patients should understand that if they are not responding to the medication within 3 months, or perhaps sooner if they are experiencing any negative side effects, their physician may elect to discontinue it. Importantly, they should only be given lorcaserin if they are also willing to undertake the behavioral changes necessary to promote weight loss, and it should be underscored that their chances of successful weight loss with or without the medication will be greatly enhanced by doing so.

 —Kristina Lewis, MD, MPH

Study Overview

Objective. To determine whether having a body mass index (BMI) in the obese range (30 kg/m²) as an older adult woman is associated with changes in late-age survival and morbidity.

Design. Observational cohort study.

Setting and participants. This study relied upon data collected as part of the Women’s Health Initiative (WHI), an observational study and clinical trial focusing on the health of postmenopausal women aged 50–79 years at enrollment. For the purposes of the WHI, women were recruited from centers across the United States between 1993 and 1998 and could participate in several intervention studies (hormone replacement therapy, low-fat diet, calcium/vitamin D supplementation) or an observational study [1].

For this paper, the authors utilized data from those WHI participants who, based on their age at enrollment, could have reached age 85 years by September of 2012. The authors excluded women who did not provide follow-up health information within 18 months of their 85th birthdays or who reported mobility disabilities at their baseline data collection. This resulted in a total of 36,611 women for analysis.

There were a number of baseline measures collected on the study participants. Via written survey, participants self-reported their race and ethnicity, hormone use status, smoking status, alcohol consumption, physical activity level, depressive symptoms, and a number of demographic characteristics. Study personnel objectively measured height and weight to calculate baseline BMI and also measured waist circumference (WC, in cm).

The primary exposure measure for this study was BMI category at trial entry categorized as follows: underweight (< 18.5 kg/m²), healthy weight (18.5–24.9 kg/m²), overweight (25.0–29.9 kg/m²) or obese class I (30–34.9 kg/m²), II (35–39.9 kg/m²) or III (≥ 40 kg/m²), using standard accepted cut-points except for Asian/Pacific Islander participants, where alternative World Health Organization (WHO) cut-points were used. The WHO cut-points are slightly lower to account for usual body habitus and disease risk in that population. BMI changes over study follow-up were not included in the exposure measure for this study. WC (dichotomized around 88 cm) was also used as an exposure measure.

Main outcome measures. Disease-free survival status during the follow-up period. In the year at which participants were supposed to reach their 85th birthdays, they were categorized as to whether they had survived or not. Survival status was ascertained by hospital record review, autopsy reports, death certificates and review of the National Death Index. Those who survived were sub-grouped according to type of survival into 1 of the following categories: (1) no incident disease and no mobility disability (healthy), (2) baseline disease present but no incident disease or mobility disability during follow-up (prevalent disease), (3) incident disease but no mobility disability during follow-up (incident disease), and (4) incident mobility disability with or without incident disease (disabled).

Diseases of interest (prevalent and incident) included coronary and cerebrovascular disease, cancer, diabetes and hip fracture—the conditions the investigators felt most increased risk of death or morbidity and mobility disability in this population of aging women. Baseline disease status was defined using self-report, but incident disease in follow-up was more rigorously defined using self-report plus medical record review, except for incident diabetes, which required only self-report of diagnosis plus report of new oral hypoglycemic or insulin use.

Because the outcome of interest (survival status) had 5 possible categories, multinomial logistic regression was used as the analytic technique, with baseline BMI category and WC categories as predictors. The authors adjusted for baseline characteristics including age, race/ethnicity, study arm (intervention or observational for WHI), educational level, marital status, smoking status, ethanol use, self-reported physical activity and depression symptoms. Because of the possibly interrelated predictors (BMI and WC), the authors built BMI models with and without WC, and when WC was the primary predictor they adjusted for a participant’s BMI in order to try to isolate the impact of central adiposity. Additionally, they performed the analyses stratified by race and ethnicity as well as by smoking status.

Results. The mean (SD) baseline age of participants was 72.4 (3) years, and the vast majority (88.5%) self-identified as non-Hispanic white. At the end of the follow-up period, of the initial 36,611 participants, 9079 (24.8%) had died, 6702 (18.3%) had become disabled, 8512 (23.2%) had developed incident disease without disability, 5366 (14.6%) had prevalent but no incident disease, and 6952 (18.9%) were categorized as healthy. There were a number of potentially confounding baseline characteristics that differed between the survival categories. Importantly, race was associated with survival status—non-Hispanic white women were more likely to be in the “healthy” category at follow-up than their counterparts from other races/ethnicities. Baseline smokers were more likely not to live to 85 years, and those with less than a high school education were also more likely not to live to 85 years.

In models adjusting for baseline covariates, with BMI category as the primary predictor, women with an obese baseline BMI had significantly increased odds of not living to 85 years of age, relative to women in a healthy baseline BMI category, with increasing odds of death among those with higher baseline BMI levels (class I obesity odds ratio [OR] 1.72 [95% CI 1.55–1.92], class II obesity OR 3.28 [95% CI 2.69–4.01], class III obesity OR 3.48 [95% CI 2.52–4.80]). Amongst survivors, baseline obesity was also associated with greater odds of developing incident disease, relative to healthy weight women (class I obesity OR 1.65 [95% CI 1.48–1.84], class II obesity OR 2.44 (95% CI 2.02–2.96), class III obesity OR 1.73 [95% CI 1.21–2.46]). There was a striking relationship between baseline obesity and the odds of incident disability during follow-up (class I obesity OR 3.22 [95% CI 2.87–3.61], class II obesity OR 6.62 [95% CI 5.41–8.09], class III obesity OR 6.65 [95% CI 4.80–9.21]).

Women who were overweight at baseline also displayed statistically significant but more modestly increased odds of incident disease, mobility disability, and death relative to their normal-weight counterparts. Importantly, even in multivariable models, being underweight at baseline was also associated with significantly increased odds of death before age 85 relative to healthy weight individuals (OR 2.09 [95% CI 1.54–2.85]) but not with increased odds of incident disease or disability.

When WC status was adjusted for in the “BMI-outcome” models, the odds of death, disability, and incident disease were attenuated for obese women but remained elevated, particularly for women with class II or III obesity. When WC was examined as a primary predictor in multivariable models (adjusted for BMI category), those women with baseline WC ≥ 88 cm experienced increased odds of incident disease (OR 1.47 [95% CI 1.33–1.62]), mobility disability (OR 1.64 [95% CI 1.49–1.84]) and death (OR 1.83 [95% CI 1.66–2.03]) compared to women with smaller baseline WC.

When participants were stratified by race/ethnicity, the relationships for increasing odds of incident disease/disability with baseline obesity persisted for non-Hispanic white and black/African-American participants. Hispanic/Latina participants who were obese at baseline, however, did not have significantly increased odds of death before 85 years relative to healthy weight counterparts, although there were far fewer of these women represented in the cohort (n = 600). Asian/Pacific Islander (API) participants (n = 781), the majority of whom were in the healthy weight range at baseline (57%), showed a somewhat different pattern. Odds ratios for incident disease and death among obese API women were not significantly elevated relative to healthy weight women (although the “n ”s for these groups was relatively small), however the odds of incident disability was significantly elevated amongst API women who were obese at baseline (OR 4.95 [95% CI 1.51–16.23]).

Conclusion. Compared to older women with a healthy BMI, obese women and those with increased abdominal circumference had a lower chance of surviving to age 85 years. Those who did survive were more likely to develop incident disease and/or disability than their healthy weight counterparts.

Commentary

The prevalence of obesity has risen substantially over the past several decades, and few demographic groups have found themselves spared from the epidemic [2]. Although much focus is placed on obesity incidence and prevalence among children and young adults, adults over age 60, a growing segment of the US population, are heavily impacted by the rising rates of obesity as well, with 42% of women and 37% of men in this group characterized as obese in 2010 [2]. This trend has potentially major implications for policy makers who are tasked with cutting the cost of programs such as Medicare.

Obesity has only recently been recognized as a disease by the American Medical Association, and yet it has long been associated with costly and debilitating chronic conditions such as type 2 diabetes, hypertension, sleep apnea, and degenerative joint disease [3]. Despite this fact, several epidemiologic studies have suggested an “obesity paradox”—older adults who are mildly obese have mortality rates similar to normal weight adults, and those who are overweight appear to have lower mortality [4]. These papers have generated controversy among obesity researchers and epidemiologists who have grappled with the following question: How is it possible that overweight and obesity, while clearly linked to so many chronic conditions that increase mortality and morbidity, might be a good thing? Is there such a thing as a “healthy level of obesity,” or, can you be “fit and fat?” In the midst of these discussions and the media storm that inevitably surrounds them, patients are confronted with confusing mixed messages, possibly making them less likely to attempt to maintain a healthy body weight. Unfortunately, as many prior authors have asserted, most of the epidemiologic studies that assert this protective effect of overweight and obesity have not accounted for potentially important confounders of the “weight category–mortality” relationship, such as smoking status [5]. Among older adults, a substantial fraction of those in the normal weight category are at a so-called healthy BMI for very unhealthy reasons, such as cigarette smoking, cancer, or other chronic conditions (ie, they were heavier but lost weight due to underlying illness). Including these sick (but so-called “healthy weight”) people alongside those who are truly healthy and in a healthy BMI range muddies the picture and does not effectively isolate the impact of weight status on morbidity and mortality.

This cohort study by Rillamas-Sun et al makes an important contribution to the discussion by relying on a very large and comprehensive dataset, with an impressive follow-up period of nearly 2 decades, to more fully isolate the relationship between BMI category and survival for postmenopausal women. By adjusting for important potential confounders such as baseline smoking status, alcohol use, chronic disease status and a number of sociodemographic factors, and by separating out the chronically ill patients from the beginning, the investigators reached conclusions that seem to align better with all that we know about the increased health risks conferred by obesity. They found that postmenopausal women who were obese but without prevalent disease at baseline had increased odds of death before age 85, as well as increased odds of incident chronic disease (such as cardiovascular disease or diabetes) and increased odds of incident disability relative to postmenopausal women starting out in a healthy BMI range. Degree of obesity seemed to matter as well; those with class II and III obesity had significantly increased odds of developing mobility impairment, in particular, relative to normal weight women. This is particularly important when viewed through the lens of caring for an aging population—those who have significant mobility impairment will have a much harder time caring for themselves as they age. Furthermore, they found that overweight women also faced slightly increased odds of these outcomes relative to normal weight women. Abdominal adiposity, in particular, appeared to confer risk of death and disease, as elevated odds of mortality and incident disease or disability persisted in women with waist circumference ≥ 88 cm even after adjusting for BMI. As has been suggested by prior research on this topic, this study also supported the finding that being underweight increases ones odds of death, however, there was no increased incidence of disease or mobility disability for underweight women (relative to healthy starting weight).

The authors of the study made a wise decision in separating women with baseline chronic illness from those who had not yet been diagnosed with diabetes, cardiovascular disease or other chronic condition at baseline. As is pointed out in an editorial accompanying this study [6], this creates a scenario where the exposure (obesity) clearly predates the outcome (chronic illness), helping to avoid contamination of risk estimates by reverse causation (ie, is chronic illness leading to increased obesity, with the downstream increase in mortality actually due to the chronic illness?).

Despite the clear strengths of the study, there are several important limitations that must be acknowledged in interpreting the results. The most obvious is that BMI status was only measured at baseline. There is no way of knowing either what a participant’s weight trajectory had been in their younger years, or what happened to the BMI during the study follow-up period, both of which could certainly impact a participant’s risk of morbidity or mortality. Given a follow-up period of nearly 20 years, it is possible that there was crossover between BMI (exposure) categories after baseline assignment. Furthermore, the study does not address the very important question of how an intervention to promote weight loss in older women might impact morbidity and mortality—it is possible that encouraging weight loss in this population may in fact worsen health outcomes for some patients [6].

The generalizability of the study may be somewhat limited. The study population itself represented a group of women who were likely relatively healthy and motivated, having self-selected to participate in the WHI, thus they could have been healthier than groups studied in previous population-based samples. Furthermore, the study results may not generalize to men, however other similar cohort studies with male participants have reached similar conclusions [7].

Applications for Clinical Practice

To promote longevity and maintenance of independence in our growing population of postmenopausal women, it is important that physicians continue to educate and assist their patients in maintaining a healthy weight as they age. Although the impact of intentional weight loss in obese older women is not addressed by this paper, it does support the idea that obese postmenopausal women are at higher risk of death before age 85 years and disability. Therefore, for these patients, physicians should take particular care to reinforce healthy lifestyle choices such as good nutrition and regular physical activity.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine whether having a body mass index (BMI) in the obese range (30 kg/m²) as an older adult woman is associated with changes in late-age survival and morbidity.

Design. Observational cohort study.

Setting and participants. This study relied upon data collected as part of the Women’s Health Initiative (WHI), an observational study and clinical trial focusing on the health of postmenopausal women aged 50–79 years at enrollment. For the purposes of the WHI, women were recruited from centers across the United States between 1993 and 1998 and could participate in several intervention studies (hormone replacement therapy, low-fat diet, calcium/vitamin D supplementation) or an observational study [1].

For this paper, the authors utilized data from those WHI participants who, based on their age at enrollment, could have reached age 85 years by September of 2012. The authors excluded women who did not provide follow-up health information within 18 months of their 85th birthdays or who reported mobility disabilities at their baseline data collection. This resulted in a total of 36,611 women for analysis.

There were a number of baseline measures collected on the study participants. Via written survey, participants self-reported their race and ethnicity, hormone use status, smoking status, alcohol consumption, physical activity level, depressive symptoms, and a number of demographic characteristics. Study personnel objectively measured height and weight to calculate baseline BMI and also measured waist circumference (WC, in cm).

The primary exposure measure for this study was BMI category at trial entry categorized as follows: underweight (< 18.5 kg/m²), healthy weight (18.5–24.9 kg/m²), overweight (25.0–29.9 kg/m²) or obese class I (30–34.9 kg/m²), II (35–39.9 kg/m²) or III (≥ 40 kg/m²), using standard accepted cut-points except for Asian/Pacific Islander participants, where alternative World Health Organization (WHO) cut-points were used. The WHO cut-points are slightly lower to account for usual body habitus and disease risk in that population. BMI changes over study follow-up were not included in the exposure measure for this study. WC (dichotomized around 88 cm) was also used as an exposure measure.

Main outcome measures. Disease-free survival status during the follow-up period. In the year at which participants were supposed to reach their 85th birthdays, they were categorized as to whether they had survived or not. Survival status was ascertained by hospital record review, autopsy reports, death certificates and review of the National Death Index. Those who survived were sub-grouped according to type of survival into 1 of the following categories: (1) no incident disease and no mobility disability (healthy), (2) baseline disease present but no incident disease or mobility disability during follow-up (prevalent disease), (3) incident disease but no mobility disability during follow-up (incident disease), and (4) incident mobility disability with or without incident disease (disabled).

Diseases of interest (prevalent and incident) included coronary and cerebrovascular disease, cancer, diabetes and hip fracture—the conditions the investigators felt most increased risk of death or morbidity and mobility disability in this population of aging women. Baseline disease status was defined using self-report, but incident disease in follow-up was more rigorously defined using self-report plus medical record review, except for incident diabetes, which required only self-report of diagnosis plus report of new oral hypoglycemic or insulin use.

Because the outcome of interest (survival status) had 5 possible categories, multinomial logistic regression was used as the analytic technique, with baseline BMI category and WC categories as predictors. The authors adjusted for baseline characteristics including age, race/ethnicity, study arm (intervention or observational for WHI), educational level, marital status, smoking status, ethanol use, self-reported physical activity and depression symptoms. Because of the possibly interrelated predictors (BMI and WC), the authors built BMI models with and without WC, and when WC was the primary predictor they adjusted for a participant’s BMI in order to try to isolate the impact of central adiposity. Additionally, they performed the analyses stratified by race and ethnicity as well as by smoking status.

Results. The mean (SD) baseline age of participants was 72.4 (3) years, and the vast majority (88.5%) self-identified as non-Hispanic white. At the end of the follow-up period, of the initial 36,611 participants, 9079 (24.8%) had died, 6702 (18.3%) had become disabled, 8512 (23.2%) had developed incident disease without disability, 5366 (14.6%) had prevalent but no incident disease, and 6952 (18.9%) were categorized as healthy. There were a number of potentially confounding baseline characteristics that differed between the survival categories. Importantly, race was associated with survival status—non-Hispanic white women were more likely to be in the “healthy” category at follow-up than their counterparts from other races/ethnicities. Baseline smokers were more likely not to live to 85 years, and those with less than a high school education were also more likely not to live to 85 years.

In models adjusting for baseline covariates, with BMI category as the primary predictor, women with an obese baseline BMI had significantly increased odds of not living to 85 years of age, relative to women in a healthy baseline BMI category, with increasing odds of death among those with higher baseline BMI levels (class I obesity odds ratio [OR] 1.72 [95% CI 1.55–1.92], class II obesity OR 3.28 [95% CI 2.69–4.01], class III obesity OR 3.48 [95% CI 2.52–4.80]). Amongst survivors, baseline obesity was also associated with greater odds of developing incident disease, relative to healthy weight women (class I obesity OR 1.65 [95% CI 1.48–1.84], class II obesity OR 2.44 (95% CI 2.02–2.96), class III obesity OR 1.73 [95% CI 1.21–2.46]). There was a striking relationship between baseline obesity and the odds of incident disability during follow-up (class I obesity OR 3.22 [95% CI 2.87–3.61], class II obesity OR 6.62 [95% CI 5.41–8.09], class III obesity OR 6.65 [95% CI 4.80–9.21]).

Women who were overweight at baseline also displayed statistically significant but more modestly increased odds of incident disease, mobility disability, and death relative to their normal-weight counterparts. Importantly, even in multivariable models, being underweight at baseline was also associated with significantly increased odds of death before age 85 relative to healthy weight individuals (OR 2.09 [95% CI 1.54–2.85]) but not with increased odds of incident disease or disability.

When WC status was adjusted for in the “BMI-outcome” models, the odds of death, disability, and incident disease were attenuated for obese women but remained elevated, particularly for women with class II or III obesity. When WC was examined as a primary predictor in multivariable models (adjusted for BMI category), those women with baseline WC ≥ 88 cm experienced increased odds of incident disease (OR 1.47 [95% CI 1.33–1.62]), mobility disability (OR 1.64 [95% CI 1.49–1.84]) and death (OR 1.83 [95% CI 1.66–2.03]) compared to women with smaller baseline WC.

When participants were stratified by race/ethnicity, the relationships for increasing odds of incident disease/disability with baseline obesity persisted for non-Hispanic white and black/African-American participants. Hispanic/Latina participants who were obese at baseline, however, did not have significantly increased odds of death before 85 years relative to healthy weight counterparts, although there were far fewer of these women represented in the cohort (n = 600). Asian/Pacific Islander (API) participants (n = 781), the majority of whom were in the healthy weight range at baseline (57%), showed a somewhat different pattern. Odds ratios for incident disease and death among obese API women were not significantly elevated relative to healthy weight women (although the “n ”s for these groups was relatively small), however the odds of incident disability was significantly elevated amongst API women who were obese at baseline (OR 4.95 [95% CI 1.51–16.23]).

Conclusion. Compared to older women with a healthy BMI, obese women and those with increased abdominal circumference had a lower chance of surviving to age 85 years. Those who did survive were more likely to develop incident disease and/or disability than their healthy weight counterparts.

Commentary

The prevalence of obesity has risen substantially over the past several decades, and few demographic groups have found themselves spared from the epidemic [2]. Although much focus is placed on obesity incidence and prevalence among children and young adults, adults over age 60, a growing segment of the US population, are heavily impacted by the rising rates of obesity as well, with 42% of women and 37% of men in this group characterized as obese in 2010 [2]. This trend has potentially major implications for policy makers who are tasked with cutting the cost of programs such as Medicare.

Obesity has only recently been recognized as a disease by the American Medical Association, and yet it has long been associated with costly and debilitating chronic conditions such as type 2 diabetes, hypertension, sleep apnea, and degenerative joint disease [3]. Despite this fact, several epidemiologic studies have suggested an “obesity paradox”—older adults who are mildly obese have mortality rates similar to normal weight adults, and those who are overweight appear to have lower mortality [4]. These papers have generated controversy among obesity researchers and epidemiologists who have grappled with the following question: How is it possible that overweight and obesity, while clearly linked to so many chronic conditions that increase mortality and morbidity, might be a good thing? Is there such a thing as a “healthy level of obesity,” or, can you be “fit and fat?” In the midst of these discussions and the media storm that inevitably surrounds them, patients are confronted with confusing mixed messages, possibly making them less likely to attempt to maintain a healthy body weight. Unfortunately, as many prior authors have asserted, most of the epidemiologic studies that assert this protective effect of overweight and obesity have not accounted for potentially important confounders of the “weight category–mortality” relationship, such as smoking status [5]. Among older adults, a substantial fraction of those in the normal weight category are at a so-called healthy BMI for very unhealthy reasons, such as cigarette smoking, cancer, or other chronic conditions (ie, they were heavier but lost weight due to underlying illness). Including these sick (but so-called “healthy weight”) people alongside those who are truly healthy and in a healthy BMI range muddies the picture and does not effectively isolate the impact of weight status on morbidity and mortality.

This cohort study by Rillamas-Sun et al makes an important contribution to the discussion by relying on a very large and comprehensive dataset, with an impressive follow-up period of nearly 2 decades, to more fully isolate the relationship between BMI category and survival for postmenopausal women. By adjusting for important potential confounders such as baseline smoking status, alcohol use, chronic disease status and a number of sociodemographic factors, and by separating out the chronically ill patients from the beginning, the investigators reached conclusions that seem to align better with all that we know about the increased health risks conferred by obesity. They found that postmenopausal women who were obese but without prevalent disease at baseline had increased odds of death before age 85, as well as increased odds of incident chronic disease (such as cardiovascular disease or diabetes) and increased odds of incident disability relative to postmenopausal women starting out in a healthy BMI range. Degree of obesity seemed to matter as well; those with class II and III obesity had significantly increased odds of developing mobility impairment, in particular, relative to normal weight women. This is particularly important when viewed through the lens of caring for an aging population—those who have significant mobility impairment will have a much harder time caring for themselves as they age. Furthermore, they found that overweight women also faced slightly increased odds of these outcomes relative to normal weight women. Abdominal adiposity, in particular, appeared to confer risk of death and disease, as elevated odds of mortality and incident disease or disability persisted in women with waist circumference ≥ 88 cm even after adjusting for BMI. As has been suggested by prior research on this topic, this study also supported the finding that being underweight increases ones odds of death, however, there was no increased incidence of disease or mobility disability for underweight women (relative to healthy starting weight).

The authors of the study made a wise decision in separating women with baseline chronic illness from those who had not yet been diagnosed with diabetes, cardiovascular disease or other chronic condition at baseline. As is pointed out in an editorial accompanying this study [6], this creates a scenario where the exposure (obesity) clearly predates the outcome (chronic illness), helping to avoid contamination of risk estimates by reverse causation (ie, is chronic illness leading to increased obesity, with the downstream increase in mortality actually due to the chronic illness?).

Despite the clear strengths of the study, there are several important limitations that must be acknowledged in interpreting the results. The most obvious is that BMI status was only measured at baseline. There is no way of knowing either what a participant’s weight trajectory had been in their younger years, or what happened to the BMI during the study follow-up period, both of which could certainly impact a participant’s risk of morbidity or mortality. Given a follow-up period of nearly 20 years, it is possible that there was crossover between BMI (exposure) categories after baseline assignment. Furthermore, the study does not address the very important question of how an intervention to promote weight loss in older women might impact morbidity and mortality—it is possible that encouraging weight loss in this population may in fact worsen health outcomes for some patients [6].

The generalizability of the study may be somewhat limited. The study population itself represented a group of women who were likely relatively healthy and motivated, having self-selected to participate in the WHI, thus they could have been healthier than groups studied in previous population-based samples. Furthermore, the study results may not generalize to men, however other similar cohort studies with male participants have reached similar conclusions [7].

Applications for Clinical Practice

To promote longevity and maintenance of independence in our growing population of postmenopausal women, it is important that physicians continue to educate and assist their patients in maintaining a healthy weight as they age. Although the impact of intentional weight loss in obese older women is not addressed by this paper, it does support the idea that obese postmenopausal women are at higher risk of death before age 85 years and disability. Therefore, for these patients, physicians should take particular care to reinforce healthy lifestyle choices such as good nutrition and regular physical activity.

—Kristina Lewis, MD, MPH

Study Overview

Objective. To determine whether having a body mass index (BMI) in the obese range (30 kg/m²) as an older adult woman is associated with changes in late-age survival and morbidity.

Design. Observational cohort study.

Setting and participants. This study relied upon data collected as part of the Women’s Health Initiative (WHI), an observational study and clinical trial focusing on the health of postmenopausal women aged 50–79 years at enrollment. For the purposes of the WHI, women were recruited from centers across the United States between 1993 and 1998 and could participate in several intervention studies (hormone replacement therapy, low-fat diet, calcium/vitamin D supplementation) or an observational study [1].

For this paper, the authors utilized data from those WHI participants who, based on their age at enrollment, could have reached age 85 years by September of 2012. The authors excluded women who did not provide follow-up health information within 18 months of their 85th birthdays or who reported mobility disabilities at their baseline data collection. This resulted in a total of 36,611 women for analysis.

There were a number of baseline measures collected on the study participants. Via written survey, participants self-reported their race and ethnicity, hormone use status, smoking status, alcohol consumption, physical activity level, depressive symptoms, and a number of demographic characteristics. Study personnel objectively measured height and weight to calculate baseline BMI and also measured waist circumference (WC, in cm).

The primary exposure measure for this study was BMI category at trial entry categorized as follows: underweight (< 18.5 kg/m²), healthy weight (18.5–24.9 kg/m²), overweight (25.0–29.9 kg/m²) or obese class I (30–34.9 kg/m²), II (35–39.9 kg/m²) or III (≥ 40 kg/m²), using standard accepted cut-points except for Asian/Pacific Islander participants, where alternative World Health Organization (WHO) cut-points were used. The WHO cut-points are slightly lower to account for usual body habitus and disease risk in that population. BMI changes over study follow-up were not included in the exposure measure for this study. WC (dichotomized around 88 cm) was also used as an exposure measure.

Main outcome measures. Disease-free survival status during the follow-up period. In the year at which participants were supposed to reach their 85th birthdays, they were categorized as to whether they had survived or not. Survival status was ascertained by hospital record review, autopsy reports, death certificates and review of the National Death Index. Those who survived were sub-grouped according to type of survival into 1 of the following categories: (1) no incident disease and no mobility disability (healthy), (2) baseline disease present but no incident disease or mobility disability during follow-up (prevalent disease), (3) incident disease but no mobility disability during follow-up (incident disease), and (4) incident mobility disability with or without incident disease (disabled).

Diseases of interest (prevalent and incident) included coronary and cerebrovascular disease, cancer, diabetes and hip fracture—the conditions the investigators felt most increased risk of death or morbidity and mobility disability in this population of aging women. Baseline disease status was defined using self-report, but incident disease in follow-up was more rigorously defined using self-report plus medical record review, except for incident diabetes, which required only self-report of diagnosis plus report of new oral hypoglycemic or insulin use.

Because the outcome of interest (survival status) had 5 possible categories, multinomial logistic regression was used as the analytic technique, with baseline BMI category and WC categories as predictors. The authors adjusted for baseline characteristics including age, race/ethnicity, study arm (intervention or observational for WHI), educational level, marital status, smoking status, ethanol use, self-reported physical activity and depression symptoms. Because of the possibly interrelated predictors (BMI and WC), the authors built BMI models with and without WC, and when WC was the primary predictor they adjusted for a participant’s BMI in order to try to isolate the impact of central adiposity. Additionally, they performed the analyses stratified by race and ethnicity as well as by smoking status.

Results. The mean (SD) baseline age of participants was 72.4 (3) years, and the vast majority (88.5%) self-identified as non-Hispanic white. At the end of the follow-up period, of the initial 36,611 participants, 9079 (24.8%) had died, 6702 (18.3%) had become disabled, 8512 (23.2%) had developed incident disease without disability, 5366 (14.6%) had prevalent but no incident disease, and 6952 (18.9%) were categorized as healthy. There were a number of potentially confounding baseline characteristics that differed between the survival categories. Importantly, race was associated with survival status—non-Hispanic white women were more likely to be in the “healthy” category at follow-up than their counterparts from other races/ethnicities. Baseline smokers were more likely not to live to 85 years, and those with less than a high school education were also more likely not to live to 85 years.

In models adjusting for baseline covariates, with BMI category as the primary predictor, women with an obese baseline BMI had significantly increased odds of not living to 85 years of age, relative to women in a healthy baseline BMI category, with increasing odds of death among those with higher baseline BMI levels (class I obesity odds ratio [OR] 1.72 [95% CI 1.55–1.92], class II obesity OR 3.28 [95% CI 2.69–4.01], class III obesity OR 3.48 [95% CI 2.52–4.80]). Amongst survivors, baseline obesity was also associated with greater odds of developing incident disease, relative to healthy weight women (class I obesity OR 1.65 [95% CI 1.48–1.84], class II obesity OR 2.44 (95% CI 2.02–2.96), class III obesity OR 1.73 [95% CI 1.21–2.46]). There was a striking relationship between baseline obesity and the odds of incident disability during follow-up (class I obesity OR 3.22 [95% CI 2.87–3.61], class II obesity OR 6.62 [95% CI 5.41–8.09], class III obesity OR 6.65 [95% CI 4.80–9.21]).

Women who were overweight at baseline also displayed statistically significant but more modestly increased odds of incident disease, mobility disability, and death relative to their normal-weight counterparts. Importantly, even in multivariable models, being underweight at baseline was also associated with significantly increased odds of death before age 85 relative to healthy weight individuals (OR 2.09 [95% CI 1.54–2.85]) but not with increased odds of incident disease or disability.

When WC status was adjusted for in the “BMI-outcome” models, the odds of death, disability, and incident disease were attenuated for obese women but remained elevated, particularly for women with class II or III obesity. When WC was examined as a primary predictor in multivariable models (adjusted for BMI category), those women with baseline WC ≥ 88 cm experienced increased odds of incident disease (OR 1.47 [95% CI 1.33–1.62]), mobility disability (OR 1.64 [95% CI 1.49–1.84]) and death (OR 1.83 [95% CI 1.66–2.03]) compared to women with smaller baseline WC.

When participants were stratified by race/ethnicity, the relationships for increasing odds of incident disease/disability with baseline obesity persisted for non-Hispanic white and black/African-American participants. Hispanic/Latina participants who were obese at baseline, however, did not have significantly increased odds of death before 85 years relative to healthy weight counterparts, although there were far fewer of these women represented in the cohort (n = 600). Asian/Pacific Islander (API) participants (n = 781), the majority of whom were in the healthy weight range at baseline (57%), showed a somewhat different pattern. Odds ratios for incident disease and death among obese API women were not significantly elevated relative to healthy weight women (although the “n ”s for these groups was relatively small), however the odds of incident disability was significantly elevated amongst API women who were obese at baseline (OR 4.95 [95% CI 1.51–16.23]).

Conclusion. Compared to older women with a healthy BMI, obese women and those with increased abdominal circumference had a lower chance of surviving to age 85 years. Those who did survive were more likely to develop incident disease and/or disability than their healthy weight counterparts.

Commentary

The prevalence of obesity has risen substantially over the past several decades, and few demographic groups have found themselves spared from the epidemic [2]. Although much focus is placed on obesity incidence and prevalence among children and young adults, adults over age 60, a growing segment of the US population, are heavily impacted by the rising rates of obesity as well, with 42% of women and 37% of men in this group characterized as obese in 2010 [2]. This trend has potentially major implications for policy makers who are tasked with cutting the cost of programs such as Medicare.

Obesity has only recently been recognized as a disease by the American Medical Association, and yet it has long been associated with costly and debilitating chronic conditions such as type 2 diabetes, hypertension, sleep apnea, and degenerative joint disease [3]. Despite this fact, several epidemiologic studies have suggested an “obesity paradox”—older adults who are mildly obese have mortality rates similar to normal weight adults, and those who are overweight appear to have lower mortality [4]. These papers have generated controversy among obesity researchers and epidemiologists who have grappled with the following question: How is it possible that overweight and obesity, while clearly linked to so many chronic conditions that increase mortality and morbidity, might be a good thing? Is there such a thing as a “healthy level of obesity,” or, can you be “fit and fat?” In the midst of these discussions and the media storm that inevitably surrounds them, patients are confronted with confusing mixed messages, possibly making them less likely to attempt to maintain a healthy body weight. Unfortunately, as many prior authors have asserted, most of the epidemiologic studies that assert this protective effect of overweight and obesity have not accounted for potentially important confounders of the “weight category–mortality” relationship, such as smoking status [5]. Among older adults, a substantial fraction of those in the normal weight category are at a so-called healthy BMI for very unhealthy reasons, such as cigarette smoking, cancer, or other chronic conditions (ie, they were heavier but lost weight due to underlying illness). Including these sick (but so-called “healthy weight”) people alongside those who are truly healthy and in a healthy BMI range muddies the picture and does not effectively isolate the impact of weight status on morbidity and mortality.

This cohort study by Rillamas-Sun et al makes an important contribution to the discussion by relying on a very large and comprehensive dataset, with an impressive follow-up period of nearly 2 decades, to more fully isolate the relationship between BMI category and survival for postmenopausal women. By adjusting for important potential confounders such as baseline smoking status, alcohol use, chronic disease status and a number of sociodemographic factors, and by separating out the chronically ill patients from the beginning, the investigators reached conclusions that seem to align better with all that we know about the increased health risks conferred by obesity. They found that postmenopausal women who were obese but without prevalent disease at baseline had increased odds of death before age 85, as well as increased odds of incident chronic disease (such as cardiovascular disease or diabetes) and increased odds of incident disability relative to postmenopausal women starting out in a healthy BMI range. Degree of obesity seemed to matter as well; those with class II and III obesity had significantly increased odds of developing mobility impairment, in particular, relative to normal weight women. This is particularly important when viewed through the lens of caring for an aging population—those who have significant mobility impairment will have a much harder time caring for themselves as they age. Furthermore, they found that overweight women also faced slightly increased odds of these outcomes relative to normal weight women. Abdominal adiposity, in particular, appeared to confer risk of death and disease, as elevated odds of mortality and incident disease or disability persisted in women with waist circumference ≥ 88 cm even after adjusting for BMI. As has been suggested by prior research on this topic, this study also supported the finding that being underweight increases ones odds of death, however, there was no increased incidence of disease or mobility disability for underweight women (relative to healthy starting weight).

The authors of the study made a wise decision in separating women with baseline chronic illness from those who had not yet been diagnosed with diabetes, cardiovascular disease or other chronic condition at baseline. As is pointed out in an editorial accompanying this study [6], this creates a scenario where the exposure (obesity) clearly predates the outcome (chronic illness), helping to avoid contamination of risk estimates by reverse causation (ie, is chronic illness leading to increased obesity, with the downstream increase in mortality actually due to the chronic illness?).

Despite the clear strengths of the study, there are several important limitations that must be acknowledged in interpreting the results. The most obvious is that BMI status was only measured at baseline. There is no way of knowing either what a participant’s weight trajectory had been in their younger years, or what happened to the BMI during the study follow-up period, both of which could certainly impact a participant’s risk of morbidity or mortality. Given a follow-up period of nearly 20 years, it is possible that there was crossover between BMI (exposure) categories after baseline assignment. Furthermore, the study does not address the very important question of how an intervention to promote weight loss in older women might impact morbidity and mortality—it is possible that encouraging weight loss in this population may in fact worsen health outcomes for some patients [6].

The generalizability of the study may be somewhat limited. The study population itself represented a group of women who were likely relatively healthy and motivated, having self-selected to participate in the WHI, thus they could have been healthier than groups studied in previous population-based samples. Furthermore, the study results may not generalize to men, however other similar cohort studies with male participants have reached similar conclusions [7].

Applications for Clinical Practice

To promote longevity and maintenance of independence in our growing population of postmenopausal women, it is important that physicians continue to educate and assist their patients in maintaining a healthy weight as they age. Although the impact of intentional weight loss in obese older women is not addressed by this paper, it does support the idea that obese postmenopausal women are at higher risk of death before age 85 years and disability. Therefore, for these patients, physicians should take particular care to reinforce healthy lifestyle choices such as good nutrition and regular physical activity.

—Kristina Lewis, MD, MPH