Drug Therapy

Cigna and Express Scripts have announced a new patient assistance program aimed at ensuring that patients with diabetes are not prevented by cost-related issues from getting access to insulin.

KatarzynaBialasiewicz/Thinkstock

The new program, open to Cigna members who are covered in commercial plans, would cap out-of-pocket costs for a 30-day supply of insulin at $25. For plan members, the only eligibility requirement is having an out-of-pocket cost higher than $25, according to a press release.

For a member to participate in the program, the plan administrator at the member’s place of employment has to opt in to it. There are no eligibility requirements imposed on the employer, other than a willingness to opt in.

A spokeswoman for Express Scripts said that there is no charge to sign up for the program, and most plans will not see an additional cost to get the copayment to $25 for the patient.

The announcement comes in the wake of the first of two hearings by the House Committee on Energy & Commerce aimed at understanding why insulin prices have spiked in recent years. The first hearing, held on April 2, examined the impact that the high list price of insulin is having on patients, and how out-of-pocket expenses are limiting access to this life-saving drug. The second hearing, expected to occur during the week of April 8 (the date had not been scheduled as of press time), will bring together various players in the supply chain, including the three major manufacturers of insulin.

“We are confident that our new program will remove cost as a barrier for people in participating plans who need insulin,” Steve Miller, MD, executive vice president and chief clinical officer at Cigna, said in a statement.

The Express Scripts spokeswoman noted that there were more than 700,000 people in a commercially insured plan across Cigna and Express Scripts who had a claim for insulin in 2018. The average out-of-pocket cost of a 30-day supply of insulin in 2018 across this population was $41.50.

[email protected]

Cigna and Express Scripts have announced a new patient assistance program aimed at ensuring that patients with diabetes are not prevented by cost-related issues from getting access to insulin.

KatarzynaBialasiewicz/Thinkstock

The new program, open to Cigna members who are covered in commercial plans, would cap out-of-pocket costs for a 30-day supply of insulin at $25. For plan members, the only eligibility requirement is having an out-of-pocket cost higher than $25, according to a press release.

For a member to participate in the program, the plan administrator at the member’s place of employment has to opt in to it. There are no eligibility requirements imposed on the employer, other than a willingness to opt in.

A spokeswoman for Express Scripts said that there is no charge to sign up for the program, and most plans will not see an additional cost to get the copayment to $25 for the patient.

The announcement comes in the wake of the first of two hearings by the House Committee on Energy & Commerce aimed at understanding why insulin prices have spiked in recent years. The first hearing, held on April 2, examined the impact that the high list price of insulin is having on patients, and how out-of-pocket expenses are limiting access to this life-saving drug. The second hearing, expected to occur during the week of April 8 (the date had not been scheduled as of press time), will bring together various players in the supply chain, including the three major manufacturers of insulin.

“We are confident that our new program will remove cost as a barrier for people in participating plans who need insulin,” Steve Miller, MD, executive vice president and chief clinical officer at Cigna, said in a statement.

The Express Scripts spokeswoman noted that there were more than 700,000 people in a commercially insured plan across Cigna and Express Scripts who had a claim for insulin in 2018. The average out-of-pocket cost of a 30-day supply of insulin in 2018 across this population was $41.50.

[email protected]

Cigna and Express Scripts have announced a new patient assistance program aimed at ensuring that patients with diabetes are not prevented by cost-related issues from getting access to insulin.

KatarzynaBialasiewicz/Thinkstock

The new program, open to Cigna members who are covered in commercial plans, would cap out-of-pocket costs for a 30-day supply of insulin at $25. For plan members, the only eligibility requirement is having an out-of-pocket cost higher than $25, according to a press release.

For a member to participate in the program, the plan administrator at the member’s place of employment has to opt in to it. There are no eligibility requirements imposed on the employer, other than a willingness to opt in.

A spokeswoman for Express Scripts said that there is no charge to sign up for the program, and most plans will not see an additional cost to get the copayment to $25 for the patient.

The announcement comes in the wake of the first of two hearings by the House Committee on Energy & Commerce aimed at understanding why insulin prices have spiked in recent years. The first hearing, held on April 2, examined the impact that the high list price of insulin is having on patients, and how out-of-pocket expenses are limiting access to this life-saving drug. The second hearing, expected to occur during the week of April 8 (the date had not been scheduled as of press time), will bring together various players in the supply chain, including the three major manufacturers of insulin.

“We are confident that our new program will remove cost as a barrier for people in participating plans who need insulin,” Steve Miller, MD, executive vice president and chief clinical officer at Cigna, said in a statement.

The Express Scripts spokeswoman noted that there were more than 700,000 people in a commercially insured plan across Cigna and Express Scripts who had a claim for insulin in 2018. The average out-of-pocket cost of a 30-day supply of insulin in 2018 across this population was $41.50.

[email protected]

Which patients are likely to benefit from using aspirin for primary prevention? In this article, we review the evidence to date, summarized for primary care settings in guidelines issued by the US Preventive Services Task Force (USPSTF). We supplement this summary with a rundown of the risks associated with aspirin use. And then we wrap up by identifying a clinical decision tool that is available to help make personalized decisions in a busy clinic setting, where determining an individual’s potential cardiovascular benefits and bleeding risk can be challenging.

The “roadmap” from the guidelines. In 2014, after performing a review of the literature, the US Food and Drug Administration recommended against the routine use of aspirin for primary prevention of cardiovascular disease (CVD).¹ In 2016, the USPSTF published 4 separate systematic reviews along with a decision analysis using a microsimulation model, which informed their position statement on aspirin for primary prevention.^2-6 These USPSTF reviews and recommendations incorporated both CVD and colorectal cancer (CRC) benefits with the bleeding risks from aspirin. Generally, for individuals 50 to 59 years old, the benefits are deemed to outweigh the harms; shared decision making is advised with those 60 to 69 years of age. For patients younger than 50 or 70 and older, evidence is inconclusive.

The benefits of primary prevention with aspirin

Cardiovascular disease

The Antithrombotic Trialists’ (ATT) Collaboration was one of the first meta-analyses that addressed the benefit-to-harm balance and called into question the routine use of aspirin for primary prevention.⁷ The USPSTF systematic review included the studies from the ATT Collaboration as well as trials performed after its publication, bringing the total number of eligible randomized controlled trials reviewed to 11.²

The benefit of aspirin for primary prevention of nonfatal myocardial infarction (MI) has been shown in multiple randomized controlled trials. The USPSTF systematic review showed a statistically significant relative risk reduction of 17% in patients taking low-dose aspirin (≤ 100 mg; relative risk [RR] = 0.83; confidence interval [95% CI], 0.74-0.94), although the heterogeneity of the studies was high. The same low dose of aspirin showed a statistically significant reduction in nonfatal stroke (RR = 0.86; 95% CI, 0.76-0.98), although the same benefit was not observed when all doses of aspirin were included. Cardiovascular disease mortality and all-cause mortality were not statistically different for patients taking low-dose aspirin when compared with placebo (RR = 0.97; 95% CI, 0.85-1.10 for CVD mortality; RR = 0.95; 95% CI, 0.89-1.01 for all-cause mortality).²

One study of more than 14,000 older (≥ 60 years) Japanese patients showed a statistically significant reduction in nonfatal MI (hazard ratio [HR] = 0.53; 95% CI, 0.31-0.91, P = .02) and nonfatal strokes (HR = 0.57; 95% CI, 0.32-0.99; P = .04). The study was stopped early because at 5 years of follow-up there was no statistically significant difference in a composite primary outcome, which included death from cardiovascular causes, nonfatal MI, and nonfatal stroke (HR = 0.94; 95% CI, 0.77-1.15; P = .54).⁸

Preventive benefits of aspirin outweigh risks for those 50-59 years of age who have a 10-year cardiovascular disease risk of ≥ 10%.

Several recent landmark studies have called into question the benefit of aspirin for cardiovascular primary prevention, especially in obese individuals, patients with diabetes, and the elderly. A meta-analysis of 10 trials showed that the effectiveness of aspirin doses between 75 mg and 100 mg for primary prevention decreased as weight increased; patients weighing 70 kg or more received no benefit.⁹ The ASCEND (A Study of Cardiovascular Events in Diabetes) trial included more than 15,000 patients with diabetes but no cardiovascular disease. Patients randomized to receive the low-dose aspirin did have fewer serious vascular events (incidence rate ratio [IRR] = 0.88; 95% CI, 0.79-0.97; P = .01), but they also had high risk of major bleeding events (IRR = 1.29; 95% CI, 1.09-1.52; P = .003).¹⁰ The ASPREE (Aspirin in Reducing Events in the Elderly) trial included more than 19,000 patients ages 70 years and older with no cardiovascular disease and compared low-dose aspirin to placebo. There was no statistically significant cardiovascular benefit, although there was an increase of major hemorrhage (HR = 1.38; 95% CI, 1.18-1.62; P < .001).¹¹ The ARRIVE (A Randomized Trial of Induction Versus Expectant Management) trial included 12,546 moderate atherosclerotic CVD (ASCVD) risk patients. Although a per-protocol analysis showed a decrease in rates of fatal and nonfatal MI (HR = 0.53; 95% CI, 0.36-0.79; P = .0014), the more reliable intention-to-treat analysis showed no improvement for any outcomes.¹²

[polldaddy:10286821]

Colorectal cancer

The literature base on prevention of cancer has been growing rapidly. However, the deluge of findings over the past 2 decades of trials and analyses has also introduced ambiguity and, often, conflicting results. The first journal article suggesting aspirin for primary prevention of cancer, published in 1988, was a case-control study wherein a population with CRC was matched to controls to look for potential protective factors.¹³ The most notable finding was the CRC risk reduction for those taking aspirin or aspirin-containing medications. Since then numerous studies and analyses have explored aspirin’s potential in primary prevention of many types of cancer, with overall unclear findings as denoted in the 2016 USPSTF systemic reviews and recommendations.

Continue to: One major limiting factor...

One major limiting factor is that most data come from CVD prevention trials, and only a limited number of trials have focused specifically on cancer prevention. For the USPSTF, these data showed no statistically significant risk reduction in overall cancer mortality (RR = 0.96; 95% CI, 0.87-1.06) or in total cancer incidence (RR = 0.98; 95% CI, 0.93-1.04).⁴ Other ongoing trials may yield more definitive data.¹⁴

The particular interest in CRC was due to it being the first cancer found to be preventable with aspirin therapy. The USPSTF, while acknowledging the homogeneous nature of supporting studies, noted that their significant number and resulting evidence made CRC the only cancer warranting evaluation. Population studies have now shown more benefit than the few randomized control trials. The Women’s Health Study and the Physicians’ Health Study were both limited by their duration. But such studies conducted over a longer period revealed notable benefits in the second decade of use, with a statistically significant lower CRC incidence (RR = 0.60; 95% CI, 0.47-0.76). Additionally, CRC mortality at 20 years was decreased in patients taking aspirin regularly (RR = 0.67; 95% CI, 0.52-0.86).⁴ Multiple studies are in progress to better establish aspirin’s CRC benefit.

While not directly applicable to the general population, use of aspirin for patients with Lynch syndrome to prevent CRC has strong supporting evidence.¹⁵ Beyond CRC, there is nascent evidence from limited observational studies that aspirin may have a preventive effect on melanoma and ovarian and pancreatic cancers.^16-18 Further studies or compilations of data would be needed to draw more significant conclusions on other types of cancers. Larger studies would prove more difficult to do, given the smaller incidences of these cancers.

Interestingly, a recent study showed that for individuals 70 years and older, aspirin might increase the risk for all-cause mortality, primarily due to increased cancer mortality across all types.¹⁹ Although this result was unexpected, caution should be used when prescribing aspirin particularly for patients 70 or older with active cancer.

A look at the harms associated with aspirin use

Aspirin has long been known to cause clinically significant bleeding. Aspirin inhibits platelet-derived cyclooxygenase-1 (COX-1), a potent vasoconstrictor, and thereby decreases platelet aggregation, reducing thromboembolic potential and prolonging bleeding time. These effects can confer health benefits but also carry the potential for risks. A decision to initiate aspirin therapy for primary prevention relies on an understanding of the benefit-to-harm balance.

Continue to: Initial aspirin studies...

Initial aspirin studies did not show a statistically significant increase in bleeding, likely due to too few events and inadequate powering. Subsequent meta-analyses from multiple evaluations have consistently shown bleeding to be a risk.^3,7 The risk for bleeding with aspirin has also been examined in multiple cohort studies, which has helped elucidate the risk in greater detail.

Gastrointestinal bleeding

Epidemiologic data show that among patients who do not use nonsteroidal anti-inflammatory drugs (NSAIDs), the rate of upper gastrointestinal (GI) complications is 1 case per 1000 patient-years.²⁰ Multiple studies have consistently shown that aspirin use increases the rate of significant upper GI bleeding over baseline risk (odds ratio [OR] = 1.54-1.58).^3,21,22 Interestingly, these increases seem not to be influenced by other factors, such as comorbidities that increase the risk for ASCVD. Analysis of cancer prevention studies showed similar epidemiologic trends, with aspirin use exceeding a baseline bleeding risk of 0.7 cases of upper GI complications per 1000 patient-years (OR = 1.31-1.73).²³

Other risk factors. Evaluation of risk factors for bleeding primarily comes from 2 studies.^3,7 Most data concern the impact of individual factors on significant GI bleeding, with fewer data available for evaluating risk for intracerebral hemorrhage (ICH). Initial analysis of individual prospective studies showed little or no correlation between risk for bleeding and such factors as gender, age, or history of hypertension or ASCVD.²¹ Subsequent analysis of meta-data and large cohorts did show statistically significant impact on rates of bleeding across several factors (TABLE 1^3,7).

Enteric coating on aspirin does appear to lower the rates of gastric mucosal injury.

Of note is a large heterogeneous cohort study conducted in Spain. Data showed significant increases in baseline risk for GI bleeding in older men with a history of GI bleeding and NSAID use. The absolute risk for GI bleed in this group was potentially as high as 150 cases per 1000 patient-years, well above the risk level assumed for the average patient.²⁴ A seemingly small OR of 1.5 could dramatically increase the absolute risk for bleeding in such patients, and it suggests that a generalized risk for bleeding probably shouldn’t be applied to all patients. Individuals may be better served by a baseline risk calculation reflecting multiple factors.

Intracerebral hemorrhage

Due to the comparatively uncommon nature of ICH, fewer data are available to support definitive conclusions about its increased risk with aspirin use. Aspirin use appeared to increase the risk for ICH with ratios between 1.27 and 1.32 in meta-analyses (measured as an OR or as an RR),^3,7,21 with an IRR of 1.54 in a cohort study.²² The only statistically significant factors suspected to increase the risk of ICH at baseline were smoking (RR = 2.18) and mean BP > 20 mm Hg over normal (OR = 2.18). Age, gender, and diabetes all showed a nonsignificant trend toward risk increase.⁷

Continue to: Risk based on dose and formulation

The effect of aspirin dose and formulation on bleeding risk is uncertain. Some studies have shown an increased risk for bleeding with daily doses of aspirin ≥ 300 mg, while others have shown no significant increase in rates for bleeding with differing doses.^21,25 Enteric coating does appear to lower the rates of gastric mucosal injury, although there are few data on the effect toward reducing clinically significant bleeding.²⁶ Currently, several prospective studies are underway to help clarify the evidence.²⁷

Putting it all together

For the general population, the evidence shows that the benefits and harms of aspirin for primary prevention are relatively even. The USPSTF guidelines are the first to recommend aspirin for both CVD and cancer prevention while taking into account the bleeding risk. According to the findings of the USPSTF, the balance of benefits and harms of aspirin use is contingent on 4 factors: age, baseline CVD risk, risk for bleeding, and preferences about taking aspirin.⁶ The complete recommendations from the USPSTF, along with other leading organizations, are outlined in TABLE 2.^6,28-31

Applying the evidence and varying guidelines in practice can feel daunting. Some practical tools have been developed to help clinicians understand patients’ bleeding risk and potential benefits with aspirin use. One such tool is highlighted below. Others are also available, and each has its own strengths and weaknesses.

Aspirin-Guide (www.aspiringuide.com) is a Web-based clinical decision support tool with an associated mobile application. It uses internal calculators (including the pooled cohort calculator prepared jointly by the American College of Cardiology and the American Heart Association) to assess CVD risk as well as bleeding risk. This tool gives clinicians patient-specific numbers-needed-to-treat and numbers-needed-to-harm when considering starting aspirin for primary prevention. It gives specific recommendations for aspirin use based on the data entered, and it also gives providers information to help guide shared decision-making with patients.³² Unfortunately, this decision support tool and others do not take into account the data from the most recent trials, so they should be used with caution.

CORRESPONDENCE
LCDR Dustin K. Smith, DO, Naval Branch Clinic Diego Garcia, PSC 466, Box 301, FPO, AP 96595; [email protected].

Which patients are likely to benefit from using aspirin for primary prevention? In this article, we review the evidence to date, summarized for primary care settings in guidelines issued by the US Preventive Services Task Force (USPSTF). We supplement this summary with a rundown of the risks associated with aspirin use. And then we wrap up by identifying a clinical decision tool that is available to help make personalized decisions in a busy clinic setting, where determining an individual’s potential cardiovascular benefits and bleeding risk can be challenging.

The “roadmap” from the guidelines. In 2014, after performing a review of the literature, the US Food and Drug Administration recommended against the routine use of aspirin for primary prevention of cardiovascular disease (CVD).¹ In 2016, the USPSTF published 4 separate systematic reviews along with a decision analysis using a microsimulation model, which informed their position statement on aspirin for primary prevention.^2-6 These USPSTF reviews and recommendations incorporated both CVD and colorectal cancer (CRC) benefits with the bleeding risks from aspirin. Generally, for individuals 50 to 59 years old, the benefits are deemed to outweigh the harms; shared decision making is advised with those 60 to 69 years of age. For patients younger than 50 or 70 and older, evidence is inconclusive.

The benefits of primary prevention with aspirin

Cardiovascular disease

The Antithrombotic Trialists’ (ATT) Collaboration was one of the first meta-analyses that addressed the benefit-to-harm balance and called into question the routine use of aspirin for primary prevention.⁷ The USPSTF systematic review included the studies from the ATT Collaboration as well as trials performed after its publication, bringing the total number of eligible randomized controlled trials reviewed to 11.²

The benefit of aspirin for primary prevention of nonfatal myocardial infarction (MI) has been shown in multiple randomized controlled trials. The USPSTF systematic review showed a statistically significant relative risk reduction of 17% in patients taking low-dose aspirin (≤ 100 mg; relative risk [RR] = 0.83; confidence interval [95% CI], 0.74-0.94), although the heterogeneity of the studies was high. The same low dose of aspirin showed a statistically significant reduction in nonfatal stroke (RR = 0.86; 95% CI, 0.76-0.98), although the same benefit was not observed when all doses of aspirin were included. Cardiovascular disease mortality and all-cause mortality were not statistically different for patients taking low-dose aspirin when compared with placebo (RR = 0.97; 95% CI, 0.85-1.10 for CVD mortality; RR = 0.95; 95% CI, 0.89-1.01 for all-cause mortality).²

One study of more than 14,000 older (≥ 60 years) Japanese patients showed a statistically significant reduction in nonfatal MI (hazard ratio [HR] = 0.53; 95% CI, 0.31-0.91, P = .02) and nonfatal strokes (HR = 0.57; 95% CI, 0.32-0.99; P = .04). The study was stopped early because at 5 years of follow-up there was no statistically significant difference in a composite primary outcome, which included death from cardiovascular causes, nonfatal MI, and nonfatal stroke (HR = 0.94; 95% CI, 0.77-1.15; P = .54).⁸

Preventive benefits of aspirin outweigh risks for those 50-59 years of age who have a 10-year cardiovascular disease risk of ≥ 10%.

Several recent landmark studies have called into question the benefit of aspirin for cardiovascular primary prevention, especially in obese individuals, patients with diabetes, and the elderly. A meta-analysis of 10 trials showed that the effectiveness of aspirin doses between 75 mg and 100 mg for primary prevention decreased as weight increased; patients weighing 70 kg or more received no benefit.⁹ The ASCEND (A Study of Cardiovascular Events in Diabetes) trial included more than 15,000 patients with diabetes but no cardiovascular disease. Patients randomized to receive the low-dose aspirin did have fewer serious vascular events (incidence rate ratio [IRR] = 0.88; 95% CI, 0.79-0.97; P = .01), but they also had high risk of major bleeding events (IRR = 1.29; 95% CI, 1.09-1.52; P = .003).¹⁰ The ASPREE (Aspirin in Reducing Events in the Elderly) trial included more than 19,000 patients ages 70 years and older with no cardiovascular disease and compared low-dose aspirin to placebo. There was no statistically significant cardiovascular benefit, although there was an increase of major hemorrhage (HR = 1.38; 95% CI, 1.18-1.62; P < .001).¹¹ The ARRIVE (A Randomized Trial of Induction Versus Expectant Management) trial included 12,546 moderate atherosclerotic CVD (ASCVD) risk patients. Although a per-protocol analysis showed a decrease in rates of fatal and nonfatal MI (HR = 0.53; 95% CI, 0.36-0.79; P = .0014), the more reliable intention-to-treat analysis showed no improvement for any outcomes.¹²

[polldaddy:10286821]

Colorectal cancer

The literature base on prevention of cancer has been growing rapidly. However, the deluge of findings over the past 2 decades of trials and analyses has also introduced ambiguity and, often, conflicting results. The first journal article suggesting aspirin for primary prevention of cancer, published in 1988, was a case-control study wherein a population with CRC was matched to controls to look for potential protective factors.¹³ The most notable finding was the CRC risk reduction for those taking aspirin or aspirin-containing medications. Since then numerous studies and analyses have explored aspirin’s potential in primary prevention of many types of cancer, with overall unclear findings as denoted in the 2016 USPSTF systemic reviews and recommendations.

Continue to: One major limiting factor...

One major limiting factor is that most data come from CVD prevention trials, and only a limited number of trials have focused specifically on cancer prevention. For the USPSTF, these data showed no statistically significant risk reduction in overall cancer mortality (RR = 0.96; 95% CI, 0.87-1.06) or in total cancer incidence (RR = 0.98; 95% CI, 0.93-1.04).⁴ Other ongoing trials may yield more definitive data.¹⁴

The particular interest in CRC was due to it being the first cancer found to be preventable with aspirin therapy. The USPSTF, while acknowledging the homogeneous nature of supporting studies, noted that their significant number and resulting evidence made CRC the only cancer warranting evaluation. Population studies have now shown more benefit than the few randomized control trials. The Women’s Health Study and the Physicians’ Health Study were both limited by their duration. But such studies conducted over a longer period revealed notable benefits in the second decade of use, with a statistically significant lower CRC incidence (RR = 0.60; 95% CI, 0.47-0.76). Additionally, CRC mortality at 20 years was decreased in patients taking aspirin regularly (RR = 0.67; 95% CI, 0.52-0.86).⁴ Multiple studies are in progress to better establish aspirin’s CRC benefit.

While not directly applicable to the general population, use of aspirin for patients with Lynch syndrome to prevent CRC has strong supporting evidence.¹⁵ Beyond CRC, there is nascent evidence from limited observational studies that aspirin may have a preventive effect on melanoma and ovarian and pancreatic cancers.^16-18 Further studies or compilations of data would be needed to draw more significant conclusions on other types of cancers. Larger studies would prove more difficult to do, given the smaller incidences of these cancers.

Interestingly, a recent study showed that for individuals 70 years and older, aspirin might increase the risk for all-cause mortality, primarily due to increased cancer mortality across all types.¹⁹ Although this result was unexpected, caution should be used when prescribing aspirin particularly for patients 70 or older with active cancer.

A look at the harms associated with aspirin use

Aspirin has long been known to cause clinically significant bleeding. Aspirin inhibits platelet-derived cyclooxygenase-1 (COX-1), a potent vasoconstrictor, and thereby decreases platelet aggregation, reducing thromboembolic potential and prolonging bleeding time. These effects can confer health benefits but also carry the potential for risks. A decision to initiate aspirin therapy for primary prevention relies on an understanding of the benefit-to-harm balance.

Continue to: Initial aspirin studies...

Initial aspirin studies did not show a statistically significant increase in bleeding, likely due to too few events and inadequate powering. Subsequent meta-analyses from multiple evaluations have consistently shown bleeding to be a risk.^3,7 The risk for bleeding with aspirin has also been examined in multiple cohort studies, which has helped elucidate the risk in greater detail.

Gastrointestinal bleeding

Epidemiologic data show that among patients who do not use nonsteroidal anti-inflammatory drugs (NSAIDs), the rate of upper gastrointestinal (GI) complications is 1 case per 1000 patient-years.²⁰ Multiple studies have consistently shown that aspirin use increases the rate of significant upper GI bleeding over baseline risk (odds ratio [OR] = 1.54-1.58).^3,21,22 Interestingly, these increases seem not to be influenced by other factors, such as comorbidities that increase the risk for ASCVD. Analysis of cancer prevention studies showed similar epidemiologic trends, with aspirin use exceeding a baseline bleeding risk of 0.7 cases of upper GI complications per 1000 patient-years (OR = 1.31-1.73).²³

Other risk factors. Evaluation of risk factors for bleeding primarily comes from 2 studies.^3,7 Most data concern the impact of individual factors on significant GI bleeding, with fewer data available for evaluating risk for intracerebral hemorrhage (ICH). Initial analysis of individual prospective studies showed little or no correlation between risk for bleeding and such factors as gender, age, or history of hypertension or ASCVD.²¹ Subsequent analysis of meta-data and large cohorts did show statistically significant impact on rates of bleeding across several factors (TABLE 1^3,7).

Enteric coating on aspirin does appear to lower the rates of gastric mucosal injury.

Of note is a large heterogeneous cohort study conducted in Spain. Data showed significant increases in baseline risk for GI bleeding in older men with a history of GI bleeding and NSAID use. The absolute risk for GI bleed in this group was potentially as high as 150 cases per 1000 patient-years, well above the risk level assumed for the average patient.²⁴ A seemingly small OR of 1.5 could dramatically increase the absolute risk for bleeding in such patients, and it suggests that a generalized risk for bleeding probably shouldn’t be applied to all patients. Individuals may be better served by a baseline risk calculation reflecting multiple factors.

Intracerebral hemorrhage

Due to the comparatively uncommon nature of ICH, fewer data are available to support definitive conclusions about its increased risk with aspirin use. Aspirin use appeared to increase the risk for ICH with ratios between 1.27 and 1.32 in meta-analyses (measured as an OR or as an RR),^3,7,21 with an IRR of 1.54 in a cohort study.²² The only statistically significant factors suspected to increase the risk of ICH at baseline were smoking (RR = 2.18) and mean BP > 20 mm Hg over normal (OR = 2.18). Age, gender, and diabetes all showed a nonsignificant trend toward risk increase.⁷

Continue to: Risk based on dose and formulation

The effect of aspirin dose and formulation on bleeding risk is uncertain. Some studies have shown an increased risk for bleeding with daily doses of aspirin ≥ 300 mg, while others have shown no significant increase in rates for bleeding with differing doses.^21,25 Enteric coating does appear to lower the rates of gastric mucosal injury, although there are few data on the effect toward reducing clinically significant bleeding.²⁶ Currently, several prospective studies are underway to help clarify the evidence.²⁷

Putting it all together

For the general population, the evidence shows that the benefits and harms of aspirin for primary prevention are relatively even. The USPSTF guidelines are the first to recommend aspirin for both CVD and cancer prevention while taking into account the bleeding risk. According to the findings of the USPSTF, the balance of benefits and harms of aspirin use is contingent on 4 factors: age, baseline CVD risk, risk for bleeding, and preferences about taking aspirin.⁶ The complete recommendations from the USPSTF, along with other leading organizations, are outlined in TABLE 2.^6,28-31

Applying the evidence and varying guidelines in practice can feel daunting. Some practical tools have been developed to help clinicians understand patients’ bleeding risk and potential benefits with aspirin use. One such tool is highlighted below. Others are also available, and each has its own strengths and weaknesses.

Aspirin-Guide (www.aspiringuide.com) is a Web-based clinical decision support tool with an associated mobile application. It uses internal calculators (including the pooled cohort calculator prepared jointly by the American College of Cardiology and the American Heart Association) to assess CVD risk as well as bleeding risk. This tool gives clinicians patient-specific numbers-needed-to-treat and numbers-needed-to-harm when considering starting aspirin for primary prevention. It gives specific recommendations for aspirin use based on the data entered, and it also gives providers information to help guide shared decision-making with patients.³² Unfortunately, this decision support tool and others do not take into account the data from the most recent trials, so they should be used with caution.

CORRESPONDENCE
LCDR Dustin K. Smith, DO, Naval Branch Clinic Diego Garcia, PSC 466, Box 301, FPO, AP 96595; [email protected].

Which patients are likely to benefit from using aspirin for primary prevention? In this article, we review the evidence to date, summarized for primary care settings in guidelines issued by the US Preventive Services Task Force (USPSTF). We supplement this summary with a rundown of the risks associated with aspirin use. And then we wrap up by identifying a clinical decision tool that is available to help make personalized decisions in a busy clinic setting, where determining an individual’s potential cardiovascular benefits and bleeding risk can be challenging.

The “roadmap” from the guidelines. In 2014, after performing a review of the literature, the US Food and Drug Administration recommended against the routine use of aspirin for primary prevention of cardiovascular disease (CVD).¹ In 2016, the USPSTF published 4 separate systematic reviews along with a decision analysis using a microsimulation model, which informed their position statement on aspirin for primary prevention.^2-6 These USPSTF reviews and recommendations incorporated both CVD and colorectal cancer (CRC) benefits with the bleeding risks from aspirin. Generally, for individuals 50 to 59 years old, the benefits are deemed to outweigh the harms; shared decision making is advised with those 60 to 69 years of age. For patients younger than 50 or 70 and older, evidence is inconclusive.

The benefits of primary prevention with aspirin

Cardiovascular disease

The Antithrombotic Trialists’ (ATT) Collaboration was one of the first meta-analyses that addressed the benefit-to-harm balance and called into question the routine use of aspirin for primary prevention.⁷ The USPSTF systematic review included the studies from the ATT Collaboration as well as trials performed after its publication, bringing the total number of eligible randomized controlled trials reviewed to 11.²

The benefit of aspirin for primary prevention of nonfatal myocardial infarction (MI) has been shown in multiple randomized controlled trials. The USPSTF systematic review showed a statistically significant relative risk reduction of 17% in patients taking low-dose aspirin (≤ 100 mg; relative risk [RR] = 0.83; confidence interval [95% CI], 0.74-0.94), although the heterogeneity of the studies was high. The same low dose of aspirin showed a statistically significant reduction in nonfatal stroke (RR = 0.86; 95% CI, 0.76-0.98), although the same benefit was not observed when all doses of aspirin were included. Cardiovascular disease mortality and all-cause mortality were not statistically different for patients taking low-dose aspirin when compared with placebo (RR = 0.97; 95% CI, 0.85-1.10 for CVD mortality; RR = 0.95; 95% CI, 0.89-1.01 for all-cause mortality).²

One study of more than 14,000 older (≥ 60 years) Japanese patients showed a statistically significant reduction in nonfatal MI (hazard ratio [HR] = 0.53; 95% CI, 0.31-0.91, P = .02) and nonfatal strokes (HR = 0.57; 95% CI, 0.32-0.99; P = .04). The study was stopped early because at 5 years of follow-up there was no statistically significant difference in a composite primary outcome, which included death from cardiovascular causes, nonfatal MI, and nonfatal stroke (HR = 0.94; 95% CI, 0.77-1.15; P = .54).⁸

Preventive benefits of aspirin outweigh risks for those 50-59 years of age who have a 10-year cardiovascular disease risk of ≥ 10%.

Several recent landmark studies have called into question the benefit of aspirin for cardiovascular primary prevention, especially in obese individuals, patients with diabetes, and the elderly. A meta-analysis of 10 trials showed that the effectiveness of aspirin doses between 75 mg and 100 mg for primary prevention decreased as weight increased; patients weighing 70 kg or more received no benefit.⁹ The ASCEND (A Study of Cardiovascular Events in Diabetes) trial included more than 15,000 patients with diabetes but no cardiovascular disease. Patients randomized to receive the low-dose aspirin did have fewer serious vascular events (incidence rate ratio [IRR] = 0.88; 95% CI, 0.79-0.97; P = .01), but they also had high risk of major bleeding events (IRR = 1.29; 95% CI, 1.09-1.52; P = .003).¹⁰ The ASPREE (Aspirin in Reducing Events in the Elderly) trial included more than 19,000 patients ages 70 years and older with no cardiovascular disease and compared low-dose aspirin to placebo. There was no statistically significant cardiovascular benefit, although there was an increase of major hemorrhage (HR = 1.38; 95% CI, 1.18-1.62; P < .001).¹¹ The ARRIVE (A Randomized Trial of Induction Versus Expectant Management) trial included 12,546 moderate atherosclerotic CVD (ASCVD) risk patients. Although a per-protocol analysis showed a decrease in rates of fatal and nonfatal MI (HR = 0.53; 95% CI, 0.36-0.79; P = .0014), the more reliable intention-to-treat analysis showed no improvement for any outcomes.¹²

[polldaddy:10286821]

Colorectal cancer

The literature base on prevention of cancer has been growing rapidly. However, the deluge of findings over the past 2 decades of trials and analyses has also introduced ambiguity and, often, conflicting results. The first journal article suggesting aspirin for primary prevention of cancer, published in 1988, was a case-control study wherein a population with CRC was matched to controls to look for potential protective factors.¹³ The most notable finding was the CRC risk reduction for those taking aspirin or aspirin-containing medications. Since then numerous studies and analyses have explored aspirin’s potential in primary prevention of many types of cancer, with overall unclear findings as denoted in the 2016 USPSTF systemic reviews and recommendations.

Continue to: One major limiting factor...

One major limiting factor is that most data come from CVD prevention trials, and only a limited number of trials have focused specifically on cancer prevention. For the USPSTF, these data showed no statistically significant risk reduction in overall cancer mortality (RR = 0.96; 95% CI, 0.87-1.06) or in total cancer incidence (RR = 0.98; 95% CI, 0.93-1.04).⁴ Other ongoing trials may yield more definitive data.¹⁴

The particular interest in CRC was due to it being the first cancer found to be preventable with aspirin therapy. The USPSTF, while acknowledging the homogeneous nature of supporting studies, noted that their significant number and resulting evidence made CRC the only cancer warranting evaluation. Population studies have now shown more benefit than the few randomized control trials. The Women’s Health Study and the Physicians’ Health Study were both limited by their duration. But such studies conducted over a longer period revealed notable benefits in the second decade of use, with a statistically significant lower CRC incidence (RR = 0.60; 95% CI, 0.47-0.76). Additionally, CRC mortality at 20 years was decreased in patients taking aspirin regularly (RR = 0.67; 95% CI, 0.52-0.86).⁴ Multiple studies are in progress to better establish aspirin’s CRC benefit.

While not directly applicable to the general population, use of aspirin for patients with Lynch syndrome to prevent CRC has strong supporting evidence.¹⁵ Beyond CRC, there is nascent evidence from limited observational studies that aspirin may have a preventive effect on melanoma and ovarian and pancreatic cancers.^16-18 Further studies or compilations of data would be needed to draw more significant conclusions on other types of cancers. Larger studies would prove more difficult to do, given the smaller incidences of these cancers.

Interestingly, a recent study showed that for individuals 70 years and older, aspirin might increase the risk for all-cause mortality, primarily due to increased cancer mortality across all types.¹⁹ Although this result was unexpected, caution should be used when prescribing aspirin particularly for patients 70 or older with active cancer.

A look at the harms associated with aspirin use

Aspirin has long been known to cause clinically significant bleeding. Aspirin inhibits platelet-derived cyclooxygenase-1 (COX-1), a potent vasoconstrictor, and thereby decreases platelet aggregation, reducing thromboembolic potential and prolonging bleeding time. These effects can confer health benefits but also carry the potential for risks. A decision to initiate aspirin therapy for primary prevention relies on an understanding of the benefit-to-harm balance.

Continue to: Initial aspirin studies...

Initial aspirin studies did not show a statistically significant increase in bleeding, likely due to too few events and inadequate powering. Subsequent meta-analyses from multiple evaluations have consistently shown bleeding to be a risk.^3,7 The risk for bleeding with aspirin has also been examined in multiple cohort studies, which has helped elucidate the risk in greater detail.

Gastrointestinal bleeding

Epidemiologic data show that among patients who do not use nonsteroidal anti-inflammatory drugs (NSAIDs), the rate of upper gastrointestinal (GI) complications is 1 case per 1000 patient-years.²⁰ Multiple studies have consistently shown that aspirin use increases the rate of significant upper GI bleeding over baseline risk (odds ratio [OR] = 1.54-1.58).^3,21,22 Interestingly, these increases seem not to be influenced by other factors, such as comorbidities that increase the risk for ASCVD. Analysis of cancer prevention studies showed similar epidemiologic trends, with aspirin use exceeding a baseline bleeding risk of 0.7 cases of upper GI complications per 1000 patient-years (OR = 1.31-1.73).²³

Other risk factors. Evaluation of risk factors for bleeding primarily comes from 2 studies.^3,7 Most data concern the impact of individual factors on significant GI bleeding, with fewer data available for evaluating risk for intracerebral hemorrhage (ICH). Initial analysis of individual prospective studies showed little or no correlation between risk for bleeding and such factors as gender, age, or history of hypertension or ASCVD.²¹ Subsequent analysis of meta-data and large cohorts did show statistically significant impact on rates of bleeding across several factors (TABLE 1^3,7).

Enteric coating on aspirin does appear to lower the rates of gastric mucosal injury.

Of note is a large heterogeneous cohort study conducted in Spain. Data showed significant increases in baseline risk for GI bleeding in older men with a history of GI bleeding and NSAID use. The absolute risk for GI bleed in this group was potentially as high as 150 cases per 1000 patient-years, well above the risk level assumed for the average patient.²⁴ A seemingly small OR of 1.5 could dramatically increase the absolute risk for bleeding in such patients, and it suggests that a generalized risk for bleeding probably shouldn’t be applied to all patients. Individuals may be better served by a baseline risk calculation reflecting multiple factors.

Intracerebral hemorrhage

Due to the comparatively uncommon nature of ICH, fewer data are available to support definitive conclusions about its increased risk with aspirin use. Aspirin use appeared to increase the risk for ICH with ratios between 1.27 and 1.32 in meta-analyses (measured as an OR or as an RR),^3,7,21 with an IRR of 1.54 in a cohort study.²² The only statistically significant factors suspected to increase the risk of ICH at baseline were smoking (RR = 2.18) and mean BP > 20 mm Hg over normal (OR = 2.18). Age, gender, and diabetes all showed a nonsignificant trend toward risk increase.⁷

Continue to: Risk based on dose and formulation

The effect of aspirin dose and formulation on bleeding risk is uncertain. Some studies have shown an increased risk for bleeding with daily doses of aspirin ≥ 300 mg, while others have shown no significant increase in rates for bleeding with differing doses.^21,25 Enteric coating does appear to lower the rates of gastric mucosal injury, although there are few data on the effect toward reducing clinically significant bleeding.²⁶ Currently, several prospective studies are underway to help clarify the evidence.²⁷

Putting it all together

For the general population, the evidence shows that the benefits and harms of aspirin for primary prevention are relatively even. The USPSTF guidelines are the first to recommend aspirin for both CVD and cancer prevention while taking into account the bleeding risk. According to the findings of the USPSTF, the balance of benefits and harms of aspirin use is contingent on 4 factors: age, baseline CVD risk, risk for bleeding, and preferences about taking aspirin.⁶ The complete recommendations from the USPSTF, along with other leading organizations, are outlined in TABLE 2.^6,28-31

Applying the evidence and varying guidelines in practice can feel daunting. Some practical tools have been developed to help clinicians understand patients’ bleeding risk and potential benefits with aspirin use. One such tool is highlighted below. Others are also available, and each has its own strengths and weaknesses.

Aspirin-Guide (www.aspiringuide.com) is a Web-based clinical decision support tool with an associated mobile application. It uses internal calculators (including the pooled cohort calculator prepared jointly by the American College of Cardiology and the American Heart Association) to assess CVD risk as well as bleeding risk. This tool gives clinicians patient-specific numbers-needed-to-treat and numbers-needed-to-harm when considering starting aspirin for primary prevention. It gives specific recommendations for aspirin use based on the data entered, and it also gives providers information to help guide shared decision-making with patients.³² Unfortunately, this decision support tool and others do not take into account the data from the most recent trials, so they should be used with caution.

CORRESPONDENCE
LCDR Dustin K. Smith, DO, Naval Branch Clinic Diego Garcia, PSC 466, Box 301, FPO, AP 96595; [email protected].

WASHINGTON – Although list prices remain an easy target when discussing spikes in insulin prices, reforms are needed across the supply chain for the problem to be fully resolved, panelists said at a House Committee on Energy & Commerce hearing on insulin affordability.

“Each member of the supply chain has a responsibility to help solve this problem,” said Alvin C. Powers, MD, director of the Vanderbilt Diabetes Center at Vanderbilt University, who was speaking on behalf of the Endocrine Society during the April 2 hearing of the committee’s oversight & investigations subcommittee.

Dr. Powers identified all members – manufacturers, payers, pharmacy benefit managers, patients, providers, and Congress – as having a role in developing a solution that will encourage more access to the treatment.

The hearing was the first of two in a series specifically examining the price of insulin. This one focused on the role pricing issues play in terms of access to insulin and patient outcomes.

To highlight the pricing issues, it was noted that a vial of Humalog (insulin lispro) cost $21 when it was launched by Eli Lilly in 1996. It now costs $275 even though it has gone through no changes in formulation or innovation during that time.

Kasia Lipska, MD, of Yale University School of Medicine noted that a summer 2017 survey conducted by the Yale Diabetes Center found that one in four patients took less than the prescribed dose of insulin specifically because of the cost of insulin.

[email protected]

WASHINGTON – Although list prices remain an easy target when discussing spikes in insulin prices, reforms are needed across the supply chain for the problem to be fully resolved, panelists said at a House Committee on Energy & Commerce hearing on insulin affordability.

“Each member of the supply chain has a responsibility to help solve this problem,” said Alvin C. Powers, MD, director of the Vanderbilt Diabetes Center at Vanderbilt University, who was speaking on behalf of the Endocrine Society during the April 2 hearing of the committee’s oversight & investigations subcommittee.

Dr. Powers identified all members – manufacturers, payers, pharmacy benefit managers, patients, providers, and Congress – as having a role in developing a solution that will encourage more access to the treatment.

The hearing was the first of two in a series specifically examining the price of insulin. This one focused on the role pricing issues play in terms of access to insulin and patient outcomes.

To highlight the pricing issues, it was noted that a vial of Humalog (insulin lispro) cost $21 when it was launched by Eli Lilly in 1996. It now costs $275 even though it has gone through no changes in formulation or innovation during that time.

Kasia Lipska, MD, of Yale University School of Medicine noted that a summer 2017 survey conducted by the Yale Diabetes Center found that one in four patients took less than the prescribed dose of insulin specifically because of the cost of insulin.

[email protected]

WASHINGTON – Although list prices remain an easy target when discussing spikes in insulin prices, reforms are needed across the supply chain for the problem to be fully resolved, panelists said at a House Committee on Energy & Commerce hearing on insulin affordability.

“Each member of the supply chain has a responsibility to help solve this problem,” said Alvin C. Powers, MD, director of the Vanderbilt Diabetes Center at Vanderbilt University, who was speaking on behalf of the Endocrine Society during the April 2 hearing of the committee’s oversight & investigations subcommittee.

Dr. Powers identified all members – manufacturers, payers, pharmacy benefit managers, patients, providers, and Congress – as having a role in developing a solution that will encourage more access to the treatment.

The hearing was the first of two in a series specifically examining the price of insulin. This one focused on the role pricing issues play in terms of access to insulin and patient outcomes.

To highlight the pricing issues, it was noted that a vial of Humalog (insulin lispro) cost $21 when it was launched by Eli Lilly in 1996. It now costs $275 even though it has gone through no changes in formulation or innovation during that time.

Kasia Lipska, MD, of Yale University School of Medicine noted that a summer 2017 survey conducted by the Yale Diabetes Center found that one in four patients took less than the prescribed dose of insulin specifically because of the cost of insulin.

[email protected]

A 60-year-old man is admitted for respiratory failure following a massive myocardial infarction. He develops ventilator-associated pneumonia and is treated with cefepime and vancomycin. Three days later, he develops prolonged atypical absence seizures.

What caused these seizures? The neurologist thinks it might be the cefepime. Do you agree?

Antibiotics are widely used in the United States, with 269 million courses of oral therapy prescribed in 2011.¹ Adverse effects such as rash are well known, but rare effects such as seizure, hypoglycemia, and hypoxemia may not be immediately attributed to these drugs.

In this article, we review less-recognized but potentially serious adverse effects of antibiotics commonly prescribed in the United States. We have structured our discussion by organ system for ease of reference.

NERVOUS SYSTEM

The potential adverse effects of antibiotics on the nervous system range from encephalopathy and seizure to nonconvulsive status epilepticus.

Encephalopathy and seizure

Encephalopathy has been reported with penicillins, cephalosporins, sulfamethoxazole-trimethoprim, quinolones, and oxazolidinones such as linezolid.^2,3

Seizures are known to occur with penicillins, cephalosporins, carbapenems, and quinolones.^2–4 For cephalosporins, these effects are more common at higher doses, in elderly patients, and in patients with renal impairment. Carbapenems are associated with seizure activity in elderly patients.^2–4

Encephalopathy and seizure can also occur on a continuum, as is the case with piperacillin-induced encephalopathy, with progressive dysarthria, tremor, and progressive confusion culminating in tonic-clonic seizures.²

Nonconvulsive status epilepticus

Nonconvulsive status epilepticus, marked by prolonged atypical absence seizures, has complicated the use of penicillins, quinolones, clarithromycin, and cephalosporins, specifically cefepime.^2,3,5 Diagnosis can be difficult and requires clinical awareness and confirmation with electroencephalography.

Class-specific neurologic effects

Certain antibiotics have class-specific effects:

Tetracyclines: cranial nerve toxicity, neuromuscular blockade, and intracranial hypertension.²

Sulfamethoxazole-trimethoprim: tremors and psychosis, with visual and auditory hallucinations.⁶

Macrolides: dysequilibrium and potentially irreversible hearing loss.²

Quinolones: orofacial dyskinesia and a Tourette-like syndrome, with a higher incidence reported with newer quinolones.⁷

Linezolid: optic and peripheral neuropathy²; neuropathy can be persistent and can lead to loss of vision. The package insert recommends monitoring visual function in patients taking linezolid for more than 3 months and in any patient reporting visual symptoms.⁸

Linezolid is also associated with serotonin syndrome when combined with a drug that potentiates serotonergic activity, most commonly selective serotonin reuptake inhibitors. The syndrome is characterized by a triad of cognitive or behavioral changes, autonomic instability, and neuromuscular excitability such as spontaneous clonus.⁹

Metronidazole: optic and peripheral neuropathy, in addition to cerebellar toxicity and central nervous system lesions on magnetic resonance imaging of the brain. In a series of 11 cases of cerebellar toxicity, most patients presented with ataxia and dysarthria associated with high total doses of metronidazole, and in most cases, magnetic resonance imaging showed resolution of the lesions upon discontinuation of metronidazole.¹⁰

HEMATOLOGIC AND RHEUMATOLOGIC EFFECTS

Agranulocytosis has been associated with beta-lactams, in most cases with prolonged exposure. In one report, the average exposure before onset of agranulocytosis was 22 days for nafcillin and 25 days for penicillin. For penicillins, more than 50% of cases involved high daily doses.¹¹

Likewise, most episodes of vancomycin-induced neutropenia were reported to occur after 20 days of therapy.¹²

In another study, most cases of drug-induced anemia were due to ceftriaxone and piperacillin.¹³

Drug-induced thrombocytopenia has been described with penicillins, cephalo­sporins, sulfonamides, and vancomycin¹⁴ and is a well-recognized effect of linezolid. The syndrome of drug reaction with eosinophilia and systemic symptoms, a severe and rare adverse reaction, has been reported with minocycline, sulfamethoxazole, and vancomycin.¹⁵

The tetracycline minocycline has been reported to cause drug-induced lupus and polyarteritis nodosa-like vasculitis.¹⁶ Drug-induced lupus presents as myalgias and arthralgias, serositis, constitutional symptoms, and positive antinuclear antibody titers. The effect is not dose-dependent. Penicillin, cefuroxime, and nitrofurantoin have also been implicated.¹⁶

Kermani et al¹⁷ described 9 cases of polyarteritis nodosa, in which 5 patients (56%) had systemic involvement including renal artery microaneurysm, mononeuritis multiplex, and mesenteric vasculitis, and some of these patients also had cutaneous involvement. All patients had positive antineutrophil cytoplasmic antibody in a perinuclear pattern. The median time from start of the minocycline to symptom onset was 9 months, and the median duration of use was 2 years.

Quinolones have also been reported to cause fatal hypersensitivity vasculitis.^18,19

CARDIOVASCULAR SYSTEM

Macrolides and quinolones have been reported to cause QT-interval prolongation and torsades de pointes. The risk is greatest when a  macrolide is co-administered with a CYP3A4 inhibitor.

Of the macrolides, azithromycin is the safest, as clarithromycin and erythromycin are more likely to cause QT prolongation.

While QT prolongation is a class effect of quinolones, there is variability within the class. Ciprofloxacin is thought to be the safest in terms of cardiovascular adverse effects.²⁰ In addition, Owens and Nolin²⁰ reported that quinolone-associated QT prolongation was more likely to occur in patients with pre-existing QT prolongation, electrolyte abnormalities, organic heart disease, and bradycardia, and especially in women. Other risk factors for QT prolongation with quinolone use include underlying cardiac disease and advanced age.²¹

Quinolones have also been associated with an increased risk of aortic dissection. The US Food and Drug Administration has issued a warning advising clinicians to avoid quinolones in patients who have aneurysms or are at risk for aneurysms, such as patients with advanced age, peripheral atherosclerotic vascular disease, hypertension and conditions such as Marfan and Ehlers-Danlos syndrome.²²

DIGESTIVE SYSTEM

Tetracyclines are known to cause esophagitis from direct contact with and disruption of the mucosal lining. Doxycycline is the most frequent offender.²³

Amoxicillin-clavulanate is the antibiotic most commonly associated with drug-induced liver injury, mainly attributable to the clavulanate component.²⁴ It is more common in men over age 50 and with prolonged and repeated dosing and is sometimes fatal. Other adverse effects include Stevens-Johnson syndrome, interstitial nephritis, and thrombotic thrombocytopenic purpura.²⁵

Cholestatic hepatitis has been reported with penicillins, particularly dicloxacillin, oxacillin, and amoxicillin-clavulanate; cephalosporins; doxycycline; sulfamethoxazole-trimethoprim; macrolides; and ciprofloxacin.^24–26 Hepatocellular injury is linked to amoxicillin-clavulanate and doxycycline. Drug-induced mixed liver injury has been observed with amoxicillin-clavulanate, sulfamethoxazole-trimethoprim and, rarely, cephalosporins.

Liver injury is classified as cholestatic if the alkaline phosphatase level is more than 2 times higher than normal, or if the ratio of alanine aminotransferase to alkaline phosphatase is less than 2; if the ratio is greater than 5, the injury is considered hepatocellular.²⁴ Mixed liver injury, the most common, is defined as a ratio from 2 to 5.

Nitrofurantoin has also been linked to hepatotoxicity, cirrhosis, and end-stage liver disease, and to death if the drug is continued after the onset of jaundice.²⁶ Death from liver injury has been reported with amoxicillin-clavulanate, sulfamethoxazole-trimethoprim, and erythromycin, and jaundice indicates a poor prognosis, associated with a 10% mortality rate or need for liver transplant in all patients.²⁴

ENDOCRINE SYSTEM

Clarithromycin, sulfonamides, and quinolones are known to precipitate hypoglycemia by interacting with sulfonylureas. A study of Medicare patients age 66 or older who were taking glipizide or glyburide reported that female sex, older age, and a history of hypoglycemic episodes were associated with antibiotic-related hypoglycemia.²⁷ The odds ratio for hypoglycemia was highest for clarithromycin (3.96), sulfamethoxazole-trimethoprim (2.56), metronidazole (2.11), and ciprofloxacin (1.62) when compared with antibiotics that do not cause hypoglycemia. There was no signal for levofloxacin-mediated hypoglycemia in this series.²⁷

RESPIRATORY SYSTEM

Hypersensitivity lung disease has been reported with penicillin, ampicillin, cephalosporins, ciprofloxacin, and sulfonamides including sulfamethoxazole-trimethoprim.²⁸ The lipopeptide daptomycin has been reported to cause acute eosinophilic pneumonia defined as fever for less than 5 days, pulmonary infiltrates, hypoxemia, and a bronchoalveolar lavage or biopsy study with eosinophils. Daptomycin should be stopped early in these cases, and the patient should not be rechallenged, as the reaction can be deadly.²⁹

Nitrofurantoin has a long history of hypersensitivity pneumonitis in its acute form and a chronic allergic response. While more widely recognized, nitrofurantoin pulmonary toxicity is rare, occurring in 1 in 5,000 patients.³⁰

RENAL SYSTEM

Acute interstitial nephritis has been reported with penicillins, cephalosporins, macrolides, quinolones, sulfonamides, and vancomycin.^31–33 Acute tubular necrosis has been linked to cephalosporins and tetracyclines. Crystal nephropathy has been seen with quinolones and sulfonamides.

Advanced age is an important risk factor for renal dysfunction from quinolones,¹⁸ and penicillin G has been reported to cause glomerulonephritis.³¹

MUSCULOSKELETAL SYSTEM

Quinolones have been associated with arthropathy or tendinitis at a rate of 1%, including cases of Achilles tendon rupture.¹⁸ The US Food and Drug Administration announced in 2016 that the serious adverse events with fluoroquinolones outweigh the benefits in patients with acute sinusitis, acute bronchitis, and uncomplicated urinary tract infection, and that they should be used only if there are no other options.³⁴

Daptomycin is known to cause elevations of creatine kinase.³⁴ Weekly monitoring is recommended based on postmarketing data reports of elevations in 2.5% of patients; myopathy is a rarer effect, occurring in 0.2% of patients.³⁵

REPRODUCTIVE SYSTEM

Antibiotics have long been reported to interact with oral contraceptives, but the data are not compelling for commonly used antibiotics. The strongest association is with rifampicin, which reduces oral contraceptive efficacy and warrants an alternative mode of contraception.³⁶

BACK TO OUR PATIENT

Antibiotics can have serious adverse effects, and it is important for clinicians to be cognizant of this. Our 60-year-old patient who was taking cefepime and vancomycin for pneumonia developed prolonged atypical absence seizures. When the cefepime was discontinued, his mental status improved, and no other seizures were observed.

A 60-year-old man is admitted for respiratory failure following a massive myocardial infarction. He develops ventilator-associated pneumonia and is treated with cefepime and vancomycin. Three days later, he develops prolonged atypical absence seizures.

What caused these seizures? The neurologist thinks it might be the cefepime. Do you agree?

Antibiotics are widely used in the United States, with 269 million courses of oral therapy prescribed in 2011.¹ Adverse effects such as rash are well known, but rare effects such as seizure, hypoglycemia, and hypoxemia may not be immediately attributed to these drugs.

In this article, we review less-recognized but potentially serious adverse effects of antibiotics commonly prescribed in the United States. We have structured our discussion by organ system for ease of reference.

NERVOUS SYSTEM

The potential adverse effects of antibiotics on the nervous system range from encephalopathy and seizure to nonconvulsive status epilepticus.

Encephalopathy and seizure

Encephalopathy has been reported with penicillins, cephalosporins, sulfamethoxazole-trimethoprim, quinolones, and oxazolidinones such as linezolid.^2,3

Seizures are known to occur with penicillins, cephalosporins, carbapenems, and quinolones.^2–4 For cephalosporins, these effects are more common at higher doses, in elderly patients, and in patients with renal impairment. Carbapenems are associated with seizure activity in elderly patients.^2–4

Encephalopathy and seizure can also occur on a continuum, as is the case with piperacillin-induced encephalopathy, with progressive dysarthria, tremor, and progressive confusion culminating in tonic-clonic seizures.²

Nonconvulsive status epilepticus

Nonconvulsive status epilepticus, marked by prolonged atypical absence seizures, has complicated the use of penicillins, quinolones, clarithromycin, and cephalosporins, specifically cefepime.^2,3,5 Diagnosis can be difficult and requires clinical awareness and confirmation with electroencephalography.

Class-specific neurologic effects

Certain antibiotics have class-specific effects:

Tetracyclines: cranial nerve toxicity, neuromuscular blockade, and intracranial hypertension.²

Sulfamethoxazole-trimethoprim: tremors and psychosis, with visual and auditory hallucinations.⁶

Macrolides: dysequilibrium and potentially irreversible hearing loss.²

Quinolones: orofacial dyskinesia and a Tourette-like syndrome, with a higher incidence reported with newer quinolones.⁷

Linezolid: optic and peripheral neuropathy²; neuropathy can be persistent and can lead to loss of vision. The package insert recommends monitoring visual function in patients taking linezolid for more than 3 months and in any patient reporting visual symptoms.⁸

Linezolid is also associated with serotonin syndrome when combined with a drug that potentiates serotonergic activity, most commonly selective serotonin reuptake inhibitors. The syndrome is characterized by a triad of cognitive or behavioral changes, autonomic instability, and neuromuscular excitability such as spontaneous clonus.⁹

Metronidazole: optic and peripheral neuropathy, in addition to cerebellar toxicity and central nervous system lesions on magnetic resonance imaging of the brain. In a series of 11 cases of cerebellar toxicity, most patients presented with ataxia and dysarthria associated with high total doses of metronidazole, and in most cases, magnetic resonance imaging showed resolution of the lesions upon discontinuation of metronidazole.¹⁰

HEMATOLOGIC AND RHEUMATOLOGIC EFFECTS

Agranulocytosis has been associated with beta-lactams, in most cases with prolonged exposure. In one report, the average exposure before onset of agranulocytosis was 22 days for nafcillin and 25 days for penicillin. For penicillins, more than 50% of cases involved high daily doses.¹¹

Likewise, most episodes of vancomycin-induced neutropenia were reported to occur after 20 days of therapy.¹²

In another study, most cases of drug-induced anemia were due to ceftriaxone and piperacillin.¹³

Drug-induced thrombocytopenia has been described with penicillins, cephalo­sporins, sulfonamides, and vancomycin¹⁴ and is a well-recognized effect of linezolid. The syndrome of drug reaction with eosinophilia and systemic symptoms, a severe and rare adverse reaction, has been reported with minocycline, sulfamethoxazole, and vancomycin.¹⁵

The tetracycline minocycline has been reported to cause drug-induced lupus and polyarteritis nodosa-like vasculitis.¹⁶ Drug-induced lupus presents as myalgias and arthralgias, serositis, constitutional symptoms, and positive antinuclear antibody titers. The effect is not dose-dependent. Penicillin, cefuroxime, and nitrofurantoin have also been implicated.¹⁶

Kermani et al¹⁷ described 9 cases of polyarteritis nodosa, in which 5 patients (56%) had systemic involvement including renal artery microaneurysm, mononeuritis multiplex, and mesenteric vasculitis, and some of these patients also had cutaneous involvement. All patients had positive antineutrophil cytoplasmic antibody in a perinuclear pattern. The median time from start of the minocycline to symptom onset was 9 months, and the median duration of use was 2 years.

Quinolones have also been reported to cause fatal hypersensitivity vasculitis.^18,19

CARDIOVASCULAR SYSTEM

Macrolides and quinolones have been reported to cause QT-interval prolongation and torsades de pointes. The risk is greatest when a  macrolide is co-administered with a CYP3A4 inhibitor.

Of the macrolides, azithromycin is the safest, as clarithromycin and erythromycin are more likely to cause QT prolongation.

While QT prolongation is a class effect of quinolones, there is variability within the class. Ciprofloxacin is thought to be the safest in terms of cardiovascular adverse effects.²⁰ In addition, Owens and Nolin²⁰ reported that quinolone-associated QT prolongation was more likely to occur in patients with pre-existing QT prolongation, electrolyte abnormalities, organic heart disease, and bradycardia, and especially in women. Other risk factors for QT prolongation with quinolone use include underlying cardiac disease and advanced age.²¹

Quinolones have also been associated with an increased risk of aortic dissection. The US Food and Drug Administration has issued a warning advising clinicians to avoid quinolones in patients who have aneurysms or are at risk for aneurysms, such as patients with advanced age, peripheral atherosclerotic vascular disease, hypertension and conditions such as Marfan and Ehlers-Danlos syndrome.²²

DIGESTIVE SYSTEM

Tetracyclines are known to cause esophagitis from direct contact with and disruption of the mucosal lining. Doxycycline is the most frequent offender.²³

Amoxicillin-clavulanate is the antibiotic most commonly associated with drug-induced liver injury, mainly attributable to the clavulanate component.²⁴ It is more common in men over age 50 and with prolonged and repeated dosing and is sometimes fatal. Other adverse effects include Stevens-Johnson syndrome, interstitial nephritis, and thrombotic thrombocytopenic purpura.²⁵

Cholestatic hepatitis has been reported with penicillins, particularly dicloxacillin, oxacillin, and amoxicillin-clavulanate; cephalosporins; doxycycline; sulfamethoxazole-trimethoprim; macrolides; and ciprofloxacin.^24–26 Hepatocellular injury is linked to amoxicillin-clavulanate and doxycycline. Drug-induced mixed liver injury has been observed with amoxicillin-clavulanate, sulfamethoxazole-trimethoprim and, rarely, cephalosporins.

Liver injury is classified as cholestatic if the alkaline phosphatase level is more than 2 times higher than normal, or if the ratio of alanine aminotransferase to alkaline phosphatase is less than 2; if the ratio is greater than 5, the injury is considered hepatocellular.²⁴ Mixed liver injury, the most common, is defined as a ratio from 2 to 5.

Nitrofurantoin has also been linked to hepatotoxicity, cirrhosis, and end-stage liver disease, and to death if the drug is continued after the onset of jaundice.²⁶ Death from liver injury has been reported with amoxicillin-clavulanate, sulfamethoxazole-trimethoprim, and erythromycin, and jaundice indicates a poor prognosis, associated with a 10% mortality rate or need for liver transplant in all patients.²⁴

ENDOCRINE SYSTEM

Clarithromycin, sulfonamides, and quinolones are known to precipitate hypoglycemia by interacting with sulfonylureas. A study of Medicare patients age 66 or older who were taking glipizide or glyburide reported that female sex, older age, and a history of hypoglycemic episodes were associated with antibiotic-related hypoglycemia.²⁷ The odds ratio for hypoglycemia was highest for clarithromycin (3.96), sulfamethoxazole-trimethoprim (2.56), metronidazole (2.11), and ciprofloxacin (1.62) when compared with antibiotics that do not cause hypoglycemia. There was no signal for levofloxacin-mediated hypoglycemia in this series.²⁷

RESPIRATORY SYSTEM

Hypersensitivity lung disease has been reported with penicillin, ampicillin, cephalosporins, ciprofloxacin, and sulfonamides including sulfamethoxazole-trimethoprim.²⁸ The lipopeptide daptomycin has been reported to cause acute eosinophilic pneumonia defined as fever for less than 5 days, pulmonary infiltrates, hypoxemia, and a bronchoalveolar lavage or biopsy study with eosinophils. Daptomycin should be stopped early in these cases, and the patient should not be rechallenged, as the reaction can be deadly.²⁹

Nitrofurantoin has a long history of hypersensitivity pneumonitis in its acute form and a chronic allergic response. While more widely recognized, nitrofurantoin pulmonary toxicity is rare, occurring in 1 in 5,000 patients.³⁰

RENAL SYSTEM

Acute interstitial nephritis has been reported with penicillins, cephalosporins, macrolides, quinolones, sulfonamides, and vancomycin.^31–33 Acute tubular necrosis has been linked to cephalosporins and tetracyclines. Crystal nephropathy has been seen with quinolones and sulfonamides.

Advanced age is an important risk factor for renal dysfunction from quinolones,¹⁸ and penicillin G has been reported to cause glomerulonephritis.³¹

MUSCULOSKELETAL SYSTEM

Quinolones have been associated with arthropathy or tendinitis at a rate of 1%, including cases of Achilles tendon rupture.¹⁸ The US Food and Drug Administration announced in 2016 that the serious adverse events with fluoroquinolones outweigh the benefits in patients with acute sinusitis, acute bronchitis, and uncomplicated urinary tract infection, and that they should be used only if there are no other options.³⁴

Daptomycin is known to cause elevations of creatine kinase.³⁴ Weekly monitoring is recommended based on postmarketing data reports of elevations in 2.5% of patients; myopathy is a rarer effect, occurring in 0.2% of patients.³⁵

REPRODUCTIVE SYSTEM

Antibiotics have long been reported to interact with oral contraceptives, but the data are not compelling for commonly used antibiotics. The strongest association is with rifampicin, which reduces oral contraceptive efficacy and warrants an alternative mode of contraception.³⁶

BACK TO OUR PATIENT

Antibiotics can have serious adverse effects, and it is important for clinicians to be cognizant of this. Our 60-year-old patient who was taking cefepime and vancomycin for pneumonia developed prolonged atypical absence seizures. When the cefepime was discontinued, his mental status improved, and no other seizures were observed.

A 60-year-old man is admitted for respiratory failure following a massive myocardial infarction. He develops ventilator-associated pneumonia and is treated with cefepime and vancomycin. Three days later, he develops prolonged atypical absence seizures.

What caused these seizures? The neurologist thinks it might be the cefepime. Do you agree?

Antibiotics are widely used in the United States, with 269 million courses of oral therapy prescribed in 2011.¹ Adverse effects such as rash are well known, but rare effects such as seizure, hypoglycemia, and hypoxemia may not be immediately attributed to these drugs.

In this article, we review less-recognized but potentially serious adverse effects of antibiotics commonly prescribed in the United States. We have structured our discussion by organ system for ease of reference.

NERVOUS SYSTEM

The potential adverse effects of antibiotics on the nervous system range from encephalopathy and seizure to nonconvulsive status epilepticus.

Encephalopathy and seizure

Encephalopathy has been reported with penicillins, cephalosporins, sulfamethoxazole-trimethoprim, quinolones, and oxazolidinones such as linezolid.^2,3

Seizures are known to occur with penicillins, cephalosporins, carbapenems, and quinolones.^2–4 For cephalosporins, these effects are more common at higher doses, in elderly patients, and in patients with renal impairment. Carbapenems are associated with seizure activity in elderly patients.^2–4

Encephalopathy and seizure can also occur on a continuum, as is the case with piperacillin-induced encephalopathy, with progressive dysarthria, tremor, and progressive confusion culminating in tonic-clonic seizures.²

Nonconvulsive status epilepticus

Nonconvulsive status epilepticus, marked by prolonged atypical absence seizures, has complicated the use of penicillins, quinolones, clarithromycin, and cephalosporins, specifically cefepime.^2,3,5 Diagnosis can be difficult and requires clinical awareness and confirmation with electroencephalography.

Class-specific neurologic effects

Certain antibiotics have class-specific effects:

Tetracyclines: cranial nerve toxicity, neuromuscular blockade, and intracranial hypertension.²

Sulfamethoxazole-trimethoprim: tremors and psychosis, with visual and auditory hallucinations.⁶

Macrolides: dysequilibrium and potentially irreversible hearing loss.²

Quinolones: orofacial dyskinesia and a Tourette-like syndrome, with a higher incidence reported with newer quinolones.⁷

Linezolid: optic and peripheral neuropathy²; neuropathy can be persistent and can lead to loss of vision. The package insert recommends monitoring visual function in patients taking linezolid for more than 3 months and in any patient reporting visual symptoms.⁸

Linezolid is also associated with serotonin syndrome when combined with a drug that potentiates serotonergic activity, most commonly selective serotonin reuptake inhibitors. The syndrome is characterized by a triad of cognitive or behavioral changes, autonomic instability, and neuromuscular excitability such as spontaneous clonus.⁹

Metronidazole: optic and peripheral neuropathy, in addition to cerebellar toxicity and central nervous system lesions on magnetic resonance imaging of the brain. In a series of 11 cases of cerebellar toxicity, most patients presented with ataxia and dysarthria associated with high total doses of metronidazole, and in most cases, magnetic resonance imaging showed resolution of the lesions upon discontinuation of metronidazole.¹⁰

HEMATOLOGIC AND RHEUMATOLOGIC EFFECTS

Agranulocytosis has been associated with beta-lactams, in most cases with prolonged exposure. In one report, the average exposure before onset of agranulocytosis was 22 days for nafcillin and 25 days for penicillin. For penicillins, more than 50% of cases involved high daily doses.¹¹

Likewise, most episodes of vancomycin-induced neutropenia were reported to occur after 20 days of therapy.¹²

In another study, most cases of drug-induced anemia were due to ceftriaxone and piperacillin.¹³

Drug-induced thrombocytopenia has been described with penicillins, cephalo­sporins, sulfonamides, and vancomycin¹⁴ and is a well-recognized effect of linezolid. The syndrome of drug reaction with eosinophilia and systemic symptoms, a severe and rare adverse reaction, has been reported with minocycline, sulfamethoxazole, and vancomycin.¹⁵

The tetracycline minocycline has been reported to cause drug-induced lupus and polyarteritis nodosa-like vasculitis.¹⁶ Drug-induced lupus presents as myalgias and arthralgias, serositis, constitutional symptoms, and positive antinuclear antibody titers. The effect is not dose-dependent. Penicillin, cefuroxime, and nitrofurantoin have also been implicated.¹⁶

Kermani et al¹⁷ described 9 cases of polyarteritis nodosa, in which 5 patients (56%) had systemic involvement including renal artery microaneurysm, mononeuritis multiplex, and mesenteric vasculitis, and some of these patients also had cutaneous involvement. All patients had positive antineutrophil cytoplasmic antibody in a perinuclear pattern. The median time from start of the minocycline to symptom onset was 9 months, and the median duration of use was 2 years.

Quinolones have also been reported to cause fatal hypersensitivity vasculitis.^18,19

CARDIOVASCULAR SYSTEM

Macrolides and quinolones have been reported to cause QT-interval prolongation and torsades de pointes. The risk is greatest when a  macrolide is co-administered with a CYP3A4 inhibitor.

Of the macrolides, azithromycin is the safest, as clarithromycin and erythromycin are more likely to cause QT prolongation.

While QT prolongation is a class effect of quinolones, there is variability within the class. Ciprofloxacin is thought to be the safest in terms of cardiovascular adverse effects.²⁰ In addition, Owens and Nolin²⁰ reported that quinolone-associated QT prolongation was more likely to occur in patients with pre-existing QT prolongation, electrolyte abnormalities, organic heart disease, and bradycardia, and especially in women. Other risk factors for QT prolongation with quinolone use include underlying cardiac disease and advanced age.²¹

Quinolones have also been associated with an increased risk of aortic dissection. The US Food and Drug Administration has issued a warning advising clinicians to avoid quinolones in patients who have aneurysms or are at risk for aneurysms, such as patients with advanced age, peripheral atherosclerotic vascular disease, hypertension and conditions such as Marfan and Ehlers-Danlos syndrome.²²

DIGESTIVE SYSTEM

Tetracyclines are known to cause esophagitis from direct contact with and disruption of the mucosal lining. Doxycycline is the most frequent offender.²³

Amoxicillin-clavulanate is the antibiotic most commonly associated with drug-induced liver injury, mainly attributable to the clavulanate component.²⁴ It is more common in men over age 50 and with prolonged and repeated dosing and is sometimes fatal. Other adverse effects include Stevens-Johnson syndrome, interstitial nephritis, and thrombotic thrombocytopenic purpura.²⁵

Cholestatic hepatitis has been reported with penicillins, particularly dicloxacillin, oxacillin, and amoxicillin-clavulanate; cephalosporins; doxycycline; sulfamethoxazole-trimethoprim; macrolides; and ciprofloxacin.^24–26 Hepatocellular injury is linked to amoxicillin-clavulanate and doxycycline. Drug-induced mixed liver injury has been observed with amoxicillin-clavulanate, sulfamethoxazole-trimethoprim and, rarely, cephalosporins.

Liver injury is classified as cholestatic if the alkaline phosphatase level is more than 2 times higher than normal, or if the ratio of alanine aminotransferase to alkaline phosphatase is less than 2; if the ratio is greater than 5, the injury is considered hepatocellular.²⁴ Mixed liver injury, the most common, is defined as a ratio from 2 to 5.

Nitrofurantoin has also been linked to hepatotoxicity, cirrhosis, and end-stage liver disease, and to death if the drug is continued after the onset of jaundice.²⁶ Death from liver injury has been reported with amoxicillin-clavulanate, sulfamethoxazole-trimethoprim, and erythromycin, and jaundice indicates a poor prognosis, associated with a 10% mortality rate or need for liver transplant in all patients.²⁴

ENDOCRINE SYSTEM

Clarithromycin, sulfonamides, and quinolones are known to precipitate hypoglycemia by interacting with sulfonylureas. A study of Medicare patients age 66 or older who were taking glipizide or glyburide reported that female sex, older age, and a history of hypoglycemic episodes were associated with antibiotic-related hypoglycemia.²⁷ The odds ratio for hypoglycemia was highest for clarithromycin (3.96), sulfamethoxazole-trimethoprim (2.56), metronidazole (2.11), and ciprofloxacin (1.62) when compared with antibiotics that do not cause hypoglycemia. There was no signal for levofloxacin-mediated hypoglycemia in this series.²⁷

RESPIRATORY SYSTEM

Hypersensitivity lung disease has been reported with penicillin, ampicillin, cephalosporins, ciprofloxacin, and sulfonamides including sulfamethoxazole-trimethoprim.²⁸ The lipopeptide daptomycin has been reported to cause acute eosinophilic pneumonia defined as fever for less than 5 days, pulmonary infiltrates, hypoxemia, and a bronchoalveolar lavage or biopsy study with eosinophils. Daptomycin should be stopped early in these cases, and the patient should not be rechallenged, as the reaction can be deadly.²⁹

Nitrofurantoin has a long history of hypersensitivity pneumonitis in its acute form and a chronic allergic response. While more widely recognized, nitrofurantoin pulmonary toxicity is rare, occurring in 1 in 5,000 patients.³⁰

RENAL SYSTEM

Acute interstitial nephritis has been reported with penicillins, cephalosporins, macrolides, quinolones, sulfonamides, and vancomycin.^31–33 Acute tubular necrosis has been linked to cephalosporins and tetracyclines. Crystal nephropathy has been seen with quinolones and sulfonamides.

Advanced age is an important risk factor for renal dysfunction from quinolones,¹⁸ and penicillin G has been reported to cause glomerulonephritis.³¹

MUSCULOSKELETAL SYSTEM

Quinolones have been associated with arthropathy or tendinitis at a rate of 1%, including cases of Achilles tendon rupture.¹⁸ The US Food and Drug Administration announced in 2016 that the serious adverse events with fluoroquinolones outweigh the benefits in patients with acute sinusitis, acute bronchitis, and uncomplicated urinary tract infection, and that they should be used only if there are no other options.³⁴

Daptomycin is known to cause elevations of creatine kinase.³⁴ Weekly monitoring is recommended based on postmarketing data reports of elevations in 2.5% of patients; myopathy is a rarer effect, occurring in 0.2% of patients.³⁵

REPRODUCTIVE SYSTEM

Antibiotics have long been reported to interact with oral contraceptives, but the data are not compelling for commonly used antibiotics. The strongest association is with rifampicin, which reduces oral contraceptive efficacy and warrants an alternative mode of contraception.³⁶

BACK TO OUR PATIENT

Antibiotics can have serious adverse effects, and it is important for clinicians to be cognizant of this. Our 60-year-old patient who was taking cefepime and vancomycin for pneumonia developed prolonged atypical absence seizures. When the cefepime was discontinued, his mental status improved, and no other seizures were observed.

Acute agitation in the pregnant patient should be treated as an obstetric emergency, as it jeopardizes the safety of the patient and fetus, as well as others in the emergency room. Uncontrolled agitation is associated with obstetric complications such as preterm delivery, placental abnormalities, postnatal death, and spontaneous abortion.¹

Current data on the reproductive safety of drugs commonly used to treat acute agitation—benzodiazepines, typical (first-generation) antipsychotics, atypical (second-generation) antipsychotics, and diphenhydramine—suggest no increase in risk beyond the 2% to 3% risk of congenital malformations in the general population when used in the first trimester.^2,3

FOCUS OF THE EMERGENCY EVALUATION

Agitation is defined as the physical manifestation of internal distress, due to an underlying medical condition such as delirium or to a psychiatric condition such as acute intoxication or withdrawal, psychosis, mania, or personality disorder.⁴

For the agitated pregnant woman who is not belligerent at presentation, triage should start with a basic assessment of airways, breathing, and circulation, as well as vital signs and glucose level.⁵ A thorough medical history and a description of events leading to the presentation, obtained from the patient or the patient’s family or friends, are vital for narrowing the diagnosis and deciding treatment.

The initial evaluation should include consideration of delirium, trauma, intracranial hemorrhage, coagulopathy, thrombocytopenia, amniotic and venous thromboembolism, hypoxia and hypercapnia, and signs and symptoms of intoxication or withdrawal from substances such as alcohol, cocaine, phencyclidine, methamphetamine, and substituted cathinones (“bath salts”). From 20 weeks of gestation to 6 weeks postpartum, eclampsia should also be considered in the differential diagnosis.¹ Ruling out these conditions is important since the management of each differs vastly from the protocol for agitation secondary to psychosis, mania, or delirium.

NEW SYSTEM TO DETERMINE RISK DURING PREGNANCY, LACTATION

The US Food and Drug Administration (FDA) has discontinued its pregnancy category labeling system that used the letters A, B, C, D, and X to convey reproductive and lactation safety. The new system, established under the FDA Pregnancy and Lactation Labeling Rule,⁶ provides descriptive, up-to-date explanations of risk, as well as previously absent context regarding baseline risk for major malformations in the general population to help with informed decision-making.⁷ This allows the healthcare provider to interpret the risk for an individual patient.

FIRST-GENERATION ANTIPSYCHOTICS SAFE, EFFECTIVE IN PREGNANCY

Reproductive safety of first-generation (ie, typical) neuroleptics such as haloperidol is supported by extensive data accumulated over the past 50 years.^2,3,8 No significant teratogenic effect has been documented with this drug class,⁷ although a 1996 meta-analysis found a small increase in the relative risk of congenital malformations in offspring exposed to low-potency antipsychotics compared with those exposed to high-potency antipsychotics.²

In general, mid- and high-potency antipsychotics (eg, haloperidol, perphenazine) are often recommended because they are less likely to have associated sedative or hypotensive effects than low-potency antipsychotics (eg, chlorpromazine, perphenazine), which may be a significant consideration for a pregnant patient.^2,8

There is a theoretical risk of neonatal extrapyramidal symptoms with exposure to first-generation antipsychotics in the third trimester, but the data to support this are from sparse case reports and small observational cohorts.⁹

NEWER ANTIPSYCHOTICS ALSO SAFE IN PREGNANCY

Newer antipsychotics such as the second-generation antipsychotics, available since the mid-1990s, are increasingly used as primary or adjunctive therapy across a wide range of psychiatric disorders.¹⁰ Recent data from large, prospective cohort studies investigating reproductive safety of these agents are reassuring, with no specific patterns of organ malformation.^11,12

DIPHENHYDRAMINE

Recent studies of antihistamines such as diphenhydramine have not reported any risk of major malformations with first-trimester exposure to antihistamines.^13,14 Dose-dependent anticholinergic adverse effects of antihistamines can induce or exacerbate delirium and agitation, although these effects are classically seen in elderly, nonpregnant patients.¹⁵ Thus, given the paucity of adverse effects and the low risk, diphenhydramine is considered safe to use in pregnancy.¹³

Benzodiazepines are not contraindicated for the treatment of acute agitation in pregnancy.¹⁶ Reproductive safety data from meta-analyses and large population-based cohort studies have found no evidence of increased risk of major malformations in neonates born to mothers on prescription benzodiazepines in the first trimester.^17,18 While third-trimester exposure to benzodiazepines has been associated with “floppy-baby” syndrome and neonatal withdrawal syndrome,¹⁶ these are more likely to occur in women on long-term prescription benzodiazepine therapy. No study has yet assessed the risk of these outcomes with a 1-time acute exposure in the emergency department; however, the risk is likely minimal given the aforementioned data observed in women on long-term prescription benzodiazepine therapy.

STEPWISE MANAGEMENT OF AGITATION IN PREGNANCY

If untreated, agitation in pregnancy is independently associated with outcomes that include premature delivery, low birth weight, growth retardation, postnatal death, and spontaneous abortion.¹ The risk of these outcomes greatly outweighs any potential risk from psychotropic medications during pregnancy.

Nevertheless, intervention should progress in a stepwise manner, starting with the least restrictive and progressing toward more restrictive interventions, including pharmacotherapy, use of a seclusion room, and physical restraints (Figure 1).^4,19

Before medications are considered, attempts should be made to engage with and “de-escalate” the patient in a safe, nonstimulating environment.¹⁹ If this approach is not effective, the patient should be offered oral medications to help with her agitation. However, if the patient’s behavior continues to escalate, presenting a danger to herself or staff, the use of emergency medications is clearly indicated. Providers should succinctly inform the patient of the need for immediate intervention.

If the patient has had a good response in the past to one of these medications or is currently taking one as needed, the same medication should be offered. If the patient has never been treated for agitation, it is important to consider the presenting symptoms, differential diagnosis, and the route and rapidity of administration of medication. If the patient has experienced a fall or other trauma, confirming a viable fetal heart rate between 10 to 22 weeks of gestation with Doppler ultrasonography and obstetric consultation should be considered.

DRUG THERAPY RECOMMENDATIONS

Mild to moderate agitation in pregnancy should be managed conservatively with diphenhydramine. Other options include a benzodiazepine, particularly lorazepam, if alcohol withdrawal is suspected. A second-generation antipsychotic such as olanzapine in a rapidly dissolving form or ziprasidone is another option if a rapid response is required.²⁰Table 1 provides a summary of pharmacotherapy recommendations.

Severe agitation may require a combination of agents. A commonly used, safe regimen—colloquially called the “B52 bomb”—is haloperidol 5 mg, lorazepam 2 mg, and diphenhydramine 25 to 50 mg for prophylaxis of dystonia.²⁰

The patient’s response should be monitored closely, as dosing may require modification as a result of pregnancy-related changes in drug distribution, metabolism, and clearance.²¹

Although no study to our knowledge has assessed risk associated with 1-time exposure to any of these classes of medications in pregnant women, the aforementioned data on long-term exposure provide reassurance that single exposure in emergency departments likely has little or no effect for the developing fetus.

PHYSICAL RESTRAINTS FOR AGITATION IN PREGNANCY

Physical restraints along with emergency medications (ie, chemical restraint) may be indicated when the patient poses a danger to herself or others. In some cases, both types of restraint may be required, whether in the emergency room or an inpatient setting.

However, during the second and third trimesters, physical restraints such as 4-point restraints may predispose the patient to inferior vena cava compression syndrome and compromise placental blood flow.⁴ Therefore, pregnant patients after 20 weeks of gestation should be positioned in the left lateral decubitus position, with the right hip positioned 10 to 12 cm off the bed with pillows or blankets. And when restraints are used in pregnant patients, frequent checking of vital signs and physical assessment is needed to mitigate risks.⁴

Acute agitation in the pregnant patient should be treated as an obstetric emergency, as it jeopardizes the safety of the patient and fetus, as well as others in the emergency room. Uncontrolled agitation is associated with obstetric complications such as preterm delivery, placental abnormalities, postnatal death, and spontaneous abortion.¹

Current data on the reproductive safety of drugs commonly used to treat acute agitation—benzodiazepines, typical (first-generation) antipsychotics, atypical (second-generation) antipsychotics, and diphenhydramine—suggest no increase in risk beyond the 2% to 3% risk of congenital malformations in the general population when used in the first trimester.^2,3

FOCUS OF THE EMERGENCY EVALUATION

Agitation is defined as the physical manifestation of internal distress, due to an underlying medical condition such as delirium or to a psychiatric condition such as acute intoxication or withdrawal, psychosis, mania, or personality disorder.⁴

For the agitated pregnant woman who is not belligerent at presentation, triage should start with a basic assessment of airways, breathing, and circulation, as well as vital signs and glucose level.⁵ A thorough medical history and a description of events leading to the presentation, obtained from the patient or the patient’s family or friends, are vital for narrowing the diagnosis and deciding treatment.

The initial evaluation should include consideration of delirium, trauma, intracranial hemorrhage, coagulopathy, thrombocytopenia, amniotic and venous thromboembolism, hypoxia and hypercapnia, and signs and symptoms of intoxication or withdrawal from substances such as alcohol, cocaine, phencyclidine, methamphetamine, and substituted cathinones (“bath salts”). From 20 weeks of gestation to 6 weeks postpartum, eclampsia should also be considered in the differential diagnosis.¹ Ruling out these conditions is important since the management of each differs vastly from the protocol for agitation secondary to psychosis, mania, or delirium.

NEW SYSTEM TO DETERMINE RISK DURING PREGNANCY, LACTATION

The US Food and Drug Administration (FDA) has discontinued its pregnancy category labeling system that used the letters A, B, C, D, and X to convey reproductive and lactation safety. The new system, established under the FDA Pregnancy and Lactation Labeling Rule,⁶ provides descriptive, up-to-date explanations of risk, as well as previously absent context regarding baseline risk for major malformations in the general population to help with informed decision-making.⁷ This allows the healthcare provider to interpret the risk for an individual patient.

FIRST-GENERATION ANTIPSYCHOTICS SAFE, EFFECTIVE IN PREGNANCY

Reproductive safety of first-generation (ie, typical) neuroleptics such as haloperidol is supported by extensive data accumulated over the past 50 years.^2,3,8 No significant teratogenic effect has been documented with this drug class,⁷ although a 1996 meta-analysis found a small increase in the relative risk of congenital malformations in offspring exposed to low-potency antipsychotics compared with those exposed to high-potency antipsychotics.²

In general, mid- and high-potency antipsychotics (eg, haloperidol, perphenazine) are often recommended because they are less likely to have associated sedative or hypotensive effects than low-potency antipsychotics (eg, chlorpromazine, perphenazine), which may be a significant consideration for a pregnant patient.^2,8

There is a theoretical risk of neonatal extrapyramidal symptoms with exposure to first-generation antipsychotics in the third trimester, but the data to support this are from sparse case reports and small observational cohorts.⁹

NEWER ANTIPSYCHOTICS ALSO SAFE IN PREGNANCY

Newer antipsychotics such as the second-generation antipsychotics, available since the mid-1990s, are increasingly used as primary or adjunctive therapy across a wide range of psychiatric disorders.¹⁰ Recent data from large, prospective cohort studies investigating reproductive safety of these agents are reassuring, with no specific patterns of organ malformation.^11,12

DIPHENHYDRAMINE

Recent studies of antihistamines such as diphenhydramine have not reported any risk of major malformations with first-trimester exposure to antihistamines.^13,14 Dose-dependent anticholinergic adverse effects of antihistamines can induce or exacerbate delirium and agitation, although these effects are classically seen in elderly, nonpregnant patients.¹⁵ Thus, given the paucity of adverse effects and the low risk, diphenhydramine is considered safe to use in pregnancy.¹³

Benzodiazepines are not contraindicated for the treatment of acute agitation in pregnancy.¹⁶ Reproductive safety data from meta-analyses and large population-based cohort studies have found no evidence of increased risk of major malformations in neonates born to mothers on prescription benzodiazepines in the first trimester.^17,18 While third-trimester exposure to benzodiazepines has been associated with “floppy-baby” syndrome and neonatal withdrawal syndrome,¹⁶ these are more likely to occur in women on long-term prescription benzodiazepine therapy. No study has yet assessed the risk of these outcomes with a 1-time acute exposure in the emergency department; however, the risk is likely minimal given the aforementioned data observed in women on long-term prescription benzodiazepine therapy.

STEPWISE MANAGEMENT OF AGITATION IN PREGNANCY

If untreated, agitation in pregnancy is independently associated with outcomes that include premature delivery, low birth weight, growth retardation, postnatal death, and spontaneous abortion.¹ The risk of these outcomes greatly outweighs any potential risk from psychotropic medications during pregnancy.

Nevertheless, intervention should progress in a stepwise manner, starting with the least restrictive and progressing toward more restrictive interventions, including pharmacotherapy, use of a seclusion room, and physical restraints (Figure 1).^4,19

Before medications are considered, attempts should be made to engage with and “de-escalate” the patient in a safe, nonstimulating environment.¹⁹ If this approach is not effective, the patient should be offered oral medications to help with her agitation. However, if the patient’s behavior continues to escalate, presenting a danger to herself or staff, the use of emergency medications is clearly indicated. Providers should succinctly inform the patient of the need for immediate intervention.

If the patient has had a good response in the past to one of these medications or is currently taking one as needed, the same medication should be offered. If the patient has never been treated for agitation, it is important to consider the presenting symptoms, differential diagnosis, and the route and rapidity of administration of medication. If the patient has experienced a fall or other trauma, confirming a viable fetal heart rate between 10 to 22 weeks of gestation with Doppler ultrasonography and obstetric consultation should be considered.

DRUG THERAPY RECOMMENDATIONS

Mild to moderate agitation in pregnancy should be managed conservatively with diphenhydramine. Other options include a benzodiazepine, particularly lorazepam, if alcohol withdrawal is suspected. A second-generation antipsychotic such as olanzapine in a rapidly dissolving form or ziprasidone is another option if a rapid response is required.²⁰Table 1 provides a summary of pharmacotherapy recommendations.

Severe agitation may require a combination of agents. A commonly used, safe regimen—colloquially called the “B52 bomb”—is haloperidol 5 mg, lorazepam 2 mg, and diphenhydramine 25 to 50 mg for prophylaxis of dystonia.²⁰

The patient’s response should be monitored closely, as dosing may require modification as a result of pregnancy-related changes in drug distribution, metabolism, and clearance.²¹

Although no study to our knowledge has assessed risk associated with 1-time exposure to any of these classes of medications in pregnant women, the aforementioned data on long-term exposure provide reassurance that single exposure in emergency departments likely has little or no effect for the developing fetus.

PHYSICAL RESTRAINTS FOR AGITATION IN PREGNANCY

Physical restraints along with emergency medications (ie, chemical restraint) may be indicated when the patient poses a danger to herself or others. In some cases, both types of restraint may be required, whether in the emergency room or an inpatient setting.

However, during the second and third trimesters, physical restraints such as 4-point restraints may predispose the patient to inferior vena cava compression syndrome and compromise placental blood flow.⁴ Therefore, pregnant patients after 20 weeks of gestation should be positioned in the left lateral decubitus position, with the right hip positioned 10 to 12 cm off the bed with pillows or blankets. And when restraints are used in pregnant patients, frequent checking of vital signs and physical assessment is needed to mitigate risks.⁴

Acute agitation in the pregnant patient should be treated as an obstetric emergency, as it jeopardizes the safety of the patient and fetus, as well as others in the emergency room. Uncontrolled agitation is associated with obstetric complications such as preterm delivery, placental abnormalities, postnatal death, and spontaneous abortion.¹

Current data on the reproductive safety of drugs commonly used to treat acute agitation—benzodiazepines, typical (first-generation) antipsychotics, atypical (second-generation) antipsychotics, and diphenhydramine—suggest no increase in risk beyond the 2% to 3% risk of congenital malformations in the general population when used in the first trimester.^2,3

FOCUS OF THE EMERGENCY EVALUATION

Agitation is defined as the physical manifestation of internal distress, due to an underlying medical condition such as delirium or to a psychiatric condition such as acute intoxication or withdrawal, psychosis, mania, or personality disorder.⁴

For the agitated pregnant woman who is not belligerent at presentation, triage should start with a basic assessment of airways, breathing, and circulation, as well as vital signs and glucose level.⁵ A thorough medical history and a description of events leading to the presentation, obtained from the patient or the patient’s family or friends, are vital for narrowing the diagnosis and deciding treatment.

The initial evaluation should include consideration of delirium, trauma, intracranial hemorrhage, coagulopathy, thrombocytopenia, amniotic and venous thromboembolism, hypoxia and hypercapnia, and signs and symptoms of intoxication or withdrawal from substances such as alcohol, cocaine, phencyclidine, methamphetamine, and substituted cathinones (“bath salts”). From 20 weeks of gestation to 6 weeks postpartum, eclampsia should also be considered in the differential diagnosis.¹ Ruling out these conditions is important since the management of each differs vastly from the protocol for agitation secondary to psychosis, mania, or delirium.

NEW SYSTEM TO DETERMINE RISK DURING PREGNANCY, LACTATION

The US Food and Drug Administration (FDA) has discontinued its pregnancy category labeling system that used the letters A, B, C, D, and X to convey reproductive and lactation safety. The new system, established under the FDA Pregnancy and Lactation Labeling Rule,⁶ provides descriptive, up-to-date explanations of risk, as well as previously absent context regarding baseline risk for major malformations in the general population to help with informed decision-making.⁷ This allows the healthcare provider to interpret the risk for an individual patient.

FIRST-GENERATION ANTIPSYCHOTICS SAFE, EFFECTIVE IN PREGNANCY

Reproductive safety of first-generation (ie, typical) neuroleptics such as haloperidol is supported by extensive data accumulated over the past 50 years.^2,3,8 No significant teratogenic effect has been documented with this drug class,⁷ although a 1996 meta-analysis found a small increase in the relative risk of congenital malformations in offspring exposed to low-potency antipsychotics compared with those exposed to high-potency antipsychotics.²

In general, mid- and high-potency antipsychotics (eg, haloperidol, perphenazine) are often recommended because they are less likely to have associated sedative or hypotensive effects than low-potency antipsychotics (eg, chlorpromazine, perphenazine), which may be a significant consideration for a pregnant patient.^2,8

There is a theoretical risk of neonatal extrapyramidal symptoms with exposure to first-generation antipsychotics in the third trimester, but the data to support this are from sparse case reports and small observational cohorts.⁹

NEWER ANTIPSYCHOTICS ALSO SAFE IN PREGNANCY

Newer antipsychotics such as the second-generation antipsychotics, available since the mid-1990s, are increasingly used as primary or adjunctive therapy across a wide range of psychiatric disorders.¹⁰ Recent data from large, prospective cohort studies investigating reproductive safety of these agents are reassuring, with no specific patterns of organ malformation.^11,12

DIPHENHYDRAMINE

Recent studies of antihistamines such as diphenhydramine have not reported any risk of major malformations with first-trimester exposure to antihistamines.^13,14 Dose-dependent anticholinergic adverse effects of antihistamines can induce or exacerbate delirium and agitation, although these effects are classically seen in elderly, nonpregnant patients.¹⁵ Thus, given the paucity of adverse effects and the low risk, diphenhydramine is considered safe to use in pregnancy.¹³

Benzodiazepines are not contraindicated for the treatment of acute agitation in pregnancy.¹⁶ Reproductive safety data from meta-analyses and large population-based cohort studies have found no evidence of increased risk of major malformations in neonates born to mothers on prescription benzodiazepines in the first trimester.^17,18 While third-trimester exposure to benzodiazepines has been associated with “floppy-baby” syndrome and neonatal withdrawal syndrome,¹⁶ these are more likely to occur in women on long-term prescription benzodiazepine therapy. No study has yet assessed the risk of these outcomes with a 1-time acute exposure in the emergency department; however, the risk is likely minimal given the aforementioned data observed in women on long-term prescription benzodiazepine therapy.

STEPWISE MANAGEMENT OF AGITATION IN PREGNANCY

If untreated, agitation in pregnancy is independently associated with outcomes that include premature delivery, low birth weight, growth retardation, postnatal death, and spontaneous abortion.¹ The risk of these outcomes greatly outweighs any potential risk from psychotropic medications during pregnancy.

Nevertheless, intervention should progress in a stepwise manner, starting with the least restrictive and progressing toward more restrictive interventions, including pharmacotherapy, use of a seclusion room, and physical restraints (Figure 1).^4,19

Before medications are considered, attempts should be made to engage with and “de-escalate” the patient in a safe, nonstimulating environment.¹⁹ If this approach is not effective, the patient should be offered oral medications to help with her agitation. However, if the patient’s behavior continues to escalate, presenting a danger to herself or staff, the use of emergency medications is clearly indicated. Providers should succinctly inform the patient of the need for immediate intervention.

If the patient has had a good response in the past to one of these medications or is currently taking one as needed, the same medication should be offered. If the patient has never been treated for agitation, it is important to consider the presenting symptoms, differential diagnosis, and the route and rapidity of administration of medication. If the patient has experienced a fall or other trauma, confirming a viable fetal heart rate between 10 to 22 weeks of gestation with Doppler ultrasonography and obstetric consultation should be considered.

DRUG THERAPY RECOMMENDATIONS

Mild to moderate agitation in pregnancy should be managed conservatively with diphenhydramine. Other options include a benzodiazepine, particularly lorazepam, if alcohol withdrawal is suspected. A second-generation antipsychotic such as olanzapine in a rapidly dissolving form or ziprasidone is another option if a rapid response is required.²⁰Table 1 provides a summary of pharmacotherapy recommendations.

Severe agitation may require a combination of agents. A commonly used, safe regimen—colloquially called the “B52 bomb”—is haloperidol 5 mg, lorazepam 2 mg, and diphenhydramine 25 to 50 mg for prophylaxis of dystonia.²⁰

The patient’s response should be monitored closely, as dosing may require modification as a result of pregnancy-related changes in drug distribution, metabolism, and clearance.²¹

Although no study to our knowledge has assessed risk associated with 1-time exposure to any of these classes of medications in pregnant women, the aforementioned data on long-term exposure provide reassurance that single exposure in emergency departments likely has little or no effect for the developing fetus.

PHYSICAL RESTRAINTS FOR AGITATION IN PREGNANCY

Physical restraints along with emergency medications (ie, chemical restraint) may be indicated when the patient poses a danger to herself or others. In some cases, both types of restraint may be required, whether in the emergency room or an inpatient setting.

However, during the second and third trimesters, physical restraints such as 4-point restraints may predispose the patient to inferior vena cava compression syndrome and compromise placental blood flow.⁴ Therefore, pregnant patients after 20 weeks of gestation should be positioned in the left lateral decubitus position, with the right hip positioned 10 to 12 cm off the bed with pillows or blankets. And when restraints are used in pregnant patients, frequent checking of vital signs and physical assessment is needed to mitigate risks.⁴

Click here to read the supplement.

What can be done to provide a cost-effective treatment option for von Willebrand Disease?

Topics include:

• Review of Key Clinical Studies and Surgical Procedures
• Pharmacokinetics Comparison
• Pharmacoeconomic Comparisons

Click here to read the supplement.

Click here to read the supplement.

What can be done to provide a cost-effective treatment option for von Willebrand Disease?

Topics include:

• Review of Key Clinical Studies and Surgical Procedures
• Pharmacokinetics Comparison
• Pharmacoeconomic Comparisons

Click here to read the supplement.

Click here to read the supplement.

What can be done to provide a cost-effective treatment option for von Willebrand Disease?

Topics include:

• Review of Key Clinical Studies and Surgical Procedures
• Pharmacokinetics Comparison
• Pharmacoeconomic Comparisons

Click here to read the supplement.

From the Division of Rheumatology & Immunology, University of Nebraska Medical Center, and Veterans Affairs Nebraska-Western Iowa Health Care System, Omaha, NE.

Abstract

• Objective: To review the management of traditional and nontraditional CVD cardiovascular disease risk factors in rheumatoid arthritis (RA).
• Methods: Literature review of the management of CVD risk in RA.
• Results: Because of the increased risk of CVD events and CVD mortality among RA patients, aggressive management of CVD risk is essential. Providers should follow national guidelines for the management of traditional CVD risk factors, including dyslipidemia, hypertension, and diabetes mellitus. Similar efforts are needed in counseling on lifestyle modifications, including smoking cessation, regular exercise, and maintaining a healthy body weight. Because higher RA disease activity is also linked with CVD risk, aggressive treatment of RA to a target of low disease activity or remission is critical. Furthermore, the selection of potentially “cardioprotective” agents such as methotrexate and tumor necrosis factor inhibitors, while limiting use of nonsteroidal anti-inflammatory drugs and glucocorticoids, are strategies that could be employed by rheumatologists to help mitigate CVD risk in their patients with RA.
• Conclusion: Routine assessment of CVD risk, management of traditional CVD risk factors, counseling on healthy lifestyle habits, and aggressive treatment of RA are essential to minimize CVD risk in this population.

Keywords: rheumatoid arthritis; cardiovascular disease; cardiovascular risk assessment; cardiovascular risk management.

Editor’s note: This article is part 2 of a 2-part article. “Assessment of Cardiovascular Disease Risk in Rheumatoid Arthritis” was published in the January/February 2019 issue.

Rheumatoid arthritis (RA) is a systemic autoimmune condition that contributes to an increased risk for cardiovascular disease (CVD) among affected patients. In persons with RA, the risk of incident CVD and CVD mortality are increased by approximately 50% compared with the general population.^1,2 To minimize CVD risk in this population, providers must routinely assess for CVD risk factors³ and aggressively manage both traditional and nontraditional CVD risk factors.

Managing Traditional Risk Factors

As in the general population, identification and management of traditional CVD risk factors are crucial to minimize CVD risk in the RA population. A prospective study of 201 RA patients demonstrated that traditional CVD risk factors were in fact more predictive of endothelial dysfunction and carotid atherosclerosis than were disease-related inflammatory markers in RA.⁴ Management of traditional risk factors is detailed in the following sections, and recommendations for managing all traditional CVD risk factors are summarized in the Table.

Dyslipidemia

The role of dyslipidemia in atherogenesis is well established, and as a result, lipid levels are nearly universally included in CVD risk stratification tools. However, the interpretation of lipid levels in the context of RA is challenging because of the effects of systemic inflammation on their absolute values. Compared to the general population, patients with RA have lower total cholesterol (TC) and low-density lipoprotein (LDL) levels independent of lipid-lowering therapy.^5,6 Despite this, RA patients are at increased risk for CVD. There is even some evidence to suggest a “lipid paradox” in RA, whereby lower TC (< 4 mmol/L) and LDL levels suggest an increased risk of CVD.^7,8 In contrast to LDL, higher levels of high-density lipoprotein (HDL) are typically associated with reduced CVD risk, as in the general population.^8,9 Interestingly, in a cohort of 16,085 RA patients and 48,499 age- and sex-matched controls, there was no significant difference in the relationship between LDL and CVD risk, suggesting that quantitative lipid levels alone may not entirely explain the CVD mortality gap in RA.⁹ As such, there is substantial interest in lipoprotein function within the context of CVD risk in RA. Recent investigations have identified impaired HDL function, with reduced cholesterol efflux capacity and antioxidant properties, as well as increased scavenger receptor expression and foam cell formation, in patients with RA.^10,11 More research is needed to elucidate how these alterations affect CVD morbidity and mortality and how their measurement could be integrated into improved CVD risk assessment.

Meta-analyses of randomized controlled trials have estimated that lipid-lowering therapy with HMG-CoA reductase inhibitors (statins) reduces the risk of CVD by 25% to 30%; as such, statin therapy has become the standard of care for reduction of CVD risk in the general population.¹² Benefits for primary prevention of CVD in RA have also been observed; statin therapy was associated with a reduced risk of CVD events (hazard ratio [HR], 0.45; 95% confidence interval [CI], 0.20-0.98) and all-cause mortality (HR, 0.43; 95% CI, 0.20-0.92) in a population-based cohort study.¹³ Statins appear to have similar lipid-lowering effects and result in similar CVD risk reduction when used for primary or secondary prevention in RA patients compared to non-RA controls.^14-16 Additionally, anti-inflammatory properties of statins may act in synergy with disease-modifying antirheumatic drugs (DMARDs) to improve RA disease activity. In a small study of RA patients, statin therapy improved subjective and objective markers of RA disease activity in conjunction with methotrexate.¹⁷

While statins provide robust reduction in CVD risk, some individuals cannot tolerate statin therapy or do not achieve goal LDL levels with statin therapy. Select non-statin LDL-cholesterol-lowering agents have shown promise for reducing CVD events in the general population.¹⁸ Ezetimibe, which inhibits cholesterol absorption in the small intestine, very modestly reduced CVD events when added to atorvastatin (relative risk [RR], 0.94; 95% CI, 0.89-0.99) in a double-blind randomized controlled trial.¹⁹ Novel monoclonal antibodies to proprotein convertase subtilisin/kexin type 9 (PCSK-9) inhibit the internalization of surface LDL receptors, promoting LDL clearance. Two PCSK-9 inhibitors, alirocumab and evolocumab, were approved by the US Food and Drug Administration (FDA) after randomized controlled trials demonstrated their efficacy in lowering LDL by approximately 60% and reducing CVD events by approximately 15% in patients on maximum-tolerated statin therapy.^20-22 To date, non-statin LDL-cholesterol-lowering agents have been subject to limited study in RA.²³

Identification and management of dyslipidemia offers an opportunity for substantial CVD risk reduction at the RA population level. Unfortunately, current rates of lipid screening are inadequate in this high-risk group. In a study of 3298 Medicare patients with RA, less than half of RA patients with an indication underwent appropriate lipid screening.²⁴ Additionally, statins are often underutilized for both primary and secondary prevention in RA patients. Only 27% of RA patients meeting National Cholesterol Education Program Adult Treatment Panel III criteria were initiated on statin therapy in a population-based cohort study.²⁵ Among patients discharged after a first myocardial infarction (MI), the odds of receiving lipid-lowering therapy were 31% lower for RA patients (odds ratio [OR], 0.69; 95% CI, 0.58-0.82).²⁶ Similar to the general population, adherence to statins in RA patients appears to be poor.^27-30 This raises particular concern considering that a population-based cohort study of RA patients demonstrated a 67% increased risk of MI associated with statin discontinuation, regardless of prior MI status.²⁷ Providers—rheumatologists, primary care providers, and cardiologists alike—need to remain vigilant in efforts to assess CVD risk to identify patients who will benefit from lipid-lowering therapy and to emphasize the importance to patients of statin adherence. Novel models of health-care delivery, health technologies, and patient engagement in care may prove useful for improving lipid screening and management in RA.

Tobacco Use

Cigarette smoking is a shared risk factor for both CVD and RA. Large cohort studies have identified a dose-dependent increased risk of incident RA, particularly seropositive RA, among smokers.^31-34 Tobacco smoking has also been associated with increased levels of inflammation and RA disease activity.³⁵ The consequences of tobacco use in the general population are staggering. Among individuals over the age of 30 years, tobacco use is responsible for 12% of all deaths and 10% of all CVD deaths.³⁶ Similar findings are observed in RA; a recent meta-analysis estimated there is a 50% increased risk of CVD events in RA related to smoking tobacco.³⁷ In the general population, smoking cessation markedly lowers CVD risk, and over time CVD risk may approach that of nonsmokers.^38,39 Thus, regular counseling and interventions to facilitate smoking cessation are critical to reducing CVD risk in RA patients. RA-specific smoking cessation programs have been proposed, but have yet to outperform standard smoking cessation programs.⁴⁰

Diabetes Mellitus

It is estimated that almost 10% of the US population has diabetes mellitus (DM), which in isolation portends substantial CVD risk.⁴¹ There is an increased prevalence of DM in RA, perhaps owing to factors such as physical inactivity and chronic glucocorticoid use, though a higher level of RA disease activity itself has been associated with increased insulin resistance.^42-45 In a cohort of 100 RA patients who were neither obese nor diabetic, RA patients had significantly higher fasting blood glucose and insulin levels than age- and sex-matched controls. These findings were even more pronounced in RA patients with higher levels of disease activity.⁴⁴ Similar to the general population, DM is associated with poor CVD outcomes in RA.³⁷ Therefore, both appropriate management of diabetes and control of RA disease activity are vitally important to minimize CVD risk related to DM.

Though not a universal finding, there may be an increased prevalence of hypertension in RA patients.^31,46 Nonsteroidal anti-inflammatory drug (NSAID) and glucocorticoid use may play a role in the development of hypertension, while DMARDs appear to exert a less substantial effect on blood pressure.^47,48 At least one study found that DMARD initiation (particularly for methotrexate and hydroxychloroquine) was associated with significant, albeit small, declines in both systolic and diastolic blood pressure over the first 6 months of treatment.⁴⁹

Despite its potentially higher prevalence in this population, hypertension is both underdiagnosed and undertreated in RA patients.^24,50-52 This is an important deficiency to target because, as in the general population, hypertension is associated with an increased risk of MI (RR, 1.84; 95% CI, 1.38-2.46) and composite CVD outcomes (RR, 2.24; 95% CI, 1.42-3.06) in RA.³⁷ Thresholds for initiation and escalation of antihypertensive therapy are not specific to the RA population; thus, diagnosis and management of hypertension should be informed by the American College of Cardiology/American Heart Association guidelines, treating those with in-office blood pressures exceeding 140/90 mm Hg (> 130/80 mm Hg if aged > 65 years or with concomitant CVD, DM, chronic kidney disease, or 10-year atherosclerotic cardiovascular disease risk > 10%), typically with angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, calcium channel blockers, or thiazide diuretics as comorbidities may dictate or allow.⁵³ Also, the use of NSAIDs and glucocorticoids should be minimized, particularly in those with concomitant hypertension.

Physical Activity

Likely due to factors such as articular pain and stiffness, as well as physical limitations, RA patients are more sedentary than the general population.^54,55 In a study of objectively assessed sedentary behavior in RA patients, greater average sedentary time per day and greater number of sedentary bouts (> 20 min) were associated with increased 10-year risk of CVD as assessed by the QRISK2.⁵⁶ Conversely, the beneficial effects of exercise are well documented. Light to moderate physical activity has been associated with improved cardiovascular outcomes, greater physical function, higher levels of HDL, as well as reduced systemic inflammation and disease activity, and improved endothelial function in RA patients.^57-61 While there has been concern that physical activity may result in accelerated joint damage, even high-intensity exercise was shown to be safe without causing significant progression of joint damage.⁵⁸

Obesity, Weight Loss, and Diet

While obesity is clearly associated with CVD risk in the general population, this relationship is much more complex in RA, as underweight RA patients are also at higher risk for CVD and CVD-related mortality.^62-64 One potential explanation for this finding is that pathological weight loss resulting in an underweight body mass index (BMI) is an independent predictor of CVD. In a study of US Veterans with RA, higher rates of weight loss (> 3 kg/m²/year) were associated with increased CVD mortality (HR, 2.27; 95% CI, 1.61-3.19) independent of BMI.⁶⁵ Systemic inflammation in RA can lead to “rheumatoid cachexia,” characterized by decreased muscle mass, increased adiposity, and increased CVD risk despite a normal or potentially decreased BMI.⁶⁶ Practitioners should be mindful of not only current body weight, but also patients’ weight trajectories when counseling on lifestyle practices such as healthy diet and regular exercise in RA patients. For obese individuals with RA, healthy weight loss should be encouraged. Interestingly, bariatric surgery in RA patients may improve RA disease activity in addition to its known effects on body weight and DM.⁶⁷

Counseling on healthy diet with a focus on limiting foods high in saturated- and trans-fatty acids and high glycemic index foods, and increasing consumption of fruits, vegetables, and mono-unsaturated fatty acids is a well-accepted and common practice to help minimize CVD risk in the general population.⁶⁸ No studies to date have investigated the effect of specific diets on CVD risk in RA patients, and thus we recommend adherence to general population recommendations.

Disease Activity

In addition to traditional risk factors, several studies have identified associations between the level of RA disease activity and risk of CVD. In a cohort of US Veterans with RA, CVD-related mortality increased in a dose-dependent manner with higher disease activity categories. In stark contrast, the CVD mortality rates of those in remission paralleled the rates from the general population (standardized mortality ratio [SMR], 0.68; 95% CI, 0.37-1.27).⁶⁹ In a separate cohort of 1157 RA patients without prior CVD, achieving low disease activity was associated with a lower risk of incident CVD events (HR, 0.65; 95% CI, 0.43-0.99).⁷⁰ Additionally, high disease activity has been associated with surrogate markers of CVD and other CVD risk factors including NT-proBNP and systolic blood pressure.^71,72 While no randomized controlled trial data is available to inform this recommendation, observational data suggest RA should be aggressively treated (ideally to achieve and maintain remission or low disease activity) to minimize CVD risk. While keeping this treatment goal in mind, the differential effects of specific RA therapies on CVD must also be considered.

Glucocorticoids and NSAIDs

With the expanding repertoire of DMARDs available and more aggressive treatment approaches, the role of glucocorticoids and NSAIDs in RA treatment is decreasing over time. While their efficacy for improving pain and stiffness is well established, concern regarding their contribution to CVD risk in RA patients is warranted.

Glucocorticoids are known to have detrimental effects on traditional CVD risk factors such as hypertension, insulin resistance, and dyslipidemia in the general population, as well as in RA patients.^73,74 In a meta-analysis of predominantly observational studies of RA patients, glucocorticoid use was associated with an increased risk of CVD events (RR, 1.47; 95% CI, 1.34-1.60), including MI, congestive heart failure (CHF), and cerebrovascular accident (CVA).⁷⁵ Evidence is conflicting in regards to a clear dose threshold that leads to increased CVD risk with glucocorticoids, though higher doses are associated with greater risk.^76-81 As RA patients requiring glucocorticoids typically have higher disease activity, confounding by indication remains a complicating factor in assessing the relative contributions of glucocorticoid use and RA disease activity to elevated CVD risk in many analyses.

The increased CVD risk with NSAID use is not specific to RA and has been well established in the general population.^82-84 In the previously mentioned meta-analysis, an increased overall risk of CVD events was observed with NSAID use in RA (RR, 1.18; 95% CI, 1.01-1.38). It should be noted that cyclo-oxygenase 2 (COX-2) inhibitors, in particular rofecoxib (now removed from the market), appeared to drive the majority of this risk (RR, 1.36; 95% CI, 1.10-1.67 in COX-2 inhibitors and RR 1.08, 95% CI, 0.94-1.24 in nonselective NSAIDs), suggesting a potential differential risk among NSAIDs.⁷⁵ While naproxen has been thought to carry the lowest risk of CVD based on initial studies, this has not been universally observed, including in a recent randomized controlled trial of more than 24,000 RA and osteoarthritis patients.^82,85,86

Providers should use the lowest possible dose and duration of glucocorticoids and NSAIDs to achieve symptom relief, with continual efforts to taper or discontinue. Candidates for glucocorticoid and NSAID therapy should be selected carefully, and use of these therapies should be avoided in those with prior CVD or at high risk for CVD based on traditional CVD risk factors. Most importantly, providers should focus on utilizing DMARDs for the management of RA, which more effectively treat RA as well as reduce CVD risk.

Methotrexate (MTX), a mainstay in the treatment of RA, is a conventional DMARD observed to improve overall survival and mitigate CVD risk in multiple RA cohorts.^75,87,88 In a recent meta-analysis comprised of 236,525 RA patients and 5410 CVD events, MTX use was associated with a 28% reduction in overall CVD events across 8 studies (RR, 0.72; 95% CI, 0.57-0.91), substantiating similar findings in a prior meta-analysis.^75,88 MTX use was specifically associated with a decreased risk of MI (RR, 0.81; 95% CI, 0.68-0.96). Case-control and cohort studies have cited a 20% to 50% reduced risk of CHF with MTX use.^89,90 The potential cardioprotective effect of MTX appears to be both multifactorial and complex, likely mediated through both direct and indirect mechanisms. MTX directly promotes anti-atherogenic lipoprotein function, improves endothelial function, and scavenges free radicals.^91,92 Indirectly, MTX likely reduces CVD risk by effectively reducing RA disease activity. Based on these and other data, MTX remains the cornerstone of DMARD therapy in RA patients when targeting CVD risk reduction.

Hydroxychloroquine

Emerging evidence suggests that hydroxychloroquine (HCQ), an antimalarial most often utilized in combination with alternative DMARDs in RA, prevents DM and has beneficial effects on lipid profiles. A recent meta-analysis compiled 3 homogenous observational studies that investigated the effect of HCQ on incident DM. RA patients ever exposed to HCQ had a 40% lower incidence of DM (HR, 0.59; 95% CI, 0.49-0.70).⁹³ Increased duration of HCQ use was shown to further reduce risk of incident DM.⁹⁴ The aforementioned meta-analysis also pooled 5 studies investigating the effect of HCQ on lipid profiles, with favorable mean differences in TC (–9.82 mg/dL), LDL (–10.61 mg/dL), HDL (4.13 mg/dL), and triglycerides (–19.15 mg/dL) in HCQ users compared to non-users.⁹³ Given these favorable changes to traditional CVD risk factors, it is not surprising that in a retrospective study of 1266 RA patients without prior CVD, HCQ was associated with significantly lower risk of incident CVD. While external validation of these findings is needed, HCQ is an attractive conventional DMARD to be used in RA for CVD risk reduction. Moreover, its combination with MTX and sulfasalazine also shows promise for CVD risk reduction.^95,96

TNF Inhibitors

Tumor necrosis factor (TNF) inhibitors are often the initial biologic DMARD therapy used in RA patients not responding to conventional DMARDs. In the previously described meta-analysis, TNF inhibitors were associated with similar reductions in CVD events as MTX (RR, 0.70; 95% CI, 0.54-0.90).⁷⁵ Of note, there was a trend toward reduced risk of CHF (RR, 0.75; 95% CI, 0.49-1.15) in this same meta-analysis, an area of concern with TNF inhibitor use due to a prior randomized controlled trial demonstrating worsening clinical status in patients with existing moderate-to-severe CHF treated with high-dose infliximab.⁹⁷ Current RA treatment guidelines recommend avoiding TNF inhibitor use in individuals with CHF.⁹⁸

Aside from the risk of CHF exacerbation, TNF inhibitors appear to be cardioprotective. Similar to MTX, the mechanism by which TNF inhibition reduces cardiovascular risk is complex and likely due to both direct and indirect mechanisms. Substantial research has been conducted on the effect of TNF inhibition on lipids, with a recent meta-analysis demonstrating increases in HDL and TC, with stable LDL and atherogenic index over treatment follow-up.⁹⁹ A subsequent meta-analysis not limited to RA patients yielded similar results.¹⁰⁰ In addition to quantitative lipid changes, alteration of lipoprotein function, improvement in myocardial function, reduced aortic stiffness, improved blood pressure, and reduced RA disease activity may also be responsible for cardioprotective benefits of these agents.^101,102

Non-TNF Biologic and Traditional Synthetic DMARDs

Tocilizumab, an IL-6 inhibitor, can potently increase LDL levels, but it does not appear to increase the risk of CVD events and may actually promote more favorable anti-atherogenic lipoprotein function.^103-106 Although these quantitative lipid changes received significant attention in the wake of early reports detailing this effect, similar lipid changes appear to accompany other DMARDs including TNF inhibitors and tofacitinib.¹⁰⁷ There have been few studies evaluating the risk of CVD with other non-TNF inhibitor biologic DMARDs and traditional synthetic DMARDs, warranting future study.

Conclusion

To mitigate the increased risk of CVD in RA, primary care and subspecialty providers alike must be aware of this heightened risk in RA, perform frequent assessments of CVD risk,³ and aggressively manage both traditional and nontraditional CVD risk factors. The differential roles in this effort may not be clear; thus, we have proposed a co-management strategy detailed in the Figure. Clear communication between providers is of the utmost importance to ensure effective management of CVD risk.

Given limited evidence for RA-specific CVD risk assessments and traditional risk factor treatment targets, management should follow pertinent national guidelines. The importance of lifestyle counseling should not be overlooked, with a focus on smoking cessation, healthy diet and body weight, and regular aerobic exercise. Finally, rheumatologists should aggressively manage RA using a treat-to-target approach, minimize the use of glucocorticoids and NSAIDs, and preferentially select DMARDs that have been associated with lower CVD risk. Through this comprehensive approach, recent trends of improved CVD outcomes in RA will hopefully become more widespread.¹⁰⁸

Corresponding author: Bryant R. England, MD; 986270 Nebraska Medical Center, Omaha, NE 68198-6270; [email protected].

Financial disclosures: Dr. England is supported by UNMC Internal Medicine Scientist Development Award, UNMC Physician-Scientist Training Program, the UNMC Mentored Scholars Program, and the Rheumatology Research Foundation Scientist Development Award. Dr. Mikuls is supported by a VA Merit Award (CX000896) and grants from the National Institutes of Health: National Institute of General Medical Sciences (U54GM115458), National Institute on Alcohol Abuse and Alcoholism (R25AA020818), and National Institute of Arthritis and Musculoskeletal and Skin Diseases (2P50AR60772).

From the Division of Rheumatology & Immunology, University of Nebraska Medical Center, and Veterans Affairs Nebraska-Western Iowa Health Care System, Omaha, NE.

Abstract

• Objective: To review the management of traditional and nontraditional CVD cardiovascular disease risk factors in rheumatoid arthritis (RA).
• Methods: Literature review of the management of CVD risk in RA.
• Results: Because of the increased risk of CVD events and CVD mortality among RA patients, aggressive management of CVD risk is essential. Providers should follow national guidelines for the management of traditional CVD risk factors, including dyslipidemia, hypertension, and diabetes mellitus. Similar efforts are needed in counseling on lifestyle modifications, including smoking cessation, regular exercise, and maintaining a healthy body weight. Because higher RA disease activity is also linked with CVD risk, aggressive treatment of RA to a target of low disease activity or remission is critical. Furthermore, the selection of potentially “cardioprotective” agents such as methotrexate and tumor necrosis factor inhibitors, while limiting use of nonsteroidal anti-inflammatory drugs and glucocorticoids, are strategies that could be employed by rheumatologists to help mitigate CVD risk in their patients with RA.
• Conclusion: Routine assessment of CVD risk, management of traditional CVD risk factors, counseling on healthy lifestyle habits, and aggressive treatment of RA are essential to minimize CVD risk in this population.

Keywords: rheumatoid arthritis; cardiovascular disease; cardiovascular risk assessment; cardiovascular risk management.

Editor’s note: This article is part 2 of a 2-part article. “Assessment of Cardiovascular Disease Risk in Rheumatoid Arthritis” was published in the January/February 2019 issue.

Rheumatoid arthritis (RA) is a systemic autoimmune condition that contributes to an increased risk for cardiovascular disease (CVD) among affected patients. In persons with RA, the risk of incident CVD and CVD mortality are increased by approximately 50% compared with the general population.^1,2 To minimize CVD risk in this population, providers must routinely assess for CVD risk factors³ and aggressively manage both traditional and nontraditional CVD risk factors.

Managing Traditional Risk Factors

As in the general population, identification and management of traditional CVD risk factors are crucial to minimize CVD risk in the RA population. A prospective study of 201 RA patients demonstrated that traditional CVD risk factors were in fact more predictive of endothelial dysfunction and carotid atherosclerosis than were disease-related inflammatory markers in RA.⁴ Management of traditional risk factors is detailed in the following sections, and recommendations for managing all traditional CVD risk factors are summarized in the Table.

Dyslipidemia

The role of dyslipidemia in atherogenesis is well established, and as a result, lipid levels are nearly universally included in CVD risk stratification tools. However, the interpretation of lipid levels in the context of RA is challenging because of the effects of systemic inflammation on their absolute values. Compared to the general population, patients with RA have lower total cholesterol (TC) and low-density lipoprotein (LDL) levels independent of lipid-lowering therapy.^5,6 Despite this, RA patients are at increased risk for CVD. There is even some evidence to suggest a “lipid paradox” in RA, whereby lower TC (< 4 mmol/L) and LDL levels suggest an increased risk of CVD.^7,8 In contrast to LDL, higher levels of high-density lipoprotein (HDL) are typically associated with reduced CVD risk, as in the general population.^8,9 Interestingly, in a cohort of 16,085 RA patients and 48,499 age- and sex-matched controls, there was no significant difference in the relationship between LDL and CVD risk, suggesting that quantitative lipid levels alone may not entirely explain the CVD mortality gap in RA.⁹ As such, there is substantial interest in lipoprotein function within the context of CVD risk in RA. Recent investigations have identified impaired HDL function, with reduced cholesterol efflux capacity and antioxidant properties, as well as increased scavenger receptor expression and foam cell formation, in patients with RA.^10,11 More research is needed to elucidate how these alterations affect CVD morbidity and mortality and how their measurement could be integrated into improved CVD risk assessment.

Meta-analyses of randomized controlled trials have estimated that lipid-lowering therapy with HMG-CoA reductase inhibitors (statins) reduces the risk of CVD by 25% to 30%; as such, statin therapy has become the standard of care for reduction of CVD risk in the general population.¹² Benefits for primary prevention of CVD in RA have also been observed; statin therapy was associated with a reduced risk of CVD events (hazard ratio [HR], 0.45; 95% confidence interval [CI], 0.20-0.98) and all-cause mortality (HR, 0.43; 95% CI, 0.20-0.92) in a population-based cohort study.¹³ Statins appear to have similar lipid-lowering effects and result in similar CVD risk reduction when used for primary or secondary prevention in RA patients compared to non-RA controls.^14-16 Additionally, anti-inflammatory properties of statins may act in synergy with disease-modifying antirheumatic drugs (DMARDs) to improve RA disease activity. In a small study of RA patients, statin therapy improved subjective and objective markers of RA disease activity in conjunction with methotrexate.¹⁷

While statins provide robust reduction in CVD risk, some individuals cannot tolerate statin therapy or do not achieve goal LDL levels with statin therapy. Select non-statin LDL-cholesterol-lowering agents have shown promise for reducing CVD events in the general population.¹⁸ Ezetimibe, which inhibits cholesterol absorption in the small intestine, very modestly reduced CVD events when added to atorvastatin (relative risk [RR], 0.94; 95% CI, 0.89-0.99) in a double-blind randomized controlled trial.¹⁹ Novel monoclonal antibodies to proprotein convertase subtilisin/kexin type 9 (PCSK-9) inhibit the internalization of surface LDL receptors, promoting LDL clearance. Two PCSK-9 inhibitors, alirocumab and evolocumab, were approved by the US Food and Drug Administration (FDA) after randomized controlled trials demonstrated their efficacy in lowering LDL by approximately 60% and reducing CVD events by approximately 15% in patients on maximum-tolerated statin therapy.^20-22 To date, non-statin LDL-cholesterol-lowering agents have been subject to limited study in RA.²³

Identification and management of dyslipidemia offers an opportunity for substantial CVD risk reduction at the RA population level. Unfortunately, current rates of lipid screening are inadequate in this high-risk group. In a study of 3298 Medicare patients with RA, less than half of RA patients with an indication underwent appropriate lipid screening.²⁴ Additionally, statins are often underutilized for both primary and secondary prevention in RA patients. Only 27% of RA patients meeting National Cholesterol Education Program Adult Treatment Panel III criteria were initiated on statin therapy in a population-based cohort study.²⁵ Among patients discharged after a first myocardial infarction (MI), the odds of receiving lipid-lowering therapy were 31% lower for RA patients (odds ratio [OR], 0.69; 95% CI, 0.58-0.82).²⁶ Similar to the general population, adherence to statins in RA patients appears to be poor.^27-30 This raises particular concern considering that a population-based cohort study of RA patients demonstrated a 67% increased risk of MI associated with statin discontinuation, regardless of prior MI status.²⁷ Providers—rheumatologists, primary care providers, and cardiologists alike—need to remain vigilant in efforts to assess CVD risk to identify patients who will benefit from lipid-lowering therapy and to emphasize the importance to patients of statin adherence. Novel models of health-care delivery, health technologies, and patient engagement in care may prove useful for improving lipid screening and management in RA.

Tobacco Use

Cigarette smoking is a shared risk factor for both CVD and RA. Large cohort studies have identified a dose-dependent increased risk of incident RA, particularly seropositive RA, among smokers.^31-34 Tobacco smoking has also been associated with increased levels of inflammation and RA disease activity.³⁵ The consequences of tobacco use in the general population are staggering. Among individuals over the age of 30 years, tobacco use is responsible for 12% of all deaths and 10% of all CVD deaths.³⁶ Similar findings are observed in RA; a recent meta-analysis estimated there is a 50% increased risk of CVD events in RA related to smoking tobacco.³⁷ In the general population, smoking cessation markedly lowers CVD risk, and over time CVD risk may approach that of nonsmokers.^38,39 Thus, regular counseling and interventions to facilitate smoking cessation are critical to reducing CVD risk in RA patients. RA-specific smoking cessation programs have been proposed, but have yet to outperform standard smoking cessation programs.⁴⁰

Diabetes Mellitus

It is estimated that almost 10% of the US population has diabetes mellitus (DM), which in isolation portends substantial CVD risk.⁴¹ There is an increased prevalence of DM in RA, perhaps owing to factors such as physical inactivity and chronic glucocorticoid use, though a higher level of RA disease activity itself has been associated with increased insulin resistance.^42-45 In a cohort of 100 RA patients who were neither obese nor diabetic, RA patients had significantly higher fasting blood glucose and insulin levels than age- and sex-matched controls. These findings were even more pronounced in RA patients with higher levels of disease activity.⁴⁴ Similar to the general population, DM is associated with poor CVD outcomes in RA.³⁷ Therefore, both appropriate management of diabetes and control of RA disease activity are vitally important to minimize CVD risk related to DM.

Though not a universal finding, there may be an increased prevalence of hypertension in RA patients.^31,46 Nonsteroidal anti-inflammatory drug (NSAID) and glucocorticoid use may play a role in the development of hypertension, while DMARDs appear to exert a less substantial effect on blood pressure.^47,48 At least one study found that DMARD initiation (particularly for methotrexate and hydroxychloroquine) was associated with significant, albeit small, declines in both systolic and diastolic blood pressure over the first 6 months of treatment.⁴⁹

Despite its potentially higher prevalence in this population, hypertension is both underdiagnosed and undertreated in RA patients.^24,50-52 This is an important deficiency to target because, as in the general population, hypertension is associated with an increased risk of MI (RR, 1.84; 95% CI, 1.38-2.46) and composite CVD outcomes (RR, 2.24; 95% CI, 1.42-3.06) in RA.³⁷ Thresholds for initiation and escalation of antihypertensive therapy are not specific to the RA population; thus, diagnosis and management of hypertension should be informed by the American College of Cardiology/American Heart Association guidelines, treating those with in-office blood pressures exceeding 140/90 mm Hg (> 130/80 mm Hg if aged > 65 years or with concomitant CVD, DM, chronic kidney disease, or 10-year atherosclerotic cardiovascular disease risk > 10%), typically with angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, calcium channel blockers, or thiazide diuretics as comorbidities may dictate or allow.⁵³ Also, the use of NSAIDs and glucocorticoids should be minimized, particularly in those with concomitant hypertension.

Physical Activity

Likely due to factors such as articular pain and stiffness, as well as physical limitations, RA patients are more sedentary than the general population.^54,55 In a study of objectively assessed sedentary behavior in RA patients, greater average sedentary time per day and greater number of sedentary bouts (> 20 min) were associated with increased 10-year risk of CVD as assessed by the QRISK2.⁵⁶ Conversely, the beneficial effects of exercise are well documented. Light to moderate physical activity has been associated with improved cardiovascular outcomes, greater physical function, higher levels of HDL, as well as reduced systemic inflammation and disease activity, and improved endothelial function in RA patients.^57-61 While there has been concern that physical activity may result in accelerated joint damage, even high-intensity exercise was shown to be safe without causing significant progression of joint damage.⁵⁸

Obesity, Weight Loss, and Diet

While obesity is clearly associated with CVD risk in the general population, this relationship is much more complex in RA, as underweight RA patients are also at higher risk for CVD and CVD-related mortality.^62-64 One potential explanation for this finding is that pathological weight loss resulting in an underweight body mass index (BMI) is an independent predictor of CVD. In a study of US Veterans with RA, higher rates of weight loss (> 3 kg/m²/year) were associated with increased CVD mortality (HR, 2.27; 95% CI, 1.61-3.19) independent of BMI.⁶⁵ Systemic inflammation in RA can lead to “rheumatoid cachexia,” characterized by decreased muscle mass, increased adiposity, and increased CVD risk despite a normal or potentially decreased BMI.⁶⁶ Practitioners should be mindful of not only current body weight, but also patients’ weight trajectories when counseling on lifestyle practices such as healthy diet and regular exercise in RA patients. For obese individuals with RA, healthy weight loss should be encouraged. Interestingly, bariatric surgery in RA patients may improve RA disease activity in addition to its known effects on body weight and DM.⁶⁷

Counseling on healthy diet with a focus on limiting foods high in saturated- and trans-fatty acids and high glycemic index foods, and increasing consumption of fruits, vegetables, and mono-unsaturated fatty acids is a well-accepted and common practice to help minimize CVD risk in the general population.⁶⁸ No studies to date have investigated the effect of specific diets on CVD risk in RA patients, and thus we recommend adherence to general population recommendations.

Disease Activity

In addition to traditional risk factors, several studies have identified associations between the level of RA disease activity and risk of CVD. In a cohort of US Veterans with RA, CVD-related mortality increased in a dose-dependent manner with higher disease activity categories. In stark contrast, the CVD mortality rates of those in remission paralleled the rates from the general population (standardized mortality ratio [SMR], 0.68; 95% CI, 0.37-1.27).⁶⁹ In a separate cohort of 1157 RA patients without prior CVD, achieving low disease activity was associated with a lower risk of incident CVD events (HR, 0.65; 95% CI, 0.43-0.99).⁷⁰ Additionally, high disease activity has been associated with surrogate markers of CVD and other CVD risk factors including NT-proBNP and systolic blood pressure.^71,72 While no randomized controlled trial data is available to inform this recommendation, observational data suggest RA should be aggressively treated (ideally to achieve and maintain remission or low disease activity) to minimize CVD risk. While keeping this treatment goal in mind, the differential effects of specific RA therapies on CVD must also be considered.

Glucocorticoids and NSAIDs

With the expanding repertoire of DMARDs available and more aggressive treatment approaches, the role of glucocorticoids and NSAIDs in RA treatment is decreasing over time. While their efficacy for improving pain and stiffness is well established, concern regarding their contribution to CVD risk in RA patients is warranted.

Glucocorticoids are known to have detrimental effects on traditional CVD risk factors such as hypertension, insulin resistance, and dyslipidemia in the general population, as well as in RA patients.^73,74 In a meta-analysis of predominantly observational studies of RA patients, glucocorticoid use was associated with an increased risk of CVD events (RR, 1.47; 95% CI, 1.34-1.60), including MI, congestive heart failure (CHF), and cerebrovascular accident (CVA).⁷⁵ Evidence is conflicting in regards to a clear dose threshold that leads to increased CVD risk with glucocorticoids, though higher doses are associated with greater risk.^76-81 As RA patients requiring glucocorticoids typically have higher disease activity, confounding by indication remains a complicating factor in assessing the relative contributions of glucocorticoid use and RA disease activity to elevated CVD risk in many analyses.

The increased CVD risk with NSAID use is not specific to RA and has been well established in the general population.^82-84 In the previously mentioned meta-analysis, an increased overall risk of CVD events was observed with NSAID use in RA (RR, 1.18; 95% CI, 1.01-1.38). It should be noted that cyclo-oxygenase 2 (COX-2) inhibitors, in particular rofecoxib (now removed from the market), appeared to drive the majority of this risk (RR, 1.36; 95% CI, 1.10-1.67 in COX-2 inhibitors and RR 1.08, 95% CI, 0.94-1.24 in nonselective NSAIDs), suggesting a potential differential risk among NSAIDs.⁷⁵ While naproxen has been thought to carry the lowest risk of CVD based on initial studies, this has not been universally observed, including in a recent randomized controlled trial of more than 24,000 RA and osteoarthritis patients.^82,85,86

Providers should use the lowest possible dose and duration of glucocorticoids and NSAIDs to achieve symptom relief, with continual efforts to taper or discontinue. Candidates for glucocorticoid and NSAID therapy should be selected carefully, and use of these therapies should be avoided in those with prior CVD or at high risk for CVD based on traditional CVD risk factors. Most importantly, providers should focus on utilizing DMARDs for the management of RA, which more effectively treat RA as well as reduce CVD risk.

Methotrexate (MTX), a mainstay in the treatment of RA, is a conventional DMARD observed to improve overall survival and mitigate CVD risk in multiple RA cohorts.^75,87,88 In a recent meta-analysis comprised of 236,525 RA patients and 5410 CVD events, MTX use was associated with a 28% reduction in overall CVD events across 8 studies (RR, 0.72; 95% CI, 0.57-0.91), substantiating similar findings in a prior meta-analysis.^75,88 MTX use was specifically associated with a decreased risk of MI (RR, 0.81; 95% CI, 0.68-0.96). Case-control and cohort studies have cited a 20% to 50% reduced risk of CHF with MTX use.^89,90 The potential cardioprotective effect of MTX appears to be both multifactorial and complex, likely mediated through both direct and indirect mechanisms. MTX directly promotes anti-atherogenic lipoprotein function, improves endothelial function, and scavenges free radicals.^91,92 Indirectly, MTX likely reduces CVD risk by effectively reducing RA disease activity. Based on these and other data, MTX remains the cornerstone of DMARD therapy in RA patients when targeting CVD risk reduction.

Hydroxychloroquine

Emerging evidence suggests that hydroxychloroquine (HCQ), an antimalarial most often utilized in combination with alternative DMARDs in RA, prevents DM and has beneficial effects on lipid profiles. A recent meta-analysis compiled 3 homogenous observational studies that investigated the effect of HCQ on incident DM. RA patients ever exposed to HCQ had a 40% lower incidence of DM (HR, 0.59; 95% CI, 0.49-0.70).⁹³ Increased duration of HCQ use was shown to further reduce risk of incident DM.⁹⁴ The aforementioned meta-analysis also pooled 5 studies investigating the effect of HCQ on lipid profiles, with favorable mean differences in TC (–9.82 mg/dL), LDL (–10.61 mg/dL), HDL (4.13 mg/dL), and triglycerides (–19.15 mg/dL) in HCQ users compared to non-users.⁹³ Given these favorable changes to traditional CVD risk factors, it is not surprising that in a retrospective study of 1266 RA patients without prior CVD, HCQ was associated with significantly lower risk of incident CVD. While external validation of these findings is needed, HCQ is an attractive conventional DMARD to be used in RA for CVD risk reduction. Moreover, its combination with MTX and sulfasalazine also shows promise for CVD risk reduction.^95,96

TNF Inhibitors

Tumor necrosis factor (TNF) inhibitors are often the initial biologic DMARD therapy used in RA patients not responding to conventional DMARDs. In the previously described meta-analysis, TNF inhibitors were associated with similar reductions in CVD events as MTX (RR, 0.70; 95% CI, 0.54-0.90).⁷⁵ Of note, there was a trend toward reduced risk of CHF (RR, 0.75; 95% CI, 0.49-1.15) in this same meta-analysis, an area of concern with TNF inhibitor use due to a prior randomized controlled trial demonstrating worsening clinical status in patients with existing moderate-to-severe CHF treated with high-dose infliximab.⁹⁷ Current RA treatment guidelines recommend avoiding TNF inhibitor use in individuals with CHF.⁹⁸

Aside from the risk of CHF exacerbation, TNF inhibitors appear to be cardioprotective. Similar to MTX, the mechanism by which TNF inhibition reduces cardiovascular risk is complex and likely due to both direct and indirect mechanisms. Substantial research has been conducted on the effect of TNF inhibition on lipids, with a recent meta-analysis demonstrating increases in HDL and TC, with stable LDL and atherogenic index over treatment follow-up.⁹⁹ A subsequent meta-analysis not limited to RA patients yielded similar results.¹⁰⁰ In addition to quantitative lipid changes, alteration of lipoprotein function, improvement in myocardial function, reduced aortic stiffness, improved blood pressure, and reduced RA disease activity may also be responsible for cardioprotective benefits of these agents.^101,102

Non-TNF Biologic and Traditional Synthetic DMARDs

Tocilizumab, an IL-6 inhibitor, can potently increase LDL levels, but it does not appear to increase the risk of CVD events and may actually promote more favorable anti-atherogenic lipoprotein function.^103-106 Although these quantitative lipid changes received significant attention in the wake of early reports detailing this effect, similar lipid changes appear to accompany other DMARDs including TNF inhibitors and tofacitinib.¹⁰⁷ There have been few studies evaluating the risk of CVD with other non-TNF inhibitor biologic DMARDs and traditional synthetic DMARDs, warranting future study.

Conclusion

To mitigate the increased risk of CVD in RA, primary care and subspecialty providers alike must be aware of this heightened risk in RA, perform frequent assessments of CVD risk,³ and aggressively manage both traditional and nontraditional CVD risk factors. The differential roles in this effort may not be clear; thus, we have proposed a co-management strategy detailed in the Figure. Clear communication between providers is of the utmost importance to ensure effective management of CVD risk.

Given limited evidence for RA-specific CVD risk assessments and traditional risk factor treatment targets, management should follow pertinent national guidelines. The importance of lifestyle counseling should not be overlooked, with a focus on smoking cessation, healthy diet and body weight, and regular aerobic exercise. Finally, rheumatologists should aggressively manage RA using a treat-to-target approach, minimize the use of glucocorticoids and NSAIDs, and preferentially select DMARDs that have been associated with lower CVD risk. Through this comprehensive approach, recent trends of improved CVD outcomes in RA will hopefully become more widespread.¹⁰⁸

Corresponding author: Bryant R. England, MD; 986270 Nebraska Medical Center, Omaha, NE 68198-6270; [email protected].

Financial disclosures: Dr. England is supported by UNMC Internal Medicine Scientist Development Award, UNMC Physician-Scientist Training Program, the UNMC Mentored Scholars Program, and the Rheumatology Research Foundation Scientist Development Award. Dr. Mikuls is supported by a VA Merit Award (CX000896) and grants from the National Institutes of Health: National Institute of General Medical Sciences (U54GM115458), National Institute on Alcohol Abuse and Alcoholism (R25AA020818), and National Institute of Arthritis and Musculoskeletal and Skin Diseases (2P50AR60772).

From the Division of Rheumatology & Immunology, University of Nebraska Medical Center, and Veterans Affairs Nebraska-Western Iowa Health Care System, Omaha, NE.

Abstract

• Objective: To review the management of traditional and nontraditional CVD cardiovascular disease risk factors in rheumatoid arthritis (RA).
• Methods: Literature review of the management of CVD risk in RA.
• Results: Because of the increased risk of CVD events and CVD mortality among RA patients, aggressive management of CVD risk is essential. Providers should follow national guidelines for the management of traditional CVD risk factors, including dyslipidemia, hypertension, and diabetes mellitus. Similar efforts are needed in counseling on lifestyle modifications, including smoking cessation, regular exercise, and maintaining a healthy body weight. Because higher RA disease activity is also linked with CVD risk, aggressive treatment of RA to a target of low disease activity or remission is critical. Furthermore, the selection of potentially “cardioprotective” agents such as methotrexate and tumor necrosis factor inhibitors, while limiting use of nonsteroidal anti-inflammatory drugs and glucocorticoids, are strategies that could be employed by rheumatologists to help mitigate CVD risk in their patients with RA.
• Conclusion: Routine assessment of CVD risk, management of traditional CVD risk factors, counseling on healthy lifestyle habits, and aggressive treatment of RA are essential to minimize CVD risk in this population.

Keywords: rheumatoid arthritis; cardiovascular disease; cardiovascular risk assessment; cardiovascular risk management.

Editor’s note: This article is part 2 of a 2-part article. “Assessment of Cardiovascular Disease Risk in Rheumatoid Arthritis” was published in the January/February 2019 issue.

Rheumatoid arthritis (RA) is a systemic autoimmune condition that contributes to an increased risk for cardiovascular disease (CVD) among affected patients. In persons with RA, the risk of incident CVD and CVD mortality are increased by approximately 50% compared with the general population.^1,2 To minimize CVD risk in this population, providers must routinely assess for CVD risk factors³ and aggressively manage both traditional and nontraditional CVD risk factors.

Managing Traditional Risk Factors

As in the general population, identification and management of traditional CVD risk factors are crucial to minimize CVD risk in the RA population. A prospective study of 201 RA patients demonstrated that traditional CVD risk factors were in fact more predictive of endothelial dysfunction and carotid atherosclerosis than were disease-related inflammatory markers in RA.⁴ Management of traditional risk factors is detailed in the following sections, and recommendations for managing all traditional CVD risk factors are summarized in the Table.

Dyslipidemia

The role of dyslipidemia in atherogenesis is well established, and as a result, lipid levels are nearly universally included in CVD risk stratification tools. However, the interpretation of lipid levels in the context of RA is challenging because of the effects of systemic inflammation on their absolute values. Compared to the general population, patients with RA have lower total cholesterol (TC) and low-density lipoprotein (LDL) levels independent of lipid-lowering therapy.^5,6 Despite this, RA patients are at increased risk for CVD. There is even some evidence to suggest a “lipid paradox” in RA, whereby lower TC (< 4 mmol/L) and LDL levels suggest an increased risk of CVD.^7,8 In contrast to LDL, higher levels of high-density lipoprotein (HDL) are typically associated with reduced CVD risk, as in the general population.^8,9 Interestingly, in a cohort of 16,085 RA patients and 48,499 age- and sex-matched controls, there was no significant difference in the relationship between LDL and CVD risk, suggesting that quantitative lipid levels alone may not entirely explain the CVD mortality gap in RA.⁹ As such, there is substantial interest in lipoprotein function within the context of CVD risk in RA. Recent investigations have identified impaired HDL function, with reduced cholesterol efflux capacity and antioxidant properties, as well as increased scavenger receptor expression and foam cell formation, in patients with RA.^10,11 More research is needed to elucidate how these alterations affect CVD morbidity and mortality and how their measurement could be integrated into improved CVD risk assessment.

Meta-analyses of randomized controlled trials have estimated that lipid-lowering therapy with HMG-CoA reductase inhibitors (statins) reduces the risk of CVD by 25% to 30%; as such, statin therapy has become the standard of care for reduction of CVD risk in the general population.¹² Benefits for primary prevention of CVD in RA have also been observed; statin therapy was associated with a reduced risk of CVD events (hazard ratio [HR], 0.45; 95% confidence interval [CI], 0.20-0.98) and all-cause mortality (HR, 0.43; 95% CI, 0.20-0.92) in a population-based cohort study.¹³ Statins appear to have similar lipid-lowering effects and result in similar CVD risk reduction when used for primary or secondary prevention in RA patients compared to non-RA controls.^14-16 Additionally, anti-inflammatory properties of statins may act in synergy with disease-modifying antirheumatic drugs (DMARDs) to improve RA disease activity. In a small study of RA patients, statin therapy improved subjective and objective markers of RA disease activity in conjunction with methotrexate.¹⁷

While statins provide robust reduction in CVD risk, some individuals cannot tolerate statin therapy or do not achieve goal LDL levels with statin therapy. Select non-statin LDL-cholesterol-lowering agents have shown promise for reducing CVD events in the general population.¹⁸ Ezetimibe, which inhibits cholesterol absorption in the small intestine, very modestly reduced CVD events when added to atorvastatin (relative risk [RR], 0.94; 95% CI, 0.89-0.99) in a double-blind randomized controlled trial.¹⁹ Novel monoclonal antibodies to proprotein convertase subtilisin/kexin type 9 (PCSK-9) inhibit the internalization of surface LDL receptors, promoting LDL clearance. Two PCSK-9 inhibitors, alirocumab and evolocumab, were approved by the US Food and Drug Administration (FDA) after randomized controlled trials demonstrated their efficacy in lowering LDL by approximately 60% and reducing CVD events by approximately 15% in patients on maximum-tolerated statin therapy.^20-22 To date, non-statin LDL-cholesterol-lowering agents have been subject to limited study in RA.²³

Identification and management of dyslipidemia offers an opportunity for substantial CVD risk reduction at the RA population level. Unfortunately, current rates of lipid screening are inadequate in this high-risk group. In a study of 3298 Medicare patients with RA, less than half of RA patients with an indication underwent appropriate lipid screening.²⁴ Additionally, statins are often underutilized for both primary and secondary prevention in RA patients. Only 27% of RA patients meeting National Cholesterol Education Program Adult Treatment Panel III criteria were initiated on statin therapy in a population-based cohort study.²⁵ Among patients discharged after a first myocardial infarction (MI), the odds of receiving lipid-lowering therapy were 31% lower for RA patients (odds ratio [OR], 0.69; 95% CI, 0.58-0.82).²⁶ Similar to the general population, adherence to statins in RA patients appears to be poor.^27-30 This raises particular concern considering that a population-based cohort study of RA patients demonstrated a 67% increased risk of MI associated with statin discontinuation, regardless of prior MI status.²⁷ Providers—rheumatologists, primary care providers, and cardiologists alike—need to remain vigilant in efforts to assess CVD risk to identify patients who will benefit from lipid-lowering therapy and to emphasize the importance to patients of statin adherence. Novel models of health-care delivery, health technologies, and patient engagement in care may prove useful for improving lipid screening and management in RA.

Tobacco Use

Cigarette smoking is a shared risk factor for both CVD and RA. Large cohort studies have identified a dose-dependent increased risk of incident RA, particularly seropositive RA, among smokers.^31-34 Tobacco smoking has also been associated with increased levels of inflammation and RA disease activity.³⁵ The consequences of tobacco use in the general population are staggering. Among individuals over the age of 30 years, tobacco use is responsible for 12% of all deaths and 10% of all CVD deaths.³⁶ Similar findings are observed in RA; a recent meta-analysis estimated there is a 50% increased risk of CVD events in RA related to smoking tobacco.³⁷ In the general population, smoking cessation markedly lowers CVD risk, and over time CVD risk may approach that of nonsmokers.^38,39 Thus, regular counseling and interventions to facilitate smoking cessation are critical to reducing CVD risk in RA patients. RA-specific smoking cessation programs have been proposed, but have yet to outperform standard smoking cessation programs.⁴⁰

Diabetes Mellitus

It is estimated that almost 10% of the US population has diabetes mellitus (DM), which in isolation portends substantial CVD risk.⁴¹ There is an increased prevalence of DM in RA, perhaps owing to factors such as physical inactivity and chronic glucocorticoid use, though a higher level of RA disease activity itself has been associated with increased insulin resistance.^42-45 In a cohort of 100 RA patients who were neither obese nor diabetic, RA patients had significantly higher fasting blood glucose and insulin levels than age- and sex-matched controls. These findings were even more pronounced in RA patients with higher levels of disease activity.⁴⁴ Similar to the general population, DM is associated with poor CVD outcomes in RA.³⁷ Therefore, both appropriate management of diabetes and control of RA disease activity are vitally important to minimize CVD risk related to DM.

Though not a universal finding, there may be an increased prevalence of hypertension in RA patients.^31,46 Nonsteroidal anti-inflammatory drug (NSAID) and glucocorticoid use may play a role in the development of hypertension, while DMARDs appear to exert a less substantial effect on blood pressure.^47,48 At least one study found that DMARD initiation (particularly for methotrexate and hydroxychloroquine) was associated with significant, albeit small, declines in both systolic and diastolic blood pressure over the first 6 months of treatment.⁴⁹

Despite its potentially higher prevalence in this population, hypertension is both underdiagnosed and undertreated in RA patients.^24,50-52 This is an important deficiency to target because, as in the general population, hypertension is associated with an increased risk of MI (RR, 1.84; 95% CI, 1.38-2.46) and composite CVD outcomes (RR, 2.24; 95% CI, 1.42-3.06) in RA.³⁷ Thresholds for initiation and escalation of antihypertensive therapy are not specific to the RA population; thus, diagnosis and management of hypertension should be informed by the American College of Cardiology/American Heart Association guidelines, treating those with in-office blood pressures exceeding 140/90 mm Hg (> 130/80 mm Hg if aged > 65 years or with concomitant CVD, DM, chronic kidney disease, or 10-year atherosclerotic cardiovascular disease risk > 10%), typically with angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, calcium channel blockers, or thiazide diuretics as comorbidities may dictate or allow.⁵³ Also, the use of NSAIDs and glucocorticoids should be minimized, particularly in those with concomitant hypertension.

Physical Activity

Likely due to factors such as articular pain and stiffness, as well as physical limitations, RA patients are more sedentary than the general population.^54,55 In a study of objectively assessed sedentary behavior in RA patients, greater average sedentary time per day and greater number of sedentary bouts (> 20 min) were associated with increased 10-year risk of CVD as assessed by the QRISK2.⁵⁶ Conversely, the beneficial effects of exercise are well documented. Light to moderate physical activity has been associated with improved cardiovascular outcomes, greater physical function, higher levels of HDL, as well as reduced systemic inflammation and disease activity, and improved endothelial function in RA patients.^57-61 While there has been concern that physical activity may result in accelerated joint damage, even high-intensity exercise was shown to be safe without causing significant progression of joint damage.⁵⁸

Obesity, Weight Loss, and Diet

While obesity is clearly associated with CVD risk in the general population, this relationship is much more complex in RA, as underweight RA patients are also at higher risk for CVD and CVD-related mortality.^62-64 One potential explanation for this finding is that pathological weight loss resulting in an underweight body mass index (BMI) is an independent predictor of CVD. In a study of US Veterans with RA, higher rates of weight loss (> 3 kg/m²/year) were associated with increased CVD mortality (HR, 2.27; 95% CI, 1.61-3.19) independent of BMI.⁶⁵ Systemic inflammation in RA can lead to “rheumatoid cachexia,” characterized by decreased muscle mass, increased adiposity, and increased CVD risk despite a normal or potentially decreased BMI.⁶⁶ Practitioners should be mindful of not only current body weight, but also patients’ weight trajectories when counseling on lifestyle practices such as healthy diet and regular exercise in RA patients. For obese individuals with RA, healthy weight loss should be encouraged. Interestingly, bariatric surgery in RA patients may improve RA disease activity in addition to its known effects on body weight and DM.⁶⁷

Counseling on healthy diet with a focus on limiting foods high in saturated- and trans-fatty acids and high glycemic index foods, and increasing consumption of fruits, vegetables, and mono-unsaturated fatty acids is a well-accepted and common practice to help minimize CVD risk in the general population.⁶⁸ No studies to date have investigated the effect of specific diets on CVD risk in RA patients, and thus we recommend adherence to general population recommendations.

Disease Activity

In addition to traditional risk factors, several studies have identified associations between the level of RA disease activity and risk of CVD. In a cohort of US Veterans with RA, CVD-related mortality increased in a dose-dependent manner with higher disease activity categories. In stark contrast, the CVD mortality rates of those in remission paralleled the rates from the general population (standardized mortality ratio [SMR], 0.68; 95% CI, 0.37-1.27).⁶⁹ In a separate cohort of 1157 RA patients without prior CVD, achieving low disease activity was associated with a lower risk of incident CVD events (HR, 0.65; 95% CI, 0.43-0.99).⁷⁰ Additionally, high disease activity has been associated with surrogate markers of CVD and other CVD risk factors including NT-proBNP and systolic blood pressure.^71,72 While no randomized controlled trial data is available to inform this recommendation, observational data suggest RA should be aggressively treated (ideally to achieve and maintain remission or low disease activity) to minimize CVD risk. While keeping this treatment goal in mind, the differential effects of specific RA therapies on CVD must also be considered.

Glucocorticoids and NSAIDs

With the expanding repertoire of DMARDs available and more aggressive treatment approaches, the role of glucocorticoids and NSAIDs in RA treatment is decreasing over time. While their efficacy for improving pain and stiffness is well established, concern regarding their contribution to CVD risk in RA patients is warranted.

Glucocorticoids are known to have detrimental effects on traditional CVD risk factors such as hypertension, insulin resistance, and dyslipidemia in the general population, as well as in RA patients.^73,74 In a meta-analysis of predominantly observational studies of RA patients, glucocorticoid use was associated with an increased risk of CVD events (RR, 1.47; 95% CI, 1.34-1.60), including MI, congestive heart failure (CHF), and cerebrovascular accident (CVA).⁷⁵ Evidence is conflicting in regards to a clear dose threshold that leads to increased CVD risk with glucocorticoids, though higher doses are associated with greater risk.^76-81 As RA patients requiring glucocorticoids typically have higher disease activity, confounding by indication remains a complicating factor in assessing the relative contributions of glucocorticoid use and RA disease activity to elevated CVD risk in many analyses.

The increased CVD risk with NSAID use is not specific to RA and has been well established in the general population.^82-84 In the previously mentioned meta-analysis, an increased overall risk of CVD events was observed with NSAID use in RA (RR, 1.18; 95% CI, 1.01-1.38). It should be noted that cyclo-oxygenase 2 (COX-2) inhibitors, in particular rofecoxib (now removed from the market), appeared to drive the majority of this risk (RR, 1.36; 95% CI, 1.10-1.67 in COX-2 inhibitors and RR 1.08, 95% CI, 0.94-1.24 in nonselective NSAIDs), suggesting a potential differential risk among NSAIDs.⁷⁵ While naproxen has been thought to carry the lowest risk of CVD based on initial studies, this has not been universally observed, including in a recent randomized controlled trial of more than 24,000 RA and osteoarthritis patients.^82,85,86

Providers should use the lowest possible dose and duration of glucocorticoids and NSAIDs to achieve symptom relief, with continual efforts to taper or discontinue. Candidates for glucocorticoid and NSAID therapy should be selected carefully, and use of these therapies should be avoided in those with prior CVD or at high risk for CVD based on traditional CVD risk factors. Most importantly, providers should focus on utilizing DMARDs for the management of RA, which more effectively treat RA as well as reduce CVD risk.

Methotrexate (MTX), a mainstay in the treatment of RA, is a conventional DMARD observed to improve overall survival and mitigate CVD risk in multiple RA cohorts.^75,87,88 In a recent meta-analysis comprised of 236,525 RA patients and 5410 CVD events, MTX use was associated with a 28% reduction in overall CVD events across 8 studies (RR, 0.72; 95% CI, 0.57-0.91), substantiating similar findings in a prior meta-analysis.^75,88 MTX use was specifically associated with a decreased risk of MI (RR, 0.81; 95% CI, 0.68-0.96). Case-control and cohort studies have cited a 20% to 50% reduced risk of CHF with MTX use.^89,90 The potential cardioprotective effect of MTX appears to be both multifactorial and complex, likely mediated through both direct and indirect mechanisms. MTX directly promotes anti-atherogenic lipoprotein function, improves endothelial function, and scavenges free radicals.^91,92 Indirectly, MTX likely reduces CVD risk by effectively reducing RA disease activity. Based on these and other data, MTX remains the cornerstone of DMARD therapy in RA patients when targeting CVD risk reduction.

Hydroxychloroquine

Emerging evidence suggests that hydroxychloroquine (HCQ), an antimalarial most often utilized in combination with alternative DMARDs in RA, prevents DM and has beneficial effects on lipid profiles. A recent meta-analysis compiled 3 homogenous observational studies that investigated the effect of HCQ on incident DM. RA patients ever exposed to HCQ had a 40% lower incidence of DM (HR, 0.59; 95% CI, 0.49-0.70).⁹³ Increased duration of HCQ use was shown to further reduce risk of incident DM.⁹⁴ The aforementioned meta-analysis also pooled 5 studies investigating the effect of HCQ on lipid profiles, with favorable mean differences in TC (–9.82 mg/dL), LDL (–10.61 mg/dL), HDL (4.13 mg/dL), and triglycerides (–19.15 mg/dL) in HCQ users compared to non-users.⁹³ Given these favorable changes to traditional CVD risk factors, it is not surprising that in a retrospective study of 1266 RA patients without prior CVD, HCQ was associated with significantly lower risk of incident CVD. While external validation of these findings is needed, HCQ is an attractive conventional DMARD to be used in RA for CVD risk reduction. Moreover, its combination with MTX and sulfasalazine also shows promise for CVD risk reduction.^95,96

TNF Inhibitors

Tumor necrosis factor (TNF) inhibitors are often the initial biologic DMARD therapy used in RA patients not responding to conventional DMARDs. In the previously described meta-analysis, TNF inhibitors were associated with similar reductions in CVD events as MTX (RR, 0.70; 95% CI, 0.54-0.90).⁷⁵ Of note, there was a trend toward reduced risk of CHF (RR, 0.75; 95% CI, 0.49-1.15) in this same meta-analysis, an area of concern with TNF inhibitor use due to a prior randomized controlled trial demonstrating worsening clinical status in patients with existing moderate-to-severe CHF treated with high-dose infliximab.⁹⁷ Current RA treatment guidelines recommend avoiding TNF inhibitor use in individuals with CHF.⁹⁸

Aside from the risk of CHF exacerbation, TNF inhibitors appear to be cardioprotective. Similar to MTX, the mechanism by which TNF inhibition reduces cardiovascular risk is complex and likely due to both direct and indirect mechanisms. Substantial research has been conducted on the effect of TNF inhibition on lipids, with a recent meta-analysis demonstrating increases in HDL and TC, with stable LDL and atherogenic index over treatment follow-up.⁹⁹ A subsequent meta-analysis not limited to RA patients yielded similar results.¹⁰⁰ In addition to quantitative lipid changes, alteration of lipoprotein function, improvement in myocardial function, reduced aortic stiffness, improved blood pressure, and reduced RA disease activity may also be responsible for cardioprotective benefits of these agents.^101,102

Non-TNF Biologic and Traditional Synthetic DMARDs

Tocilizumab, an IL-6 inhibitor, can potently increase LDL levels, but it does not appear to increase the risk of CVD events and may actually promote more favorable anti-atherogenic lipoprotein function.^103-106 Although these quantitative lipid changes received significant attention in the wake of early reports detailing this effect, similar lipid changes appear to accompany other DMARDs including TNF inhibitors and tofacitinib.¹⁰⁷ There have been few studies evaluating the risk of CVD with other non-TNF inhibitor biologic DMARDs and traditional synthetic DMARDs, warranting future study.

Conclusion

To mitigate the increased risk of CVD in RA, primary care and subspecialty providers alike must be aware of this heightened risk in RA, perform frequent assessments of CVD risk,³ and aggressively manage both traditional and nontraditional CVD risk factors. The differential roles in this effort may not be clear; thus, we have proposed a co-management strategy detailed in the Figure. Clear communication between providers is of the utmost importance to ensure effective management of CVD risk.

Given limited evidence for RA-specific CVD risk assessments and traditional risk factor treatment targets, management should follow pertinent national guidelines. The importance of lifestyle counseling should not be overlooked, with a focus on smoking cessation, healthy diet and body weight, and regular aerobic exercise. Finally, rheumatologists should aggressively manage RA using a treat-to-target approach, minimize the use of glucocorticoids and NSAIDs, and preferentially select DMARDs that have been associated with lower CVD risk. Through this comprehensive approach, recent trends of improved CVD outcomes in RA will hopefully become more widespread.¹⁰⁸

Corresponding author: Bryant R. England, MD; 986270 Nebraska Medical Center, Omaha, NE 68198-6270; [email protected].

Financial disclosures: Dr. England is supported by UNMC Internal Medicine Scientist Development Award, UNMC Physician-Scientist Training Program, the UNMC Mentored Scholars Program, and the Rheumatology Research Foundation Scientist Development Award. Dr. Mikuls is supported by a VA Merit Award (CX000896) and grants from the National Institutes of Health: National Institute of General Medical Sciences (U54GM115458), National Institute on Alcohol Abuse and Alcoholism (R25AA020818), and National Institute of Arthritis and Musculoskeletal and Skin Diseases (2P50AR60772).

When Erin Gilmer filled her insulin prescription at a Denver-area Walgreens in January, she paid $8.50. U.S. taxpayers paid another $280.51.

iStock/ThinkStock

“It eats at me to know that taxpayer money is being wasted,” said Gilmer, who has Medicare and was diagnosed with type 1 diabetes while a sophomore at the University of Colorado in 2002.

The diagnosis meant that for the rest of her life she’d require daily insulin shots to stay alive. But the price of that insulin is skyrocketing.

Between 2009 and 2017 the wholesale price of a single vial of Humalog, the Eli Lilly and Co.–manufactured insulin Gilmer uses, nearly tripled – rising from $92.70 to $274.70, according to data from IBM Watson Health.

Six years ago, Gilmer qualified for Social Security Disability Insurance – and thus, Medicare – because of a range of health issues. At the time, the insulin she needed cost $167.70 per vial, according to IBM Watson Health.

“When it’s taxpayer money paying for medication for someone like me, it makes it a national issue, not just a diabetic issue,” Gilmer said.

Stories about people with type 1 diabetes dying when they couldn’t afford insulin have made headlines. Patient activists like Gilmer have protested high prices outside Lilly’s headquarters in Indianapolis.

Last October in Minnesota, State Attorney General Lori Swanson sued insulin manufacturers, alleging price gouging. Pharmaceutical executives were grilled about high drug prices by the Senate Finance Committee on Feb. 26.

This is the backdrop for Lilly’s announcement March 4 that it is rolling out a half-priced, generic version of Humalog called “insulin lispro.” The list price: $137.35 per vial.

“Patients, doctors and policymakers are demanding lower list prices for medicines and lower patient costs at the pharmacy counter,” Eli Lilly CEO David Ricks wrote in a blog post about the move. “You might be surprised to hear that we agree – it’s time for change in our system and for consumer prices to come down.”
 

No panacea

When Lilly’s Humalog, the first short-acting insulin, came to market in 1996, the list price was about $21 per vial. The price didn’t reach $275 overnight, but yearly price increases added up.

In February 2009, for example, the wholesale price was $92.70, according to IBM Watson Health. It rose to $99.65 in December 2009, then to $107.60 in September 2010, $115.70 in May 2011, and so on.

“There’s no justification for why prices should keep increasing at an average rate of 10% every year,” said Inmaculada Hernandez of the University of Pittsburgh School of Pharmacy, who was lead author of a January report in Health Affairs attributing the skyrocketing cost of prescription drugs to accumulated yearly price hikes.

“The public perception that we have in general is that drugs are so expensive because we need to pay for research and development, and that’s true,” Hernandez said. “However, usually research and development is paid for in the first years of life of a drug.”

At $137.35 per vial, Lilly’s generic insulin is priced at about the same level as Humalog was in 2012, 16 years after it came to market.

“We want to recognize that this is not a panacea,” said company spokesman Greg Kueterman. “This is an option that we hope can help people in the current system that we work with.”

It’s worth noting that Humalog is a rapid-acting insulin, but that’s only one of the two types of insulin most people with type 1 diabetes use every day. The second kind is long-lasting. Lilly makes one called Basaglar. The most popular long-lasting insulin is Lantus, produced by Sanofi. Neither has a lower-cost alternative.

Still, Lilly’s move on Humalog could put pressure on the other two big makers of insulin to act.

Novo Nordisk called Lilly’s lower-priced generic insulin “an important development,” in an emailed statement.

“Bringing affordable insulin to the market requires ideas from all stakeholders,” Novo Nordisk’s Ken Inchausti said in an email, which also listed steps the company has taken, such as a patient assistance program. The statement didn’t say whether Novo Nordisk is considering offering a lower-priced version of its popular insulin Novolog, a rival of Humalog.

A statement from Sanofi, the third major insulin maker, also didn’t say whether the company would offer lower-priced versions of its insulins.

“Sanofi supports any actions that increase access to insulins for patients living with diabetes at an affordable price,” spokeswoman Ashleigh Koss said in the email, which also touted the company’s patient assistance program.
 

A different kind of generic

One twist in this story is that Lilly’s new insulin is just a repackaged version of Humalog, minus the brand name. It’s called an “authorized generic.”

“Whoever came up with the term ‘authorized generic’?” Dr. Vincent Rajkumar said, laughing. Rajkumar is a hematologist at the Mayo Clinic in Rochester, Minn.

“It’s the same exact drug” as the brand name, he continued.

Typically, Rajkumar said, authorized generics are introduced by brand-name drugmakers to compete with generic versions of their drugs made by rival companies.

But in the case of Humalog and other insulins, there are no generics made by competitors, as there are for, say, the cholesterol medicine Lipitor or even other diabetes drugs, such as metformin.

So when Lilly’s authorized generic comes to market, the company will have both Humalog insulin and the authorized generic version of that medicine on the market.

Rajkumar said it’s a public relations move.

“There’s outrage over the price of insulin that is being discussed in Congress and elsewhere. And so the company basically says, ‘Hey, we will make the identical product available at half price.’ On the surface that sounds great,” Rajkumar said.

“But you look at the problems and you think, ‘OK, how crazy is this that someone is actually going to be buying the brand-name drug?’ ”

In fact, it’s possible that Lilly could break even or profit off its authorized generic compared to the name-brand Humalog, according to University of Pittsburgh’s Hernandez.

The profit margin would depend on the rebates paid by the company to insurers and pharmacy benefit managers. Rebates are getting a lot of attention these days as one factor that pushes drug prices higher. They’re usually not disclosed and increase as a drug’s price increases, providing an incentive to some companies to raise prices.

“Doing an authorized generic is nothing else than giving insurers two options,” Hernandez said: Pay the full list price for a brand-name drug and receive a higher rebate, or pay the lower price for the authorized generic and receive a presumably smaller rebate.

“What we really need to get insulin prices down is to get generics into the market, and we need more than one,” Hernandez said, adding that previous research has shown that prices begin to go down when two or three generics are competing in the marketplace.

Even so, Lillly’s Kueterman said the authorized generic insulin “is going to help hopefully move the system toward a more sustainable model.”

“I can guarantee you the reason that we’re doing this is to help people,” Kueterman said, noting the company’s Diabetes Solution Center has also helped “10,000 people each month pay significantly less for their insulin” since it opened in August.

For Erin Gilmer, the news about an authorized generic insulin from Lilly has left her mildly encouraged.

“It sounds really good, and it will help some people, which is great,” Gilmer said. “It’s Eli Lilly and pharma starting to understand that grassroots activism has to be taken seriously, and we are at a tipping point.”
 

This story is part of a partnership that includes NPR and Kaiser Health News. Kaiser Health News is a nonprofit national health policy news service. It is an editorially independent program of the Henry J. Kaiser Family Foundation that is not affiliated with Kaiser Permanente.

When Erin Gilmer filled her insulin prescription at a Denver-area Walgreens in January, she paid $8.50. U.S. taxpayers paid another $280.51.

iStock/ThinkStock

“It eats at me to know that taxpayer money is being wasted,” said Gilmer, who has Medicare and was diagnosed with type 1 diabetes while a sophomore at the University of Colorado in 2002.

The diagnosis meant that for the rest of her life she’d require daily insulin shots to stay alive. But the price of that insulin is skyrocketing.

Between 2009 and 2017 the wholesale price of a single vial of Humalog, the Eli Lilly and Co.–manufactured insulin Gilmer uses, nearly tripled – rising from $92.70 to $274.70, according to data from IBM Watson Health.

Six years ago, Gilmer qualified for Social Security Disability Insurance – and thus, Medicare – because of a range of health issues. At the time, the insulin she needed cost $167.70 per vial, according to IBM Watson Health.

“When it’s taxpayer money paying for medication for someone like me, it makes it a national issue, not just a diabetic issue,” Gilmer said.

Stories about people with type 1 diabetes dying when they couldn’t afford insulin have made headlines. Patient activists like Gilmer have protested high prices outside Lilly’s headquarters in Indianapolis.

Last October in Minnesota, State Attorney General Lori Swanson sued insulin manufacturers, alleging price gouging. Pharmaceutical executives were grilled about high drug prices by the Senate Finance Committee on Feb. 26.

This is the backdrop for Lilly’s announcement March 4 that it is rolling out a half-priced, generic version of Humalog called “insulin lispro.” The list price: $137.35 per vial.

“Patients, doctors and policymakers are demanding lower list prices for medicines and lower patient costs at the pharmacy counter,” Eli Lilly CEO David Ricks wrote in a blog post about the move. “You might be surprised to hear that we agree – it’s time for change in our system and for consumer prices to come down.”
 

No panacea

When Lilly’s Humalog, the first short-acting insulin, came to market in 1996, the list price was about $21 per vial. The price didn’t reach $275 overnight, but yearly price increases added up.

In February 2009, for example, the wholesale price was $92.70, according to IBM Watson Health. It rose to $99.65 in December 2009, then to $107.60 in September 2010, $115.70 in May 2011, and so on.

“There’s no justification for why prices should keep increasing at an average rate of 10% every year,” said Inmaculada Hernandez of the University of Pittsburgh School of Pharmacy, who was lead author of a January report in Health Affairs attributing the skyrocketing cost of prescription drugs to accumulated yearly price hikes.

“The public perception that we have in general is that drugs are so expensive because we need to pay for research and development, and that’s true,” Hernandez said. “However, usually research and development is paid for in the first years of life of a drug.”

At $137.35 per vial, Lilly’s generic insulin is priced at about the same level as Humalog was in 2012, 16 years after it came to market.

“We want to recognize that this is not a panacea,” said company spokesman Greg Kueterman. “This is an option that we hope can help people in the current system that we work with.”

It’s worth noting that Humalog is a rapid-acting insulin, but that’s only one of the two types of insulin most people with type 1 diabetes use every day. The second kind is long-lasting. Lilly makes one called Basaglar. The most popular long-lasting insulin is Lantus, produced by Sanofi. Neither has a lower-cost alternative.

Still, Lilly’s move on Humalog could put pressure on the other two big makers of insulin to act.

Novo Nordisk called Lilly’s lower-priced generic insulin “an important development,” in an emailed statement.

“Bringing affordable insulin to the market requires ideas from all stakeholders,” Novo Nordisk’s Ken Inchausti said in an email, which also listed steps the company has taken, such as a patient assistance program. The statement didn’t say whether Novo Nordisk is considering offering a lower-priced version of its popular insulin Novolog, a rival of Humalog.

A statement from Sanofi, the third major insulin maker, also didn’t say whether the company would offer lower-priced versions of its insulins.

“Sanofi supports any actions that increase access to insulins for patients living with diabetes at an affordable price,” spokeswoman Ashleigh Koss said in the email, which also touted the company’s patient assistance program.
 

A different kind of generic

One twist in this story is that Lilly’s new insulin is just a repackaged version of Humalog, minus the brand name. It’s called an “authorized generic.”

“Whoever came up with the term ‘authorized generic’?” Dr. Vincent Rajkumar said, laughing. Rajkumar is a hematologist at the Mayo Clinic in Rochester, Minn.

“It’s the same exact drug” as the brand name, he continued.

Typically, Rajkumar said, authorized generics are introduced by brand-name drugmakers to compete with generic versions of their drugs made by rival companies.

But in the case of Humalog and other insulins, there are no generics made by competitors, as there are for, say, the cholesterol medicine Lipitor or even other diabetes drugs, such as metformin.

So when Lilly’s authorized generic comes to market, the company will have both Humalog insulin and the authorized generic version of that medicine on the market.

Rajkumar said it’s a public relations move.

“There’s outrage over the price of insulin that is being discussed in Congress and elsewhere. And so the company basically says, ‘Hey, we will make the identical product available at half price.’ On the surface that sounds great,” Rajkumar said.

“But you look at the problems and you think, ‘OK, how crazy is this that someone is actually going to be buying the brand-name drug?’ ”

In fact, it’s possible that Lilly could break even or profit off its authorized generic compared to the name-brand Humalog, according to University of Pittsburgh’s Hernandez.

The profit margin would depend on the rebates paid by the company to insurers and pharmacy benefit managers. Rebates are getting a lot of attention these days as one factor that pushes drug prices higher. They’re usually not disclosed and increase as a drug’s price increases, providing an incentive to some companies to raise prices.

“Doing an authorized generic is nothing else than giving insurers two options,” Hernandez said: Pay the full list price for a brand-name drug and receive a higher rebate, or pay the lower price for the authorized generic and receive a presumably smaller rebate.

“What we really need to get insulin prices down is to get generics into the market, and we need more than one,” Hernandez said, adding that previous research has shown that prices begin to go down when two or three generics are competing in the marketplace.

Even so, Lillly’s Kueterman said the authorized generic insulin “is going to help hopefully move the system toward a more sustainable model.”

“I can guarantee you the reason that we’re doing this is to help people,” Kueterman said, noting the company’s Diabetes Solution Center has also helped “10,000 people each month pay significantly less for their insulin” since it opened in August.

For Erin Gilmer, the news about an authorized generic insulin from Lilly has left her mildly encouraged.

“It sounds really good, and it will help some people, which is great,” Gilmer said. “It’s Eli Lilly and pharma starting to understand that grassroots activism has to be taken seriously, and we are at a tipping point.”
 

This story is part of a partnership that includes NPR and Kaiser Health News. Kaiser Health News is a nonprofit national health policy news service. It is an editorially independent program of the Henry J. Kaiser Family Foundation that is not affiliated with Kaiser Permanente.

When Erin Gilmer filled her insulin prescription at a Denver-area Walgreens in January, she paid $8.50. U.S. taxpayers paid another $280.51.

iStock/ThinkStock

“It eats at me to know that taxpayer money is being wasted,” said Gilmer, who has Medicare and was diagnosed with type 1 diabetes while a sophomore at the University of Colorado in 2002.

The diagnosis meant that for the rest of her life she’d require daily insulin shots to stay alive. But the price of that insulin is skyrocketing.

Between 2009 and 2017 the wholesale price of a single vial of Humalog, the Eli Lilly and Co.–manufactured insulin Gilmer uses, nearly tripled – rising from $92.70 to $274.70, according to data from IBM Watson Health.

Six years ago, Gilmer qualified for Social Security Disability Insurance – and thus, Medicare – because of a range of health issues. At the time, the insulin she needed cost $167.70 per vial, according to IBM Watson Health.

“When it’s taxpayer money paying for medication for someone like me, it makes it a national issue, not just a diabetic issue,” Gilmer said.

Stories about people with type 1 diabetes dying when they couldn’t afford insulin have made headlines. Patient activists like Gilmer have protested high prices outside Lilly’s headquarters in Indianapolis.

Last October in Minnesota, State Attorney General Lori Swanson sued insulin manufacturers, alleging price gouging. Pharmaceutical executives were grilled about high drug prices by the Senate Finance Committee on Feb. 26.

This is the backdrop for Lilly’s announcement March 4 that it is rolling out a half-priced, generic version of Humalog called “insulin lispro.” The list price: $137.35 per vial.

“Patients, doctors and policymakers are demanding lower list prices for medicines and lower patient costs at the pharmacy counter,” Eli Lilly CEO David Ricks wrote in a blog post about the move. “You might be surprised to hear that we agree – it’s time for change in our system and for consumer prices to come down.”
 

No panacea

When Lilly’s Humalog, the first short-acting insulin, came to market in 1996, the list price was about $21 per vial. The price didn’t reach $275 overnight, but yearly price increases added up.

In February 2009, for example, the wholesale price was $92.70, according to IBM Watson Health. It rose to $99.65 in December 2009, then to $107.60 in September 2010, $115.70 in May 2011, and so on.

“There’s no justification for why prices should keep increasing at an average rate of 10% every year,” said Inmaculada Hernandez of the University of Pittsburgh School of Pharmacy, who was lead author of a January report in Health Affairs attributing the skyrocketing cost of prescription drugs to accumulated yearly price hikes.

“The public perception that we have in general is that drugs are so expensive because we need to pay for research and development, and that’s true,” Hernandez said. “However, usually research and development is paid for in the first years of life of a drug.”

At $137.35 per vial, Lilly’s generic insulin is priced at about the same level as Humalog was in 2012, 16 years after it came to market.

“We want to recognize that this is not a panacea,” said company spokesman Greg Kueterman. “This is an option that we hope can help people in the current system that we work with.”

It’s worth noting that Humalog is a rapid-acting insulin, but that’s only one of the two types of insulin most people with type 1 diabetes use every day. The second kind is long-lasting. Lilly makes one called Basaglar. The most popular long-lasting insulin is Lantus, produced by Sanofi. Neither has a lower-cost alternative.

Still, Lilly’s move on Humalog could put pressure on the other two big makers of insulin to act.

Novo Nordisk called Lilly’s lower-priced generic insulin “an important development,” in an emailed statement.

“Bringing affordable insulin to the market requires ideas from all stakeholders,” Novo Nordisk’s Ken Inchausti said in an email, which also listed steps the company has taken, such as a patient assistance program. The statement didn’t say whether Novo Nordisk is considering offering a lower-priced version of its popular insulin Novolog, a rival of Humalog.

A statement from Sanofi, the third major insulin maker, also didn’t say whether the company would offer lower-priced versions of its insulins.

“Sanofi supports any actions that increase access to insulins for patients living with diabetes at an affordable price,” spokeswoman Ashleigh Koss said in the email, which also touted the company’s patient assistance program.
 

A different kind of generic

One twist in this story is that Lilly’s new insulin is just a repackaged version of Humalog, minus the brand name. It’s called an “authorized generic.”

“Whoever came up with the term ‘authorized generic’?” Dr. Vincent Rajkumar said, laughing. Rajkumar is a hematologist at the Mayo Clinic in Rochester, Minn.

“It’s the same exact drug” as the brand name, he continued.

Typically, Rajkumar said, authorized generics are introduced by brand-name drugmakers to compete with generic versions of their drugs made by rival companies.

But in the case of Humalog and other insulins, there are no generics made by competitors, as there are for, say, the cholesterol medicine Lipitor or even other diabetes drugs, such as metformin.

So when Lilly’s authorized generic comes to market, the company will have both Humalog insulin and the authorized generic version of that medicine on the market.

Rajkumar said it’s a public relations move.

“There’s outrage over the price of insulin that is being discussed in Congress and elsewhere. And so the company basically says, ‘Hey, we will make the identical product available at half price.’ On the surface that sounds great,” Rajkumar said.

“But you look at the problems and you think, ‘OK, how crazy is this that someone is actually going to be buying the brand-name drug?’ ”

In fact, it’s possible that Lilly could break even or profit off its authorized generic compared to the name-brand Humalog, according to University of Pittsburgh’s Hernandez.

The profit margin would depend on the rebates paid by the company to insurers and pharmacy benefit managers. Rebates are getting a lot of attention these days as one factor that pushes drug prices higher. They’re usually not disclosed and increase as a drug’s price increases, providing an incentive to some companies to raise prices.

“Doing an authorized generic is nothing else than giving insurers two options,” Hernandez said: Pay the full list price for a brand-name drug and receive a higher rebate, or pay the lower price for the authorized generic and receive a presumably smaller rebate.

“What we really need to get insulin prices down is to get generics into the market, and we need more than one,” Hernandez said, adding that previous research has shown that prices begin to go down when two or three generics are competing in the marketplace.

Even so, Lillly’s Kueterman said the authorized generic insulin “is going to help hopefully move the system toward a more sustainable model.”

“I can guarantee you the reason that we’re doing this is to help people,” Kueterman said, noting the company’s Diabetes Solution Center has also helped “10,000 people each month pay significantly less for their insulin” since it opened in August.

For Erin Gilmer, the news about an authorized generic insulin from Lilly has left her mildly encouraged.

“It sounds really good, and it will help some people, which is great,” Gilmer said. “It’s Eli Lilly and pharma starting to understand that grassroots activism has to be taken seriously, and we are at a tipping point.”
 

This story is part of a partnership that includes NPR and Kaiser Health News. Kaiser Health News is a nonprofit national health policy news service. It is an editorially independent program of the Henry J. Kaiser Family Foundation that is not affiliated with Kaiser Permanente.

A 62-year-old woman presented to our emergency department with fever, chills, hoarseness, pain on swallowing, and a painful neck. Her symptoms had begun 1 day earlier. Because acetaminophen brought no improvement, she went to an urgent care facility, where a nasal swab polymerase chain reaction test was positive for influenza A, and a throat swab rapid test was positive for group A streptococci. She was then referred to our emergency department.

She reported no pre-existing conditions predisposing her to infection. Her temperature was 99.9°F (37.7°C), pulse 112 beats per minute, and respiratory rate 24 breaths per minute. The physical examination was unremarkable except for bilateral anterior cervical adenopathy and bilateral anterior neck tenderness. Her pharynx was not injected, and no exudate, palatal edema, or petechiae were noted.

Results of initial laboratory testing were as follows:

• White blood cell count 20.5 × 10⁹/L (reference range 3.9–11)
• Neutrophils 76% (42%–75%)
• Bands 15% (0%–5%)
• Lymphocytes 3% (21%–51%)
• Erythrocyte sedimentation rate 75 mm/h (< 20 mm/h)
• C-reactive protein 247.14 mg/L (≤ 3 mg/L)
• Serum aminotransferase levels were normal.
• Polymerase chain reaction testing of a nasal swab was negative for viral infection.

Throat swabs and blood samples were sent for culture.

She was started on ceftriaxone 1 g intravenously every 24 hours, with close observation in the medical intensive care unit, where she was admitted because of epiglottitis. On hospital day 3, the throat culture was reported as negative, but the blood culture was reported as positive for Haemophilus influenzae. Thus, the clinical diagnosis was acute epiglottitis due to H influenzae, not group A streptococci.

The patient completed 10 days of ceftriaxone therapy; her recovery was uneventful, and she was discharged on hospital day 10.

INFLUENZA: CHALLENGES TO PROMPT, ACCURATE DIAGNOSIS

During influenza season, emergency departments are inundated with adults with influenza A and other viral respiratory infections. This makes prompt, accurate diagnosis a challenge,¹ given the broad differential diagnosis.^2,3 Adults with influenza and its complications as well as unrelated conditions can present a special challenge.⁴

Our patient presented with acute-onset influenza A and was then found to have acute epiglottitis, an unexpected complication of influenza A.⁵ A positive rapid test for group A streptococci done at an urgent care facility led emergency department physicians to assume that the acute epiglottitis was due to group A streptococci. Unless correlated with clinical findings, results of rapid diagnostic tests may mislead the unwary practitioner. Accurate diagnosis should be based mainly on the history and physical findings. Results of rapid diagnostic tests can be helpful if interpreted in the clinical context.^6–8

The rapid test for streptococci is appropriate for the diagnosis of pharyngitis due to group A streptococci in people under age 30 with acute-onset sore throat, fever, and bilateral acute cervical adenopathy, without fatigue or myalgias. However, the rapid test does not differentiate colonization from infection. Group A streptococci are common colonizers with viral pharyngitis. In 30% of cases of Epstein-Barr virus pharyngitis, there is colonization with group A streptococci. A positive rapid test in such cases can result in the wrong diagnosis, ie, pharyngitis due to group A streptococci rather than Epstein-Barr virus.

A 62-year-old woman presented to our emergency department with fever, chills, hoarseness, pain on swallowing, and a painful neck. Her symptoms had begun 1 day earlier. Because acetaminophen brought no improvement, she went to an urgent care facility, where a nasal swab polymerase chain reaction test was positive for influenza A, and a throat swab rapid test was positive for group A streptococci. She was then referred to our emergency department.

She reported no pre-existing conditions predisposing her to infection. Her temperature was 99.9°F (37.7°C), pulse 112 beats per minute, and respiratory rate 24 breaths per minute. The physical examination was unremarkable except for bilateral anterior cervical adenopathy and bilateral anterior neck tenderness. Her pharynx was not injected, and no exudate, palatal edema, or petechiae were noted.

Results of initial laboratory testing were as follows:

• White blood cell count 20.5 × 10⁹/L (reference range 3.9–11)
• Neutrophils 76% (42%–75%)
• Bands 15% (0%–5%)
• Lymphocytes 3% (21%–51%)
• Erythrocyte sedimentation rate 75 mm/h (< 20 mm/h)
• C-reactive protein 247.14 mg/L (≤ 3 mg/L)
• Serum aminotransferase levels were normal.
• Polymerase chain reaction testing of a nasal swab was negative for viral infection.

Throat swabs and blood samples were sent for culture.

She was started on ceftriaxone 1 g intravenously every 24 hours, with close observation in the medical intensive care unit, where she was admitted because of epiglottitis. On hospital day 3, the throat culture was reported as negative, but the blood culture was reported as positive for Haemophilus influenzae. Thus, the clinical diagnosis was acute epiglottitis due to H influenzae, not group A streptococci.

The patient completed 10 days of ceftriaxone therapy; her recovery was uneventful, and she was discharged on hospital day 10.

INFLUENZA: CHALLENGES TO PROMPT, ACCURATE DIAGNOSIS

During influenza season, emergency departments are inundated with adults with influenza A and other viral respiratory infections. This makes prompt, accurate diagnosis a challenge,¹ given the broad differential diagnosis.^2,3 Adults with influenza and its complications as well as unrelated conditions can present a special challenge.⁴

Our patient presented with acute-onset influenza A and was then found to have acute epiglottitis, an unexpected complication of influenza A.⁵ A positive rapid test for group A streptococci done at an urgent care facility led emergency department physicians to assume that the acute epiglottitis was due to group A streptococci. Unless correlated with clinical findings, results of rapid diagnostic tests may mislead the unwary practitioner. Accurate diagnosis should be based mainly on the history and physical findings. Results of rapid diagnostic tests can be helpful if interpreted in the clinical context.^6–8

The rapid test for streptococci is appropriate for the diagnosis of pharyngitis due to group A streptococci in people under age 30 with acute-onset sore throat, fever, and bilateral acute cervical adenopathy, without fatigue or myalgias. However, the rapid test does not differentiate colonization from infection. Group A streptococci are common colonizers with viral pharyngitis. In 30% of cases of Epstein-Barr virus pharyngitis, there is colonization with group A streptococci. A positive rapid test in such cases can result in the wrong diagnosis, ie, pharyngitis due to group A streptococci rather than Epstein-Barr virus.

A 62-year-old woman presented to our emergency department with fever, chills, hoarseness, pain on swallowing, and a painful neck. Her symptoms had begun 1 day earlier. Because acetaminophen brought no improvement, she went to an urgent care facility, where a nasal swab polymerase chain reaction test was positive for influenza A, and a throat swab rapid test was positive for group A streptococci. She was then referred to our emergency department.

She reported no pre-existing conditions predisposing her to infection. Her temperature was 99.9°F (37.7°C), pulse 112 beats per minute, and respiratory rate 24 breaths per minute. The physical examination was unremarkable except for bilateral anterior cervical adenopathy and bilateral anterior neck tenderness. Her pharynx was not injected, and no exudate, palatal edema, or petechiae were noted.

Results of initial laboratory testing were as follows:

• White blood cell count 20.5 × 10⁹/L (reference range 3.9–11)
• Neutrophils 76% (42%–75%)
• Bands 15% (0%–5%)
• Lymphocytes 3% (21%–51%)
• Erythrocyte sedimentation rate 75 mm/h (< 20 mm/h)
• C-reactive protein 247.14 mg/L (≤ 3 mg/L)
• Serum aminotransferase levels were normal.
• Polymerase chain reaction testing of a nasal swab was negative for viral infection.

Throat swabs and blood samples were sent for culture.

She was started on ceftriaxone 1 g intravenously every 24 hours, with close observation in the medical intensive care unit, where she was admitted because of epiglottitis. On hospital day 3, the throat culture was reported as negative, but the blood culture was reported as positive for Haemophilus influenzae. Thus, the clinical diagnosis was acute epiglottitis due to H influenzae, not group A streptococci.

The patient completed 10 days of ceftriaxone therapy; her recovery was uneventful, and she was discharged on hospital day 10.

INFLUENZA: CHALLENGES TO PROMPT, ACCURATE DIAGNOSIS

During influenza season, emergency departments are inundated with adults with influenza A and other viral respiratory infections. This makes prompt, accurate diagnosis a challenge,¹ given the broad differential diagnosis.^2,3 Adults with influenza and its complications as well as unrelated conditions can present a special challenge.⁴

Our patient presented with acute-onset influenza A and was then found to have acute epiglottitis, an unexpected complication of influenza A.⁵ A positive rapid test for group A streptococci done at an urgent care facility led emergency department physicians to assume that the acute epiglottitis was due to group A streptococci. Unless correlated with clinical findings, results of rapid diagnostic tests may mislead the unwary practitioner. Accurate diagnosis should be based mainly on the history and physical findings. Results of rapid diagnostic tests can be helpful if interpreted in the clinical context.^6–8

The rapid test for streptococci is appropriate for the diagnosis of pharyngitis due to group A streptococci in people under age 30 with acute-onset sore throat, fever, and bilateral acute cervical adenopathy, without fatigue or myalgias. However, the rapid test does not differentiate colonization from infection. Group A streptococci are common colonizers with viral pharyngitis. In 30% of cases of Epstein-Barr virus pharyngitis, there is colonization with group A streptococci. A positive rapid test in such cases can result in the wrong diagnosis, ie, pharyngitis due to group A streptococci rather than Epstein-Barr virus.

Delaying or withholding antibiotics in elderly patients with a urinary tract infection (UTI) increases the risk of sepsis and death, results of a large, population-based study suggest.

andresr/Getty Images

The risk of bloodstream infection was more than seven times greater in patients who did not receive antibiotics immediately after seeing a general practitioner for a UTI versus those who did, according to results of the study based on primary care records and other data for nearly 160,000 U.K. patients aged 65 years or older. Death rates and hospital admissions were significantly higher for these patients, according to the study published in The BMJ by Myriam Gharbi, PharmD, Phd, Imperial College London, and her colleagues.

The publication of these findings coincides with an increase in Escherichia coli bloodstream infections in England.

“Our study suggests the early initiation of antibiotics for UTI in older high risk adult populations (especially men aged [older than] 85 years) should be recommended to prevent serious complications,” Dr. Gharbi and her coauthors said in their report.

The population-based cohort study comprised 157,264 adult primary care patients at least 65 years of age who had one or more suspected or confirmed lower UTIs from November 2007 to May 2015. The researchers found that health care providers had diagnosed a total of 312,896 UTI episodes in these patients during the period they studied. In 7.2% (22,534) of the UTI episodes, the researchers were unable to find records of the patients having been prescribed antibiotics by a general practitioner within 7 days of the UTI diagnosis. These 22,534 episodes included those that occurred in patients who had a complication before an antibiotic was prescribed. An additional 6.2% (19,292) of the episodes occurred in patients who were prescribed antibiotics, but not during their first UTI-related visit to a general practitioner or on the same day of such a visit. The researchers classified this group of patients as having been prescribed antibiotics on a deferred or delayed basis, as they were not prescribed such drugs within 7 days of their visit.

Overall, there were 1,539 cases (0.5% of the total number of UTIs) of bloodstream infection within 60 days of the initial urinary tract infection diagnosis, the researchers reported.

The bloodstream infection rate was 2.9% for patients who were not prescribed antibiotics ever or prior to an infection occurring, 2.2% in those who were prescribed antibiotics on a deferred basis, and 0.2% in those who were prescribed antibiotics immediately, meaning during their first visit to a general practitioner for a UTI or on the same day of such a visit (P less than .001). After adjustment for potential confounding variables such as age, sex, and region, the patients classified as having not been prescribed antibiotics or having been prescribed antibiotics on a deferred basis were significantly more likely to have a bloodstream infection within 60 days of their visit to a health care provider, compared with those who received antibiotics immediately, with odds ratios of 8.08 (95% confidence interval, 7.12-9.16) and 7.12 (95% CI, 6.22-8.14), respectively.

Hospital admissions after a UTI episode were nearly twice as high in the no- or deferred-antibiotics groups (27.0% and 26.8%, respectively), compared with the group that received antibiotics right away (14.8%), the investigators reported. The lengths of hospital stays were 12.1 days for the group classified as having not been prescribed antibiotics, 7.7 days for the group subject to delayed antibiotic prescribing, and 6.3 days for the group who received antibiotics immediately.

Deaths within 60 days of experiencing a urinary tract infection occurred in 5.4% of patients in the no-antibiotics group, 2.8% of the deferred-antibiotics group, and 1.6% of the immediate-antibiotics group. After adjustment for covariates, a regression analysis showed the risks for all-cause mortality were 1.16 and 2.18 times higher in the deferred-antibiotics group and the no-antibiotics group, respectively, according to the paper.

In the immediate-antibiotics group, those patients who received nitrofurantoin had a “small but significant increase” in 60-day survival versus those who received trimethoprim, the investigators noted in the discussion section of their report.

“This increase could reflect either higher levels of resistance to trimethoprim or a healthier population treated with nitrofurantoin, the latest being not recommended for patients with poor kidney function,” the researchers wrote.

This study was supported by the National Institute for Health Research and other U.K. sources. One study coauthor reported working as an epidemiologist with GSK in areas not related to the study.

SOURCE: Gharbi M et al. BMJ. 2019 Feb 27. doi: 10.1136/bmj.l525.

Delaying or withholding antibiotics in elderly patients with a urinary tract infection (UTI) increases the risk of sepsis and death, results of a large, population-based study suggest.

andresr/Getty Images

The risk of bloodstream infection was more than seven times greater in patients who did not receive antibiotics immediately after seeing a general practitioner for a UTI versus those who did, according to results of the study based on primary care records and other data for nearly 160,000 U.K. patients aged 65 years or older. Death rates and hospital admissions were significantly higher for these patients, according to the study published in The BMJ by Myriam Gharbi, PharmD, Phd, Imperial College London, and her colleagues.

The publication of these findings coincides with an increase in Escherichia coli bloodstream infections in England.

“Our study suggests the early initiation of antibiotics for UTI in older high risk adult populations (especially men aged [older than] 85 years) should be recommended to prevent serious complications,” Dr. Gharbi and her coauthors said in their report.

The population-based cohort study comprised 157,264 adult primary care patients at least 65 years of age who had one or more suspected or confirmed lower UTIs from November 2007 to May 2015. The researchers found that health care providers had diagnosed a total of 312,896 UTI episodes in these patients during the period they studied. In 7.2% (22,534) of the UTI episodes, the researchers were unable to find records of the patients having been prescribed antibiotics by a general practitioner within 7 days of the UTI diagnosis. These 22,534 episodes included those that occurred in patients who had a complication before an antibiotic was prescribed. An additional 6.2% (19,292) of the episodes occurred in patients who were prescribed antibiotics, but not during their first UTI-related visit to a general practitioner or on the same day of such a visit. The researchers classified this group of patients as having been prescribed antibiotics on a deferred or delayed basis, as they were not prescribed such drugs within 7 days of their visit.

Overall, there were 1,539 cases (0.5% of the total number of UTIs) of bloodstream infection within 60 days of the initial urinary tract infection diagnosis, the researchers reported.

The bloodstream infection rate was 2.9% for patients who were not prescribed antibiotics ever or prior to an infection occurring, 2.2% in those who were prescribed antibiotics on a deferred basis, and 0.2% in those who were prescribed antibiotics immediately, meaning during their first visit to a general practitioner for a UTI or on the same day of such a visit (P less than .001). After adjustment for potential confounding variables such as age, sex, and region, the patients classified as having not been prescribed antibiotics or having been prescribed antibiotics on a deferred basis were significantly more likely to have a bloodstream infection within 60 days of their visit to a health care provider, compared with those who received antibiotics immediately, with odds ratios of 8.08 (95% confidence interval, 7.12-9.16) and 7.12 (95% CI, 6.22-8.14), respectively.

Hospital admissions after a UTI episode were nearly twice as high in the no- or deferred-antibiotics groups (27.0% and 26.8%, respectively), compared with the group that received antibiotics right away (14.8%), the investigators reported. The lengths of hospital stays were 12.1 days for the group classified as having not been prescribed antibiotics, 7.7 days for the group subject to delayed antibiotic prescribing, and 6.3 days for the group who received antibiotics immediately.

Deaths within 60 days of experiencing a urinary tract infection occurred in 5.4% of patients in the no-antibiotics group, 2.8% of the deferred-antibiotics group, and 1.6% of the immediate-antibiotics group. After adjustment for covariates, a regression analysis showed the risks for all-cause mortality were 1.16 and 2.18 times higher in the deferred-antibiotics group and the no-antibiotics group, respectively, according to the paper.

In the immediate-antibiotics group, those patients who received nitrofurantoin had a “small but significant increase” in 60-day survival versus those who received trimethoprim, the investigators noted in the discussion section of their report.

“This increase could reflect either higher levels of resistance to trimethoprim or a healthier population treated with nitrofurantoin, the latest being not recommended for patients with poor kidney function,” the researchers wrote.

This study was supported by the National Institute for Health Research and other U.K. sources. One study coauthor reported working as an epidemiologist with GSK in areas not related to the study.

SOURCE: Gharbi M et al. BMJ. 2019 Feb 27. doi: 10.1136/bmj.l525.

Delaying or withholding antibiotics in elderly patients with a urinary tract infection (UTI) increases the risk of sepsis and death, results of a large, population-based study suggest.

andresr/Getty Images

The risk of bloodstream infection was more than seven times greater in patients who did not receive antibiotics immediately after seeing a general practitioner for a UTI versus those who did, according to results of the study based on primary care records and other data for nearly 160,000 U.K. patients aged 65 years or older. Death rates and hospital admissions were significantly higher for these patients, according to the study published in The BMJ by Myriam Gharbi, PharmD, Phd, Imperial College London, and her colleagues.

The publication of these findings coincides with an increase in Escherichia coli bloodstream infections in England.

“Our study suggests the early initiation of antibiotics for UTI in older high risk adult populations (especially men aged [older than] 85 years) should be recommended to prevent serious complications,” Dr. Gharbi and her coauthors said in their report.

The population-based cohort study comprised 157,264 adult primary care patients at least 65 years of age who had one or more suspected or confirmed lower UTIs from November 2007 to May 2015. The researchers found that health care providers had diagnosed a total of 312,896 UTI episodes in these patients during the period they studied. In 7.2% (22,534) of the UTI episodes, the researchers were unable to find records of the patients having been prescribed antibiotics by a general practitioner within 7 days of the UTI diagnosis. These 22,534 episodes included those that occurred in patients who had a complication before an antibiotic was prescribed. An additional 6.2% (19,292) of the episodes occurred in patients who were prescribed antibiotics, but not during their first UTI-related visit to a general practitioner or on the same day of such a visit. The researchers classified this group of patients as having been prescribed antibiotics on a deferred or delayed basis, as they were not prescribed such drugs within 7 days of their visit.

Overall, there were 1,539 cases (0.5% of the total number of UTIs) of bloodstream infection within 60 days of the initial urinary tract infection diagnosis, the researchers reported.

The bloodstream infection rate was 2.9% for patients who were not prescribed antibiotics ever or prior to an infection occurring, 2.2% in those who were prescribed antibiotics on a deferred basis, and 0.2% in those who were prescribed antibiotics immediately, meaning during their first visit to a general practitioner for a UTI or on the same day of such a visit (P less than .001). After adjustment for potential confounding variables such as age, sex, and region, the patients classified as having not been prescribed antibiotics or having been prescribed antibiotics on a deferred basis were significantly more likely to have a bloodstream infection within 60 days of their visit to a health care provider, compared with those who received antibiotics immediately, with odds ratios of 8.08 (95% confidence interval, 7.12-9.16) and 7.12 (95% CI, 6.22-8.14), respectively.

Hospital admissions after a UTI episode were nearly twice as high in the no- or deferred-antibiotics groups (27.0% and 26.8%, respectively), compared with the group that received antibiotics right away (14.8%), the investigators reported. The lengths of hospital stays were 12.1 days for the group classified as having not been prescribed antibiotics, 7.7 days for the group subject to delayed antibiotic prescribing, and 6.3 days for the group who received antibiotics immediately.

Deaths within 60 days of experiencing a urinary tract infection occurred in 5.4% of patients in the no-antibiotics group, 2.8% of the deferred-antibiotics group, and 1.6% of the immediate-antibiotics group. After adjustment for covariates, a regression analysis showed the risks for all-cause mortality were 1.16 and 2.18 times higher in the deferred-antibiotics group and the no-antibiotics group, respectively, according to the paper.

In the immediate-antibiotics group, those patients who received nitrofurantoin had a “small but significant increase” in 60-day survival versus those who received trimethoprim, the investigators noted in the discussion section of their report.

“This increase could reflect either higher levels of resistance to trimethoprim or a healthier population treated with nitrofurantoin, the latest being not recommended for patients with poor kidney function,” the researchers wrote.

This study was supported by the National Institute for Health Research and other U.K. sources. One study coauthor reported working as an epidemiologist with GSK in areas not related to the study.

SOURCE: Gharbi M et al. BMJ. 2019 Feb 27. doi: 10.1136/bmj.l525.