Vascular

Widespread use of antiretroviral therapy has caused a remarkable decline in rates of morbidity and death related to acquired immunodeficiency syndrome (AIDS) and has effectively made human immunodeficiency virus (HIV) infection a manageable—although not yet curable— chronic condition. And as the HIV-infected population on antiretroviral therapy ages, the prevalence of chronic conditions (eg, cardiovascular disease, hepatic disease, pulmonary disease, non-AIDS cancers) and deaths attributable to these conditions have also increased.¹

Many of the traditional risk factors for cardiovascular disease in the general population, including smoking, dyslipidemia, and diabetes, are common in HIV-infected patients, and HIV infection itself independently increases the risk of coronary heart disease. In addition, different antiretroviral combinations can contribute, in varying degrees, to changes in lipid levels and insulin resistance, further increasing coronary risk.

Ultimately, however, the immunologic benefits of antiretroviral therapy for individual patients far exceed the modest increase in cardiovascular risk associated with certain regimens. In most cases, careful selection of the initial antiretroviral regimen and the addition of lipid-lowering or glucose-controlling medications (with close attention to drug interactions) can effectively manage the metabolic changes associated with antiretroviral therapy and obviate any premature modification of virologically suppressive regimens.

TRADITIONAL CARDIAC RISK FACTORS IN HIV PATIENTS

The risk of coronary heart disease in HIV patients is influenced mostly by traditional factors such as age, smoking, diabetes, and dyslipidemia, including high levels of total cholesterol and low-density lipoprotein cholesterol (LDL-C) and low levels of high-density lipoprotein cholesterol (HDL-C).²

In various large cohorts, HIV-infected men had a higher prevalence of smoking,³ a lower mean HDL-C level,^3,4 and a higher mean triglyceride level^3,4 than men without HIV infection, placing them at greater risk of coronary heart disease. However, even after adjusting for traditional risk factors, rates of atherosclerosis are still higher in people who are infected with HIV than in those who are not.⁵

EFFECT OF HIV INFECTION ON CORONARY RISK

HIV infection has been shown to increase coronary risk.

In the Kaiser Permanente database,⁶ HIV-positive patients had a significantly higher rate of hospitalizations for coronary heart disease than did people who were not infected.

Similarly, in a cohort study of almost 4,000 HIV-infected patients and more than 1 million controls, the risk of acute myocardial infarction was 75% higher for HIV-positive patients than for HIV-negative patients, even after adjusting for sex, race, hypertension, diabetes, and dyslipidemia.⁵

The Fat Redistribution and Metabolism (FRAM) cross-sectional study⁷ showed that HIV infection was associated with greater carotid intima media thickness, an established marker of atherosclerosis, independently of traditional risk factors and to virtually the same degree as smoking and male sex.

Other studies of subclinical atherosclerosis in HIV patients have yielded disparate results, likely because of differences in study design, methods of measuring carotid thickness, and characteristics of the study populations (eg, prevalence of cardiovascular risk factors and stage of HIV disease). However, a meta-analysis of six prospective cohort studies, three case-control studies, and four cross-sectional studies confirmed that HIV patients had slightly but statistically significantly greater carotid intima media thickness than HIV-negative people.⁸

MECHANISMS BY WHICH HIV MAY PROMOTE CORONARY HEART DISEASE

The pathogenesis of coronary heart disease in HIV infection has not been fully elucidated, but the virus appears to contribute directly to the accelerated development of atherosclerosis. It may do so through direct effects on cholesterol processing and transport, attraction of monocytes to the intimal wall, and activation of monocytes to induce an inflammatory response and endothelial proliferation.

Effects on lipids

In early HIV infection, levels of total cholesterol and HDL-C are lower. In more advanced infection, lower CD4+ lymphocyte counts have been associated with lower levels of apolipoprotein B and with smaller LDL-C particles, suggesting that HIV affects lipid processing and delivery to vessel walls.⁹ HIV infection is also associated with reduced clearance of LDL-C.¹⁰ HIV appears to specifically inhibit the compensatory efflux of excess cholesterol from macrophages, thus promoting the formation of foam cells in atherosclerotic plaque.¹¹

Attraction of monocytes to the vessel wall

In vitro studies also suggest that HIV enhances migration of monocytes into the vascular intima during atherosclerotic plaque development by promoting secretion of the chemokine monocyte chemoattractant protein 1¹² and the expression of endothelial cell adhesion molecules such as intercellular adhesion molecule 1, vascular cell adhesion molecule 1 (VCAM-1), and E-selectin.¹³

Inflammation

A recent study suggests that chronic inflammation may be a key contributor to the accelerated development of atherosclerosis in HIV patients. Hsue et al¹⁴ compared carotid intima media thickness and levels of C-reactive protein (a marker of systemic inflammation) in HIV-positive and HIV-negative patients. The carotid intima media thickness was greater in all groups of HIV patients, irrespective of level of viremia or exposure to antiretroviral therapy, than in healthy controls. In addition, C-reactive protein levels remained elevated in HIV-infected participants regardless of their level of viremia.

These findings suggest not only that HIV-associated atherosclerosis is determined by advanced immunodeficiency, high-level viremia, and exposure to antiretroviral drugs, but also that persistent inflammation due to HIV infection may play an important role in accelerated atherosclerosis.

 

EFFECT OF ANTIRETROVIRAL THERAPY ON CORONARY RISK

Antiretroviral therapy is associated with a small but significant increase in coronary risk.

Medi-Cal,¹⁵ a retrospective study of 28,513 patients, found antiretroviral therapy to be associated with coronary heart disease among patients 18 to 33 years of age (relative risk 2.06, P < .001).

The Data Collection on Adverse Events of Anti-HIV Drugs study¹⁶ prospectively followed 23,437 patients for 94,469 person-years. Adjusted for exposure to nonnucleoside reverse transcriptase inhibitors and for hypertension and diabetes, the relative risk of myocardial infarction per year of protease inhibitor exposure was 1.16 (95% confidence interval [CI] 1.10–1.23). The relative risk was lower after adjusting for serum lipid levels but remained significant at 1.10 (95% CI 1.04–1.18).

Reports have been mixed regarding a possible association between myocardial infarction and the nucleoside reverse transcriptase inhibitor abacavir (Ziagen): several studies found a statistically significant association,^17–20 and others did not.^21–23 Differences in study design (observational cohort studies vs prospective randomized clinical trials), populations studied (differing in age, cardiovascular risk factor prevalence, and whether the patients had already been exposed to treatment), and outcome definition probably contributed to the different conclusions.

On the other hand, several studies have shown that suppression of HIV with antiretroviral therapy actually improves some of the surrogate markers of cardiovascular disease. For example:

• Markers of endothelial function such as flow-mediated vasodilation improve significantly within 4 weeks of a patient’s starting antiretroviral therapy, regardless of the class of antiretroviral drug used.²⁴
• After viral suppression is achieved, levels of the markers of endothelial activation VCAM-1 and P-selectin decline significantly, as do levels of the adipocyte activation marker leptin and the coagulation marker D-dimer.^25,26
• Levels of the anti-inflammatory markers adiponectin and interleukin 10 increase. ^25,26

Interrupting antiretroviral therapy may increase coronary risk

Not only is uncontrolled viral replication in untreated HIV infection associated with cardiovascular disease, but interrupting antiretroviral therapy may result in a supplementary increase in coronary risk.

In the 5,472-patient Strategies for Management of Antiretroviral Therapy (SMART) trial, the rate of cardiovascular disease events was higher if treatment was interrupted than with continuous treatment, with a hazard ratio of 1.57 (95% CI 1.0–2.46, P = .05).²⁷

This association between treatment interruption and coronary events does not appear to be related to the level of viremia.²⁸ Rather, development of cardiovascular disease in HIV-infected patients who interrupt antiretroviral therapy may be mediated, to a large extent, by chronic inflammation in the setting of viral replication. In the treatment-interruption group, levels of the inflammatory cytokine interleukin 6 (IL-6) and the coagulation marker D-dimer were significantly elevated 1 month after randomization, and these differences were strongly associated with death (odds ratio [OR] 12.6, P < .0001 for IL-6; OR 13.1, P < .0001 for D-dimer). Elevated IL-6 levels were also significantly associated with the development of cardiovascular disease (OR 2.8, P = .03).²⁹

METABOLIC COMPLICATIONS OF ANTIRETROVIRAL THERAPY

Persons with HIV infection may experience metabolic complications that are due to HIV itself or to its treatment.

Cross-sectional studies that included HIV-negative patients as controls have demonstrated changes in lipid processing that are known to promote atherosclerosis. For example, persons with HIV infection have smaller LDL-C particles³⁰ and higher levels of circulating oxidized LDL-C.³¹

In the Multicenter AIDS Cohort Study (MACS), after HIV seroconversion, nonfasting total cholesterol, LDL-C, and HDL-C levels declined, which is consistent with a chronic inflammatory state. After antiretroviral therapy was started, lipid levels returned to baseline levels or slightly higher except for HDL-C, which remained low.⁹ These changes may be due to a general “return to health,” or they may be direct medication effects.

Similar patterns were seen in the SMART study.²⁸ Participants randomized to receive intermittent antiretroviral therapy had overall decreases in all lipid levels, with a marked reduction in HDL-C, while those randomized to receive continuous therapy had increased levels of all lipids, including HDL-C, at 12 months. Overall, the ratio of total cholesterol to HDL-C actually increased for participants on episodic therapy, while it decreased in the continuous-treatment group. Along with continued vascular inflammation, the low HDL-C may have contributed to the worse cardiovascular outcomes in patients who received intermittent antiretroviral therapy.

Some lipid changes associated with antiretroviral therapy may actually be beneficial. For example, nonnucleoside reverse transcriptase inhibitors may raise HDL-C levels. However, such increases alone do not necessarily offset the other lipid changes or translate to an observed improvement in coronary risk.³²

The degree of dyslipidemia and specific lipid changes differ among the different classes of antiretroviral drugs and even among the individual drugs within each class. Furthermore, the magnitude of the observed lipid changes varies widely among patients on the same antiretroviral regimen, reflecting the likely important role of host genomics.

While the protease inhibitors and nonnucleoside reverse transcriptase inhibitors have well-described effects on lipids (described in greater detail in the following sections), there have been no reported significant changes in lipid profiles or cardiovascular risk associated with the newest classes, ie, fusion inhibitors such as enfuvirtide (Fuzeon), CC chemokine receptor type 5 (CCR5) receptor inhibitors such as maraviroc (Selzentry), or integrase inhibitors such as raltegravir (Isentress).

 

 

Impact of protease inhibitors on lipids

Most protease inhibitors raise lipid levels, but the drugs in this class appear to differ in important ways (Table 1).^33–41

Ritonavir (Norvir) and ritonavir-boosted protease inhibitor combinations cause the most significant increases in lipids. Currently, ritonavir is used in low doses to boost the levels of most other protease inhibitors as the standard of care in protease inhibitor-based regimens. However, in most patients, giving ritonavir with protease inhibitors raises lipid levels, particularly triglycerides.

Most boosted protease inhibitor regimens have similar effects on lipid levels, with some exceptions.

Tipranavir (Aptivus) plus ritonavir, for example, markedly raises total cholesterol and triglyceride levels and would not be recommended for patients with dyslipidemia at baseline.³³

Atazanavir (Reyataz)^34,35 plus ritonavir and darunavir (Prezista)³⁶ plus ritonavir cause more modest lipid changes. Unboosted atazanavir raises lipid levels only minimally, if at all,^34,35 but it is no longer a preferred regimen according to US Department of Health and Human Services guidelines.⁴²

Impact of nonnucleoside reverse transcriptase inhibitors on lipids

Nonnucleoside reverse transcriptase inhibitors are also associated with lipid abnormalities, but to a lesser extent than the protease inhibitors (Table 2).^43–45

Efavirenz (Sustiva), a nonnucleoside reverse transcriptase inhibitor, when added to a regimen of two or three nucleoside reverse transcriptase inhibitors, resulted in modest increases in all lipids, including HDL-C (a potentially beneficial change) at 96 weeks compared with a regimen of three nucleoside reverse transcriptase inhibitors only.⁴³

Nevirapine (Viramune), compared with efavirenz, results in a more favorable lipid profile in previously untreated patients, as shown by larger increases in HDL-C and smaller increases in triglycerides at 48 weeks.⁴⁴

Etravirine (Intelence), the newest nonnucleoside reverse transcriptase inhibitor, does not appear to cause any further increase in lipids when added to a regimen containing darunavir-ritonavir and nucleoside agents.⁴⁵

Impact of nucleoside reverse transcriptase inhibitors on lipids

As a class, nucleoside reverse transcriptase inhibitors have been associated with mitochondrial toxicity and insulin resistance,⁴⁶ but the lipid changes associated with them are generally less significant than those caused by protease inhibitors or nonnucleoside reverse transcriptase inhibitors. Nevertheless, within the class, there is considerable variability in lipid changes associated with specific agents.

Stavudine (Zerit), for example, is associated with hypertriglyceridemia.

Tenofovir (Viread), for another example, in combination with emtricitabine (Emtriva) and the nonnucleoside reverse transcriptase inhibitor efavirenz (the three drugs are contained in a formulation called Atripla) was associated with a smaller increase in fasting total cholesterol than with zidovudine-lamivudine and efavirenz at 96 weeks.⁴⁷

A recent placebo-controlled, crossover, pilot study of 17 HIV-infected patients suggested that tenofovir may actually have independent lipid-lowering properties.⁴⁸

Abacavir, as discussed above, has been reported to be associated with a higher risk of myocardial infarction, but this is debatable.

MANAGING CORONARY RISK FACTORS IN HIV-INFECTED PATIENTS

Cardiovascular risk assessment

In HIV patients, cardiovascular risk can be assessed using models derived from large epidemiologic studies such as the Framingham Heart Study.⁴⁹

Current guidelines from the Infectious Diseases Society of America and the AIDS Clinical Trials Group (ACTG) for evaluating and managing dyslipidemia in HIV-infected adults are based on the National Cholesterol Education Program Adult Treatment Panel III.⁵⁰ They recommend obtaining a fasting lipid profile before starting antiretroviral therapy and within 3 to 6 months after starting a new regimen.

The guidelines also recommend stratifying risk by counting the number of cardiovascular risk factors, as is done for the general population. If the patient has more than two factors, the Framingham equation should be used to calculate the 10-year risk of myocardial infarction or cardiac death. Interventions should be offered for modifiable cardiovascular risk factors such as smoking, hypertension, physical inactivity, and diabetes mellitus. LDL-C goals should be determined, and lipid-lowering drugs should be initiated accordingly. If triglyceride levels are 200 to 500 mg/dL and levels of “non-HDL-C” (total cholesterol minus the HDL-C level) are high, a statin is recommended. If the triglyceride level is higher than 500 mg/dL, a fibrate should be started.⁵¹

 

 

Dyslipidemia management

In HIV patients, statin and fibrate therapy must be considered cautiously, given the important drug interactions with protease inhibitors and especially ritonavir. Many statins are metabolized by cytochrome P3A4, which protease inhibitors inhibit.

Statins generally considered safe to use with most protease inhibitors:

• Pravastatin (Pravachol)
• Rosuvastatin (Crestor)
• Atorvastatin (Lipitor).

Exceptions and caveats:

• Pravastatin should not be prescribed with boosted darunavir.
• Data for fluvastatin (Lescol) in HIV-infected patients on antiretroviral therapy are limited.
• Lovastatin (Mevacor) and simvastatin (Zocor) are contraindicated with protease inhibitor therapy.⁵²
• In contrast to the increase in statin levels seen with protease inhibitors, efavirenz lowers levels of simvastatin, pravastatin, and atorvastatin.^53,54

Table 3^50,52–57 summarizes the effects of protease inhibitors and nonnucleoside reverse transcriptase inhibitors on statin levels.

Ezetimibe (Zetia), which is metabolized independently of the cytochrome P450 system, has been shown to be safe and effective when given to HIV-infected patients on antiretroviral therapy.⁵⁸

Fenofibrate (Lofibra) is recommended by current guidelines for patients with elevated triglyceride levels (> 500 mg/dL).⁵¹ In the ACTG 5087 study, a combination of fenofibrate plus pravastatin was found to be safe and effective in improving lipid profiles.⁵⁹

Long-acting niacin resulted in significant improvements in triglycerides, total cholesterol, HDL-C, and LDL-C after 48 weeks of use, although insulin sensitivity worsened.⁶⁰

Fish oil has been shown to be an effective alternative to fibrates, or it can be used in combination with them.⁶¹

Switching antiretroviral agents vs adding lipid-lowering agents. In some patients with significant dyslipidemia, switching antiretro viral agents may lower lipid levels without compromising virologic control.⁶² However, due to the multifactorial nature of dyslipidemia in HIV patients on antiretroviral therapy, switching the HIV therapy alone may not result in sufficient improvement in the lipid profile⁴⁵ and may be associated with virologic failure, particularly among patients who have underlying treatment-resistant HIV.⁶³

In many cases, adding lipid-lowering agents may be more beneficial than switching the antiretroviral therapy. For example, a randomized trial in HIV-infected patients with hyperlipidemia found that adding a lipid-lowering agent such as pravastatin or bezafibrate to the unchanged antiretroviral regimen resulted in greater improvement in total cholesterol, LDL-C, and triglyceride levels than switching from a protease inhibitor to either nevirapine or efavirenz.⁶⁴

Given the complexity of prescribing lipid-lowering therapies to patients on antiretroviral therapy, we recommend that providers check with a pharmacist or refer to package inserts and other medical literature if they are unfamiliar with these drug interactions and responses to lipid-lowering therapies.

Managing insulin resistance

Diabetes mellitus is a well-known risk factor for coronary heart disease. The Data Collection on Adverse Events of Anti-HIV Drugs study found a higher incidence of coronary heart disease in HIV-infected patients, with higher rates in those with longer duration of diabetes.⁶⁵ The prevalence of diabetes in HIV-infected populations varies, depending on demographic characteristics,^65,66 prevalence of coinfection with hepatitis C virus,⁶⁶ and prevalence of exposure to antiretroviral drugs⁶⁷ in the study population.

Drugs that lessen insulin resistance include the thiazolidinedione rosiglitazone (Avandia) and the biguanide metformin (Glucophage). In a randomized trial, both drugs, alone or in combination, improved insulin sensitivity in HIV-infected patients, but neither lessened the amount of visceral or subcutaneous fat.⁶⁸

Smoking cessation

Smoking is another well-known modifiable risk factor for coronary heart disease.

The prevalence of smoking is usually higher in HIV patients than in HIV-negative people. For example, a French cohort study reported smoking prevalence rates of 56.6% in HIV-infected men vs 32.7% in HIV-negative men; in women, the rates were 58% vs 28.1%. The 5-year relative risk of coronary heart disease in HIV-infected vs HIV-negative persons was 1.20 for men and 1.59 for women. The estimated attributable risk due to smoking was 65% for men and 29% for women.³

Therefore, smoking cessation should be a top priority in managing cardiovascular risk in HIV-infected patients. In fact, control of modifiable risk factors through lifestyle changes such as smoking cessation, dietary changes, and exercise is likely to have a significant impact on cardiovascular risk in this population.

Widespread use of antiretroviral therapy has caused a remarkable decline in rates of morbidity and death related to acquired immunodeficiency syndrome (AIDS) and has effectively made human immunodeficiency virus (HIV) infection a manageable—although not yet curable— chronic condition. And as the HIV-infected population on antiretroviral therapy ages, the prevalence of chronic conditions (eg, cardiovascular disease, hepatic disease, pulmonary disease, non-AIDS cancers) and deaths attributable to these conditions have also increased.¹

Many of the traditional risk factors for cardiovascular disease in the general population, including smoking, dyslipidemia, and diabetes, are common in HIV-infected patients, and HIV infection itself independently increases the risk of coronary heart disease. In addition, different antiretroviral combinations can contribute, in varying degrees, to changes in lipid levels and insulin resistance, further increasing coronary risk.

Ultimately, however, the immunologic benefits of antiretroviral therapy for individual patients far exceed the modest increase in cardiovascular risk associated with certain regimens. In most cases, careful selection of the initial antiretroviral regimen and the addition of lipid-lowering or glucose-controlling medications (with close attention to drug interactions) can effectively manage the metabolic changes associated with antiretroviral therapy and obviate any premature modification of virologically suppressive regimens.

TRADITIONAL CARDIAC RISK FACTORS IN HIV PATIENTS

The risk of coronary heart disease in HIV patients is influenced mostly by traditional factors such as age, smoking, diabetes, and dyslipidemia, including high levels of total cholesterol and low-density lipoprotein cholesterol (LDL-C) and low levels of high-density lipoprotein cholesterol (HDL-C).²

In various large cohorts, HIV-infected men had a higher prevalence of smoking,³ a lower mean HDL-C level,^3,4 and a higher mean triglyceride level^3,4 than men without HIV infection, placing them at greater risk of coronary heart disease. However, even after adjusting for traditional risk factors, rates of atherosclerosis are still higher in people who are infected with HIV than in those who are not.⁵

EFFECT OF HIV INFECTION ON CORONARY RISK

HIV infection has been shown to increase coronary risk.

In the Kaiser Permanente database,⁶ HIV-positive patients had a significantly higher rate of hospitalizations for coronary heart disease than did people who were not infected.

Similarly, in a cohort study of almost 4,000 HIV-infected patients and more than 1 million controls, the risk of acute myocardial infarction was 75% higher for HIV-positive patients than for HIV-negative patients, even after adjusting for sex, race, hypertension, diabetes, and dyslipidemia.⁵

The Fat Redistribution and Metabolism (FRAM) cross-sectional study⁷ showed that HIV infection was associated with greater carotid intima media thickness, an established marker of atherosclerosis, independently of traditional risk factors and to virtually the same degree as smoking and male sex.

Other studies of subclinical atherosclerosis in HIV patients have yielded disparate results, likely because of differences in study design, methods of measuring carotid thickness, and characteristics of the study populations (eg, prevalence of cardiovascular risk factors and stage of HIV disease). However, a meta-analysis of six prospective cohort studies, three case-control studies, and four cross-sectional studies confirmed that HIV patients had slightly but statistically significantly greater carotid intima media thickness than HIV-negative people.⁸

MECHANISMS BY WHICH HIV MAY PROMOTE CORONARY HEART DISEASE

The pathogenesis of coronary heart disease in HIV infection has not been fully elucidated, but the virus appears to contribute directly to the accelerated development of atherosclerosis. It may do so through direct effects on cholesterol processing and transport, attraction of monocytes to the intimal wall, and activation of monocytes to induce an inflammatory response and endothelial proliferation.

Effects on lipids

In early HIV infection, levels of total cholesterol and HDL-C are lower. In more advanced infection, lower CD4+ lymphocyte counts have been associated with lower levels of apolipoprotein B and with smaller LDL-C particles, suggesting that HIV affects lipid processing and delivery to vessel walls.⁹ HIV infection is also associated with reduced clearance of LDL-C.¹⁰ HIV appears to specifically inhibit the compensatory efflux of excess cholesterol from macrophages, thus promoting the formation of foam cells in atherosclerotic plaque.¹¹

Attraction of monocytes to the vessel wall

In vitro studies also suggest that HIV enhances migration of monocytes into the vascular intima during atherosclerotic plaque development by promoting secretion of the chemokine monocyte chemoattractant protein 1¹² and the expression of endothelial cell adhesion molecules such as intercellular adhesion molecule 1, vascular cell adhesion molecule 1 (VCAM-1), and E-selectin.¹³

Inflammation

A recent study suggests that chronic inflammation may be a key contributor to the accelerated development of atherosclerosis in HIV patients. Hsue et al¹⁴ compared carotid intima media thickness and levels of C-reactive protein (a marker of systemic inflammation) in HIV-positive and HIV-negative patients. The carotid intima media thickness was greater in all groups of HIV patients, irrespective of level of viremia or exposure to antiretroviral therapy, than in healthy controls. In addition, C-reactive protein levels remained elevated in HIV-infected participants regardless of their level of viremia.

These findings suggest not only that HIV-associated atherosclerosis is determined by advanced immunodeficiency, high-level viremia, and exposure to antiretroviral drugs, but also that persistent inflammation due to HIV infection may play an important role in accelerated atherosclerosis.

 

EFFECT OF ANTIRETROVIRAL THERAPY ON CORONARY RISK

Antiretroviral therapy is associated with a small but significant increase in coronary risk.

Medi-Cal,¹⁵ a retrospective study of 28,513 patients, found antiretroviral therapy to be associated with coronary heart disease among patients 18 to 33 years of age (relative risk 2.06, P < .001).

The Data Collection on Adverse Events of Anti-HIV Drugs study¹⁶ prospectively followed 23,437 patients for 94,469 person-years. Adjusted for exposure to nonnucleoside reverse transcriptase inhibitors and for hypertension and diabetes, the relative risk of myocardial infarction per year of protease inhibitor exposure was 1.16 (95% confidence interval [CI] 1.10–1.23). The relative risk was lower after adjusting for serum lipid levels but remained significant at 1.10 (95% CI 1.04–1.18).

Reports have been mixed regarding a possible association between myocardial infarction and the nucleoside reverse transcriptase inhibitor abacavir (Ziagen): several studies found a statistically significant association,^17–20 and others did not.^21–23 Differences in study design (observational cohort studies vs prospective randomized clinical trials), populations studied (differing in age, cardiovascular risk factor prevalence, and whether the patients had already been exposed to treatment), and outcome definition probably contributed to the different conclusions.

On the other hand, several studies have shown that suppression of HIV with antiretroviral therapy actually improves some of the surrogate markers of cardiovascular disease. For example:

• Markers of endothelial function such as flow-mediated vasodilation improve significantly within 4 weeks of a patient’s starting antiretroviral therapy, regardless of the class of antiretroviral drug used.²⁴
• After viral suppression is achieved, levels of the markers of endothelial activation VCAM-1 and P-selectin decline significantly, as do levels of the adipocyte activation marker leptin and the coagulation marker D-dimer.^25,26
• Levels of the anti-inflammatory markers adiponectin and interleukin 10 increase. ^25,26

Interrupting antiretroviral therapy may increase coronary risk

Not only is uncontrolled viral replication in untreated HIV infection associated with cardiovascular disease, but interrupting antiretroviral therapy may result in a supplementary increase in coronary risk.

In the 5,472-patient Strategies for Management of Antiretroviral Therapy (SMART) trial, the rate of cardiovascular disease events was higher if treatment was interrupted than with continuous treatment, with a hazard ratio of 1.57 (95% CI 1.0–2.46, P = .05).²⁷

This association between treatment interruption and coronary events does not appear to be related to the level of viremia.²⁸ Rather, development of cardiovascular disease in HIV-infected patients who interrupt antiretroviral therapy may be mediated, to a large extent, by chronic inflammation in the setting of viral replication. In the treatment-interruption group, levels of the inflammatory cytokine interleukin 6 (IL-6) and the coagulation marker D-dimer were significantly elevated 1 month after randomization, and these differences were strongly associated with death (odds ratio [OR] 12.6, P < .0001 for IL-6; OR 13.1, P < .0001 for D-dimer). Elevated IL-6 levels were also significantly associated with the development of cardiovascular disease (OR 2.8, P = .03).²⁹

METABOLIC COMPLICATIONS OF ANTIRETROVIRAL THERAPY

Persons with HIV infection may experience metabolic complications that are due to HIV itself or to its treatment.

Cross-sectional studies that included HIV-negative patients as controls have demonstrated changes in lipid processing that are known to promote atherosclerosis. For example, persons with HIV infection have smaller LDL-C particles³⁰ and higher levels of circulating oxidized LDL-C.³¹

In the Multicenter AIDS Cohort Study (MACS), after HIV seroconversion, nonfasting total cholesterol, LDL-C, and HDL-C levels declined, which is consistent with a chronic inflammatory state. After antiretroviral therapy was started, lipid levels returned to baseline levels or slightly higher except for HDL-C, which remained low.⁹ These changes may be due to a general “return to health,” or they may be direct medication effects.

Similar patterns were seen in the SMART study.²⁸ Participants randomized to receive intermittent antiretroviral therapy had overall decreases in all lipid levels, with a marked reduction in HDL-C, while those randomized to receive continuous therapy had increased levels of all lipids, including HDL-C, at 12 months. Overall, the ratio of total cholesterol to HDL-C actually increased for participants on episodic therapy, while it decreased in the continuous-treatment group. Along with continued vascular inflammation, the low HDL-C may have contributed to the worse cardiovascular outcomes in patients who received intermittent antiretroviral therapy.

Some lipid changes associated with antiretroviral therapy may actually be beneficial. For example, nonnucleoside reverse transcriptase inhibitors may raise HDL-C levels. However, such increases alone do not necessarily offset the other lipid changes or translate to an observed improvement in coronary risk.³²

The degree of dyslipidemia and specific lipid changes differ among the different classes of antiretroviral drugs and even among the individual drugs within each class. Furthermore, the magnitude of the observed lipid changes varies widely among patients on the same antiretroviral regimen, reflecting the likely important role of host genomics.

While the protease inhibitors and nonnucleoside reverse transcriptase inhibitors have well-described effects on lipids (described in greater detail in the following sections), there have been no reported significant changes in lipid profiles or cardiovascular risk associated with the newest classes, ie, fusion inhibitors such as enfuvirtide (Fuzeon), CC chemokine receptor type 5 (CCR5) receptor inhibitors such as maraviroc (Selzentry), or integrase inhibitors such as raltegravir (Isentress).

 

 

Impact of protease inhibitors on lipids

Most protease inhibitors raise lipid levels, but the drugs in this class appear to differ in important ways (Table 1).^33–41

Ritonavir (Norvir) and ritonavir-boosted protease inhibitor combinations cause the most significant increases in lipids. Currently, ritonavir is used in low doses to boost the levels of most other protease inhibitors as the standard of care in protease inhibitor-based regimens. However, in most patients, giving ritonavir with protease inhibitors raises lipid levels, particularly triglycerides.

Most boosted protease inhibitor regimens have similar effects on lipid levels, with some exceptions.

Tipranavir (Aptivus) plus ritonavir, for example, markedly raises total cholesterol and triglyceride levels and would not be recommended for patients with dyslipidemia at baseline.³³

Atazanavir (Reyataz)^34,35 plus ritonavir and darunavir (Prezista)³⁶ plus ritonavir cause more modest lipid changes. Unboosted atazanavir raises lipid levels only minimally, if at all,^34,35 but it is no longer a preferred regimen according to US Department of Health and Human Services guidelines.⁴²

Impact of nonnucleoside reverse transcriptase inhibitors on lipids

Nonnucleoside reverse transcriptase inhibitors are also associated with lipid abnormalities, but to a lesser extent than the protease inhibitors (Table 2).^43–45

Efavirenz (Sustiva), a nonnucleoside reverse transcriptase inhibitor, when added to a regimen of two or three nucleoside reverse transcriptase inhibitors, resulted in modest increases in all lipids, including HDL-C (a potentially beneficial change) at 96 weeks compared with a regimen of three nucleoside reverse transcriptase inhibitors only.⁴³

Nevirapine (Viramune), compared with efavirenz, results in a more favorable lipid profile in previously untreated patients, as shown by larger increases in HDL-C and smaller increases in triglycerides at 48 weeks.⁴⁴

Etravirine (Intelence), the newest nonnucleoside reverse transcriptase inhibitor, does not appear to cause any further increase in lipids when added to a regimen containing darunavir-ritonavir and nucleoside agents.⁴⁵

Impact of nucleoside reverse transcriptase inhibitors on lipids

As a class, nucleoside reverse transcriptase inhibitors have been associated with mitochondrial toxicity and insulin resistance,⁴⁶ but the lipid changes associated with them are generally less significant than those caused by protease inhibitors or nonnucleoside reverse transcriptase inhibitors. Nevertheless, within the class, there is considerable variability in lipid changes associated with specific agents.

Stavudine (Zerit), for example, is associated with hypertriglyceridemia.

Tenofovir (Viread), for another example, in combination with emtricitabine (Emtriva) and the nonnucleoside reverse transcriptase inhibitor efavirenz (the three drugs are contained in a formulation called Atripla) was associated with a smaller increase in fasting total cholesterol than with zidovudine-lamivudine and efavirenz at 96 weeks.⁴⁷

A recent placebo-controlled, crossover, pilot study of 17 HIV-infected patients suggested that tenofovir may actually have independent lipid-lowering properties.⁴⁸

Abacavir, as discussed above, has been reported to be associated with a higher risk of myocardial infarction, but this is debatable.

MANAGING CORONARY RISK FACTORS IN HIV-INFECTED PATIENTS

Cardiovascular risk assessment

In HIV patients, cardiovascular risk can be assessed using models derived from large epidemiologic studies such as the Framingham Heart Study.⁴⁹

Current guidelines from the Infectious Diseases Society of America and the AIDS Clinical Trials Group (ACTG) for evaluating and managing dyslipidemia in HIV-infected adults are based on the National Cholesterol Education Program Adult Treatment Panel III.⁵⁰ They recommend obtaining a fasting lipid profile before starting antiretroviral therapy and within 3 to 6 months after starting a new regimen.

The guidelines also recommend stratifying risk by counting the number of cardiovascular risk factors, as is done for the general population. If the patient has more than two factors, the Framingham equation should be used to calculate the 10-year risk of myocardial infarction or cardiac death. Interventions should be offered for modifiable cardiovascular risk factors such as smoking, hypertension, physical inactivity, and diabetes mellitus. LDL-C goals should be determined, and lipid-lowering drugs should be initiated accordingly. If triglyceride levels are 200 to 500 mg/dL and levels of “non-HDL-C” (total cholesterol minus the HDL-C level) are high, a statin is recommended. If the triglyceride level is higher than 500 mg/dL, a fibrate should be started.⁵¹

 

 

Dyslipidemia management

In HIV patients, statin and fibrate therapy must be considered cautiously, given the important drug interactions with protease inhibitors and especially ritonavir. Many statins are metabolized by cytochrome P3A4, which protease inhibitors inhibit.

Statins generally considered safe to use with most protease inhibitors:

• Pravastatin (Pravachol)
• Rosuvastatin (Crestor)
• Atorvastatin (Lipitor).

Exceptions and caveats:

• Pravastatin should not be prescribed with boosted darunavir.
• Data for fluvastatin (Lescol) in HIV-infected patients on antiretroviral therapy are limited.
• Lovastatin (Mevacor) and simvastatin (Zocor) are contraindicated with protease inhibitor therapy.⁵²
• In contrast to the increase in statin levels seen with protease inhibitors, efavirenz lowers levels of simvastatin, pravastatin, and atorvastatin.^53,54

Table 3^50,52–57 summarizes the effects of protease inhibitors and nonnucleoside reverse transcriptase inhibitors on statin levels.

Ezetimibe (Zetia), which is metabolized independently of the cytochrome P450 system, has been shown to be safe and effective when given to HIV-infected patients on antiretroviral therapy.⁵⁸

Fenofibrate (Lofibra) is recommended by current guidelines for patients with elevated triglyceride levels (> 500 mg/dL).⁵¹ In the ACTG 5087 study, a combination of fenofibrate plus pravastatin was found to be safe and effective in improving lipid profiles.⁵⁹

Long-acting niacin resulted in significant improvements in triglycerides, total cholesterol, HDL-C, and LDL-C after 48 weeks of use, although insulin sensitivity worsened.⁶⁰

Fish oil has been shown to be an effective alternative to fibrates, or it can be used in combination with them.⁶¹

Switching antiretroviral agents vs adding lipid-lowering agents. In some patients with significant dyslipidemia, switching antiretro viral agents may lower lipid levels without compromising virologic control.⁶² However, due to the multifactorial nature of dyslipidemia in HIV patients on antiretroviral therapy, switching the HIV therapy alone may not result in sufficient improvement in the lipid profile⁴⁵ and may be associated with virologic failure, particularly among patients who have underlying treatment-resistant HIV.⁶³

In many cases, adding lipid-lowering agents may be more beneficial than switching the antiretroviral therapy. For example, a randomized trial in HIV-infected patients with hyperlipidemia found that adding a lipid-lowering agent such as pravastatin or bezafibrate to the unchanged antiretroviral regimen resulted in greater improvement in total cholesterol, LDL-C, and triglyceride levels than switching from a protease inhibitor to either nevirapine or efavirenz.⁶⁴

Given the complexity of prescribing lipid-lowering therapies to patients on antiretroviral therapy, we recommend that providers check with a pharmacist or refer to package inserts and other medical literature if they are unfamiliar with these drug interactions and responses to lipid-lowering therapies.

Managing insulin resistance

Diabetes mellitus is a well-known risk factor for coronary heart disease. The Data Collection on Adverse Events of Anti-HIV Drugs study found a higher incidence of coronary heart disease in HIV-infected patients, with higher rates in those with longer duration of diabetes.⁶⁵ The prevalence of diabetes in HIV-infected populations varies, depending on demographic characteristics,^65,66 prevalence of coinfection with hepatitis C virus,⁶⁶ and prevalence of exposure to antiretroviral drugs⁶⁷ in the study population.

Drugs that lessen insulin resistance include the thiazolidinedione rosiglitazone (Avandia) and the biguanide metformin (Glucophage). In a randomized trial, both drugs, alone or in combination, improved insulin sensitivity in HIV-infected patients, but neither lessened the amount of visceral or subcutaneous fat.⁶⁸

Smoking cessation

Smoking is another well-known modifiable risk factor for coronary heart disease.

The prevalence of smoking is usually higher in HIV patients than in HIV-negative people. For example, a French cohort study reported smoking prevalence rates of 56.6% in HIV-infected men vs 32.7% in HIV-negative men; in women, the rates were 58% vs 28.1%. The 5-year relative risk of coronary heart disease in HIV-infected vs HIV-negative persons was 1.20 for men and 1.59 for women. The estimated attributable risk due to smoking was 65% for men and 29% for women.³

Therefore, smoking cessation should be a top priority in managing cardiovascular risk in HIV-infected patients. In fact, control of modifiable risk factors through lifestyle changes such as smoking cessation, dietary changes, and exercise is likely to have a significant impact on cardiovascular risk in this population.

Widespread use of antiretroviral therapy has caused a remarkable decline in rates of morbidity and death related to acquired immunodeficiency syndrome (AIDS) and has effectively made human immunodeficiency virus (HIV) infection a manageable—although not yet curable— chronic condition. And as the HIV-infected population on antiretroviral therapy ages, the prevalence of chronic conditions (eg, cardiovascular disease, hepatic disease, pulmonary disease, non-AIDS cancers) and deaths attributable to these conditions have also increased.¹

Many of the traditional risk factors for cardiovascular disease in the general population, including smoking, dyslipidemia, and diabetes, are common in HIV-infected patients, and HIV infection itself independently increases the risk of coronary heart disease. In addition, different antiretroviral combinations can contribute, in varying degrees, to changes in lipid levels and insulin resistance, further increasing coronary risk.

Ultimately, however, the immunologic benefits of antiretroviral therapy for individual patients far exceed the modest increase in cardiovascular risk associated with certain regimens. In most cases, careful selection of the initial antiretroviral regimen and the addition of lipid-lowering or glucose-controlling medications (with close attention to drug interactions) can effectively manage the metabolic changes associated with antiretroviral therapy and obviate any premature modification of virologically suppressive regimens.

TRADITIONAL CARDIAC RISK FACTORS IN HIV PATIENTS

The risk of coronary heart disease in HIV patients is influenced mostly by traditional factors such as age, smoking, diabetes, and dyslipidemia, including high levels of total cholesterol and low-density lipoprotein cholesterol (LDL-C) and low levels of high-density lipoprotein cholesterol (HDL-C).²

In various large cohorts, HIV-infected men had a higher prevalence of smoking,³ a lower mean HDL-C level,^3,4 and a higher mean triglyceride level^3,4 than men without HIV infection, placing them at greater risk of coronary heart disease. However, even after adjusting for traditional risk factors, rates of atherosclerosis are still higher in people who are infected with HIV than in those who are not.⁵

EFFECT OF HIV INFECTION ON CORONARY RISK

HIV infection has been shown to increase coronary risk.

In the Kaiser Permanente database,⁶ HIV-positive patients had a significantly higher rate of hospitalizations for coronary heart disease than did people who were not infected.

Similarly, in a cohort study of almost 4,000 HIV-infected patients and more than 1 million controls, the risk of acute myocardial infarction was 75% higher for HIV-positive patients than for HIV-negative patients, even after adjusting for sex, race, hypertension, diabetes, and dyslipidemia.⁵

The Fat Redistribution and Metabolism (FRAM) cross-sectional study⁷ showed that HIV infection was associated with greater carotid intima media thickness, an established marker of atherosclerosis, independently of traditional risk factors and to virtually the same degree as smoking and male sex.

Other studies of subclinical atherosclerosis in HIV patients have yielded disparate results, likely because of differences in study design, methods of measuring carotid thickness, and characteristics of the study populations (eg, prevalence of cardiovascular risk factors and stage of HIV disease). However, a meta-analysis of six prospective cohort studies, three case-control studies, and four cross-sectional studies confirmed that HIV patients had slightly but statistically significantly greater carotid intima media thickness than HIV-negative people.⁸

MECHANISMS BY WHICH HIV MAY PROMOTE CORONARY HEART DISEASE

The pathogenesis of coronary heart disease in HIV infection has not been fully elucidated, but the virus appears to contribute directly to the accelerated development of atherosclerosis. It may do so through direct effects on cholesterol processing and transport, attraction of monocytes to the intimal wall, and activation of monocytes to induce an inflammatory response and endothelial proliferation.

Effects on lipids

In early HIV infection, levels of total cholesterol and HDL-C are lower. In more advanced infection, lower CD4+ lymphocyte counts have been associated with lower levels of apolipoprotein B and with smaller LDL-C particles, suggesting that HIV affects lipid processing and delivery to vessel walls.⁹ HIV infection is also associated with reduced clearance of LDL-C.¹⁰ HIV appears to specifically inhibit the compensatory efflux of excess cholesterol from macrophages, thus promoting the formation of foam cells in atherosclerotic plaque.¹¹

Attraction of monocytes to the vessel wall

In vitro studies also suggest that HIV enhances migration of monocytes into the vascular intima during atherosclerotic plaque development by promoting secretion of the chemokine monocyte chemoattractant protein 1¹² and the expression of endothelial cell adhesion molecules such as intercellular adhesion molecule 1, vascular cell adhesion molecule 1 (VCAM-1), and E-selectin.¹³

Inflammation

A recent study suggests that chronic inflammation may be a key contributor to the accelerated development of atherosclerosis in HIV patients. Hsue et al¹⁴ compared carotid intima media thickness and levels of C-reactive protein (a marker of systemic inflammation) in HIV-positive and HIV-negative patients. The carotid intima media thickness was greater in all groups of HIV patients, irrespective of level of viremia or exposure to antiretroviral therapy, than in healthy controls. In addition, C-reactive protein levels remained elevated in HIV-infected participants regardless of their level of viremia.

These findings suggest not only that HIV-associated atherosclerosis is determined by advanced immunodeficiency, high-level viremia, and exposure to antiretroviral drugs, but also that persistent inflammation due to HIV infection may play an important role in accelerated atherosclerosis.

 

EFFECT OF ANTIRETROVIRAL THERAPY ON CORONARY RISK

Antiretroviral therapy is associated with a small but significant increase in coronary risk.

Medi-Cal,¹⁵ a retrospective study of 28,513 patients, found antiretroviral therapy to be associated with coronary heart disease among patients 18 to 33 years of age (relative risk 2.06, P < .001).

The Data Collection on Adverse Events of Anti-HIV Drugs study¹⁶ prospectively followed 23,437 patients for 94,469 person-years. Adjusted for exposure to nonnucleoside reverse transcriptase inhibitors and for hypertension and diabetes, the relative risk of myocardial infarction per year of protease inhibitor exposure was 1.16 (95% confidence interval [CI] 1.10–1.23). The relative risk was lower after adjusting for serum lipid levels but remained significant at 1.10 (95% CI 1.04–1.18).

Reports have been mixed regarding a possible association between myocardial infarction and the nucleoside reverse transcriptase inhibitor abacavir (Ziagen): several studies found a statistically significant association,^17–20 and others did not.^21–23 Differences in study design (observational cohort studies vs prospective randomized clinical trials), populations studied (differing in age, cardiovascular risk factor prevalence, and whether the patients had already been exposed to treatment), and outcome definition probably contributed to the different conclusions.

On the other hand, several studies have shown that suppression of HIV with antiretroviral therapy actually improves some of the surrogate markers of cardiovascular disease. For example:

• Markers of endothelial function such as flow-mediated vasodilation improve significantly within 4 weeks of a patient’s starting antiretroviral therapy, regardless of the class of antiretroviral drug used.²⁴
• After viral suppression is achieved, levels of the markers of endothelial activation VCAM-1 and P-selectin decline significantly, as do levels of the adipocyte activation marker leptin and the coagulation marker D-dimer.^25,26
• Levels of the anti-inflammatory markers adiponectin and interleukin 10 increase. ^25,26

Interrupting antiretroviral therapy may increase coronary risk

Not only is uncontrolled viral replication in untreated HIV infection associated with cardiovascular disease, but interrupting antiretroviral therapy may result in a supplementary increase in coronary risk.

In the 5,472-patient Strategies for Management of Antiretroviral Therapy (SMART) trial, the rate of cardiovascular disease events was higher if treatment was interrupted than with continuous treatment, with a hazard ratio of 1.57 (95% CI 1.0–2.46, P = .05).²⁷

This association between treatment interruption and coronary events does not appear to be related to the level of viremia.²⁸ Rather, development of cardiovascular disease in HIV-infected patients who interrupt antiretroviral therapy may be mediated, to a large extent, by chronic inflammation in the setting of viral replication. In the treatment-interruption group, levels of the inflammatory cytokine interleukin 6 (IL-6) and the coagulation marker D-dimer were significantly elevated 1 month after randomization, and these differences were strongly associated with death (odds ratio [OR] 12.6, P < .0001 for IL-6; OR 13.1, P < .0001 for D-dimer). Elevated IL-6 levels were also significantly associated with the development of cardiovascular disease (OR 2.8, P = .03).²⁹

METABOLIC COMPLICATIONS OF ANTIRETROVIRAL THERAPY

Persons with HIV infection may experience metabolic complications that are due to HIV itself or to its treatment.

Cross-sectional studies that included HIV-negative patients as controls have demonstrated changes in lipid processing that are known to promote atherosclerosis. For example, persons with HIV infection have smaller LDL-C particles³⁰ and higher levels of circulating oxidized LDL-C.³¹

In the Multicenter AIDS Cohort Study (MACS), after HIV seroconversion, nonfasting total cholesterol, LDL-C, and HDL-C levels declined, which is consistent with a chronic inflammatory state. After antiretroviral therapy was started, lipid levels returned to baseline levels or slightly higher except for HDL-C, which remained low.⁹ These changes may be due to a general “return to health,” or they may be direct medication effects.

Similar patterns were seen in the SMART study.²⁸ Participants randomized to receive intermittent antiretroviral therapy had overall decreases in all lipid levels, with a marked reduction in HDL-C, while those randomized to receive continuous therapy had increased levels of all lipids, including HDL-C, at 12 months. Overall, the ratio of total cholesterol to HDL-C actually increased for participants on episodic therapy, while it decreased in the continuous-treatment group. Along with continued vascular inflammation, the low HDL-C may have contributed to the worse cardiovascular outcomes in patients who received intermittent antiretroviral therapy.

Some lipid changes associated with antiretroviral therapy may actually be beneficial. For example, nonnucleoside reverse transcriptase inhibitors may raise HDL-C levels. However, such increases alone do not necessarily offset the other lipid changes or translate to an observed improvement in coronary risk.³²

The degree of dyslipidemia and specific lipid changes differ among the different classes of antiretroviral drugs and even among the individual drugs within each class. Furthermore, the magnitude of the observed lipid changes varies widely among patients on the same antiretroviral regimen, reflecting the likely important role of host genomics.

While the protease inhibitors and nonnucleoside reverse transcriptase inhibitors have well-described effects on lipids (described in greater detail in the following sections), there have been no reported significant changes in lipid profiles or cardiovascular risk associated with the newest classes, ie, fusion inhibitors such as enfuvirtide (Fuzeon), CC chemokine receptor type 5 (CCR5) receptor inhibitors such as maraviroc (Selzentry), or integrase inhibitors such as raltegravir (Isentress).

 

 

Impact of protease inhibitors on lipids

Most protease inhibitors raise lipid levels, but the drugs in this class appear to differ in important ways (Table 1).^33–41

Ritonavir (Norvir) and ritonavir-boosted protease inhibitor combinations cause the most significant increases in lipids. Currently, ritonavir is used in low doses to boost the levels of most other protease inhibitors as the standard of care in protease inhibitor-based regimens. However, in most patients, giving ritonavir with protease inhibitors raises lipid levels, particularly triglycerides.

Most boosted protease inhibitor regimens have similar effects on lipid levels, with some exceptions.

Tipranavir (Aptivus) plus ritonavir, for example, markedly raises total cholesterol and triglyceride levels and would not be recommended for patients with dyslipidemia at baseline.³³

Atazanavir (Reyataz)^34,35 plus ritonavir and darunavir (Prezista)³⁶ plus ritonavir cause more modest lipid changes. Unboosted atazanavir raises lipid levels only minimally, if at all,^34,35 but it is no longer a preferred regimen according to US Department of Health and Human Services guidelines.⁴²

Impact of nonnucleoside reverse transcriptase inhibitors on lipids

Nonnucleoside reverse transcriptase inhibitors are also associated with lipid abnormalities, but to a lesser extent than the protease inhibitors (Table 2).^43–45

Efavirenz (Sustiva), a nonnucleoside reverse transcriptase inhibitor, when added to a regimen of two or three nucleoside reverse transcriptase inhibitors, resulted in modest increases in all lipids, including HDL-C (a potentially beneficial change) at 96 weeks compared with a regimen of three nucleoside reverse transcriptase inhibitors only.⁴³

Nevirapine (Viramune), compared with efavirenz, results in a more favorable lipid profile in previously untreated patients, as shown by larger increases in HDL-C and smaller increases in triglycerides at 48 weeks.⁴⁴

Etravirine (Intelence), the newest nonnucleoside reverse transcriptase inhibitor, does not appear to cause any further increase in lipids when added to a regimen containing darunavir-ritonavir and nucleoside agents.⁴⁵

Impact of nucleoside reverse transcriptase inhibitors on lipids

As a class, nucleoside reverse transcriptase inhibitors have been associated with mitochondrial toxicity and insulin resistance,⁴⁶ but the lipid changes associated with them are generally less significant than those caused by protease inhibitors or nonnucleoside reverse transcriptase inhibitors. Nevertheless, within the class, there is considerable variability in lipid changes associated with specific agents.

Stavudine (Zerit), for example, is associated with hypertriglyceridemia.

Tenofovir (Viread), for another example, in combination with emtricitabine (Emtriva) and the nonnucleoside reverse transcriptase inhibitor efavirenz (the three drugs are contained in a formulation called Atripla) was associated with a smaller increase in fasting total cholesterol than with zidovudine-lamivudine and efavirenz at 96 weeks.⁴⁷

A recent placebo-controlled, crossover, pilot study of 17 HIV-infected patients suggested that tenofovir may actually have independent lipid-lowering properties.⁴⁸

Abacavir, as discussed above, has been reported to be associated with a higher risk of myocardial infarction, but this is debatable.

MANAGING CORONARY RISK FACTORS IN HIV-INFECTED PATIENTS

Cardiovascular risk assessment

In HIV patients, cardiovascular risk can be assessed using models derived from large epidemiologic studies such as the Framingham Heart Study.⁴⁹

Current guidelines from the Infectious Diseases Society of America and the AIDS Clinical Trials Group (ACTG) for evaluating and managing dyslipidemia in HIV-infected adults are based on the National Cholesterol Education Program Adult Treatment Panel III.⁵⁰ They recommend obtaining a fasting lipid profile before starting antiretroviral therapy and within 3 to 6 months after starting a new regimen.

The guidelines also recommend stratifying risk by counting the number of cardiovascular risk factors, as is done for the general population. If the patient has more than two factors, the Framingham equation should be used to calculate the 10-year risk of myocardial infarction or cardiac death. Interventions should be offered for modifiable cardiovascular risk factors such as smoking, hypertension, physical inactivity, and diabetes mellitus. LDL-C goals should be determined, and lipid-lowering drugs should be initiated accordingly. If triglyceride levels are 200 to 500 mg/dL and levels of “non-HDL-C” (total cholesterol minus the HDL-C level) are high, a statin is recommended. If the triglyceride level is higher than 500 mg/dL, a fibrate should be started.⁵¹

 

 

Dyslipidemia management

In HIV patients, statin and fibrate therapy must be considered cautiously, given the important drug interactions with protease inhibitors and especially ritonavir. Many statins are metabolized by cytochrome P3A4, which protease inhibitors inhibit.

Statins generally considered safe to use with most protease inhibitors:

• Pravastatin (Pravachol)
• Rosuvastatin (Crestor)
• Atorvastatin (Lipitor).

Exceptions and caveats:

• Pravastatin should not be prescribed with boosted darunavir.
• Data for fluvastatin (Lescol) in HIV-infected patients on antiretroviral therapy are limited.
• Lovastatin (Mevacor) and simvastatin (Zocor) are contraindicated with protease inhibitor therapy.⁵²
• In contrast to the increase in statin levels seen with protease inhibitors, efavirenz lowers levels of simvastatin, pravastatin, and atorvastatin.^53,54

Table 3^50,52–57 summarizes the effects of protease inhibitors and nonnucleoside reverse transcriptase inhibitors on statin levels.

Ezetimibe (Zetia), which is metabolized independently of the cytochrome P450 system, has been shown to be safe and effective when given to HIV-infected patients on antiretroviral therapy.⁵⁸

Fenofibrate (Lofibra) is recommended by current guidelines for patients with elevated triglyceride levels (> 500 mg/dL).⁵¹ In the ACTG 5087 study, a combination of fenofibrate plus pravastatin was found to be safe and effective in improving lipid profiles.⁵⁹

Long-acting niacin resulted in significant improvements in triglycerides, total cholesterol, HDL-C, and LDL-C after 48 weeks of use, although insulin sensitivity worsened.⁶⁰

Fish oil has been shown to be an effective alternative to fibrates, or it can be used in combination with them.⁶¹

Switching antiretroviral agents vs adding lipid-lowering agents. In some patients with significant dyslipidemia, switching antiretro viral agents may lower lipid levels without compromising virologic control.⁶² However, due to the multifactorial nature of dyslipidemia in HIV patients on antiretroviral therapy, switching the HIV therapy alone may not result in sufficient improvement in the lipid profile⁴⁵ and may be associated with virologic failure, particularly among patients who have underlying treatment-resistant HIV.⁶³

In many cases, adding lipid-lowering agents may be more beneficial than switching the antiretroviral therapy. For example, a randomized trial in HIV-infected patients with hyperlipidemia found that adding a lipid-lowering agent such as pravastatin or bezafibrate to the unchanged antiretroviral regimen resulted in greater improvement in total cholesterol, LDL-C, and triglyceride levels than switching from a protease inhibitor to either nevirapine or efavirenz.⁶⁴

Given the complexity of prescribing lipid-lowering therapies to patients on antiretroviral therapy, we recommend that providers check with a pharmacist or refer to package inserts and other medical literature if they are unfamiliar with these drug interactions and responses to lipid-lowering therapies.

Managing insulin resistance

Diabetes mellitus is a well-known risk factor for coronary heart disease. The Data Collection on Adverse Events of Anti-HIV Drugs study found a higher incidence of coronary heart disease in HIV-infected patients, with higher rates in those with longer duration of diabetes.⁶⁵ The prevalence of diabetes in HIV-infected populations varies, depending on demographic characteristics,^65,66 prevalence of coinfection with hepatitis C virus,⁶⁶ and prevalence of exposure to antiretroviral drugs⁶⁷ in the study population.

Drugs that lessen insulin resistance include the thiazolidinedione rosiglitazone (Avandia) and the biguanide metformin (Glucophage). In a randomized trial, both drugs, alone or in combination, improved insulin sensitivity in HIV-infected patients, but neither lessened the amount of visceral or subcutaneous fat.⁶⁸

Smoking cessation

Smoking is another well-known modifiable risk factor for coronary heart disease.

The prevalence of smoking is usually higher in HIV patients than in HIV-negative people. For example, a French cohort study reported smoking prevalence rates of 56.6% in HIV-infected men vs 32.7% in HIV-negative men; in women, the rates were 58% vs 28.1%. The 5-year relative risk of coronary heart disease in HIV-infected vs HIV-negative persons was 1.20 for men and 1.59 for women. The estimated attributable risk due to smoking was 65% for men and 29% for women.³

Therefore, smoking cessation should be a top priority in managing cardiovascular risk in HIV-infected patients. In fact, control of modifiable risk factors through lifestyle changes such as smoking cessation, dietary changes, and exercise is likely to have a significant impact on cardiovascular risk in this population.

Coenzyme Q10 supplements have been purported to be effective for treating a variety of disorders,^1,2 in particular hypertension and statin-induced myalgia.

Several studies^3–7 found that coenzyme Q10 supplementation significantly lowered blood pressure in hypertensive patients. Moreover, some trials have demonstrated that statin therapy reduces serum or muscle levels of coenzyme Q10,^8–14 prompting investigations to determine whether coenzyme Q10 deficiency is related to statin-induced muscle pain.^15–17

In this review, we discuss the efficacy and safety of coenzyme Q10 supplementation in patients with hypertension and those taking statins, and some of the caveats about using supplements that are not approved by the US Food and Drug Administration (FDA), as well as the bioavailability and quality of available formulations.

WHAT IS COENZYME Q10?

Coenzyme Q10, also known as coenzyme Q, ubidecarenone, and ubiquinone, is found in all human cells, with the highest concentrations in the heart, liver, kidney, and pancreas.^1,2 It is a potent antioxidant, a membrane stabilizer, and an integral cofactor in the mitochondrial respiratory chain, helping to generate adenosine triphosphate, the major cellular energy source.^1,2,18 It may also regulate genes associated with cell metabolism.¹⁹

RATIONALE FOR SUPPLEMENTATION

Coenzyme Q10 supplementation has been used, recommended, or studied in heart failure, hypertension, parkinsonism, mitochondrial encephalomyopathies, and other ailments.

In hypertension

Depending on the class, various antihypertensive drugs can have adverse effects such as depression, cough, and cardiac and renal dysfunction. ^20,21 Furthermore, many patients need to take more than one drug to control their blood pressure, increasing their risk of side effects. Some researchers believe coenzyme Q10 supplementation may reduce the need to take multiple antihypertensive drugs.⁵

Coenzyme Q10 appears to lower blood pressure. The exact mechanism is not known, but one theory is that it reduces peripheral resistance by preserving nitric oxide.²¹ Nitric oxide relaxes peripheral arteries, lowering blood pressure. In some forms of hypertension, superoxide radicals that inactivate nitric oxide are overproduced; coenzyme Q10, with its antioxidant effects, may prevent the inactivation of nitric oxide by these free radicals. Alternatively, coenzyme Q10 may boost the production of the prostaglandin prostacyclin (PGI₂) a potent vasodilator and inhibitor of platelet aggregation, or it may enhance the sensitivity of arterial smooth muscle to PGI₂, or both.^1,22

In patients taking statins

Hydroxymethylglutaryl coenzyme A reductase inhibitors (statins), first-line agents for lowering cholesterol levels to prevent cardiovascular disease, are some of the most commonly prescribed medications.^23,24 However, statin therapy carries a risk of myopathy, which can range from muscle aches to rhabdomyolysis. ^23,24

In a clinical advisory,²⁵ the American College of Cardiology, the American Heart Association, and the National Heart, Lung, and Blood Institute recommend that patients on statin therapy who experience muscle soreness, tenderness, or pain with serum creatine kinase levels 3 to 10 times the upper limit of normal should have their creatine kinase level checked weekly. If the level is 3 to 10 times the upper limit of normal, statin therapy may be continued, but if it exceeds 10 times the upper limit, then statins and other potential offending agents (eg, niacin, fibrate) need to be discontinued.

Statins inhibit the synthesis of cholesterol by reducing the production of mevalonate, a precursor of both cholesterol and coenzyme Q10. Since both cholesterol and coenzyme Q10 are produced by the same pathway, it is not surprising that statins have been reported to reduce serum and muscle coenzyme Q10 levels.^9–14 However, one study did not report a significant reduction of coenzyme Q10 levels in muscle tissue in patients treated with simvastatin 20 mg for 6 months.²⁶

Nonetheless, researchers have hypothesized that a reduction in coenzyme Q10 levels in muscle tissue causes mitochondrial dysfunction, which could increase the risk of statininduced myopathy,^13–17 and some believe that treatment with coenzyme Q10 may reduce myalgic symptoms and allow patients to remain on statin therapy.^13,24

Researchers have investigated the potential of coenzyme Q10 supplementation to reduce or prevent statin-induced myopathy.^15–17 (More on this below.)

Interestingly, a randomized, placebo-controlled trial²⁷ found that 6 months of daily therapy with simvastatin (Zocor) 20 mg or pravastatin (Pravachol) 40 mg lowered systolic and diastolic blood pressure significantly in patients with no documented history of cardiovascular disease or diabetes. A possible mechanism of statin-induced blood pressure reduction is the up-regulation of endothelial nitric oxide synthetase, a potent vasodilator. Coenzyme Q10 levels were not assessed during this study. Whether coenzyme Q10 supplementation used to treat statin-induced myalgia enhances or inhibits the antihypertensive effects of statins is not yet known.

 

EVIDENCE OF EFFECTIVENESS IN HYPERTENSION

A number of trials provide clinical evidence that some patients with high blood pressure may benefit from coenzyme Q10 supplementation (Table 1).^{3–7,28–31}

Rosenfeldt et al²⁸ performed a meta-analysis and found that some trials documented statistically significant reductions in diastolic or systolic blood pressure or both, while others reported negligible effects.^3,29 In one small trial,³⁰ blood pressures actually went up in patients taking coenzyme Q10. Coenzyme Q10 dosages and length of therapy varied from study to study in the meta-analysis. Only minor adverse effects such as gastrointestinal upset and headache were reported.

Yamagami et al³ randomly assigned 20 patients with hypertension and a low coenzyme Q10 level to receive 100 mg of coenzyme Q10 or placebo daily for 12 weeks. Patients continued their usual antihypertensive regimen during the study period. Blood pressures, coenzyme Q10 levels, and antihypertensive drugs used were comparable between the study groups.

After 12 weeks of therapy, the mean coenzyme Q10 level in the active-treatment group had more than doubled, from 0.704 to 1.597 μg/mL. This group also experienced a statistically significant drop in systolic blood pressure, from 167 mm Hg at baseline to 148 mm Hg at 12 weeks. In the placebo group, the systolic blood pressure was 168 mm Hg at baseline and 164 mm Hg at 12 weeks; the change was not statistically significant. Diastolic pressure was not significantly lower at 12 weeks than at baseline in either group.

The authors concluded that coenzyme Q10 supplementation brought a mild reduction in high blood pressure in patients who had low coenzyme Q10 serum levels.

Digiesi et al³¹ randomized 18 patients with essential hypertension to receive either coenzyme Q10 100 mg or placebo daily for 10 weeks. All antihypertensive therapy was discontinued at baseline. After the first 10 weeks, patients went through a 2-week washout period and then were switched to the opposite therapy for an additional 10 weeks. Mean baseline blood pressure values were 167 mm Hg systolic and 103 mm Hg diastolic.

Those taking the supplement had a statistically significant decrease in systolic and diastolic pressures (P < .001). The antihypertensive effect was noted in the 3rd or 4th week of active treatment and persisted for the duration of therapy. The effects dissipated 7 to 10 days after coenzyme Q10 was stopped.

Langsjoen et al⁵ evaluated the effects of adding coenzyme Q10 to the antihypertensive drug regimen of 109 patients who had a primary diagnosis of essential hypertension in a prospective observational study. Patients with hypertension as a secondary diagnosis and other cardiovascular diseases were excluded. Variable doses of coenzyme Q10 were given, adjusted according to clinical response and to achieve serum levels greater than 2.0 μg/mL. The average dose was 225 mg/day; the mean serum level attained was 3.02 μg/mL.

Over several months, patients taking the supplement had a reduction in mean systolic pressure from 159 mm Hg at baseline to 147 mm Hg (P < .001), and a reduction in mean diastolic pressure from 94 to 85 mm Hg (P < .001). Thirty-seven percent of patients were able to discontinue one antihypertensive drug, 11% discontinued two drugs, and 4% were able to stop taking three drugs. However, 46% remained on the same antihypertensive regimen, and 3% needed an additional drug.

Singh et al⁶ randomized 64 patients who had coronary artery disease and who had been on antihypertensive drugs for more than 1 year to receive either B-complex vitamins or coenzyme Q10 (hydrosoluble Q-Gel) 60 mg orally once daily for 8 weeks. Five patients were not available for follow-up; therefore, only 59 patients were evaluated. Fifty-five (93%) of the 59 patients were taking only one antihypertensive drug. Initial antihypertensive drug use was similar between study groups and was continued throughout the trial.

After 8 weeks of therapy, the coenzyme Q10 group had significantly lower systolic and diastolic blood pressure than the placebo group (P < .05 for both). There was also a statistically significant decrease in the dosage of antihypertensive drugs in the coenzyme Q10 group but not in the placebo group (P < .05), reflecting coenzyme Q10’s additive antihypertensive effect.

Burke et al⁷ randomized 41 men and 35 women with isolated systolic hypertension (systolic pressure 150–170 mm Hg, diastolic pressure < 90 mm Hg) to receive a twice-daily dose of 60 mg of emulsified coenzyme Q10 (hydrosoluble Q-Gel) with 150 IU of vitamin E or placebo containing vitamin E alone for 12 weeks. The study also included 5 men and 4 women with normal blood pressure, all of whom received coenzyme Q10. A total of 80 patients completed treatment. The primary goal of the study was to determine the efficacy of coenzyme Q10 in the treatment of isolated systolic hypertension in patients without comorbid conditions. Blood pressures were monitored twice a week during the trial, by the same nurse.

After 12 weeks of treatment, the mean reduction in systolic pressure in hypertensive patients on coenzyme Q10 was 17.8 ± 7.3 mm Hg. There were no significant changes in diastolic pressure in any study group with treatment. Patients with isolated systolic hypertension who were taking coenzyme Q10 had a statistically significant reduction in systolic pressure compared with baseline and placebo (P < .01 for both). Approximately 55% of patients on coenzyme Q10 achieved a reduction in systolic pressure of 4 mm Hg or greater, while 45% did not respond to therapy. The mean plasma coenzyme Q10 level of the treatment group increased from 0.47 ± 0.19 μg/mL to 2.69 ± 0.54 μg/mL after 12 weeks; however, the study did not have the statistical power to demonstrate a relationship between coenzyme Q10 levels and changes in blood pressure. Twenty-seven (34%) of the 80 patients were taking a statin while on coenzyme Q10 therapy.

 

STUDIES IN STATIN-INDUCED MYOPATHY

Thibault et al³² and Kim et al³³ reported that patients taking lovastatin (Mevacor) at dosages as high as 35 mg/kg/day to inhibit tumor growth achieved symptomatic relief of statin-induced musculoskeletal toxicity after coenzyme Q10 supplementation.

Caso et al¹⁵ performed a small pilot study in 32 patients to determine if coenzyme Q10 supplementation would improve myalgic symptoms in patients treated with statins. In this double-blind, randomized trial, patients received either coenzyme Q10 100 mg/day or vitamin E 400 IU/day for 30 days. The extent of muscle pain and its interference with daily activities were determined before and after therapy using the Brief Pain Inventory Questionnaire. The statins were atorvastatin (Lipitor) 10 mg or 20 mg, lovastatin 40 mg, pravastatin 40 mg, and simvastatin 10, 20, 40, and 80 mg. Five patients in the coenzyme Q10 group and four patients in the vitamin E group were taking nonsteroidal anti-inflammatory drugs before and during the trial. The intensity of muscle pain and its interference with daily activities were similar between study groups before the start of therapy.

After 30 days of treatment with coenzyme Q10, the pain intensity had decreased significantly from baseline (P < .001). In contrast, no change in pain intensity from baseline was noted in patients receiving vitamin E. The Pain Severity Score was significantly different between study groups, favoring the coenzyme Q10 group (P < .001). Sixteen of 18 patients on coenzyme Q10 reported a reduction in pain, while only 3 of 14 patients on vitamin E reported a similar response. Also, the interference of pain with daily activities significantly improved with coenzyme Q10 (P < .02), whereas vitamin E did not have a significant impact on this.

Young et al¹⁷ randomized 44 patients with prior statin-induced myalgia to receive increasing doses of simvastatin (10–40 mg/day) in combination with either coenzyme Q10 (Q-Gel) 200 mg/day or placebo. The primary goal was to determine if coenzyme Q10 supplementation would help improve statin tolerance in patients with a history of statininduced myalgia. Plasma coenzyme Q10 and lipid levels were measured at baseline and at the end of the study. The intensity of myalgia was assessed with a visual analogue scale.

At 12 weeks, the coenzyme Q10 plasma level was significantly higher in the treatment group than in the placebo group (P < .001). However, no differences were noted between groups in the number of patients who tolerated the 40-mg/day simvastatin dose (P = .34) or in the number of patients who remained on any simvastatin dose (P = .47). Additionally, myalgia scores did not differ between groups (P = .63). The authors acknowledged that there were only small increases in the myalgia pain scores reported in either group. Therefore, patients in the treatment group may not have experienced sufficiently severe muscle pain to have benefited from coenzyme Q10 supplementation.

IS COENZYME Q10 SAFE?

Studies have indicated that these supplements are well tolerated, with relatively few adverse effects or potential drug interactions.^1,2,34

The FDA does not routinely assess the purity or quality of over-the-counter coenzyme Q10 products.³⁵ However, the United States Pharmacopeia (USP) does test dietary supplements to make sure that they are not mislabeled and that they do not contain contaminants. ³⁶

A USP-verified dietary supplement should:

• Contain the exact ingredients listed on the label in the listed potency and amounts
• Not include harmful levels of certain contaminants such as lead, mercury, pesticides, or bacteria
• Appropriately disintegrate and release its contents into the body within a specified period of time
• Be produced using the FDA’s current Good Manufacturing Practices.³⁶

Side effects, contraindications, warnings

Coenzyme Q10 is a relatively safe dietary supplement. It is contraindicated in patients who are allergic to it or to any of its components.² Most clinical trials have not reported significant adverse effects that necessitated stopping therapy.³⁴ However, gastrointestinal effects such as abdominal discomfort, nausea, vomiting, diarrhea, and anorexia have occurred.^1,2,34 Allergic rash and headache have also been reported.^1,2,34 In addition, coenzyme Q10’s antiplatelet effect may increase the risk of bleeding. ^37,38 It undergoes biotransformation in the liver and is eliminated primarily via the biliary tract,³⁹ so it can accumulate in patients with hepatic impairment or biliary obstruction.

Interactions with drugs

Coenzyme Q10’s effects on platelet function may increase the risk of bleeding in patients taking antiplatelet drugs such as aspirin or clopidogrel (Plavix).^37,38 On the other hand, since it acts like vitamin K, it may counteract the anticoagulant effects of warfarin (Coumadin). ^1,2,40

Coenzyme Q10 may have an additive antihypertensive effect when given with antihypertensive drugs.⁴¹

Coenzyme Q10 may improve beta-cell function and enhance insulin sensitivity, which may reduce insulin requirements for diabetic patients.^42,43

SLOWLY ABSORBED

Coenzyme Q10 is absorbed slowly from the gastrointestinal tract, possibly because it has a high molecular weight and is not very watersoluble. ³⁹

One pharmacokinetic study found that after a single 100-mg oral dose of coenzyme Q10, the mean peak plasma levels of about 1 μg/mL occurred between 5 and 10 hours (mean 6.5 hours).⁴⁴ Coenzyme Q10 100 mg given orally three times daily produced a mean steadystate plasma level of 5.4 μg/mL; about 90% of this steady-state concentration was achieved after 4 days.³⁹

Some formulations have significantly better oral bioavailability and therefore produce higher plasma levels. Soft-gel capsules, especially those with vegetable oil or vitamin E, may have better absorption.⁴³

A pharmacokinetic study showed that the area under the curve of the plasma coenzyme Q10 concentration was more than twice as high with Q-Gel soft-gel capsules, a completely solubilized formulation, than with softgel capsules with an oil suspension, powderfilled hard-shell capsules, or regular tablets.⁴⁵ Another study reported that colloidal-Q10, a formulation contained in VESIsorb (a novel drug delivery system sold as CoQsource) had greater bioavailability than solubilized and oil-based preparations.⁴⁶ Commercially available solubilized preparations containing ubiquinol, a metabolized form of coenzyme Q10, have been shown to produce higher serum levels than solubilized products.⁴⁷

Of note: unless the manufacturer claims that its product is water-soluble, the USP does not evaluate its dissolution rate.⁴⁸ Therefore, USP-verified coenzyme Q10 products that are not water-soluble may have lower bioavailability than their solubilized counterparts.

Dry dosage forms of coenzyme Q10 (eg, tablets, capsules) may be more readily absorbed if taken with a fatty meal.⁴³

 

SLOWLY ELIMINATED

Taken orally, coenzyme Q10 has a low clearance rate, with an elimination half-life of about 34 hours.³⁹

After absorption, exogenous coenzyme Q10 is taken up by chylomicrons that transport it to the liver, where it is incorporated into verylow-density lipoproteins. It is then distributed to various organs, including the adrenal glands, spleen, kidneys, lungs, and heart. Coenzyme Q10 is eliminated primarily via the biliary tract. About 60% of an oral dose is eliminated in the feces during chronic oral administration.³⁹

TWICE-DAILY DOSING

A typical daily dose of coenzyme Q10 for treating hypertension is 120 to 200 mg, usually given orally in two divided doses.¹ For statininduced myopathy, 100 to 200 mg orally daily has been used.¹

Coenzyme Q10 is given in divided doses to enhance its absorption and to minimize gastrointestinal effects.^1,43 Taking it with a fatty meal may also increase its absorption.⁴³

Since solubilized forms of coenzyme Q10 and ubiquinol have significantly greater bioavailability than nonsolubilized forms, the therapeutic dose of these formulations may be lower.⁴⁷

MONITORING DURING TREATMENT

Without supplementation, the mean serum level of endogenous coenzyme Q10 has been reported to be 0.99 ± 0.30 mg/L (range 0.55– 1.87).¹⁸ Serum levels above 2 μg/mL have been associated with significant reductions in blood pressure.^5,7,28

The possible effects of coenzyme Q10 on blood pressure, blood glucose levels, serum creatine kinase levels, and myopathic symptoms should be kept in mind when monitoring patients who have hypertension,⁴¹ diabetes,^41,42 or statin-induced myalgia.^15,17 Coenzyme Q10’s possible potentiating effects on antiplatelet drugs and its inhibitory effect on warfarin should be kept in mind as well.

COST VARIES

Coenzyme Q10 is available in different dosage forms (eg, regular and rapid-release softgel capsules, regular and chewable tablets, chewable wafers, and liquid) from a variety of manufacturers. Products come in different strengths, typically ranging from 30 to 400 mg. USP-verified formulations are listed at www.usp.org/USPVerified/dietarySupplements/under “Verified Supplements.” Only USP-verified products that claim to be water-soluble meet USP dissolution requirements.

The cost varies, depending on the vendor. In general, dosage forms with greater bioavailability, such as Q-Gel and ubiquinol supplements, are more expensive. For example, a regimen of 60 mg twice daily of regular-release coenzyme Q capsules may cost approximately $20 per month, compared with $60 per month for the same supply of Q-Gel Ultra capsules. However, in some cases, supplements that produce higher serum levels may be more cost-effective.

CURRENT ROLE IN THERAPY

As an antihypertensive adjunct

Several small clinical trials have shown that coenzyme Q10 supplementation can lower blood pressure. The supplements were reported to be safe and well tolerated. Moreover, some patients with essential hypertension who were taking coenzyme Q10 were able to discontinue one or more antihypertensive drugs. A significant reduction in blood pressure with use of coenzyme Q10 would be expected to reduce the adverse consequences of hypertension in the same manner as conventional antihypertensive agents.

However, no large, double-blind, randomized study has evaluated the impact of coenzyme Q10 when taken with other antihypertensive drugs (eg, angiotensin-converting enzyme inhibitors, beta-blockers, diuretics) on specific clinical end points such as the incidence of stroke or death from a major cardiac event. Furthermore, its effects on cardiac function, exercise tolerance, and quality of life have not been determined.

The bottom line. In some cases, it seems reasonable to recommend this product as an adjunct to conventional antihypertensive therapy. Larger, well-designed clinical trials of coenzyme Q10’s antihypertensive effects on specific clinical end points such as the risk of stroke or myocardial infarction are needed to define its true therapeutic value.

As a treatment for statin-induced myalgia

Clinical evidence supporting coenzyme Q10’s use in the treatment of statin-induced myopathy is limited. Whether coenzyme Q10 is depleted from muscle tissue during statin therapy has not been confirmed. Supplementation helped reduce the severity of musculoskeletal effects of megadoses of lovastatin. However, clinical trials of coenzyme Q10 in the treatment of myalgia associated with antilipidemic statin doses did not consistently report significant improvement. Nevertheless, coenzyme Q10 has been shown to be relatively safe, with few adverse effects.

The bottom line. In some cases, coenzyme Q could be considered as a possible treatment for statin-induced myalgia, pending large-scale studies to determine if it is truly effective for this purpose.

Coenzyme Q10 supplements have been purported to be effective for treating a variety of disorders,^1,2 in particular hypertension and statin-induced myalgia.

Several studies^3–7 found that coenzyme Q10 supplementation significantly lowered blood pressure in hypertensive patients. Moreover, some trials have demonstrated that statin therapy reduces serum or muscle levels of coenzyme Q10,^8–14 prompting investigations to determine whether coenzyme Q10 deficiency is related to statin-induced muscle pain.^15–17

In this review, we discuss the efficacy and safety of coenzyme Q10 supplementation in patients with hypertension and those taking statins, and some of the caveats about using supplements that are not approved by the US Food and Drug Administration (FDA), as well as the bioavailability and quality of available formulations.

WHAT IS COENZYME Q10?

Coenzyme Q10, also known as coenzyme Q, ubidecarenone, and ubiquinone, is found in all human cells, with the highest concentrations in the heart, liver, kidney, and pancreas.^1,2 It is a potent antioxidant, a membrane stabilizer, and an integral cofactor in the mitochondrial respiratory chain, helping to generate adenosine triphosphate, the major cellular energy source.^1,2,18 It may also regulate genes associated with cell metabolism.¹⁹

RATIONALE FOR SUPPLEMENTATION

Coenzyme Q10 supplementation has been used, recommended, or studied in heart failure, hypertension, parkinsonism, mitochondrial encephalomyopathies, and other ailments.

In hypertension

Depending on the class, various antihypertensive drugs can have adverse effects such as depression, cough, and cardiac and renal dysfunction. ^20,21 Furthermore, many patients need to take more than one drug to control their blood pressure, increasing their risk of side effects. Some researchers believe coenzyme Q10 supplementation may reduce the need to take multiple antihypertensive drugs.⁵

Coenzyme Q10 appears to lower blood pressure. The exact mechanism is not known, but one theory is that it reduces peripheral resistance by preserving nitric oxide.²¹ Nitric oxide relaxes peripheral arteries, lowering blood pressure. In some forms of hypertension, superoxide radicals that inactivate nitric oxide are overproduced; coenzyme Q10, with its antioxidant effects, may prevent the inactivation of nitric oxide by these free radicals. Alternatively, coenzyme Q10 may boost the production of the prostaglandin prostacyclin (PGI₂) a potent vasodilator and inhibitor of platelet aggregation, or it may enhance the sensitivity of arterial smooth muscle to PGI₂, or both.^1,22

In patients taking statins

Hydroxymethylglutaryl coenzyme A reductase inhibitors (statins), first-line agents for lowering cholesterol levels to prevent cardiovascular disease, are some of the most commonly prescribed medications.^23,24 However, statin therapy carries a risk of myopathy, which can range from muscle aches to rhabdomyolysis. ^23,24

In a clinical advisory,²⁵ the American College of Cardiology, the American Heart Association, and the National Heart, Lung, and Blood Institute recommend that patients on statin therapy who experience muscle soreness, tenderness, or pain with serum creatine kinase levels 3 to 10 times the upper limit of normal should have their creatine kinase level checked weekly. If the level is 3 to 10 times the upper limit of normal, statin therapy may be continued, but if it exceeds 10 times the upper limit, then statins and other potential offending agents (eg, niacin, fibrate) need to be discontinued.

Statins inhibit the synthesis of cholesterol by reducing the production of mevalonate, a precursor of both cholesterol and coenzyme Q10. Since both cholesterol and coenzyme Q10 are produced by the same pathway, it is not surprising that statins have been reported to reduce serum and muscle coenzyme Q10 levels.^9–14 However, one study did not report a significant reduction of coenzyme Q10 levels in muscle tissue in patients treated with simvastatin 20 mg for 6 months.²⁶

Nonetheless, researchers have hypothesized that a reduction in coenzyme Q10 levels in muscle tissue causes mitochondrial dysfunction, which could increase the risk of statininduced myopathy,^13–17 and some believe that treatment with coenzyme Q10 may reduce myalgic symptoms and allow patients to remain on statin therapy.^13,24

Researchers have investigated the potential of coenzyme Q10 supplementation to reduce or prevent statin-induced myopathy.^15–17 (More on this below.)

Interestingly, a randomized, placebo-controlled trial²⁷ found that 6 months of daily therapy with simvastatin (Zocor) 20 mg or pravastatin (Pravachol) 40 mg lowered systolic and diastolic blood pressure significantly in patients with no documented history of cardiovascular disease or diabetes. A possible mechanism of statin-induced blood pressure reduction is the up-regulation of endothelial nitric oxide synthetase, a potent vasodilator. Coenzyme Q10 levels were not assessed during this study. Whether coenzyme Q10 supplementation used to treat statin-induced myalgia enhances or inhibits the antihypertensive effects of statins is not yet known.

 

EVIDENCE OF EFFECTIVENESS IN HYPERTENSION

A number of trials provide clinical evidence that some patients with high blood pressure may benefit from coenzyme Q10 supplementation (Table 1).^{3–7,28–31}

Rosenfeldt et al²⁸ performed a meta-analysis and found that some trials documented statistically significant reductions in diastolic or systolic blood pressure or both, while others reported negligible effects.^3,29 In one small trial,³⁰ blood pressures actually went up in patients taking coenzyme Q10. Coenzyme Q10 dosages and length of therapy varied from study to study in the meta-analysis. Only minor adverse effects such as gastrointestinal upset and headache were reported.

Yamagami et al³ randomly assigned 20 patients with hypertension and a low coenzyme Q10 level to receive 100 mg of coenzyme Q10 or placebo daily for 12 weeks. Patients continued their usual antihypertensive regimen during the study period. Blood pressures, coenzyme Q10 levels, and antihypertensive drugs used were comparable between the study groups.

After 12 weeks of therapy, the mean coenzyme Q10 level in the active-treatment group had more than doubled, from 0.704 to 1.597 μg/mL. This group also experienced a statistically significant drop in systolic blood pressure, from 167 mm Hg at baseline to 148 mm Hg at 12 weeks. In the placebo group, the systolic blood pressure was 168 mm Hg at baseline and 164 mm Hg at 12 weeks; the change was not statistically significant. Diastolic pressure was not significantly lower at 12 weeks than at baseline in either group.

The authors concluded that coenzyme Q10 supplementation brought a mild reduction in high blood pressure in patients who had low coenzyme Q10 serum levels.

Digiesi et al³¹ randomized 18 patients with essential hypertension to receive either coenzyme Q10 100 mg or placebo daily for 10 weeks. All antihypertensive therapy was discontinued at baseline. After the first 10 weeks, patients went through a 2-week washout period and then were switched to the opposite therapy for an additional 10 weeks. Mean baseline blood pressure values were 167 mm Hg systolic and 103 mm Hg diastolic.

Those taking the supplement had a statistically significant decrease in systolic and diastolic pressures (P < .001). The antihypertensive effect was noted in the 3rd or 4th week of active treatment and persisted for the duration of therapy. The effects dissipated 7 to 10 days after coenzyme Q10 was stopped.

Langsjoen et al⁵ evaluated the effects of adding coenzyme Q10 to the antihypertensive drug regimen of 109 patients who had a primary diagnosis of essential hypertension in a prospective observational study. Patients with hypertension as a secondary diagnosis and other cardiovascular diseases were excluded. Variable doses of coenzyme Q10 were given, adjusted according to clinical response and to achieve serum levels greater than 2.0 μg/mL. The average dose was 225 mg/day; the mean serum level attained was 3.02 μg/mL.

Over several months, patients taking the supplement had a reduction in mean systolic pressure from 159 mm Hg at baseline to 147 mm Hg (P < .001), and a reduction in mean diastolic pressure from 94 to 85 mm Hg (P < .001). Thirty-seven percent of patients were able to discontinue one antihypertensive drug, 11% discontinued two drugs, and 4% were able to stop taking three drugs. However, 46% remained on the same antihypertensive regimen, and 3% needed an additional drug.

Singh et al⁶ randomized 64 patients who had coronary artery disease and who had been on antihypertensive drugs for more than 1 year to receive either B-complex vitamins or coenzyme Q10 (hydrosoluble Q-Gel) 60 mg orally once daily for 8 weeks. Five patients were not available for follow-up; therefore, only 59 patients were evaluated. Fifty-five (93%) of the 59 patients were taking only one antihypertensive drug. Initial antihypertensive drug use was similar between study groups and was continued throughout the trial.

After 8 weeks of therapy, the coenzyme Q10 group had significantly lower systolic and diastolic blood pressure than the placebo group (P < .05 for both). There was also a statistically significant decrease in the dosage of antihypertensive drugs in the coenzyme Q10 group but not in the placebo group (P < .05), reflecting coenzyme Q10’s additive antihypertensive effect.

Burke et al⁷ randomized 41 men and 35 women with isolated systolic hypertension (systolic pressure 150–170 mm Hg, diastolic pressure < 90 mm Hg) to receive a twice-daily dose of 60 mg of emulsified coenzyme Q10 (hydrosoluble Q-Gel) with 150 IU of vitamin E or placebo containing vitamin E alone for 12 weeks. The study also included 5 men and 4 women with normal blood pressure, all of whom received coenzyme Q10. A total of 80 patients completed treatment. The primary goal of the study was to determine the efficacy of coenzyme Q10 in the treatment of isolated systolic hypertension in patients without comorbid conditions. Blood pressures were monitored twice a week during the trial, by the same nurse.

After 12 weeks of treatment, the mean reduction in systolic pressure in hypertensive patients on coenzyme Q10 was 17.8 ± 7.3 mm Hg. There were no significant changes in diastolic pressure in any study group with treatment. Patients with isolated systolic hypertension who were taking coenzyme Q10 had a statistically significant reduction in systolic pressure compared with baseline and placebo (P < .01 for both). Approximately 55% of patients on coenzyme Q10 achieved a reduction in systolic pressure of 4 mm Hg or greater, while 45% did not respond to therapy. The mean plasma coenzyme Q10 level of the treatment group increased from 0.47 ± 0.19 μg/mL to 2.69 ± 0.54 μg/mL after 12 weeks; however, the study did not have the statistical power to demonstrate a relationship between coenzyme Q10 levels and changes in blood pressure. Twenty-seven (34%) of the 80 patients were taking a statin while on coenzyme Q10 therapy.

 

STUDIES IN STATIN-INDUCED MYOPATHY

Thibault et al³² and Kim et al³³ reported that patients taking lovastatin (Mevacor) at dosages as high as 35 mg/kg/day to inhibit tumor growth achieved symptomatic relief of statin-induced musculoskeletal toxicity after coenzyme Q10 supplementation.

Caso et al¹⁵ performed a small pilot study in 32 patients to determine if coenzyme Q10 supplementation would improve myalgic symptoms in patients treated with statins. In this double-blind, randomized trial, patients received either coenzyme Q10 100 mg/day or vitamin E 400 IU/day for 30 days. The extent of muscle pain and its interference with daily activities were determined before and after therapy using the Brief Pain Inventory Questionnaire. The statins were atorvastatin (Lipitor) 10 mg or 20 mg, lovastatin 40 mg, pravastatin 40 mg, and simvastatin 10, 20, 40, and 80 mg. Five patients in the coenzyme Q10 group and four patients in the vitamin E group were taking nonsteroidal anti-inflammatory drugs before and during the trial. The intensity of muscle pain and its interference with daily activities were similar between study groups before the start of therapy.

After 30 days of treatment with coenzyme Q10, the pain intensity had decreased significantly from baseline (P < .001). In contrast, no change in pain intensity from baseline was noted in patients receiving vitamin E. The Pain Severity Score was significantly different between study groups, favoring the coenzyme Q10 group (P < .001). Sixteen of 18 patients on coenzyme Q10 reported a reduction in pain, while only 3 of 14 patients on vitamin E reported a similar response. Also, the interference of pain with daily activities significantly improved with coenzyme Q10 (P < .02), whereas vitamin E did not have a significant impact on this.

Young et al¹⁷ randomized 44 patients with prior statin-induced myalgia to receive increasing doses of simvastatin (10–40 mg/day) in combination with either coenzyme Q10 (Q-Gel) 200 mg/day or placebo. The primary goal was to determine if coenzyme Q10 supplementation would help improve statin tolerance in patients with a history of statininduced myalgia. Plasma coenzyme Q10 and lipid levels were measured at baseline and at the end of the study. The intensity of myalgia was assessed with a visual analogue scale.

At 12 weeks, the coenzyme Q10 plasma level was significantly higher in the treatment group than in the placebo group (P < .001). However, no differences were noted between groups in the number of patients who tolerated the 40-mg/day simvastatin dose (P = .34) or in the number of patients who remained on any simvastatin dose (P = .47). Additionally, myalgia scores did not differ between groups (P = .63). The authors acknowledged that there were only small increases in the myalgia pain scores reported in either group. Therefore, patients in the treatment group may not have experienced sufficiently severe muscle pain to have benefited from coenzyme Q10 supplementation.

IS COENZYME Q10 SAFE?

Studies have indicated that these supplements are well tolerated, with relatively few adverse effects or potential drug interactions.^1,2,34

The FDA does not routinely assess the purity or quality of over-the-counter coenzyme Q10 products.³⁵ However, the United States Pharmacopeia (USP) does test dietary supplements to make sure that they are not mislabeled and that they do not contain contaminants. ³⁶

A USP-verified dietary supplement should:

• Contain the exact ingredients listed on the label in the listed potency and amounts
• Not include harmful levels of certain contaminants such as lead, mercury, pesticides, or bacteria
• Appropriately disintegrate and release its contents into the body within a specified period of time
• Be produced using the FDA’s current Good Manufacturing Practices.³⁶

Side effects, contraindications, warnings

Coenzyme Q10 is a relatively safe dietary supplement. It is contraindicated in patients who are allergic to it or to any of its components.² Most clinical trials have not reported significant adverse effects that necessitated stopping therapy.³⁴ However, gastrointestinal effects such as abdominal discomfort, nausea, vomiting, diarrhea, and anorexia have occurred.^1,2,34 Allergic rash and headache have also been reported.^1,2,34 In addition, coenzyme Q10’s antiplatelet effect may increase the risk of bleeding. ^37,38 It undergoes biotransformation in the liver and is eliminated primarily via the biliary tract,³⁹ so it can accumulate in patients with hepatic impairment or biliary obstruction.

Interactions with drugs

Coenzyme Q10’s effects on platelet function may increase the risk of bleeding in patients taking antiplatelet drugs such as aspirin or clopidogrel (Plavix).^37,38 On the other hand, since it acts like vitamin K, it may counteract the anticoagulant effects of warfarin (Coumadin). ^1,2,40

Coenzyme Q10 may have an additive antihypertensive effect when given with antihypertensive drugs.⁴¹

Coenzyme Q10 may improve beta-cell function and enhance insulin sensitivity, which may reduce insulin requirements for diabetic patients.^42,43

SLOWLY ABSORBED

Coenzyme Q10 is absorbed slowly from the gastrointestinal tract, possibly because it has a high molecular weight and is not very watersoluble. ³⁹

One pharmacokinetic study found that after a single 100-mg oral dose of coenzyme Q10, the mean peak plasma levels of about 1 μg/mL occurred between 5 and 10 hours (mean 6.5 hours).⁴⁴ Coenzyme Q10 100 mg given orally three times daily produced a mean steadystate plasma level of 5.4 μg/mL; about 90% of this steady-state concentration was achieved after 4 days.³⁹

Some formulations have significantly better oral bioavailability and therefore produce higher plasma levels. Soft-gel capsules, especially those with vegetable oil or vitamin E, may have better absorption.⁴³

A pharmacokinetic study showed that the area under the curve of the plasma coenzyme Q10 concentration was more than twice as high with Q-Gel soft-gel capsules, a completely solubilized formulation, than with softgel capsules with an oil suspension, powderfilled hard-shell capsules, or regular tablets.⁴⁵ Another study reported that colloidal-Q10, a formulation contained in VESIsorb (a novel drug delivery system sold as CoQsource) had greater bioavailability than solubilized and oil-based preparations.⁴⁶ Commercially available solubilized preparations containing ubiquinol, a metabolized form of coenzyme Q10, have been shown to produce higher serum levels than solubilized products.⁴⁷

Of note: unless the manufacturer claims that its product is water-soluble, the USP does not evaluate its dissolution rate.⁴⁸ Therefore, USP-verified coenzyme Q10 products that are not water-soluble may have lower bioavailability than their solubilized counterparts.

Dry dosage forms of coenzyme Q10 (eg, tablets, capsules) may be more readily absorbed if taken with a fatty meal.⁴³

 

SLOWLY ELIMINATED

Taken orally, coenzyme Q10 has a low clearance rate, with an elimination half-life of about 34 hours.³⁹

After absorption, exogenous coenzyme Q10 is taken up by chylomicrons that transport it to the liver, where it is incorporated into verylow-density lipoproteins. It is then distributed to various organs, including the adrenal glands, spleen, kidneys, lungs, and heart. Coenzyme Q10 is eliminated primarily via the biliary tract. About 60% of an oral dose is eliminated in the feces during chronic oral administration.³⁹

TWICE-DAILY DOSING

A typical daily dose of coenzyme Q10 for treating hypertension is 120 to 200 mg, usually given orally in two divided doses.¹ For statininduced myopathy, 100 to 200 mg orally daily has been used.¹

Coenzyme Q10 is given in divided doses to enhance its absorption and to minimize gastrointestinal effects.^1,43 Taking it with a fatty meal may also increase its absorption.⁴³

Since solubilized forms of coenzyme Q10 and ubiquinol have significantly greater bioavailability than nonsolubilized forms, the therapeutic dose of these formulations may be lower.⁴⁷

MONITORING DURING TREATMENT

Without supplementation, the mean serum level of endogenous coenzyme Q10 has been reported to be 0.99 ± 0.30 mg/L (range 0.55– 1.87).¹⁸ Serum levels above 2 μg/mL have been associated with significant reductions in blood pressure.^5,7,28

The possible effects of coenzyme Q10 on blood pressure, blood glucose levels, serum creatine kinase levels, and myopathic symptoms should be kept in mind when monitoring patients who have hypertension,⁴¹ diabetes,^41,42 or statin-induced myalgia.^15,17 Coenzyme Q10’s possible potentiating effects on antiplatelet drugs and its inhibitory effect on warfarin should be kept in mind as well.

COST VARIES

Coenzyme Q10 is available in different dosage forms (eg, regular and rapid-release softgel capsules, regular and chewable tablets, chewable wafers, and liquid) from a variety of manufacturers. Products come in different strengths, typically ranging from 30 to 400 mg. USP-verified formulations are listed at www.usp.org/USPVerified/dietarySupplements/under “Verified Supplements.” Only USP-verified products that claim to be water-soluble meet USP dissolution requirements.

The cost varies, depending on the vendor. In general, dosage forms with greater bioavailability, such as Q-Gel and ubiquinol supplements, are more expensive. For example, a regimen of 60 mg twice daily of regular-release coenzyme Q capsules may cost approximately $20 per month, compared with $60 per month for the same supply of Q-Gel Ultra capsules. However, in some cases, supplements that produce higher serum levels may be more cost-effective.

CURRENT ROLE IN THERAPY

As an antihypertensive adjunct

Several small clinical trials have shown that coenzyme Q10 supplementation can lower blood pressure. The supplements were reported to be safe and well tolerated. Moreover, some patients with essential hypertension who were taking coenzyme Q10 were able to discontinue one or more antihypertensive drugs. A significant reduction in blood pressure with use of coenzyme Q10 would be expected to reduce the adverse consequences of hypertension in the same manner as conventional antihypertensive agents.

However, no large, double-blind, randomized study has evaluated the impact of coenzyme Q10 when taken with other antihypertensive drugs (eg, angiotensin-converting enzyme inhibitors, beta-blockers, diuretics) on specific clinical end points such as the incidence of stroke or death from a major cardiac event. Furthermore, its effects on cardiac function, exercise tolerance, and quality of life have not been determined.

The bottom line. In some cases, it seems reasonable to recommend this product as an adjunct to conventional antihypertensive therapy. Larger, well-designed clinical trials of coenzyme Q10’s antihypertensive effects on specific clinical end points such as the risk of stroke or myocardial infarction are needed to define its true therapeutic value.

As a treatment for statin-induced myalgia

Clinical evidence supporting coenzyme Q10’s use in the treatment of statin-induced myopathy is limited. Whether coenzyme Q10 is depleted from muscle tissue during statin therapy has not been confirmed. Supplementation helped reduce the severity of musculoskeletal effects of megadoses of lovastatin. However, clinical trials of coenzyme Q10 in the treatment of myalgia associated with antilipidemic statin doses did not consistently report significant improvement. Nevertheless, coenzyme Q10 has been shown to be relatively safe, with few adverse effects.

The bottom line. In some cases, coenzyme Q could be considered as a possible treatment for statin-induced myalgia, pending large-scale studies to determine if it is truly effective for this purpose.

Coenzyme Q10 supplements have been purported to be effective for treating a variety of disorders,^1,2 in particular hypertension and statin-induced myalgia.

Several studies^3–7 found that coenzyme Q10 supplementation significantly lowered blood pressure in hypertensive patients. Moreover, some trials have demonstrated that statin therapy reduces serum or muscle levels of coenzyme Q10,^8–14 prompting investigations to determine whether coenzyme Q10 deficiency is related to statin-induced muscle pain.^15–17

In this review, we discuss the efficacy and safety of coenzyme Q10 supplementation in patients with hypertension and those taking statins, and some of the caveats about using supplements that are not approved by the US Food and Drug Administration (FDA), as well as the bioavailability and quality of available formulations.

WHAT IS COENZYME Q10?

Coenzyme Q10, also known as coenzyme Q, ubidecarenone, and ubiquinone, is found in all human cells, with the highest concentrations in the heart, liver, kidney, and pancreas.^1,2 It is a potent antioxidant, a membrane stabilizer, and an integral cofactor in the mitochondrial respiratory chain, helping to generate adenosine triphosphate, the major cellular energy source.^1,2,18 It may also regulate genes associated with cell metabolism.¹⁹

RATIONALE FOR SUPPLEMENTATION

Coenzyme Q10 supplementation has been used, recommended, or studied in heart failure, hypertension, parkinsonism, mitochondrial encephalomyopathies, and other ailments.

In hypertension

Depending on the class, various antihypertensive drugs can have adverse effects such as depression, cough, and cardiac and renal dysfunction. ^20,21 Furthermore, many patients need to take more than one drug to control their blood pressure, increasing their risk of side effects. Some researchers believe coenzyme Q10 supplementation may reduce the need to take multiple antihypertensive drugs.⁵

Coenzyme Q10 appears to lower blood pressure. The exact mechanism is not known, but one theory is that it reduces peripheral resistance by preserving nitric oxide.²¹ Nitric oxide relaxes peripheral arteries, lowering blood pressure. In some forms of hypertension, superoxide radicals that inactivate nitric oxide are overproduced; coenzyme Q10, with its antioxidant effects, may prevent the inactivation of nitric oxide by these free radicals. Alternatively, coenzyme Q10 may boost the production of the prostaglandin prostacyclin (PGI₂) a potent vasodilator and inhibitor of platelet aggregation, or it may enhance the sensitivity of arterial smooth muscle to PGI₂, or both.^1,22

In patients taking statins

Hydroxymethylglutaryl coenzyme A reductase inhibitors (statins), first-line agents for lowering cholesterol levels to prevent cardiovascular disease, are some of the most commonly prescribed medications.^23,24 However, statin therapy carries a risk of myopathy, which can range from muscle aches to rhabdomyolysis. ^23,24

In a clinical advisory,²⁵ the American College of Cardiology, the American Heart Association, and the National Heart, Lung, and Blood Institute recommend that patients on statin therapy who experience muscle soreness, tenderness, or pain with serum creatine kinase levels 3 to 10 times the upper limit of normal should have their creatine kinase level checked weekly. If the level is 3 to 10 times the upper limit of normal, statin therapy may be continued, but if it exceeds 10 times the upper limit, then statins and other potential offending agents (eg, niacin, fibrate) need to be discontinued.

Statins inhibit the synthesis of cholesterol by reducing the production of mevalonate, a precursor of both cholesterol and coenzyme Q10. Since both cholesterol and coenzyme Q10 are produced by the same pathway, it is not surprising that statins have been reported to reduce serum and muscle coenzyme Q10 levels.^9–14 However, one study did not report a significant reduction of coenzyme Q10 levels in muscle tissue in patients treated with simvastatin 20 mg for 6 months.²⁶

Nonetheless, researchers have hypothesized that a reduction in coenzyme Q10 levels in muscle tissue causes mitochondrial dysfunction, which could increase the risk of statininduced myopathy,^13–17 and some believe that treatment with coenzyme Q10 may reduce myalgic symptoms and allow patients to remain on statin therapy.^13,24

Researchers have investigated the potential of coenzyme Q10 supplementation to reduce or prevent statin-induced myopathy.^15–17 (More on this below.)

Interestingly, a randomized, placebo-controlled trial²⁷ found that 6 months of daily therapy with simvastatin (Zocor) 20 mg or pravastatin (Pravachol) 40 mg lowered systolic and diastolic blood pressure significantly in patients with no documented history of cardiovascular disease or diabetes. A possible mechanism of statin-induced blood pressure reduction is the up-regulation of endothelial nitric oxide synthetase, a potent vasodilator. Coenzyme Q10 levels were not assessed during this study. Whether coenzyme Q10 supplementation used to treat statin-induced myalgia enhances or inhibits the antihypertensive effects of statins is not yet known.

 

EVIDENCE OF EFFECTIVENESS IN HYPERTENSION

A number of trials provide clinical evidence that some patients with high blood pressure may benefit from coenzyme Q10 supplementation (Table 1).^{3–7,28–31}

Rosenfeldt et al²⁸ performed a meta-analysis and found that some trials documented statistically significant reductions in diastolic or systolic blood pressure or both, while others reported negligible effects.^3,29 In one small trial,³⁰ blood pressures actually went up in patients taking coenzyme Q10. Coenzyme Q10 dosages and length of therapy varied from study to study in the meta-analysis. Only minor adverse effects such as gastrointestinal upset and headache were reported.

Yamagami et al³ randomly assigned 20 patients with hypertension and a low coenzyme Q10 level to receive 100 mg of coenzyme Q10 or placebo daily for 12 weeks. Patients continued their usual antihypertensive regimen during the study period. Blood pressures, coenzyme Q10 levels, and antihypertensive drugs used were comparable between the study groups.

After 12 weeks of therapy, the mean coenzyme Q10 level in the active-treatment group had more than doubled, from 0.704 to 1.597 μg/mL. This group also experienced a statistically significant drop in systolic blood pressure, from 167 mm Hg at baseline to 148 mm Hg at 12 weeks. In the placebo group, the systolic blood pressure was 168 mm Hg at baseline and 164 mm Hg at 12 weeks; the change was not statistically significant. Diastolic pressure was not significantly lower at 12 weeks than at baseline in either group.

The authors concluded that coenzyme Q10 supplementation brought a mild reduction in high blood pressure in patients who had low coenzyme Q10 serum levels.

Digiesi et al³¹ randomized 18 patients with essential hypertension to receive either coenzyme Q10 100 mg or placebo daily for 10 weeks. All antihypertensive therapy was discontinued at baseline. After the first 10 weeks, patients went through a 2-week washout period and then were switched to the opposite therapy for an additional 10 weeks. Mean baseline blood pressure values were 167 mm Hg systolic and 103 mm Hg diastolic.

Those taking the supplement had a statistically significant decrease in systolic and diastolic pressures (P < .001). The antihypertensive effect was noted in the 3rd or 4th week of active treatment and persisted for the duration of therapy. The effects dissipated 7 to 10 days after coenzyme Q10 was stopped.

Langsjoen et al⁵ evaluated the effects of adding coenzyme Q10 to the antihypertensive drug regimen of 109 patients who had a primary diagnosis of essential hypertension in a prospective observational study. Patients with hypertension as a secondary diagnosis and other cardiovascular diseases were excluded. Variable doses of coenzyme Q10 were given, adjusted according to clinical response and to achieve serum levels greater than 2.0 μg/mL. The average dose was 225 mg/day; the mean serum level attained was 3.02 μg/mL.

Over several months, patients taking the supplement had a reduction in mean systolic pressure from 159 mm Hg at baseline to 147 mm Hg (P < .001), and a reduction in mean diastolic pressure from 94 to 85 mm Hg (P < .001). Thirty-seven percent of patients were able to discontinue one antihypertensive drug, 11% discontinued two drugs, and 4% were able to stop taking three drugs. However, 46% remained on the same antihypertensive regimen, and 3% needed an additional drug.

Singh et al⁶ randomized 64 patients who had coronary artery disease and who had been on antihypertensive drugs for more than 1 year to receive either B-complex vitamins or coenzyme Q10 (hydrosoluble Q-Gel) 60 mg orally once daily for 8 weeks. Five patients were not available for follow-up; therefore, only 59 patients were evaluated. Fifty-five (93%) of the 59 patients were taking only one antihypertensive drug. Initial antihypertensive drug use was similar between study groups and was continued throughout the trial.

After 8 weeks of therapy, the coenzyme Q10 group had significantly lower systolic and diastolic blood pressure than the placebo group (P < .05 for both). There was also a statistically significant decrease in the dosage of antihypertensive drugs in the coenzyme Q10 group but not in the placebo group (P < .05), reflecting coenzyme Q10’s additive antihypertensive effect.

Burke et al⁷ randomized 41 men and 35 women with isolated systolic hypertension (systolic pressure 150–170 mm Hg, diastolic pressure < 90 mm Hg) to receive a twice-daily dose of 60 mg of emulsified coenzyme Q10 (hydrosoluble Q-Gel) with 150 IU of vitamin E or placebo containing vitamin E alone for 12 weeks. The study also included 5 men and 4 women with normal blood pressure, all of whom received coenzyme Q10. A total of 80 patients completed treatment. The primary goal of the study was to determine the efficacy of coenzyme Q10 in the treatment of isolated systolic hypertension in patients without comorbid conditions. Blood pressures were monitored twice a week during the trial, by the same nurse.

After 12 weeks of treatment, the mean reduction in systolic pressure in hypertensive patients on coenzyme Q10 was 17.8 ± 7.3 mm Hg. There were no significant changes in diastolic pressure in any study group with treatment. Patients with isolated systolic hypertension who were taking coenzyme Q10 had a statistically significant reduction in systolic pressure compared with baseline and placebo (P < .01 for both). Approximately 55% of patients on coenzyme Q10 achieved a reduction in systolic pressure of 4 mm Hg or greater, while 45% did not respond to therapy. The mean plasma coenzyme Q10 level of the treatment group increased from 0.47 ± 0.19 μg/mL to 2.69 ± 0.54 μg/mL after 12 weeks; however, the study did not have the statistical power to demonstrate a relationship between coenzyme Q10 levels and changes in blood pressure. Twenty-seven (34%) of the 80 patients were taking a statin while on coenzyme Q10 therapy.

 

STUDIES IN STATIN-INDUCED MYOPATHY

Thibault et al³² and Kim et al³³ reported that patients taking lovastatin (Mevacor) at dosages as high as 35 mg/kg/day to inhibit tumor growth achieved symptomatic relief of statin-induced musculoskeletal toxicity after coenzyme Q10 supplementation.

Caso et al¹⁵ performed a small pilot study in 32 patients to determine if coenzyme Q10 supplementation would improve myalgic symptoms in patients treated with statins. In this double-blind, randomized trial, patients received either coenzyme Q10 100 mg/day or vitamin E 400 IU/day for 30 days. The extent of muscle pain and its interference with daily activities were determined before and after therapy using the Brief Pain Inventory Questionnaire. The statins were atorvastatin (Lipitor) 10 mg or 20 mg, lovastatin 40 mg, pravastatin 40 mg, and simvastatin 10, 20, 40, and 80 mg. Five patients in the coenzyme Q10 group and four patients in the vitamin E group were taking nonsteroidal anti-inflammatory drugs before and during the trial. The intensity of muscle pain and its interference with daily activities were similar between study groups before the start of therapy.

After 30 days of treatment with coenzyme Q10, the pain intensity had decreased significantly from baseline (P < .001). In contrast, no change in pain intensity from baseline was noted in patients receiving vitamin E. The Pain Severity Score was significantly different between study groups, favoring the coenzyme Q10 group (P < .001). Sixteen of 18 patients on coenzyme Q10 reported a reduction in pain, while only 3 of 14 patients on vitamin E reported a similar response. Also, the interference of pain with daily activities significantly improved with coenzyme Q10 (P < .02), whereas vitamin E did not have a significant impact on this.

Young et al¹⁷ randomized 44 patients with prior statin-induced myalgia to receive increasing doses of simvastatin (10–40 mg/day) in combination with either coenzyme Q10 (Q-Gel) 200 mg/day or placebo. The primary goal was to determine if coenzyme Q10 supplementation would help improve statin tolerance in patients with a history of statininduced myalgia. Plasma coenzyme Q10 and lipid levels were measured at baseline and at the end of the study. The intensity of myalgia was assessed with a visual analogue scale.

At 12 weeks, the coenzyme Q10 plasma level was significantly higher in the treatment group than in the placebo group (P < .001). However, no differences were noted between groups in the number of patients who tolerated the 40-mg/day simvastatin dose (P = .34) or in the number of patients who remained on any simvastatin dose (P = .47). Additionally, myalgia scores did not differ between groups (P = .63). The authors acknowledged that there were only small increases in the myalgia pain scores reported in either group. Therefore, patients in the treatment group may not have experienced sufficiently severe muscle pain to have benefited from coenzyme Q10 supplementation.

IS COENZYME Q10 SAFE?

Studies have indicated that these supplements are well tolerated, with relatively few adverse effects or potential drug interactions.^1,2,34

The FDA does not routinely assess the purity or quality of over-the-counter coenzyme Q10 products.³⁵ However, the United States Pharmacopeia (USP) does test dietary supplements to make sure that they are not mislabeled and that they do not contain contaminants. ³⁶

A USP-verified dietary supplement should:

• Contain the exact ingredients listed on the label in the listed potency and amounts
• Not include harmful levels of certain contaminants such as lead, mercury, pesticides, or bacteria
• Appropriately disintegrate and release its contents into the body within a specified period of time
• Be produced using the FDA’s current Good Manufacturing Practices.³⁶

Side effects, contraindications, warnings

Coenzyme Q10 is a relatively safe dietary supplement. It is contraindicated in patients who are allergic to it or to any of its components.² Most clinical trials have not reported significant adverse effects that necessitated stopping therapy.³⁴ However, gastrointestinal effects such as abdominal discomfort, nausea, vomiting, diarrhea, and anorexia have occurred.^1,2,34 Allergic rash and headache have also been reported.^1,2,34 In addition, coenzyme Q10’s antiplatelet effect may increase the risk of bleeding. ^37,38 It undergoes biotransformation in the liver and is eliminated primarily via the biliary tract,³⁹ so it can accumulate in patients with hepatic impairment or biliary obstruction.

Interactions with drugs

Coenzyme Q10’s effects on platelet function may increase the risk of bleeding in patients taking antiplatelet drugs such as aspirin or clopidogrel (Plavix).^37,38 On the other hand, since it acts like vitamin K, it may counteract the anticoagulant effects of warfarin (Coumadin). ^1,2,40

Coenzyme Q10 may have an additive antihypertensive effect when given with antihypertensive drugs.⁴¹

Coenzyme Q10 may improve beta-cell function and enhance insulin sensitivity, which may reduce insulin requirements for diabetic patients.^42,43

SLOWLY ABSORBED

Coenzyme Q10 is absorbed slowly from the gastrointestinal tract, possibly because it has a high molecular weight and is not very watersoluble. ³⁹

One pharmacokinetic study found that after a single 100-mg oral dose of coenzyme Q10, the mean peak plasma levels of about 1 μg/mL occurred between 5 and 10 hours (mean 6.5 hours).⁴⁴ Coenzyme Q10 100 mg given orally three times daily produced a mean steadystate plasma level of 5.4 μg/mL; about 90% of this steady-state concentration was achieved after 4 days.³⁹

Some formulations have significantly better oral bioavailability and therefore produce higher plasma levels. Soft-gel capsules, especially those with vegetable oil or vitamin E, may have better absorption.⁴³

A pharmacokinetic study showed that the area under the curve of the plasma coenzyme Q10 concentration was more than twice as high with Q-Gel soft-gel capsules, a completely solubilized formulation, than with softgel capsules with an oil suspension, powderfilled hard-shell capsules, or regular tablets.⁴⁵ Another study reported that colloidal-Q10, a formulation contained in VESIsorb (a novel drug delivery system sold as CoQsource) had greater bioavailability than solubilized and oil-based preparations.⁴⁶ Commercially available solubilized preparations containing ubiquinol, a metabolized form of coenzyme Q10, have been shown to produce higher serum levels than solubilized products.⁴⁷

Of note: unless the manufacturer claims that its product is water-soluble, the USP does not evaluate its dissolution rate.⁴⁸ Therefore, USP-verified coenzyme Q10 products that are not water-soluble may have lower bioavailability than their solubilized counterparts.

Dry dosage forms of coenzyme Q10 (eg, tablets, capsules) may be more readily absorbed if taken with a fatty meal.⁴³

 

SLOWLY ELIMINATED

Taken orally, coenzyme Q10 has a low clearance rate, with an elimination half-life of about 34 hours.³⁹

After absorption, exogenous coenzyme Q10 is taken up by chylomicrons that transport it to the liver, where it is incorporated into verylow-density lipoproteins. It is then distributed to various organs, including the adrenal glands, spleen, kidneys, lungs, and heart. Coenzyme Q10 is eliminated primarily via the biliary tract. About 60% of an oral dose is eliminated in the feces during chronic oral administration.³⁹

TWICE-DAILY DOSING

A typical daily dose of coenzyme Q10 for treating hypertension is 120 to 200 mg, usually given orally in two divided doses.¹ For statininduced myopathy, 100 to 200 mg orally daily has been used.¹

Coenzyme Q10 is given in divided doses to enhance its absorption and to minimize gastrointestinal effects.^1,43 Taking it with a fatty meal may also increase its absorption.⁴³

Since solubilized forms of coenzyme Q10 and ubiquinol have significantly greater bioavailability than nonsolubilized forms, the therapeutic dose of these formulations may be lower.⁴⁷

MONITORING DURING TREATMENT

Without supplementation, the mean serum level of endogenous coenzyme Q10 has been reported to be 0.99 ± 0.30 mg/L (range 0.55– 1.87).¹⁸ Serum levels above 2 μg/mL have been associated with significant reductions in blood pressure.^5,7,28

The possible effects of coenzyme Q10 on blood pressure, blood glucose levels, serum creatine kinase levels, and myopathic symptoms should be kept in mind when monitoring patients who have hypertension,⁴¹ diabetes,^41,42 or statin-induced myalgia.^15,17 Coenzyme Q10’s possible potentiating effects on antiplatelet drugs and its inhibitory effect on warfarin should be kept in mind as well.

COST VARIES

Coenzyme Q10 is available in different dosage forms (eg, regular and rapid-release softgel capsules, regular and chewable tablets, chewable wafers, and liquid) from a variety of manufacturers. Products come in different strengths, typically ranging from 30 to 400 mg. USP-verified formulations are listed at www.usp.org/USPVerified/dietarySupplements/under “Verified Supplements.” Only USP-verified products that claim to be water-soluble meet USP dissolution requirements.

The cost varies, depending on the vendor. In general, dosage forms with greater bioavailability, such as Q-Gel and ubiquinol supplements, are more expensive. For example, a regimen of 60 mg twice daily of regular-release coenzyme Q capsules may cost approximately $20 per month, compared with $60 per month for the same supply of Q-Gel Ultra capsules. However, in some cases, supplements that produce higher serum levels may be more cost-effective.

CURRENT ROLE IN THERAPY

As an antihypertensive adjunct

Several small clinical trials have shown that coenzyme Q10 supplementation can lower blood pressure. The supplements were reported to be safe and well tolerated. Moreover, some patients with essential hypertension who were taking coenzyme Q10 were able to discontinue one or more antihypertensive drugs. A significant reduction in blood pressure with use of coenzyme Q10 would be expected to reduce the adverse consequences of hypertension in the same manner as conventional antihypertensive agents.

However, no large, double-blind, randomized study has evaluated the impact of coenzyme Q10 when taken with other antihypertensive drugs (eg, angiotensin-converting enzyme inhibitors, beta-blockers, diuretics) on specific clinical end points such as the incidence of stroke or death from a major cardiac event. Furthermore, its effects on cardiac function, exercise tolerance, and quality of life have not been determined.

The bottom line. In some cases, it seems reasonable to recommend this product as an adjunct to conventional antihypertensive therapy. Larger, well-designed clinical trials of coenzyme Q10’s antihypertensive effects on specific clinical end points such as the risk of stroke or myocardial infarction are needed to define its true therapeutic value.

As a treatment for statin-induced myalgia

Clinical evidence supporting coenzyme Q10’s use in the treatment of statin-induced myopathy is limited. Whether coenzyme Q10 is depleted from muscle tissue during statin therapy has not been confirmed. Supplementation helped reduce the severity of musculoskeletal effects of megadoses of lovastatin. However, clinical trials of coenzyme Q10 in the treatment of myalgia associated with antilipidemic statin doses did not consistently report significant improvement. Nevertheless, coenzyme Q10 has been shown to be relatively safe, with few adverse effects.

The bottom line. In some cases, coenzyme Q could be considered as a possible treatment for statin-induced myalgia, pending large-scale studies to determine if it is truly effective for this purpose.

The medical management of non-ST-elevation acute coronary syndromes focuses on blocking the coagulation cascade and inhibiting platelets. This—plus diagnostic angiography followed, if needed, by revascularization—has reduced the rates of death and recurrent ischemic events.¹ However, the combination of potent antithrombotic drugs and invasive procedures also increases the risk of bleeding.

This review discusses the incidence and complications associated with bleeding during the treatment of acute coronary syndromes and summarizes recommendations for preventing and managing bleeding in this setting.

THE TRUE INCIDENCE OF BLEEDING IS HARD TO DETERMINE

The optimal way to detect and analyze bleeding events in clinical trials and registries is highly debated. The reported incidences of bleeding during antithrombotic and antiplatelet therapy for non-ST-elevation acute coronary syndromes depend on how bleeding was defined, how the acute coronary syndromes were treated, and on other factors such as how the study was designed.

How was bleeding defined?

The first bleeding classification schemes were the GUSTO² and the TIMI³ scales (Table 1), both of which were developed for studies of thrombolytic therapy for ST-elevation myocardial infarction. The GUSTO classification is based on clinical events and categorizes bleeding as severe, moderate, or mild. In contrast, the TIMI classification is based on laboratory values and categorizes bleeding as major, moderate, or minor.

Since these classification schemes are based on different types of data, they yield different numbers when applied to the same study population. For instance, Rao et al⁴ pooled the data from the PURSUIT and PARAGON B trials (15,454 patients in all) and found that the incidence of severe bleeding (by the GUSTO criteria) was 1.2%, while the rate of major bleeding (by the TIMI criteria) was 8.2%.

What was the treatment strategy?

Another reason that the true incidence of bleeding is hard to determine is that different studies used treatment strategies that differed in the type, timing, and dose of antithrombotic agents and whether invasive procedures were used early. For example, if unfractionated heparin is used aggressively in regimens that are not adjusted for weight and with a higher target for the activated clotting time, the risk of bleeding is higher than with conservative dosing.^5–7

Subherwal et al⁸ evaluated the effect of treatment strategy on the incidence of bleeding in patients with non-ST-elevation acute coronary syndromes who received two or more antithrombotic drugs in the CRUSADE Quality Improvement Initiative. The risk of bleeding was higher with an invasive approach (catheterization) than with a conservative approach (no catheterization), regardless of baseline bleeding risk.

What type of study was it?

Another source of variation is the design of the study. Registries differ from clinical trials in patient characteristics and in the way data are gathered (prospectively vs retrospectively).

In registries, data are often collected retrospectively, whereas in clinical trials the data are prospectively collected. For this reason, the definition of bleeding in registries is often based on events that are easily identified through chart review, such as transfusion. This may lead to a lower reported rate of bleeding, since other, less serious bleeding events such as access-site hematomas and epistaxis may not be documented in the medical record.

On the other hand, registries often include older and sicker patients, who may be more prone to bleeding and who are often excluded from clinical trials. This may lead to a higher rate of reported bleeding.⁹

Where the study was conducted makes a difference as well, owing to regional practice differences. For example, Moscucci et al¹⁰ reported that the incidence of major bleeding in 24,045 patients with non-ST-elevation acute coronary syndromes in the GRACE registry (in 14 countries worldwide) was 3.9%. In contrast, Yang et al¹¹ reported that the rate of bleeding in the CRUSADE registry (in the United States) was 10.3%.

This difference was partly influenced by different definitions of bleeding. The GRACE registry defined major bleeding as life-threatening events requiring transfusion of two or more units of packed red blood cells, or resulting in an absolute decrease in the hematocrit of 10% or more or death, or hemorrhagic subdural hematoma. In contrast, the CRUSADE data reflect bleeding requiring transfusion. However, practice patterns such as greater use of invasive procedures in the United States may also be responsible.

Rao and colleagues¹² examined international variation in blood transfusion rates among patients with acute coronary syndromes. Patients outside the United States were significantly less likely to receive transfusions, even after adjusting for patient and practice differences.

Taking these confounders into account, it is reasonable to estimate that the frequency of bleeding in patients with non-ST-elevation acute coronary syndromes ranges from less than 1% to 10%.¹³

 

BLEEDING IS ASSOCIATED WITH POOR OUTCOMES

Regardless of the definition or the data source, hemorrhagic complications are associated with a higher risk of death and nonfatal adverse events, both in the short term and in the long term.

Short-term outcomes

A higher risk of death. In the GRACE registry study by Moscucci et al¹⁰ discussed above, patients who had major bleeding were significantly more likely to die during their hospitalization than those who did not (odds ratio [OR] 1.64, 95% confidence interval [CI] 1.18–2.28).

Rao et al¹⁴ evaluated pooled data from the multicenter international GUSTO IIb, PURSUIT, and PARAGON A and B trials and found that the effects of bleeding in non-ST-elevation acute coronary syndromes extended beyond the hospital stay. The more severe the bleeding (by the GUSTO criteria), the greater the adjusted hazard ratio (HR) for death within 30 days:

• With mild bleeding—HR 1.6, 95% CI 1.3–1.9
• With moderate bleeding—HR 2.7, 95% CI 2.3–3.4
• With severe bleeding—HR 10.6, 95% CI 8.3–13.6.

The pattern was the same for death within 6 months:

• With mild bleeding—HR 1.4, 95% CI 1.2–1.6
• With moderate bleeding—HR 2.1, 95% CI 1.8–2.4
• With severe bleeding, HR 7.5, 95% CI 6.1–9.3.

These findings were confirmed by Eikelboom et al¹⁵ in 34,146 patients with acute coronary syndromes in the OASIS registry, the OASIS-2 trial, and the CURE randomized trial. In the first 30 days, five times as many patients died (12.8% vs 2.5%; P < .0009) among those who developed major bleeding compared with those who did not. These investigators defined major bleeding as bleeding that was life-threatening or significantly disabling or that required transfusion of two or more units of packed red blood cells.

A higher risk of nonfatal adverse events. Bleeding after antithrombotic therapy for non-ST-elevation acute coronary syndromes has also been associated with nonfatal adverse events such as stroke and stent thrombosis.

For example, in the study by Eikelboom et al,¹⁵ major bleeding was associated with a higher risk of recurrent ischemic events. Approximately 1 in 5 patients in the OASIS trials who developed major bleeding during the first 30 days died or had a myocardial infarction or stroke by 30 days, compared with 1 in 20 of those who did not develop major bleeding during the first 30 days. However, after events that occurred during the first 30 days were excluded, the association between major bleeding and both myocardial infarction and stroke was no longer evident between 30 days and 6 months.

Manoukian et al¹⁶ evaluated the impact of major bleeding in 13,819 patients with highrisk acute coronary syndromes undergoing treatment with an early invasive strategy in the ACUITY trial. At 30 days, patients with major bleeding had higher rates of the composite end point of death, myocardial infarction, or unplanned revascularization for ischemia (23.1% vs 6.8%, P < .0001) and of stent thrombosis (3.4% vs 0.6%, P < .0001).

Long-term outcomes

The association between bleeding and adverse outcomes persists in the long term as well, although the mechanisms underlying this association are not well studied.

Kinnaird et al¹⁷ examined the data from 10,974 unselected patients who underwent percutaneous coronary intervention. At 1 year, the following percentages of patients had died:

• After TIMI major bleeding—17.2%
• After TIMI minor bleeding—9.1%
• After no bleeding—5.5%.

However, after adjustment for potential confounders, only transfusion remained a significant predictor of 1-year mortality.

Mehran et al¹⁸ evaluated 1-year mortality data from the ACUITY trial. Compared with the rate in patients who had no major bleeding and no myocardial infarction, the hazard ratios for death were:

• After major bleeding—HR 3.5, 95% CI 2.7–4.4
• After myocardial infarction—HR 3.1, 95% CI 2.4–3.9.

Interestingly, the risk of death associated with myocardial infarction abated after 7 days, while the risk associated with bleeding persisted beyond 30 days and remained constant throughout the first year following the bleeding event.

Similarly, Ndrepepa and colleagues¹⁹ examined pooled data from four ISAR trials using the TIMI bleeding scale and found that myocardial infarction, target vessel revascularization, and major bleeding all had similar discriminatory ability at predicting 1-year mortality.

In patients undergoing elective or urgent percutaneous coronary intervention in the REPLACE-2 trial,²⁰ independent predictors of death by 1 year were²¹:

• Major hemorrhage (OR 2.66, 95% CI 1.44–4.92)
• Periprocedural myocardial infarction (OR 2.46, 95% CI 1.44–4.20).

THEORIES OF HOW BLEEDING MAY CAUSE ADVERSE OUTCOMES

Several mechanisms have been proposed to explain the association between bleeding during treatment for acute coronary syndromes and adverse clinical outcomes.^13,22

The immediate effects of bleeding are thought to be hypotension and a reflex hyperadrenergic state to compensate for the loss of intravascular volume.²³ This physiologic response is believed to contribute to myocardial ischemia by further decreasing myocardial oxygen supply in obstructive coronary disease.

Trying to minimize blood loss, physicians may withhold anticoagulation and antiplatelet therapy, which in turn may lead to further ischemia.²⁴ To compensate for blood loss, physicians may also resort to blood transfusion. However, depletion of 2,3-diphosphoglycerate and nitric oxide in stored donor red blood cells is postulated to reduce oxygen delivery by increasing hemoglobin’s affinity for oxygen, leading to induced microvascular obstruction and adverse inflammatory reactions.^15,25

Recent data have also begun to elucidate the long-term effects of bleeding during acute coronary syndrome management. Patients with anemia during the acute phase of infarction have greater neurohormonal activation.²⁶ These adaptive responses to anemia may lead to eccentric left ventricular remodeling that may lead to higher oxygen consumption, increased diastolic wall stress, interstitial fibrosis, and accelerated myocyte loss.^27–30

Nevertheless, we must point out that although strong associations between bleeding and adverse outcomes have been established, direct causality has not.

 

TO PREVENT BLEEDING, START BY ASSESSING RISK

Preventing bleeding is a key step in balancing the safety and efficacy of aggressive management of non-ST-elevation acute coronary syndromes. Current guidelines^1,31 call for assessing the risk of both thrombosis and bleeding in patients presenting with these syndromes (Figure 1). Doing so may allow clinicians to tailor therapy by adjusting the treatment regimen in patients at risk of bleeding to include medications associated with favorable bleeding profiles and by using radial access as the point of entry at the time of coronary artery angiography.

The CRUSADE bleeding risk score

The CRUSADE bleeding score (calculator available at http://www.crusadebleedingscore.org/) was developed and validated in more than 89,000 community-treated patients with non-ST-elevation acute coronary syndromes.⁸ It is based on eight variables:

• Sex (higher risk in women)
• History of diabetes (higher risk)
• Prior vascular disease (higher risk)
• Heart rate (the higher the rate, the higher the risk)
• Systolic blood pressure (higher risk with pressures above or below the 121–180 mm Hg range)
• Signs of congestive heart failure (higher risk)
• Baseline hematocrit (the lower the hematocrit, the higher the risk)
• Creatinine clearance (by the Cockcroft-Gault formula; the lower the creatinine clearance, the higher the risk).

Patients who are found to have bleeding scores suggesting a moderate or higher risk of bleeding should be considered for medications associated with a favorable bleeding profile, and for radial access at the time of coronary angiography. Scores are graded as follows⁸:

• < 21: Very low risk
• 21–30: Low risk
• 31–40: Moderate risk
• 41–50: High risk
• > 50: Very high risk.

The CRUSADE bleeding score is unique in that, unlike earlier risk stratification tools, it was developed in a “real world” population, not in subgroups or in a clinical trial. It can be calculated at baseline to help guide the selection of treatment.⁸

Adjusting the heparin regimen in patients at risk of bleeding

Both the joint American College of Cardiology/American Heart Association¹ and the European Society of Cardiology guidelines³¹ for the treatment of non-ST-elevation acute coronary syndromes recommend taking steps to prevent bleeding, such as adjusting the dosage of unfractionated heparin, using safer drugs, reducing the duration of antithrombotic treatment, and using combinations of antithrombotic and antiplatelet agents according to proven indications.³¹

In the CRUSADE registry, 42% of patients with non-ST-elevation acute coronary syndromes received at least one initial dose of antithrombotic drug outside the recommended range, resulting in an estimated 15% excess of bleeding events.³² Thus, proper dosing is a target for prevention.

Appropriate antithrombotic dosing takes into account the patient’s age, weight, and renal function. However, heparin dosage in the current guidelines¹ is based on weight only: a loading dose of 60 U/kg (maximum 4,000 U) by intravenous bolus, then 12 U/kg/hour (maximum 1,000 U/hour) to maintain an activated partial thromboplastin time of 50 to 70 seconds.¹

Renal dysfunction is particularly worrisome in patients with non-ST-elevation acute coronary syndromes because it is associated with higher rates of major bleeding and death. In the OASIS-5 trial,³³ the overall risk of death was approximately five times higher in patients in the lowest quartile of renal function (glomerular filtration rate < 58 mL/min/1.73 m²) than in the highest quartile (glomerular filtration rate ≥ 86 mL/min/1.73 m²).

Renal function must be evaluated not only on admission but also throughout the hospital stay. Patients presenting with acute coronary syndromes often experience fluctuations in renal function that would call for adjustment of heparin dosing, either increasing the dose to maximize the drug’s efficacy if renal function is recovering or decreasing the dose to prevent bleeding if renal function is deteriorating.

Clopidogrel vs prasugrel

Certain medications should be avoided when the risk of bleeding outweighs any potential benefit in terms of ischemia.

For example, in a randomized trial,³⁴ prasugrel (Effient), a potent thienopyridine, was associated with a significantly lower rate of the composite end point of stroke, myocardial infarction, or death than clopidogrel (Plavix) in patients with acute coronary syndromes undergoing percutaneous coronary interventions. However, it did not seem to offer any advantage in patients 75 years old and older, those with prior stroke or transient ischemic attack, or those weighing less than 60 kg, and it posed a substantially higher risk of bleeding.

With clopidogrel, the risk of acute bleeding is primarily in patients who undergo coronary artery bypass grafting within 5 days of receiving a dose.^35,36 Therefore, clopidogrel should be stopped 5 to 7 days before bypass surgery.¹ Importantly, there is no increased risk of recurrent ischemic events during this 5-day waiting period in patients who receive clopidogrel early. Therefore, the recommendation to stop clopidogrel before surgery does not negate the benefits of early treatment.³⁶

Lower-risk drugs: Fondaparinux and bivalirudin

At this time, only two agents have been studied in clinical trials that have specifically focused on reducing bleeding risk: fondaparinux (Arixtra) and bivalirudin (Angiomax).^20,37–39

Fondaparinux

OASIS-5 was a randomized, double-blind trial that compared fondaparinux and enoxaparin (Lovenox) in patients with acute coronary syndromes.³⁸ Fondaparinux was similar to enoxaparin in terms of the combined end point of death, myocardial infarction, or refractory ischemia at 9 days, and fewer patients on fondaparinux developed bleeding (2.2% vs 4.1%, HR 0.52; 95% CI 0.44–0.61). This difference persisted during long-term follow-up.

Importantly, fewer patients died in the fondaparinux group. At 180 days, 638 (6.5%) of the patients in the enoxaparin group had died, compared with 574 (5.8%) in the fondaparinux group, a difference of 64 deaths (P = .05). The authors found that 41 fewer patients in the fondaparinux group than in the enoxaparin group died after major bleeding, and 20 fewer patients in the fondaparinux group died after minor bleeding.³⁸ Thus, most of the difference in mortality rates between the two groups was attributed to a lower incidence of bleeding with fondaparinux.

Unfortunately, despite its safe bleeding profile, fondaparinux has fallen out of favor for use in acute coronary syndromes, owing to a higher risk of catheter thrombosis in the fondaparinux group (0.9%) than in those undergoing percutaneous coronary interventions with enoxaparin alone (0.4%) in the OASIS-5 trial.⁴⁰

 

 

Bivalirudin

The direct thrombin inhibitor bivalirudin has been studied in three large randomized trials in patients undergoing percutaneous coronary interventions.^20,37,41

The ACUITY trial³⁷ was a prospective, open-label, randomized, multicenter trial that compared three regimens in patients with moderate or high-risk non-ST-elevation acute coronary syndromes:

• Heparin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin alone.

Bivalirudin alone was as effective as heparin plus a glycoprotein IIb/IIIa inhibitor with respect to the composite ischemia end point, which at 30 days had occurred in 7.8% vs 7.3% of the patients in these treatment groups (P = .32, RR 1.08; 95% CI 0.93–1.24), and it was superior with respect to major bleeding (3.0% vs 5.7%, P < .001, RR 0.53; 95% CI 0.43–0.65).

The HORIZONS-AMI study⁴¹ was a prospective, open-label, randomized, multicenter trial that compared bivalirudin alone vs heparin plus a glycoprotein IIb/IIIa inhibitor in patients with ST-elevation acute coronary syndromes who were undergoing primary percutaneous coronary interventions. The two primary end points were major bleeding and net adverse events.

At 1 year, patients assigned to bivalirudin had a lower rate of major bleeding than did controls (5.8% vs 9.2%, HR 0.61, 95% CI 0.48–0.78, P < .0001), with similar rates of major adverse cardiac events in both groups (11.9% vs 11.9%, HR 1.00, 95% CI 0.82– 1.21, P = .98).⁴¹

Both OASIS 5 and HORIZONS-AMI are examples of clinical trials in which strategies that reduced bleeding risk were also associated with improved survival.

For cardiac catheterization, inserting the catheter in the wrist poses less risk

Bleeding is currently the most common noncardiac complication in patients undergoing percutaneous coronary interventions, and it most often occurs at the vascular access site.¹⁷

Rao et al¹² evaluated data from 593,094 procedures in the National Cardiovascular Data Registry and found that, compared with the femoral approach, the use of transradial percutaneous coronary intervention was associated with a similar rate of procedural success (OR 1.02, 95% CI 0.93–1.12) but a significantly lower risk of bleeding complications (OR 0.42, 95% CI 0.31–0.56) after multivariable adjustment.

The use of smaller sheath sizes (4F–6F) and preferential use of bivalirudin over unfractionated heparin and glycoprotein IIb/IIIa inhibitor therapy are other methods described to decrease the risk of bleeding after percutaneous coronary interventions.^20,41–49

IF BLEEDING OCCURS

Once a bleeding complication occurs, cessation of therapy is a potential option. Stopping or reversing antithrombotic and antiplatelet therapy is warranted in the event of major bleeding (eg, gastrointestinal, retroperitoneal, intracranial).³¹

Stopping antithrombotic and antiplatelet therapy

Whether bleeding is minor or major, the risk of a recurrent thrombotic event must be considered, especially in patients who have undergone revascularization, stent implantation, or both. The risk of acute thrombotic events after interrupting antithrombotic or antiplatelet agents is considered greatest 4 to 5 days following revascularization or percutaneous coronary intervention.¹⁵ If bleeding can be controlled with local treatment such as pressure, packing, or dressing, antithrombotic and antiplatelet therapy need not be interrupted.⁵⁰

Current guidelines recommend strict control of hemorrhage for at least 24 hours before reintroducing antiplatelet or antithrombotic agents.

It is also important to remember that in the setting of gastrointestinal bleeding due to peptic ulcer disease, adjunctive proton pump inhibitors are recommended after restarting antiplatelet or antithrombotic therapy or both.

Importantly, evidence-based antithrombotic medications (especially dual antiplatelet therapy) should be restarted once the acute bleeding event has resolved.³¹

Reversal of anticoagulant and antiplatelet therapies

Reversal of antithrombotic therapy is occasionally necessary (Table 2).

Unfractionated heparin is reversed with infusion of protamine sulfate at a dose of 1 mg per 100 U of unfractionated heparin given over the previous 4 hours.^51,52 The rate of protamine sulfate infusion should be less than 100 mg over 2 hours, with 50% of the dose given initially and subsequent doses titrated according to bleeding response.^52,53 Protamine sulfate is associated with a risk of hypotension and bradycardia, and for this reason it should be given no faster than 5 mg/min.

Low-molecular-weight heparin (LMWH) can be inhibited by 1 mg of protamine sulfate for each 1 mg of LMWH given over the previous 4 hours.^51,52

However, protamine sulfate only partially neutralizes the anticoagulant effect of LMWH. In cases in which protamine sulfate is unsuccessful in abating bleeding associated with LMWH use, guidelines allow for the use of recombinant factor VIIa (NovoSeven).³¹ In healthy volunteers given fondaparinux, recombinant factor VIIa normalized coagulation times and thrombin generation within 1.5 hours, with a sustained effect for 6 hours.⁵²

It is important to note that the use of this agent has not been fully studied, it is very costly (a single dose of 40 μg/kg costs from $3,000 to $4,000), and it is linked to reports of increased risk of thrombotic complications.^54,55

Antiplatelet agents are more complex to reverse. The antiplatelet actions of aspirin and clopidogrel wear off as new platelets are produced. Approximately 10% of a patient’s platelet count is produced daily; thus, the antiplatelet effects of aspirin and clopidogrel can persist for 5 to 10 days.^31,56

If these agents need to be reversed quickly to stop bleeding, according to expert consensus the aspirin effect can be reversed by transfusion of one unit of platelets. The antiplatelet effect of clopidogrel is more significant than that of aspirin; thus, two units of platelets are recommended.⁵⁶

Glycoprotein IIb/IIIa inhibitors. If a major bleeding event requires the reversal of glycoprotein IIb/IIIa inhibitor therapy, the treatment must take into consideration the pharmacodynamics of the target drug. Both eptifibatide (Integrilin) and tirofiban (Aggrastat) competitively inhibit glycoprotein IIb/IIIa receptors; thus, their effects depend on dosing, elimination, and time. Due to the stoichiometry of both drugs, transfusion of platelets is ineffective. Both eptifibatide and tirofiban are eliminated by the kidney; thus, normal renal function is key to the amount of time it takes for platelet function to return to baseline.⁵⁷ Evidence suggests that fibrinogen-rich plasma can be administered to restore platelet function.^31,58,59

Abciximab (ReoPro). Whereas reversal of eptifibatide and tirofiban focuses on overcoming competitive inhibition, neutralization of abciximab involves overcoming its high receptor affinity. At 24 hours after abciximab infusion is stopped, platelet aggregation may still be inhibited by up to 50%. Fortunately, owing to abciximab’s short plasma half-life and its dilution in serum, platelet transfusion is effective in reversing its antiplatelet effects.^31,57

 

 

Blood transfusion

Long considered beneficial to critically ill patients, blood transfusion to maintain hematocrit levels during acute coronary syndromes has come under intense scrutiny. Randomized trials have shown that transfusion should not be given aggressively to critically ill patients.⁶⁰ In acute coronary syndromes, there are only observational data.

Rao et al⁶¹ used detailed clinical data from 24,112 patients with acute coronary syndromes in the GUSTO IIb, PURSUIT, and PARAGON B trials to determine the association between blood transfusion and outcomes in patients who developed moderate to severe bleeding, anemia, or both during their hospitalization. The rates of death in the hospital and at 30 days were significantly higher in patients who received a transfusion (30-day mortality HR 3.94; 95% CI 3.36–4.75). However, there was no significant association between transfusion and the 30-day mortality rate if the nadir hematocrit was 25% or less.

Of note: no randomized clinical trial has evaluated transfusion strategies in acute coronary syndromes at this time. Until such data are available, it is reasonable to follow published guidelines and to avoid transfusion in stable patients with ischemic heart disease unless the hematocrit is 25% or less.³¹ The risks and benefits of blood transfusion should be carefully weighed. Routine use of transfusion to maintain predefined hemoglobin levels is not recommended in stable patients.

The medical management of non-ST-elevation acute coronary syndromes focuses on blocking the coagulation cascade and inhibiting platelets. This—plus diagnostic angiography followed, if needed, by revascularization—has reduced the rates of death and recurrent ischemic events.¹ However, the combination of potent antithrombotic drugs and invasive procedures also increases the risk of bleeding.

This review discusses the incidence and complications associated with bleeding during the treatment of acute coronary syndromes and summarizes recommendations for preventing and managing bleeding in this setting.

THE TRUE INCIDENCE OF BLEEDING IS HARD TO DETERMINE

The optimal way to detect and analyze bleeding events in clinical trials and registries is highly debated. The reported incidences of bleeding during antithrombotic and antiplatelet therapy for non-ST-elevation acute coronary syndromes depend on how bleeding was defined, how the acute coronary syndromes were treated, and on other factors such as how the study was designed.

How was bleeding defined?

The first bleeding classification schemes were the GUSTO² and the TIMI³ scales (Table 1), both of which were developed for studies of thrombolytic therapy for ST-elevation myocardial infarction. The GUSTO classification is based on clinical events and categorizes bleeding as severe, moderate, or mild. In contrast, the TIMI classification is based on laboratory values and categorizes bleeding as major, moderate, or minor.

Since these classification schemes are based on different types of data, they yield different numbers when applied to the same study population. For instance, Rao et al⁴ pooled the data from the PURSUIT and PARAGON B trials (15,454 patients in all) and found that the incidence of severe bleeding (by the GUSTO criteria) was 1.2%, while the rate of major bleeding (by the TIMI criteria) was 8.2%.

What was the treatment strategy?

Another reason that the true incidence of bleeding is hard to determine is that different studies used treatment strategies that differed in the type, timing, and dose of antithrombotic agents and whether invasive procedures were used early. For example, if unfractionated heparin is used aggressively in regimens that are not adjusted for weight and with a higher target for the activated clotting time, the risk of bleeding is higher than with conservative dosing.^5–7

Subherwal et al⁸ evaluated the effect of treatment strategy on the incidence of bleeding in patients with non-ST-elevation acute coronary syndromes who received two or more antithrombotic drugs in the CRUSADE Quality Improvement Initiative. The risk of bleeding was higher with an invasive approach (catheterization) than with a conservative approach (no catheterization), regardless of baseline bleeding risk.

What type of study was it?

Another source of variation is the design of the study. Registries differ from clinical trials in patient characteristics and in the way data are gathered (prospectively vs retrospectively).

In registries, data are often collected retrospectively, whereas in clinical trials the data are prospectively collected. For this reason, the definition of bleeding in registries is often based on events that are easily identified through chart review, such as transfusion. This may lead to a lower reported rate of bleeding, since other, less serious bleeding events such as access-site hematomas and epistaxis may not be documented in the medical record.

On the other hand, registries often include older and sicker patients, who may be more prone to bleeding and who are often excluded from clinical trials. This may lead to a higher rate of reported bleeding.⁹

Where the study was conducted makes a difference as well, owing to regional practice differences. For example, Moscucci et al¹⁰ reported that the incidence of major bleeding in 24,045 patients with non-ST-elevation acute coronary syndromes in the GRACE registry (in 14 countries worldwide) was 3.9%. In contrast, Yang et al¹¹ reported that the rate of bleeding in the CRUSADE registry (in the United States) was 10.3%.

This difference was partly influenced by different definitions of bleeding. The GRACE registry defined major bleeding as life-threatening events requiring transfusion of two or more units of packed red blood cells, or resulting in an absolute decrease in the hematocrit of 10% or more or death, or hemorrhagic subdural hematoma. In contrast, the CRUSADE data reflect bleeding requiring transfusion. However, practice patterns such as greater use of invasive procedures in the United States may also be responsible.

Rao and colleagues¹² examined international variation in blood transfusion rates among patients with acute coronary syndromes. Patients outside the United States were significantly less likely to receive transfusions, even after adjusting for patient and practice differences.

Taking these confounders into account, it is reasonable to estimate that the frequency of bleeding in patients with non-ST-elevation acute coronary syndromes ranges from less than 1% to 10%.¹³

 

BLEEDING IS ASSOCIATED WITH POOR OUTCOMES

Regardless of the definition or the data source, hemorrhagic complications are associated with a higher risk of death and nonfatal adverse events, both in the short term and in the long term.

Short-term outcomes

A higher risk of death. In the GRACE registry study by Moscucci et al¹⁰ discussed above, patients who had major bleeding were significantly more likely to die during their hospitalization than those who did not (odds ratio [OR] 1.64, 95% confidence interval [CI] 1.18–2.28).

Rao et al¹⁴ evaluated pooled data from the multicenter international GUSTO IIb, PURSUIT, and PARAGON A and B trials and found that the effects of bleeding in non-ST-elevation acute coronary syndromes extended beyond the hospital stay. The more severe the bleeding (by the GUSTO criteria), the greater the adjusted hazard ratio (HR) for death within 30 days:

• With mild bleeding—HR 1.6, 95% CI 1.3–1.9
• With moderate bleeding—HR 2.7, 95% CI 2.3–3.4
• With severe bleeding—HR 10.6, 95% CI 8.3–13.6.

The pattern was the same for death within 6 months:

• With mild bleeding—HR 1.4, 95% CI 1.2–1.6
• With moderate bleeding—HR 2.1, 95% CI 1.8–2.4
• With severe bleeding, HR 7.5, 95% CI 6.1–9.3.

These findings were confirmed by Eikelboom et al¹⁵ in 34,146 patients with acute coronary syndromes in the OASIS registry, the OASIS-2 trial, and the CURE randomized trial. In the first 30 days, five times as many patients died (12.8% vs 2.5%; P < .0009) among those who developed major bleeding compared with those who did not. These investigators defined major bleeding as bleeding that was life-threatening or significantly disabling or that required transfusion of two or more units of packed red blood cells.

A higher risk of nonfatal adverse events. Bleeding after antithrombotic therapy for non-ST-elevation acute coronary syndromes has also been associated with nonfatal adverse events such as stroke and stent thrombosis.

For example, in the study by Eikelboom et al,¹⁵ major bleeding was associated with a higher risk of recurrent ischemic events. Approximately 1 in 5 patients in the OASIS trials who developed major bleeding during the first 30 days died or had a myocardial infarction or stroke by 30 days, compared with 1 in 20 of those who did not develop major bleeding during the first 30 days. However, after events that occurred during the first 30 days were excluded, the association between major bleeding and both myocardial infarction and stroke was no longer evident between 30 days and 6 months.

Manoukian et al¹⁶ evaluated the impact of major bleeding in 13,819 patients with highrisk acute coronary syndromes undergoing treatment with an early invasive strategy in the ACUITY trial. At 30 days, patients with major bleeding had higher rates of the composite end point of death, myocardial infarction, or unplanned revascularization for ischemia (23.1% vs 6.8%, P < .0001) and of stent thrombosis (3.4% vs 0.6%, P < .0001).

Long-term outcomes

The association between bleeding and adverse outcomes persists in the long term as well, although the mechanisms underlying this association are not well studied.

Kinnaird et al¹⁷ examined the data from 10,974 unselected patients who underwent percutaneous coronary intervention. At 1 year, the following percentages of patients had died:

• After TIMI major bleeding—17.2%
• After TIMI minor bleeding—9.1%
• After no bleeding—5.5%.

However, after adjustment for potential confounders, only transfusion remained a significant predictor of 1-year mortality.

Mehran et al¹⁸ evaluated 1-year mortality data from the ACUITY trial. Compared with the rate in patients who had no major bleeding and no myocardial infarction, the hazard ratios for death were:

• After major bleeding—HR 3.5, 95% CI 2.7–4.4
• After myocardial infarction—HR 3.1, 95% CI 2.4–3.9.

Interestingly, the risk of death associated with myocardial infarction abated after 7 days, while the risk associated with bleeding persisted beyond 30 days and remained constant throughout the first year following the bleeding event.

Similarly, Ndrepepa and colleagues¹⁹ examined pooled data from four ISAR trials using the TIMI bleeding scale and found that myocardial infarction, target vessel revascularization, and major bleeding all had similar discriminatory ability at predicting 1-year mortality.

In patients undergoing elective or urgent percutaneous coronary intervention in the REPLACE-2 trial,²⁰ independent predictors of death by 1 year were²¹:

• Major hemorrhage (OR 2.66, 95% CI 1.44–4.92)
• Periprocedural myocardial infarction (OR 2.46, 95% CI 1.44–4.20).

THEORIES OF HOW BLEEDING MAY CAUSE ADVERSE OUTCOMES

Several mechanisms have been proposed to explain the association between bleeding during treatment for acute coronary syndromes and adverse clinical outcomes.^13,22

The immediate effects of bleeding are thought to be hypotension and a reflex hyperadrenergic state to compensate for the loss of intravascular volume.²³ This physiologic response is believed to contribute to myocardial ischemia by further decreasing myocardial oxygen supply in obstructive coronary disease.

Trying to minimize blood loss, physicians may withhold anticoagulation and antiplatelet therapy, which in turn may lead to further ischemia.²⁴ To compensate for blood loss, physicians may also resort to blood transfusion. However, depletion of 2,3-diphosphoglycerate and nitric oxide in stored donor red blood cells is postulated to reduce oxygen delivery by increasing hemoglobin’s affinity for oxygen, leading to induced microvascular obstruction and adverse inflammatory reactions.^15,25

Recent data have also begun to elucidate the long-term effects of bleeding during acute coronary syndrome management. Patients with anemia during the acute phase of infarction have greater neurohormonal activation.²⁶ These adaptive responses to anemia may lead to eccentric left ventricular remodeling that may lead to higher oxygen consumption, increased diastolic wall stress, interstitial fibrosis, and accelerated myocyte loss.^27–30

Nevertheless, we must point out that although strong associations between bleeding and adverse outcomes have been established, direct causality has not.

 

TO PREVENT BLEEDING, START BY ASSESSING RISK

Preventing bleeding is a key step in balancing the safety and efficacy of aggressive management of non-ST-elevation acute coronary syndromes. Current guidelines^1,31 call for assessing the risk of both thrombosis and bleeding in patients presenting with these syndromes (Figure 1). Doing so may allow clinicians to tailor therapy by adjusting the treatment regimen in patients at risk of bleeding to include medications associated with favorable bleeding profiles and by using radial access as the point of entry at the time of coronary artery angiography.

The CRUSADE bleeding risk score

The CRUSADE bleeding score (calculator available at http://www.crusadebleedingscore.org/) was developed and validated in more than 89,000 community-treated patients with non-ST-elevation acute coronary syndromes.⁸ It is based on eight variables:

• Sex (higher risk in women)
• History of diabetes (higher risk)
• Prior vascular disease (higher risk)
• Heart rate (the higher the rate, the higher the risk)
• Systolic blood pressure (higher risk with pressures above or below the 121–180 mm Hg range)
• Signs of congestive heart failure (higher risk)
• Baseline hematocrit (the lower the hematocrit, the higher the risk)
• Creatinine clearance (by the Cockcroft-Gault formula; the lower the creatinine clearance, the higher the risk).

Patients who are found to have bleeding scores suggesting a moderate or higher risk of bleeding should be considered for medications associated with a favorable bleeding profile, and for radial access at the time of coronary angiography. Scores are graded as follows⁸:

• < 21: Very low risk
• 21–30: Low risk
• 31–40: Moderate risk
• 41–50: High risk
• > 50: Very high risk.

The CRUSADE bleeding score is unique in that, unlike earlier risk stratification tools, it was developed in a “real world” population, not in subgroups or in a clinical trial. It can be calculated at baseline to help guide the selection of treatment.⁸

Adjusting the heparin regimen in patients at risk of bleeding

Both the joint American College of Cardiology/American Heart Association¹ and the European Society of Cardiology guidelines³¹ for the treatment of non-ST-elevation acute coronary syndromes recommend taking steps to prevent bleeding, such as adjusting the dosage of unfractionated heparin, using safer drugs, reducing the duration of antithrombotic treatment, and using combinations of antithrombotic and antiplatelet agents according to proven indications.³¹

In the CRUSADE registry, 42% of patients with non-ST-elevation acute coronary syndromes received at least one initial dose of antithrombotic drug outside the recommended range, resulting in an estimated 15% excess of bleeding events.³² Thus, proper dosing is a target for prevention.

Appropriate antithrombotic dosing takes into account the patient’s age, weight, and renal function. However, heparin dosage in the current guidelines¹ is based on weight only: a loading dose of 60 U/kg (maximum 4,000 U) by intravenous bolus, then 12 U/kg/hour (maximum 1,000 U/hour) to maintain an activated partial thromboplastin time of 50 to 70 seconds.¹

Renal dysfunction is particularly worrisome in patients with non-ST-elevation acute coronary syndromes because it is associated with higher rates of major bleeding and death. In the OASIS-5 trial,³³ the overall risk of death was approximately five times higher in patients in the lowest quartile of renal function (glomerular filtration rate < 58 mL/min/1.73 m²) than in the highest quartile (glomerular filtration rate ≥ 86 mL/min/1.73 m²).

Renal function must be evaluated not only on admission but also throughout the hospital stay. Patients presenting with acute coronary syndromes often experience fluctuations in renal function that would call for adjustment of heparin dosing, either increasing the dose to maximize the drug’s efficacy if renal function is recovering or decreasing the dose to prevent bleeding if renal function is deteriorating.

Clopidogrel vs prasugrel

Certain medications should be avoided when the risk of bleeding outweighs any potential benefit in terms of ischemia.

For example, in a randomized trial,³⁴ prasugrel (Effient), a potent thienopyridine, was associated with a significantly lower rate of the composite end point of stroke, myocardial infarction, or death than clopidogrel (Plavix) in patients with acute coronary syndromes undergoing percutaneous coronary interventions. However, it did not seem to offer any advantage in patients 75 years old and older, those with prior stroke or transient ischemic attack, or those weighing less than 60 kg, and it posed a substantially higher risk of bleeding.

With clopidogrel, the risk of acute bleeding is primarily in patients who undergo coronary artery bypass grafting within 5 days of receiving a dose.^35,36 Therefore, clopidogrel should be stopped 5 to 7 days before bypass surgery.¹ Importantly, there is no increased risk of recurrent ischemic events during this 5-day waiting period in patients who receive clopidogrel early. Therefore, the recommendation to stop clopidogrel before surgery does not negate the benefits of early treatment.³⁶

Lower-risk drugs: Fondaparinux and bivalirudin

At this time, only two agents have been studied in clinical trials that have specifically focused on reducing bleeding risk: fondaparinux (Arixtra) and bivalirudin (Angiomax).^20,37–39

Fondaparinux

OASIS-5 was a randomized, double-blind trial that compared fondaparinux and enoxaparin (Lovenox) in patients with acute coronary syndromes.³⁸ Fondaparinux was similar to enoxaparin in terms of the combined end point of death, myocardial infarction, or refractory ischemia at 9 days, and fewer patients on fondaparinux developed bleeding (2.2% vs 4.1%, HR 0.52; 95% CI 0.44–0.61). This difference persisted during long-term follow-up.

Importantly, fewer patients died in the fondaparinux group. At 180 days, 638 (6.5%) of the patients in the enoxaparin group had died, compared with 574 (5.8%) in the fondaparinux group, a difference of 64 deaths (P = .05). The authors found that 41 fewer patients in the fondaparinux group than in the enoxaparin group died after major bleeding, and 20 fewer patients in the fondaparinux group died after minor bleeding.³⁸ Thus, most of the difference in mortality rates between the two groups was attributed to a lower incidence of bleeding with fondaparinux.

Unfortunately, despite its safe bleeding profile, fondaparinux has fallen out of favor for use in acute coronary syndromes, owing to a higher risk of catheter thrombosis in the fondaparinux group (0.9%) than in those undergoing percutaneous coronary interventions with enoxaparin alone (0.4%) in the OASIS-5 trial.⁴⁰

 

 

Bivalirudin

The direct thrombin inhibitor bivalirudin has been studied in three large randomized trials in patients undergoing percutaneous coronary interventions.^20,37,41

The ACUITY trial³⁷ was a prospective, open-label, randomized, multicenter trial that compared three regimens in patients with moderate or high-risk non-ST-elevation acute coronary syndromes:

• Heparin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin alone.

Bivalirudin alone was as effective as heparin plus a glycoprotein IIb/IIIa inhibitor with respect to the composite ischemia end point, which at 30 days had occurred in 7.8% vs 7.3% of the patients in these treatment groups (P = .32, RR 1.08; 95% CI 0.93–1.24), and it was superior with respect to major bleeding (3.0% vs 5.7%, P < .001, RR 0.53; 95% CI 0.43–0.65).

The HORIZONS-AMI study⁴¹ was a prospective, open-label, randomized, multicenter trial that compared bivalirudin alone vs heparin plus a glycoprotein IIb/IIIa inhibitor in patients with ST-elevation acute coronary syndromes who were undergoing primary percutaneous coronary interventions. The two primary end points were major bleeding and net adverse events.

At 1 year, patients assigned to bivalirudin had a lower rate of major bleeding than did controls (5.8% vs 9.2%, HR 0.61, 95% CI 0.48–0.78, P < .0001), with similar rates of major adverse cardiac events in both groups (11.9% vs 11.9%, HR 1.00, 95% CI 0.82– 1.21, P = .98).⁴¹

Both OASIS 5 and HORIZONS-AMI are examples of clinical trials in which strategies that reduced bleeding risk were also associated with improved survival.

For cardiac catheterization, inserting the catheter in the wrist poses less risk

Bleeding is currently the most common noncardiac complication in patients undergoing percutaneous coronary interventions, and it most often occurs at the vascular access site.¹⁷

Rao et al¹² evaluated data from 593,094 procedures in the National Cardiovascular Data Registry and found that, compared with the femoral approach, the use of transradial percutaneous coronary intervention was associated with a similar rate of procedural success (OR 1.02, 95% CI 0.93–1.12) but a significantly lower risk of bleeding complications (OR 0.42, 95% CI 0.31–0.56) after multivariable adjustment.

The use of smaller sheath sizes (4F–6F) and preferential use of bivalirudin over unfractionated heparin and glycoprotein IIb/IIIa inhibitor therapy are other methods described to decrease the risk of bleeding after percutaneous coronary interventions.^20,41–49

IF BLEEDING OCCURS

Once a bleeding complication occurs, cessation of therapy is a potential option. Stopping or reversing antithrombotic and antiplatelet therapy is warranted in the event of major bleeding (eg, gastrointestinal, retroperitoneal, intracranial).³¹

Stopping antithrombotic and antiplatelet therapy

Whether bleeding is minor or major, the risk of a recurrent thrombotic event must be considered, especially in patients who have undergone revascularization, stent implantation, or both. The risk of acute thrombotic events after interrupting antithrombotic or antiplatelet agents is considered greatest 4 to 5 days following revascularization or percutaneous coronary intervention.¹⁵ If bleeding can be controlled with local treatment such as pressure, packing, or dressing, antithrombotic and antiplatelet therapy need not be interrupted.⁵⁰

Current guidelines recommend strict control of hemorrhage for at least 24 hours before reintroducing antiplatelet or antithrombotic agents.

It is also important to remember that in the setting of gastrointestinal bleeding due to peptic ulcer disease, adjunctive proton pump inhibitors are recommended after restarting antiplatelet or antithrombotic therapy or both.

Importantly, evidence-based antithrombotic medications (especially dual antiplatelet therapy) should be restarted once the acute bleeding event has resolved.³¹

Reversal of anticoagulant and antiplatelet therapies

Reversal of antithrombotic therapy is occasionally necessary (Table 2).

Unfractionated heparin is reversed with infusion of protamine sulfate at a dose of 1 mg per 100 U of unfractionated heparin given over the previous 4 hours.^51,52 The rate of protamine sulfate infusion should be less than 100 mg over 2 hours, with 50% of the dose given initially and subsequent doses titrated according to bleeding response.^52,53 Protamine sulfate is associated with a risk of hypotension and bradycardia, and for this reason it should be given no faster than 5 mg/min.

Low-molecular-weight heparin (LMWH) can be inhibited by 1 mg of protamine sulfate for each 1 mg of LMWH given over the previous 4 hours.^51,52

However, protamine sulfate only partially neutralizes the anticoagulant effect of LMWH. In cases in which protamine sulfate is unsuccessful in abating bleeding associated with LMWH use, guidelines allow for the use of recombinant factor VIIa (NovoSeven).³¹ In healthy volunteers given fondaparinux, recombinant factor VIIa normalized coagulation times and thrombin generation within 1.5 hours, with a sustained effect for 6 hours.⁵²

It is important to note that the use of this agent has not been fully studied, it is very costly (a single dose of 40 μg/kg costs from $3,000 to $4,000), and it is linked to reports of increased risk of thrombotic complications.^54,55

Antiplatelet agents are more complex to reverse. The antiplatelet actions of aspirin and clopidogrel wear off as new platelets are produced. Approximately 10% of a patient’s platelet count is produced daily; thus, the antiplatelet effects of aspirin and clopidogrel can persist for 5 to 10 days.^31,56

If these agents need to be reversed quickly to stop bleeding, according to expert consensus the aspirin effect can be reversed by transfusion of one unit of platelets. The antiplatelet effect of clopidogrel is more significant than that of aspirin; thus, two units of platelets are recommended.⁵⁶

Glycoprotein IIb/IIIa inhibitors. If a major bleeding event requires the reversal of glycoprotein IIb/IIIa inhibitor therapy, the treatment must take into consideration the pharmacodynamics of the target drug. Both eptifibatide (Integrilin) and tirofiban (Aggrastat) competitively inhibit glycoprotein IIb/IIIa receptors; thus, their effects depend on dosing, elimination, and time. Due to the stoichiometry of both drugs, transfusion of platelets is ineffective. Both eptifibatide and tirofiban are eliminated by the kidney; thus, normal renal function is key to the amount of time it takes for platelet function to return to baseline.⁵⁷ Evidence suggests that fibrinogen-rich plasma can be administered to restore platelet function.^31,58,59

Abciximab (ReoPro). Whereas reversal of eptifibatide and tirofiban focuses on overcoming competitive inhibition, neutralization of abciximab involves overcoming its high receptor affinity. At 24 hours after abciximab infusion is stopped, platelet aggregation may still be inhibited by up to 50%. Fortunately, owing to abciximab’s short plasma half-life and its dilution in serum, platelet transfusion is effective in reversing its antiplatelet effects.^31,57

 

 

Blood transfusion

Long considered beneficial to critically ill patients, blood transfusion to maintain hematocrit levels during acute coronary syndromes has come under intense scrutiny. Randomized trials have shown that transfusion should not be given aggressively to critically ill patients.⁶⁰ In acute coronary syndromes, there are only observational data.

Rao et al⁶¹ used detailed clinical data from 24,112 patients with acute coronary syndromes in the GUSTO IIb, PURSUIT, and PARAGON B trials to determine the association between blood transfusion and outcomes in patients who developed moderate to severe bleeding, anemia, or both during their hospitalization. The rates of death in the hospital and at 30 days were significantly higher in patients who received a transfusion (30-day mortality HR 3.94; 95% CI 3.36–4.75). However, there was no significant association between transfusion and the 30-day mortality rate if the nadir hematocrit was 25% or less.

Of note: no randomized clinical trial has evaluated transfusion strategies in acute coronary syndromes at this time. Until such data are available, it is reasonable to follow published guidelines and to avoid transfusion in stable patients with ischemic heart disease unless the hematocrit is 25% or less.³¹ The risks and benefits of blood transfusion should be carefully weighed. Routine use of transfusion to maintain predefined hemoglobin levels is not recommended in stable patients.

The medical management of non-ST-elevation acute coronary syndromes focuses on blocking the coagulation cascade and inhibiting platelets. This—plus diagnostic angiography followed, if needed, by revascularization—has reduced the rates of death and recurrent ischemic events.¹ However, the combination of potent antithrombotic drugs and invasive procedures also increases the risk of bleeding.

This review discusses the incidence and complications associated with bleeding during the treatment of acute coronary syndromes and summarizes recommendations for preventing and managing bleeding in this setting.

THE TRUE INCIDENCE OF BLEEDING IS HARD TO DETERMINE

The optimal way to detect and analyze bleeding events in clinical trials and registries is highly debated. The reported incidences of bleeding during antithrombotic and antiplatelet therapy for non-ST-elevation acute coronary syndromes depend on how bleeding was defined, how the acute coronary syndromes were treated, and on other factors such as how the study was designed.

How was bleeding defined?

The first bleeding classification schemes were the GUSTO² and the TIMI³ scales (Table 1), both of which were developed for studies of thrombolytic therapy for ST-elevation myocardial infarction. The GUSTO classification is based on clinical events and categorizes bleeding as severe, moderate, or mild. In contrast, the TIMI classification is based on laboratory values and categorizes bleeding as major, moderate, or minor.

Since these classification schemes are based on different types of data, they yield different numbers when applied to the same study population. For instance, Rao et al⁴ pooled the data from the PURSUIT and PARAGON B trials (15,454 patients in all) and found that the incidence of severe bleeding (by the GUSTO criteria) was 1.2%, while the rate of major bleeding (by the TIMI criteria) was 8.2%.

What was the treatment strategy?

Another reason that the true incidence of bleeding is hard to determine is that different studies used treatment strategies that differed in the type, timing, and dose of antithrombotic agents and whether invasive procedures were used early. For example, if unfractionated heparin is used aggressively in regimens that are not adjusted for weight and with a higher target for the activated clotting time, the risk of bleeding is higher than with conservative dosing.^5–7

Subherwal et al⁸ evaluated the effect of treatment strategy on the incidence of bleeding in patients with non-ST-elevation acute coronary syndromes who received two or more antithrombotic drugs in the CRUSADE Quality Improvement Initiative. The risk of bleeding was higher with an invasive approach (catheterization) than with a conservative approach (no catheterization), regardless of baseline bleeding risk.

What type of study was it?

Another source of variation is the design of the study. Registries differ from clinical trials in patient characteristics and in the way data are gathered (prospectively vs retrospectively).

In registries, data are often collected retrospectively, whereas in clinical trials the data are prospectively collected. For this reason, the definition of bleeding in registries is often based on events that are easily identified through chart review, such as transfusion. This may lead to a lower reported rate of bleeding, since other, less serious bleeding events such as access-site hematomas and epistaxis may not be documented in the medical record.

On the other hand, registries often include older and sicker patients, who may be more prone to bleeding and who are often excluded from clinical trials. This may lead to a higher rate of reported bleeding.⁹

Where the study was conducted makes a difference as well, owing to regional practice differences. For example, Moscucci et al¹⁰ reported that the incidence of major bleeding in 24,045 patients with non-ST-elevation acute coronary syndromes in the GRACE registry (in 14 countries worldwide) was 3.9%. In contrast, Yang et al¹¹ reported that the rate of bleeding in the CRUSADE registry (in the United States) was 10.3%.

This difference was partly influenced by different definitions of bleeding. The GRACE registry defined major bleeding as life-threatening events requiring transfusion of two or more units of packed red blood cells, or resulting in an absolute decrease in the hematocrit of 10% or more or death, or hemorrhagic subdural hematoma. In contrast, the CRUSADE data reflect bleeding requiring transfusion. However, practice patterns such as greater use of invasive procedures in the United States may also be responsible.

Rao and colleagues¹² examined international variation in blood transfusion rates among patients with acute coronary syndromes. Patients outside the United States were significantly less likely to receive transfusions, even after adjusting for patient and practice differences.

Taking these confounders into account, it is reasonable to estimate that the frequency of bleeding in patients with non-ST-elevation acute coronary syndromes ranges from less than 1% to 10%.¹³

 

BLEEDING IS ASSOCIATED WITH POOR OUTCOMES

Regardless of the definition or the data source, hemorrhagic complications are associated with a higher risk of death and nonfatal adverse events, both in the short term and in the long term.

Short-term outcomes

A higher risk of death. In the GRACE registry study by Moscucci et al¹⁰ discussed above, patients who had major bleeding were significantly more likely to die during their hospitalization than those who did not (odds ratio [OR] 1.64, 95% confidence interval [CI] 1.18–2.28).

Rao et al¹⁴ evaluated pooled data from the multicenter international GUSTO IIb, PURSUIT, and PARAGON A and B trials and found that the effects of bleeding in non-ST-elevation acute coronary syndromes extended beyond the hospital stay. The more severe the bleeding (by the GUSTO criteria), the greater the adjusted hazard ratio (HR) for death within 30 days:

• With mild bleeding—HR 1.6, 95% CI 1.3–1.9
• With moderate bleeding—HR 2.7, 95% CI 2.3–3.4
• With severe bleeding—HR 10.6, 95% CI 8.3–13.6.

The pattern was the same for death within 6 months:

• With mild bleeding—HR 1.4, 95% CI 1.2–1.6
• With moderate bleeding—HR 2.1, 95% CI 1.8–2.4
• With severe bleeding, HR 7.5, 95% CI 6.1–9.3.

These findings were confirmed by Eikelboom et al¹⁵ in 34,146 patients with acute coronary syndromes in the OASIS registry, the OASIS-2 trial, and the CURE randomized trial. In the first 30 days, five times as many patients died (12.8% vs 2.5%; P < .0009) among those who developed major bleeding compared with those who did not. These investigators defined major bleeding as bleeding that was life-threatening or significantly disabling or that required transfusion of two or more units of packed red blood cells.

A higher risk of nonfatal adverse events. Bleeding after antithrombotic therapy for non-ST-elevation acute coronary syndromes has also been associated with nonfatal adverse events such as stroke and stent thrombosis.

For example, in the study by Eikelboom et al,¹⁵ major bleeding was associated with a higher risk of recurrent ischemic events. Approximately 1 in 5 patients in the OASIS trials who developed major bleeding during the first 30 days died or had a myocardial infarction or stroke by 30 days, compared with 1 in 20 of those who did not develop major bleeding during the first 30 days. However, after events that occurred during the first 30 days were excluded, the association between major bleeding and both myocardial infarction and stroke was no longer evident between 30 days and 6 months.

Manoukian et al¹⁶ evaluated the impact of major bleeding in 13,819 patients with highrisk acute coronary syndromes undergoing treatment with an early invasive strategy in the ACUITY trial. At 30 days, patients with major bleeding had higher rates of the composite end point of death, myocardial infarction, or unplanned revascularization for ischemia (23.1% vs 6.8%, P < .0001) and of stent thrombosis (3.4% vs 0.6%, P < .0001).

Long-term outcomes

The association between bleeding and adverse outcomes persists in the long term as well, although the mechanisms underlying this association are not well studied.

Kinnaird et al¹⁷ examined the data from 10,974 unselected patients who underwent percutaneous coronary intervention. At 1 year, the following percentages of patients had died:

• After TIMI major bleeding—17.2%
• After TIMI minor bleeding—9.1%
• After no bleeding—5.5%.

However, after adjustment for potential confounders, only transfusion remained a significant predictor of 1-year mortality.

Mehran et al¹⁸ evaluated 1-year mortality data from the ACUITY trial. Compared with the rate in patients who had no major bleeding and no myocardial infarction, the hazard ratios for death were:

• After major bleeding—HR 3.5, 95% CI 2.7–4.4
• After myocardial infarction—HR 3.1, 95% CI 2.4–3.9.

Interestingly, the risk of death associated with myocardial infarction abated after 7 days, while the risk associated with bleeding persisted beyond 30 days and remained constant throughout the first year following the bleeding event.

Similarly, Ndrepepa and colleagues¹⁹ examined pooled data from four ISAR trials using the TIMI bleeding scale and found that myocardial infarction, target vessel revascularization, and major bleeding all had similar discriminatory ability at predicting 1-year mortality.

In patients undergoing elective or urgent percutaneous coronary intervention in the REPLACE-2 trial,²⁰ independent predictors of death by 1 year were²¹:

• Major hemorrhage (OR 2.66, 95% CI 1.44–4.92)
• Periprocedural myocardial infarction (OR 2.46, 95% CI 1.44–4.20).

THEORIES OF HOW BLEEDING MAY CAUSE ADVERSE OUTCOMES

Several mechanisms have been proposed to explain the association between bleeding during treatment for acute coronary syndromes and adverse clinical outcomes.^13,22

The immediate effects of bleeding are thought to be hypotension and a reflex hyperadrenergic state to compensate for the loss of intravascular volume.²³ This physiologic response is believed to contribute to myocardial ischemia by further decreasing myocardial oxygen supply in obstructive coronary disease.

Trying to minimize blood loss, physicians may withhold anticoagulation and antiplatelet therapy, which in turn may lead to further ischemia.²⁴ To compensate for blood loss, physicians may also resort to blood transfusion. However, depletion of 2,3-diphosphoglycerate and nitric oxide in stored donor red blood cells is postulated to reduce oxygen delivery by increasing hemoglobin’s affinity for oxygen, leading to induced microvascular obstruction and adverse inflammatory reactions.^15,25

Recent data have also begun to elucidate the long-term effects of bleeding during acute coronary syndrome management. Patients with anemia during the acute phase of infarction have greater neurohormonal activation.²⁶ These adaptive responses to anemia may lead to eccentric left ventricular remodeling that may lead to higher oxygen consumption, increased diastolic wall stress, interstitial fibrosis, and accelerated myocyte loss.^27–30

Nevertheless, we must point out that although strong associations between bleeding and adverse outcomes have been established, direct causality has not.

 

TO PREVENT BLEEDING, START BY ASSESSING RISK

Preventing bleeding is a key step in balancing the safety and efficacy of aggressive management of non-ST-elevation acute coronary syndromes. Current guidelines^1,31 call for assessing the risk of both thrombosis and bleeding in patients presenting with these syndromes (Figure 1). Doing so may allow clinicians to tailor therapy by adjusting the treatment regimen in patients at risk of bleeding to include medications associated with favorable bleeding profiles and by using radial access as the point of entry at the time of coronary artery angiography.

The CRUSADE bleeding risk score

The CRUSADE bleeding score (calculator available at http://www.crusadebleedingscore.org/) was developed and validated in more than 89,000 community-treated patients with non-ST-elevation acute coronary syndromes.⁸ It is based on eight variables:

• Sex (higher risk in women)
• History of diabetes (higher risk)
• Prior vascular disease (higher risk)
• Heart rate (the higher the rate, the higher the risk)
• Systolic blood pressure (higher risk with pressures above or below the 121–180 mm Hg range)
• Signs of congestive heart failure (higher risk)
• Baseline hematocrit (the lower the hematocrit, the higher the risk)
• Creatinine clearance (by the Cockcroft-Gault formula; the lower the creatinine clearance, the higher the risk).

Patients who are found to have bleeding scores suggesting a moderate or higher risk of bleeding should be considered for medications associated with a favorable bleeding profile, and for radial access at the time of coronary angiography. Scores are graded as follows⁸:

• < 21: Very low risk
• 21–30: Low risk
• 31–40: Moderate risk
• 41–50: High risk
• > 50: Very high risk.

The CRUSADE bleeding score is unique in that, unlike earlier risk stratification tools, it was developed in a “real world” population, not in subgroups or in a clinical trial. It can be calculated at baseline to help guide the selection of treatment.⁸

Adjusting the heparin regimen in patients at risk of bleeding

Both the joint American College of Cardiology/American Heart Association¹ and the European Society of Cardiology guidelines³¹ for the treatment of non-ST-elevation acute coronary syndromes recommend taking steps to prevent bleeding, such as adjusting the dosage of unfractionated heparin, using safer drugs, reducing the duration of antithrombotic treatment, and using combinations of antithrombotic and antiplatelet agents according to proven indications.³¹

In the CRUSADE registry, 42% of patients with non-ST-elevation acute coronary syndromes received at least one initial dose of antithrombotic drug outside the recommended range, resulting in an estimated 15% excess of bleeding events.³² Thus, proper dosing is a target for prevention.

Appropriate antithrombotic dosing takes into account the patient’s age, weight, and renal function. However, heparin dosage in the current guidelines¹ is based on weight only: a loading dose of 60 U/kg (maximum 4,000 U) by intravenous bolus, then 12 U/kg/hour (maximum 1,000 U/hour) to maintain an activated partial thromboplastin time of 50 to 70 seconds.¹

Renal dysfunction is particularly worrisome in patients with non-ST-elevation acute coronary syndromes because it is associated with higher rates of major bleeding and death. In the OASIS-5 trial,³³ the overall risk of death was approximately five times higher in patients in the lowest quartile of renal function (glomerular filtration rate < 58 mL/min/1.73 m²) than in the highest quartile (glomerular filtration rate ≥ 86 mL/min/1.73 m²).

Renal function must be evaluated not only on admission but also throughout the hospital stay. Patients presenting with acute coronary syndromes often experience fluctuations in renal function that would call for adjustment of heparin dosing, either increasing the dose to maximize the drug’s efficacy if renal function is recovering or decreasing the dose to prevent bleeding if renal function is deteriorating.

Clopidogrel vs prasugrel

Certain medications should be avoided when the risk of bleeding outweighs any potential benefit in terms of ischemia.

For example, in a randomized trial,³⁴ prasugrel (Effient), a potent thienopyridine, was associated with a significantly lower rate of the composite end point of stroke, myocardial infarction, or death than clopidogrel (Plavix) in patients with acute coronary syndromes undergoing percutaneous coronary interventions. However, it did not seem to offer any advantage in patients 75 years old and older, those with prior stroke or transient ischemic attack, or those weighing less than 60 kg, and it posed a substantially higher risk of bleeding.

With clopidogrel, the risk of acute bleeding is primarily in patients who undergo coronary artery bypass grafting within 5 days of receiving a dose.^35,36 Therefore, clopidogrel should be stopped 5 to 7 days before bypass surgery.¹ Importantly, there is no increased risk of recurrent ischemic events during this 5-day waiting period in patients who receive clopidogrel early. Therefore, the recommendation to stop clopidogrel before surgery does not negate the benefits of early treatment.³⁶

Lower-risk drugs: Fondaparinux and bivalirudin

At this time, only two agents have been studied in clinical trials that have specifically focused on reducing bleeding risk: fondaparinux (Arixtra) and bivalirudin (Angiomax).^20,37–39

Fondaparinux

OASIS-5 was a randomized, double-blind trial that compared fondaparinux and enoxaparin (Lovenox) in patients with acute coronary syndromes.³⁸ Fondaparinux was similar to enoxaparin in terms of the combined end point of death, myocardial infarction, or refractory ischemia at 9 days, and fewer patients on fondaparinux developed bleeding (2.2% vs 4.1%, HR 0.52; 95% CI 0.44–0.61). This difference persisted during long-term follow-up.

Importantly, fewer patients died in the fondaparinux group. At 180 days, 638 (6.5%) of the patients in the enoxaparin group had died, compared with 574 (5.8%) in the fondaparinux group, a difference of 64 deaths (P = .05). The authors found that 41 fewer patients in the fondaparinux group than in the enoxaparin group died after major bleeding, and 20 fewer patients in the fondaparinux group died after minor bleeding.³⁸ Thus, most of the difference in mortality rates between the two groups was attributed to a lower incidence of bleeding with fondaparinux.

Unfortunately, despite its safe bleeding profile, fondaparinux has fallen out of favor for use in acute coronary syndromes, owing to a higher risk of catheter thrombosis in the fondaparinux group (0.9%) than in those undergoing percutaneous coronary interventions with enoxaparin alone (0.4%) in the OASIS-5 trial.⁴⁰

 

 

Bivalirudin

The direct thrombin inhibitor bivalirudin has been studied in three large randomized trials in patients undergoing percutaneous coronary interventions.^20,37,41

The ACUITY trial³⁷ was a prospective, open-label, randomized, multicenter trial that compared three regimens in patients with moderate or high-risk non-ST-elevation acute coronary syndromes:

• Heparin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin alone.

Bivalirudin alone was as effective as heparin plus a glycoprotein IIb/IIIa inhibitor with respect to the composite ischemia end point, which at 30 days had occurred in 7.8% vs 7.3% of the patients in these treatment groups (P = .32, RR 1.08; 95% CI 0.93–1.24), and it was superior with respect to major bleeding (3.0% vs 5.7%, P < .001, RR 0.53; 95% CI 0.43–0.65).

The HORIZONS-AMI study⁴¹ was a prospective, open-label, randomized, multicenter trial that compared bivalirudin alone vs heparin plus a glycoprotein IIb/IIIa inhibitor in patients with ST-elevation acute coronary syndromes who were undergoing primary percutaneous coronary interventions. The two primary end points were major bleeding and net adverse events.

At 1 year, patients assigned to bivalirudin had a lower rate of major bleeding than did controls (5.8% vs 9.2%, HR 0.61, 95% CI 0.48–0.78, P < .0001), with similar rates of major adverse cardiac events in both groups (11.9% vs 11.9%, HR 1.00, 95% CI 0.82– 1.21, P = .98).⁴¹

Both OASIS 5 and HORIZONS-AMI are examples of clinical trials in which strategies that reduced bleeding risk were also associated with improved survival.

For cardiac catheterization, inserting the catheter in the wrist poses less risk

Bleeding is currently the most common noncardiac complication in patients undergoing percutaneous coronary interventions, and it most often occurs at the vascular access site.¹⁷

Rao et al¹² evaluated data from 593,094 procedures in the National Cardiovascular Data Registry and found that, compared with the femoral approach, the use of transradial percutaneous coronary intervention was associated with a similar rate of procedural success (OR 1.02, 95% CI 0.93–1.12) but a significantly lower risk of bleeding complications (OR 0.42, 95% CI 0.31–0.56) after multivariable adjustment.

The use of smaller sheath sizes (4F–6F) and preferential use of bivalirudin over unfractionated heparin and glycoprotein IIb/IIIa inhibitor therapy are other methods described to decrease the risk of bleeding after percutaneous coronary interventions.^20,41–49

IF BLEEDING OCCURS

Once a bleeding complication occurs, cessation of therapy is a potential option. Stopping or reversing antithrombotic and antiplatelet therapy is warranted in the event of major bleeding (eg, gastrointestinal, retroperitoneal, intracranial).³¹

Stopping antithrombotic and antiplatelet therapy

Whether bleeding is minor or major, the risk of a recurrent thrombotic event must be considered, especially in patients who have undergone revascularization, stent implantation, or both. The risk of acute thrombotic events after interrupting antithrombotic or antiplatelet agents is considered greatest 4 to 5 days following revascularization or percutaneous coronary intervention.¹⁵ If bleeding can be controlled with local treatment such as pressure, packing, or dressing, antithrombotic and antiplatelet therapy need not be interrupted.⁵⁰

Current guidelines recommend strict control of hemorrhage for at least 24 hours before reintroducing antiplatelet or antithrombotic agents.

It is also important to remember that in the setting of gastrointestinal bleeding due to peptic ulcer disease, adjunctive proton pump inhibitors are recommended after restarting antiplatelet or antithrombotic therapy or both.

Importantly, evidence-based antithrombotic medications (especially dual antiplatelet therapy) should be restarted once the acute bleeding event has resolved.³¹

Reversal of anticoagulant and antiplatelet therapies

Reversal of antithrombotic therapy is occasionally necessary (Table 2).

Unfractionated heparin is reversed with infusion of protamine sulfate at a dose of 1 mg per 100 U of unfractionated heparin given over the previous 4 hours.^51,52 The rate of protamine sulfate infusion should be less than 100 mg over 2 hours, with 50% of the dose given initially and subsequent doses titrated according to bleeding response.^52,53 Protamine sulfate is associated with a risk of hypotension and bradycardia, and for this reason it should be given no faster than 5 mg/min.

Low-molecular-weight heparin (LMWH) can be inhibited by 1 mg of protamine sulfate for each 1 mg of LMWH given over the previous 4 hours.^51,52

However, protamine sulfate only partially neutralizes the anticoagulant effect of LMWH. In cases in which protamine sulfate is unsuccessful in abating bleeding associated with LMWH use, guidelines allow for the use of recombinant factor VIIa (NovoSeven).³¹ In healthy volunteers given fondaparinux, recombinant factor VIIa normalized coagulation times and thrombin generation within 1.5 hours, with a sustained effect for 6 hours.⁵²

It is important to note that the use of this agent has not been fully studied, it is very costly (a single dose of 40 μg/kg costs from $3,000 to $4,000), and it is linked to reports of increased risk of thrombotic complications.^54,55

Antiplatelet agents are more complex to reverse. The antiplatelet actions of aspirin and clopidogrel wear off as new platelets are produced. Approximately 10% of a patient’s platelet count is produced daily; thus, the antiplatelet effects of aspirin and clopidogrel can persist for 5 to 10 days.^31,56

If these agents need to be reversed quickly to stop bleeding, according to expert consensus the aspirin effect can be reversed by transfusion of one unit of platelets. The antiplatelet effect of clopidogrel is more significant than that of aspirin; thus, two units of platelets are recommended.⁵⁶

Glycoprotein IIb/IIIa inhibitors. If a major bleeding event requires the reversal of glycoprotein IIb/IIIa inhibitor therapy, the treatment must take into consideration the pharmacodynamics of the target drug. Both eptifibatide (Integrilin) and tirofiban (Aggrastat) competitively inhibit glycoprotein IIb/IIIa receptors; thus, their effects depend on dosing, elimination, and time. Due to the stoichiometry of both drugs, transfusion of platelets is ineffective. Both eptifibatide and tirofiban are eliminated by the kidney; thus, normal renal function is key to the amount of time it takes for platelet function to return to baseline.⁵⁷ Evidence suggests that fibrinogen-rich plasma can be administered to restore platelet function.^31,58,59

Abciximab (ReoPro). Whereas reversal of eptifibatide and tirofiban focuses on overcoming competitive inhibition, neutralization of abciximab involves overcoming its high receptor affinity. At 24 hours after abciximab infusion is stopped, platelet aggregation may still be inhibited by up to 50%. Fortunately, owing to abciximab’s short plasma half-life and its dilution in serum, platelet transfusion is effective in reversing its antiplatelet effects.^31,57

 

 

Blood transfusion

Long considered beneficial to critically ill patients, blood transfusion to maintain hematocrit levels during acute coronary syndromes has come under intense scrutiny. Randomized trials have shown that transfusion should not be given aggressively to critically ill patients.⁶⁰ In acute coronary syndromes, there are only observational data.

Rao et al⁶¹ used detailed clinical data from 24,112 patients with acute coronary syndromes in the GUSTO IIb, PURSUIT, and PARAGON B trials to determine the association between blood transfusion and outcomes in patients who developed moderate to severe bleeding, anemia, or both during their hospitalization. The rates of death in the hospital and at 30 days were significantly higher in patients who received a transfusion (30-day mortality HR 3.94; 95% CI 3.36–4.75). However, there was no significant association between transfusion and the 30-day mortality rate if the nadir hematocrit was 25% or less.

Of note: no randomized clinical trial has evaluated transfusion strategies in acute coronary syndromes at this time. Until such data are available, it is reasonable to follow published guidelines and to avoid transfusion in stable patients with ischemic heart disease unless the hematocrit is 25% or less.³¹ The risks and benefits of blood transfusion should be carefully weighed. Routine use of transfusion to maintain predefined hemoglobin levels is not recommended in stable patients.