Alcohol and hypertension

An association between heavier alcohol consumption and increased systemic hypertension has joined other correlates of hypertension, such as obesity and salt intake, as a major research focus on hypertension risk factors.^1,2 In epidemiologic studies, higher blood pressure among persons reporting a usual intake of 3 or more drinks per day has been found in men and women, in several ethnic groups, and across all adult ages. Although plausible mechanisms have not been established, several indications, including short-term clinical experiments, suggest a causal relationship. Heavier drinking may, in fact, be the most common cause of reversible hypertension.

If drinking alcohol plays a causal role in hypertension, one might expect that alcohol-associated hypertension would be followed by the usual adverse health sequelae of increased blood pressure, including coronary heart disease (CHD), stroke, heart failure, and renal failure. With the possible exception of renal failure, drinking alcohol has complex, independent relationships with each of these outcomes. For example, population-based studies have established that light-moderate drinking is associated with a lower CHD risk³ and with lower risk of heart failure associated with CHD.⁴ Because hypertension is a major CHD risk factor, the adverse effects of alcohol-related hypertension in heavy drinkers might mitigate the benefits of alcohol. The epidemiologic relationship between stroke risk and alcohol is more complex than that for CHD; hemorrhagic and ischemic stroke each have different risk factors and associations with alcohol. However, hypertension is a powerful predictor of both major stroke types.

In an earlier Kaiser Permanente study, we examined the alcohol—hypertension interactions with risk of mortality and selected causes of hospitalization by comparing hypertensive heavy drinkers with matched hypertensive controls who were not heavy drinkers. Unfortunately, alcohol's harmful and beneficial effects dominated the outcomes, obscuring the clinical outcomes related to alcohol-associated hypertension.² We recently performed a new cohort study of several cardiovascular outcomes in which the risk of sequelae was examined within specific stratified alcohol intake categories. This method of analysis permitted ascertainment of risk related to hypertension independent of baseline alcohol intake.

Subjects and methods

The study included 128,934 members of a Northern California prepaid comprehensive health care program who supplied baseline data at health examinations from 1978 to 1985. Sociodemographic status, habits, medical history, and the usual number of drinks (wine, liquor, beer, or cocktails) in the past year were documented.

The blood pressure categories of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure VI were used for analyses.⁵ The referent group (n = 46,728) had "optimal" blood pressure (< 120/80 mm Hg); the other categories were "normal" (120-129/80-84 mm Hg; n = 28,454), "high normal" (130- 139/85-89 mm; n = 21&thinsp;157), and "Hypertension" (≥140/90 mm Hg; n = 30,873).

Follow-up end points studied included death from cardiovascular causes (n = 8358), hospitalizations attributed to cardiovascular causes (n = 18,962), and an outpatient diagnosis of hypertension (n = 26,340). Analytic Cox proportional hazards and logistic models were controlled for age, sex, ethnicity, education, body mass index, cigarette smoking, and alcohol intake. The alcohol intake categories were never drinkers, ex-drinkers, and usual number of drinks of < 1, 1-2, and ≥ 3 per day. Some study models of the outcomes included all alcohol categories compared with never drinkers. The key analyses were limited to specific alcohol intake groups so that the risk associated with hypertension could be assessed in subjects with those drinking habits.

Results

P

The study substantiated several expected correlates of hypertension,^5-7 such as greater prevalence with increasing age and body mass index and lower prevalence with increasing education (data not shown). Importantly, the association between an alcohol intake of ≥ 3 drinks per day and hypertension (blood pressure ≥ 140/90 mm Hg) was confirmed, with a 70% increased prevalence of hypertension in this group of drinkers. The increased prevalence of hypertension among these heavier drinkers was 90% for men and 60% for women ( < .001 for each sex).

The risk of death attributed to any cardiovascular cause or to CHD showed a progressive association with increasing blood pressure for all alcohol strata (Table 1). A similar increased risk of hospitalization for cardiovascular diagnoses was associated with hypertension and with "high normal" blood pressure (130-139/85-89 mm Hg) for each condition (Table 2). Results were similar for hospitalization diagnoses of hypertension, heart failure, and stroke (not shown in Table 2). Table 3 shows that the results were similar for outpatient diagnoses of hypertension many years later. These risks were present in both sexes and in each major ethnic group. There was only a slight lessening of strength from baseline measurements after 10 years.

Table 1. Risk of cardiovascular and CHD mortality for blood pressure

categories "stratified by alcohol intake."

The usual associations of CHD, stroke, and heart failure with the covariates were found, with progressive risks for increasing baseline cigarette smoking, body mass index, and age. African Americans were at greater risk than whites for most outcomes, particularly hypertension, and men were at significantly greater risk for all outcomes except outpatient hypertension.

Table 2. Risk of selected cardiovascular hospitalizations for blood pressure

categories* stratified by alcohol intake.†

Table 3. Odds ratio of outpatient hypertension diagnosis for blood pressure

categories* stratified by alcohol intake.†

We confirmed the known association between alcohol intake and cardiovascular outcomes by comparing the risks of subjects reporting various alcohol intake amounts with those of lifelong abstainers. For example, as previously described,³ among persons reporting 1 to 2 drinks per day, the risks were lower for all cardiovascular and CHD deaths and for all hospitalization outcomes than among abstainers. It is important to note that this level of alcohol intake (1-2 drinks per day) was related to an increased risk of an outpatient diagnosis of hypertension.

Discussion

Our study confirmed the known increased risk of cardiovascular events among individuals with a blood pressure ≥ 120/80 mm Hg.^5-7 This increase is progressive with all blood pressure categories; above optimal blood pressure, there is no specific cut point below which higher pressures have no adverse effects. The second important finding is that these adverse cardiovascular consequences are independent of alcohol intake reported at the time of blood pressure measurement. These blood pressure relationships are unrelated to the beneficial or adverse effects of alcohol.

The large study population of subjects with various drinking habits, the diversity of the end points studied, and the long follow-up period are strengths of this study. Interpretation of the data is limited by the fact that we could not identify the specific individuals whose blood pressure was raised by an alcohol effect. This limitation makes it impossible to ascertain the specific effects of hypertension caused by alcohol.

An ideal study would compare sequelae in groups with known serial blood pressure changes related to serial drinking changes. Such a study, even a randomized controlled trial, would require a very large population with a long follow-up period, which is probably unlikely to ever be performed.

As people reach older ages, alcohol intake tends to diminish.⁸ Because older persons drink less, the persistence of the association of sequelae with increased baseline blood pressure among daily drinkers at ≥ 10 years suggests that the associations are valid. We know that in this study population, 8 mostly new heavy drinkers come from those reporting 1 to 2 drinks per day, a group also likely to include under-reporters.⁹ Changes to or from heavy drinking are thus likely to involve persons reporting an intake of 1 to 2 drinks per day, a category related to higher hypertension prevalence. A final problem, common to all observational studies, is the inability to control for all potential risk factors. Uncontrolled traits could hypothetically be related to both drinking habits and to hypertension-associated outcomes.

In our study, more than one third of heavy drinkers had hypertension, compared with less than one quarter in the entire population. Most heavy drinkers do not have hypertension, a fact supporting speculation about existence of alcohol "sensitivity" analogous to the genetically based salt sensitivity phenomenon. ¹⁰ Alcohol—blood pressure responsiveness might indicate predilection to subsequent hypertension, even without continuation of heavy drinking.

The alcohol—hypertension facts have nothing to do with the now widely known possible benefits of light-moderate drinking.^3,11 Whether or not an individual may benefit from such alcohol intake depends almost entirely on individual risk—benefit considerations.^3,11,12 The great prevalence of hypertension plus its established role in disease mandate attention to any modifiable etiologic traits.^13-14 Part of the management of hypertension in a heavier drinker is reduction of alcohol intake or abstinence. The results of our study suggest that hypertension at any drinking level is far from benign and that alcohol-associated hypertension belongs on the list of reasons to avoid heavy drinking.

Conclusions

We evaluated cardiovascular sequelae in heavy drinkers, light drinkers, and abstainers and found that the risk of all outcomes was progressively greater for increasing blood pressure categories, with a similarly increased risk for abstainers, light drinkers, and heavy drinkers. Alcohol—blood pressure interactions were not statistically significant for death and hospitalization. The study was limited by the inability to separate subjects with increased blood pressure due to alcohol intake from subjects with other causes of hypertension. These results show that hypertension risks are comparable regardless of alcohol intake.