Reviews

Discuss this article at www.facebook.com/CurrentPsychiatry

In addition to increasing patients’ risk for cardiovascular disease, stroke, and cancer, obesity and metabolic disturbance contribute to age-related cognitive decline and dementia. In particular, insulin resistance and hyperinsulinemia promote neurocognitive dysfunction and neurodegenerative changes during the extended, preclinical phase of Alzheimer’s disease (AD). However, with dietary modification it may be possible to resensitize insulin receptors, correct hyperinsulinemia, and improve memory function.

Metabolic disturbance and neurodegeneration

In the United States, 5.4 million people have AD, and there will be an estimated 16 million cases by 2050.¹ Simultaneously we are experiencing an epidemic of metabolic disturbance and obesity. Approximately, 64% of adults in the United States are overweight (body mass index [BMI]: 25.0 to 29.9 kg/m²) and 34% are obese (BMI: ≥30 kg/m²).² By 2030, 86% of adults will be overweight and 51% will be obese.³ This confluence of epidemics is not coincidental but instead reflects the fact that metabolic disturbance is a fundamental factor contributing to cognitive decline and neurodegeneration.⁴

Ninety-six percent of AD cases are classified as late onset, sporadic AD, occurring after age 64.¹ Mild cognitive impairment (MCI) is a clinical construct that entails greater than expected memory impairment for the patient’s age and identifies older adults who are at increased risk for dementia. MCI represents the first clinical manifestation of neurodegeneration for a subset of patients who will progress to AD.^5,6 MCI is distinguished from age-associated memory impairment (AAMI), which originally was conceptualized as normal or benign memory decline with aging.^7,8 Recent data indicate that Alzheimer’s-type neuropathologic changes are the basis for subjective memory complaints and objectively assessed age-related cognitive decline,⁹ and early neurodegeneration is present in many patients with AAMI or MCI.¹⁰ This is consistent with the idea that an extended preclinical phase precedes AD onset. The preclinical phase can persist for a decade or more and precedes MCI and overt functional decline. However, neuropathologic changes accumulate during the preclinical phase of AD¹¹ and during the preclinical phase of type 2 diabetes mellitus (T2DM).

Hyperinsulinemia and dementia

Insulin resistance and hyperinsulinemia occur in >40% of individuals age ≥60 and prevalence increases with age.^4,12 Hyperinsulinemia develops to compensate for insulin resistance to overcome receptor insensitivity and maintain glucose homeostasis. Insulin receptors are densely expressed in brain regions vulnerable to neurodegeneration, including the medial temporal lobe and prefrontal cortex, which mediate long-term memory and working memory. However, insulin must be transported into the CNS from the periphery because little is synthesized in the brain. Paradoxically, peripheral compensatory hyperinsulinemia resulting from insulin resistance is associated with central (brain) hypoinsulinemia because of insensitivity and saturation of the receptor-mediated blood-brain barrier transport mechanism.^13-15

Hyperinsulinemia is the precursor to T2DM. However, hyperinsulinemia is not well recognized in clinical contexts and generally is not a treatment target. Nonetheless, it contributes to several health problems, and insulin resistance in middle age is associated with age-related diseases such as hypertension, coronary artery disease, stroke, and cancer, while insulin sensitivity protects against such disorders.¹⁶

Chronic insulin resistance may contribute more to dementia development than T2DM because of the extended period of hyperinsulinemia that precedes T2DM onset. In population studies,¹⁷ insulin resistance syndrome increases risk for developing AD independent of apolipoprotein E (APOE e4) allele status, and in a longitudinal study,¹⁸ the risk for AD solely attributable to peripheral hyperinsulinemia was up to 39%. Being overweight in midlife increases risk for dementia in late life, and APOE e4 allele status does not contribute additional risk after accounting for BMI.¹⁹ Middle-aged individuals with hyperinsulinemia show memory decline, and obesity in middle age was associated with greater cognitive impairment after 6-year follow-up.²⁰ Even in older adults who seem cognitively unimpaired, BMI and fasting insulin are positively correlated with atrophy in frontal, temporal, and subcortical brain regions, and obesity is an independent risk for atrophy in several brain regions, including the hippocampus.²¹

Compared with healthy older adults, individuals with AD have lower ratios of cerebrospinal fluid to plasma insulin.²² This lower ratio reflects the peripheral-to-central gradient of insulin levels in AD and suggests an etiological role for such metabolic disturbance. Insulin resistance has downstream effects that potentiate neurodegenerative factors, and central hypoinsulinemia can accelerate neurodegenerative processes and cognitive decline.^4,23 Brain insulin plays a direct role in regulating proinflammatory cytokines and neurotrophic and neuroplastic factors essential for memory function. Insulin degrading enzyme, which varies with insulin levels,²⁴ regulates the generation and clearance of amyloid β (Aβ) from the brain.²⁵

Hyperinsulinemia typically is evident in increasing waist circumference and body weight.²⁶ Waist circumference of ≥100 cm (39 inches) is a sensitive, specific, and independent predictor of hyperinsulinemia for men and women and a stronger predictor than BMI, waist-to-hip ratio, and other measures of body fat.²⁷ Unpublished data derived from our clinical research with MCI subjects supports the association of metabolic disturbance with age-related cognitive decline. Our subjects are recruited from the community on the basis of mild memory decline and—other than excluding those with diabetes—weight and metabolic status are not considered in evaluating individuals for enrollment. The Table contains data on waist circumference and metabolic function in 122 older adults (age ≥68) with MCI. On average, these individuals exhibited fasting insulin values in the hyperinsulinemia range and elevated fasting glucose levels that indicated borderline diabetes. Waist circumference also was high, indicating excessive visceral fat deposition. We also observed a relationship between waist circumference and insulin, a consistent observation in older adults with memory decline. These data would not be surprising in any sample of older adults because of the population base rates for these conditions. However, we also found that waist circumference was a significant predictor of memory performance in patients with MCI. Abdominal adiposity is highly correlated with intrahepatic fat.²⁸ Given this and recent indications that Alzheimer’s-type neuropathologic factors are generated in the liver,^29,30 the predictive value of waist circumference to memory performance may reflect the fact that it is a proxy for downstream actions of liver fat.

Table

Waist circumference and metabolic factors in 122 older adults with MCI^a

Dietary interventions

There is no cure for dementia, and it is not clear when effective therapy might be developed. Prevention and risk mitigation represent the best means of reducing the impact of this public health problem. Researchers have proposed that interventions initiated when individuals have predementia conditions such as AAMI and MCI might stall progression of cognitive decline, and MCI may be the last point when interventions might be effective because of the self-reinforcing neuropathologic cascades of AD.³¹ Because central hypoinsulinemia may promote central inflammation, Aβ generation, and reduced neuroplasticity, approaches aimed at improving metabolic function (and in particular correcting hyperinsulinemia) could influence fundamental neurodegenerative processes. Dietary approaches to preventing dementia are effective, low-risk, yet underutilized interventions. Reducing insulin by restricting calories³² or maintaining a ketogenic diet³³ has been associated with improved memory function in middle-aged and older adults.

Carbohydrate consumption is the principal determinant of insulin secretion. Eliminating high-glycemic foods, including processed carbohydrates and sweets, would sensitize insulin receptors and correct hyperinsulinemia. In addition, replacing high glycemic foods with fruits and vegetables would increase polyphenol intake. Epidemiologic evidence supports the idea that greater consumption of polyphenol-containing vegetables and fruits mitigates risk for neurocognitive decline and dementia.^34,35 Preclinical evidence suggests that such protection may be related to neuronal signaling effects and anti- inflammatory and antioxidant actions.³⁶ In addition, certain polyphenol compounds, such as those found in berries, enhance metabolic function.^37,38 In a 12-week pilot trial, older adults with early memory changes (N=9, mean age 76) who drank supplemental blueberry juice showed enhanced memory and improved metabolic parameters.³⁹

Dietary changes that preserve insulin receptor sensitivity can help ensure general health with aging and substantially mitigate risk for neurodegeneration. The Western diet is particularly insulinogenic and dietary habits are difficult to change. However, the substantial benefits, absence of adverse effects, and low cost make dietary intervention the optimal means of protecting against neurodegeneration and other age-related diseases. Embarking on such a program early in life would be best, although late-life intervention can be effective.

Related Resources

• Craft S, Watson GS. Insulin and neurodegenerative disease: shared and specific mechanisms. Lancet Neurol. 2004;3(3):169-178.
• Luchsinger JA, Tang MX, Shea S, et al. Hyperinsulinemia and risk of Alzheimer’s disease. Neurology. 2004; 63(7):1187-1192.
• Krikorian R, Shidler MD, Dangelo K, et al. Dietary ketosis enhances memory in mild cognitive impairment. Neurbiol Aging. 2012;33(2):425.e19-e27.

Disclosure

Dr. Krikorian receives grant support from the National Institutes of Health, 1R01AG034617-01.

Discuss this article at www.facebook.com/CurrentPsychiatry

In addition to increasing patients’ risk for cardiovascular disease, stroke, and cancer, obesity and metabolic disturbance contribute to age-related cognitive decline and dementia. In particular, insulin resistance and hyperinsulinemia promote neurocognitive dysfunction and neurodegenerative changes during the extended, preclinical phase of Alzheimer’s disease (AD). However, with dietary modification it may be possible to resensitize insulin receptors, correct hyperinsulinemia, and improve memory function.

Metabolic disturbance and neurodegeneration

In the United States, 5.4 million people have AD, and there will be an estimated 16 million cases by 2050.¹ Simultaneously we are experiencing an epidemic of metabolic disturbance and obesity. Approximately, 64% of adults in the United States are overweight (body mass index [BMI]: 25.0 to 29.9 kg/m²) and 34% are obese (BMI: ≥30 kg/m²).² By 2030, 86% of adults will be overweight and 51% will be obese.³ This confluence of epidemics is not coincidental but instead reflects the fact that metabolic disturbance is a fundamental factor contributing to cognitive decline and neurodegeneration.⁴

Ninety-six percent of AD cases are classified as late onset, sporadic AD, occurring after age 64.¹ Mild cognitive impairment (MCI) is a clinical construct that entails greater than expected memory impairment for the patient’s age and identifies older adults who are at increased risk for dementia. MCI represents the first clinical manifestation of neurodegeneration for a subset of patients who will progress to AD.^5,6 MCI is distinguished from age-associated memory impairment (AAMI), which originally was conceptualized as normal or benign memory decline with aging.^7,8 Recent data indicate that Alzheimer’s-type neuropathologic changes are the basis for subjective memory complaints and objectively assessed age-related cognitive decline,⁹ and early neurodegeneration is present in many patients with AAMI or MCI.¹⁰ This is consistent with the idea that an extended preclinical phase precedes AD onset. The preclinical phase can persist for a decade or more and precedes MCI and overt functional decline. However, neuropathologic changes accumulate during the preclinical phase of AD¹¹ and during the preclinical phase of type 2 diabetes mellitus (T2DM).

Hyperinsulinemia and dementia

Insulin resistance and hyperinsulinemia occur in >40% of individuals age ≥60 and prevalence increases with age.^4,12 Hyperinsulinemia develops to compensate for insulin resistance to overcome receptor insensitivity and maintain glucose homeostasis. Insulin receptors are densely expressed in brain regions vulnerable to neurodegeneration, including the medial temporal lobe and prefrontal cortex, which mediate long-term memory and working memory. However, insulin must be transported into the CNS from the periphery because little is synthesized in the brain. Paradoxically, peripheral compensatory hyperinsulinemia resulting from insulin resistance is associated with central (brain) hypoinsulinemia because of insensitivity and saturation of the receptor-mediated blood-brain barrier transport mechanism.^13-15

Hyperinsulinemia is the precursor to T2DM. However, hyperinsulinemia is not well recognized in clinical contexts and generally is not a treatment target. Nonetheless, it contributes to several health problems, and insulin resistance in middle age is associated with age-related diseases such as hypertension, coronary artery disease, stroke, and cancer, while insulin sensitivity protects against such disorders.¹⁶

Chronic insulin resistance may contribute more to dementia development than T2DM because of the extended period of hyperinsulinemia that precedes T2DM onset. In population studies,¹⁷ insulin resistance syndrome increases risk for developing AD independent of apolipoprotein E (APOE e4) allele status, and in a longitudinal study,¹⁸ the risk for AD solely attributable to peripheral hyperinsulinemia was up to 39%. Being overweight in midlife increases risk for dementia in late life, and APOE e4 allele status does not contribute additional risk after accounting for BMI.¹⁹ Middle-aged individuals with hyperinsulinemia show memory decline, and obesity in middle age was associated with greater cognitive impairment after 6-year follow-up.²⁰ Even in older adults who seem cognitively unimpaired, BMI and fasting insulin are positively correlated with atrophy in frontal, temporal, and subcortical brain regions, and obesity is an independent risk for atrophy in several brain regions, including the hippocampus.²¹

Compared with healthy older adults, individuals with AD have lower ratios of cerebrospinal fluid to plasma insulin.²² This lower ratio reflects the peripheral-to-central gradient of insulin levels in AD and suggests an etiological role for such metabolic disturbance. Insulin resistance has downstream effects that potentiate neurodegenerative factors, and central hypoinsulinemia can accelerate neurodegenerative processes and cognitive decline.^4,23 Brain insulin plays a direct role in regulating proinflammatory cytokines and neurotrophic and neuroplastic factors essential for memory function. Insulin degrading enzyme, which varies with insulin levels,²⁴ regulates the generation and clearance of amyloid β (Aβ) from the brain.²⁵

Hyperinsulinemia typically is evident in increasing waist circumference and body weight.²⁶ Waist circumference of ≥100 cm (39 inches) is a sensitive, specific, and independent predictor of hyperinsulinemia for men and women and a stronger predictor than BMI, waist-to-hip ratio, and other measures of body fat.²⁷ Unpublished data derived from our clinical research with MCI subjects supports the association of metabolic disturbance with age-related cognitive decline. Our subjects are recruited from the community on the basis of mild memory decline and—other than excluding those with diabetes—weight and metabolic status are not considered in evaluating individuals for enrollment. The Table contains data on waist circumference and metabolic function in 122 older adults (age ≥68) with MCI. On average, these individuals exhibited fasting insulin values in the hyperinsulinemia range and elevated fasting glucose levels that indicated borderline diabetes. Waist circumference also was high, indicating excessive visceral fat deposition. We also observed a relationship between waist circumference and insulin, a consistent observation in older adults with memory decline. These data would not be surprising in any sample of older adults because of the population base rates for these conditions. However, we also found that waist circumference was a significant predictor of memory performance in patients with MCI. Abdominal adiposity is highly correlated with intrahepatic fat.²⁸ Given this and recent indications that Alzheimer’s-type neuropathologic factors are generated in the liver,^29,30 the predictive value of waist circumference to memory performance may reflect the fact that it is a proxy for downstream actions of liver fat.

Table

Waist circumference and metabolic factors in 122 older adults with MCI^a

Dietary interventions

There is no cure for dementia, and it is not clear when effective therapy might be developed. Prevention and risk mitigation represent the best means of reducing the impact of this public health problem. Researchers have proposed that interventions initiated when individuals have predementia conditions such as AAMI and MCI might stall progression of cognitive decline, and MCI may be the last point when interventions might be effective because of the self-reinforcing neuropathologic cascades of AD.³¹ Because central hypoinsulinemia may promote central inflammation, Aβ generation, and reduced neuroplasticity, approaches aimed at improving metabolic function (and in particular correcting hyperinsulinemia) could influence fundamental neurodegenerative processes. Dietary approaches to preventing dementia are effective, low-risk, yet underutilized interventions. Reducing insulin by restricting calories³² or maintaining a ketogenic diet³³ has been associated with improved memory function in middle-aged and older adults.

Carbohydrate consumption is the principal determinant of insulin secretion. Eliminating high-glycemic foods, including processed carbohydrates and sweets, would sensitize insulin receptors and correct hyperinsulinemia. In addition, replacing high glycemic foods with fruits and vegetables would increase polyphenol intake. Epidemiologic evidence supports the idea that greater consumption of polyphenol-containing vegetables and fruits mitigates risk for neurocognitive decline and dementia.^34,35 Preclinical evidence suggests that such protection may be related to neuronal signaling effects and anti- inflammatory and antioxidant actions.³⁶ In addition, certain polyphenol compounds, such as those found in berries, enhance metabolic function.^37,38 In a 12-week pilot trial, older adults with early memory changes (N=9, mean age 76) who drank supplemental blueberry juice showed enhanced memory and improved metabolic parameters.³⁹

Dietary changes that preserve insulin receptor sensitivity can help ensure general health with aging and substantially mitigate risk for neurodegeneration. The Western diet is particularly insulinogenic and dietary habits are difficult to change. However, the substantial benefits, absence of adverse effects, and low cost make dietary intervention the optimal means of protecting against neurodegeneration and other age-related diseases. Embarking on such a program early in life would be best, although late-life intervention can be effective.

Related Resources

• Craft S, Watson GS. Insulin and neurodegenerative disease: shared and specific mechanisms. Lancet Neurol. 2004;3(3):169-178.
• Luchsinger JA, Tang MX, Shea S, et al. Hyperinsulinemia and risk of Alzheimer’s disease. Neurology. 2004; 63(7):1187-1192.
• Krikorian R, Shidler MD, Dangelo K, et al. Dietary ketosis enhances memory in mild cognitive impairment. Neurbiol Aging. 2012;33(2):425.e19-e27.

Disclosure

Dr. Krikorian receives grant support from the National Institutes of Health, 1R01AG034617-01.

Discuss this article at www.facebook.com/CurrentPsychiatry

In addition to increasing patients’ risk for cardiovascular disease, stroke, and cancer, obesity and metabolic disturbance contribute to age-related cognitive decline and dementia. In particular, insulin resistance and hyperinsulinemia promote neurocognitive dysfunction and neurodegenerative changes during the extended, preclinical phase of Alzheimer’s disease (AD). However, with dietary modification it may be possible to resensitize insulin receptors, correct hyperinsulinemia, and improve memory function.

Metabolic disturbance and neurodegeneration

In the United States, 5.4 million people have AD, and there will be an estimated 16 million cases by 2050.¹ Simultaneously we are experiencing an epidemic of metabolic disturbance and obesity. Approximately, 64% of adults in the United States are overweight (body mass index [BMI]: 25.0 to 29.9 kg/m²) and 34% are obese (BMI: ≥30 kg/m²).² By 2030, 86% of adults will be overweight and 51% will be obese.³ This confluence of epidemics is not coincidental but instead reflects the fact that metabolic disturbance is a fundamental factor contributing to cognitive decline and neurodegeneration.⁴

Ninety-six percent of AD cases are classified as late onset, sporadic AD, occurring after age 64.¹ Mild cognitive impairment (MCI) is a clinical construct that entails greater than expected memory impairment for the patient’s age and identifies older adults who are at increased risk for dementia. MCI represents the first clinical manifestation of neurodegeneration for a subset of patients who will progress to AD.^5,6 MCI is distinguished from age-associated memory impairment (AAMI), which originally was conceptualized as normal or benign memory decline with aging.^7,8 Recent data indicate that Alzheimer’s-type neuropathologic changes are the basis for subjective memory complaints and objectively assessed age-related cognitive decline,⁹ and early neurodegeneration is present in many patients with AAMI or MCI.¹⁰ This is consistent with the idea that an extended preclinical phase precedes AD onset. The preclinical phase can persist for a decade or more and precedes MCI and overt functional decline. However, neuropathologic changes accumulate during the preclinical phase of AD¹¹ and during the preclinical phase of type 2 diabetes mellitus (T2DM).

Hyperinsulinemia and dementia

Insulin resistance and hyperinsulinemia occur in >40% of individuals age ≥60 and prevalence increases with age.^4,12 Hyperinsulinemia develops to compensate for insulin resistance to overcome receptor insensitivity and maintain glucose homeostasis. Insulin receptors are densely expressed in brain regions vulnerable to neurodegeneration, including the medial temporal lobe and prefrontal cortex, which mediate long-term memory and working memory. However, insulin must be transported into the CNS from the periphery because little is synthesized in the brain. Paradoxically, peripheral compensatory hyperinsulinemia resulting from insulin resistance is associated with central (brain) hypoinsulinemia because of insensitivity and saturation of the receptor-mediated blood-brain barrier transport mechanism.^13-15

Hyperinsulinemia is the precursor to T2DM. However, hyperinsulinemia is not well recognized in clinical contexts and generally is not a treatment target. Nonetheless, it contributes to several health problems, and insulin resistance in middle age is associated with age-related diseases such as hypertension, coronary artery disease, stroke, and cancer, while insulin sensitivity protects against such disorders.¹⁶

Chronic insulin resistance may contribute more to dementia development than T2DM because of the extended period of hyperinsulinemia that precedes T2DM onset. In population studies,¹⁷ insulin resistance syndrome increases risk for developing AD independent of apolipoprotein E (APOE e4) allele status, and in a longitudinal study,¹⁸ the risk for AD solely attributable to peripheral hyperinsulinemia was up to 39%. Being overweight in midlife increases risk for dementia in late life, and APOE e4 allele status does not contribute additional risk after accounting for BMI.¹⁹ Middle-aged individuals with hyperinsulinemia show memory decline, and obesity in middle age was associated with greater cognitive impairment after 6-year follow-up.²⁰ Even in older adults who seem cognitively unimpaired, BMI and fasting insulin are positively correlated with atrophy in frontal, temporal, and subcortical brain regions, and obesity is an independent risk for atrophy in several brain regions, including the hippocampus.²¹

Compared with healthy older adults, individuals with AD have lower ratios of cerebrospinal fluid to plasma insulin.²² This lower ratio reflects the peripheral-to-central gradient of insulin levels in AD and suggests an etiological role for such metabolic disturbance. Insulin resistance has downstream effects that potentiate neurodegenerative factors, and central hypoinsulinemia can accelerate neurodegenerative processes and cognitive decline.^4,23 Brain insulin plays a direct role in regulating proinflammatory cytokines and neurotrophic and neuroplastic factors essential for memory function. Insulin degrading enzyme, which varies with insulin levels,²⁴ regulates the generation and clearance of amyloid β (Aβ) from the brain.²⁵

Hyperinsulinemia typically is evident in increasing waist circumference and body weight.²⁶ Waist circumference of ≥100 cm (39 inches) is a sensitive, specific, and independent predictor of hyperinsulinemia for men and women and a stronger predictor than BMI, waist-to-hip ratio, and other measures of body fat.²⁷ Unpublished data derived from our clinical research with MCI subjects supports the association of metabolic disturbance with age-related cognitive decline. Our subjects are recruited from the community on the basis of mild memory decline and—other than excluding those with diabetes—weight and metabolic status are not considered in evaluating individuals for enrollment. The Table contains data on waist circumference and metabolic function in 122 older adults (age ≥68) with MCI. On average, these individuals exhibited fasting insulin values in the hyperinsulinemia range and elevated fasting glucose levels that indicated borderline diabetes. Waist circumference also was high, indicating excessive visceral fat deposition. We also observed a relationship between waist circumference and insulin, a consistent observation in older adults with memory decline. These data would not be surprising in any sample of older adults because of the population base rates for these conditions. However, we also found that waist circumference was a significant predictor of memory performance in patients with MCI. Abdominal adiposity is highly correlated with intrahepatic fat.²⁸ Given this and recent indications that Alzheimer’s-type neuropathologic factors are generated in the liver,^29,30 the predictive value of waist circumference to memory performance may reflect the fact that it is a proxy for downstream actions of liver fat.

Table

Waist circumference and metabolic factors in 122 older adults with MCI^a

Dietary interventions

There is no cure for dementia, and it is not clear when effective therapy might be developed. Prevention and risk mitigation represent the best means of reducing the impact of this public health problem. Researchers have proposed that interventions initiated when individuals have predementia conditions such as AAMI and MCI might stall progression of cognitive decline, and MCI may be the last point when interventions might be effective because of the self-reinforcing neuropathologic cascades of AD.³¹ Because central hypoinsulinemia may promote central inflammation, Aβ generation, and reduced neuroplasticity, approaches aimed at improving metabolic function (and in particular correcting hyperinsulinemia) could influence fundamental neurodegenerative processes. Dietary approaches to preventing dementia are effective, low-risk, yet underutilized interventions. Reducing insulin by restricting calories³² or maintaining a ketogenic diet³³ has been associated with improved memory function in middle-aged and older adults.

Carbohydrate consumption is the principal determinant of insulin secretion. Eliminating high-glycemic foods, including processed carbohydrates and sweets, would sensitize insulin receptors and correct hyperinsulinemia. In addition, replacing high glycemic foods with fruits and vegetables would increase polyphenol intake. Epidemiologic evidence supports the idea that greater consumption of polyphenol-containing vegetables and fruits mitigates risk for neurocognitive decline and dementia.^34,35 Preclinical evidence suggests that such protection may be related to neuronal signaling effects and anti- inflammatory and antioxidant actions.³⁶ In addition, certain polyphenol compounds, such as those found in berries, enhance metabolic function.^37,38 In a 12-week pilot trial, older adults with early memory changes (N=9, mean age 76) who drank supplemental blueberry juice showed enhanced memory and improved metabolic parameters.³⁹

Dietary changes that preserve insulin receptor sensitivity can help ensure general health with aging and substantially mitigate risk for neurodegeneration. The Western diet is particularly insulinogenic and dietary habits are difficult to change. However, the substantial benefits, absence of adverse effects, and low cost make dietary intervention the optimal means of protecting against neurodegeneration and other age-related diseases. Embarking on such a program early in life would be best, although late-life intervention can be effective.

Related Resources

• Craft S, Watson GS. Insulin and neurodegenerative disease: shared and specific mechanisms. Lancet Neurol. 2004;3(3):169-178.
• Luchsinger JA, Tang MX, Shea S, et al. Hyperinsulinemia and risk of Alzheimer’s disease. Neurology. 2004; 63(7):1187-1192.
• Krikorian R, Shidler MD, Dangelo K, et al. Dietary ketosis enhances memory in mild cognitive impairment. Neurbiol Aging. 2012;33(2):425.e19-e27.

Disclosure

Dr. Krikorian receives grant support from the National Institutes of Health, 1R01AG034617-01.

Re-experiencing a previous life-threatening stress through nightmares or recurrent memories is a hallmark of posttraumatic stress disorder (PTSD). In the United States, the lifetime risk of PTSD is 10.1% and the 12-month prevalence is 3.7%.¹ The selective serotonin reuptake inhibitors (SSRIs) sertraline and paroxetine are FDA-approved for treating PTSD; clinicians commonly use any SSRI for this disorder. Although SSRIs can alleviate many PTSD symptoms, at times patients experience only a partial response, which necessitates other interventions.

Rationale for using topiramate

The anticonvulsant topiramate blocks voltage-sensitive sodium channels, augments γ-aminobutyric acid type A, antagonizes the glutamate receptor, and inhibits carbonic anhydrase. Researchers have hypothesized that limbic nuclei become sensitized and “kindled” after exposure to a traumatic event. Anticonvulsants such as topiramate may help mitigate stress-activated kindling in PTSD.^2,3

What does the evidence say?

Although less compelling than double-blind, placebo-controlled trials, small open-label studies and some case reports indicate a potential role for topiramate in PTSD for specific populations.^4,5 In an 8-week open- label study, Alderman et al⁶ found adjunctive topiramate led to a statistically significant reduction in Clinician-Administered PTSD Scale (CAPS) scores and nightmares in 43 male veterans with combat-related PTSD. There was a nonsignificant decrease in high-risk alcohol use.

In a 2002 retrospective case series, Berlant et al⁷ found topiramate as monotherapy or adjunctive therapy reduced nightmares in 35 patients with chronic, non-combat PTSD. Nightmares decreased in 79% of patients and flashbacks decreased in 86%, with symptom improvement in a median of 4 days. Limitations of this study included lack of placebo control, a low number of participants, and a high dropout rate (9/35).

Two years later, Berlant⁸ used the PTSD Checklist-Civilian version (PCL-C) to assess response to topiramate in an open-label study of 33 patients with chronic, non-hallucinatory PTSD. Twenty-eight patients used topiramate as add-on therapy. PCL-C scores decreased by ≥30% in 77% of patients in 4 weeks, with a median dose of 50 mg/d and a median response time of 9 days.

In a double-blind, placebo-controlled trial, Tucker et al⁹ assessed 38 civilian patients who took topiramate monotherapy for PTSD. Using the CAPS, researchers concluded that topiramate reduced re-experiencing symptoms, but the effect was not statistically significant.⁹

Lindley et al¹⁰ conducted a randomized, double-blind, placebo-controlled trial to study the effect of add-on topiramate in 40 patients with chronic, combat-related PTSD. Because many patients in this study had a history of depression and substance use disorders, topiramate was added to antidepressants; no anticonvulsants, antipsychotics, or benzodiazepines were used. Similar to previous studies, researchers found no statistically significant effect on PTSD symptom severity or global symptom improvement. However, the small number of participants and a high dropout rate limited this study.¹⁰

In a 12-week, double-blind, placebo-controlled study of 35 men and women age 18 to 62 with PTSD, Yeh et al¹¹ found that topiramate (mean dose: 102.94 mg/d) lead to a statistically significant overall CAPS score reduction, with significant improvements in re-experiencing symptoms, such as nightmares.

Our opinion

FDA-approved treatments such as SSRIs should be the first pharmacologic intervention for PTSD. If a patient’s response is partial or inadequate, consider additional treatment options. For patients with persistent re-experiencing symptoms, evidence and experience with prazosin and trazodone are more robust than that for topiramate.¹²

Using topiramate to reduce re-experiencing symptoms such as nightmares in PTSD is not supported by statistically significant evidence from double-blind, placebo- controlled trials. However, numerous open-label studies and case reports suggest that there may be a role for topiramate in PTSD patients who do not respond to other treatments. Data indicate that topiramate may be helpful for PTSD patients who have high-risk alcohol use⁶ or migraine headaches.¹³ Because some patients who take topiramate lose weight, the medication may be useful for PTSD patients who are overweight.¹³

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Related Resource

• U.S. Department of Veterans Affairs. Nightmares and PTSD. www.ptsd.va.gov/public/pages/nightmares.asp.

Drug Brand Names

• Paroxetine • Paxil
• Sertraline • Zoloft
• Prazosin • Minipress
• Topiramate • Topamax
• Trazodone • Desyrel, Oleptro

Re-experiencing a previous life-threatening stress through nightmares or recurrent memories is a hallmark of posttraumatic stress disorder (PTSD). In the United States, the lifetime risk of PTSD is 10.1% and the 12-month prevalence is 3.7%.¹ The selective serotonin reuptake inhibitors (SSRIs) sertraline and paroxetine are FDA-approved for treating PTSD; clinicians commonly use any SSRI for this disorder. Although SSRIs can alleviate many PTSD symptoms, at times patients experience only a partial response, which necessitates other interventions.

Rationale for using topiramate

The anticonvulsant topiramate blocks voltage-sensitive sodium channels, augments γ-aminobutyric acid type A, antagonizes the glutamate receptor, and inhibits carbonic anhydrase. Researchers have hypothesized that limbic nuclei become sensitized and “kindled” after exposure to a traumatic event. Anticonvulsants such as topiramate may help mitigate stress-activated kindling in PTSD.^2,3

What does the evidence say?

Although less compelling than double-blind, placebo-controlled trials, small open-label studies and some case reports indicate a potential role for topiramate in PTSD for specific populations.^4,5 In an 8-week open- label study, Alderman et al⁶ found adjunctive topiramate led to a statistically significant reduction in Clinician-Administered PTSD Scale (CAPS) scores and nightmares in 43 male veterans with combat-related PTSD. There was a nonsignificant decrease in high-risk alcohol use.

In a 2002 retrospective case series, Berlant et al⁷ found topiramate as monotherapy or adjunctive therapy reduced nightmares in 35 patients with chronic, non-combat PTSD. Nightmares decreased in 79% of patients and flashbacks decreased in 86%, with symptom improvement in a median of 4 days. Limitations of this study included lack of placebo control, a low number of participants, and a high dropout rate (9/35).

Two years later, Berlant⁸ used the PTSD Checklist-Civilian version (PCL-C) to assess response to topiramate in an open-label study of 33 patients with chronic, non-hallucinatory PTSD. Twenty-eight patients used topiramate as add-on therapy. PCL-C scores decreased by ≥30% in 77% of patients in 4 weeks, with a median dose of 50 mg/d and a median response time of 9 days.

In a double-blind, placebo-controlled trial, Tucker et al⁹ assessed 38 civilian patients who took topiramate monotherapy for PTSD. Using the CAPS, researchers concluded that topiramate reduced re-experiencing symptoms, but the effect was not statistically significant.⁹

Lindley et al¹⁰ conducted a randomized, double-blind, placebo-controlled trial to study the effect of add-on topiramate in 40 patients with chronic, combat-related PTSD. Because many patients in this study had a history of depression and substance use disorders, topiramate was added to antidepressants; no anticonvulsants, antipsychotics, or benzodiazepines were used. Similar to previous studies, researchers found no statistically significant effect on PTSD symptom severity or global symptom improvement. However, the small number of participants and a high dropout rate limited this study.¹⁰

In a 12-week, double-blind, placebo-controlled study of 35 men and women age 18 to 62 with PTSD, Yeh et al¹¹ found that topiramate (mean dose: 102.94 mg/d) lead to a statistically significant overall CAPS score reduction, with significant improvements in re-experiencing symptoms, such as nightmares.

Our opinion

FDA-approved treatments such as SSRIs should be the first pharmacologic intervention for PTSD. If a patient’s response is partial or inadequate, consider additional treatment options. For patients with persistent re-experiencing symptoms, evidence and experience with prazosin and trazodone are more robust than that for topiramate.¹²

Using topiramate to reduce re-experiencing symptoms such as nightmares in PTSD is not supported by statistically significant evidence from double-blind, placebo- controlled trials. However, numerous open-label studies and case reports suggest that there may be a role for topiramate in PTSD patients who do not respond to other treatments. Data indicate that topiramate may be helpful for PTSD patients who have high-risk alcohol use⁶ or migraine headaches.¹³ Because some patients who take topiramate lose weight, the medication may be useful for PTSD patients who are overweight.¹³

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Related Resource

• U.S. Department of Veterans Affairs. Nightmares and PTSD. www.ptsd.va.gov/public/pages/nightmares.asp.

Drug Brand Names

• Paroxetine • Paxil
• Sertraline • Zoloft
• Prazosin • Minipress
• Topiramate • Topamax
• Trazodone • Desyrel, Oleptro

Re-experiencing a previous life-threatening stress through nightmares or recurrent memories is a hallmark of posttraumatic stress disorder (PTSD). In the United States, the lifetime risk of PTSD is 10.1% and the 12-month prevalence is 3.7%.¹ The selective serotonin reuptake inhibitors (SSRIs) sertraline and paroxetine are FDA-approved for treating PTSD; clinicians commonly use any SSRI for this disorder. Although SSRIs can alleviate many PTSD symptoms, at times patients experience only a partial response, which necessitates other interventions.

Rationale for using topiramate

The anticonvulsant topiramate blocks voltage-sensitive sodium channels, augments γ-aminobutyric acid type A, antagonizes the glutamate receptor, and inhibits carbonic anhydrase. Researchers have hypothesized that limbic nuclei become sensitized and “kindled” after exposure to a traumatic event. Anticonvulsants such as topiramate may help mitigate stress-activated kindling in PTSD.^2,3

What does the evidence say?

Although less compelling than double-blind, placebo-controlled trials, small open-label studies and some case reports indicate a potential role for topiramate in PTSD for specific populations.^4,5 In an 8-week open- label study, Alderman et al⁶ found adjunctive topiramate led to a statistically significant reduction in Clinician-Administered PTSD Scale (CAPS) scores and nightmares in 43 male veterans with combat-related PTSD. There was a nonsignificant decrease in high-risk alcohol use.

In a 2002 retrospective case series, Berlant et al⁷ found topiramate as monotherapy or adjunctive therapy reduced nightmares in 35 patients with chronic, non-combat PTSD. Nightmares decreased in 79% of patients and flashbacks decreased in 86%, with symptom improvement in a median of 4 days. Limitations of this study included lack of placebo control, a low number of participants, and a high dropout rate (9/35).

Two years later, Berlant⁸ used the PTSD Checklist-Civilian version (PCL-C) to assess response to topiramate in an open-label study of 33 patients with chronic, non-hallucinatory PTSD. Twenty-eight patients used topiramate as add-on therapy. PCL-C scores decreased by ≥30% in 77% of patients in 4 weeks, with a median dose of 50 mg/d and a median response time of 9 days.

In a double-blind, placebo-controlled trial, Tucker et al⁹ assessed 38 civilian patients who took topiramate monotherapy for PTSD. Using the CAPS, researchers concluded that topiramate reduced re-experiencing symptoms, but the effect was not statistically significant.⁹

Lindley et al¹⁰ conducted a randomized, double-blind, placebo-controlled trial to study the effect of add-on topiramate in 40 patients with chronic, combat-related PTSD. Because many patients in this study had a history of depression and substance use disorders, topiramate was added to antidepressants; no anticonvulsants, antipsychotics, or benzodiazepines were used. Similar to previous studies, researchers found no statistically significant effect on PTSD symptom severity or global symptom improvement. However, the small number of participants and a high dropout rate limited this study.¹⁰

In a 12-week, double-blind, placebo-controlled study of 35 men and women age 18 to 62 with PTSD, Yeh et al¹¹ found that topiramate (mean dose: 102.94 mg/d) lead to a statistically significant overall CAPS score reduction, with significant improvements in re-experiencing symptoms, such as nightmares.

Our opinion

FDA-approved treatments such as SSRIs should be the first pharmacologic intervention for PTSD. If a patient’s response is partial or inadequate, consider additional treatment options. For patients with persistent re-experiencing symptoms, evidence and experience with prazosin and trazodone are more robust than that for topiramate.¹²

Using topiramate to reduce re-experiencing symptoms such as nightmares in PTSD is not supported by statistically significant evidence from double-blind, placebo- controlled trials. However, numerous open-label studies and case reports suggest that there may be a role for topiramate in PTSD patients who do not respond to other treatments. Data indicate that topiramate may be helpful for PTSD patients who have high-risk alcohol use⁶ or migraine headaches.¹³ Because some patients who take topiramate lose weight, the medication may be useful for PTSD patients who are overweight.¹³

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Related Resource

• U.S. Department of Veterans Affairs. Nightmares and PTSD. www.ptsd.va.gov/public/pages/nightmares.asp.

Drug Brand Names

• Paroxetine • Paxil
• Sertraline • Zoloft
• Prazosin • Minipress
• Topiramate • Topamax
• Trazodone • Desyrel, Oleptro

Some older patients with depression, anxiety, or insomnia may be reluctant to turn to pharmacotherapy and may prefer psychotherapeutic treatments.¹ Evidence has established cognitive-behavioral therapy (CBT) as an effective intervention for several psychiatric disorders and CBT should be considered when treating geriatric patients (Table 1).²

Table 1

Indications for CBT

Research evaluating the efficacy of CBT for depression in older adults was first published in the early 1980s. Since then, research and application of CBT with older adults has expanded to include other psychiatric disorders and researchers have suggested changes to increase the efficacy of CBT for these patients. This article provides:

• an overview of CBT’s efficacy for older adults with depression, anxiety, and insomnia
• modifications to employ when providing CBT to older patients.

The cognitive model of CBT

In the 1970s, Aaron T. Beck, MD, developed CBT while working with depressed patients. Beck’s patients reported thoughts characterized by inaccuracies and distortions in association with their depressed mood. He found these thoughts could be brought to the patient’s conscious attention and modified to improve the patient’s depression. This finding led to the development of CBT.

CBT is based on a cognitive model of the relationship among cognition, emotion, and behavior. Mood and behavior are viewed as determined by a person’s perception and interpretation of events, which manifest as a stream of automatically generated thoughts (Figure).³ These automatic thoughts have their origins in an underlying network of beliefs or schema. Patients with psychiatric disorders such as anxiety and depression typically have frequent automatic thoughts that characteristically lack validity because they arise from dysfunctional beliefs. The therapeutic process consists of helping the patient become aware of his or her internal stream of thoughts when distressed, and to identify and modify the dysfunctional thoughts. Behavioral techniques are used to bring about functional changes in behavior, regulate emotion, and help the cognitive restructuring process. Modifying the patient’s underlying dysfunctional beliefs leads to lasting improvements. In this structured therapy, the therapist and patient work collaboratively to use an approach that features reality testing and experimentation.⁴

Figure

The cognitive model of CBT

CBT: cognitive-behavioral therapy
Source: Adapted from reference 3

Indications for CBT in older adults

Depression. Among psychotherapies used in older adults, CBT has received the most research for late-life depression.⁵ Randomized controlled trials (RCTs) have found CBT is superior to treatment as usual in depressed adults age ≥60.⁶ It also has been found to be superior to wait-list control⁷ and talking as control.^6,8 Meta-analyses have shown above-average effect sizes for CBT in treating late-life depression.^9,10 A follow-up study found improvement was maintained up to 2 years after CBT, which suggests CBT’s impact is likely to be long lasting.¹¹

Thompson et al¹² compared 102 depressed patients age >60 who were treated with CBT alone, desipramine alone, or a combination of the 2. A combination of medication and CBT worked best for severely depressed patients; CBT alone or a combination of CBT and medication worked best for moderately depressed patients.

CBT is an option when treating depressed medically ill older adults. Research indicates that CBT could reduce depression in older patients with Parkinson’s disease¹³ and chronic obstructive pulmonary disease.¹⁴

As patients get older, cognitive impairment with comorbid depression can make treatment challenging. Limited research suggests CBT applied in a modified format that involves caregivers and uses problem solving and behavioral strategies can significantly reduce depression in patients with dementia.¹⁵

Anxiety. Researchers have examined the efficacy of variants of CBT in treating older adults with anxiety disorders—commonly, generalized anxiety disorder (GAD), panic disorder, agoraphobia, subjective anxiety, or a combination of these illnesses.^16,17 Randomized trials have supported CBT’s efficacy for older patients with GAD and mixed anxiety states; gains made in CBT were maintained over a 1-year follow-up.^18,19 In a meta-analysis of 15 studies using cognitive and behavioral methods of treating anxiety in older patients, Nordhus and Pallesen¹⁶ reported a significant effect size of 0.55. In a 2008 meta-analysis that included only RCTs, CBT was superior to wait-list conditions as well as active control conditions in treating anxious older patients.²⁰

However, some research suggests that CBT for GAD may not be as effective for older adults as it is for younger adults. In a study of CBT for GAD in older adults, Stanley et al¹⁹ reported smaller effect sizes compared with CBT for younger adults. Researchers have found relatively few differences between CBT and comparison conditions—supportive psychotherapy or active control conditions—in treating GAD in older adults.²¹ Modified, more effective formats of CBT for GAD in older adults need to be established.²² Mohlman et al²³ supplemented standard CBT for late-life GAD with memory and learning aids—weekly reading assignments, graphing exercises to chart mood ratings, reminder phone calls from therapists, and homework compliance requirement. This approach improved the response rate from 40% to 75%.²³

Insomnia. Studies have found CBT to be an effective means of treating insomnia in geriatric patients. Although sleep problems occur more frequently among older patients, only 15% of chronic insomnia patients receive treatment; psychotherapy rarely is used.²⁴ CBT for insomnia (CBT-I) should be considered for older adults because managing insomnia with medications may be problematic and these patients may prefer nonpharmacologic treatment.² CBT-I typically incorporates cognitive strategies with established behavioral techniques, including sleep hygiene education, cognitive restructuring, relaxation training, stimulus control, and/or sleep restriction. The CBT-I multicomponent treatment package meets all criteria to be considered an evidence-based treatment for late-life insomnia.²⁵

RCTs have reported significant improvements in late-life insomnia with CBT-I.^26,27 Reviews and meta-analyses have also concluded that cognitive-behavioral treatments are effective for treating insomnia in older adults.^25,28 Most insomnia cases in geriatric patients are reported to occur secondary to other medical or psychiatric conditions that are judged as causing the insomnia.²⁵ In these cases, direct treatment of the insomnia usually is delayed or omitted.²⁸ Studies evaluating the efficacy of CBT packages for treating insomnia occurring in conjunction with other medical or psychiatric illnesses have reported significant improvement of insomnia.^28,29 Because insomnia frequently occurs in older patients with medical illnesses and psychiatric disorders, CBT-I could be beneficial for such patients.

Good candidates for CBT

Clinical experience indicates that older adults in relatively good health with no significant cognitive decline are good candidates for CBT. These patients tend to comply with their assignments, are interested in applying the learned strategies, and are motivated to read self-help books. CBT’s structured, goal-oriented approach makes it a short-term treatment, which makes it cost effective. Insomnia patients may improve after 6 to 8 CBT-I sessions and patients with anxiety or depression may need to undergo 15 to 20 CBT sessions. Patients age ≥65 have basic Medicare coverage that includes mental health care and psychotherapy.

There are no absolute contraindications for CBT, but the greater the cognitive impairment, the less the patient will benefit from CBT (Table 2). Similarly, severe depression and anxiety might make it difficult for patients to participate meaningfully, although CBT may be incorporated gradually as patients improve with medication. Severe medical illnesses and sensory losses such as visual and hearing loss would make it difficult to carry out CBT effectively.

Table 2

Contraindications for CBT

Adapting CBT for older patients

When using CBT with older patients, it is important to keep in mind characteristics that define the geriatric population. Laidlaw et al³⁰ developed a model to help clinicians develop a more appropriate conceptualization of older patients that focuses on significant events and related cognitions associated with physical health, changes in role investments, and interactions with younger generations. It emphasizes the need to explore beliefs about aging viewed through each patient’s socio-cultural lens and examine cognitions in the context of the time period in which the individual has lived.

Losses and transitions. For many older patients, the latter years of life are characterized by losses and transitions.³¹ According to Thompson,³¹ these losses and transitions can trigger thoughts of missed opportunities or unresolved relationships and reflection on unachieved goals.³¹ CBT for older adults should focus on the meaning the patient gives to these losses and transitions. For example, depressed patients could view their retirement as a loss of self worth as they become less productive. CBT can help patients identify ways of thinking about the situation that will enable them to adapt to these losses and transitions.

Changes in cognition. Changes in cognitive functioning with aging are not universal and there’s considerable variability, but it’s important to make appropriate adaptations when needed. Patients may experience a decline in cognitive speed, working memory, selective attention, and fluid intelligence. This would require that information be presented slowly, with frequent repetitions and summaries. Also, it might be helpful to present information in alternate ways and to encourage patients to take notes during sessions. To accommodate for a decline in fluid intelligence, presenting new information in the context of previous experiences will help promote learning. Recordings of important information and conclusions from cognitive restructuring that patients can listen to between sessions could serve as helpful reminders that will help patients progress. Phone prompts or alarms can remind patients to carry out certain therapeutic measures, such as breathing exercises. Caretakers can attend sessions to become familiar with strategies performed during CBT and act as a co-therapist at home; however, their inclusion must be done with the consent of both parties and only if it’s viewed as necessary for the patient’s progress.

Additional strategies. For patients with substantial cognitive decline, cognitive restructuring might not be as effective as behavioral strategies—activity scheduling, graded task assignment, graded exposure, and rehearsals. Because older adults often have strengthened dysfunctional beliefs over a long time, modifying them takes longer, which is why the tapering process usually takes longer for older patients than for younger patients. The lengthier tapering ensures learning is well established and the process of modifying dysfunctional beliefs to functional beliefs continues. Collaborating with other professionals—physicians, social workers, and case managers—will help ensure a shared care process in which common goals are met.

The websites of the Academy of Cognitive Therapy, American Psychological Association, and Association for Behavioral and Cognitive Therapies can help clinicians who do not offer CBT to locate a qualified therapist for their patients (Related Resources).

Related Resources

• Academy of Cognitive Therapy. www.academyofct.org.
• American Psychological Association. www.apa.org.
• Association for Behavioral and Cognitive Therapies. www.abct.org.
• Laidlaw K, Thompson LW, Dick-Siskin L, et al. Cognitive behaviour therapy with older people. West Sussex, England: John Wiley & Sons, Ltd; 2003.

Drug Brand Name

• Desipramine • Norpramin

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Some older patients with depression, anxiety, or insomnia may be reluctant to turn to pharmacotherapy and may prefer psychotherapeutic treatments.¹ Evidence has established cognitive-behavioral therapy (CBT) as an effective intervention for several psychiatric disorders and CBT should be considered when treating geriatric patients (Table 1).²

Table 1

Indications for CBT

Research evaluating the efficacy of CBT for depression in older adults was first published in the early 1980s. Since then, research and application of CBT with older adults has expanded to include other psychiatric disorders and researchers have suggested changes to increase the efficacy of CBT for these patients. This article provides:

• an overview of CBT’s efficacy for older adults with depression, anxiety, and insomnia
• modifications to employ when providing CBT to older patients.

The cognitive model of CBT

In the 1970s, Aaron T. Beck, MD, developed CBT while working with depressed patients. Beck’s patients reported thoughts characterized by inaccuracies and distortions in association with their depressed mood. He found these thoughts could be brought to the patient’s conscious attention and modified to improve the patient’s depression. This finding led to the development of CBT.

CBT is based on a cognitive model of the relationship among cognition, emotion, and behavior. Mood and behavior are viewed as determined by a person’s perception and interpretation of events, which manifest as a stream of automatically generated thoughts (Figure).³ These automatic thoughts have their origins in an underlying network of beliefs or schema. Patients with psychiatric disorders such as anxiety and depression typically have frequent automatic thoughts that characteristically lack validity because they arise from dysfunctional beliefs. The therapeutic process consists of helping the patient become aware of his or her internal stream of thoughts when distressed, and to identify and modify the dysfunctional thoughts. Behavioral techniques are used to bring about functional changes in behavior, regulate emotion, and help the cognitive restructuring process. Modifying the patient’s underlying dysfunctional beliefs leads to lasting improvements. In this structured therapy, the therapist and patient work collaboratively to use an approach that features reality testing and experimentation.⁴

Figure

The cognitive model of CBT

CBT: cognitive-behavioral therapy
Source: Adapted from reference 3

Indications for CBT in older adults

Depression. Among psychotherapies used in older adults, CBT has received the most research for late-life depression.⁵ Randomized controlled trials (RCTs) have found CBT is superior to treatment as usual in depressed adults age ≥60.⁶ It also has been found to be superior to wait-list control⁷ and talking as control.^6,8 Meta-analyses have shown above-average effect sizes for CBT in treating late-life depression.^9,10 A follow-up study found improvement was maintained up to 2 years after CBT, which suggests CBT’s impact is likely to be long lasting.¹¹

Thompson et al¹² compared 102 depressed patients age >60 who were treated with CBT alone, desipramine alone, or a combination of the 2. A combination of medication and CBT worked best for severely depressed patients; CBT alone or a combination of CBT and medication worked best for moderately depressed patients.

CBT is an option when treating depressed medically ill older adults. Research indicates that CBT could reduce depression in older patients with Parkinson’s disease¹³ and chronic obstructive pulmonary disease.¹⁴

As patients get older, cognitive impairment with comorbid depression can make treatment challenging. Limited research suggests CBT applied in a modified format that involves caregivers and uses problem solving and behavioral strategies can significantly reduce depression in patients with dementia.¹⁵

Anxiety. Researchers have examined the efficacy of variants of CBT in treating older adults with anxiety disorders—commonly, generalized anxiety disorder (GAD), panic disorder, agoraphobia, subjective anxiety, or a combination of these illnesses.^16,17 Randomized trials have supported CBT’s efficacy for older patients with GAD and mixed anxiety states; gains made in CBT were maintained over a 1-year follow-up.^18,19 In a meta-analysis of 15 studies using cognitive and behavioral methods of treating anxiety in older patients, Nordhus and Pallesen¹⁶ reported a significant effect size of 0.55. In a 2008 meta-analysis that included only RCTs, CBT was superior to wait-list conditions as well as active control conditions in treating anxious older patients.²⁰

However, some research suggests that CBT for GAD may not be as effective for older adults as it is for younger adults. In a study of CBT for GAD in older adults, Stanley et al¹⁹ reported smaller effect sizes compared with CBT for younger adults. Researchers have found relatively few differences between CBT and comparison conditions—supportive psychotherapy or active control conditions—in treating GAD in older adults.²¹ Modified, more effective formats of CBT for GAD in older adults need to be established.²² Mohlman et al²³ supplemented standard CBT for late-life GAD with memory and learning aids—weekly reading assignments, graphing exercises to chart mood ratings, reminder phone calls from therapists, and homework compliance requirement. This approach improved the response rate from 40% to 75%.²³

Insomnia. Studies have found CBT to be an effective means of treating insomnia in geriatric patients. Although sleep problems occur more frequently among older patients, only 15% of chronic insomnia patients receive treatment; psychotherapy rarely is used.²⁴ CBT for insomnia (CBT-I) should be considered for older adults because managing insomnia with medications may be problematic and these patients may prefer nonpharmacologic treatment.² CBT-I typically incorporates cognitive strategies with established behavioral techniques, including sleep hygiene education, cognitive restructuring, relaxation training, stimulus control, and/or sleep restriction. The CBT-I multicomponent treatment package meets all criteria to be considered an evidence-based treatment for late-life insomnia.²⁵

RCTs have reported significant improvements in late-life insomnia with CBT-I.^26,27 Reviews and meta-analyses have also concluded that cognitive-behavioral treatments are effective for treating insomnia in older adults.^25,28 Most insomnia cases in geriatric patients are reported to occur secondary to other medical or psychiatric conditions that are judged as causing the insomnia.²⁵ In these cases, direct treatment of the insomnia usually is delayed or omitted.²⁸ Studies evaluating the efficacy of CBT packages for treating insomnia occurring in conjunction with other medical or psychiatric illnesses have reported significant improvement of insomnia.^28,29 Because insomnia frequently occurs in older patients with medical illnesses and psychiatric disorders, CBT-I could be beneficial for such patients.

Good candidates for CBT

Clinical experience indicates that older adults in relatively good health with no significant cognitive decline are good candidates for CBT. These patients tend to comply with their assignments, are interested in applying the learned strategies, and are motivated to read self-help books. CBT’s structured, goal-oriented approach makes it a short-term treatment, which makes it cost effective. Insomnia patients may improve after 6 to 8 CBT-I sessions and patients with anxiety or depression may need to undergo 15 to 20 CBT sessions. Patients age ≥65 have basic Medicare coverage that includes mental health care and psychotherapy.

There are no absolute contraindications for CBT, but the greater the cognitive impairment, the less the patient will benefit from CBT (Table 2). Similarly, severe depression and anxiety might make it difficult for patients to participate meaningfully, although CBT may be incorporated gradually as patients improve with medication. Severe medical illnesses and sensory losses such as visual and hearing loss would make it difficult to carry out CBT effectively.

Table 2

Contraindications for CBT

Adapting CBT for older patients

When using CBT with older patients, it is important to keep in mind characteristics that define the geriatric population. Laidlaw et al³⁰ developed a model to help clinicians develop a more appropriate conceptualization of older patients that focuses on significant events and related cognitions associated with physical health, changes in role investments, and interactions with younger generations. It emphasizes the need to explore beliefs about aging viewed through each patient’s socio-cultural lens and examine cognitions in the context of the time period in which the individual has lived.

Losses and transitions. For many older patients, the latter years of life are characterized by losses and transitions.³¹ According to Thompson,³¹ these losses and transitions can trigger thoughts of missed opportunities or unresolved relationships and reflection on unachieved goals.³¹ CBT for older adults should focus on the meaning the patient gives to these losses and transitions. For example, depressed patients could view their retirement as a loss of self worth as they become less productive. CBT can help patients identify ways of thinking about the situation that will enable them to adapt to these losses and transitions.

Changes in cognition. Changes in cognitive functioning with aging are not universal and there’s considerable variability, but it’s important to make appropriate adaptations when needed. Patients may experience a decline in cognitive speed, working memory, selective attention, and fluid intelligence. This would require that information be presented slowly, with frequent repetitions and summaries. Also, it might be helpful to present information in alternate ways and to encourage patients to take notes during sessions. To accommodate for a decline in fluid intelligence, presenting new information in the context of previous experiences will help promote learning. Recordings of important information and conclusions from cognitive restructuring that patients can listen to between sessions could serve as helpful reminders that will help patients progress. Phone prompts or alarms can remind patients to carry out certain therapeutic measures, such as breathing exercises. Caretakers can attend sessions to become familiar with strategies performed during CBT and act as a co-therapist at home; however, their inclusion must be done with the consent of both parties and only if it’s viewed as necessary for the patient’s progress.

Additional strategies. For patients with substantial cognitive decline, cognitive restructuring might not be as effective as behavioral strategies—activity scheduling, graded task assignment, graded exposure, and rehearsals. Because older adults often have strengthened dysfunctional beliefs over a long time, modifying them takes longer, which is why the tapering process usually takes longer for older patients than for younger patients. The lengthier tapering ensures learning is well established and the process of modifying dysfunctional beliefs to functional beliefs continues. Collaborating with other professionals—physicians, social workers, and case managers—will help ensure a shared care process in which common goals are met.

The websites of the Academy of Cognitive Therapy, American Psychological Association, and Association for Behavioral and Cognitive Therapies can help clinicians who do not offer CBT to locate a qualified therapist for their patients (Related Resources).

Related Resources

• Academy of Cognitive Therapy. www.academyofct.org.
• American Psychological Association. www.apa.org.
• Association for Behavioral and Cognitive Therapies. www.abct.org.
• Laidlaw K, Thompson LW, Dick-Siskin L, et al. Cognitive behaviour therapy with older people. West Sussex, England: John Wiley & Sons, Ltd; 2003.

Drug Brand Name

• Desipramine • Norpramin

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Some older patients with depression, anxiety, or insomnia may be reluctant to turn to pharmacotherapy and may prefer psychotherapeutic treatments.¹ Evidence has established cognitive-behavioral therapy (CBT) as an effective intervention for several psychiatric disorders and CBT should be considered when treating geriatric patients (Table 1).²

Table 1

Indications for CBT

Research evaluating the efficacy of CBT for depression in older adults was first published in the early 1980s. Since then, research and application of CBT with older adults has expanded to include other psychiatric disorders and researchers have suggested changes to increase the efficacy of CBT for these patients. This article provides:

• an overview of CBT’s efficacy for older adults with depression, anxiety, and insomnia
• modifications to employ when providing CBT to older patients.

The cognitive model of CBT

In the 1970s, Aaron T. Beck, MD, developed CBT while working with depressed patients. Beck’s patients reported thoughts characterized by inaccuracies and distortions in association with their depressed mood. He found these thoughts could be brought to the patient’s conscious attention and modified to improve the patient’s depression. This finding led to the development of CBT.

CBT is based on a cognitive model of the relationship among cognition, emotion, and behavior. Mood and behavior are viewed as determined by a person’s perception and interpretation of events, which manifest as a stream of automatically generated thoughts (Figure).³ These automatic thoughts have their origins in an underlying network of beliefs or schema. Patients with psychiatric disorders such as anxiety and depression typically have frequent automatic thoughts that characteristically lack validity because they arise from dysfunctional beliefs. The therapeutic process consists of helping the patient become aware of his or her internal stream of thoughts when distressed, and to identify and modify the dysfunctional thoughts. Behavioral techniques are used to bring about functional changes in behavior, regulate emotion, and help the cognitive restructuring process. Modifying the patient’s underlying dysfunctional beliefs leads to lasting improvements. In this structured therapy, the therapist and patient work collaboratively to use an approach that features reality testing and experimentation.⁴

Figure

The cognitive model of CBT

CBT: cognitive-behavioral therapy
Source: Adapted from reference 3

Indications for CBT in older adults

Depression. Among psychotherapies used in older adults, CBT has received the most research for late-life depression.⁵ Randomized controlled trials (RCTs) have found CBT is superior to treatment as usual in depressed adults age ≥60.⁶ It also has been found to be superior to wait-list control⁷ and talking as control.^6,8 Meta-analyses have shown above-average effect sizes for CBT in treating late-life depression.^9,10 A follow-up study found improvement was maintained up to 2 years after CBT, which suggests CBT’s impact is likely to be long lasting.¹¹

Thompson et al¹² compared 102 depressed patients age >60 who were treated with CBT alone, desipramine alone, or a combination of the 2. A combination of medication and CBT worked best for severely depressed patients; CBT alone or a combination of CBT and medication worked best for moderately depressed patients.

CBT is an option when treating depressed medically ill older adults. Research indicates that CBT could reduce depression in older patients with Parkinson’s disease¹³ and chronic obstructive pulmonary disease.¹⁴

As patients get older, cognitive impairment with comorbid depression can make treatment challenging. Limited research suggests CBT applied in a modified format that involves caregivers and uses problem solving and behavioral strategies can significantly reduce depression in patients with dementia.¹⁵

Anxiety. Researchers have examined the efficacy of variants of CBT in treating older adults with anxiety disorders—commonly, generalized anxiety disorder (GAD), panic disorder, agoraphobia, subjective anxiety, or a combination of these illnesses.^16,17 Randomized trials have supported CBT’s efficacy for older patients with GAD and mixed anxiety states; gains made in CBT were maintained over a 1-year follow-up.^18,19 In a meta-analysis of 15 studies using cognitive and behavioral methods of treating anxiety in older patients, Nordhus and Pallesen¹⁶ reported a significant effect size of 0.55. In a 2008 meta-analysis that included only RCTs, CBT was superior to wait-list conditions as well as active control conditions in treating anxious older patients.²⁰

However, some research suggests that CBT for GAD may not be as effective for older adults as it is for younger adults. In a study of CBT for GAD in older adults, Stanley et al¹⁹ reported smaller effect sizes compared with CBT for younger adults. Researchers have found relatively few differences between CBT and comparison conditions—supportive psychotherapy or active control conditions—in treating GAD in older adults.²¹ Modified, more effective formats of CBT for GAD in older adults need to be established.²² Mohlman et al²³ supplemented standard CBT for late-life GAD with memory and learning aids—weekly reading assignments, graphing exercises to chart mood ratings, reminder phone calls from therapists, and homework compliance requirement. This approach improved the response rate from 40% to 75%.²³

Insomnia. Studies have found CBT to be an effective means of treating insomnia in geriatric patients. Although sleep problems occur more frequently among older patients, only 15% of chronic insomnia patients receive treatment; psychotherapy rarely is used.²⁴ CBT for insomnia (CBT-I) should be considered for older adults because managing insomnia with medications may be problematic and these patients may prefer nonpharmacologic treatment.² CBT-I typically incorporates cognitive strategies with established behavioral techniques, including sleep hygiene education, cognitive restructuring, relaxation training, stimulus control, and/or sleep restriction. The CBT-I multicomponent treatment package meets all criteria to be considered an evidence-based treatment for late-life insomnia.²⁵

RCTs have reported significant improvements in late-life insomnia with CBT-I.^26,27 Reviews and meta-analyses have also concluded that cognitive-behavioral treatments are effective for treating insomnia in older adults.^25,28 Most insomnia cases in geriatric patients are reported to occur secondary to other medical or psychiatric conditions that are judged as causing the insomnia.²⁵ In these cases, direct treatment of the insomnia usually is delayed or omitted.²⁸ Studies evaluating the efficacy of CBT packages for treating insomnia occurring in conjunction with other medical or psychiatric illnesses have reported significant improvement of insomnia.^28,29 Because insomnia frequently occurs in older patients with medical illnesses and psychiatric disorders, CBT-I could be beneficial for such patients.

Good candidates for CBT

Clinical experience indicates that older adults in relatively good health with no significant cognitive decline are good candidates for CBT. These patients tend to comply with their assignments, are interested in applying the learned strategies, and are motivated to read self-help books. CBT’s structured, goal-oriented approach makes it a short-term treatment, which makes it cost effective. Insomnia patients may improve after 6 to 8 CBT-I sessions and patients with anxiety or depression may need to undergo 15 to 20 CBT sessions. Patients age ≥65 have basic Medicare coverage that includes mental health care and psychotherapy.

There are no absolute contraindications for CBT, but the greater the cognitive impairment, the less the patient will benefit from CBT (Table 2). Similarly, severe depression and anxiety might make it difficult for patients to participate meaningfully, although CBT may be incorporated gradually as patients improve with medication. Severe medical illnesses and sensory losses such as visual and hearing loss would make it difficult to carry out CBT effectively.

Table 2

Contraindications for CBT

Adapting CBT for older patients

When using CBT with older patients, it is important to keep in mind characteristics that define the geriatric population. Laidlaw et al³⁰ developed a model to help clinicians develop a more appropriate conceptualization of older patients that focuses on significant events and related cognitions associated with physical health, changes in role investments, and interactions with younger generations. It emphasizes the need to explore beliefs about aging viewed through each patient’s socio-cultural lens and examine cognitions in the context of the time period in which the individual has lived.

Losses and transitions. For many older patients, the latter years of life are characterized by losses and transitions.³¹ According to Thompson,³¹ these losses and transitions can trigger thoughts of missed opportunities or unresolved relationships and reflection on unachieved goals.³¹ CBT for older adults should focus on the meaning the patient gives to these losses and transitions. For example, depressed patients could view their retirement as a loss of self worth as they become less productive. CBT can help patients identify ways of thinking about the situation that will enable them to adapt to these losses and transitions.

Changes in cognition. Changes in cognitive functioning with aging are not universal and there’s considerable variability, but it’s important to make appropriate adaptations when needed. Patients may experience a decline in cognitive speed, working memory, selective attention, and fluid intelligence. This would require that information be presented slowly, with frequent repetitions and summaries. Also, it might be helpful to present information in alternate ways and to encourage patients to take notes during sessions. To accommodate for a decline in fluid intelligence, presenting new information in the context of previous experiences will help promote learning. Recordings of important information and conclusions from cognitive restructuring that patients can listen to between sessions could serve as helpful reminders that will help patients progress. Phone prompts or alarms can remind patients to carry out certain therapeutic measures, such as breathing exercises. Caretakers can attend sessions to become familiar with strategies performed during CBT and act as a co-therapist at home; however, their inclusion must be done with the consent of both parties and only if it’s viewed as necessary for the patient’s progress.

Additional strategies. For patients with substantial cognitive decline, cognitive restructuring might not be as effective as behavioral strategies—activity scheduling, graded task assignment, graded exposure, and rehearsals. Because older adults often have strengthened dysfunctional beliefs over a long time, modifying them takes longer, which is why the tapering process usually takes longer for older patients than for younger patients. The lengthier tapering ensures learning is well established and the process of modifying dysfunctional beliefs to functional beliefs continues. Collaborating with other professionals—physicians, social workers, and case managers—will help ensure a shared care process in which common goals are met.

The websites of the Academy of Cognitive Therapy, American Psychological Association, and Association for Behavioral and Cognitive Therapies can help clinicians who do not offer CBT to locate a qualified therapist for their patients (Related Resources).

Related Resources

• Academy of Cognitive Therapy. www.academyofct.org.
• American Psychological Association. www.apa.org.
• Association for Behavioral and Cognitive Therapies. www.abct.org.
• Laidlaw K, Thompson LW, Dick-Siskin L, et al. Cognitive behaviour therapy with older people. West Sussex, England: John Wiley & Sons, Ltd; 2003.

Drug Brand Name

• Desipramine • Norpramin

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

The autoimmune hemolytic anemias (AIHA) are rare but important hematologic diseases. They can range in severity from mildly symptomatic illness to a rapidly fatal syndrome. The incidence of AIHA is estimated to be between 0.6 and 3 cases per 100,000 persons. AIHA is mediated by antibodies, and in the majority of cases immunglobulin (Ig) G is the mediating antibody. This type of AIHA is referred to as "warm" AIHA because IgG antibodies bind best at body temperature. "Cold" AIHA is mediated by IgM antibodies, which bind maximally at temperatures below 37°C. This manual reviews the most common types of AIHA, with emphasis on diagnosis and treatment.

To read the full article in PDF:

Click here

The autoimmune hemolytic anemias (AIHA) are rare but important hematologic diseases. They can range in severity from mildly symptomatic illness to a rapidly fatal syndrome. The incidence of AIHA is estimated to be between 0.6 and 3 cases per 100,000 persons. AIHA is mediated by antibodies, and in the majority of cases immunglobulin (Ig) G is the mediating antibody. This type of AIHA is referred to as "warm" AIHA because IgG antibodies bind best at body temperature. "Cold" AIHA is mediated by IgM antibodies, which bind maximally at temperatures below 37°C. This manual reviews the most common types of AIHA, with emphasis on diagnosis and treatment.

To read the full article in PDF:

Click here

The autoimmune hemolytic anemias (AIHA) are rare but important hematologic diseases. They can range in severity from mildly symptomatic illness to a rapidly fatal syndrome. The incidence of AIHA is estimated to be between 0.6 and 3 cases per 100,000 persons. AIHA is mediated by antibodies, and in the majority of cases immunglobulin (Ig) G is the mediating antibody. This type of AIHA is referred to as "warm" AIHA because IgG antibodies bind best at body temperature. "Cold" AIHA is mediated by IgM antibodies, which bind maximally at temperatures below 37°C. This manual reviews the most common types of AIHA, with emphasis on diagnosis and treatment.

To read the full article in PDF:

Click here

Whenever a patient begins treatment with a glucocorticoid drug, we need to think about bone loss.

The American College of Rheumatology (ACR) issued recommendations for preventing and treating glucocorticoid-induced osteoporosis in 2010.¹ Compared with its previous guidelines,² the new ones are more tailored and nuanced but may be more difficult for physicians to follow. The guidelines call for assessing fracture risk using the computer-based Fracture Risk Assessment Tool, or FRAX (www/shef.ac.uk/FRAX), developed by the World Health Organization (WHO). For those without a computer or ready access to the Web, an application of FRAX is available for download on smartphones.

In this article, my purpose is to review the new recommendations and to offer my perspective, which does not necessarily reflect the opinions of the ACR.

DESPITE EVIDENCE, MANY PATIENTS RECEIVE NO INTERVENTION

Use of glucocorticoids is the most common cause of secondary osteoporosis. During the first 6 to 12 months of use, these drugs can cause a rapid loss of bone mass due to increased bone resorption; with continued use, they cause a slower but steady decline in bone mass due to reduced bone formation.³ Epidemiologic studies have found that the risk of fractures increases with dose, starting with doses as low as 2.5 mg per day of prednisone or its equivalent.⁴

Numerous clinical trials have evaluated the effect of bisphosphonates and teriparatide (Forteo) on bone mass and fracture risk in patients on glucocorticoid therapy. The bisphosphonates alendronate (Fosamax) and risedronate (Actonel) have both been shown to increase bone mass and reduce vertebral fracture risk in glucocorticoid recipients.^5–8 Zoledronic acid (Reclast), a parenteral bisphosphonate given in one annual dose, was shown to increase bone mass more than oral risedronate taken daily,⁹ and teriparatide, a formulation of parathyroid hormone, was better than alendronate.¹⁰

However, despite the known risk of fractures with glucocorticoid use and the demonstrated efficacy of available agents in preventing bone loss and fracture, many patients do not receive any intervention.^11,12

WHAT HAS HAPPENED SINCE 2001?

In the interval since 2001, several guidelines for managing glucocorticoid-induced osteoporosis have been published in other countries.^13–17 Broadly speaking, they recommend starting preventive drug therapy for patients at risk of fracture at the same time glucocorticoid drugs are started if the patient is expected to take glucocorticoids for more than 3 to 6 months in doses higher than 5 to 7.5 mg of prednisone or its equivalent daily.

Recommendations for patients who have been on glucocorticoids for longer than 3 to 6 months at initial evaluation have been based largely on T scores derived from dual-energy x-ray absorptiometry (DXA). Thresholds for initiating therapy have varied: the ACR in 2001 recommended preventive treatment if the T score is lower than −1.0, whereas British guidelines said −1.5 and Dutch guidelines said −2.5.

In the United States, since 2001 when the ACR published its last guidelines,² zoledronic acid and teriparatide have been approved for use in glucocorticoid-induced osteoporosis. In addition, guideline-development methodology has evolved and now is more scientifically rigorous. Finally, a risk-assessment tool has been developed that enables a more tailored approach (see below).

FRAX (www.shef.ac.uk/FRAX)

FRAX is a tool developed by the WHO to calculate the risk of fracture. If you go to the FRAX Web site and enter the required clinical information (race, age, sex, weight, height, previous fracture, family history of a fractured hip in a parent, current smoking, use of glucocorticoids, rheumatoid arthritis, secondary osteoporosis, consumption of three or more units of alcohol per day, and bone mineral density of the femoral neck), it will tell you the patient’s 10-year absolute (not relative) risk of major osteoporotic fracture and of hip fracture.

Since FRAX was unveiled in 2008, calculation of absolute fracture risk has become the standard method for making treatment decisions in patients with low bone mass who have not yet received any fracture-preventing treatment.¹⁸ The use of clinical risk factors in FRAX increases its ability to predict risk over and above the use of bone density by itself. And glucocorticoids are one of the clinical risk factors in FRAX.

But in which patients is treatment with a bisphosphonate or teriparatide cost-effective?

Thresholds for cost-effectiveness have been developed on the basis of economic assumptions that are country-specific. In the United States, the National Osteoporosis Foundation recommends drug therapy if the 10-year absolute risk of a major osteoporotic fracture of the hip, spine (clinical, not radiographic), wrist, or humerus is greater than 20% or if the risk of a hip fracture is greater than 3%.¹⁹

At equivalent bone densities, women taking glucocorticoids are at considerably higher risk of fracture than nonusers.²⁰ For example, consider a 65-year-old white woman, weight 59 kg, height 163 cm, no previous fractures, no parent with a fractured hip, no current smoking, no rheumatoid arthritis, no secondary osteoporosis, no excessive alcohol use, and a T score of −2.2 in the femoral neck. (Try this on the FRAX Web site.) If she does not use glucocorticoids, her 10-year risk of hip fracture is 2.0%; using glucocorticoids increases the risk to 3.6%. This is higher than the 3% National Osteoporosis Foundation guideline; thus, treatment would be recommended.

Also using FRAX, a 55-year-old white woman with a T score of −1.8 and on glucocorticoid therapy has a 67% higher risk of major osteoporotic fracture and an 80% higher risk of hip fracture.

For a third example, a white woman age 60, weight 70 kg, height 168 cm, negative for all the other risk factors but with a T score of −2.1 and on glucocorticoids has a calculated 10-year fracture risk of 2.1%, which is below the National Osteoporosis Foundation treatment threshold. However, most clinicians would probably recommend treatment for her, depending on the anticipated dose and duration of glucocorticoid therapy.

A caveat. In FRAX, glucocorticoid therapy is a categorical variable—a yes-or-no question—and yes is defined as having ever used a glucocorticoid in a dose greater than 5 mg for more than 3 months. Therefore, according to FRAX, a patient who took 5 mg of prednisone for 3 months 5 years ago has the same fracture risk as a patient on 60 mg of prednisone after a diagnosis of temporal arteritis. For this reason, the FRAX tool is likely to underestimate fracture risk, especially in patients currently taking glucocorticoids and those on higher doses of these drugs.

Kanis et al used the General Practice Research Database to adjust the fracture risk for glucocorticoid use in FRAX.²¹ At doses higher than 7.5 mg, the fracture risk had to be revised upward by 10% to 25% depending on the fracture site (hip vs any major osteoporotic fracture) and age (greater at age 40 than at age 90).

The underestimation of fracture risk led the ACR Expert Advisory Panel to create risk strata for major osteoporotic fractures, ie, low (< 10% risk per 10 years), medium (10%–20%), and high (> 20%) and uses these cut points to make treatment recommendations.

HOW THE 2010 GUIDELINES WERE DEVELOPED

Whereas the 2001 recommendations were based on a more informal consensus approach, the 2010 recommendations use a more scientifically rigorous methodology for guideline development, the Research and Development/University of California at Los Angeles (RAND/UCLA) Appropriateness Method. The RAND/UCLA method combines the best available scientific evidence with expert opinion to develop practice guidelines.

In drawing up the 2010 recommendations the ACR used three panels of experts. The Core Executive Panel conducted a systematic review of controlled clinical trials of therapies currently approved for treating glucocorticoid-induced osteoporosis in the United States, Canada, or the European Union. They found 53 articles meeting their inclusion criteria; an evidence report was produced that informed the development of the recommendations. This evidence report and guideline development process is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658. The Expert Advisory Panel framed the recommendations, and the Task Force Panel voted on them. The Core Executive Panel and Expert Advisory Panel constructed 48 patient-specific clinical scenarios using four variables: sex, age, race/ethnicity, and femoral neck T scores.

The members of the Task Force Panel were asked to use the evidence report and their expert judgment to vote on and rate the appropriateness of using a specific therapy in the context of each scenario on a 9-point Likert scale (1 = appropriate; 9 = not appropriate). Agreement occurred when 7 or more of the 10 panel members rated a scenario 1, 2, or 3. Disagreements were defined as 3 or more of the 10 members rating the scenario between 4 and 9 while the other members rated it lower.

Disagreements in voting were discussed in an attempt to achieve consensus, and a second vote was conducted which determined the final recommendations. If disagreement remained after the vote, no recommendation was made.

No attempt was made to assign priority of one drug over another when multiple drugs were deemed appropriate, although the final recommendations did differentiate drugs based on patient categories.

START WITH COUNSELING, ASSESSMENT

For patients starting or already on glucocorticoid therapy that is expected to last at least 3 months, the first step is to counsel them on lifestyle modifications (Table 1) and to assess their risk factors (Figure 1). Recommendations for monitoring patients receiving glucocorticoid therapy for at least 3 months are presented in Table 2.

These recommendations are based on literature review, and the strength of evidence is graded:

• Grade A—derived from multiple randomized controlled trials or a meta-analysis
• Grade B—derived from a single randomized controlled trial or nonrandomized study
• Grade C—derived from consensus, expert opinion, or case series.

This system is the same one used by the American College of Cardiology and is based on clinical trial data.²²

Recommendations for calcium intake and vitamin D supplementation were graded A; all other recommendations were graded C (Tables 1 and 2). It is important to note that practices that receive a grade of C may still be accepted as standard of care, such as fall assessment and smoking cessation.

FOR POSTMENOPAUSAL WOMEN AND FOR MEN AGE 50 AND OLDER

FRAX low-risk group

Recall that “low risk” based on the new ACR guidelines means that the 10-year absolute risk of a major osteoporotic fracture, as calculated with FRAX, is less than 10%.

• If glucocorticoid use is expected to last or has already lasted at least 3 months and the dose is less than 7.5 mg/day, no pharmacologic treatment is recommended.
• If glucocorticoid use is expected to last or has already lasted at least 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended.

Comment. These are the most straightforward of the recommendations. All three bisphosphonates are recommended as treatment options if the glucocorticoid dose is at least 7.5 mg/day and the duration at least 3 months. Ibandronate (Boniva) was not included because it has no data from clinical trials.

“Medium risk” means that the 10-year absolute fracture risk of major osteoporotic fractures is 10% to 20%.

• If glucocorticoid use is anticipated to last or has lasted at least 3 months and the dose is less than 7.5 mg/day, alendronate or risedronate is recommended.
• If glucocorticoid use is anticipated to last or has lasted at least 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended.

Comment. Treatment is recommended at all glucocorticoid doses for patients in the medium-risk category if the duration of glucocorticoid treatment is at least 3 months, with one difference: zoledronic acid is recommended only if the glucocorticoid dose is 7.5 mg/day or higher. This inconsistency persisted after a second round of voting by the Task Force Panel.

FRAX high-risk group

In this group, the 10-year risk of major osteoporotic fractures is higher than 20%.

• If the glucocorticoid dose is less than 5 mg/day for up to 1 month, alendronate, risedronate, or zoledronic acid is recommended.
• If the dose is 5 mg/day or more for up to 1 month, or any dose for more than 1 month, alendronate, risedronate, zoledronic acid or teriparatide is recommended.

Comment. Based on current National Osteoporosis Foundation guidelines, all patients with a 10-year risk greater than 20% are recommended for treatment for any duration and dose of glucocorticoid use. However, teriparatide is recommended only if the duration of glucocorticoid therapy is more than 1 month.

FOR PREMENOPAUSAL WOMEN AND FOR MEN YOUNGER THAN AGE 50

Use of FRAX is not appropriate in premenopausal women or in men younger than 50 years.

Younger patients with no prevalent fracture

For men younger than 50 and premenopausal women who have not had a previous fracture, data were considered inadequate to make a recommendation, and no votes were taken.

Prevalent fracture in premenopausal women of nonchildbearing potential

In premenopausal women of nonchildbearing potential who have had a fracture:

• If the glucocorticoid duration is 1 to 3 months and the dose is 5 mg/day or higher, alendronate or risedronate is recommended.
• If the duration is 1 to 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended
• If the duration is more than 3 months, alendronate, risedronate, zoledronic acid, or teriparatide is recommended.

Comment. Treatment is recommended with any of the four medications in patients with a fracture and treated with glucocorticoids for more than 3 months. For shorter-duration glucocorticoid use (1–3 months) at 5 mg/day or higher, only alendronate and risedronate are recommended. If the dose is 7.5 mg/day or higher, any bisphosphonate is recommended. Zoledronic acid was consistently differentiated by the expert panel on the basis of dose and duration of glucocorticoid use, in view of its 1-year duration of effect after one dose.

Prevalent fracture in women of childbearing potential

• If the glucocorticoid duration is 1 to 3 months, there was no consensus (ie, voting disagreements could not be resolved).
• If the glucocorticoid duration is more than 3 months and the dose is 7.5 mg/day or more, alendronate, risedronate, or teriparatide is recommended.
• If the glucocorticoid duration is more than 3 months and the dose is less than 7.5 mg/day, there was no consensus.

Comment. Childbearing potential creates further complexities because of concern about fetal toxicity with bisphosphonates. For short-term glucocorticoid therapy at any dose and for therapy longer than 3 months at less than 7.5 mg, no consensus could be reached. For therapy longer than 3 months and with 7.5 mg/day or higher, treatment is recommended but not with zoledronic acid, based on the long half-life of the drug and concern for fetal toxicity.

Additional risk stratification

The panel recommended that if the following were present, a shift to a higher fracture risk category should be considered (low to medium, or medium to high):

• High daily dose of glucocorticoid
• High cumulative glucocorticoid dose
• Declining bone mineral density on serial DXA.

These are known risk factors that increase fracture risk but would not affect fracture risk in the FRAX model.

WHAT IS NEW IN THE 2010 RECOMMENDATIONS?

Recommendations for counseling now include fall risk assessment, height measurement, 25-hydroxyvitamin D measurement, and evaluation of patients for prevalent and incident fractures using vertebral fracture assessment by DXA or radiographic imaging of the spine.

Recommended drugs now include teriparatide and zoledronic acid, while estrogen and testosterone are no longer recommended as therapies for glucocorticoid-induced osteoporosis. Ibandronate is not included, since there have been no randomized controlled trials of this bisphosphonate in glucocorticoid-induced osteoporosis.

Recommendations for treatment in 2001 were based on T scores alone, while the 2010 recommendations use an assessment of absolute fracture risk based on FRAX for postmenopausal women and for men age 50 and older.

A clinician’s guide that summarizes the ACR recommendations is available at www.rheumatology.org/practice/clinical/guidelines/.

RECOMMENDATIONS DO NOT REPLACE CLINICAL JUDGMENT

Although the 2010 recommendations were more rigorous in their development process than those of 2001, they have limitations and they should not replace clinical judgment. Rather, they are intended to provide an evidence-based approach to guide clinicians in making treatment choices in patients on glucocorticoid therapy.

CONSIDERING ABSOLUTE FRACTURE RISK IN TREATMENT DECISIONS

The 2001 ACR guidelines recommended fracture-preventing treatment in all patients starting glucocorticoid therapy at more than 5 mg/day if the planned duration of treatment was at least 3 months, and in patients on long-term glucocorticoid therapy if the T score was less than −1.0. While these guidelines were simple and easy to use, they were not specific enough to provide useful guidance in specific scenarios.

A model of absolute fracture risk was not available in 2001. A 55-year old white woman with a T score of −1.1 who smoked, who had been using 5 mg of prednisone for the last 12 months, and who had stable bone mass on serial DXA scans would have been recommended for treatment based on the 2001 recommendations. If this patient’s FRAX-calculated 10-year absolute risk of a major osteoporotic fracture is less than 10%, that would be well below the National Osteoporosis Foundation’s cost-effective treatment threshold of 20%. The new guidelines suggest no treatment is needed, since the risk category is low and the dose is less than 7.5 mg. However, if on serial DXA this patient had a significant decline in bone mass, the guidelines suggest shifting the patient to a higher risk category, ie, from low to medium risk, which would result in a recommendation in favor of treatment.

The 2010 recommendations are not as simple to use as those from 2001. They encourage using FRAX to calculate fracture risk; thus, knowledge of the strengths and limitations of FRAX is required. Access to the internet in the examination room or use of the FRAX tool on a smartphone as well as willingness to spend a minute to calculate fracture risk are needed. For those who cannot or choose not to use the FRAX tool, the ACR publication provides tables for patient risk assessment based on age and T score. However, the tables would have to be readily available in the clinic, which may not be practical.

The 2010 recommendation provide a more nuanced approach to treatment in patients on glucocorticoid therapy and are likely to change treatment decisions based on their use, just as FRAX has altered treatment decisions in patients with primary osteoporosis.²³

FRAX has limitations

FRAX underestimates the effect of glucocorticoids on fracture risk because steroid use is a yes-or-no question and its weight represents the average risk in a population that has ever used steroids, most of whom were using doses between 2.5 and 7.5 mg.

The WHO recognized this limitation and suggested an upward adjustment of risk for patients on 7.5 mg or more, ranging from 10% to 25%.²¹ For patients on high doses of steroids, this adjustment is still likely to result in underestimation of fracture risk and undertreatment of glucocorticoid-treated patients.

The 2010 recommendations adjust for this limitation, recommending treatment in the low-risk and medium-risk categories if the glucocorticoid dose is 7.5 mg or higher. If a patient is using high daily doses of steroids or has a declining bone density, the 2010 recommendations suggest increasing the risk category from low to medium or medium to high.

FRAX risk factors are dichotomous (yes/no) and are not adjusted for dose effects such as multiple fractures (vs a single fracture), heavy smoking (vs light smoking), heavy alcohol use (6 units per day vs 3 units), or severe rheumatoid arthritis (vs mild disease). Family history of osteoporosis in the FRAX is limited to parents with a hip fracture—vertebral fractures in a family member do not count.

Since FRAX uses the bone mineral density in the hip, it underestimates fracture risk in patients with low spine density but normal hip density. It may also underestimate fracture risk in patients with declining bone mass; the 2010 recommendations suggest the clinician should increase the risk category in this situation.

LIMITATIONS OF THE GUIDELINES

The 2010 recommendations do not include several important groups in which steroids are used, including transplant recipients, children, and patients on inhaled corticosteroids. The panel thought that there were insufficient data to make recommendations for these populations, as well as for premenopausal women and men younger than 50 years who did not have a prevalent fracture. The absence of a recommendation in these situations should not be considered a recommendation for no treatment; it is an acknowledgment of a lack of evidence, a lack of consensus among experts, and the need for additional clinical trials.

For premenopausal women and men under age 50 with a fracture, the recommendations are complicated and not intuitive. Zoledronic acid is not recommended for women of non-childbearing potential with a glucocorticoid duration of 1 to 3 months unless the steroid dose is at least 7.5 mg. This recommendation was based on panel voting and consensus that giving zoledronic acid, a medication with a 1-year duration of effect, in a patient on steroids for only 1 to 3 months was not warranted.

Teriparatide was recommended only if glucocorticoids are used for at least 3 months, although anyone who already has a fracture might be considered at high enough risk to warrant anabolic therapy regardless of steroid use or duration.

Zoledronic acid was excluded in women of childbearing potential, based on panel voting and consensus that drugs given in smaller amounts over 1 year might be less harmful to a fetus than one with a longer half-life given in a larger bolus once a year.

The panel could reach no consensus on women of childbearing potential with a prevalent fracture who were using less than 7.5 mg/day of glucocorticoids. A lack of consensus was the result of insufficient data to make evidence-based decisions and a disagreement among experts on the correct treatment.

The guidelines do not address the duration of treatment with bisphosphonates, a topic of importance because of concern for the potential long-term side effects of these medications.

THE BOTTOM LINE

The 2010 recommendations add a degree of complexity, with different medications recommended on the basis of glucocorticoid dose and duration as well as patient age, menopausal status, and childbearing potential. Guideline developers and clinicians face a difficult trade-off: easy-to-follow guidelines or more targeted guidelines that are more complex and therefore more difficult to use than previous guidelines.

This criticism is reasonable. The complexity is a result of insufficient evidence from clinical trials to make more exact and user-friendly recommendations, and also a result of the RAND/UCLA methodology. In cases that lack sufficient evidence on which to make a decision, the guideline development uses voting among experts in an attempt to develop consensus. This often results in complexity, lack of consensus, or inconsistencies.

The guidelines are straightforward for postmenopausal women and men age 50 and older on at least 7.5 mg prednisone for more than 3 months.

Since there is substantial evidence that many patients on glucocorticoid therapy go untreated, the risk of fracture in this population would be substantially reduced if clinicians would adhere to the recommendations.

Whenever a patient begins treatment with a glucocorticoid drug, we need to think about bone loss.

The American College of Rheumatology (ACR) issued recommendations for preventing and treating glucocorticoid-induced osteoporosis in 2010.¹ Compared with its previous guidelines,² the new ones are more tailored and nuanced but may be more difficult for physicians to follow. The guidelines call for assessing fracture risk using the computer-based Fracture Risk Assessment Tool, or FRAX (www/shef.ac.uk/FRAX), developed by the World Health Organization (WHO). For those without a computer or ready access to the Web, an application of FRAX is available for download on smartphones.

In this article, my purpose is to review the new recommendations and to offer my perspective, which does not necessarily reflect the opinions of the ACR.

DESPITE EVIDENCE, MANY PATIENTS RECEIVE NO INTERVENTION

Use of glucocorticoids is the most common cause of secondary osteoporosis. During the first 6 to 12 months of use, these drugs can cause a rapid loss of bone mass due to increased bone resorption; with continued use, they cause a slower but steady decline in bone mass due to reduced bone formation.³ Epidemiologic studies have found that the risk of fractures increases with dose, starting with doses as low as 2.5 mg per day of prednisone or its equivalent.⁴

Numerous clinical trials have evaluated the effect of bisphosphonates and teriparatide (Forteo) on bone mass and fracture risk in patients on glucocorticoid therapy. The bisphosphonates alendronate (Fosamax) and risedronate (Actonel) have both been shown to increase bone mass and reduce vertebral fracture risk in glucocorticoid recipients.^5–8 Zoledronic acid (Reclast), a parenteral bisphosphonate given in one annual dose, was shown to increase bone mass more than oral risedronate taken daily,⁹ and teriparatide, a formulation of parathyroid hormone, was better than alendronate.¹⁰

However, despite the known risk of fractures with glucocorticoid use and the demonstrated efficacy of available agents in preventing bone loss and fracture, many patients do not receive any intervention.^11,12

WHAT HAS HAPPENED SINCE 2001?

In the interval since 2001, several guidelines for managing glucocorticoid-induced osteoporosis have been published in other countries.^13–17 Broadly speaking, they recommend starting preventive drug therapy for patients at risk of fracture at the same time glucocorticoid drugs are started if the patient is expected to take glucocorticoids for more than 3 to 6 months in doses higher than 5 to 7.5 mg of prednisone or its equivalent daily.

Recommendations for patients who have been on glucocorticoids for longer than 3 to 6 months at initial evaluation have been based largely on T scores derived from dual-energy x-ray absorptiometry (DXA). Thresholds for initiating therapy have varied: the ACR in 2001 recommended preventive treatment if the T score is lower than −1.0, whereas British guidelines said −1.5 and Dutch guidelines said −2.5.

In the United States, since 2001 when the ACR published its last guidelines,² zoledronic acid and teriparatide have been approved for use in glucocorticoid-induced osteoporosis. In addition, guideline-development methodology has evolved and now is more scientifically rigorous. Finally, a risk-assessment tool has been developed that enables a more tailored approach (see below).

FRAX (www.shef.ac.uk/FRAX)

FRAX is a tool developed by the WHO to calculate the risk of fracture. If you go to the FRAX Web site and enter the required clinical information (race, age, sex, weight, height, previous fracture, family history of a fractured hip in a parent, current smoking, use of glucocorticoids, rheumatoid arthritis, secondary osteoporosis, consumption of three or more units of alcohol per day, and bone mineral density of the femoral neck), it will tell you the patient’s 10-year absolute (not relative) risk of major osteoporotic fracture and of hip fracture.

Since FRAX was unveiled in 2008, calculation of absolute fracture risk has become the standard method for making treatment decisions in patients with low bone mass who have not yet received any fracture-preventing treatment.¹⁸ The use of clinical risk factors in FRAX increases its ability to predict risk over and above the use of bone density by itself. And glucocorticoids are one of the clinical risk factors in FRAX.

But in which patients is treatment with a bisphosphonate or teriparatide cost-effective?

Thresholds for cost-effectiveness have been developed on the basis of economic assumptions that are country-specific. In the United States, the National Osteoporosis Foundation recommends drug therapy if the 10-year absolute risk of a major osteoporotic fracture of the hip, spine (clinical, not radiographic), wrist, or humerus is greater than 20% or if the risk of a hip fracture is greater than 3%.¹⁹

At equivalent bone densities, women taking glucocorticoids are at considerably higher risk of fracture than nonusers.²⁰ For example, consider a 65-year-old white woman, weight 59 kg, height 163 cm, no previous fractures, no parent with a fractured hip, no current smoking, no rheumatoid arthritis, no secondary osteoporosis, no excessive alcohol use, and a T score of −2.2 in the femoral neck. (Try this on the FRAX Web site.) If she does not use glucocorticoids, her 10-year risk of hip fracture is 2.0%; using glucocorticoids increases the risk to 3.6%. This is higher than the 3% National Osteoporosis Foundation guideline; thus, treatment would be recommended.

Also using FRAX, a 55-year-old white woman with a T score of −1.8 and on glucocorticoid therapy has a 67% higher risk of major osteoporotic fracture and an 80% higher risk of hip fracture.

For a third example, a white woman age 60, weight 70 kg, height 168 cm, negative for all the other risk factors but with a T score of −2.1 and on glucocorticoids has a calculated 10-year fracture risk of 2.1%, which is below the National Osteoporosis Foundation treatment threshold. However, most clinicians would probably recommend treatment for her, depending on the anticipated dose and duration of glucocorticoid therapy.

A caveat. In FRAX, glucocorticoid therapy is a categorical variable—a yes-or-no question—and yes is defined as having ever used a glucocorticoid in a dose greater than 5 mg for more than 3 months. Therefore, according to FRAX, a patient who took 5 mg of prednisone for 3 months 5 years ago has the same fracture risk as a patient on 60 mg of prednisone after a diagnosis of temporal arteritis. For this reason, the FRAX tool is likely to underestimate fracture risk, especially in patients currently taking glucocorticoids and those on higher doses of these drugs.

Kanis et al used the General Practice Research Database to adjust the fracture risk for glucocorticoid use in FRAX.²¹ At doses higher than 7.5 mg, the fracture risk had to be revised upward by 10% to 25% depending on the fracture site (hip vs any major osteoporotic fracture) and age (greater at age 40 than at age 90).

The underestimation of fracture risk led the ACR Expert Advisory Panel to create risk strata for major osteoporotic fractures, ie, low (< 10% risk per 10 years), medium (10%–20%), and high (> 20%) and uses these cut points to make treatment recommendations.

HOW THE 2010 GUIDELINES WERE DEVELOPED

Whereas the 2001 recommendations were based on a more informal consensus approach, the 2010 recommendations use a more scientifically rigorous methodology for guideline development, the Research and Development/University of California at Los Angeles (RAND/UCLA) Appropriateness Method. The RAND/UCLA method combines the best available scientific evidence with expert opinion to develop practice guidelines.

In drawing up the 2010 recommendations the ACR used three panels of experts. The Core Executive Panel conducted a systematic review of controlled clinical trials of therapies currently approved for treating glucocorticoid-induced osteoporosis in the United States, Canada, or the European Union. They found 53 articles meeting their inclusion criteria; an evidence report was produced that informed the development of the recommendations. This evidence report and guideline development process is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658. The Expert Advisory Panel framed the recommendations, and the Task Force Panel voted on them. The Core Executive Panel and Expert Advisory Panel constructed 48 patient-specific clinical scenarios using four variables: sex, age, race/ethnicity, and femoral neck T scores.

The members of the Task Force Panel were asked to use the evidence report and their expert judgment to vote on and rate the appropriateness of using a specific therapy in the context of each scenario on a 9-point Likert scale (1 = appropriate; 9 = not appropriate). Agreement occurred when 7 or more of the 10 panel members rated a scenario 1, 2, or 3. Disagreements were defined as 3 or more of the 10 members rating the scenario between 4 and 9 while the other members rated it lower.

Disagreements in voting were discussed in an attempt to achieve consensus, and a second vote was conducted which determined the final recommendations. If disagreement remained after the vote, no recommendation was made.

No attempt was made to assign priority of one drug over another when multiple drugs were deemed appropriate, although the final recommendations did differentiate drugs based on patient categories.

START WITH COUNSELING, ASSESSMENT

For patients starting or already on glucocorticoid therapy that is expected to last at least 3 months, the first step is to counsel them on lifestyle modifications (Table 1) and to assess their risk factors (Figure 1). Recommendations for monitoring patients receiving glucocorticoid therapy for at least 3 months are presented in Table 2.

These recommendations are based on literature review, and the strength of evidence is graded:

• Grade A—derived from multiple randomized controlled trials or a meta-analysis
• Grade B—derived from a single randomized controlled trial or nonrandomized study
• Grade C—derived from consensus, expert opinion, or case series.

This system is the same one used by the American College of Cardiology and is based on clinical trial data.²²

Recommendations for calcium intake and vitamin D supplementation were graded A; all other recommendations were graded C (Tables 1 and 2). It is important to note that practices that receive a grade of C may still be accepted as standard of care, such as fall assessment and smoking cessation.

FOR POSTMENOPAUSAL WOMEN AND FOR MEN AGE 50 AND OLDER

FRAX low-risk group

Recall that “low risk” based on the new ACR guidelines means that the 10-year absolute risk of a major osteoporotic fracture, as calculated with FRAX, is less than 10%.

• If glucocorticoid use is expected to last or has already lasted at least 3 months and the dose is less than 7.5 mg/day, no pharmacologic treatment is recommended.
• If glucocorticoid use is expected to last or has already lasted at least 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended.

Comment. These are the most straightforward of the recommendations. All three bisphosphonates are recommended as treatment options if the glucocorticoid dose is at least 7.5 mg/day and the duration at least 3 months. Ibandronate (Boniva) was not included because it has no data from clinical trials.

“Medium risk” means that the 10-year absolute fracture risk of major osteoporotic fractures is 10% to 20%.

• If glucocorticoid use is anticipated to last or has lasted at least 3 months and the dose is less than 7.5 mg/day, alendronate or risedronate is recommended.
• If glucocorticoid use is anticipated to last or has lasted at least 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended.

Comment. Treatment is recommended at all glucocorticoid doses for patients in the medium-risk category if the duration of glucocorticoid treatment is at least 3 months, with one difference: zoledronic acid is recommended only if the glucocorticoid dose is 7.5 mg/day or higher. This inconsistency persisted after a second round of voting by the Task Force Panel.

FRAX high-risk group

In this group, the 10-year risk of major osteoporotic fractures is higher than 20%.

• If the glucocorticoid dose is less than 5 mg/day for up to 1 month, alendronate, risedronate, or zoledronic acid is recommended.
• If the dose is 5 mg/day or more for up to 1 month, or any dose for more than 1 month, alendronate, risedronate, zoledronic acid or teriparatide is recommended.

Comment. Based on current National Osteoporosis Foundation guidelines, all patients with a 10-year risk greater than 20% are recommended for treatment for any duration and dose of glucocorticoid use. However, teriparatide is recommended only if the duration of glucocorticoid therapy is more than 1 month.

FOR PREMENOPAUSAL WOMEN AND FOR MEN YOUNGER THAN AGE 50

Use of FRAX is not appropriate in premenopausal women or in men younger than 50 years.

Younger patients with no prevalent fracture

For men younger than 50 and premenopausal women who have not had a previous fracture, data were considered inadequate to make a recommendation, and no votes were taken.

Prevalent fracture in premenopausal women of nonchildbearing potential

In premenopausal women of nonchildbearing potential who have had a fracture:

• If the glucocorticoid duration is 1 to 3 months and the dose is 5 mg/day or higher, alendronate or risedronate is recommended.
• If the duration is 1 to 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended
• If the duration is more than 3 months, alendronate, risedronate, zoledronic acid, or teriparatide is recommended.

Comment. Treatment is recommended with any of the four medications in patients with a fracture and treated with glucocorticoids for more than 3 months. For shorter-duration glucocorticoid use (1–3 months) at 5 mg/day or higher, only alendronate and risedronate are recommended. If the dose is 7.5 mg/day or higher, any bisphosphonate is recommended. Zoledronic acid was consistently differentiated by the expert panel on the basis of dose and duration of glucocorticoid use, in view of its 1-year duration of effect after one dose.

Prevalent fracture in women of childbearing potential

• If the glucocorticoid duration is 1 to 3 months, there was no consensus (ie, voting disagreements could not be resolved).
• If the glucocorticoid duration is more than 3 months and the dose is 7.5 mg/day or more, alendronate, risedronate, or teriparatide is recommended.
• If the glucocorticoid duration is more than 3 months and the dose is less than 7.5 mg/day, there was no consensus.

Comment. Childbearing potential creates further complexities because of concern about fetal toxicity with bisphosphonates. For short-term glucocorticoid therapy at any dose and for therapy longer than 3 months at less than 7.5 mg, no consensus could be reached. For therapy longer than 3 months and with 7.5 mg/day or higher, treatment is recommended but not with zoledronic acid, based on the long half-life of the drug and concern for fetal toxicity.

Additional risk stratification

The panel recommended that if the following were present, a shift to a higher fracture risk category should be considered (low to medium, or medium to high):

• High daily dose of glucocorticoid
• High cumulative glucocorticoid dose
• Declining bone mineral density on serial DXA.

These are known risk factors that increase fracture risk but would not affect fracture risk in the FRAX model.

WHAT IS NEW IN THE 2010 RECOMMENDATIONS?

Recommendations for counseling now include fall risk assessment, height measurement, 25-hydroxyvitamin D measurement, and evaluation of patients for prevalent and incident fractures using vertebral fracture assessment by DXA or radiographic imaging of the spine.

Recommended drugs now include teriparatide and zoledronic acid, while estrogen and testosterone are no longer recommended as therapies for glucocorticoid-induced osteoporosis. Ibandronate is not included, since there have been no randomized controlled trials of this bisphosphonate in glucocorticoid-induced osteoporosis.

Recommendations for treatment in 2001 were based on T scores alone, while the 2010 recommendations use an assessment of absolute fracture risk based on FRAX for postmenopausal women and for men age 50 and older.

A clinician’s guide that summarizes the ACR recommendations is available at www.rheumatology.org/practice/clinical/guidelines/.

RECOMMENDATIONS DO NOT REPLACE CLINICAL JUDGMENT

Although the 2010 recommendations were more rigorous in their development process than those of 2001, they have limitations and they should not replace clinical judgment. Rather, they are intended to provide an evidence-based approach to guide clinicians in making treatment choices in patients on glucocorticoid therapy.

CONSIDERING ABSOLUTE FRACTURE RISK IN TREATMENT DECISIONS

The 2001 ACR guidelines recommended fracture-preventing treatment in all patients starting glucocorticoid therapy at more than 5 mg/day if the planned duration of treatment was at least 3 months, and in patients on long-term glucocorticoid therapy if the T score was less than −1.0. While these guidelines were simple and easy to use, they were not specific enough to provide useful guidance in specific scenarios.

A model of absolute fracture risk was not available in 2001. A 55-year old white woman with a T score of −1.1 who smoked, who had been using 5 mg of prednisone for the last 12 months, and who had stable bone mass on serial DXA scans would have been recommended for treatment based on the 2001 recommendations. If this patient’s FRAX-calculated 10-year absolute risk of a major osteoporotic fracture is less than 10%, that would be well below the National Osteoporosis Foundation’s cost-effective treatment threshold of 20%. The new guidelines suggest no treatment is needed, since the risk category is low and the dose is less than 7.5 mg. However, if on serial DXA this patient had a significant decline in bone mass, the guidelines suggest shifting the patient to a higher risk category, ie, from low to medium risk, which would result in a recommendation in favor of treatment.

The 2010 recommendations are not as simple to use as those from 2001. They encourage using FRAX to calculate fracture risk; thus, knowledge of the strengths and limitations of FRAX is required. Access to the internet in the examination room or use of the FRAX tool on a smartphone as well as willingness to spend a minute to calculate fracture risk are needed. For those who cannot or choose not to use the FRAX tool, the ACR publication provides tables for patient risk assessment based on age and T score. However, the tables would have to be readily available in the clinic, which may not be practical.

The 2010 recommendation provide a more nuanced approach to treatment in patients on glucocorticoid therapy and are likely to change treatment decisions based on their use, just as FRAX has altered treatment decisions in patients with primary osteoporosis.²³

FRAX has limitations

FRAX underestimates the effect of glucocorticoids on fracture risk because steroid use is a yes-or-no question and its weight represents the average risk in a population that has ever used steroids, most of whom were using doses between 2.5 and 7.5 mg.

The WHO recognized this limitation and suggested an upward adjustment of risk for patients on 7.5 mg or more, ranging from 10% to 25%.²¹ For patients on high doses of steroids, this adjustment is still likely to result in underestimation of fracture risk and undertreatment of glucocorticoid-treated patients.

The 2010 recommendations adjust for this limitation, recommending treatment in the low-risk and medium-risk categories if the glucocorticoid dose is 7.5 mg or higher. If a patient is using high daily doses of steroids or has a declining bone density, the 2010 recommendations suggest increasing the risk category from low to medium or medium to high.

FRAX risk factors are dichotomous (yes/no) and are not adjusted for dose effects such as multiple fractures (vs a single fracture), heavy smoking (vs light smoking), heavy alcohol use (6 units per day vs 3 units), or severe rheumatoid arthritis (vs mild disease). Family history of osteoporosis in the FRAX is limited to parents with a hip fracture—vertebral fractures in a family member do not count.

Since FRAX uses the bone mineral density in the hip, it underestimates fracture risk in patients with low spine density but normal hip density. It may also underestimate fracture risk in patients with declining bone mass; the 2010 recommendations suggest the clinician should increase the risk category in this situation.

LIMITATIONS OF THE GUIDELINES

The 2010 recommendations do not include several important groups in which steroids are used, including transplant recipients, children, and patients on inhaled corticosteroids. The panel thought that there were insufficient data to make recommendations for these populations, as well as for premenopausal women and men younger than 50 years who did not have a prevalent fracture. The absence of a recommendation in these situations should not be considered a recommendation for no treatment; it is an acknowledgment of a lack of evidence, a lack of consensus among experts, and the need for additional clinical trials.

For premenopausal women and men under age 50 with a fracture, the recommendations are complicated and not intuitive. Zoledronic acid is not recommended for women of non-childbearing potential with a glucocorticoid duration of 1 to 3 months unless the steroid dose is at least 7.5 mg. This recommendation was based on panel voting and consensus that giving zoledronic acid, a medication with a 1-year duration of effect, in a patient on steroids for only 1 to 3 months was not warranted.

Teriparatide was recommended only if glucocorticoids are used for at least 3 months, although anyone who already has a fracture might be considered at high enough risk to warrant anabolic therapy regardless of steroid use or duration.

Zoledronic acid was excluded in women of childbearing potential, based on panel voting and consensus that drugs given in smaller amounts over 1 year might be less harmful to a fetus than one with a longer half-life given in a larger bolus once a year.

The panel could reach no consensus on women of childbearing potential with a prevalent fracture who were using less than 7.5 mg/day of glucocorticoids. A lack of consensus was the result of insufficient data to make evidence-based decisions and a disagreement among experts on the correct treatment.

The guidelines do not address the duration of treatment with bisphosphonates, a topic of importance because of concern for the potential long-term side effects of these medications.

THE BOTTOM LINE

The 2010 recommendations add a degree of complexity, with different medications recommended on the basis of glucocorticoid dose and duration as well as patient age, menopausal status, and childbearing potential. Guideline developers and clinicians face a difficult trade-off: easy-to-follow guidelines or more targeted guidelines that are more complex and therefore more difficult to use than previous guidelines.

This criticism is reasonable. The complexity is a result of insufficient evidence from clinical trials to make more exact and user-friendly recommendations, and also a result of the RAND/UCLA methodology. In cases that lack sufficient evidence on which to make a decision, the guideline development uses voting among experts in an attempt to develop consensus. This often results in complexity, lack of consensus, or inconsistencies.

The guidelines are straightforward for postmenopausal women and men age 50 and older on at least 7.5 mg prednisone for more than 3 months.

Since there is substantial evidence that many patients on glucocorticoid therapy go untreated, the risk of fracture in this population would be substantially reduced if clinicians would adhere to the recommendations.

Whenever a patient begins treatment with a glucocorticoid drug, we need to think about bone loss.

The American College of Rheumatology (ACR) issued recommendations for preventing and treating glucocorticoid-induced osteoporosis in 2010.¹ Compared with its previous guidelines,² the new ones are more tailored and nuanced but may be more difficult for physicians to follow. The guidelines call for assessing fracture risk using the computer-based Fracture Risk Assessment Tool, or FRAX (www/shef.ac.uk/FRAX), developed by the World Health Organization (WHO). For those without a computer or ready access to the Web, an application of FRAX is available for download on smartphones.

In this article, my purpose is to review the new recommendations and to offer my perspective, which does not necessarily reflect the opinions of the ACR.

DESPITE EVIDENCE, MANY PATIENTS RECEIVE NO INTERVENTION

Use of glucocorticoids is the most common cause of secondary osteoporosis. During the first 6 to 12 months of use, these drugs can cause a rapid loss of bone mass due to increased bone resorption; with continued use, they cause a slower but steady decline in bone mass due to reduced bone formation.³ Epidemiologic studies have found that the risk of fractures increases with dose, starting with doses as low as 2.5 mg per day of prednisone or its equivalent.⁴

Numerous clinical trials have evaluated the effect of bisphosphonates and teriparatide (Forteo) on bone mass and fracture risk in patients on glucocorticoid therapy. The bisphosphonates alendronate (Fosamax) and risedronate (Actonel) have both been shown to increase bone mass and reduce vertebral fracture risk in glucocorticoid recipients.^5–8 Zoledronic acid (Reclast), a parenteral bisphosphonate given in one annual dose, was shown to increase bone mass more than oral risedronate taken daily,⁹ and teriparatide, a formulation of parathyroid hormone, was better than alendronate.¹⁰

However, despite the known risk of fractures with glucocorticoid use and the demonstrated efficacy of available agents in preventing bone loss and fracture, many patients do not receive any intervention.^11,12

WHAT HAS HAPPENED SINCE 2001?

In the interval since 2001, several guidelines for managing glucocorticoid-induced osteoporosis have been published in other countries.^13–17 Broadly speaking, they recommend starting preventive drug therapy for patients at risk of fracture at the same time glucocorticoid drugs are started if the patient is expected to take glucocorticoids for more than 3 to 6 months in doses higher than 5 to 7.5 mg of prednisone or its equivalent daily.

Recommendations for patients who have been on glucocorticoids for longer than 3 to 6 months at initial evaluation have been based largely on T scores derived from dual-energy x-ray absorptiometry (DXA). Thresholds for initiating therapy have varied: the ACR in 2001 recommended preventive treatment if the T score is lower than −1.0, whereas British guidelines said −1.5 and Dutch guidelines said −2.5.

In the United States, since 2001 when the ACR published its last guidelines,² zoledronic acid and teriparatide have been approved for use in glucocorticoid-induced osteoporosis. In addition, guideline-development methodology has evolved and now is more scientifically rigorous. Finally, a risk-assessment tool has been developed that enables a more tailored approach (see below).

FRAX (www.shef.ac.uk/FRAX)

FRAX is a tool developed by the WHO to calculate the risk of fracture. If you go to the FRAX Web site and enter the required clinical information (race, age, sex, weight, height, previous fracture, family history of a fractured hip in a parent, current smoking, use of glucocorticoids, rheumatoid arthritis, secondary osteoporosis, consumption of three or more units of alcohol per day, and bone mineral density of the femoral neck), it will tell you the patient’s 10-year absolute (not relative) risk of major osteoporotic fracture and of hip fracture.

Since FRAX was unveiled in 2008, calculation of absolute fracture risk has become the standard method for making treatment decisions in patients with low bone mass who have not yet received any fracture-preventing treatment.¹⁸ The use of clinical risk factors in FRAX increases its ability to predict risk over and above the use of bone density by itself. And glucocorticoids are one of the clinical risk factors in FRAX.

But in which patients is treatment with a bisphosphonate or teriparatide cost-effective?

Thresholds for cost-effectiveness have been developed on the basis of economic assumptions that are country-specific. In the United States, the National Osteoporosis Foundation recommends drug therapy if the 10-year absolute risk of a major osteoporotic fracture of the hip, spine (clinical, not radiographic), wrist, or humerus is greater than 20% or if the risk of a hip fracture is greater than 3%.¹⁹

At equivalent bone densities, women taking glucocorticoids are at considerably higher risk of fracture than nonusers.²⁰ For example, consider a 65-year-old white woman, weight 59 kg, height 163 cm, no previous fractures, no parent with a fractured hip, no current smoking, no rheumatoid arthritis, no secondary osteoporosis, no excessive alcohol use, and a T score of −2.2 in the femoral neck. (Try this on the FRAX Web site.) If she does not use glucocorticoids, her 10-year risk of hip fracture is 2.0%; using glucocorticoids increases the risk to 3.6%. This is higher than the 3% National Osteoporosis Foundation guideline; thus, treatment would be recommended.

Also using FRAX, a 55-year-old white woman with a T score of −1.8 and on glucocorticoid therapy has a 67% higher risk of major osteoporotic fracture and an 80% higher risk of hip fracture.

For a third example, a white woman age 60, weight 70 kg, height 168 cm, negative for all the other risk factors but with a T score of −2.1 and on glucocorticoids has a calculated 10-year fracture risk of 2.1%, which is below the National Osteoporosis Foundation treatment threshold. However, most clinicians would probably recommend treatment for her, depending on the anticipated dose and duration of glucocorticoid therapy.

A caveat. In FRAX, glucocorticoid therapy is a categorical variable—a yes-or-no question—and yes is defined as having ever used a glucocorticoid in a dose greater than 5 mg for more than 3 months. Therefore, according to FRAX, a patient who took 5 mg of prednisone for 3 months 5 years ago has the same fracture risk as a patient on 60 mg of prednisone after a diagnosis of temporal arteritis. For this reason, the FRAX tool is likely to underestimate fracture risk, especially in patients currently taking glucocorticoids and those on higher doses of these drugs.

Kanis et al used the General Practice Research Database to adjust the fracture risk for glucocorticoid use in FRAX.²¹ At doses higher than 7.5 mg, the fracture risk had to be revised upward by 10% to 25% depending on the fracture site (hip vs any major osteoporotic fracture) and age (greater at age 40 than at age 90).

The underestimation of fracture risk led the ACR Expert Advisory Panel to create risk strata for major osteoporotic fractures, ie, low (< 10% risk per 10 years), medium (10%–20%), and high (> 20%) and uses these cut points to make treatment recommendations.

HOW THE 2010 GUIDELINES WERE DEVELOPED

Whereas the 2001 recommendations were based on a more informal consensus approach, the 2010 recommendations use a more scientifically rigorous methodology for guideline development, the Research and Development/University of California at Los Angeles (RAND/UCLA) Appropriateness Method. The RAND/UCLA method combines the best available scientific evidence with expert opinion to develop practice guidelines.

In drawing up the 2010 recommendations the ACR used three panels of experts. The Core Executive Panel conducted a systematic review of controlled clinical trials of therapies currently approved for treating glucocorticoid-induced osteoporosis in the United States, Canada, or the European Union. They found 53 articles meeting their inclusion criteria; an evidence report was produced that informed the development of the recommendations. This evidence report and guideline development process is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658. The Expert Advisory Panel framed the recommendations, and the Task Force Panel voted on them. The Core Executive Panel and Expert Advisory Panel constructed 48 patient-specific clinical scenarios using four variables: sex, age, race/ethnicity, and femoral neck T scores.

The members of the Task Force Panel were asked to use the evidence report and their expert judgment to vote on and rate the appropriateness of using a specific therapy in the context of each scenario on a 9-point Likert scale (1 = appropriate; 9 = not appropriate). Agreement occurred when 7 or more of the 10 panel members rated a scenario 1, 2, or 3. Disagreements were defined as 3 or more of the 10 members rating the scenario between 4 and 9 while the other members rated it lower.

Disagreements in voting were discussed in an attempt to achieve consensus, and a second vote was conducted which determined the final recommendations. If disagreement remained after the vote, no recommendation was made.

No attempt was made to assign priority of one drug over another when multiple drugs were deemed appropriate, although the final recommendations did differentiate drugs based on patient categories.

START WITH COUNSELING, ASSESSMENT

For patients starting or already on glucocorticoid therapy that is expected to last at least 3 months, the first step is to counsel them on lifestyle modifications (Table 1) and to assess their risk factors (Figure 1). Recommendations for monitoring patients receiving glucocorticoid therapy for at least 3 months are presented in Table 2.

These recommendations are based on literature review, and the strength of evidence is graded:

• Grade A—derived from multiple randomized controlled trials or a meta-analysis
• Grade B—derived from a single randomized controlled trial or nonrandomized study
• Grade C—derived from consensus, expert opinion, or case series.

This system is the same one used by the American College of Cardiology and is based on clinical trial data.²²

Recommendations for calcium intake and vitamin D supplementation were graded A; all other recommendations were graded C (Tables 1 and 2). It is important to note that practices that receive a grade of C may still be accepted as standard of care, such as fall assessment and smoking cessation.

FOR POSTMENOPAUSAL WOMEN AND FOR MEN AGE 50 AND OLDER

FRAX low-risk group

Recall that “low risk” based on the new ACR guidelines means that the 10-year absolute risk of a major osteoporotic fracture, as calculated with FRAX, is less than 10%.

• If glucocorticoid use is expected to last or has already lasted at least 3 months and the dose is less than 7.5 mg/day, no pharmacologic treatment is recommended.
• If glucocorticoid use is expected to last or has already lasted at least 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended.

Comment. These are the most straightforward of the recommendations. All three bisphosphonates are recommended as treatment options if the glucocorticoid dose is at least 7.5 mg/day and the duration at least 3 months. Ibandronate (Boniva) was not included because it has no data from clinical trials.

“Medium risk” means that the 10-year absolute fracture risk of major osteoporotic fractures is 10% to 20%.

• If glucocorticoid use is anticipated to last or has lasted at least 3 months and the dose is less than 7.5 mg/day, alendronate or risedronate is recommended.
• If glucocorticoid use is anticipated to last or has lasted at least 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended.

Comment. Treatment is recommended at all glucocorticoid doses for patients in the medium-risk category if the duration of glucocorticoid treatment is at least 3 months, with one difference: zoledronic acid is recommended only if the glucocorticoid dose is 7.5 mg/day or higher. This inconsistency persisted after a second round of voting by the Task Force Panel.

FRAX high-risk group

In this group, the 10-year risk of major osteoporotic fractures is higher than 20%.

• If the glucocorticoid dose is less than 5 mg/day for up to 1 month, alendronate, risedronate, or zoledronic acid is recommended.
• If the dose is 5 mg/day or more for up to 1 month, or any dose for more than 1 month, alendronate, risedronate, zoledronic acid or teriparatide is recommended.

Comment. Based on current National Osteoporosis Foundation guidelines, all patients with a 10-year risk greater than 20% are recommended for treatment for any duration and dose of glucocorticoid use. However, teriparatide is recommended only if the duration of glucocorticoid therapy is more than 1 month.

FOR PREMENOPAUSAL WOMEN AND FOR MEN YOUNGER THAN AGE 50

Use of FRAX is not appropriate in premenopausal women or in men younger than 50 years.

Younger patients with no prevalent fracture

For men younger than 50 and premenopausal women who have not had a previous fracture, data were considered inadequate to make a recommendation, and no votes were taken.

Prevalent fracture in premenopausal women of nonchildbearing potential

In premenopausal women of nonchildbearing potential who have had a fracture:

• If the glucocorticoid duration is 1 to 3 months and the dose is 5 mg/day or higher, alendronate or risedronate is recommended.
• If the duration is 1 to 3 months and the dose is 7.5 mg/day or higher, alendronate, risedronate, or zoledronic acid is recommended
• If the duration is more than 3 months, alendronate, risedronate, zoledronic acid, or teriparatide is recommended.

Comment. Treatment is recommended with any of the four medications in patients with a fracture and treated with glucocorticoids for more than 3 months. For shorter-duration glucocorticoid use (1–3 months) at 5 mg/day or higher, only alendronate and risedronate are recommended. If the dose is 7.5 mg/day or higher, any bisphosphonate is recommended. Zoledronic acid was consistently differentiated by the expert panel on the basis of dose and duration of glucocorticoid use, in view of its 1-year duration of effect after one dose.

Prevalent fracture in women of childbearing potential

• If the glucocorticoid duration is 1 to 3 months, there was no consensus (ie, voting disagreements could not be resolved).
• If the glucocorticoid duration is more than 3 months and the dose is 7.5 mg/day or more, alendronate, risedronate, or teriparatide is recommended.
• If the glucocorticoid duration is more than 3 months and the dose is less than 7.5 mg/day, there was no consensus.

Comment. Childbearing potential creates further complexities because of concern about fetal toxicity with bisphosphonates. For short-term glucocorticoid therapy at any dose and for therapy longer than 3 months at less than 7.5 mg, no consensus could be reached. For therapy longer than 3 months and with 7.5 mg/day or higher, treatment is recommended but not with zoledronic acid, based on the long half-life of the drug and concern for fetal toxicity.

Additional risk stratification

The panel recommended that if the following were present, a shift to a higher fracture risk category should be considered (low to medium, or medium to high):

• High daily dose of glucocorticoid
• High cumulative glucocorticoid dose
• Declining bone mineral density on serial DXA.

These are known risk factors that increase fracture risk but would not affect fracture risk in the FRAX model.

WHAT IS NEW IN THE 2010 RECOMMENDATIONS?

Recommendations for counseling now include fall risk assessment, height measurement, 25-hydroxyvitamin D measurement, and evaluation of patients for prevalent and incident fractures using vertebral fracture assessment by DXA or radiographic imaging of the spine.

Recommended drugs now include teriparatide and zoledronic acid, while estrogen and testosterone are no longer recommended as therapies for glucocorticoid-induced osteoporosis. Ibandronate is not included, since there have been no randomized controlled trials of this bisphosphonate in glucocorticoid-induced osteoporosis.

Recommendations for treatment in 2001 were based on T scores alone, while the 2010 recommendations use an assessment of absolute fracture risk based on FRAX for postmenopausal women and for men age 50 and older.

A clinician’s guide that summarizes the ACR recommendations is available at www.rheumatology.org/practice/clinical/guidelines/.

RECOMMENDATIONS DO NOT REPLACE CLINICAL JUDGMENT

Although the 2010 recommendations were more rigorous in their development process than those of 2001, they have limitations and they should not replace clinical judgment. Rather, they are intended to provide an evidence-based approach to guide clinicians in making treatment choices in patients on glucocorticoid therapy.

CONSIDERING ABSOLUTE FRACTURE RISK IN TREATMENT DECISIONS

The 2001 ACR guidelines recommended fracture-preventing treatment in all patients starting glucocorticoid therapy at more than 5 mg/day if the planned duration of treatment was at least 3 months, and in patients on long-term glucocorticoid therapy if the T score was less than −1.0. While these guidelines were simple and easy to use, they were not specific enough to provide useful guidance in specific scenarios.

A model of absolute fracture risk was not available in 2001. A 55-year old white woman with a T score of −1.1 who smoked, who had been using 5 mg of prednisone for the last 12 months, and who had stable bone mass on serial DXA scans would have been recommended for treatment based on the 2001 recommendations. If this patient’s FRAX-calculated 10-year absolute risk of a major osteoporotic fracture is less than 10%, that would be well below the National Osteoporosis Foundation’s cost-effective treatment threshold of 20%. The new guidelines suggest no treatment is needed, since the risk category is low and the dose is less than 7.5 mg. However, if on serial DXA this patient had a significant decline in bone mass, the guidelines suggest shifting the patient to a higher risk category, ie, from low to medium risk, which would result in a recommendation in favor of treatment.

The 2010 recommendations are not as simple to use as those from 2001. They encourage using FRAX to calculate fracture risk; thus, knowledge of the strengths and limitations of FRAX is required. Access to the internet in the examination room or use of the FRAX tool on a smartphone as well as willingness to spend a minute to calculate fracture risk are needed. For those who cannot or choose not to use the FRAX tool, the ACR publication provides tables for patient risk assessment based on age and T score. However, the tables would have to be readily available in the clinic, which may not be practical.

The 2010 recommendation provide a more nuanced approach to treatment in patients on glucocorticoid therapy and are likely to change treatment decisions based on their use, just as FRAX has altered treatment decisions in patients with primary osteoporosis.²³

FRAX has limitations

FRAX underestimates the effect of glucocorticoids on fracture risk because steroid use is a yes-or-no question and its weight represents the average risk in a population that has ever used steroids, most of whom were using doses between 2.5 and 7.5 mg.

The WHO recognized this limitation and suggested an upward adjustment of risk for patients on 7.5 mg or more, ranging from 10% to 25%.²¹ For patients on high doses of steroids, this adjustment is still likely to result in underestimation of fracture risk and undertreatment of glucocorticoid-treated patients.

The 2010 recommendations adjust for this limitation, recommending treatment in the low-risk and medium-risk categories if the glucocorticoid dose is 7.5 mg or higher. If a patient is using high daily doses of steroids or has a declining bone density, the 2010 recommendations suggest increasing the risk category from low to medium or medium to high.

FRAX risk factors are dichotomous (yes/no) and are not adjusted for dose effects such as multiple fractures (vs a single fracture), heavy smoking (vs light smoking), heavy alcohol use (6 units per day vs 3 units), or severe rheumatoid arthritis (vs mild disease). Family history of osteoporosis in the FRAX is limited to parents with a hip fracture—vertebral fractures in a family member do not count.

Since FRAX uses the bone mineral density in the hip, it underestimates fracture risk in patients with low spine density but normal hip density. It may also underestimate fracture risk in patients with declining bone mass; the 2010 recommendations suggest the clinician should increase the risk category in this situation.

LIMITATIONS OF THE GUIDELINES

The 2010 recommendations do not include several important groups in which steroids are used, including transplant recipients, children, and patients on inhaled corticosteroids. The panel thought that there were insufficient data to make recommendations for these populations, as well as for premenopausal women and men younger than 50 years who did not have a prevalent fracture. The absence of a recommendation in these situations should not be considered a recommendation for no treatment; it is an acknowledgment of a lack of evidence, a lack of consensus among experts, and the need for additional clinical trials.

For premenopausal women and men under age 50 with a fracture, the recommendations are complicated and not intuitive. Zoledronic acid is not recommended for women of non-childbearing potential with a glucocorticoid duration of 1 to 3 months unless the steroid dose is at least 7.5 mg. This recommendation was based on panel voting and consensus that giving zoledronic acid, a medication with a 1-year duration of effect, in a patient on steroids for only 1 to 3 months was not warranted.

Teriparatide was recommended only if glucocorticoids are used for at least 3 months, although anyone who already has a fracture might be considered at high enough risk to warrant anabolic therapy regardless of steroid use or duration.

Zoledronic acid was excluded in women of childbearing potential, based on panel voting and consensus that drugs given in smaller amounts over 1 year might be less harmful to a fetus than one with a longer half-life given in a larger bolus once a year.

The panel could reach no consensus on women of childbearing potential with a prevalent fracture who were using less than 7.5 mg/day of glucocorticoids. A lack of consensus was the result of insufficient data to make evidence-based decisions and a disagreement among experts on the correct treatment.

The guidelines do not address the duration of treatment with bisphosphonates, a topic of importance because of concern for the potential long-term side effects of these medications.

THE BOTTOM LINE

The 2010 recommendations add a degree of complexity, with different medications recommended on the basis of glucocorticoid dose and duration as well as patient age, menopausal status, and childbearing potential. Guideline developers and clinicians face a difficult trade-off: easy-to-follow guidelines or more targeted guidelines that are more complex and therefore more difficult to use than previous guidelines.

This criticism is reasonable. The complexity is a result of insufficient evidence from clinical trials to make more exact and user-friendly recommendations, and also a result of the RAND/UCLA methodology. In cases that lack sufficient evidence on which to make a decision, the guideline development uses voting among experts in an attempt to develop consensus. This often results in complexity, lack of consensus, or inconsistencies.

The guidelines are straightforward for postmenopausal women and men age 50 and older on at least 7.5 mg prednisone for more than 3 months.

Since there is substantial evidence that many patients on glucocorticoid therapy go untreated, the risk of fracture in this population would be substantially reduced if clinicians would adhere to the recommendations.

Cardiac tamponade is a life-threatening condition that can be palliated or cured, depending on its cause and on the timeliness of treatment. Making a timely diagnosis and providing the appropriate treatment can be gratifying for both patient and physician.

Cardiac tamponade occurs when fluid in the pericardial space reaches a pressure exceeding central venous pressure. This leads to jugular venous distention, visceral organ engorgement, edema, and elevated pulmonary venous pressure that causes dyspnea. Despite compensatory tachycardia, the decrease in cardiac filling leads to a fall in cardiac output and to arterial hypoperfusion of vital organs.

PEARL 1: SLOW ACCUMULATION LEADS TO EDEMA

The rate at which pericardial fluid accumulates influences the clinical presentation of cardiac tamponade, in particular whether or not there is edema. Whereas rapid accumulation is characterized more by hypotension than by edema, the slow accumulation of pericardial fluid affords the patient time to drink enough liquid to keep the central venous pressure higher than the rising pericardial pressure. Thus, edema and dyspnea are more prominent features of cardiac tamponade when there is a slow rise in pericardial pressure.

PEARL 2: EDEMA IS NOT ALWAYS TREATED WITH A DIURETIC

Edema is not always treated with a diuretic. In a patient who has a pericardial effusion that has developed slowly and who has been drinking enough fluid to keep the central venous pressure higher than the pericardial pressure, a diuretic can remove enough volume from the circulation to lower the central venous pressure below the intrapericardial pressure and thus convert a benign pericardial effusion to potentially lethal cardiac tamponade.

One must understand the cause of edema or low urine output before treating it. This underscores the importance of the history and the physical examination. All of the following must be assessed:

• Symptoms and time course of the illness
• Concurrent medical illnesses
• Neck veins
• Blood pressure and its response to inspiration
• Heart sounds
• Heart rate and rhythm
• Abdominal organ engorgement
• Edema (or its absence).

PEARL 3: UNDERSTANDING THE CAUSE IS ESSENTIAL

Understanding the cause of cardiac tamponade is essential.

A trauma patient first encountered in the emergency department may have an underlying disease, but the focus is squarely on the effects of trauma or violent injury. In a patient with multiple trauma, hypotension and tachycardia that do not respond to intravenous volume replacement when there is an obvious rise in central venous pressure should be clues to cardiac tamponade.¹

If the patient has recently undergone a cardiac procedure (for example, cardiac surgery, myocardial biopsy, coronary intervention, electrophysiologic study with intracardiac electrodes, transvenous pacemaker placement, pacemaker lead extraction, or radiofrequency ablation), knowing about the procedure narrows the differential diagnosis when hypotension, tachycardia, and jugular venous distention develop.

PEARL 4: CARDIAC OR AORTIC RUPTURE REQUIRES SURGERY

When the etiology of cardiac tamponade is cardiac or aortic rupture, the treatment is surgical.

Painful sudden causes of cardiac tamponade include hemopericardium due to rupture of the free wall after myocardial infarction, and spontaneous or posttraumatic dissection and rupture of the ascending aorta. Prompt diagnosis is necessary, but since these lesions will not close and heal spontaneously, the definitive treatment should be surgery. Moreover, needle removal of intrapericardial blood that has been opposing further bleeding is sure to permit bleeding to recur, often with lethal consequences.²

Causes of cardiac tamponade that have a less-acute onset are likely to be complications of medical problems. Medical illnesses known to be associated with cardiac tamponade include:

• Infectious disease (idiopathic or viral, associated with smallpox vaccination, mycobacterial, purulent bacterial, fungal)
• Metastatic cancer (lung, breast, esophagus, lymphoma, pancreas, liver, leukemia, stomach, melanoma)³
• Connective tissue disease (rheumatoid arthritis, systemic lupus erythematosus, ankylosing spondylitis, scleroderma, Wegener granulomatosis, acute rheumatic fever)
• Endocrine disease (hypothyroidism)
• Drug side effects (procainamide, isoniazid, hydralazine, minoxidil, phenytoin, anticoagulants, methysergide)
• Inflammatory bowel disease (Crohn disease, ulcerative colitis)
• Congestive heart failure
• Uremia
• Radiation therapy
• Postmyocardial infarction syndrome (Dressler syndrome)
• Postpericardiotomy syndrome.

PEARL 5: REVIEW IMAGING BEFORE DIAGNOSING

What often brings a patient with cardiac tamponade to the attention of the physician is a finding on echocardiography, computed tomography, or magnetic resonance imaging of the chest.

Always review the imaging studies before making the diagnosis of cardiac tamponade. These tests must be reviewed to assess the anatomy and the size and location of the effusion. Particularly, one must look for atrial and right ventricular collapse and inferior vena caval plethora, which are echocardiographic signs of cardiac tamponade.⁴ Figures 1, 2, and 3 show imaging studies in a patient who presented with worsening cough 2 weeks after undergoing a cardiac procedure and who was found to have cardiac tamponade.

When the history and these imaging studies place cardiac tamponade high in the differential diagnosis as the cause of edema or dyspnea, it is time to reexamine the patient. The best first step is to measure pulsus paradoxus.

HOW PULSUS PARADOXUS OCCURS

To fully appreciate the subtleties of the next pearls, it is necessary to understand the pathophysiology of cardiac tamponade.

When pericardial fluid accumulation raises the pericardial pressure above the central venous pressure and pulmonary venous pressure (intravascular pressure), blood will not passively return to the right side of the heart from the venae cavae nor to the left side of the heart from the pulmonary veins unless it is influenced by the effects of respiration on intrathoracic pressure. During respiration, the right and left sides of the heart are alternately filled and deprived of their respective venous return.

During inspiration, as the intrathoracic pressure decreases, blood in the venae cavae empties into the right side of the heart, while blood in the pulmonary veins preferentially remains in the pulmonary veins, underfilling the left side of the heart. Since the right ventricle is more filled than the left ventricle during inspiration, the ventricular septum shifts from right to left, further opposing pulmonary venous return. As a result, during cardiac tamponade, the systemic blood pressure falls with inspiration.

During expiration the opposite occurs. Expiration decreases the intrathoracic volume, so the intrathoracic pressure rises. This tends to oppose vena caval return to the right side of the heart and to favor pulmonary venous return to the left side of the heart. The ventricular septum shifts from left to right, further accommodating left ventricular filling, raising stroke volume, and increasing blood pressure. This exaggerated alternate filling of the right and left sides of the heart during cardiac tamponade is what accounts for pulsus paradoxus, an inspiratory fall in systolic blood pressure of greater than 10 mm Hg.

If intravascular pressure is low (due to hemorrhage, dehydration, or diuretic therapy), the pressure in the pericardial space needed to oppose venous return is much less. In this low-pressure scenario, the results are low cardiac output and hypotension, which are treated by giving intravenous fluids to maintain intravascular volume.

PEARL 6: MEASURE PULSUS PARADOXUS

When cardiac tamponade is considered, one must always measure the pulsus paradoxus.

The term pulsus paradoxus was coined by Adolph Kussmaul in 1873, before physicians could even measure blood pressure. All they could do at that time was palpate the pulse and listen to the heart. Kussmaul described his observation as a conspicuous discrepancy between the cardiac action and the arterial pulse.

Although not described by Kussmaul, another explanation for this finding might be more suited to the use of the word “paradoxical.” When the pulse is palpated in a normal patient, with inspiration the pulse rate will increase via the Bainbridge reflex, and with expiration it will decrease. But in a patient with cardiac tamponade, there is a paradoxical inspiratory slowing of the pulse (because the decreased magnitude of the pulse at times makes it imperceptible) and an expiratory increase in pulse rate as the magnitude of the pulse again makes it palpable.

The magnitude of the fall in systolic blood pressure during inspiration has been used to estimate the level of hemodynamic impairment resulting from pericardial effusion.⁵ A rapidly accumulating pericardial effusion can have more hemodynamic impact than a much larger one that accumulates slowly. Thus, the intrapericardial pressure must be considered more than the volume of pericardial fluid.

When there is severe cardiac tamponade and overt pulsus paradoxus, simple palpation of a proximal arterial pulse can detect a marked inspiratory decrease or loss of the pulse, which returns with expiration. Tachycardia is almost always present, unless the cause is hypothyroidism.⁶

How to measure pulsus paradoxus with a manual sphygmomanometer

A stethoscope and manual sphygmomanometer are all that is needed to measure pulsus paradoxus. A noninvasive blood pressure monitor that averages multiple measurements cannot detect or quantify pulsus paradoxus.

The patient should be supine with the head elevated 30° to 45°, and the examiner should be seated comfortably at the patient’s side. The manometer should be on the opposite side of the patient in plain view of the examiner. Position the cuff on the arm above the elbow and place your stethoscope on the antecubital fossa. Then:

• Inflate the cuff 20 mm Hg above the highest systolic pressure previously auscultated.
• Slowly decrease the manometer pressure by 5 mm Hg and hold it there through two or three respiratory cycles while listening for the first Korotkoff (systolic) sound. Repeat this until you can hear the systolic sound (but only during expiration) and mentally note the pressure.
• Continue to decrease the manometer pressure by 5-mm Hg increments while listening. When the Korotkoff sounds no longer disappear with inspiration, mentally note this second value as well. The pulsus paradoxus is the difference between these values.
• When the Korotkoff sounds disappear as the manometer pressure is decreased, note this final value. This is the diastolic blood pressure.

PEARL 7: THE PLETHYSMOGRAM WAVE-FORM PARALLELS PULSUS PARADOXUS

Manual measurement of blood pressure and pulsus paradoxus can be difficult, especially in an obese patient or one with a fat-distorted arm on which the cuff does not maintain its position. In such patients, increased girth of the neck and abdomen also make it difficult to assess the jugular venous distention and visceral organ engorgement that characterize cardiac tamponade.

When the use of a sphygmomanometer is not possible, an arterial catheter can be inserted to demonstrate pulsus paradoxus. Simpler, however, is the novel use of another noninvasive instrument to detect and coarsely quantify pulsus paradoxus.⁷ The waveform on finger pulse oximetry can demonstrate pulsus paradoxus. The plethysmogram of the finger pulse oximeter can demonstrate the decrease in magnitude of the waveform with each inspiration (Figure 4).

Caution must be taken when interpreting this waveform, as with any measurement of pulsus paradoxus, to exclude a concomitant arrhythmia.

PEARL 8: PULSUS PARADOXUS WITHOUT CARDIAC TAMPONADE

Pulsus paradoxus can be present in the absence of cardiac tamponade. Once pulsus paradoxus of more than 10 mm Hg is measured, one must be sure the patient does not have a condition that can cause pulsus paradoxus without cardiac tamponade. Most of these are pulmonary conditions that necessitate an exaggerated inspiratory effort that can lower intrathoracic pressure sufficiently to oppose pulmonary venous return and cause a fall in systemic blood pressure:

• Chronic bronchitis
• Emphysema
• Mucus plug
• Pneumothorax
• Pulmonary embolism
• Stridor.

In these, there may be pulsus paradoxus, but not due to cardiac tamponade.

PEARL 9: CARDIAC TAMPONADE CAN BE PRESENT WITHOUT PULSUS PARADOXUS

Cardiac tamponade can be present without pulsus paradoxus. This occurs when certain conditions prevent inspiratory underfilling of the left ventricle relative to the filling of the right ventricle.⁸

How does this work? In cardiac tamponade, factors that drive the exaggerated fall in arterial pressure with inspiration (pulsus paradoxus) are the augmented right ventricular filling and the decreased left ventricular filling, both due to the lowering of the intrathoracic pressure. As the vena caval emptying is augmented, the right ventricular filling is increased, the ventricular septum shifts to the left, and pulmonary venous return to the heart is decreased.

Factors that can oppose pulsus paradoxus:

• Positive pressure ventilation prevents pulsus paradoxus by preventing the fall in intrathoracic pressure.
• Severe aortic regurgitation does not permit underfilling of the left ventricle during inspiration.
• An atrial septal defect will always equalize the right and left atrial pressures, preventing differential right ventricular and left ventricular filling with inspiration.
• Severe left ventricular hypertrophy does not permit the inspiratory shift of the ventricular septum from right to left that would otherwise lead to decreased left ventricular filling.
• Severe left ventricular dysfunction, with its low stroke volume and severe elevation of left ventricular end-diastolic pressure, never permits underfilling of the left ventricle, despite cardiac tamponade and an inspiratory decrease in intrathoracic pressure.
• Intravascular volume depletion due to hemorrhage, hemodialysis, or mistaken use of diuretics to treat edema can cause marked hypotension, making pulsus paradoxus impossible to detect.

Knowledge of underlying medical conditions, the likelihood of their causing cardiac tamponade, and the appearance of the echocardiogram prompt the physician to look further when the presence or absence of pulsus paradoxus does not fit with the working diagnosis.

The echocardiogram can give hints to the etiology of a pericardial effusion, such as clotted blood after trauma or a cardiac-perforating procedure, tumor studding of the epicardium,⁹ or fibrin strands indicating chronicity or an inflammatory process.¹⁰ Diastolic collapse of the right ventricle, more than collapse of the right atrium or left atrium, speaks for the severity of cardiac tamponade. With hemodynamically significant pericardial effusion and cardiac tamponade, the inferior vena cava is distended and does not decrease in size with inspiration unless there is severe intravascular volume depletion, at which time the inferior vena cava is underfilled throughout the respiratory cycle.

PEARL 10: PLAN HOW TO DRAIN

The size and location of the pericardial effusion and the patient’s hemodynamics must be integrated when deciding how to relieve cardiac tamponade. When cardiac tamponade is indeed severe and the patient and physician agree that it must be drained, the options are percutaneous needle aspiration (pericardiocentesis) and surgical pericardiostomy (creation of a pericardial window). Here again, as assessed by echocardiography, the access to the pericardial fluid should influence the choice.

Pericardiocentesis can be safely done if certain criteria are met. The patient must be able to lie still in the supine position, perhaps with the head of the bed elevated 30 degrees. Anticoagulation must be reversed or allowed time to resolve if drainage is not an emergency.

Pericardiocentesis can be risky or unsuccessful if there is not enough pericardial fluid to permit respiratory cardiac motion without perforating the heart with the needle; if the effusion is loculated (confined to a pocket) posteriorly; or if it is too far from the skin to permit precise control and placement of a spinal needle into the pericardial space. In cases of cardiac tamponade in which the anatomy indicates surgical pericardiostomy but severe hypotension prevents the induction of anesthesia and positive-pressure ventilation—which can result in profound, irreversible hypotension—percutaneous needle drainage (pericardiocentesis) should be performed in the operating room to relieve the tamponade before the induction of anesthesia and the surgical drainage.¹¹

To reiterate, a suspected cardiac or aortic rupture that causes cardiac tamponade is usually large and not apt to self-seal. In such cases, the halt in the accumulation of pericardial blood is due to hypotension and not due to spontaneous resolution. Open surgical drainage is required from the outset because an initial success of pericardiocentesis yields to the recurrence of cardiac tamponade.

PEARL 11: ANTICIPATE WHAT THE FLUID SHOULD LOOK LIKE

Before performing pericardiocentesis, anticipate the appearance of the pericardial fluid on the basis of the presumed etiology, ie:

• Sanguinous—trauma, heart surgery, cardiac perforation from a procedure, anticoagulation, uremia, or malignancy
• Serous—congestive heart failure, acute radiation therapy
• Purulent—infections (natural or postoperative)
• Turbid (like gold paint)—mycobacterial infection, rheumatoid arthritis, myxedema
• Chylous—pericardium fistulized to the thoracic duct by a natural or postsurgical cause.

Sanguinous pericardial effusion encountered during a pericardiocentesis, if not anticipated, can be daunting and can cause the operator to question if it is the result of inadvertent needle placement in a cardiac chamber. If the needle is indeed in the heart, blood often surges out under pressure in pulses, which strongly suggests that the needle is not in the pericardial space and should be removed; but if confirmation of the location is needed before removing the needle, it can be done by injecting 2 mL of agitated sterile saline through the pericardiocentesis needle during echocardiographic imaging.¹²

Before inserting the needle, the ideal access location and needle angle must be determined by the operator with echocardiographic transducer in hand. The distance from skin to a point just through the parietal pericardium can also be measured at this time.

Once the needle is in the pericardial fluid (and you are confident of its placement), removal of 50 to 100 mL of the fluid with a large syringe can be enough to afford the patient easier breathing, higher blood pressure, and lower pulsus paradoxus—and even the physician will breathe easier. The same syringe can be filled and emptied multiple times. Less traumatic and more complete removal of pericardial fluid requires insertion of a multihole pigtail catheter over a J-tipped guidewire that is introduced through the needle.

PEARL 12: DRAIN SLOWLY TO AVOID PULMONARY EDEMA

Pulmonary edema is an uncommon complication of pericardiocentesis that might be avoidable. Heralded by sudden coughing and pink, frothy sputum, it can rapidly deteriorate into respiratory failure. The mechanism has been attributed to a sudden increase in right ventricular stroke volume and resultant left ventricular filling after the excess pericardial fluid has been removed, before the systemic arteries, which constrict to keep the systemic blood pressure up during cardiac tamponade, have had time to relax.¹³

To avoid this complication, if the volume of pericardial fluid responsible for cardiac tamponade is large, it should be removed slowly,¹⁴ stopping for a several-minute rest after each 250 mL. Catheter removal of pericardial fluid by gravity drainage over 24 hours has been suggested.¹⁵ A drawback to this approach is catheter clotting or sludging before all the fluid has been removed. It is helpful to keep the drainage catheter close to the patient’s body temperature to make the fluid less viscous. Output should be monitored hourly.

When the pericardial fluid has been completely drained, one must decide how long to leave the catheter in. One reason to remove the catheter at this time is that it causes pleuritic pain; another is to avoid introducing infection. A reason to leave the catheter in is to observe the effect of medical treatment on the hourly pericardial fluid output. Nonsteroidal anti-inflammatory drugs are the drugs of first choice when treating pericardial inflammation and suppressing production of pericardial fluid.¹⁶ In most cases the catheter should not be left in place for more than 3 days.

Laboratory analysis of the pericardial fluid should shed light on its suspected cause. Analysis usually involves chemistry testing, microscopic inspection of blood cell smears, cytology, microbiologic stains and cultures, and immunologic tests. Results often take days. Meyers and colleagues¹⁷ expound on this subject.

Cardiac tamponade is a life-threatening condition that can be palliated or cured, depending on its cause and on the timeliness of treatment. Making a timely diagnosis and providing the appropriate treatment can be gratifying for both patient and physician.

Cardiac tamponade occurs when fluid in the pericardial space reaches a pressure exceeding central venous pressure. This leads to jugular venous distention, visceral organ engorgement, edema, and elevated pulmonary venous pressure that causes dyspnea. Despite compensatory tachycardia, the decrease in cardiac filling leads to a fall in cardiac output and to arterial hypoperfusion of vital organs.

PEARL 1: SLOW ACCUMULATION LEADS TO EDEMA

The rate at which pericardial fluid accumulates influences the clinical presentation of cardiac tamponade, in particular whether or not there is edema. Whereas rapid accumulation is characterized more by hypotension than by edema, the slow accumulation of pericardial fluid affords the patient time to drink enough liquid to keep the central venous pressure higher than the rising pericardial pressure. Thus, edema and dyspnea are more prominent features of cardiac tamponade when there is a slow rise in pericardial pressure.

PEARL 2: EDEMA IS NOT ALWAYS TREATED WITH A DIURETIC

Edema is not always treated with a diuretic. In a patient who has a pericardial effusion that has developed slowly and who has been drinking enough fluid to keep the central venous pressure higher than the pericardial pressure, a diuretic can remove enough volume from the circulation to lower the central venous pressure below the intrapericardial pressure and thus convert a benign pericardial effusion to potentially lethal cardiac tamponade.

One must understand the cause of edema or low urine output before treating it. This underscores the importance of the history and the physical examination. All of the following must be assessed:

• Symptoms and time course of the illness
• Concurrent medical illnesses
• Neck veins
• Blood pressure and its response to inspiration
• Heart sounds
• Heart rate and rhythm
• Abdominal organ engorgement
• Edema (or its absence).

PEARL 3: UNDERSTANDING THE CAUSE IS ESSENTIAL

Understanding the cause of cardiac tamponade is essential.

A trauma patient first encountered in the emergency department may have an underlying disease, but the focus is squarely on the effects of trauma or violent injury. In a patient with multiple trauma, hypotension and tachycardia that do not respond to intravenous volume replacement when there is an obvious rise in central venous pressure should be clues to cardiac tamponade.¹

If the patient has recently undergone a cardiac procedure (for example, cardiac surgery, myocardial biopsy, coronary intervention, electrophysiologic study with intracardiac electrodes, transvenous pacemaker placement, pacemaker lead extraction, or radiofrequency ablation), knowing about the procedure narrows the differential diagnosis when hypotension, tachycardia, and jugular venous distention develop.

PEARL 4: CARDIAC OR AORTIC RUPTURE REQUIRES SURGERY

When the etiology of cardiac tamponade is cardiac or aortic rupture, the treatment is surgical.

Painful sudden causes of cardiac tamponade include hemopericardium due to rupture of the free wall after myocardial infarction, and spontaneous or posttraumatic dissection and rupture of the ascending aorta. Prompt diagnosis is necessary, but since these lesions will not close and heal spontaneously, the definitive treatment should be surgery. Moreover, needle removal of intrapericardial blood that has been opposing further bleeding is sure to permit bleeding to recur, often with lethal consequences.²

Causes of cardiac tamponade that have a less-acute onset are likely to be complications of medical problems. Medical illnesses known to be associated with cardiac tamponade include:

• Infectious disease (idiopathic or viral, associated with smallpox vaccination, mycobacterial, purulent bacterial, fungal)
• Metastatic cancer (lung, breast, esophagus, lymphoma, pancreas, liver, leukemia, stomach, melanoma)³
• Connective tissue disease (rheumatoid arthritis, systemic lupus erythematosus, ankylosing spondylitis, scleroderma, Wegener granulomatosis, acute rheumatic fever)
• Endocrine disease (hypothyroidism)
• Drug side effects (procainamide, isoniazid, hydralazine, minoxidil, phenytoin, anticoagulants, methysergide)
• Inflammatory bowel disease (Crohn disease, ulcerative colitis)
• Congestive heart failure
• Uremia
• Radiation therapy
• Postmyocardial infarction syndrome (Dressler syndrome)
• Postpericardiotomy syndrome.

PEARL 5: REVIEW IMAGING BEFORE DIAGNOSING

What often brings a patient with cardiac tamponade to the attention of the physician is a finding on echocardiography, computed tomography, or magnetic resonance imaging of the chest.

Always review the imaging studies before making the diagnosis of cardiac tamponade. These tests must be reviewed to assess the anatomy and the size and location of the effusion. Particularly, one must look for atrial and right ventricular collapse and inferior vena caval plethora, which are echocardiographic signs of cardiac tamponade.⁴ Figures 1, 2, and 3 show imaging studies in a patient who presented with worsening cough 2 weeks after undergoing a cardiac procedure and who was found to have cardiac tamponade.

When the history and these imaging studies place cardiac tamponade high in the differential diagnosis as the cause of edema or dyspnea, it is time to reexamine the patient. The best first step is to measure pulsus paradoxus.

HOW PULSUS PARADOXUS OCCURS

To fully appreciate the subtleties of the next pearls, it is necessary to understand the pathophysiology of cardiac tamponade.

When pericardial fluid accumulation raises the pericardial pressure above the central venous pressure and pulmonary venous pressure (intravascular pressure), blood will not passively return to the right side of the heart from the venae cavae nor to the left side of the heart from the pulmonary veins unless it is influenced by the effects of respiration on intrathoracic pressure. During respiration, the right and left sides of the heart are alternately filled and deprived of their respective venous return.

During inspiration, as the intrathoracic pressure decreases, blood in the venae cavae empties into the right side of the heart, while blood in the pulmonary veins preferentially remains in the pulmonary veins, underfilling the left side of the heart. Since the right ventricle is more filled than the left ventricle during inspiration, the ventricular septum shifts from right to left, further opposing pulmonary venous return. As a result, during cardiac tamponade, the systemic blood pressure falls with inspiration.

During expiration the opposite occurs. Expiration decreases the intrathoracic volume, so the intrathoracic pressure rises. This tends to oppose vena caval return to the right side of the heart and to favor pulmonary venous return to the left side of the heart. The ventricular septum shifts from left to right, further accommodating left ventricular filling, raising stroke volume, and increasing blood pressure. This exaggerated alternate filling of the right and left sides of the heart during cardiac tamponade is what accounts for pulsus paradoxus, an inspiratory fall in systolic blood pressure of greater than 10 mm Hg.

If intravascular pressure is low (due to hemorrhage, dehydration, or diuretic therapy), the pressure in the pericardial space needed to oppose venous return is much less. In this low-pressure scenario, the results are low cardiac output and hypotension, which are treated by giving intravenous fluids to maintain intravascular volume.

PEARL 6: MEASURE PULSUS PARADOXUS

When cardiac tamponade is considered, one must always measure the pulsus paradoxus.

The term pulsus paradoxus was coined by Adolph Kussmaul in 1873, before physicians could even measure blood pressure. All they could do at that time was palpate the pulse and listen to the heart. Kussmaul described his observation as a conspicuous discrepancy between the cardiac action and the arterial pulse.

Although not described by Kussmaul, another explanation for this finding might be more suited to the use of the word “paradoxical.” When the pulse is palpated in a normal patient, with inspiration the pulse rate will increase via the Bainbridge reflex, and with expiration it will decrease. But in a patient with cardiac tamponade, there is a paradoxical inspiratory slowing of the pulse (because the decreased magnitude of the pulse at times makes it imperceptible) and an expiratory increase in pulse rate as the magnitude of the pulse again makes it palpable.

The magnitude of the fall in systolic blood pressure during inspiration has been used to estimate the level of hemodynamic impairment resulting from pericardial effusion.⁵ A rapidly accumulating pericardial effusion can have more hemodynamic impact than a much larger one that accumulates slowly. Thus, the intrapericardial pressure must be considered more than the volume of pericardial fluid.

When there is severe cardiac tamponade and overt pulsus paradoxus, simple palpation of a proximal arterial pulse can detect a marked inspiratory decrease or loss of the pulse, which returns with expiration. Tachycardia is almost always present, unless the cause is hypothyroidism.⁶

How to measure pulsus paradoxus with a manual sphygmomanometer

A stethoscope and manual sphygmomanometer are all that is needed to measure pulsus paradoxus. A noninvasive blood pressure monitor that averages multiple measurements cannot detect or quantify pulsus paradoxus.

The patient should be supine with the head elevated 30° to 45°, and the examiner should be seated comfortably at the patient’s side. The manometer should be on the opposite side of the patient in plain view of the examiner. Position the cuff on the arm above the elbow and place your stethoscope on the antecubital fossa. Then:

• Inflate the cuff 20 mm Hg above the highest systolic pressure previously auscultated.
• Slowly decrease the manometer pressure by 5 mm Hg and hold it there through two or three respiratory cycles while listening for the first Korotkoff (systolic) sound. Repeat this until you can hear the systolic sound (but only during expiration) and mentally note the pressure.
• Continue to decrease the manometer pressure by 5-mm Hg increments while listening. When the Korotkoff sounds no longer disappear with inspiration, mentally note this second value as well. The pulsus paradoxus is the difference between these values.
• When the Korotkoff sounds disappear as the manometer pressure is decreased, note this final value. This is the diastolic blood pressure.

PEARL 7: THE PLETHYSMOGRAM WAVE-FORM PARALLELS PULSUS PARADOXUS

Manual measurement of blood pressure and pulsus paradoxus can be difficult, especially in an obese patient or one with a fat-distorted arm on which the cuff does not maintain its position. In such patients, increased girth of the neck and abdomen also make it difficult to assess the jugular venous distention and visceral organ engorgement that characterize cardiac tamponade.

When the use of a sphygmomanometer is not possible, an arterial catheter can be inserted to demonstrate pulsus paradoxus. Simpler, however, is the novel use of another noninvasive instrument to detect and coarsely quantify pulsus paradoxus.⁷ The waveform on finger pulse oximetry can demonstrate pulsus paradoxus. The plethysmogram of the finger pulse oximeter can demonstrate the decrease in magnitude of the waveform with each inspiration (Figure 4).

Caution must be taken when interpreting this waveform, as with any measurement of pulsus paradoxus, to exclude a concomitant arrhythmia.

PEARL 8: PULSUS PARADOXUS WITHOUT CARDIAC TAMPONADE

Pulsus paradoxus can be present in the absence of cardiac tamponade. Once pulsus paradoxus of more than 10 mm Hg is measured, one must be sure the patient does not have a condition that can cause pulsus paradoxus without cardiac tamponade. Most of these are pulmonary conditions that necessitate an exaggerated inspiratory effort that can lower intrathoracic pressure sufficiently to oppose pulmonary venous return and cause a fall in systemic blood pressure:

• Chronic bronchitis
• Emphysema
• Mucus plug
• Pneumothorax
• Pulmonary embolism
• Stridor.

In these, there may be pulsus paradoxus, but not due to cardiac tamponade.

PEARL 9: CARDIAC TAMPONADE CAN BE PRESENT WITHOUT PULSUS PARADOXUS

Cardiac tamponade can be present without pulsus paradoxus. This occurs when certain conditions prevent inspiratory underfilling of the left ventricle relative to the filling of the right ventricle.⁸

How does this work? In cardiac tamponade, factors that drive the exaggerated fall in arterial pressure with inspiration (pulsus paradoxus) are the augmented right ventricular filling and the decreased left ventricular filling, both due to the lowering of the intrathoracic pressure. As the vena caval emptying is augmented, the right ventricular filling is increased, the ventricular septum shifts to the left, and pulmonary venous return to the heart is decreased.

Factors that can oppose pulsus paradoxus:

• Positive pressure ventilation prevents pulsus paradoxus by preventing the fall in intrathoracic pressure.
• Severe aortic regurgitation does not permit underfilling of the left ventricle during inspiration.
• An atrial septal defect will always equalize the right and left atrial pressures, preventing differential right ventricular and left ventricular filling with inspiration.
• Severe left ventricular hypertrophy does not permit the inspiratory shift of the ventricular septum from right to left that would otherwise lead to decreased left ventricular filling.
• Severe left ventricular dysfunction, with its low stroke volume and severe elevation of left ventricular end-diastolic pressure, never permits underfilling of the left ventricle, despite cardiac tamponade and an inspiratory decrease in intrathoracic pressure.
• Intravascular volume depletion due to hemorrhage, hemodialysis, or mistaken use of diuretics to treat edema can cause marked hypotension, making pulsus paradoxus impossible to detect.

Knowledge of underlying medical conditions, the likelihood of their causing cardiac tamponade, and the appearance of the echocardiogram prompt the physician to look further when the presence or absence of pulsus paradoxus does not fit with the working diagnosis.

The echocardiogram can give hints to the etiology of a pericardial effusion, such as clotted blood after trauma or a cardiac-perforating procedure, tumor studding of the epicardium,⁹ or fibrin strands indicating chronicity or an inflammatory process.¹⁰ Diastolic collapse of the right ventricle, more than collapse of the right atrium or left atrium, speaks for the severity of cardiac tamponade. With hemodynamically significant pericardial effusion and cardiac tamponade, the inferior vena cava is distended and does not decrease in size with inspiration unless there is severe intravascular volume depletion, at which time the inferior vena cava is underfilled throughout the respiratory cycle.

PEARL 10: PLAN HOW TO DRAIN

The size and location of the pericardial effusion and the patient’s hemodynamics must be integrated when deciding how to relieve cardiac tamponade. When cardiac tamponade is indeed severe and the patient and physician agree that it must be drained, the options are percutaneous needle aspiration (pericardiocentesis) and surgical pericardiostomy (creation of a pericardial window). Here again, as assessed by echocardiography, the access to the pericardial fluid should influence the choice.

Pericardiocentesis can be safely done if certain criteria are met. The patient must be able to lie still in the supine position, perhaps with the head of the bed elevated 30 degrees. Anticoagulation must be reversed or allowed time to resolve if drainage is not an emergency.

Pericardiocentesis can be risky or unsuccessful if there is not enough pericardial fluid to permit respiratory cardiac motion without perforating the heart with the needle; if the effusion is loculated (confined to a pocket) posteriorly; or if it is too far from the skin to permit precise control and placement of a spinal needle into the pericardial space. In cases of cardiac tamponade in which the anatomy indicates surgical pericardiostomy but severe hypotension prevents the induction of anesthesia and positive-pressure ventilation—which can result in profound, irreversible hypotension—percutaneous needle drainage (pericardiocentesis) should be performed in the operating room to relieve the tamponade before the induction of anesthesia and the surgical drainage.¹¹

To reiterate, a suspected cardiac or aortic rupture that causes cardiac tamponade is usually large and not apt to self-seal. In such cases, the halt in the accumulation of pericardial blood is due to hypotension and not due to spontaneous resolution. Open surgical drainage is required from the outset because an initial success of pericardiocentesis yields to the recurrence of cardiac tamponade.

PEARL 11: ANTICIPATE WHAT THE FLUID SHOULD LOOK LIKE

Before performing pericardiocentesis, anticipate the appearance of the pericardial fluid on the basis of the presumed etiology, ie:

• Sanguinous—trauma, heart surgery, cardiac perforation from a procedure, anticoagulation, uremia, or malignancy
• Serous—congestive heart failure, acute radiation therapy
• Purulent—infections (natural or postoperative)
• Turbid (like gold paint)—mycobacterial infection, rheumatoid arthritis, myxedema
• Chylous—pericardium fistulized to the thoracic duct by a natural or postsurgical cause.

Sanguinous pericardial effusion encountered during a pericardiocentesis, if not anticipated, can be daunting and can cause the operator to question if it is the result of inadvertent needle placement in a cardiac chamber. If the needle is indeed in the heart, blood often surges out under pressure in pulses, which strongly suggests that the needle is not in the pericardial space and should be removed; but if confirmation of the location is needed before removing the needle, it can be done by injecting 2 mL of agitated sterile saline through the pericardiocentesis needle during echocardiographic imaging.¹²

Before inserting the needle, the ideal access location and needle angle must be determined by the operator with echocardiographic transducer in hand. The distance from skin to a point just through the parietal pericardium can also be measured at this time.

Once the needle is in the pericardial fluid (and you are confident of its placement), removal of 50 to 100 mL of the fluid with a large syringe can be enough to afford the patient easier breathing, higher blood pressure, and lower pulsus paradoxus—and even the physician will breathe easier. The same syringe can be filled and emptied multiple times. Less traumatic and more complete removal of pericardial fluid requires insertion of a multihole pigtail catheter over a J-tipped guidewire that is introduced through the needle.

PEARL 12: DRAIN SLOWLY TO AVOID PULMONARY EDEMA

Pulmonary edema is an uncommon complication of pericardiocentesis that might be avoidable. Heralded by sudden coughing and pink, frothy sputum, it can rapidly deteriorate into respiratory failure. The mechanism has been attributed to a sudden increase in right ventricular stroke volume and resultant left ventricular filling after the excess pericardial fluid has been removed, before the systemic arteries, which constrict to keep the systemic blood pressure up during cardiac tamponade, have had time to relax.¹³

To avoid this complication, if the volume of pericardial fluid responsible for cardiac tamponade is large, it should be removed slowly,¹⁴ stopping for a several-minute rest after each 250 mL. Catheter removal of pericardial fluid by gravity drainage over 24 hours has been suggested.¹⁵ A drawback to this approach is catheter clotting or sludging before all the fluid has been removed. It is helpful to keep the drainage catheter close to the patient’s body temperature to make the fluid less viscous. Output should be monitored hourly.

When the pericardial fluid has been completely drained, one must decide how long to leave the catheter in. One reason to remove the catheter at this time is that it causes pleuritic pain; another is to avoid introducing infection. A reason to leave the catheter in is to observe the effect of medical treatment on the hourly pericardial fluid output. Nonsteroidal anti-inflammatory drugs are the drugs of first choice when treating pericardial inflammation and suppressing production of pericardial fluid.¹⁶ In most cases the catheter should not be left in place for more than 3 days.

Laboratory analysis of the pericardial fluid should shed light on its suspected cause. Analysis usually involves chemistry testing, microscopic inspection of blood cell smears, cytology, microbiologic stains and cultures, and immunologic tests. Results often take days. Meyers and colleagues¹⁷ expound on this subject.

Cardiac tamponade is a life-threatening condition that can be palliated or cured, depending on its cause and on the timeliness of treatment. Making a timely diagnosis and providing the appropriate treatment can be gratifying for both patient and physician.

Cardiac tamponade occurs when fluid in the pericardial space reaches a pressure exceeding central venous pressure. This leads to jugular venous distention, visceral organ engorgement, edema, and elevated pulmonary venous pressure that causes dyspnea. Despite compensatory tachycardia, the decrease in cardiac filling leads to a fall in cardiac output and to arterial hypoperfusion of vital organs.

PEARL 1: SLOW ACCUMULATION LEADS TO EDEMA

The rate at which pericardial fluid accumulates influences the clinical presentation of cardiac tamponade, in particular whether or not there is edema. Whereas rapid accumulation is characterized more by hypotension than by edema, the slow accumulation of pericardial fluid affords the patient time to drink enough liquid to keep the central venous pressure higher than the rising pericardial pressure. Thus, edema and dyspnea are more prominent features of cardiac tamponade when there is a slow rise in pericardial pressure.

PEARL 2: EDEMA IS NOT ALWAYS TREATED WITH A DIURETIC

Edema is not always treated with a diuretic. In a patient who has a pericardial effusion that has developed slowly and who has been drinking enough fluid to keep the central venous pressure higher than the pericardial pressure, a diuretic can remove enough volume from the circulation to lower the central venous pressure below the intrapericardial pressure and thus convert a benign pericardial effusion to potentially lethal cardiac tamponade.

One must understand the cause of edema or low urine output before treating it. This underscores the importance of the history and the physical examination. All of the following must be assessed:

• Symptoms and time course of the illness
• Concurrent medical illnesses
• Neck veins
• Blood pressure and its response to inspiration
• Heart sounds
• Heart rate and rhythm
• Abdominal organ engorgement
• Edema (or its absence).

PEARL 3: UNDERSTANDING THE CAUSE IS ESSENTIAL

Understanding the cause of cardiac tamponade is essential.

A trauma patient first encountered in the emergency department may have an underlying disease, but the focus is squarely on the effects of trauma or violent injury. In a patient with multiple trauma, hypotension and tachycardia that do not respond to intravenous volume replacement when there is an obvious rise in central venous pressure should be clues to cardiac tamponade.¹

If the patient has recently undergone a cardiac procedure (for example, cardiac surgery, myocardial biopsy, coronary intervention, electrophysiologic study with intracardiac electrodes, transvenous pacemaker placement, pacemaker lead extraction, or radiofrequency ablation), knowing about the procedure narrows the differential diagnosis when hypotension, tachycardia, and jugular venous distention develop.

PEARL 4: CARDIAC OR AORTIC RUPTURE REQUIRES SURGERY

When the etiology of cardiac tamponade is cardiac or aortic rupture, the treatment is surgical.

Painful sudden causes of cardiac tamponade include hemopericardium due to rupture of the free wall after myocardial infarction, and spontaneous or posttraumatic dissection and rupture of the ascending aorta. Prompt diagnosis is necessary, but since these lesions will not close and heal spontaneously, the definitive treatment should be surgery. Moreover, needle removal of intrapericardial blood that has been opposing further bleeding is sure to permit bleeding to recur, often with lethal consequences.²

Causes of cardiac tamponade that have a less-acute onset are likely to be complications of medical problems. Medical illnesses known to be associated with cardiac tamponade include:

• Infectious disease (idiopathic or viral, associated with smallpox vaccination, mycobacterial, purulent bacterial, fungal)
• Metastatic cancer (lung, breast, esophagus, lymphoma, pancreas, liver, leukemia, stomach, melanoma)³
• Connective tissue disease (rheumatoid arthritis, systemic lupus erythematosus, ankylosing spondylitis, scleroderma, Wegener granulomatosis, acute rheumatic fever)
• Endocrine disease (hypothyroidism)
• Drug side effects (procainamide, isoniazid, hydralazine, minoxidil, phenytoin, anticoagulants, methysergide)
• Inflammatory bowel disease (Crohn disease, ulcerative colitis)
• Congestive heart failure
• Uremia
• Radiation therapy
• Postmyocardial infarction syndrome (Dressler syndrome)
• Postpericardiotomy syndrome.

PEARL 5: REVIEW IMAGING BEFORE DIAGNOSING

What often brings a patient with cardiac tamponade to the attention of the physician is a finding on echocardiography, computed tomography, or magnetic resonance imaging of the chest.

Always review the imaging studies before making the diagnosis of cardiac tamponade. These tests must be reviewed to assess the anatomy and the size and location of the effusion. Particularly, one must look for atrial and right ventricular collapse and inferior vena caval plethora, which are echocardiographic signs of cardiac tamponade.⁴ Figures 1, 2, and 3 show imaging studies in a patient who presented with worsening cough 2 weeks after undergoing a cardiac procedure and who was found to have cardiac tamponade.

When the history and these imaging studies place cardiac tamponade high in the differential diagnosis as the cause of edema or dyspnea, it is time to reexamine the patient. The best first step is to measure pulsus paradoxus.

HOW PULSUS PARADOXUS OCCURS

To fully appreciate the subtleties of the next pearls, it is necessary to understand the pathophysiology of cardiac tamponade.

When pericardial fluid accumulation raises the pericardial pressure above the central venous pressure and pulmonary venous pressure (intravascular pressure), blood will not passively return to the right side of the heart from the venae cavae nor to the left side of the heart from the pulmonary veins unless it is influenced by the effects of respiration on intrathoracic pressure. During respiration, the right and left sides of the heart are alternately filled and deprived of their respective venous return.

During inspiration, as the intrathoracic pressure decreases, blood in the venae cavae empties into the right side of the heart, while blood in the pulmonary veins preferentially remains in the pulmonary veins, underfilling the left side of the heart. Since the right ventricle is more filled than the left ventricle during inspiration, the ventricular septum shifts from right to left, further opposing pulmonary venous return. As a result, during cardiac tamponade, the systemic blood pressure falls with inspiration.

During expiration the opposite occurs. Expiration decreases the intrathoracic volume, so the intrathoracic pressure rises. This tends to oppose vena caval return to the right side of the heart and to favor pulmonary venous return to the left side of the heart. The ventricular septum shifts from left to right, further accommodating left ventricular filling, raising stroke volume, and increasing blood pressure. This exaggerated alternate filling of the right and left sides of the heart during cardiac tamponade is what accounts for pulsus paradoxus, an inspiratory fall in systolic blood pressure of greater than 10 mm Hg.

If intravascular pressure is low (due to hemorrhage, dehydration, or diuretic therapy), the pressure in the pericardial space needed to oppose venous return is much less. In this low-pressure scenario, the results are low cardiac output and hypotension, which are treated by giving intravenous fluids to maintain intravascular volume.

PEARL 6: MEASURE PULSUS PARADOXUS

When cardiac tamponade is considered, one must always measure the pulsus paradoxus.

The term pulsus paradoxus was coined by Adolph Kussmaul in 1873, before physicians could even measure blood pressure. All they could do at that time was palpate the pulse and listen to the heart. Kussmaul described his observation as a conspicuous discrepancy between the cardiac action and the arterial pulse.

Although not described by Kussmaul, another explanation for this finding might be more suited to the use of the word “paradoxical.” When the pulse is palpated in a normal patient, with inspiration the pulse rate will increase via the Bainbridge reflex, and with expiration it will decrease. But in a patient with cardiac tamponade, there is a paradoxical inspiratory slowing of the pulse (because the decreased magnitude of the pulse at times makes it imperceptible) and an expiratory increase in pulse rate as the magnitude of the pulse again makes it palpable.

The magnitude of the fall in systolic blood pressure during inspiration has been used to estimate the level of hemodynamic impairment resulting from pericardial effusion.⁵ A rapidly accumulating pericardial effusion can have more hemodynamic impact than a much larger one that accumulates slowly. Thus, the intrapericardial pressure must be considered more than the volume of pericardial fluid.

When there is severe cardiac tamponade and overt pulsus paradoxus, simple palpation of a proximal arterial pulse can detect a marked inspiratory decrease or loss of the pulse, which returns with expiration. Tachycardia is almost always present, unless the cause is hypothyroidism.⁶

How to measure pulsus paradoxus with a manual sphygmomanometer

A stethoscope and manual sphygmomanometer are all that is needed to measure pulsus paradoxus. A noninvasive blood pressure monitor that averages multiple measurements cannot detect or quantify pulsus paradoxus.

The patient should be supine with the head elevated 30° to 45°, and the examiner should be seated comfortably at the patient’s side. The manometer should be on the opposite side of the patient in plain view of the examiner. Position the cuff on the arm above the elbow and place your stethoscope on the antecubital fossa. Then:

• Inflate the cuff 20 mm Hg above the highest systolic pressure previously auscultated.
• Slowly decrease the manometer pressure by 5 mm Hg and hold it there through two or three respiratory cycles while listening for the first Korotkoff (systolic) sound. Repeat this until you can hear the systolic sound (but only during expiration) and mentally note the pressure.
• Continue to decrease the manometer pressure by 5-mm Hg increments while listening. When the Korotkoff sounds no longer disappear with inspiration, mentally note this second value as well. The pulsus paradoxus is the difference between these values.
• When the Korotkoff sounds disappear as the manometer pressure is decreased, note this final value. This is the diastolic blood pressure.

PEARL 7: THE PLETHYSMOGRAM WAVE-FORM PARALLELS PULSUS PARADOXUS

Manual measurement of blood pressure and pulsus paradoxus can be difficult, especially in an obese patient or one with a fat-distorted arm on which the cuff does not maintain its position. In such patients, increased girth of the neck and abdomen also make it difficult to assess the jugular venous distention and visceral organ engorgement that characterize cardiac tamponade.

When the use of a sphygmomanometer is not possible, an arterial catheter can be inserted to demonstrate pulsus paradoxus. Simpler, however, is the novel use of another noninvasive instrument to detect and coarsely quantify pulsus paradoxus.⁷ The waveform on finger pulse oximetry can demonstrate pulsus paradoxus. The plethysmogram of the finger pulse oximeter can demonstrate the decrease in magnitude of the waveform with each inspiration (Figure 4).

Caution must be taken when interpreting this waveform, as with any measurement of pulsus paradoxus, to exclude a concomitant arrhythmia.

PEARL 8: PULSUS PARADOXUS WITHOUT CARDIAC TAMPONADE

Pulsus paradoxus can be present in the absence of cardiac tamponade. Once pulsus paradoxus of more than 10 mm Hg is measured, one must be sure the patient does not have a condition that can cause pulsus paradoxus without cardiac tamponade. Most of these are pulmonary conditions that necessitate an exaggerated inspiratory effort that can lower intrathoracic pressure sufficiently to oppose pulmonary venous return and cause a fall in systemic blood pressure:

• Chronic bronchitis
• Emphysema
• Mucus plug
• Pneumothorax
• Pulmonary embolism
• Stridor.

In these, there may be pulsus paradoxus, but not due to cardiac tamponade.

PEARL 9: CARDIAC TAMPONADE CAN BE PRESENT WITHOUT PULSUS PARADOXUS

Cardiac tamponade can be present without pulsus paradoxus. This occurs when certain conditions prevent inspiratory underfilling of the left ventricle relative to the filling of the right ventricle.⁸

How does this work? In cardiac tamponade, factors that drive the exaggerated fall in arterial pressure with inspiration (pulsus paradoxus) are the augmented right ventricular filling and the decreased left ventricular filling, both due to the lowering of the intrathoracic pressure. As the vena caval emptying is augmented, the right ventricular filling is increased, the ventricular septum shifts to the left, and pulmonary venous return to the heart is decreased.

Factors that can oppose pulsus paradoxus:

• Positive pressure ventilation prevents pulsus paradoxus by preventing the fall in intrathoracic pressure.
• Severe aortic regurgitation does not permit underfilling of the left ventricle during inspiration.
• An atrial septal defect will always equalize the right and left atrial pressures, preventing differential right ventricular and left ventricular filling with inspiration.
• Severe left ventricular hypertrophy does not permit the inspiratory shift of the ventricular septum from right to left that would otherwise lead to decreased left ventricular filling.
• Severe left ventricular dysfunction, with its low stroke volume and severe elevation of left ventricular end-diastolic pressure, never permits underfilling of the left ventricle, despite cardiac tamponade and an inspiratory decrease in intrathoracic pressure.
• Intravascular volume depletion due to hemorrhage, hemodialysis, or mistaken use of diuretics to treat edema can cause marked hypotension, making pulsus paradoxus impossible to detect.

Knowledge of underlying medical conditions, the likelihood of their causing cardiac tamponade, and the appearance of the echocardiogram prompt the physician to look further when the presence or absence of pulsus paradoxus does not fit with the working diagnosis.

The echocardiogram can give hints to the etiology of a pericardial effusion, such as clotted blood after trauma or a cardiac-perforating procedure, tumor studding of the epicardium,⁹ or fibrin strands indicating chronicity or an inflammatory process.¹⁰ Diastolic collapse of the right ventricle, more than collapse of the right atrium or left atrium, speaks for the severity of cardiac tamponade. With hemodynamically significant pericardial effusion and cardiac tamponade, the inferior vena cava is distended and does not decrease in size with inspiration unless there is severe intravascular volume depletion, at which time the inferior vena cava is underfilled throughout the respiratory cycle.

PEARL 10: PLAN HOW TO DRAIN

The size and location of the pericardial effusion and the patient’s hemodynamics must be integrated when deciding how to relieve cardiac tamponade. When cardiac tamponade is indeed severe and the patient and physician agree that it must be drained, the options are percutaneous needle aspiration (pericardiocentesis) and surgical pericardiostomy (creation of a pericardial window). Here again, as assessed by echocardiography, the access to the pericardial fluid should influence the choice.

Pericardiocentesis can be safely done if certain criteria are met. The patient must be able to lie still in the supine position, perhaps with the head of the bed elevated 30 degrees. Anticoagulation must be reversed or allowed time to resolve if drainage is not an emergency.

Pericardiocentesis can be risky or unsuccessful if there is not enough pericardial fluid to permit respiratory cardiac motion without perforating the heart with the needle; if the effusion is loculated (confined to a pocket) posteriorly; or if it is too far from the skin to permit precise control and placement of a spinal needle into the pericardial space. In cases of cardiac tamponade in which the anatomy indicates surgical pericardiostomy but severe hypotension prevents the induction of anesthesia and positive-pressure ventilation—which can result in profound, irreversible hypotension—percutaneous needle drainage (pericardiocentesis) should be performed in the operating room to relieve the tamponade before the induction of anesthesia and the surgical drainage.¹¹

To reiterate, a suspected cardiac or aortic rupture that causes cardiac tamponade is usually large and not apt to self-seal. In such cases, the halt in the accumulation of pericardial blood is due to hypotension and not due to spontaneous resolution. Open surgical drainage is required from the outset because an initial success of pericardiocentesis yields to the recurrence of cardiac tamponade.

PEARL 11: ANTICIPATE WHAT THE FLUID SHOULD LOOK LIKE

Before performing pericardiocentesis, anticipate the appearance of the pericardial fluid on the basis of the presumed etiology, ie:

• Sanguinous—trauma, heart surgery, cardiac perforation from a procedure, anticoagulation, uremia, or malignancy
• Serous—congestive heart failure, acute radiation therapy
• Purulent—infections (natural or postoperative)
• Turbid (like gold paint)—mycobacterial infection, rheumatoid arthritis, myxedema
• Chylous—pericardium fistulized to the thoracic duct by a natural or postsurgical cause.

Sanguinous pericardial effusion encountered during a pericardiocentesis, if not anticipated, can be daunting and can cause the operator to question if it is the result of inadvertent needle placement in a cardiac chamber. If the needle is indeed in the heart, blood often surges out under pressure in pulses, which strongly suggests that the needle is not in the pericardial space and should be removed; but if confirmation of the location is needed before removing the needle, it can be done by injecting 2 mL of agitated sterile saline through the pericardiocentesis needle during echocardiographic imaging.¹²

Before inserting the needle, the ideal access location and needle angle must be determined by the operator with echocardiographic transducer in hand. The distance from skin to a point just through the parietal pericardium can also be measured at this time.

Once the needle is in the pericardial fluid (and you are confident of its placement), removal of 50 to 100 mL of the fluid with a large syringe can be enough to afford the patient easier breathing, higher blood pressure, and lower pulsus paradoxus—and even the physician will breathe easier. The same syringe can be filled and emptied multiple times. Less traumatic and more complete removal of pericardial fluid requires insertion of a multihole pigtail catheter over a J-tipped guidewire that is introduced through the needle.

PEARL 12: DRAIN SLOWLY TO AVOID PULMONARY EDEMA

Pulmonary edema is an uncommon complication of pericardiocentesis that might be avoidable. Heralded by sudden coughing and pink, frothy sputum, it can rapidly deteriorate into respiratory failure. The mechanism has been attributed to a sudden increase in right ventricular stroke volume and resultant left ventricular filling after the excess pericardial fluid has been removed, before the systemic arteries, which constrict to keep the systemic blood pressure up during cardiac tamponade, have had time to relax.¹³

To avoid this complication, if the volume of pericardial fluid responsible for cardiac tamponade is large, it should be removed slowly,¹⁴ stopping for a several-minute rest after each 250 mL. Catheter removal of pericardial fluid by gravity drainage over 24 hours has been suggested.¹⁵ A drawback to this approach is catheter clotting or sludging before all the fluid has been removed. It is helpful to keep the drainage catheter close to the patient’s body temperature to make the fluid less viscous. Output should be monitored hourly.

When the pericardial fluid has been completely drained, one must decide how long to leave the catheter in. One reason to remove the catheter at this time is that it causes pleuritic pain; another is to avoid introducing infection. A reason to leave the catheter in is to observe the effect of medical treatment on the hourly pericardial fluid output. Nonsteroidal anti-inflammatory drugs are the drugs of first choice when treating pericardial inflammation and suppressing production of pericardial fluid.¹⁶ In most cases the catheter should not be left in place for more than 3 days.

Laboratory analysis of the pericardial fluid should shed light on its suspected cause. Analysis usually involves chemistry testing, microscopic inspection of blood cell smears, cytology, microbiologic stains and cultures, and immunologic tests. Results often take days. Meyers and colleagues¹⁷ expound on this subject.

If you had a serious disease, would you agree to an alternative treatment that was cheap, safe, and effective—but seemed disgusting? Would you recommend it to patients?

Such a disease is recurrent Clostridium difficile infection, and such a treatment is fecal microbiota transplantation—instillation of blenderized feces from a healthy donor (ideally, the patient’s spouse or “significant other”) into the patient’s colon to restore a healthy population of bacteria.^1,2 The rationale behind this procedure is simple: antibiotics and other factors disrupt the normal balance of the colonic flora, allowing C difficile to proliferate, but the imbalance can be corrected by reintroducing the normal flora.¹

In this article, we will review how recurrent C difficile infection occurs and the importance of the gut microbiota in resisting colonization with this pathogen. We will also describe the protocol used for fecal microbiota transplantation.

C DIFFICILE INFECTION OFTEN RECURS

C difficile is the most common cause of hospital-acquired diarrhea and an important cause of morbidity and death in hospitalized patients.^3,4 The cost of this infection is estimated to be more than $1.1 billion per year and its incidence is rising, partly because of the emergence of more-virulent strains that make treatment of recurrent infection more difficult.^5,6

C difficile infection is characterized by diarrhea associated with findings suggestive of pseudomembranous colitis or, in fulminant cases, ileus or megacolon.⁷ Recurrent C difficile infection is defined as the return of symptoms within 8 weeks after successful treatment.⁷

C difficile produces two types of toxins. Toxin A is an enterotoxin, causing increased intestinal permeability and fluid secretion, while toxin B is a cytotoxin, causing intense colonic inflammation. People who have a poor host immune response to these toxins tend to develop more diarrhea and colonic inflammation.⁸

A more virulent strain of C difficile has emerged. Known as BI/NAP1/027, this strain is resistant to quinolones, and it also produces a binary toxin that has a partial gene deletion that allows for increased production of toxins A and B in vitro.^9,10 More cases of severe and recurrent C difficile infection have been associated with the increasing number of people infected with this hypervirulent strain.^9,10

C difficile infection recurs in about 20% to 30% of cases after antibiotic treatment for it, usually within 30 days, and the risk of a subsequent episode doubles after two or more occurrences.^10,11 Metronidazole (Flagyl) and vancomycin are the primary treatments; alternative treatments include fidaxomicin (Dificid), ¹⁰ rifaximin (Xifaxan),¹² nitazoxanide,¹³ and tolevamer (a novel polymer that binds C difficile toxins).¹⁴

Table 1 summarizes the treatment regimen for C difficile infection in adults, based on clinical practice guidelines from the US Centers for Disease Control and Prevention (CDC).⁷

THE NORMAL GUT MICROBIOTA KEEPS PATHOGENS OUT

Immediately after birth, the sterile human gut becomes colonized by a diverse community of microorganisms.¹⁵ This gut microbiota performs various functions, such as synthesizing vitamin K and vitamin B complex, helping digest food, maintaining the mucosal integrity of the gut, and priming the mucosal immune response to maintain homeostasis of commensal microbiota.¹⁶

However, the most important role of the gut microbiota is “colonization resistance” or preventing exogenous or potentially pathogenic organisms from establishing a colony within the gut.¹⁷ It involves competition for nutrients and occupation of binding sites on the gut epithelium by indigenous flora.¹⁶ Other factors such as the mucosal barrier, salivation, swallowing, gastric acidity, desquamation of mucosal membrane cells, intestinal motility, and secretion of antibodies also play major roles in colonization resistance.¹⁷

ANTIBIOTICS DISRUPT THE GUT FLORA

Physical or chemical injuries (the latter by antimicrobial or antineoplastic agents, eg) may disrupt the gut microbiota. In this situation, opportunistic pathogens such as C difficile colonize the gut mucosa, stimulate an immune reaction, and release toxins that cause diarrhea and inflammation.¹⁸C difficile will try to compete for nutrients and adhesion sites until it dominates the intestinal tract.

When C difficile spores are ingested, they replicate in the gut and eventually release toxins. Antibiotic therapy may eliminate C difficile bacteria but not the spores; hence, C difficile infection can recur after the antibiotic is discontinued unless the indigenous bacteria can restrain C difficile from spreading.¹⁹

HOW DOES FECAL MICROBIOTA TRANSPLANTATION WORK?

Fecal microbiota transplantation involves instilling processed stool that contains essential intestinal bacteria (eg, Bacteroides species) from a healthy screened donor into the diseased gastrointestinal tract of a suitable recipient (Figure 1).¹

The aim of this procedure is to reestablish the normal composition of the gut flora, restore balance in metabolism, and stimulate both the acquired and the humoral immune responses in the intestinal mucosa after disruption of the normal flora.20–23 One study showed that patients who have recurrent C difficile infections have fewer protective microorganisms (ie, Firmicutes and Bacteriodetes) in their gut, but after fecal microbiota transplantation their microbiota was found to be similar to that of the donor, and their symptoms promptly resolved.18

STUDIES UP TO NOW

The principle of transplanting donor stool to treat various gastrointestinal diseases has been practiced in veterinary medicine for decades in a process known as transfaunation.²⁴ Fecal microbiota transplantation was first performed in humans in the late 1950s in patients with fulminant pseudomembranous colitis that did not respond to standard antibiotic therapy for C difficile infection.²⁵ Since then, a number of case reports and case series have described instillation of donor stool via nasogastric tube,²⁶ via colonoscope,^27–31 and via enema.³² Regardless of the protocols used, disease resolution has been shown in 92% of cases and few adverse effects have been reported, even though transmission of infectious pathogens is theoretically possible.³³

A recent multicenter long-term follow-up study³⁴ showed that diarrhea resolved within 90 days after fecal microbiota transplantation in 70 (91%) of 77 patients, while resolution of C difficile infection after a further course of antibiotics with or without repeating fecal microbiota transplantation was seen in 76 (98%) of 77 patients.³⁴ Some patients were reported to have improvement of preexisting allergies, and a few patients developed peripheral neuropathy and autoimmune diseases such as Sjögren syndrome, idiopathic thrombocytopenic purpura, and rheumatoid arthritis.³³

As the important role of the gut microbiota in resisting colonization by C difficile is becoming more recognized, scientists are beginning to understand and explore the additional potential benefits of fecal microbiota transplantation on other microbiotarelated dysfunctions.² The Human Microbiome Project is focusing on characterizing and understanding the role of the microbial components of the human genetic and metabolic landscape in relation to human health and disease.³⁵ Earlier observational studies showed fecal microbiota transplantation to be beneficial in inflammatory bowel disease, ^36,37 irritable bowel syndrome,^38,39 multiple sclerosis,⁴⁰ rheumatologic⁴⁰ and autoimmune diseases,⁴¹ and metabolic syndrome,⁴² likely owing to the role of the microbiota in immunity and energy metabolism. Although these reports may provide insight into the unexplored possibilities of fecal microbiota transplantation, further clinical investigations with randomized controlled trials are still necessary.

THE CURRENT PROTOCOL FOR FECAL MICROBIOTA TRANSPLANTATION

As yet, there is no standardized protocol for fecal microbiota transplantation, since no completed randomized trial supporting its efficacy and safety has been published. However, a group of experts in infectious disease and gastroenterology have published a formal standard practice guideline,¹⁹ as summarized below.

Primary indications for fecal microbiota transplantation

• Recurrent C difficile infection—at least three episodes of mild to moderate C difficile infection and failure of a 6- to 8-week taper with vancomycin with or without an alternative antibiotic such as rifaximin or nitazoxanide, or at least two episodes of severe C difficile infection resulting in hospitalization and associated with significant morbidity
• Mild to moderate C difficile infection not responding to standard therapy for at least 1 week
• Severe or fulminant C difficile colitis that has not responded to standard therapy after 48 hours.

Who is a likely donor?

The gut microbiota is continuously replenished with bacteria from the environment in which we live, and we constantly acquire organisms from people who live in that same environment. Hence, the preferred donor is someone who has intimate physical contact with the recipient.^33,43,44 The preferred stool donor (in order of preference) is a spouse or significant partner, a family household member, or any other healthy donor.^26,36

Who should not be a donor?

It is the responsibility of the physician performing the fecal microbiota transplantation to make sure that the possibility of transmitting disease to the recipient is minimized. Extensive history-taking and physical examination must never be omitted, since not all diseases or conditions can be detected by laboratory screening alone, especially if testing was done during the early stage or window period of a given disease.¹⁹ Nevertheless, the donor’s blood and stool should be screened for transmissible diseases such as human immunodeficiency virus (HIV), hepatitis, syphilis, enteric bacteria, parasites, and C difficile.

The recipient has the option to be tested for transmissible diseases such as HIV and hepatitis in order to avoid future questions about transmission after fecal microbiota transplantation. A positive screening test must always be verified with confirmatory testing.¹⁹

Table 2 summarizes the exclusion criteria and screening tests performed for donors according to the practice guidelines for fecal microbiota transplantation formulated by Bakken et al.¹⁹

Preprocedure instructions and stool preparation

The physician should orient both the donor and recipient regarding “do’s and don’ts” before fecal microbiota transplantation. Table 3 summarizes the preprocedure instructions and steps for stool preparation.

Route of administration

The route of administration may vary depending on the clinical situation. Upper-gastrointestinal administration is performed via nasogastric or nasojejunal tube or gastroscopy. Lower-gastrointestinal administration is performed via colonoscopy (the route of choice) or retention enema.

The upper-gastrointestinal route (nasogastric tube, jejunal catheter, or gastroscope). The nasogastric or nasojejunal tube or gastroscope is inserted into the upper-gastrointestinal tract, and positioning is confirmed by radiography. From 25 to 50 mL of stool suspension is drawn up in a syringe and instilled into the tubing followed by flushing with 25 mL of normal saline.²⁶ Immediately after instillation, the tube is removed and the patient is allowed to go home and continue with his or her usual diet.

This approach is easier to perform, costs less, and poses lower risk of intestinal perforation than the colonoscopic approach. Disadvantages include the possibility that stool suspension may not reach distal areas of the colon, especially in patients with ileus and small-bowel obstruction. There is also a higher risk of bacterial overgrowth in elderly patients who have lower gastric acid levels.³³

The lower-gastrointestinal route (colonoscopy, retention enema). Colonoscopy is currently considered the first-line approach for fecal microbiota transplantation.⁴⁵ After giving informed consent, the patient undergoes standard colonoscopy under sedation. An initial colonoscopic examination is performed, and biopsy specimans are obtained if necessary. Approximately 20 mL of stool suspension is drawn up in a syringe and injected via the biopsy channel of the colonoscope every 5 to 10 cm as the scope is withdrawn, for a total volume of 250 to 500 mL.^19,27 The patient should be advised to refrain from defecating for 30 to 45 minutes after fecal microbiota transplantation.⁴⁶

This approach allows direct visualization of the entire colon, allowing instillation of stool suspension in certain areas where C difficile may predominate or hide (eg, in diverticuli).^27,47 One disadvantage to this route of administration is the risk of colon perforation, especially if the patient has toxic colitis.

Instillation via retention enema may be done at home with a standard enema kit.³² Disadvantages include the need for multiple instillations over 3 to 5 days,³⁶ back-leakage of stool suspension causing discomfort to patients, and stool suspension reaching only to the splenic flexure.⁴⁸

MEASUREMENT OF OUTCOME

Fecal microbiota transplantation is considered successful if symptoms resolve and there is no relapse within 8 weeks. Testing for C difficile in asymptomatic patients is not recommended since patients can be colonized with C difficile without necessarily developing disease.¹⁹ There is currently no consensus on treatment recommendations for patients who do not respond to fecal microbiota transplantation, although some reports showed resolution of diarrhea after a repeat 2-week standard course of oral vancomycin²⁶ or repeated instillation of feces collected from new donors.⁴⁹

IS IT READY FOR PRIME TIME?

Fecal microbiota transplantation has been used primarily as an alternative treatment for recurrent C difficile infection, although other indications for its use are currently being identified and studied. This procedure is now being done in several specialized centers in the United States and abroad, and although the protocol may vary by institution, the clinical outcomes have been consistently promising.

The Fecal Therapy to Eliminate Associated Long-standing Diarrhea (FECAL) trial, currently underway, is the first randomized trial to assess the efficacy of fecal microbiota transplantation for treatment of recurrent C difficile infection.⁵⁰ Clinical trials such as this one should satisfy our doubts about the efficacy of fecal microbiota transplantation and hopefully pave the way for its application in the near future.

An increasing number of patients are learning to overcome the “yuck factor” associated with fecal microbiota transplantation once they understand its safety and benefits.⁵¹ Moreover, the Human Microbiome Project is attempting to identify specific organisms in stool that may specifically treat C difficile infection, hence eliminating the need for whole-stool transplantation in the near future. Although fecal microbiota transplantation is still in its infancy, its low cost, safety, and effectiveness in treating recurrent C difficile infection will likely lead to the procedure becoming widely adopted in mainstream clinical practice.

Editor’s note: On January 16, 2013, after this article was completed, a randomized controlled trial of fecal microbiota transplantation was published in the New England Journal of Medicine. That trial, “Duodenal infusion of donor feces for recurrent Clostridium difficile,” found: “The infusion of donor feces was significantly more effective for the treatment of recurrent C difficile infection than the use of vancomycin.” The study is available online at http://www.nejm.org/doi/full/10.1056/NEJMoa1205037 (subscription required).