Cleveland Clinic Journal of Medicine

Invented by Andreas Grüntzig in 1977, percutaneous coronary intervention (PCI) has revolutionized the management of coronary artery disease.¹ Initially, PCI was more attractive than conventional revascularization with coronary artery bypass grafting because it was less invasive, but as time went on PCI acquired its own evidence base of improved clinical outcomes. In fact, for ST-elevation myocardial infarction, non-ST-elevation myocardial infarction, and cardiogenic shock, there is clear evidence that PCI saves lives in both the short and long term.^2,3

See related article

But PCI is also used widely in stable coronary artery disease, and in contrast to the clear-cut benefit in the acute conditions noted above, a series of reports culminating in the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial has shown that PCI in stable coronary artery disease does not reduce the risk of death or of subsequent myocardial infarction.^4,5 Cardiologists have heeded the COURAGE trial findings in their clinical decision-making, and the rate of PCI for stable coronary artery disease dropped by 60% from 2006 to 2011.⁶

In an article in this issue,⁷ Dr. Michael Rothberg describes a 55-year-old man who develops new-onset angina and then undergoes a Bruce protocol stress test that is stopped at 6 minutes due to chest pain and ST-segment depression. Dr. Rothberg argues that, based on COURAGE trial data, this patient and other patients with stable coronary artery disease should not be treated with PCI but instead should receive optimal medical therapy.

KEY ISSUES ABOUT THE COURAGE TRIAL

To understand the applicability of the results of the COURAGE trial to patient care, it is important to examine a number of key issues about this trial.

First, COURAGE enrolled a narrow group of patients with stable coronary artery disease and excluded many common patient subgroups, such as those with heart failure, severe anginal symptoms, or left main artery stenosis, who would benefit from revascularization.⁵ Specifically, for every 100 patients enrolled in COURAGE, 161 were excluded for having heart failure, 39 were excluded for class IV angina, and 31 were excluded for left main stenosis.

Second, although COURAGE has been described as a trial of PCI vs optimal medical therapy, it was not. Rather, it was a trial of optimal medical therapy with PCI first vs optimal medical therapy with crossover PCI if medical therapy failed.⁵ The crossover rate was not insubstantial: 16.1% of the patients in the medical therapy group underwent PCI by the end of the first year, increasing to 32.6% at a median of 4.6 years of follow-up.^5,7 And patients with more intense and frequent angina and resulting worse quality of life were the ones who required crossover PCI.⁸

Third, it has been proposed that patients with suspicious cardiac symptoms or abnormal stress test findings can be managed with optimal medical therapy initially, based on the COURAGE findings. However, the COURAGE trial required diagnostic angiography both to confirm underlying coronary artery disease and to exclude left main disease.⁵ Thus, regardless of one’s position on the role of PCI in stable coronary artery disease, diagnostic investigation by cardiac catheterization or computed tomographic angiography to confirm the presence or absence of coronary artery disease remains mandatory.

Fourth, optimal medical therapy was prescribed by the trial’s protocol, so that one would expect that both treatment groups received similar levels of optimal medical therapy. However, the optimal medical therapy group required more medications to achieve the same outcome as the PCI group.⁵

Finally, although it has been reported that the COURAGE trial showed no benefit for PCI, in fact, for the outcome of symptom relief, initial PCI was clearly superior to optimal medical therapy beginning at 3 months and extending out to 24 months—a result for which the magnitude of benefit is underestimated due to the occurrence of crossover PCI.⁹ In particular, women and patients with a high frequency of angina derived improvement in angina-related quality of life from PCI compared with optimal medical therapy.^8,10

For these reasons, the role of PCI in stable coronary disease is more nuanced than simply stating that the COURAGE trial results were “negative” for PCI. It is more accurate to say that in selected patients with moderate symptoms of angina and without heart failure or left main artery disease, a PCI-first strategy has no advantage over an optimal medical treatment-first strategy for the risk of death and myocardial infarction but does lead to earlier angina relief and less long-term need for medication. In addition, in up to one-third of cases, an optimal medical treatment-first strategy fails and requires crossover to PCI.⁵

Dr. Rothberg is correct in highlighting the crucial importance of optimal medical therapy in the management of stable coronary artery disease. In fact, cardiologists strive to prescribe optimal medical treatment for all coronary artery disease patients irrespective of treatment strategy. However, 3 important issues in his analysis need to be highlighted.

Controlling symptoms is important, and we should not underrate it. The patient described in Dr. Rothberg’s article could exercise for only 6 minutes on a Bruce treadmill test, indicating a quite limited functional capacity of only 5.8 metabolic equivalents of the task (METs).¹¹ (A healthy 55-year-old man should be able to achieve 10.5 METs.¹²) Inability to achieve 6 METs precludes the ability to dance, to ride a bike at a moderate pace, or to go on a hike.¹³ For many patients, these limitations are serious and important concerns for their lifestyle and quality of life. PCI has been shown to be superior to medical therapy in improving functional capacity, improving it by 20% vs 2% in one trial¹⁴ and 26% vs 7% in a second trial.¹⁴ Patients undergoing PCI were twice as likely to have a greater than 2-minute increase in exercise capacity.¹⁵ Recognizing the importance of symptom control in stable coronary artery disease is patient-centered care.

Patient decision-making is complicated, and we should not assume that patients choose PCI primarily to reduce their risk of death. A randomized trial showed that patients continued to select PCI as initial treatment even when they clearly knew that it would not prevent death or myocardial infarction.¹⁶ As noted above, patients may value earlier symptom relief, particularly if their angina is frequent or limiting. In addition, patients strongly desire to minimize medical therapy and may be willing to trade decreased life expectancy to reduce the need to take medications.¹⁷ Finally, some patients may want to be able to continue to participate in certain lifestyle activities.

PCI is expensive, but less so over the long run. With a PCI-first strategy, costs are front-loaded, and studies with short-term follow-up show a marked increase in cost. However, long-term follow-up shows that the cost differences diminish dramatically due to high rates of crossover to revascularization and increased medical care in the optimal medical therapy arm. The cumulative lifetime costs in the COURAGE trial with a PCI-first strategy, although statistically significant, were only 10% higher than with the optimal medical treatment-first strategy ($99,820 vs $90,370).¹⁸ Therefore, substantial long-term cost-savings by shifting from an initial PCI strategy to initial optimal medical therapy are unlikely to be delivered when measured over the long term.

NEWER TRIALS SUPPORT A BALANCED APPROACH

The most recent studies of the management of stable coronary artery disease support a balanced approach.

The ORBITA trial (Objective Randomised Blinded Investigation With Optimal Medical Therapy of Angioplasty in Stable Angina), on one hand, showed limited benefit of PCI vs medical therapy in patients with single-vessel coronary artery disease, preserved functional capacity, and mild symptoms.¹⁹ There was no significant improvement in exercise capacity or angina frequency, although baseline angina frequency after medical stabilization was quite low.

The FAME 2 trial (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation), on the other hand, studied patients with positive fractional flow reserve coronary artery disease (ie, using an invasive technique to confirm the hemodynamic significance of the coronary stenosis) and showed markedly better outcomes with PCI than with medical therapy.²⁰ Specifically, the PCI-first group had improved quality of life and dramatically less need for urgent revascularization.

Furthermore, as in the COURAGE trial, the optimal medical therapy group had a high crossover rate to PCI (44.2%), leading to the complete elimination of the early cost advantage of medical therapy by 3 years. The initial costs with PCI vs medical therapy were $9,944 vs $4,439 (P < .001); the 3-year costs were $16,792 vs $16,737 (P = .94).

For these reasons, a balanced approach to recommending PCI first vs optimal medical treatment first remains the best strategy.

TOWARD PATIENT-CENTERED CARE

For the 55-year-old patient in Dr. Rothberg’s article, the first step in making an appropriate decision would be to understand the severity of symptoms relative to the patient’s lifestyle. The second step is to assess the patient’s interest in an invasive procedure such as PCI relative to optimal medical therapy, as the patient may have a strong preference for one option or the other.

Finally, with the understanding that there is no difference in hard end points of myocardial infarction and death, a balanced discussion of the advantages and disadvantages of both PCI and optimal medical therapy would be needed. For PCI, advantages include earlier symptom control and improved quality of the life, particularly if symptoms are severe, with disadvantages of an invasive procedure with its attendant risks. For optimal medical therapy, advantages include improved symptom control and avoidance of an invasive procedure, while disadvantages include increased medication use and a high rate of eventual crossover to PCI. This important discussion integrating both patient and medical perspectives ultimately leads to the best decision for the individual patient.

A patient-centered approach to clinical decision-making mandates inclusion of PCI first as an option in the management of stable coronary artery disease. After confirming the patient has coronary artery disease, patients with heart failure, class IV angina at rest, or left main artery stenosis should be referred for revascularization. In the remaining patients with confirmed coronary artery disease and moderate angina symptoms, either PCI first or optimal medical therapy first is an appropriate initial strategy that considers coronary anatomy, symptom burden, and patient desires.

Invented by Andreas Grüntzig in 1977, percutaneous coronary intervention (PCI) has revolutionized the management of coronary artery disease.¹ Initially, PCI was more attractive than conventional revascularization with coronary artery bypass grafting because it was less invasive, but as time went on PCI acquired its own evidence base of improved clinical outcomes. In fact, for ST-elevation myocardial infarction, non-ST-elevation myocardial infarction, and cardiogenic shock, there is clear evidence that PCI saves lives in both the short and long term.^2,3

See related article

But PCI is also used widely in stable coronary artery disease, and in contrast to the clear-cut benefit in the acute conditions noted above, a series of reports culminating in the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial has shown that PCI in stable coronary artery disease does not reduce the risk of death or of subsequent myocardial infarction.^4,5 Cardiologists have heeded the COURAGE trial findings in their clinical decision-making, and the rate of PCI for stable coronary artery disease dropped by 60% from 2006 to 2011.⁶

In an article in this issue,⁷ Dr. Michael Rothberg describes a 55-year-old man who develops new-onset angina and then undergoes a Bruce protocol stress test that is stopped at 6 minutes due to chest pain and ST-segment depression. Dr. Rothberg argues that, based on COURAGE trial data, this patient and other patients with stable coronary artery disease should not be treated with PCI but instead should receive optimal medical therapy.

KEY ISSUES ABOUT THE COURAGE TRIAL

To understand the applicability of the results of the COURAGE trial to patient care, it is important to examine a number of key issues about this trial.

First, COURAGE enrolled a narrow group of patients with stable coronary artery disease and excluded many common patient subgroups, such as those with heart failure, severe anginal symptoms, or left main artery stenosis, who would benefit from revascularization.⁵ Specifically, for every 100 patients enrolled in COURAGE, 161 were excluded for having heart failure, 39 were excluded for class IV angina, and 31 were excluded for left main stenosis.

Second, although COURAGE has been described as a trial of PCI vs optimal medical therapy, it was not. Rather, it was a trial of optimal medical therapy with PCI first vs optimal medical therapy with crossover PCI if medical therapy failed.⁵ The crossover rate was not insubstantial: 16.1% of the patients in the medical therapy group underwent PCI by the end of the first year, increasing to 32.6% at a median of 4.6 years of follow-up.^5,7 And patients with more intense and frequent angina and resulting worse quality of life were the ones who required crossover PCI.⁸

Third, it has been proposed that patients with suspicious cardiac symptoms or abnormal stress test findings can be managed with optimal medical therapy initially, based on the COURAGE findings. However, the COURAGE trial required diagnostic angiography both to confirm underlying coronary artery disease and to exclude left main disease.⁵ Thus, regardless of one’s position on the role of PCI in stable coronary artery disease, diagnostic investigation by cardiac catheterization or computed tomographic angiography to confirm the presence or absence of coronary artery disease remains mandatory.

Fourth, optimal medical therapy was prescribed by the trial’s protocol, so that one would expect that both treatment groups received similar levels of optimal medical therapy. However, the optimal medical therapy group required more medications to achieve the same outcome as the PCI group.⁵

Finally, although it has been reported that the COURAGE trial showed no benefit for PCI, in fact, for the outcome of symptom relief, initial PCI was clearly superior to optimal medical therapy beginning at 3 months and extending out to 24 months—a result for which the magnitude of benefit is underestimated due to the occurrence of crossover PCI.⁹ In particular, women and patients with a high frequency of angina derived improvement in angina-related quality of life from PCI compared with optimal medical therapy.^8,10

For these reasons, the role of PCI in stable coronary disease is more nuanced than simply stating that the COURAGE trial results were “negative” for PCI. It is more accurate to say that in selected patients with moderate symptoms of angina and without heart failure or left main artery disease, a PCI-first strategy has no advantage over an optimal medical treatment-first strategy for the risk of death and myocardial infarction but does lead to earlier angina relief and less long-term need for medication. In addition, in up to one-third of cases, an optimal medical treatment-first strategy fails and requires crossover to PCI.⁵

Dr. Rothberg is correct in highlighting the crucial importance of optimal medical therapy in the management of stable coronary artery disease. In fact, cardiologists strive to prescribe optimal medical treatment for all coronary artery disease patients irrespective of treatment strategy. However, 3 important issues in his analysis need to be highlighted.

Controlling symptoms is important, and we should not underrate it. The patient described in Dr. Rothberg’s article could exercise for only 6 minutes on a Bruce treadmill test, indicating a quite limited functional capacity of only 5.8 metabolic equivalents of the task (METs).¹¹ (A healthy 55-year-old man should be able to achieve 10.5 METs.¹²) Inability to achieve 6 METs precludes the ability to dance, to ride a bike at a moderate pace, or to go on a hike.¹³ For many patients, these limitations are serious and important concerns for their lifestyle and quality of life. PCI has been shown to be superior to medical therapy in improving functional capacity, improving it by 20% vs 2% in one trial¹⁴ and 26% vs 7% in a second trial.¹⁴ Patients undergoing PCI were twice as likely to have a greater than 2-minute increase in exercise capacity.¹⁵ Recognizing the importance of symptom control in stable coronary artery disease is patient-centered care.

Patient decision-making is complicated, and we should not assume that patients choose PCI primarily to reduce their risk of death. A randomized trial showed that patients continued to select PCI as initial treatment even when they clearly knew that it would not prevent death or myocardial infarction.¹⁶ As noted above, patients may value earlier symptom relief, particularly if their angina is frequent or limiting. In addition, patients strongly desire to minimize medical therapy and may be willing to trade decreased life expectancy to reduce the need to take medications.¹⁷ Finally, some patients may want to be able to continue to participate in certain lifestyle activities.

PCI is expensive, but less so over the long run. With a PCI-first strategy, costs are front-loaded, and studies with short-term follow-up show a marked increase in cost. However, long-term follow-up shows that the cost differences diminish dramatically due to high rates of crossover to revascularization and increased medical care in the optimal medical therapy arm. The cumulative lifetime costs in the COURAGE trial with a PCI-first strategy, although statistically significant, were only 10% higher than with the optimal medical treatment-first strategy ($99,820 vs $90,370).¹⁸ Therefore, substantial long-term cost-savings by shifting from an initial PCI strategy to initial optimal medical therapy are unlikely to be delivered when measured over the long term.

NEWER TRIALS SUPPORT A BALANCED APPROACH

The most recent studies of the management of stable coronary artery disease support a balanced approach.

The ORBITA trial (Objective Randomised Blinded Investigation With Optimal Medical Therapy of Angioplasty in Stable Angina), on one hand, showed limited benefit of PCI vs medical therapy in patients with single-vessel coronary artery disease, preserved functional capacity, and mild symptoms.¹⁹ There was no significant improvement in exercise capacity or angina frequency, although baseline angina frequency after medical stabilization was quite low.

The FAME 2 trial (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation), on the other hand, studied patients with positive fractional flow reserve coronary artery disease (ie, using an invasive technique to confirm the hemodynamic significance of the coronary stenosis) and showed markedly better outcomes with PCI than with medical therapy.²⁰ Specifically, the PCI-first group had improved quality of life and dramatically less need for urgent revascularization.

Furthermore, as in the COURAGE trial, the optimal medical therapy group had a high crossover rate to PCI (44.2%), leading to the complete elimination of the early cost advantage of medical therapy by 3 years. The initial costs with PCI vs medical therapy were $9,944 vs $4,439 (P < .001); the 3-year costs were $16,792 vs $16,737 (P = .94).

For these reasons, a balanced approach to recommending PCI first vs optimal medical treatment first remains the best strategy.

TOWARD PATIENT-CENTERED CARE

For the 55-year-old patient in Dr. Rothberg’s article, the first step in making an appropriate decision would be to understand the severity of symptoms relative to the patient’s lifestyle. The second step is to assess the patient’s interest in an invasive procedure such as PCI relative to optimal medical therapy, as the patient may have a strong preference for one option or the other.

Finally, with the understanding that there is no difference in hard end points of myocardial infarction and death, a balanced discussion of the advantages and disadvantages of both PCI and optimal medical therapy would be needed. For PCI, advantages include earlier symptom control and improved quality of the life, particularly if symptoms are severe, with disadvantages of an invasive procedure with its attendant risks. For optimal medical therapy, advantages include improved symptom control and avoidance of an invasive procedure, while disadvantages include increased medication use and a high rate of eventual crossover to PCI. This important discussion integrating both patient and medical perspectives ultimately leads to the best decision for the individual patient.

A patient-centered approach to clinical decision-making mandates inclusion of PCI first as an option in the management of stable coronary artery disease. After confirming the patient has coronary artery disease, patients with heart failure, class IV angina at rest, or left main artery stenosis should be referred for revascularization. In the remaining patients with confirmed coronary artery disease and moderate angina symptoms, either PCI first or optimal medical therapy first is an appropriate initial strategy that considers coronary anatomy, symptom burden, and patient desires.

Invented by Andreas Grüntzig in 1977, percutaneous coronary intervention (PCI) has revolutionized the management of coronary artery disease.¹ Initially, PCI was more attractive than conventional revascularization with coronary artery bypass grafting because it was less invasive, but as time went on PCI acquired its own evidence base of improved clinical outcomes. In fact, for ST-elevation myocardial infarction, non-ST-elevation myocardial infarction, and cardiogenic shock, there is clear evidence that PCI saves lives in both the short and long term.^2,3

See related article

But PCI is also used widely in stable coronary artery disease, and in contrast to the clear-cut benefit in the acute conditions noted above, a series of reports culminating in the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial has shown that PCI in stable coronary artery disease does not reduce the risk of death or of subsequent myocardial infarction.^4,5 Cardiologists have heeded the COURAGE trial findings in their clinical decision-making, and the rate of PCI for stable coronary artery disease dropped by 60% from 2006 to 2011.⁶

In an article in this issue,⁷ Dr. Michael Rothberg describes a 55-year-old man who develops new-onset angina and then undergoes a Bruce protocol stress test that is stopped at 6 minutes due to chest pain and ST-segment depression. Dr. Rothberg argues that, based on COURAGE trial data, this patient and other patients with stable coronary artery disease should not be treated with PCI but instead should receive optimal medical therapy.

KEY ISSUES ABOUT THE COURAGE TRIAL

To understand the applicability of the results of the COURAGE trial to patient care, it is important to examine a number of key issues about this trial.

First, COURAGE enrolled a narrow group of patients with stable coronary artery disease and excluded many common patient subgroups, such as those with heart failure, severe anginal symptoms, or left main artery stenosis, who would benefit from revascularization.⁵ Specifically, for every 100 patients enrolled in COURAGE, 161 were excluded for having heart failure, 39 were excluded for class IV angina, and 31 were excluded for left main stenosis.

Second, although COURAGE has been described as a trial of PCI vs optimal medical therapy, it was not. Rather, it was a trial of optimal medical therapy with PCI first vs optimal medical therapy with crossover PCI if medical therapy failed.⁵ The crossover rate was not insubstantial: 16.1% of the patients in the medical therapy group underwent PCI by the end of the first year, increasing to 32.6% at a median of 4.6 years of follow-up.^5,7 And patients with more intense and frequent angina and resulting worse quality of life were the ones who required crossover PCI.⁸

Third, it has been proposed that patients with suspicious cardiac symptoms or abnormal stress test findings can be managed with optimal medical therapy initially, based on the COURAGE findings. However, the COURAGE trial required diagnostic angiography both to confirm underlying coronary artery disease and to exclude left main disease.⁵ Thus, regardless of one’s position on the role of PCI in stable coronary artery disease, diagnostic investigation by cardiac catheterization or computed tomographic angiography to confirm the presence or absence of coronary artery disease remains mandatory.

Fourth, optimal medical therapy was prescribed by the trial’s protocol, so that one would expect that both treatment groups received similar levels of optimal medical therapy. However, the optimal medical therapy group required more medications to achieve the same outcome as the PCI group.⁵

Finally, although it has been reported that the COURAGE trial showed no benefit for PCI, in fact, for the outcome of symptom relief, initial PCI was clearly superior to optimal medical therapy beginning at 3 months and extending out to 24 months—a result for which the magnitude of benefit is underestimated due to the occurrence of crossover PCI.⁹ In particular, women and patients with a high frequency of angina derived improvement in angina-related quality of life from PCI compared with optimal medical therapy.^8,10

For these reasons, the role of PCI in stable coronary disease is more nuanced than simply stating that the COURAGE trial results were “negative” for PCI. It is more accurate to say that in selected patients with moderate symptoms of angina and without heart failure or left main artery disease, a PCI-first strategy has no advantage over an optimal medical treatment-first strategy for the risk of death and myocardial infarction but does lead to earlier angina relief and less long-term need for medication. In addition, in up to one-third of cases, an optimal medical treatment-first strategy fails and requires crossover to PCI.⁵

Dr. Rothberg is correct in highlighting the crucial importance of optimal medical therapy in the management of stable coronary artery disease. In fact, cardiologists strive to prescribe optimal medical treatment for all coronary artery disease patients irrespective of treatment strategy. However, 3 important issues in his analysis need to be highlighted.

Controlling symptoms is important, and we should not underrate it. The patient described in Dr. Rothberg’s article could exercise for only 6 minutes on a Bruce treadmill test, indicating a quite limited functional capacity of only 5.8 metabolic equivalents of the task (METs).¹¹ (A healthy 55-year-old man should be able to achieve 10.5 METs.¹²) Inability to achieve 6 METs precludes the ability to dance, to ride a bike at a moderate pace, or to go on a hike.¹³ For many patients, these limitations are serious and important concerns for their lifestyle and quality of life. PCI has been shown to be superior to medical therapy in improving functional capacity, improving it by 20% vs 2% in one trial¹⁴ and 26% vs 7% in a second trial.¹⁴ Patients undergoing PCI were twice as likely to have a greater than 2-minute increase in exercise capacity.¹⁵ Recognizing the importance of symptom control in stable coronary artery disease is patient-centered care.

Patient decision-making is complicated, and we should not assume that patients choose PCI primarily to reduce their risk of death. A randomized trial showed that patients continued to select PCI as initial treatment even when they clearly knew that it would not prevent death or myocardial infarction.¹⁶ As noted above, patients may value earlier symptom relief, particularly if their angina is frequent or limiting. In addition, patients strongly desire to minimize medical therapy and may be willing to trade decreased life expectancy to reduce the need to take medications.¹⁷ Finally, some patients may want to be able to continue to participate in certain lifestyle activities.

PCI is expensive, but less so over the long run. With a PCI-first strategy, costs are front-loaded, and studies with short-term follow-up show a marked increase in cost. However, long-term follow-up shows that the cost differences diminish dramatically due to high rates of crossover to revascularization and increased medical care in the optimal medical therapy arm. The cumulative lifetime costs in the COURAGE trial with a PCI-first strategy, although statistically significant, were only 10% higher than with the optimal medical treatment-first strategy ($99,820 vs $90,370).¹⁸ Therefore, substantial long-term cost-savings by shifting from an initial PCI strategy to initial optimal medical therapy are unlikely to be delivered when measured over the long term.

NEWER TRIALS SUPPORT A BALANCED APPROACH

The most recent studies of the management of stable coronary artery disease support a balanced approach.

The ORBITA trial (Objective Randomised Blinded Investigation With Optimal Medical Therapy of Angioplasty in Stable Angina), on one hand, showed limited benefit of PCI vs medical therapy in patients with single-vessel coronary artery disease, preserved functional capacity, and mild symptoms.¹⁹ There was no significant improvement in exercise capacity or angina frequency, although baseline angina frequency after medical stabilization was quite low.

The FAME 2 trial (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation), on the other hand, studied patients with positive fractional flow reserve coronary artery disease (ie, using an invasive technique to confirm the hemodynamic significance of the coronary stenosis) and showed markedly better outcomes with PCI than with medical therapy.²⁰ Specifically, the PCI-first group had improved quality of life and dramatically less need for urgent revascularization.

Furthermore, as in the COURAGE trial, the optimal medical therapy group had a high crossover rate to PCI (44.2%), leading to the complete elimination of the early cost advantage of medical therapy by 3 years. The initial costs with PCI vs medical therapy were $9,944 vs $4,439 (P < .001); the 3-year costs were $16,792 vs $16,737 (P = .94).

For these reasons, a balanced approach to recommending PCI first vs optimal medical treatment first remains the best strategy.

TOWARD PATIENT-CENTERED CARE

For the 55-year-old patient in Dr. Rothberg’s article, the first step in making an appropriate decision would be to understand the severity of symptoms relative to the patient’s lifestyle. The second step is to assess the patient’s interest in an invasive procedure such as PCI relative to optimal medical therapy, as the patient may have a strong preference for one option or the other.

Finally, with the understanding that there is no difference in hard end points of myocardial infarction and death, a balanced discussion of the advantages and disadvantages of both PCI and optimal medical therapy would be needed. For PCI, advantages include earlier symptom control and improved quality of the life, particularly if symptoms are severe, with disadvantages of an invasive procedure with its attendant risks. For optimal medical therapy, advantages include improved symptom control and avoidance of an invasive procedure, while disadvantages include increased medication use and a high rate of eventual crossover to PCI. This important discussion integrating both patient and medical perspectives ultimately leads to the best decision for the individual patient.

A patient-centered approach to clinical decision-making mandates inclusion of PCI first as an option in the management of stable coronary artery disease. After confirming the patient has coronary artery disease, patients with heart failure, class IV angina at rest, or left main artery stenosis should be referred for revascularization. In the remaining patients with confirmed coronary artery disease and moderate angina symptoms, either PCI first or optimal medical therapy first is an appropriate initial strategy that considers coronary anatomy, symptom burden, and patient desires.

It is difficult to provide enough information for informed consent and to ensure that the patient and his or her family fully understand what we are saying. Patients often come in with their own preferences and biases based on anecdote, dinner conversations, or the Internet. The physician must push hard to dispel a patient’s bias with the facts, while recognizing that we too regularly present “facts” and recommendations colored by our own biases based on anecdotal experience, professional ritual, and intellectual hubris.

Two articles in this issue of the Journal, one by Dr. Michael Rothberg¹ and the other by Dr. Umesh Khot,² examine percutaneous coronary intervention (PCI) in patients with stable chronic angina. Both discuss the findings of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial³ and how to use these findings in helping patients decide whether to undergo PCI.

Rothberg and Khot agree that in the COURAGE trial, PCI effectively if not completely reduced angina but did not decrease the likelihood of death or subsequent myocardial infarction (MI). Patients were excluded from the study if they had a likelihood of left main disease, heart failure, or severe angina. All underwent catheterization, and all were given optimal medical therapy. Thus, the trial results do not directly relate to every patient with stable angina.

While the patient may find it confusing that angina and the risk of MI are not reduced in parallel, since both are due to atherosclerosis, their dynamic pathophysiology is different. The COURAGE results are the mirror image of those in some early studies of aspirin in coronary disease, in which aspirin reduced the incidence of MI but did not significantly affect angina.

In view of the COURAGE results, Rothberg seems surprised that PCI continues to be frequently used in patients with stable angina. He points out that according to some surveys,⁴ not all cardiologists have embraced these (and other similar study) results. But as Khot notes, the use of PCI in stable angina has decreased. More interesting to me were the results of an online study conducted by Rothberg and colleagues in which participants were provided different background information about PCI.⁵ Even if given explicit information that PCI did not prevent MI, a fair number still said they would choose it and still believed it would prevent this outcome. Bias clearly influences what patients read and hear, and they bring these biases into the shared decision-making process.

While some patients may not fully understand PCI’s risks and putative benefits, others may choose it because of their personal knowledge of others’ experience or perhaps because the “softer” benefits demonstrated in COURAGE and other trials are important to them. As outlined by Khot, patients who underwent PCI had more rapid relief of angina symptoms, possibly experienced greater relief of symptoms even if incomplete, and needed less medication. More patients needed urgent revascularization in the medical group than in the PCI group. Rothberg appropriately notes that this did not “equate to a reduction in the rate of MI,” but to some patients (eg, international travelers, caregivers) this higher possibility of needing an urgent procedure may be enough to make them want the initial elective procedure. While patients should be told that many of the patients in the medical therapy group in COURAGE crossed over to get PCI (16% at 1 year, and about 1/3 after a median of 4.6 years of follow-up), a patient for whom avoiding invasive procedures is the highest priority will likely “hear” that he or she has a 2/3 likelihood of not needing PCI without being at increased risk of death or MI with medical therapy.

As Rothberg points out, “providing information alone is not enough.” The patient needs to recognize, verbalize, and perhaps rank his or her own biases, fears, and desires. Equally important, we need to recognize our own biases and not let them overshadow the patient’s concerns.

I urge you to read both articles, not only because they offer excellent critiques of the COURAGE results and what they mean in practice, but also because they should make us reflect on how often and well we engage in shared decision-making with our patients. Reading these made me realize that I need to better understand my patients’ concerns. Discussing my interpretation of clinical study results, no matter how sophisticated or correct, and then offering a recommendation without fully understanding the patient’s treatment goals is not shared decision-making. The seemingly “wrong” decision may be right for the patient.

It is difficult to provide enough information for informed consent and to ensure that the patient and his or her family fully understand what we are saying. Patients often come in with their own preferences and biases based on anecdote, dinner conversations, or the Internet. The physician must push hard to dispel a patient’s bias with the facts, while recognizing that we too regularly present “facts” and recommendations colored by our own biases based on anecdotal experience, professional ritual, and intellectual hubris.

Two articles in this issue of the Journal, one by Dr. Michael Rothberg¹ and the other by Dr. Umesh Khot,² examine percutaneous coronary intervention (PCI) in patients with stable chronic angina. Both discuss the findings of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial³ and how to use these findings in helping patients decide whether to undergo PCI.

Rothberg and Khot agree that in the COURAGE trial, PCI effectively if not completely reduced angina but did not decrease the likelihood of death or subsequent myocardial infarction (MI). Patients were excluded from the study if they had a likelihood of left main disease, heart failure, or severe angina. All underwent catheterization, and all were given optimal medical therapy. Thus, the trial results do not directly relate to every patient with stable angina.

While the patient may find it confusing that angina and the risk of MI are not reduced in parallel, since both are due to atherosclerosis, their dynamic pathophysiology is different. The COURAGE results are the mirror image of those in some early studies of aspirin in coronary disease, in which aspirin reduced the incidence of MI but did not significantly affect angina.

In view of the COURAGE results, Rothberg seems surprised that PCI continues to be frequently used in patients with stable angina. He points out that according to some surveys,⁴ not all cardiologists have embraced these (and other similar study) results. But as Khot notes, the use of PCI in stable angina has decreased. More interesting to me were the results of an online study conducted by Rothberg and colleagues in which participants were provided different background information about PCI.⁵ Even if given explicit information that PCI did not prevent MI, a fair number still said they would choose it and still believed it would prevent this outcome. Bias clearly influences what patients read and hear, and they bring these biases into the shared decision-making process.

While some patients may not fully understand PCI’s risks and putative benefits, others may choose it because of their personal knowledge of others’ experience or perhaps because the “softer” benefits demonstrated in COURAGE and other trials are important to them. As outlined by Khot, patients who underwent PCI had more rapid relief of angina symptoms, possibly experienced greater relief of symptoms even if incomplete, and needed less medication. More patients needed urgent revascularization in the medical group than in the PCI group. Rothberg appropriately notes that this did not “equate to a reduction in the rate of MI,” but to some patients (eg, international travelers, caregivers) this higher possibility of needing an urgent procedure may be enough to make them want the initial elective procedure. While patients should be told that many of the patients in the medical therapy group in COURAGE crossed over to get PCI (16% at 1 year, and about 1/3 after a median of 4.6 years of follow-up), a patient for whom avoiding invasive procedures is the highest priority will likely “hear” that he or she has a 2/3 likelihood of not needing PCI without being at increased risk of death or MI with medical therapy.

As Rothberg points out, “providing information alone is not enough.” The patient needs to recognize, verbalize, and perhaps rank his or her own biases, fears, and desires. Equally important, we need to recognize our own biases and not let them overshadow the patient’s concerns.

I urge you to read both articles, not only because they offer excellent critiques of the COURAGE results and what they mean in practice, but also because they should make us reflect on how often and well we engage in shared decision-making with our patients. Reading these made me realize that I need to better understand my patients’ concerns. Discussing my interpretation of clinical study results, no matter how sophisticated or correct, and then offering a recommendation without fully understanding the patient’s treatment goals is not shared decision-making. The seemingly “wrong” decision may be right for the patient.

It is difficult to provide enough information for informed consent and to ensure that the patient and his or her family fully understand what we are saying. Patients often come in with their own preferences and biases based on anecdote, dinner conversations, or the Internet. The physician must push hard to dispel a patient’s bias with the facts, while recognizing that we too regularly present “facts” and recommendations colored by our own biases based on anecdotal experience, professional ritual, and intellectual hubris.

Two articles in this issue of the Journal, one by Dr. Michael Rothberg¹ and the other by Dr. Umesh Khot,² examine percutaneous coronary intervention (PCI) in patients with stable chronic angina. Both discuss the findings of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial³ and how to use these findings in helping patients decide whether to undergo PCI.

Rothberg and Khot agree that in the COURAGE trial, PCI effectively if not completely reduced angina but did not decrease the likelihood of death or subsequent myocardial infarction (MI). Patients were excluded from the study if they had a likelihood of left main disease, heart failure, or severe angina. All underwent catheterization, and all were given optimal medical therapy. Thus, the trial results do not directly relate to every patient with stable angina.

While the patient may find it confusing that angina and the risk of MI are not reduced in parallel, since both are due to atherosclerosis, their dynamic pathophysiology is different. The COURAGE results are the mirror image of those in some early studies of aspirin in coronary disease, in which aspirin reduced the incidence of MI but did not significantly affect angina.

In view of the COURAGE results, Rothberg seems surprised that PCI continues to be frequently used in patients with stable angina. He points out that according to some surveys,⁴ not all cardiologists have embraced these (and other similar study) results. But as Khot notes, the use of PCI in stable angina has decreased. More interesting to me were the results of an online study conducted by Rothberg and colleagues in which participants were provided different background information about PCI.⁵ Even if given explicit information that PCI did not prevent MI, a fair number still said they would choose it and still believed it would prevent this outcome. Bias clearly influences what patients read and hear, and they bring these biases into the shared decision-making process.

While some patients may not fully understand PCI’s risks and putative benefits, others may choose it because of their personal knowledge of others’ experience or perhaps because the “softer” benefits demonstrated in COURAGE and other trials are important to them. As outlined by Khot, patients who underwent PCI had more rapid relief of angina symptoms, possibly experienced greater relief of symptoms even if incomplete, and needed less medication. More patients needed urgent revascularization in the medical group than in the PCI group. Rothberg appropriately notes that this did not “equate to a reduction in the rate of MI,” but to some patients (eg, international travelers, caregivers) this higher possibility of needing an urgent procedure may be enough to make them want the initial elective procedure. While patients should be told that many of the patients in the medical therapy group in COURAGE crossed over to get PCI (16% at 1 year, and about 1/3 after a median of 4.6 years of follow-up), a patient for whom avoiding invasive procedures is the highest priority will likely “hear” that he or she has a 2/3 likelihood of not needing PCI without being at increased risk of death or MI with medical therapy.

As Rothberg points out, “providing information alone is not enough.” The patient needs to recognize, verbalize, and perhaps rank his or her own biases, fears, and desires. Equally important, we need to recognize our own biases and not let them overshadow the patient’s concerns.

I urge you to read both articles, not only because they offer excellent critiques of the COURAGE results and what they mean in practice, but also because they should make us reflect on how often and well we engage in shared decision-making with our patients. Reading these made me realize that I need to better understand my patients’ concerns. Discussing my interpretation of clinical study results, no matter how sophisticated or correct, and then offering a recommendation without fully understanding the patient’s treatment goals is not shared decision-making. The seemingly “wrong” decision may be right for the patient.

Medications started for appropriate indications in middle age may need to be monitored more closely as the patient ages. Some drugs may become unnecessary or even dangerous as the patient ages, functional status and renal function decline, and goals of care change.

See related editorial

Older adults tend to have multiple illnesses and therefore take more drugs, and polypharmacy increases the risk of poor outcomes. The number of medications a person uses is a risk factor for adverse drug reactions, nonadherence, financial burden, drug-drug interactions, and worse outcomes.¹

The prevalence of polypharmacy increased from an estimated 8.2% to 15% from 1999 to 2011 based on the National Health and Nutrition Examination Survey.² Guideline-based therapy for specific diseases may lead to the addition of more medications to reach disease targets.³ Most older adults in the United States compound the risk of prescribed medications by also taking over-the-counter medications and dietary supplements.⁴

In addition, medications are often used in older adults based on studies of younger persons without significant comorbidities. Applying clinical guidelines based on these studies to older adults with comorbidity and functional impairment is challenging.⁵ Age-related pharmacokinetic and pharmacodynamic changes increase the risk of adverse drug reactions.⁶

In this article, we review commonly used medications that are potentially inappropriate based on clinical practice. We also review tools to evaluate appropriate drug therapy in older adults.

DRUGS THAT ARE COMMONLY USED, BUT POTENTIALLY INAPPROPRIATE

Statins

Statins are effective when used as secondary prevention in older adults,⁷ but their efficacy when used as primary prevention of atherosclerotic cardiovascular disease in people age 75 and older is questionable.⁸ Nevertheless, they are widely used for this purpose. For example, before the 2013 joint guidelines of the American College of Cardiology and the American Heart Association (ACC/AHA) were released, 22% of patients age 80 and older in the Geisinger health system were taking a statin for primary prevention.⁹

The 2013 ACC/AHA guidelines included a limited recommendation for statins for primary prevention of atherosclerotic cardiovascular disease in adults age 75 and older.¹⁰ The guideline noted, however, that few data were available to support this recommendation.¹⁰

In a systematic review of 18 randomized clinical trials of statins for primary prevention of atherosclerotic cardiovascular disease, the mean age was 57, yet conclusions were extrapolated to an older patient population.¹¹ The estimated 10-year risk of atherosclerotic cardiovascular disease based on pooled cohort risk equations of adults age 75 and older always exceeds the 7.5% treatment threshold recommended by the guidelines.⁸

Myopathy is a common adverse effect of statins. In addition, statins interact with other drugs that inhibit the cytochrome P450 3A4 isoenzyme, such as amlodipine, amiodarone, and diltiazem.^8,12 If statin therapy caused no functional limitation due to muscle pain or weakness, statins for primary prevention would be cost-effective, but even a small increase in adverse effects in an elderly patient can offset the cardiovascular benefit.¹³ A recent post hoc secondary analysis found no benefit of pravastatin for primary prevention in adults age 75 and older.¹⁴

Thus, statin treatment for primary prevention in older patients should be individualized, based on life expectancy, function, and cardiovascular risk. Statin therapy does not replace modification of other risk factors.

Anticholinergics

Drugs with anticholinergic properties are commonly prescribed in the elderly for conditions such as muscle spasm, overactive bladder, psychiatric disorders, insomnia, extrapyramidal symptoms, vertigo, pruritus, peptic ulcer disease, seasonal allergies, and even the common cold,¹⁵ and many of the drugs often prescribed have strong anticholinergic properties (Table 1). Taking multiple medications with anticholinergic properties results in a high “anticholinergic burden,” which is associated with falls, impulsive behavior, poor physical performance, loss of independence, dementia, delirium, and brain atrophy.^15–18

The 2014 American College of Physicians guideline on nonsurgical management of urinary incontinence in women recommends pharmacologic treatment for urgency and stress urinary incontinence after failure of nonpharmacologic therapy,¹⁹ and many drugs for these urinary symptoms have anticholinergic properties. If an anticholinergic is necessary, an agent that results in a lower anticholinergic burden should be considered in older patients.

A pharmacist-initiated medication review and intervention may be another way to adjust medications to reduce the patient’s anticholinergic burden.^20,21 The common use of anticholinergic drugs in older adults reminds us to monitor their use closely.²²

Benzodiazepines are among the most commonly prescribed psychotropics in developed countries and are prescribed mainly by primary care physicians rather than psychiatrists.²³

In 2008, 5.2% of US adults ages 18 to 80 used a benzodiazepine, and long-term use was more prevalent in older patients (ages 65–80).²³

Benzodiazepines are prescribed for anxiety,²⁴ insomnia,²⁵ and agitation. They can cause withdrawal²⁶ and have potential for abuse.²⁷ Benzodiazepines are associated with cognitive decline,²⁸ impaired driving,²⁹ falls,³⁰ and hip fractures³¹ in older adults.

In addition, use of nonbenzodiazepine hypnotics (eg, zolpidem) is on the rise,³² and these drugs are known to increase the risk of hip fracture in nursing home residents.³³

The American Geriatrics Society, through the American Board of Internal Medicine’s Choosing Wisely campaign, recommends avoiding benzodiazepines as a first-line treatment for insomnia, agitation, or delirium in older adults.³⁴ Yet prescribing practices with these drugs in primary care settings conflict with guidelines, partly due to lack of training in constructive strategies regarding appropriate use of benzodiazepines.³⁵ Educating patients on the risks and benefits of benzodiazepine treatment, especially long-term use, has been shown to reduce the rate of benzodiazepine-associated secondary events.³⁶

Antipsychotics

Off-label use of antipsychotics is common and is increasing in the United States. In 2008, off-label use of antipsychotic drugs accounted for an estimated $6 billion.³⁷ A common off-label use is to manage behavioral symptoms of dementia, despite a black-box warning about an increased risk of death in patients with dementia who are treated with antipsychotics.^38,39 The Choosing Wisely campaign recommends against prescribing antipsychotics as a first-line treatment of behavioral and psychological symptoms of dementia.³⁴

Antipsychotic drugs are associated with risk of acute kidney injury,⁴⁰ as well as increased risk of falls and fractures (eg, a 52% higher risk of a serious fall, and a 50% higher risk of a nonvertebral osteoporotic fracture).⁴¹

Patients with dementia often exhibit aggression, resistance to care, and other challenging or disruptive behaviors. In such instances, antipsychotic drugs are often prescribed, but they provide limited and inconsistent benefits, while causing oversedation and worsening of cognitive function and increasing the likelihood of falling, stroke, and death.^38,39,41

Because pharmacologic treatments for dementia are only modestly effective, have notable risks, and do not treat some of the behaviors that family members and caregivers find most distressing, nonpharmacologic measures are recommended as first-line treatment.⁴² These include caregiver education and support, training in problem-solving, and targeted therapy directed at the underlying causes of specific behaviors (eg, implementing nighttime routines to address sleep disturbances).⁴² Nonpharmacologic management of behavioral symptoms in dementia can significantly improve quality of life for patients and caregivers.⁴² Use of antipsychotic drugs in patients with dementia should be limited to cases in which nonpharmacologic measures have failed and patients pose an imminent threat to themselves or others.⁴³

Proton pump inhibitors

Proton pump inhibitors are among the most commonly prescribed medications in the United States, and their use has increased significantly over the decade. It has been estimated that between 25% and 70% of these prescriptions have no appropriate indication.⁴⁴

There is considerable excess use of acid suppressants in both inpatient and outpatient settings.^45,46 In one study, at discharge from an internal medicine service, almost half of patients were taking a proton pump inhibitor.⁴⁷

Evidence-based guidelines recommend these drugs to treat gastroesophageal reflux disease, nonerosive reflux disease, erosive esophagitis, dyspepsia, and peptic ulcer disease. However, long-term use (ie, beyond 8 weeks) is recommended only for patients with erosive esophagitis, Barrett esophagus, a pathologic hypersecretory condition, or a demonstrated need for maintenance treatment for reflux disease.⁴⁸

Although proton pump inhibitors are highly effective and have low toxicity, there are reports of an association with Clostridium difficile infection,⁴⁹ community-acquired pneumonia,⁵⁰ hip fracture,⁵¹ vitamin B₁₂ deficiency,⁵² atrophic gastritis,⁵³ kidney disease,⁵⁴ and dementia.⁵⁵

Nondrug therapies such as weight loss and elevation of the head of the bed may improve esophageal pH levels and reflux symptoms.⁵⁶

Deprescribing.org has practical advice for healthcare providers, patients, and caregivers on how to discontinue proton pump inhibitors, including videos, algorithms, and guidelines.

TOOLS TO EVALUATE APPROPRIATE DRUG THERAPY

Beers criteria

The Beers criteria (Table 2), developed in 1991 by a geriatrician as an approach to safer, more effective drug therapy in frail elderly nursing home patients,⁵⁷ were updated by the American Geriatrics Society in 2015 for use in any clinical setting.⁵⁸ (The criteria are also available as a smartphone application through the American Geriatrics Society at www.americangeriatrics.org.)

The Beers criteria offer evidence-based recommendations on drugs to avoid in the elderly, along with the rationale for use, the quality of evidence behind the recommendation, and the graded strength of the recommendation. The Beers criteria should be viewed through the lens of clinical judgment to offer safer nonpharmacologic and pharmacologic treatments.

The Joint Commission recommends medication reconciliation at every transition of care.⁵⁹ The Beers criteria are a good starting point for a comprehensive medication review.

STOPP/START criteria

Another tool to aid safe prescribing in older adults is the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions (STOPP), used in conjuction with the Screening Tool to Alert Doctors to Right Treatment (START). The STOPP/START criteria^60,61 are based on an up-to-date literature review and consensus (Table 3).

THE BOTTOM LINE

Physicians caring for older adults need to diligently weigh the benefits of drug therapy and consider the patient’s care goals, current level of functioning, life expectancy, values, and preferences. Statin therapy for primary prevention, anticholinergics, benzodiazepines, antipsychotics, and proton pump inhibitors are widely used without proper indications, pointing to the need for a periodic comprehensive review of medications to reevaluate the risks vs the benefits of the patient’s medications. The Beers criteria and the STOPP/ START criteria can be useful tools for this purpose.

Medications started for appropriate indications in middle age may need to be monitored more closely as the patient ages. Some drugs may become unnecessary or even dangerous as the patient ages, functional status and renal function decline, and goals of care change.

See related editorial

Older adults tend to have multiple illnesses and therefore take more drugs, and polypharmacy increases the risk of poor outcomes. The number of medications a person uses is a risk factor for adverse drug reactions, nonadherence, financial burden, drug-drug interactions, and worse outcomes.¹

The prevalence of polypharmacy increased from an estimated 8.2% to 15% from 1999 to 2011 based on the National Health and Nutrition Examination Survey.² Guideline-based therapy for specific diseases may lead to the addition of more medications to reach disease targets.³ Most older adults in the United States compound the risk of prescribed medications by also taking over-the-counter medications and dietary supplements.⁴

In addition, medications are often used in older adults based on studies of younger persons without significant comorbidities. Applying clinical guidelines based on these studies to older adults with comorbidity and functional impairment is challenging.⁵ Age-related pharmacokinetic and pharmacodynamic changes increase the risk of adverse drug reactions.⁶

In this article, we review commonly used medications that are potentially inappropriate based on clinical practice. We also review tools to evaluate appropriate drug therapy in older adults.

DRUGS THAT ARE COMMONLY USED, BUT POTENTIALLY INAPPROPRIATE

Statins

Statins are effective when used as secondary prevention in older adults,⁷ but their efficacy when used as primary prevention of atherosclerotic cardiovascular disease in people age 75 and older is questionable.⁸ Nevertheless, they are widely used for this purpose. For example, before the 2013 joint guidelines of the American College of Cardiology and the American Heart Association (ACC/AHA) were released, 22% of patients age 80 and older in the Geisinger health system were taking a statin for primary prevention.⁹

The 2013 ACC/AHA guidelines included a limited recommendation for statins for primary prevention of atherosclerotic cardiovascular disease in adults age 75 and older.¹⁰ The guideline noted, however, that few data were available to support this recommendation.¹⁰

In a systematic review of 18 randomized clinical trials of statins for primary prevention of atherosclerotic cardiovascular disease, the mean age was 57, yet conclusions were extrapolated to an older patient population.¹¹ The estimated 10-year risk of atherosclerotic cardiovascular disease based on pooled cohort risk equations of adults age 75 and older always exceeds the 7.5% treatment threshold recommended by the guidelines.⁸

Myopathy is a common adverse effect of statins. In addition, statins interact with other drugs that inhibit the cytochrome P450 3A4 isoenzyme, such as amlodipine, amiodarone, and diltiazem.^8,12 If statin therapy caused no functional limitation due to muscle pain or weakness, statins for primary prevention would be cost-effective, but even a small increase in adverse effects in an elderly patient can offset the cardiovascular benefit.¹³ A recent post hoc secondary analysis found no benefit of pravastatin for primary prevention in adults age 75 and older.¹⁴

Thus, statin treatment for primary prevention in older patients should be individualized, based on life expectancy, function, and cardiovascular risk. Statin therapy does not replace modification of other risk factors.

Anticholinergics

Drugs with anticholinergic properties are commonly prescribed in the elderly for conditions such as muscle spasm, overactive bladder, psychiatric disorders, insomnia, extrapyramidal symptoms, vertigo, pruritus, peptic ulcer disease, seasonal allergies, and even the common cold,¹⁵ and many of the drugs often prescribed have strong anticholinergic properties (Table 1). Taking multiple medications with anticholinergic properties results in a high “anticholinergic burden,” which is associated with falls, impulsive behavior, poor physical performance, loss of independence, dementia, delirium, and brain atrophy.^15–18

The 2014 American College of Physicians guideline on nonsurgical management of urinary incontinence in women recommends pharmacologic treatment for urgency and stress urinary incontinence after failure of nonpharmacologic therapy,¹⁹ and many drugs for these urinary symptoms have anticholinergic properties. If an anticholinergic is necessary, an agent that results in a lower anticholinergic burden should be considered in older patients.

A pharmacist-initiated medication review and intervention may be another way to adjust medications to reduce the patient’s anticholinergic burden.^20,21 The common use of anticholinergic drugs in older adults reminds us to monitor their use closely.²²

Benzodiazepines are among the most commonly prescribed psychotropics in developed countries and are prescribed mainly by primary care physicians rather than psychiatrists.²³

In 2008, 5.2% of US adults ages 18 to 80 used a benzodiazepine, and long-term use was more prevalent in older patients (ages 65–80).²³

Benzodiazepines are prescribed for anxiety,²⁴ insomnia,²⁵ and agitation. They can cause withdrawal²⁶ and have potential for abuse.²⁷ Benzodiazepines are associated with cognitive decline,²⁸ impaired driving,²⁹ falls,³⁰ and hip fractures³¹ in older adults.

In addition, use of nonbenzodiazepine hypnotics (eg, zolpidem) is on the rise,³² and these drugs are known to increase the risk of hip fracture in nursing home residents.³³

The American Geriatrics Society, through the American Board of Internal Medicine’s Choosing Wisely campaign, recommends avoiding benzodiazepines as a first-line treatment for insomnia, agitation, or delirium in older adults.³⁴ Yet prescribing practices with these drugs in primary care settings conflict with guidelines, partly due to lack of training in constructive strategies regarding appropriate use of benzodiazepines.³⁵ Educating patients on the risks and benefits of benzodiazepine treatment, especially long-term use, has been shown to reduce the rate of benzodiazepine-associated secondary events.³⁶

Antipsychotics

Off-label use of antipsychotics is common and is increasing in the United States. In 2008, off-label use of antipsychotic drugs accounted for an estimated $6 billion.³⁷ A common off-label use is to manage behavioral symptoms of dementia, despite a black-box warning about an increased risk of death in patients with dementia who are treated with antipsychotics.^38,39 The Choosing Wisely campaign recommends against prescribing antipsychotics as a first-line treatment of behavioral and psychological symptoms of dementia.³⁴

Antipsychotic drugs are associated with risk of acute kidney injury,⁴⁰ as well as increased risk of falls and fractures (eg, a 52% higher risk of a serious fall, and a 50% higher risk of a nonvertebral osteoporotic fracture).⁴¹

Patients with dementia often exhibit aggression, resistance to care, and other challenging or disruptive behaviors. In such instances, antipsychotic drugs are often prescribed, but they provide limited and inconsistent benefits, while causing oversedation and worsening of cognitive function and increasing the likelihood of falling, stroke, and death.^38,39,41

Because pharmacologic treatments for dementia are only modestly effective, have notable risks, and do not treat some of the behaviors that family members and caregivers find most distressing, nonpharmacologic measures are recommended as first-line treatment.⁴² These include caregiver education and support, training in problem-solving, and targeted therapy directed at the underlying causes of specific behaviors (eg, implementing nighttime routines to address sleep disturbances).⁴² Nonpharmacologic management of behavioral symptoms in dementia can significantly improve quality of life for patients and caregivers.⁴² Use of antipsychotic drugs in patients with dementia should be limited to cases in which nonpharmacologic measures have failed and patients pose an imminent threat to themselves or others.⁴³

Proton pump inhibitors

Proton pump inhibitors are among the most commonly prescribed medications in the United States, and their use has increased significantly over the decade. It has been estimated that between 25% and 70% of these prescriptions have no appropriate indication.⁴⁴

There is considerable excess use of acid suppressants in both inpatient and outpatient settings.^45,46 In one study, at discharge from an internal medicine service, almost half of patients were taking a proton pump inhibitor.⁴⁷

Evidence-based guidelines recommend these drugs to treat gastroesophageal reflux disease, nonerosive reflux disease, erosive esophagitis, dyspepsia, and peptic ulcer disease. However, long-term use (ie, beyond 8 weeks) is recommended only for patients with erosive esophagitis, Barrett esophagus, a pathologic hypersecretory condition, or a demonstrated need for maintenance treatment for reflux disease.⁴⁸

Although proton pump inhibitors are highly effective and have low toxicity, there are reports of an association with Clostridium difficile infection,⁴⁹ community-acquired pneumonia,⁵⁰ hip fracture,⁵¹ vitamin B₁₂ deficiency,⁵² atrophic gastritis,⁵³ kidney disease,⁵⁴ and dementia.⁵⁵

Nondrug therapies such as weight loss and elevation of the head of the bed may improve esophageal pH levels and reflux symptoms.⁵⁶

Deprescribing.org has practical advice for healthcare providers, patients, and caregivers on how to discontinue proton pump inhibitors, including videos, algorithms, and guidelines.

TOOLS TO EVALUATE APPROPRIATE DRUG THERAPY

Beers criteria

The Beers criteria (Table 2), developed in 1991 by a geriatrician as an approach to safer, more effective drug therapy in frail elderly nursing home patients,⁵⁷ were updated by the American Geriatrics Society in 2015 for use in any clinical setting.⁵⁸ (The criteria are also available as a smartphone application through the American Geriatrics Society at www.americangeriatrics.org.)

The Beers criteria offer evidence-based recommendations on drugs to avoid in the elderly, along with the rationale for use, the quality of evidence behind the recommendation, and the graded strength of the recommendation. The Beers criteria should be viewed through the lens of clinical judgment to offer safer nonpharmacologic and pharmacologic treatments.

The Joint Commission recommends medication reconciliation at every transition of care.⁵⁹ The Beers criteria are a good starting point for a comprehensive medication review.

STOPP/START criteria

Another tool to aid safe prescribing in older adults is the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions (STOPP), used in conjuction with the Screening Tool to Alert Doctors to Right Treatment (START). The STOPP/START criteria^60,61 are based on an up-to-date literature review and consensus (Table 3).

THE BOTTOM LINE

Physicians caring for older adults need to diligently weigh the benefits of drug therapy and consider the patient’s care goals, current level of functioning, life expectancy, values, and preferences. Statin therapy for primary prevention, anticholinergics, benzodiazepines, antipsychotics, and proton pump inhibitors are widely used without proper indications, pointing to the need for a periodic comprehensive review of medications to reevaluate the risks vs the benefits of the patient’s medications. The Beers criteria and the STOPP/ START criteria can be useful tools for this purpose.

Medications started for appropriate indications in middle age may need to be monitored more closely as the patient ages. Some drugs may become unnecessary or even dangerous as the patient ages, functional status and renal function decline, and goals of care change.

See related editorial

Older adults tend to have multiple illnesses and therefore take more drugs, and polypharmacy increases the risk of poor outcomes. The number of medications a person uses is a risk factor for adverse drug reactions, nonadherence, financial burden, drug-drug interactions, and worse outcomes.¹

The prevalence of polypharmacy increased from an estimated 8.2% to 15% from 1999 to 2011 based on the National Health and Nutrition Examination Survey.² Guideline-based therapy for specific diseases may lead to the addition of more medications to reach disease targets.³ Most older adults in the United States compound the risk of prescribed medications by also taking over-the-counter medications and dietary supplements.⁴

In addition, medications are often used in older adults based on studies of younger persons without significant comorbidities. Applying clinical guidelines based on these studies to older adults with comorbidity and functional impairment is challenging.⁵ Age-related pharmacokinetic and pharmacodynamic changes increase the risk of adverse drug reactions.⁶

In this article, we review commonly used medications that are potentially inappropriate based on clinical practice. We also review tools to evaluate appropriate drug therapy in older adults.

DRUGS THAT ARE COMMONLY USED, BUT POTENTIALLY INAPPROPRIATE

Statins

Statins are effective when used as secondary prevention in older adults,⁷ but their efficacy when used as primary prevention of atherosclerotic cardiovascular disease in people age 75 and older is questionable.⁸ Nevertheless, they are widely used for this purpose. For example, before the 2013 joint guidelines of the American College of Cardiology and the American Heart Association (ACC/AHA) were released, 22% of patients age 80 and older in the Geisinger health system were taking a statin for primary prevention.⁹

The 2013 ACC/AHA guidelines included a limited recommendation for statins for primary prevention of atherosclerotic cardiovascular disease in adults age 75 and older.¹⁰ The guideline noted, however, that few data were available to support this recommendation.¹⁰

In a systematic review of 18 randomized clinical trials of statins for primary prevention of atherosclerotic cardiovascular disease, the mean age was 57, yet conclusions were extrapolated to an older patient population.¹¹ The estimated 10-year risk of atherosclerotic cardiovascular disease based on pooled cohort risk equations of adults age 75 and older always exceeds the 7.5% treatment threshold recommended by the guidelines.⁸

Myopathy is a common adverse effect of statins. In addition, statins interact with other drugs that inhibit the cytochrome P450 3A4 isoenzyme, such as amlodipine, amiodarone, and diltiazem.^8,12 If statin therapy caused no functional limitation due to muscle pain or weakness, statins for primary prevention would be cost-effective, but even a small increase in adverse effects in an elderly patient can offset the cardiovascular benefit.¹³ A recent post hoc secondary analysis found no benefit of pravastatin for primary prevention in adults age 75 and older.¹⁴

Thus, statin treatment for primary prevention in older patients should be individualized, based on life expectancy, function, and cardiovascular risk. Statin therapy does not replace modification of other risk factors.

Anticholinergics

Drugs with anticholinergic properties are commonly prescribed in the elderly for conditions such as muscle spasm, overactive bladder, psychiatric disorders, insomnia, extrapyramidal symptoms, vertigo, pruritus, peptic ulcer disease, seasonal allergies, and even the common cold,¹⁵ and many of the drugs often prescribed have strong anticholinergic properties (Table 1). Taking multiple medications with anticholinergic properties results in a high “anticholinergic burden,” which is associated with falls, impulsive behavior, poor physical performance, loss of independence, dementia, delirium, and brain atrophy.^15–18

The 2014 American College of Physicians guideline on nonsurgical management of urinary incontinence in women recommends pharmacologic treatment for urgency and stress urinary incontinence after failure of nonpharmacologic therapy,¹⁹ and many drugs for these urinary symptoms have anticholinergic properties. If an anticholinergic is necessary, an agent that results in a lower anticholinergic burden should be considered in older patients.

A pharmacist-initiated medication review and intervention may be another way to adjust medications to reduce the patient’s anticholinergic burden.^20,21 The common use of anticholinergic drugs in older adults reminds us to monitor their use closely.²²

Benzodiazepines are among the most commonly prescribed psychotropics in developed countries and are prescribed mainly by primary care physicians rather than psychiatrists.²³

In 2008, 5.2% of US adults ages 18 to 80 used a benzodiazepine, and long-term use was more prevalent in older patients (ages 65–80).²³

Benzodiazepines are prescribed for anxiety,²⁴ insomnia,²⁵ and agitation. They can cause withdrawal²⁶ and have potential for abuse.²⁷ Benzodiazepines are associated with cognitive decline,²⁸ impaired driving,²⁹ falls,³⁰ and hip fractures³¹ in older adults.

In addition, use of nonbenzodiazepine hypnotics (eg, zolpidem) is on the rise,³² and these drugs are known to increase the risk of hip fracture in nursing home residents.³³

The American Geriatrics Society, through the American Board of Internal Medicine’s Choosing Wisely campaign, recommends avoiding benzodiazepines as a first-line treatment for insomnia, agitation, or delirium in older adults.³⁴ Yet prescribing practices with these drugs in primary care settings conflict with guidelines, partly due to lack of training in constructive strategies regarding appropriate use of benzodiazepines.³⁵ Educating patients on the risks and benefits of benzodiazepine treatment, especially long-term use, has been shown to reduce the rate of benzodiazepine-associated secondary events.³⁶

Antipsychotics

Off-label use of antipsychotics is common and is increasing in the United States. In 2008, off-label use of antipsychotic drugs accounted for an estimated $6 billion.³⁷ A common off-label use is to manage behavioral symptoms of dementia, despite a black-box warning about an increased risk of death in patients with dementia who are treated with antipsychotics.^38,39 The Choosing Wisely campaign recommends against prescribing antipsychotics as a first-line treatment of behavioral and psychological symptoms of dementia.³⁴

Antipsychotic drugs are associated with risk of acute kidney injury,⁴⁰ as well as increased risk of falls and fractures (eg, a 52% higher risk of a serious fall, and a 50% higher risk of a nonvertebral osteoporotic fracture).⁴¹

Patients with dementia often exhibit aggression, resistance to care, and other challenging or disruptive behaviors. In such instances, antipsychotic drugs are often prescribed, but they provide limited and inconsistent benefits, while causing oversedation and worsening of cognitive function and increasing the likelihood of falling, stroke, and death.^38,39,41

Because pharmacologic treatments for dementia are only modestly effective, have notable risks, and do not treat some of the behaviors that family members and caregivers find most distressing, nonpharmacologic measures are recommended as first-line treatment.⁴² These include caregiver education and support, training in problem-solving, and targeted therapy directed at the underlying causes of specific behaviors (eg, implementing nighttime routines to address sleep disturbances).⁴² Nonpharmacologic management of behavioral symptoms in dementia can significantly improve quality of life for patients and caregivers.⁴² Use of antipsychotic drugs in patients with dementia should be limited to cases in which nonpharmacologic measures have failed and patients pose an imminent threat to themselves or others.⁴³

Proton pump inhibitors

Proton pump inhibitors are among the most commonly prescribed medications in the United States, and their use has increased significantly over the decade. It has been estimated that between 25% and 70% of these prescriptions have no appropriate indication.⁴⁴

There is considerable excess use of acid suppressants in both inpatient and outpatient settings.^45,46 In one study, at discharge from an internal medicine service, almost half of patients were taking a proton pump inhibitor.⁴⁷

Evidence-based guidelines recommend these drugs to treat gastroesophageal reflux disease, nonerosive reflux disease, erosive esophagitis, dyspepsia, and peptic ulcer disease. However, long-term use (ie, beyond 8 weeks) is recommended only for patients with erosive esophagitis, Barrett esophagus, a pathologic hypersecretory condition, or a demonstrated need for maintenance treatment for reflux disease.⁴⁸

Although proton pump inhibitors are highly effective and have low toxicity, there are reports of an association with Clostridium difficile infection,⁴⁹ community-acquired pneumonia,⁵⁰ hip fracture,⁵¹ vitamin B₁₂ deficiency,⁵² atrophic gastritis,⁵³ kidney disease,⁵⁴ and dementia.⁵⁵

Nondrug therapies such as weight loss and elevation of the head of the bed may improve esophageal pH levels and reflux symptoms.⁵⁶

Deprescribing.org has practical advice for healthcare providers, patients, and caregivers on how to discontinue proton pump inhibitors, including videos, algorithms, and guidelines.

TOOLS TO EVALUATE APPROPRIATE DRUG THERAPY

Beers criteria

The Beers criteria (Table 2), developed in 1991 by a geriatrician as an approach to safer, more effective drug therapy in frail elderly nursing home patients,⁵⁷ were updated by the American Geriatrics Society in 2015 for use in any clinical setting.⁵⁸ (The criteria are also available as a smartphone application through the American Geriatrics Society at www.americangeriatrics.org.)

The Beers criteria offer evidence-based recommendations on drugs to avoid in the elderly, along with the rationale for use, the quality of evidence behind the recommendation, and the graded strength of the recommendation. The Beers criteria should be viewed through the lens of clinical judgment to offer safer nonpharmacologic and pharmacologic treatments.

The Joint Commission recommends medication reconciliation at every transition of care.⁵⁹ The Beers criteria are a good starting point for a comprehensive medication review.

STOPP/START criteria

Another tool to aid safe prescribing in older adults is the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions (STOPP), used in conjuction with the Screening Tool to Alert Doctors to Right Treatment (START). The STOPP/START criteria^60,61 are based on an up-to-date literature review and consensus (Table 3).

THE BOTTOM LINE

Physicians caring for older adults need to diligently weigh the benefits of drug therapy and consider the patient’s care goals, current level of functioning, life expectancy, values, and preferences. Statin therapy for primary prevention, anticholinergics, benzodiazepines, antipsychotics, and proton pump inhibitors are widely used without proper indications, pointing to the need for a periodic comprehensive review of medications to reevaluate the risks vs the benefits of the patient’s medications. The Beers criteria and the STOPP/ START criteria can be useful tools for this purpose.

According to a 2016 study, more than one-third of older adults in the United States take 5 or more medications.¹ This is a growing problem. Not only do older patients take more drugs than younger patients, they are also at higher risk of adverse drug events, drug-drug interactions, geriatric syndromes, and lower adherence.²

See related article

Many drugs that older patients are given are potentially inappropriate, ie, their risks outweigh the expected benefits, particularly when effective and safer alternative therapies exist. Although many clinicians are aware of the risks of polypharmacy, they may not be confident in discontinuing potentially inappropriate medications. The process of deliberately tapering, stopping, or reducing doses of medications with the goal of reducing harm and improving patient outcomes is known as deprescribing.³

In this issue, Kim et al⁴ review several medications that are overused or often used inappropriately in older adults: statins for primary prevention of atherosclerotic cardiovascular disease, anticholinergic drugs, benzodiazepines, antipsychotics, and proton pump inhibitors. They offer guidance about the situations in which these drugs may be inappropriate as well as alternative drug and nondrug treatments. Further, they suggest that, when prescribing or deprescribing drugs in older adults, clinicians consult tools such as the Beers criteria and the STOPP/START criteria (the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions, and the Screening Tool to Alert Doctors to Right Treatment).

The issues Kim et al review are highly relevant and may increase awareness of specific potentially inappropriate medications. They also remind us that nonpharmacologic treatments are first-line for many medical conditions. In an era of a pill for every ill and a quick-fix mentality among both patients and providers, lifestyle changes and other nonpharmacologic treatments may be overlooked. Similarly, the STOPP/START criteria, which are concrete, evidence-based recommendations that can be applied to patient care, are likely underused in clinical practice.

Although necessary and valuable, simply arming clinicians with knowledge is insufficient to tackle the problems of polypharmacy and inappropriate prescribing. As the authors note in their discussion of benzodiazepines, practice guidelines exist regarding prescribing these agents, and data from randomized trials support specific interventions to deprescribe them.⁵ Nevertheless, clinicians report feeling inadequately prepared to discontinue benzodiazepines, particularly when patients perceive benefit from them. As such, user-friendly tools and specific strategies for weighing risks vs benefits are critical for clinicians.

PUTTING KNOWLEDGE INTO PRACTICE

How do we translate knowledge into practice with regard to deprescribing potentially inappropriate medications in older patients—or prescribing drugs only if appropriate in the first place?

An opportunity arises when patients are in the hospital. Taking a medication history on admission and matching medications with indications are key starting points. Clinical pharmacists can help screen for side effects and potential interactions and can provide deprescribing recommendations. Meticulous discharge medication reconciliation, patient education, and communication of the updated medication list to the outpatient provider are central to ensuring that patients adhere to medication adjustments after they go home.

A MATTER OF BALANCE

Another factor to consider is the patient’s physiologic age compared with his or her chronologic age. If a patient has multiple comorbidities, frailty, limited life expectancy, or poor renal function, we may consider her older than her chronologic age. In this case, a drug’s risks may outweigh its benefit, which is something to be discussed. On the other hand, a high-functioning and relatively healthy elderly patient may be a candidate for medications known to reduce the risk of death or control a chronic disease better. Incorporating a patient’s goals of care and using shared decision-making are also likely to yield an optimal medication regimen.

Smartphone apps and resources embedded in electronic health records provide additional decision support. Used when prescribing or reconciling medications, these supplemental brains offer instant feedback and information on dose adjustments, drug interactions, clinical guidelines, and even potentially inappropriate medications. While the impacts of these electronic tools on prescribing patterns and outcomes in geriatric populations remain unclear, new ones are being developed and studied.⁶ This may be the most promising way to translate knowledge into practice, as it is more easily integrated with existing clinician workflows.

AN OPPORTUNITY TO IMPROVE

There is significant opportunity to reduce polypharmacy and optimize medication prescribing practices for older adults. Awareness of potentially inappropriate medications and clinical situations in which the use of certain classes of medications should be minimized is the first step in addressing this problem. Using tools such as the STOPP/START criteria, reviewing medications at critical transition points, prioritizing patient function and goals, and using electronic clinical decision support should aid prescribing decisions.

Whenever possible, collaborating with other care team members such as pharmacists may increase efficiency and effectiveness of medication management. Ultimately, inclusion of more older adults in clinical trials may provide data-driven guidance for weighing risks and benefits. Finally, further study of the effects of deprescribing on clinical outcomes may be the missing piece to help clinicians and patients find balance in prescription management.

According to a 2016 study, more than one-third of older adults in the United States take 5 or more medications.¹ This is a growing problem. Not only do older patients take more drugs than younger patients, they are also at higher risk of adverse drug events, drug-drug interactions, geriatric syndromes, and lower adherence.²

See related article

Many drugs that older patients are given are potentially inappropriate, ie, their risks outweigh the expected benefits, particularly when effective and safer alternative therapies exist. Although many clinicians are aware of the risks of polypharmacy, they may not be confident in discontinuing potentially inappropriate medications. The process of deliberately tapering, stopping, or reducing doses of medications with the goal of reducing harm and improving patient outcomes is known as deprescribing.³

In this issue, Kim et al⁴ review several medications that are overused or often used inappropriately in older adults: statins for primary prevention of atherosclerotic cardiovascular disease, anticholinergic drugs, benzodiazepines, antipsychotics, and proton pump inhibitors. They offer guidance about the situations in which these drugs may be inappropriate as well as alternative drug and nondrug treatments. Further, they suggest that, when prescribing or deprescribing drugs in older adults, clinicians consult tools such as the Beers criteria and the STOPP/START criteria (the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions, and the Screening Tool to Alert Doctors to Right Treatment).

The issues Kim et al review are highly relevant and may increase awareness of specific potentially inappropriate medications. They also remind us that nonpharmacologic treatments are first-line for many medical conditions. In an era of a pill for every ill and a quick-fix mentality among both patients and providers, lifestyle changes and other nonpharmacologic treatments may be overlooked. Similarly, the STOPP/START criteria, which are concrete, evidence-based recommendations that can be applied to patient care, are likely underused in clinical practice.

Although necessary and valuable, simply arming clinicians with knowledge is insufficient to tackle the problems of polypharmacy and inappropriate prescribing. As the authors note in their discussion of benzodiazepines, practice guidelines exist regarding prescribing these agents, and data from randomized trials support specific interventions to deprescribe them.⁵ Nevertheless, clinicians report feeling inadequately prepared to discontinue benzodiazepines, particularly when patients perceive benefit from them. As such, user-friendly tools and specific strategies for weighing risks vs benefits are critical for clinicians.

PUTTING KNOWLEDGE INTO PRACTICE

How do we translate knowledge into practice with regard to deprescribing potentially inappropriate medications in older patients—or prescribing drugs only if appropriate in the first place?

An opportunity arises when patients are in the hospital. Taking a medication history on admission and matching medications with indications are key starting points. Clinical pharmacists can help screen for side effects and potential interactions and can provide deprescribing recommendations. Meticulous discharge medication reconciliation, patient education, and communication of the updated medication list to the outpatient provider are central to ensuring that patients adhere to medication adjustments after they go home.

A MATTER OF BALANCE

Another factor to consider is the patient’s physiologic age compared with his or her chronologic age. If a patient has multiple comorbidities, frailty, limited life expectancy, or poor renal function, we may consider her older than her chronologic age. In this case, a drug’s risks may outweigh its benefit, which is something to be discussed. On the other hand, a high-functioning and relatively healthy elderly patient may be a candidate for medications known to reduce the risk of death or control a chronic disease better. Incorporating a patient’s goals of care and using shared decision-making are also likely to yield an optimal medication regimen.

Smartphone apps and resources embedded in electronic health records provide additional decision support. Used when prescribing or reconciling medications, these supplemental brains offer instant feedback and information on dose adjustments, drug interactions, clinical guidelines, and even potentially inappropriate medications. While the impacts of these electronic tools on prescribing patterns and outcomes in geriatric populations remain unclear, new ones are being developed and studied.⁶ This may be the most promising way to translate knowledge into practice, as it is more easily integrated with existing clinician workflows.

AN OPPORTUNITY TO IMPROVE

There is significant opportunity to reduce polypharmacy and optimize medication prescribing practices for older adults. Awareness of potentially inappropriate medications and clinical situations in which the use of certain classes of medications should be minimized is the first step in addressing this problem. Using tools such as the STOPP/START criteria, reviewing medications at critical transition points, prioritizing patient function and goals, and using electronic clinical decision support should aid prescribing decisions.

Whenever possible, collaborating with other care team members such as pharmacists may increase efficiency and effectiveness of medication management. Ultimately, inclusion of more older adults in clinical trials may provide data-driven guidance for weighing risks and benefits. Finally, further study of the effects of deprescribing on clinical outcomes may be the missing piece to help clinicians and patients find balance in prescription management.

According to a 2016 study, more than one-third of older adults in the United States take 5 or more medications.¹ This is a growing problem. Not only do older patients take more drugs than younger patients, they are also at higher risk of adverse drug events, drug-drug interactions, geriatric syndromes, and lower adherence.²

See related article

Many drugs that older patients are given are potentially inappropriate, ie, their risks outweigh the expected benefits, particularly when effective and safer alternative therapies exist. Although many clinicians are aware of the risks of polypharmacy, they may not be confident in discontinuing potentially inappropriate medications. The process of deliberately tapering, stopping, or reducing doses of medications with the goal of reducing harm and improving patient outcomes is known as deprescribing.³

In this issue, Kim et al⁴ review several medications that are overused or often used inappropriately in older adults: statins for primary prevention of atherosclerotic cardiovascular disease, anticholinergic drugs, benzodiazepines, antipsychotics, and proton pump inhibitors. They offer guidance about the situations in which these drugs may be inappropriate as well as alternative drug and nondrug treatments. Further, they suggest that, when prescribing or deprescribing drugs in older adults, clinicians consult tools such as the Beers criteria and the STOPP/START criteria (the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions, and the Screening Tool to Alert Doctors to Right Treatment).

The issues Kim et al review are highly relevant and may increase awareness of specific potentially inappropriate medications. They also remind us that nonpharmacologic treatments are first-line for many medical conditions. In an era of a pill for every ill and a quick-fix mentality among both patients and providers, lifestyle changes and other nonpharmacologic treatments may be overlooked. Similarly, the STOPP/START criteria, which are concrete, evidence-based recommendations that can be applied to patient care, are likely underused in clinical practice.

Although necessary and valuable, simply arming clinicians with knowledge is insufficient to tackle the problems of polypharmacy and inappropriate prescribing. As the authors note in their discussion of benzodiazepines, practice guidelines exist regarding prescribing these agents, and data from randomized trials support specific interventions to deprescribe them.⁵ Nevertheless, clinicians report feeling inadequately prepared to discontinue benzodiazepines, particularly when patients perceive benefit from them. As such, user-friendly tools and specific strategies for weighing risks vs benefits are critical for clinicians.

PUTTING KNOWLEDGE INTO PRACTICE

How do we translate knowledge into practice with regard to deprescribing potentially inappropriate medications in older patients—or prescribing drugs only if appropriate in the first place?

An opportunity arises when patients are in the hospital. Taking a medication history on admission and matching medications with indications are key starting points. Clinical pharmacists can help screen for side effects and potential interactions and can provide deprescribing recommendations. Meticulous discharge medication reconciliation, patient education, and communication of the updated medication list to the outpatient provider are central to ensuring that patients adhere to medication adjustments after they go home.

A MATTER OF BALANCE

Another factor to consider is the patient’s physiologic age compared with his or her chronologic age. If a patient has multiple comorbidities, frailty, limited life expectancy, or poor renal function, we may consider her older than her chronologic age. In this case, a drug’s risks may outweigh its benefit, which is something to be discussed. On the other hand, a high-functioning and relatively healthy elderly patient may be a candidate for medications known to reduce the risk of death or control a chronic disease better. Incorporating a patient’s goals of care and using shared decision-making are also likely to yield an optimal medication regimen.

Smartphone apps and resources embedded in electronic health records provide additional decision support. Used when prescribing or reconciling medications, these supplemental brains offer instant feedback and information on dose adjustments, drug interactions, clinical guidelines, and even potentially inappropriate medications. While the impacts of these electronic tools on prescribing patterns and outcomes in geriatric populations remain unclear, new ones are being developed and studied.⁶ This may be the most promising way to translate knowledge into practice, as it is more easily integrated with existing clinician workflows.

AN OPPORTUNITY TO IMPROVE

There is significant opportunity to reduce polypharmacy and optimize medication prescribing practices for older adults. Awareness of potentially inappropriate medications and clinical situations in which the use of certain classes of medications should be minimized is the first step in addressing this problem. Using tools such as the STOPP/START criteria, reviewing medications at critical transition points, prioritizing patient function and goals, and using electronic clinical decision support should aid prescribing decisions.

Whenever possible, collaborating with other care team members such as pharmacists may increase efficiency and effectiveness of medication management. Ultimately, inclusion of more older adults in clinical trials may provide data-driven guidance for weighing risks and benefits. Finally, further study of the effects of deprescribing on clinical outcomes may be the missing piece to help clinicians and patients find balance in prescription management.

A 66-year-old man presented to the hospital with 3 days of nausea, vomiting, and abdominal pain. He had come to the emergency department several times during this period, but the cause of his symptoms had not been determined.

The patient was successfully treated with urgent right hemicolectomy.

THE CHILAIDITI SIGN AND SYNDROME

The Chilaiditi sign is an infrequent anomaly found incidentally on chest or abdominal radiography as a colonic interposition between the liver and right hemidiaphragm.¹ It is often asymptomatic but is sometimes accompanied by nausea, vomiting, abdominal pain, and constipation, ie, Chilaiditi syndrome.

Generally, after conservative treatment with fasting and pain control, symptoms may subside and follow-up should be sufficient. However, nasogastric decompression and laxatives are occasionally needed and are often effective in patients with Chilaiditi syndrome. Urgent abdominal surgery is indicated for patients with symptoms of volvulus of the colon, stomach, or small intestine.²

DISTINGUISHING CHARACTERISTICS

The Chilaiditi sign is often confused with pneumoperitoneum, which usually requires urgent abdominal surgery. But the presence of haustration or valvulae conniventes (folds in the small bowel mucosa) in the hepatodiaphragmatic space helps distinguish between intraluminal gas and free air. If the patient presents with abdominal pain without signs of peritonitis, and if imaging indicates the Chilaiditi sign, then supplementary imaging (eg, decubitus radiography, chest CT, abdominal CT) is recommended to make the definitive diagnosis and to avoid unnecessary surgery.

Gas under the diaphragm on standing chest radiography without signs of peritonitis may also be seen after laparotomy and after scuba diving as well as in cases of biliary enteric fistula, incompetent sphincter of Oddi, gallstone ileus, and pneumatosis cystoides intestinalis. The incidence rate of the Chilaiditi sign detected by radiography is between 0.025% and 0.28%.³

PREDISPOSING FACTORS

The cause of the Chilaiditi sign remains unknown. Predisposing factors can be categorized as diaphragmatic (diaphragmatic thinning, phrenic nerve injury, expanded thoracic cavity), intestinal (megacolon, increased intra-abdominal pressure), and hepatic (hepatic atrophy, cirrhosis, ascites).

In healthy people, Chilaiditi syndrome is usually attributed to a congenital abnormal lengthening of the colon or to undue looseness of ligaments of the colon and liver.

Recognizing the Chilaiditi sign is particularly important in patients scheduled to undergo a percutaneous transhepatic procedure or colonoscopic examination, as these procedures increase the risk of perforation.

A 66-year-old man presented to the hospital with 3 days of nausea, vomiting, and abdominal pain. He had come to the emergency department several times during this period, but the cause of his symptoms had not been determined.

The patient was successfully treated with urgent right hemicolectomy.

THE CHILAIDITI SIGN AND SYNDROME

The Chilaiditi sign is an infrequent anomaly found incidentally on chest or abdominal radiography as a colonic interposition between the liver and right hemidiaphragm.¹ It is often asymptomatic but is sometimes accompanied by nausea, vomiting, abdominal pain, and constipation, ie, Chilaiditi syndrome.

Generally, after conservative treatment with fasting and pain control, symptoms may subside and follow-up should be sufficient. However, nasogastric decompression and laxatives are occasionally needed and are often effective in patients with Chilaiditi syndrome. Urgent abdominal surgery is indicated for patients with symptoms of volvulus of the colon, stomach, or small intestine.²

DISTINGUISHING CHARACTERISTICS

The Chilaiditi sign is often confused with pneumoperitoneum, which usually requires urgent abdominal surgery. But the presence of haustration or valvulae conniventes (folds in the small bowel mucosa) in the hepatodiaphragmatic space helps distinguish between intraluminal gas and free air. If the patient presents with abdominal pain without signs of peritonitis, and if imaging indicates the Chilaiditi sign, then supplementary imaging (eg, decubitus radiography, chest CT, abdominal CT) is recommended to make the definitive diagnosis and to avoid unnecessary surgery.

Gas under the diaphragm on standing chest radiography without signs of peritonitis may also be seen after laparotomy and after scuba diving as well as in cases of biliary enteric fistula, incompetent sphincter of Oddi, gallstone ileus, and pneumatosis cystoides intestinalis. The incidence rate of the Chilaiditi sign detected by radiography is between 0.025% and 0.28%.³

PREDISPOSING FACTORS

The cause of the Chilaiditi sign remains unknown. Predisposing factors can be categorized as diaphragmatic (diaphragmatic thinning, phrenic nerve injury, expanded thoracic cavity), intestinal (megacolon, increased intra-abdominal pressure), and hepatic (hepatic atrophy, cirrhosis, ascites).

In healthy people, Chilaiditi syndrome is usually attributed to a congenital abnormal lengthening of the colon or to undue looseness of ligaments of the colon and liver.

Recognizing the Chilaiditi sign is particularly important in patients scheduled to undergo a percutaneous transhepatic procedure or colonoscopic examination, as these procedures increase the risk of perforation.

A 66-year-old man presented to the hospital with 3 days of nausea, vomiting, and abdominal pain. He had come to the emergency department several times during this period, but the cause of his symptoms had not been determined.

The patient was successfully treated with urgent right hemicolectomy.

THE CHILAIDITI SIGN AND SYNDROME

The Chilaiditi sign is an infrequent anomaly found incidentally on chest or abdominal radiography as a colonic interposition between the liver and right hemidiaphragm.¹ It is often asymptomatic but is sometimes accompanied by nausea, vomiting, abdominal pain, and constipation, ie, Chilaiditi syndrome.

Generally, after conservative treatment with fasting and pain control, symptoms may subside and follow-up should be sufficient. However, nasogastric decompression and laxatives are occasionally needed and are often effective in patients with Chilaiditi syndrome. Urgent abdominal surgery is indicated for patients with symptoms of volvulus of the colon, stomach, or small intestine.²

DISTINGUISHING CHARACTERISTICS

The Chilaiditi sign is often confused with pneumoperitoneum, which usually requires urgent abdominal surgery. But the presence of haustration or valvulae conniventes (folds in the small bowel mucosa) in the hepatodiaphragmatic space helps distinguish between intraluminal gas and free air. If the patient presents with abdominal pain without signs of peritonitis, and if imaging indicates the Chilaiditi sign, then supplementary imaging (eg, decubitus radiography, chest CT, abdominal CT) is recommended to make the definitive diagnosis and to avoid unnecessary surgery.

Gas under the diaphragm on standing chest radiography without signs of peritonitis may also be seen after laparotomy and after scuba diving as well as in cases of biliary enteric fistula, incompetent sphincter of Oddi, gallstone ileus, and pneumatosis cystoides intestinalis. The incidence rate of the Chilaiditi sign detected by radiography is between 0.025% and 0.28%.³

PREDISPOSING FACTORS

The cause of the Chilaiditi sign remains unknown. Predisposing factors can be categorized as diaphragmatic (diaphragmatic thinning, phrenic nerve injury, expanded thoracic cavity), intestinal (megacolon, increased intra-abdominal pressure), and hepatic (hepatic atrophy, cirrhosis, ascites).

In healthy people, Chilaiditi syndrome is usually attributed to a congenital abnormal lengthening of the colon or to undue looseness of ligaments of the colon and liver.

Recognizing the Chilaiditi sign is particularly important in patients scheduled to undergo a percutaneous transhepatic procedure or colonoscopic examination, as these procedures increase the risk of perforation.

A 66-year-old woman presented to the   emergency room with pain and bluish discoloration of her left great toe for the past 2 days. She had no history of trauma to the toe or of peripheral vascular disease. She also complained of intermittent abdominal pain for the last 5 months and unintentional weight loss of 8 pounds.

Vascular disease was suspected, and the patient was started on systemic anticoagulation. However, chest computed tomography (CT) and conventional angiography showed no aortic disease; vessels were of normal caliber, and no distal filling defects were noted.

A routine evaluation with complete blood cell count, peripheral smear, renal function testing, and urinalysis with eosinophil smear was also unrevealing. An extensive investigation followed, including serum complement testing, assays for antinuclear, antineutrophil cytoplasmic, and anti-Sjögren syndrome A and B antibodies, cryoglobulin testing, and hepatitis B and C virus serology, as well as screening for syphilis and lupus anticoagulant, anticardiolipin, and beta-2 glycoprotein antibodies. The results for all these tests were also unremarkable.

Results of coagulation testing and venous duplex ultrasonography of the legs were normal, and electrocardiography and echocardiography showed no signs of valvular vegetation, myxoma, or patent foramen ovale.

Given our patient’s age, abdominal pain, and weight loss but negative vascular evaluation, we considered a diagnosis of paraneoplastic acral vascular syndrome. Abdominal CT revealed a tumor of the pancreatic head with multiple liver lesions, and cytologic study confirmed pancreatic adenocarcinoma.

DISTANT MARKERS OF MALIGNANCY

Causes of blue toe syndrome to consider in the differential diagnosis include arterial thromboembolism, vasoconstrictive drug use or disorders, vasculitis, and venous thrombosis.¹ These are common and deserve prompt investigation. However, if they are ruled out, a peripheral acral vascular syndrome should be considered. These syndromes present as Raynaud phenomenon, gangrene, or acrocyanosis of the fingers or toes with underlying neoplasia. Unusual features such as sudden onset in a patient over age 50, an acral distribution, and associated symptoms such as unrelated pain and weight loss should spark concern for underlying malignancy.

Paraneoplastic syndromes are defined as signs and symptoms that present distant from the site of malignancy. Dermatoses as markers of internal malignancy are well-established but perplexing clinical entities whose exact causes remain unknown.²

Paraneoplastic dermatoses are well recognized as harbingers of metastatic disease.^5,6 Our patient’s story demonstrates need for a thorough diagnostic investigation.

A 66-year-old woman presented to the   emergency room with pain and bluish discoloration of her left great toe for the past 2 days. She had no history of trauma to the toe or of peripheral vascular disease. She also complained of intermittent abdominal pain for the last 5 months and unintentional weight loss of 8 pounds.

Vascular disease was suspected, and the patient was started on systemic anticoagulation. However, chest computed tomography (CT) and conventional angiography showed no aortic disease; vessels were of normal caliber, and no distal filling defects were noted.

A routine evaluation with complete blood cell count, peripheral smear, renal function testing, and urinalysis with eosinophil smear was also unrevealing. An extensive investigation followed, including serum complement testing, assays for antinuclear, antineutrophil cytoplasmic, and anti-Sjögren syndrome A and B antibodies, cryoglobulin testing, and hepatitis B and C virus serology, as well as screening for syphilis and lupus anticoagulant, anticardiolipin, and beta-2 glycoprotein antibodies. The results for all these tests were also unremarkable.

Results of coagulation testing and venous duplex ultrasonography of the legs were normal, and electrocardiography and echocardiography showed no signs of valvular vegetation, myxoma, or patent foramen ovale.

Given our patient’s age, abdominal pain, and weight loss but negative vascular evaluation, we considered a diagnosis of paraneoplastic acral vascular syndrome. Abdominal CT revealed a tumor of the pancreatic head with multiple liver lesions, and cytologic study confirmed pancreatic adenocarcinoma.

DISTANT MARKERS OF MALIGNANCY

Causes of blue toe syndrome to consider in the differential diagnosis include arterial thromboembolism, vasoconstrictive drug use or disorders, vasculitis, and venous thrombosis.¹ These are common and deserve prompt investigation. However, if they are ruled out, a peripheral acral vascular syndrome should be considered. These syndromes present as Raynaud phenomenon, gangrene, or acrocyanosis of the fingers or toes with underlying neoplasia. Unusual features such as sudden onset in a patient over age 50, an acral distribution, and associated symptoms such as unrelated pain and weight loss should spark concern for underlying malignancy.

Paraneoplastic syndromes are defined as signs and symptoms that present distant from the site of malignancy. Dermatoses as markers of internal malignancy are well-established but perplexing clinical entities whose exact causes remain unknown.²

Paraneoplastic dermatoses are well recognized as harbingers of metastatic disease.^5,6 Our patient’s story demonstrates need for a thorough diagnostic investigation.

A 66-year-old woman presented to the   emergency room with pain and bluish discoloration of her left great toe for the past 2 days. She had no history of trauma to the toe or of peripheral vascular disease. She also complained of intermittent abdominal pain for the last 5 months and unintentional weight loss of 8 pounds.

Vascular disease was suspected, and the patient was started on systemic anticoagulation. However, chest computed tomography (CT) and conventional angiography showed no aortic disease; vessels were of normal caliber, and no distal filling defects were noted.

A routine evaluation with complete blood cell count, peripheral smear, renal function testing, and urinalysis with eosinophil smear was also unrevealing. An extensive investigation followed, including serum complement testing, assays for antinuclear, antineutrophil cytoplasmic, and anti-Sjögren syndrome A and B antibodies, cryoglobulin testing, and hepatitis B and C virus serology, as well as screening for syphilis and lupus anticoagulant, anticardiolipin, and beta-2 glycoprotein antibodies. The results for all these tests were also unremarkable.

Results of coagulation testing and venous duplex ultrasonography of the legs were normal, and electrocardiography and echocardiography showed no signs of valvular vegetation, myxoma, or patent foramen ovale.

Given our patient’s age, abdominal pain, and weight loss but negative vascular evaluation, we considered a diagnosis of paraneoplastic acral vascular syndrome. Abdominal CT revealed a tumor of the pancreatic head with multiple liver lesions, and cytologic study confirmed pancreatic adenocarcinoma.

DISTANT MARKERS OF MALIGNANCY

Causes of blue toe syndrome to consider in the differential diagnosis include arterial thromboembolism, vasoconstrictive drug use or disorders, vasculitis, and venous thrombosis.¹ These are common and deserve prompt investigation. However, if they are ruled out, a peripheral acral vascular syndrome should be considered. These syndromes present as Raynaud phenomenon, gangrene, or acrocyanosis of the fingers or toes with underlying neoplasia. Unusual features such as sudden onset in a patient over age 50, an acral distribution, and associated symptoms such as unrelated pain and weight loss should spark concern for underlying malignancy.

Paraneoplastic syndromes are defined as signs and symptoms that present distant from the site of malignancy. Dermatoses as markers of internal malignancy are well-established but perplexing clinical entities whose exact causes remain unknown.²

Paraneoplastic dermatoses are well recognized as harbingers of metastatic disease.^5,6 Our patient’s story demonstrates need for a thorough diagnostic investigation.

A 75-year-old man presented to the emergency department for evaluation of abdominal pain. He had stage 3 chronic obstructive pulmonary disease (COPD), with a forced expiratory volume in 1 second of 33%.

PREVIOUS HOSPITALIZATION

Aside from his COPD, he had been healthy until 1 month earlier, when he had been hospitalized because of shortness of breath and chest pressure with exertion. His troponin T level had been elevated, peaking at 0.117 ng/mL (reference range 0–0.029).

Left heart catheterization had shown no significant coronary artery disease. A myocardial bridge of the distal left anterior descending coronary artery had been seen, so that the artery appeared to be narrowed by 50% to 60% with ventricular contraction. But this was not thought to have been the cause of his presentation.

On discharge, he required oxygen 4 L/min by nasal cannula. Previously, he had not needed supplemental oxygen.

CURRENT PRESENTATION

The patient described persistent and severe periumbilical abdominal pain during the previous day. It was not associated with eating, and he denied diarrhea, constipation, hematemesis, hematochezia, bright red blood per rectum, or melena. He continued to describe persistent shortness of breath and pleuritic chest pain. His vital signs were as follows:

• Heart rate 104 beats per minute
• Respiratory rate 16 to 20 breaths per minute
• Blood pressure 101–142/62–84 mm Hg
• Oxygen saturation 78% on room air.

His laboratory findings on presentation are shown in Table 1, and his electrocardiogram is shown in Figure 1.

WHAT DOES HIS ELECTROCARDIOGRAM SHOW?

1. Which of the following is the most accurate description of this patient’s electrocardiogram?

• Sinus tachycardia, peaked P waves (P pulmonale) in lead II, and T-wave inversions in the right precordial leads
• Sinus tachycardia and left bundle branch block
• Sinus tachycardia and poor R-wave progression
• Sinus tachycardia and ST elevation in the precordial leads

Our patient’s electrocardiogram shows sinus tachycardia, P pulmonale, T-wave inversion in the right precordial leads (V₁–V₃), and biphasic T waves in lead V_4,, which suggest right ventricular strain.

The rhythm most commonly seen in patients with pulmonary embolism is sinus tachycardia, followed by nonspecific ST-segment or T-wave abnormalities. In one series of patients with acute pulmonary embolism, the classic findings of P pulmonale, right ventricular hypertrophy, right axis deviation, and right bundle branch block were rare (< 6%).¹ Thus, these classic findings are not sensitive for the diagnosis of pulmonary embolism, and their absence does not rule it out.

Further studies for our patient

Transthoracic echocardiography was performed to look for evidence of right ventricular strain secondary to the pulmonary embolism.

2. Which of the following findings on transthoracic echocardiography would not suggest acute pulmonary embolism?

• Midright ventricular wall hypokinesis with apical sparing
• Severe tricuspid regurgitation
• Left ventricular dilation
• Lack of respiratory variation of the inferior vena cava
• Septal wall motion toward the left ventricle

Left ventricular dilation does not suggest acute pulmonary embolism. Echocardiograms of patients with acute submassive pulmonary embolism typically show evidence of right ventricular strain, such as the other entities listed above (midright ventricular hypokinesis with apical sparing, severe tricuspid regurgitation, lack of respiratory variation of the inferior vena cava, and septal wall motion toward the left ventricle).

The degree of right ventricular dysfunction is related to the extent of acute pulmonary vascular occlusion and aids in risk-stratification of patients with acute pulmonary embolism. Midright ventricular wall hypokinesis with apical sparing has been termed the McConnell sign.²

In our patient, transthoracic echocardiography showed:

• Normal left ventricular ejection fraction
• Mild diastolic dysfunction
• Right ventricular dilation with moderately decreased right ventricular systolic function and apical sparing
• Right ventricular systolic pressure 54 mm Hg, consistent with moderate pulmonary hypertension
• Right atrial pressure 10 mm Hg
• No inspiratory collapse of a dilated inferior vena cava
• Mild tricuspid valve regurgitation.

CLASSIFICATION OF ACUTE PULMONARY EMBOLISM

3. Given the above information, how would you classify the patient’s pulmonary embolism?

• Massive
• Submassive
• Low-risk
• Clinically stable

The patient’s pulmonary embolism is submassive.

Historically, the classification of pulmonary embolism was determined by the angiographic thrombus burden. However, this has limited utility because clinical factors (eg, hypotension on initial presentation) have been shown to be better predictors of short-term mortality risk.³

Our patient is characterized as having a submassive pulmonary embolism based on elevated biomarkers (troponin T, N-terminal pro-B-type natriuretic peptide) and right ventricular dysfunction in the absence of hypotension.

ULTRASONOGRAPHY FOR DIAGNOSIS OF DEEP VEIN THROMBOSIS

Venous duplex ultrasonography has become the standard for diagnosis of lower extremity deep vein thrombosis. However, its quality and diagnostic accuracy depend on the skill of the person performing the examination. It is further limited by certain patient characteristics, including severe obesity, edema, and wounds and dressings at the site being examined.⁵

Our patient underwent duplex ultrasonography of the lower extremities, which demonstrated acute proximal and calf deep vein thrombosis in the right femoral, popliteal, and peroneal veins and no deep vein thrombosis in the left leg.

Multiple models exist to estimate the risk of complications in patients with acute pulmonary embolism.

The Bova score⁶ is based on the following factors:

• Systolic blood pressure 90–100 mm Hg (2 points) (patients with systolic blood pressure lower than 90 mm Hg were excluded from the study from which this score was derived)
• Cardiac troponin elevation (2 points)
• Right ventricular dysfunction on echocardiography or computed tomography (2 points)
• Heart rate 100 beats/min or greater (1 point).

A total score of 0, 1, or 2 (stage I) denotes low risk, 3 or 4 points (stage II) intermediate risk, and more than 4 points (stage III) high risk.

The PESI score (Pulmonary Embolism Severity Index)⁷ is based on:

• Age (1 point per year)
• Sex (10 points for being male)
• Heart rate 110 per minute or greater (20 points)
• Cancer (30 points)
• Heart failure (10 points)
• Chronic lung disease (10 points)
• Systolic blood pressure less than 100 mm Hg (30 points)
• Respiratory rate at least 30 per minute (20 points)
• Temperature less than 36ºC (20 points)
• Altered mental status (60 points)
• Arterial oxygen saturation less than 90% (20 points).

The total score is broken down into 5 classes: I (< 65 points), II (65–85), III (86–105), IV (106–125), and V (> 126). Classes I and II are low risk, and the higher ones are high risk.

The simplified PESI score⁸ was developed to more rapidly risk-stratify patients and has been found to be similar to the PESI score in prognostic accuracy. Patients get 1 point for each of the following:

• Age over 80
• Cancer
• Chronic cardiopulmonary disease (heart failure or chronic lung disease)
• Heart rate 110 per minute or greater
• Systolic blood pressure less than 100 mm Hg
• Arterial oxygen saturation less than 90%.

A total score of 0 is low risk; anything higher is high risk.

Back to our patient

Our patient had proximal and calf deep vein thrombosis of the right leg, bilateral submassive pulmonary emboli with associated biomarker elevation and right ventricular dysfunction, and left renal artery thrombosis with infarction. Using the PESI score, his risk of death in the next 30 days was 13.7% and his 30-day risk of a complicated course was 27%. Using the Bova score, his 30-day risk of death was 15.5% and his 30-day risk of a complicated course was 29.2%.^6,7

Notably, the patient’s right ventricular function had also been impaired on the echocardiogram performed during his admission 1 month previously. On transthoracic echocardiography during the current admission, the patient was found to have a similar degree of right ventricular dysfunction. This finding, along with the oxygen requirement that developed during the earlier admission, suggested that his pulmonary embolism may have been subacute and that the diagnosis may have been missed during the earlier hospital stay.

The patient was treated with unfractionated heparin. After the hospital’s multidisciplinary pulmonary embolism response team discussed and weighed the above factors, they recommended to not pursue thrombolytic therapy or inferior vena cava filter placement.

Of note, the patient’s pulses in the left lower extremity continued to be weak but palpable, and the left leg was cooler to touch than the right leg.

ASSESSING PERIPHERAL ARTERY DISEASE

4. How should the finding of weak pulses in this patient’s left leg be initially investigated?

• Computed tomographic angiography with runoff
• Ankle-brachial indices with pulse-volume recordings
• Arterial duplex ultrasonography
• Magnetic resonance angiography of the lower extremities

The ankle-brachial index is the initial diagnostic test for assessment of pulse abnormalities and for diagnosis of lower-extremity peripheral artery disease. It is calculated by dividing the higher of the ankle systolic pressures (posterior tibial or dorsalis pedis) by the higher of the 2 brachial pressures (left or right).⁹ Normal values are between 1.00 and 1.40.

Ankle-brachial indices in our patient

Our patient underwent measurement of his brachial, dorsalis pedis, and posterior tibial artery systolic pressures using blood pressure cuffs and continuous-wave Doppler. Ankle pulse-volume recordings were also obtained.

Given the patient’s poor renal function and concern for acute renal infarction, we thought it best to avoid iodinated or gadolinium contrast, such as with magnetic resonance or computed tomographic angiography.

Segmental leg pressures and pulse-volume recordings can be performed to help localize the level of arterial disease in the extremities, but were not done in this case because of the extensive deep vein thrombosis in the right leg.^10,11

Arterial ultrasonography in our patient

Arterial duplex ultrasonography was performed to help determine the location of arterial disease. It showed patent arteries in the right leg. In the left lower extremity there was slow, monophasic blood flow in the distal superficial femoral artery. The popliteal artery was occluded. The posterior tibial artery was occluded at the origin, with reconstitution distally. The peroneal artery was occluded throughout. The anterior tibial artery was patent throughout. The ultrasonographic findings were thought to be suspicious for arterial thromboembolism.

5. Given that the patient had both arterial thrombosis (renal artery, lower-extremity arteries) and venous thromboembolism (deep vein thrombosis and pulmonary embolism), which of the following would be included in the differential diagnosis?

• Antiphospholipid antibody syndrome
• Protein C or protein S deficiency
• Malignancy
• Paradoxical embolization
• Factor V Leiden mutation

Correct answers include antiphospholipid antibody syndrome, malignancy, and paradoxical embolization.

The differential diagnosis for concomitant venous and arterial thrombosis is broad,¹² and includes the following:

• Structural factors: patent foramen ovale, popliteal artery aneurysm
• Malignancy
• Inflammatory diseases: Behçet disease, Buerger disease, inflammatory bowel disease, antiphospholipid antibody syndrome, elevated lipoprotein(a), elevated homocysteine
• Hematologic diseases: myelodysplastic syndrome, disseminated intravascular coagulation, paroxysmal nocturnal hemoglobinuria, heparin-induced thrombocytopenia.

Traditional risk factors for venous thromboembolism include protein C deficiency, protein S deficiency, factor V Leiden mutation, the prothrombin G20210A gene mutation, and others. These are relatively minor risk factors for venous thrombosis and do not pose a risk for arterial thrombosis.¹² In contrast, antiphospholipid antibody syndrome and malignancy pose a risk for both venous and arterial thrombosis. Paradoxical embolism is a mechanism by which arterial thrombosis (emboli) can develop in the setting of existing venous thrombosis.¹²

Our patient underwent testing for antiphospholipid antibodies and lupus anticoagulant, and he was encouraged to undergo age-appropriate cancer screening as an outpatient.¹²

ANTIPHOSPHOLIPID ANTIBODY SYNDROME

Antiphospholipid antibody syndrome is defined by both clinical and laboratory criteria. Clinical symptoms include vascular thrombosis (arterial, venous, or both) and pregnancy-related complications.¹³

Laboratory criteria require the presence of antiphospholipid antibodies or lupus anticoagulant. These must be confirmed with repeat testing in 12 weeks. Antiphospholipid antibodies are detected by an enzyme-linked immunosorbent assay; laboratory assessment for the presence of lupus anticoagulant is a stepwise process and relies on 4 criteria:

• There should be prolongation of a phospholipid-dependent clotting test (eg, activated partial thromboplastin time, dilute Russell viper venom time test).
• There must be evidence of an inhibitory activity with mixing study.
• The inhibitor must exhibit phospholipid dependence; that is, with more phospholipid there is shortening of clotting time.
• Specific inhibitors must be excluded, including factor VIII and anticoagulant drugs such as heparin.^14–17

From Houghton DE, Moll S. Antiphospholipid antibodies. Vasc Med 2017; 22:545–550.

Clinically, one should consider antiphospholipid antibody syndrome in patients who have arterial thrombosis, a history of pregnancy morbidity, or unexplained prolongation of activated partial thromboplastin time.¹³

Antiphospholipid antibodies may be present in up to a quarter of patients with venous thromboembolism, but it is persistent positivity of antibody assays that is associated with increased future risk of venous thromboembolism.¹⁹ Of note, the risk of venous thromboembolism in patients with confirmed antiphospholipid antibody syndrome is 10 times higher than in the general population.²⁰

ANTIPHOSPHOLIPID ANTIBODIES ARE NOT ALL THE SAME

6. Which of the following antiphospholipid antibodies have not been associated with an increased thrombotic risk?

• Anti-beta-2-glycoprotein I IgG
• Lupus anticoagulant
• Antiphosphatidylserine
• Anticardiolipin IgM
• Anticardiolipin IgG

The correct answer is antiphosphatidylserine.¹⁵

Antiphospholipid antibodies are directed against a portion of select plasma proteins that are uncovered upon phospholipid binding. While lupus anticoagulant, anti-beta-2-glycoprotein I, and anticardiolipin antibodies are associated with thrombosis, antiprothrombin antibodies (including antiprothrombin and antiphosphatidylserine antibodies) are not.^15,21

Patent foramen ovale, a communication between the right and left atrium in the interatrial septum, is associated with an increased risk of paradoxical embolization. The prevalence of patent foramen ovale is estimated to be 27% to 29% in the general population.²² Noncerebral systemic paradoxical embolism occurs less frequently than cerebral embolism, accounting for approximately 5% to 10% of paradoxical emboli.²²

To evaluate for patent foramen ovale, transthoracic echocardiography is performed with a bubble (agitated saline contrast) study to assess for interatrial shunting. Transesophageal echocardiography or transcranial Doppler bubble studies may also be performed.

Although patent foramen ovale is most commonly associated with cerebral embolism, peripheral emboli can occur. Some research suggests that this may be a more common cause of arterial thromboembolism in younger patients. There have also been reports of other sites of systemic embolization, including the renal artery.¹²

Back to our patient

Initial antiphospholipid antibody testing was positive for lupus anticoagulant. Anticardiolipin and anti-beta-2-glycoprotein I antibodies were not detected.

Transesophageal echocardiography revealed a patent foramen ovale with a highly mobile atrial septum (atrial septal aneurysm).

The patient was treated with intravenous unfractionated heparin with bridging to warfarin with a target international normalized ratio (INR) of 2 to 3. His renal artery infarction and his lower-extremity arterial thromboembolic event were conservatively managed. His respiratory status improved, and he no longer required supplemental oxygen. His creatinine peaked at 1.7 mg/dL during his admission and improved to 1.2 mg/dL before he was discharged.

At follow-up, repeat echocardiography showed that his right ventricular systolic pressure had improved (decreased) to 37 mm Hg from 54 mm Hg. Repeat confirmatory testing was positive for lupus anticoagulant 12 weeks later. He has been maintained on warfarin with an INR goal of 2 to 3 as well as low-dose aspirin with plans for long-term anticoagulation. We decided to keep the patient on anticoagulation indefinitely with warfarin; he was not a candidate for a direct oral anticoagulant, given limited data on the use of these agents in the setting of lupus anticoagulant and antiphospholipid antibody syndrome.

SUMMARY OF CASE

In summary, this patient was a 75-year-old man with COPD who presented with abdominal pain. He was noted to have a left renal infarction, extensive unprovoked lower-extremity deep vein thrombosis with pulmonary emboli, and lower limb arterial thromboembolism.

He also had an underlying hypercoagulable state—antiphospholipid antibody syndrome—that predisposed him to both arterial and venous thrombosis. He was ultimately found to have a patent foramen ovale, which further increased the risk of arterial thrombosis by facilitating paradoxical embolization of venous thrombi. It is not certain whether the renal infarction and leg artery thrombi were due to paradoxical embolism or to in situ thrombosis, but we believe that it was most likely paradoxical embolization.        

A 75-year-old man presented to the emergency department for evaluation of abdominal pain. He had stage 3 chronic obstructive pulmonary disease (COPD), with a forced expiratory volume in 1 second of 33%.

PREVIOUS HOSPITALIZATION

Aside from his COPD, he had been healthy until 1 month earlier, when he had been hospitalized because of shortness of breath and chest pressure with exertion. His troponin T level had been elevated, peaking at 0.117 ng/mL (reference range 0–0.029).

Left heart catheterization had shown no significant coronary artery disease. A myocardial bridge of the distal left anterior descending coronary artery had been seen, so that the artery appeared to be narrowed by 50% to 60% with ventricular contraction. But this was not thought to have been the cause of his presentation.

On discharge, he required oxygen 4 L/min by nasal cannula. Previously, he had not needed supplemental oxygen.

CURRENT PRESENTATION

The patient described persistent and severe periumbilical abdominal pain during the previous day. It was not associated with eating, and he denied diarrhea, constipation, hematemesis, hematochezia, bright red blood per rectum, or melena. He continued to describe persistent shortness of breath and pleuritic chest pain. His vital signs were as follows:

• Heart rate 104 beats per minute
• Respiratory rate 16 to 20 breaths per minute
• Blood pressure 101–142/62–84 mm Hg
• Oxygen saturation 78% on room air.