Alcohol Intake and Noncoronary Cardiovascular Diseases

Stroke 

The clinical syndrome of stroke is a quintessential example of the heterogeneity among cardiovascular diseases. Stroke comprises two main types of cardiovascular disease — ischemic and hemorrhagic — and several subtypes of each (e.g., atherothrombotic, lacunar, embolic, subarachnoid hemorrhage, intracerebral hemorrhage). Alcohol appears to have decidedly different associations with these various types.

Ischemic stroke, the most common type in the United States and Europe, is the result of vascular occlusion of an intracerebral artery, either through local thrombosis or distal embolism. It is noteworthy that two of the strongest chronic risk factors for ischemic stroke are hypertension and AF (8), both of which have been attributed to heavy drinking (see below). Thus, even though ischemic stroke shares an atherosclerotic and thrombotic etiology with CHD, the particular factors that most strongly predict ischemic stroke risk appear to be in the subset of cardiovascular risk factors that are exacerbated by heavy alcohol use. As a consequence, it may not be surprising that the apparent inverse association of moderate alcohol intake with ischemic stroke risk occurs at a lower dose of alcohol consumption and with a lower magnitude of risk reduction than does the corresponding association with CHD risk 9, 10. It is also not surprising that heavy drinking is consistently associated with higher stroke risk.

For many years, observational (and largely case-control) studies of the relationship between alcohol consumption and risk for ischemic stroke have suggested a relatively strong inverse relationship 4, 11. For example, in the Nurses' Health Study cohort, Stampfer and colleagues documented 66 incident ischemic strokes and found relative risks of 0.3 (95% confidence interval [CI], 0.1–0.7) for intake of 5 to 14 g of alcohol per day, and 0.5 (95% CI, 0.2–1.1) for intake levels of 15 or more g/d (12). However, more recent studies have called the magnitude of lower risk into question. In a meta-analysis restricted to 19 cohort studies, Reynolds and colleagues found that light drinking (<12 g/d) was associated with a relative risk of 0.82 (95% CI, 0.73–0.92), but that consumption of 12 to 60 g/d was not associated with risk (4). In a representative study, Klatsky and colleagues reported a U-shaped association between alcohol intake and hospitalization for ischemic stroke among members of a single insurance plan. Relative risks were 0.8 (95% CI, 0.7–1.0) among those who consumed between one drink a month and one drink a day and 1.0 (95% CI, 0.8–1.2) among those who consumed 3 or more drinks per day (10). For comparison, these authors found a direct inverse association between alcohol use and risk of CHD hospitalization, with relative risks of 0.6 among women and 0.7 among men who consumed 3 or more drinks per day (9). In an analysis of the Cardiovascular Health Study, a lower risk of ischemic stroke was confined to consumers of 1 to 6 drinks per week (hazard ratio 0.75; 95% CI, 0.53–1.06), although even this level of intake modestly but not significantly increased the risk among carriers of the APOE4 allele (13). In the same population (14), the corresponding magnitude of lower risk for CHD was lower and occurred at an intake of 14 or more drinks per week.

Dulli has emphasized how the heterogeneity of ischemic stroke may influence its association with alcohol intake (15). To address this, we examined ischemic stroke risk in the Health Professionals Follow-up Study (16). Concordant with previous work, lower risk was restricted to consumers of an average of 0.1 to 9.9 g of ethanol per day, and particularly to those whose consumption was spread out over 3 to 4 days per week (hazard ratio 0.56; 95% CI, 0.31–1.02). However, stroke subtype appeared to modify risk substantially. Compared with abstention, the hazard ratios associated with intake of 0.1 to 9.9 g/d were 0.76 (95% CI, 0.50–1.15) for confirmed thrombotic stroke, but 2.20 (95% CI, 0.84–5.76) for confirmed embolic stroke. Larger studies with effective classification of ischemic stroke subtypes are needed to confirm these findings.

It should be repeated that little controversy exists regarding the relationship of heavier drinking with a higher risk of ischemic stroke 4, 16. Moreover, even single episodes of intake of 3 or more drinks or drinking to intoxication may acutely and transiently increase the risk of ischemic stroke for the ensuing 24 hours 17, 18, 19, 20, 21.

Hemorrhagic stroke is a common stroke type in Asia, where most studies addressing it have been performed, and includes both subarachnoid and intracerebral subtypes. In a meta-analysis of cohort studies of subarachnoid hemorrhage, Feigin found increased risk restricted to heavier intake (>150 g/wk), with a summary relative risk of 2.1 (95% CI, 1.5–2.8) (22). In a comparable meta-analysis of case-control studies of intracerebral hemorrhage, Ariesen found summary odds ratios of 2.05 (95% CI, 1.35–3.11) for intake of 56 g or less per day, and 4.11 (95% CI, 2.54–6.65) for intake of more than 56 g/d (23), although a large cohort study published subsequently found an increase in risk to be confined to consumers of 6 or more drinks per day (24). These results are largely consistent with Renaud's hypothesis that the effects of moderate drinking are mediated, at least in part, through activity of platelets and prothrombotic factors (25). If this hypothesis is correct, one would anticipate that moderate drinking would decrease the risk of thrombotic events, such as myocardial infarction and ischemic stroke, but increase the risk of bleeding in the brain and possibly other sites (26), exactly as has been observed.

Congestive heart failure 

For decades (27), clinical investigators interested in the cardiac effects of ethanol focused their efforts on alcoholic cardiomyopathy, a dilated cardiomyopathy attributed primarily to cardiotoxic effects of ethanol and secondarily also associated with malnutrition (28). Interestingly, although it has long been recognized that only substantial ethanol intake leads to cardiomyopathy (although less so in women) (29), only recently have studies demonstrated that in subjects with cardiomyopathy, controlled drinking could be as effective as abstinence in stabilizing or restoring myocardial function (30).

More recent interest in CHF has been spurred by 3 observations. First, with the aging of Western populations, CHF has rapidly become a serious public health concern. In 2003 alone, more than one million Americans were hospitalized with a primary diagnosis of CHF (31). Second, a substantial portion of the burden of CHF in the United States and Europe is the consequence of CHD and ischemic cardiomyopathy. If moderate alcohol use is associated with a lower risk of CHD, one might anticipate that it would also be associated with a lower risk of CHF. Third, physiological experiments have suggested that alcohol consumption may induce effective atrial natriuretic peptide (ANP)–mediated diuresis 32, 33 and lower pulmonary artery pressure (34), both of which would be expected to reduce the incidence of CHF; the association of alcohol intake with ANP has now been confirmed in a large observational study (35).

The observed association of alcohol use with risk for CHF has justified this interest but has also added new complexity. In at least 4 large prospective studies 7, 36, 37, 38, investigators have found that moderate drinking is associated with approximately a 50% lower risk of CHF, although the level of intake linked to lowest risk varies somewhat. Interestingly, in all 4 studies, accounting for incident coronary disease did not eliminate the relationship, so that even where previous myocardial infarction was not involved, risk appeared to be, to some degree, inversely related to moderate drinking. Some (39), but not all 40, 41, studies of alcohol intake among patients with established left ventricular dysfunction have also suggested a benefit from light or moderate drinking; one study has suggested an association with lower mortality in men but higher mortality in women (42).

Atrial fibrillation 

Atrial fibrillation (AF) is the most common chronic cardiac arrhythmia and is closely linked with hypertension and CHF. Clinicians have long recognized the temporal association of episodic heavy alcohol use with onset of AF, dubbed the “holiday heart” syndrome 43, 44. However, the association of regular alcohol intake with risk for AF has only recently begun to receive attention. Given the association of heavy drinking with risk for AF, one might hypothesize that even moderate drinking would be associated with higher risk, although animal models have suggested the opposite 45, 46. In human populations, some studies have shown positive relationships in men but not in women (47), whereas some have even shown inverse relationships (48). However, most studies have suggested a threshold effect 6, 44, 49, 50, 51. In these studies, there has generally been no discernible difference in risk between abstainers and moderate drinkers, but a higher risk for the heaviest drinkers. Although the absolute threshold of risk is uncertain, the largest study to date suggests that risk for AF does not increase below a long-term intake level of at least 28 and probably 35 drinks per week (51).

An interesting area for future investigation is the effect of alcohol consumption on individuals with chronic AF, most of whom are likely to be prescribed warfarin or similar anticoagulants. Although guidelines have often explicitly recommended that individuals taking warfarin abstain entirely from drinking (52), the limited available data suggest that the risk for excessive anticoagulation is not higher in moderate drinkers than it is in abstainers, although it may be elevated in binge drinkers 53, 54.

Peripheral arterial disease 

The final manifestation of cardiovascular disease to be discussed herein is peripheral arterial disease (PAD), most commonly localized to the lower extremities; for practical purposes, abdominal aortic aneurysm (AAA) is also considered in this category. Although PAD has generally received the least attention from researchers among the major atherosclerotic vascular diseases, it is nonetheless widely prevalent and a major cause of disability, primarily related to walking.

The body of evidence linking moderate drinking to PAD is modest. The Strong Heart Study and Rotterdam Study both found inverse cross-sectional associations of moderate drinking with PAD (defined as a low ankle-brachial index) that tended to be strongest in women 55, 56, although other cross-sectional studies have not consistently confirmed this finding 57, 58. Surprisingly, a subsequent analysis of the progression of the ankle-brachial index in one of these latter studies has suggested an inverse relationship between alcohol use and a 5-year change in the ankle-brachial index (59). To our knowledge, two prospective cohort studies have assessed the risk of PAD in the general population. In an analysis of symptomatic claudication from the Framingham Heart Study (5), Djoussé and colleagues found hazard ratios of 0.67 (95% CI, 0.42–0.99) associated with intake of 13–24 g/d for men and 0.44 (95% CI, 0.23–0.80) associated with intake levels of 7 to 12 g/d for women. In the Physicians' Health Study (60), Camargo and colleagues found a hazard ratio of 0.74 (95% CI, 0.57–0.97) associated with an intake of 7 or more drinks per week, although the composite outcome included only 66 cases of lower-extremity arterial disease that required revascularization. In contrast, Ciccarone and colleagues found no association of alcohol intake with risk of PAD among diabetic adults (61).

There has been relatively little investigation into the relationship of moderate alcohol consumption with risk of AAA. In an analysis of male smokers, Tornwall and colleagues found a U-shaped relationship, with relative risks for AAA of 0.74 (95% CI, 0.49–1.13) among consumers of up to 15 g of ethanol per day and 0.60 (95% CI, 0.36–1.02) among consumers of 16 to 30 g of ethanol per day (62).