Alcohol-associated hypertension: when one drinks makes a difference.

The high prevalence of systemic hypertension (HTN) and its progressive association with multiple adverse sequelae1 confer great potential importance on any modifiable trait influencing blood pressure (BP). Established modifiable factors include obesity and sodium intake. Less solidly established are physical inactivity and low potassium or calcium intake. With a relationship to HTN seen nearly unanimously in a substantial number of observational crosssectional and prospective studies, alcohol drinking has become a major focus of interest in recent decades as a modifiable HTN risk factor.2–3 Even low estimates of 5% or 7% of HTN attributable to alcohol imply that there are more patients with HTN caused by alcohol than by conventional causes of remediable secondary HTN.3 Present in both sexes and several ethnic groups, the alcohol–HTN association is independent of a number of potential confounders including nutritional factors.2–3 Available evidence suggests substantial regression of HTN in several days with alcohol abstinence. Although heavier alcohol intake ( 2 drinks daily) has been more consistently related, an alcohol–BP relationship with lighter drinking in some studies raises a possibility of great potential pubic health importance. Most observational analyses used average amount of alcohol consumed daily as the alcohol measure, with presumed underestimation of alcohol amount by some heavier drinkers impairing definition of a threshold dose level.3 To add further confusion, a few studies show lower pressures in light drinkers than in abstainers, perhaps especially in women.3 Lack of a proven mechanism limits causal interpretation of the alcohol–HTN link. There are other unresolved issues, including: (1) is the relationship linear at all drinking levels or is there a threshold alcohol dose; (2) is choice of wine, liquor, or beer a factor; (3) is drinking pattern a factor; (4) what is the time sequence in drinking–BP changes; (5) are diet and/or behavioral factors confounders; and (6) does alcoholassociated HTN result in the usual HTN sequelae? All of these queries are relevant to the role of reduction of alcohol intake in HTN treatment or prevention. The presentation by Stranges et al in this issue of Hypertension4 contributes importantly to several of these issues. This is one of a series of analyses from this Western New York State group characterized by unusually meticulous collection of data about drinking patterns. The analysis confirms the known risk of HTN associated with recent daily intake of 2 drinks per day, which was more than doubled compared with abstainers. The report’s most noteworthy feature is the robust independent relationship found for alcohol taken separately from meals. The 12% of drinkers whose predominant drinking pattern was ingestion separate from food had 64% greater risk of HTN. Although persons drinking predominantly liquor had borderline higher HTN risk versus predominant beer/wine drinkers, the independent effect of beverage type observed in the Stranges et al report4 is not impressive. The analysis is especially interesting on this point because wine is the beverage type most often consumed with food. Of historic interest is the fact that the first report5 of an alcohol–HTN association involved French soldiers who were heavy wine drinkers. Wine’s modest potassium content offers hypothetical benefit for HTN, and wine drinkers have lifestyle traits favorable for HTN risk.6–7 Yet, in agreement with this report, most available data3 suggest no important role of beverage choice in HTN risk. Weekend-only drinking, independent of total amount of alcohol, had little apparent independent BP effect in the analysis. Although related, weekend-only drinking and binge drinking are not synonymous. As Stranges et al point out,4 there are inconsistencies in BP relationships to weekend drinking in previous reports. One relevant ecological study8 showed higher BPs on Mondays in Ireland, with a heavy weekend drinking pattern, but no relation to day of week in France, with a steadier drinking pattern. One might hypothesize that repeated pressor effects of alcohol might ultimately lead to irreversible changes. Data from controlled clinical experiments verify an alcohol–HTN association and support a causal interpretation. The landmark experiment of Potter and Beevers9 used a crossover design in hospital in men with HTN and usual intake of 4 pints of British beer. The data suggested both a pressor effect of alcohol in hypertensive men and short-term BP reductions after alcohol withdrawal, with changes requiring several days to appear and not present in all subjects. No withdrawal symptoms or blood pressure rebound (overshoot) was observed. Australian reports of controlled trials of alcohol restriction in ambulatory men with HTN and normotensives showed lower pressures during alcohol withdrawal.3 Other clinical trials suggest that regular drinking may antagonize efficacy of drug treatment and weight reduction, that alcohol The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association. From the Kaiser Permanente Medical Center, Oakland, Calif. Correspondence to Arthur L. Klatsky, Kaiser Permanente Medical Center, 280 West MacArthur Boulevard, Oakland, CA 94611. E-mail [email protected] or [email protected] (Hypertension. 2004;44:805-806.) © 2004 American Heart Association, Inc.