Alcohol and hypertension: implications for management*

Update/Le point Alcohol and hypertension: implications for management* World Hypertension League' This article presents clear-cut evidence for a pressor effect of alcohol in regular drinkers. The effect is seen in both men and women, appears to increase in magnitude with the amounts consumed daily, and is additive to that of obesity and oral contraceptive use. Alcohol may also contribute to refractoriness in anti-hypertensive drug treatment. This pressor effect is at least partially reversible by reducing the alcohol consumption. Further research is required on the mechanisms involved. * A French translation of this article will appear in a later issue of the Bulletin. ' Prepared by L. Beilin, Perth, Australia, whose work is gratefully acknowledged. The present text has been accepted by the member organizations of the World Hypertension League, which in 1990 comprised the following: National Heart Foundation of Australia; Belgium Hypertension Committee; British Hypertension Society; Canadian Coalition for High Blood Pressure Prevention and Control; Chinese Hypertension Society; Danish Society of Hypertension; French National Comittee against Arterial Hyper- tension; German League against Hypertension; Hellenic Society of Hypertension; Hypertension Committee of the Hungarian So- ciety of Cardiology; All India Heart Foundation; Israel Hyperten- sion Society; Italian League against Hypertension; Japan Heart Foundation; Hypertension League of Morocco; Netherlands Heart Foundation-Hypertension Committee; Norwegian Society of Hypertension; Polish Society of Hypertension; Portuguese League against Hypertension; Senegal Foundation for Hyperten- sion and Heart Disease Control; Southern African Hypertension Society; Spanish League against Hypertension; Swiss Associa- tion against High Blood Pressure; Uruguayan Hypertension Lea- gue; U.S. National High Blood Pressure Education Program; Zaire League against Arterial Hypertension; and International Green Cross. Requests for reprints should be sent to: World Hypertension League, 20 Avenue du Bouchet, 1209 Geneva, Switzerland. This article has also been published in the Journal of human hypertension, 5: 227-232 (1991). It is published here with permission. Previous articles in this series have been: Self-measure- ment of blood pressure, Journal of hypertension, 6: 257-261 (1988), and Bulletin of the World Health Organization, 66: 155-159 (1988); Weight control in the managment of hyperten- sion, Bulletin of the World Health Organization, 67: 245-252 (1989); and Physical exercise in the management of hyperten- sion, Joumal of hypertension, 9: 283-287 (1991), and Bulletin of the World Health Organization, 69 (2): 149-153 (1991). Reprint No. 5186 Introduction A cause and effect relationship between regular alcohol consumption and blood pressure elevation was first suggested in 1915 by Lian (1), who described an increased prevalence of hypertension in French servicemen who drank more than 2.5 litres of wine per day. Since then, sporadic reports linking alcohol and hypertension have continued to appear (2), but more definitive evidence on this relationship comes from a variety of sources, including popula- tion studies, cross-sectional studies on blood pres- sure in problem drinkers, detection of an increased prevalence of abnormal liver function tests in hypertensives, blood pressure measurements follow- ing abstention in problem drinkers, concordance studies in twins, mortality statistics for drinkers and non-drinkers, and, most recently, randomized con- trolled intervention trials of varying alcohol intake in normotensive and hypertensive subjects. Many of these studies have been reviewed elsewhere and all the evidence, taken together, suggests that regular alcohol consumption can now be regarded as a major contributor to blood pressure elevation in drinking communities (3-6). Epidemiological evidence There are now over forty population studies (3, 6) demonstrating an association between regular alco- hol consumption, blood pressure levels and the Bulletin of the World Health Organization, 69(4): 377-382 (1991) 377 World Hypertension League prevalence of hypertension. The largest of these, the Kaiser-Permanente Health Screening Survey (7), involved more than 80 000 persons aged 15 to 79 years and showed an average rise of 1 millimetre mercury systolic pressure for each standard glass of alcohol drunk per day, with an effect which was independent of age, sex, obesity, race, cigarette consumption, and educational attainment. Diastolic pressure showed a similar though less marked relationship. Those taking three or more drinks a day had roughly a twofold higher prevalence of hypertension (more than 160/95 mmHg) compared with non-drinkers. With minor variations similar data have now been produced from several conti- nents (7-12), with the vast majority showing an independent relationship between alcohol and blood pressure levels, which in some instances equalled in magnitude the relative effect of obesity (9, 13) as the strongest identifiable factor contributing to blood pressure variations. A number of the studies, particularly those from North America (7, 8, 14, 15), have suggested a threshold for this effect of alcohol at around three standard drinks per day, with something of a J-shaped relation with blood pres- sure. However, many other studies, including those from different centres in Australia (9, 16, 17), have shown a linear relationship throughout the entire range of alcohol consumption from one standard drink a day upwards. Possible reasons for these discrepancies are the inclusion of some heavy drinkers and ex-heavy drinkers in the so-called teetotal population and failure to adequately control for other factors such as obesity in some of the studies. There are particular difficulties in quantitat- ing alcohol intake accurately from questionnaires, but despite this the estimated alcohol/blood pres- sure relationships overall between many different communities remain consistent, in stark contrast with the difficulty in demonstrating within-popula- tion relationships between dietary sodium intake and blood pressure. For example, the 'Intersalt' study of dietary factors influencing blood pressure levels in 52 populations worldwide showed a greater effect from alcohol consumption and obesity, than from salt intake. Only when the data were 'doc- tored' by being expressed in mmHg/year of age did a salt effect become significant (13). Studies in Perth (Western Australia) have focused on the role of alcohol in relation to other life-style factors, and have used carefully designed 7-day retrospective diaries to assess drinking habits. In the study by Arkwright et al. (9) in 500 Caucasian men aged 20-44 years working in local government, systolic blood pressures increased progressively with increasing alcohol consumption with no obvious threshold effect, so that even those drinking 1 to 2 standard glasses of beer-equivalent per day had higher blood pressures than non-drinkers. There was an increasing prevalence of both systolic and diastolic hypertension (more than 140 mmHg and 90 mmHg, respectively) with increasing alcohol con- sumption, so that the 53% of the population who drank the equivalent of three or more glasses per day showed a fourfold increase in the prevalence of systolic hypertension (10.4%) compared with non- drinkers (2.5%) and a threefold excess of diastolic hypertension. The effects of body weight excess and alcohol consumption were independent but additive, while smokers had slighty lower blood pressures overall, but again showed a drinking/blood pressure relationship even after adjusting for the effects of obesity. In women the effects of alcohol use appear to be additive to that of the blood-pressure-raising effects of oral contraceptives, as demonstrated by Wallace et al. (8) in the Lipid Clinic Study in nine North American populations. Cooke et al. (16) in Sydney (Australia) examined their data specifically for the effects of lower levels of alcohol consumption in 20000 men and women who attended health screening clinics and found a progressively linear relationship for both systolic and diastolic pressure in both men and women, with an apparent threshold for an effect of alcohol which disappeared after adjusting for the effects of age, obesity and smoking habits. Even those consuming as little as 6 standard drinks (60 grams of ethanol) a week had significantly higher blood pressures than non-drinkers. The influence of the type of alcoholic beverage was addressed in the Lipid Clinic Prevalence Study (18) in which beer and spirit consumptions were independently associated with increasing blood pres- sure levels. Although most other studies were on populations drinking predominantly one type of liquor, the fact that they demonstrated an associ- ation with blood pressure levels regardless of whether the beverage consumed was mainly beer (17,19, 20), wine (10), or mixed spirits, sake and beer (11), suggests that ethanol per se is the responsible agent. A specific pressure effect of ethanol has been confirmed by controlled trials in man, as discussed below. Whether other constitu- ents of alcoholic beverages can modify this effect remains to be demonstrated. Additional evidence Circumstantial evidence supporting a causative ef- fect of alcohol on hypertension comes from reports of a relatively high prevalence of abnormal liver function tests in hypertensive patients attending WHO Bulletin OMS. Vol 69 1991.378 Alcohol and hypertension blood pressure clinics in Scotland (Glasgow) (21) and Sweden (22), and by similar findings in a case-controlled study of hypertensives in the Scot- tish Renfrew blood pressure screening surveys (23). Longitudinal studies have demonstrated an in- creased risk of developing hypertension in heavy drinkers (24), and the Framingham survey showed that changes in blood pressure over four years were in the same direction as changes in alcohol consump- tion (25). Controlled studies The most conclusive evidence for a direct pressor effect of regular alcohol consumption comes from the first two randomized controlled studies carried out by Puddey and colleagues in Perth. In the first (26) of these, 44 normotensive subjects were ran- domized into two groups for a two-period crossover trial in which alcohol consumption was varied between normal intake and 20% of normal intake by asking volunteers to consume either 5% or 0.9% ethanol beer. As the latter was a distillate of the standard beer it provided similar fluid intake and the same constituents of beer apart from ethanol per se. The subjects who had been drinking low-alcohol beer showed a highly significant independent effect of alcohol on systolic blood pressure with an average fall of around 3.4 mmHg; the effect appeared within two weeks of reducing the alcohol intake and persisted over six weeks, with a rise in blood pressure again when intake of normal beer was resumed. Changes in alcohol consumption were confirmed by corresponding changes in mean cor- puscular volume, HDL cholesterol and both its sub-fractions, and gamma glutamyl transferase levels. A further trial of similar design was carried out in treated hypertensives (27) and again showed significant pressor effects of alcohol with changes in blood pressure of around 5 mmHg systolic and 2-3 mmHg diastolic with changes in alcohol con- sumption. Subsequent to this, similar effects have been reported in untreated hypertensives (28) and in young normotensive student volunteers put on to programmed drinking over periods of 1-2 weeks. Potter and Beevers (30) admitted groups of drinking hypertensives to hospital and demonstrated a fall in blood pressure in eight subjects over 3-4 days following abstention, followed by a rise on resump- tion of their normal alcohol consumption. In all of these intervention studies the changes in blood pressure were gradual over a period of 1-2 weeks as a rule, clearly indicating that the effect of alcohol was not simply an acute withdrawal phenomenon on the morning after drinking the night before. Unresolved issues A number of important issues remain to be resolved. The first of these is the question of individual susceptibility to the blood-pressure-elevating effects of regular alcohol consumption. The epidemiolo- gical studies clearly showed that there are additive effects of alcohol with obesity (7, 9) and oral contraceptive use (8). The effects generally seem to be greater with increasing age (10), although the data on this are not entirely clear cut. Analysis of personality factors in the data of Arkwright et al. suggests that introverted drinkers are far more prone to blood pressure elevation than extroverts, particularly if they are non-smokers (31). The effects of factors such as family history of high blood pressure or of differences in metabolism of alcohol on the pressor effect of ethanol have not been adequately explored. It is also unclear whether short concentrated bouts of drinking, say over 1-3 days at the weekend, are more likely to lead to hyperten- sion, than if the same amount of alcohol is consumed evenly over the week and combined with meals. The question of the mechanism of the blood- pressure-raising effect of alcohol is also still unre- solved (32-34). There is evidence that regular drinking increases blood pressure and heart rate variability (35), possibly by an effect on barorecep- tor control (36). Only a very small minority of heavy drinkers show the pseudo-Cushing syndrome (37) and in most cases cortisol metabolism appears to have been normal, as judged by crude plasma cortisol measurements or 24-hour-urine steroid ex- cretion. Studies of matched pairs of moderate drinkers (30 grams of ethanol/day on average) compared with non-drinkers have shown no differ- ences in plasma renin activity, angiotensin II, aldosterone, cortisol, adrenaline or noradrenaline levels (32). The relatively greater effect of alcohol on systolic compared with diastolic pressure, in younger subjects in particular, may be a clue to the mechanism and perhaps indicates a relative imbal- ance between the central nervous factors influencing cardiac output and the peripheral vascular effects of alcohol. Paradoxically the programmed administra- tion of alcohol over 1 to 2 weeks in young males leads to a decreased pressor reactivity to infused noradrenaline (34-38). The magnitude of the effects of alcohol on the prevalence of high blood pressure in different communities is difficult to evaluate, with estimates ranging from 11% (39) to 30% (40). Part of the problem is that most population studies tend to underestimate alcohol consumption, and that in most studies the pressure measurements were recorded over relatively short periods. Regular drinking also contributes indirectly to hypertension WHO Bulletin OMS. Vol 69 1991. 379 World Hypertension League by predisposing to obesity by virtue of the additional calories provided from both alcohol and other constituents of alcoholic beverages. It is therefore likely that the overall contribution of alcohol is significantly higher than that suggested by lower estimates based on multiple regression analyses. The influence of alcohol on hypertensive car- diovascular disease is also a crucial problem. Some of the effects of alcohol on conditions such as haemorrhagic stroke and left ventricular failure may well be masked in populations that are regularly screened for hypertension and in which treatment is introduced early. Studies from Yugoslavia (41) and USA (Honolulu) (42) suggest an increase in the incidence of stroke with higher levels of alcohol consumption. However, there are a number of studies in which people drinking up to three stan- dard glasses a day have a lower prevalence of coronary heart deaths (43, 44) and ischaemic stroke (44) compared with so-called non-drinkers. At present it is not clear whether these observations represent a true protective effect of lower doses of alcohol against occlusive vascular disease, or whet- her they are spurious. Possible effects which might result in a U- or J-shaped curve relating cardiovascu- lar mortality or stroke rates (45) and reported drinking habits are (i) misreporting of alcohol consumption, particularly denial by problem or former heavy drinkers, (ii) other differences in life-style distinguishing light drinkers from non- drinkers, e.g., with respect to exercise, diet and smoking habits, and (iii) previous reductions in drinking levels in those already informed that they have cardiovascular or related problems (46, 47). The last point has been argued strongly by Shaper et al. (46), on the basis of a nine-year follow-up of 7000 males seen in general practice. That population as a whole showed a U-shaped relationship between cardiovascular mortality and reported drinking habits. However, this only applied to those who already had cardiovascular disease, gout or diabetes at the outset and, in contrast, the initially disease- free subjects showed a much lower mortality rate which was weakly and positively related to drinking levels. It could, however, be equally argued from these data that alcohol was protective only in those most prone to cardiovascular disease. This impor- tant problem remains unresolved. Conclusions and recommendations In summary, there is now clear-cut evidence for a pressor effect of alcohol in regular drinkers. This effect is probably linear throughout the entire range of alcohol consumption, but appears to increase in magnitude, particularly in those drinking three or more standard drinks a day. The effect is seen in both men and women and is additive to that of obesity and oral contraceptive use. This pressor effect is at least partially reversible over 1 to 6 weeks by reducing the alcohol consumption. Possible interactions between alcohol and other factors in- fluencing blood pressure, such as dietary sodium, exercise and obesity (48), are of interest. Three large factorial studies in Perth suggest that only moderating the alcohol intake (but not sodium restriction (50)) or exercise training (51) lowers blood pressures in regular drinkers, at least in the relatively short term of 4-6 weeks. On the other hand, weight reduction and alcohol moderation had independent and additive effects on blood pressure over 5 months (52). For the individual hypertensive who drinks, alcohol reduction to less than three drinks a day may diminish or abolish his or her requirements for antihypertensive drug therapy. Doctors and health professionals should be aware that alcohol may be contributing to blood pressure elevation and to refractoriness in anti-hypertensive drug therapy. In addition, the problem drinker is especially likely to fail to take the prescribed treatment on a regular basis. A detailed drinking history should be recorded on all hypertensive patients as the problem is not just one for the "problem" drinker. The term "social" drinking is not adequate to describe drinking habits which may be relevant to blood pressure elevation. Laboratory investigation of mean corpuscular volume and liver enzymes, although not particularly sensitive or specific, nevertheless can help identify some heavier drinkers (22, 49) among the hypertensive popula- tion. Further research is required on the mechan- isms of alcohol-related hypertension and on the question whether or not there is any protective effect from lower levels of drinking on coronary heart disease or ischaemic strokes. References 1. Lian, C. L'alcoolisme: cause d'hypertension arterielle. Bulletin de l'Acad6mie de M6decine, 74: 525-528 (1915). 2. Ashley, M.J. & Rankin, J.G Alcohol consumption and hypertension: the evidence for hazardous drinking and alcohol populations. Austral. NZ j. med., 9: 201-206 (1979). 3. Beilin, L.J. & Arkwright, P.D. Alcohol and hyperten- sion. In: Robertson, J.I.S., ed. Clinical aspects of essen- tial hypertension-Handbook of hypertension, vol. 1. Amsterdam, Elsevier/North-Holland, 1983, pp. 44-63. 4. Beilin, L.J. et al. Alcohol-related hypertension. In: Strasser, T. & Ganten, D., ed. Mild hypertension: from 380 WHO Bulletin OMS. Vol 691991. Alcohol and hypertension drug trials to practice. New York, Raven Press, 1987, pp. 147-157. 5. MacMahon, S. Alcohol consumption and hypertension. Hypertension, 9: 111-121 (1987). 6. Freidman, G.D. et al. Alcohol intake and hypertension. Ann. int. med., 98: 846-849 (1983). 7. Klatsky, A.L. et al. Alcohol consumption and blood pressure. New Eng. j. med., 296: 1194-1200 (1977). 8. Wallace, R.B. et al. Alteration in blood pressures asso- ciated with combined alcohol and oral contraceptive use-the Lipid Research Clinics Prevalence Study. J. chron. dis., 35: 251-257 (1982). 9. Arkwright, P.D. et al. Effects of alcohol use and other aspects of life-style on blood pressure levels and preval- ence of hypertension in a working population. Circula- tion, 66: 60-66 (1982). 10. Milon, H. et al. Alcohol consumption and blood pres- sure in a French epidemiological study. Europ. heart j., 3(Supp.C): 59-64 (1982). 11. Ueshima, H. et al. Alcohol intake and hypertension among urban and rural Japanese populations. J. chron, dis., 37: 585-592 (1984). 12. Klatsky, A.L. et al. The relationships of alcoholic bever- age use and other traits to blood pressure: a new Kaiser Permanent study. Circulation, 73: 628-636 (1986). 13. Intersalt Cooperative Research Group. Intersalt: an international study of electrolyte excretion and blood pressure. Results for 24-hour urinary sodium and potas- sium excretion. Brit. med. j., 297: 319-328 (1988). 14. Kannel, W.B. & Sorlie, P. Hypertension in Framing- ham. In: Oglesby, P., ed. Epidemiology and control of hypertension. New York, Stratton, 1975, pp. 553-593. 15. Harburg, E. et al. Community norms of alcohol usage and blood pressure: Tecumseh, Michigan. Am. j. pub. hlth, 70: 813-820 (1980). 16. Cooke, K.M. et al. Blood pressure and its relationship to low levels of alcohol consumption. Clin. exp. pharma- col. physiol., 10: 229-233 (1983). 17. Mitchell, P.l. et al. Role of alcohol in the aetiology of hypertension. Med. j. Austral., 2: 198-200 (1980). 18. Criqui, M.H. et al. Alcohol consumption and blood pressure: the Lipid Research Clinics' Prevalence Study. Hypertension, 3: 557-565 (1981). 19. Cairns, V. et al. Alcohol consumption as a risk factor for high blood pressure-Munich blood pressure study. Hypertension, 6: 124-131 (1984). 20. Salonen, J.T. et al. Relation of blood pressure to reported intake of salt, saturated fats, and alcohol in healthy middle-aged population. J. epidemiol. commun- ity health, 37: 32-37 (1983). 21. Beevers, D.G. Alcohol and hypertension. Lancet, 2: 114-115 (1983). 22. Henningsen, N.C. et al. Hypertension, levels of serum gamma glutamyl transpeptidase and degree of blood pressure control in middle-aged males. Acta medica Scandinavica, 207: 245-251 (1980). 23. Ramsay, L.E. Alcohol and hypertension. Lancet, 2: 300 (1977). 24. Dyer, A.R. et al. Alcohol, cardiovascular risk factors and mortality-the Chicago experience. Circulation, 64: 20-27 (1981). 25. Gordon, T. & Kannel, W.B. Drinking and its relation to smoking, BP, blood lipids and uric acid (the Framing- ham Study). Arch. intern. med., 143: 1366-1374 (1983). 26. Puddey, I.B. et al. Evidence for a direct effect of alcohol consumption on blood pressure in normotensive man-a randomised controlled trial. Hypertension, 7: 707-713 (1985). 27. Puddey, I.B. et al. Regular alcohol use raises blood pressure in treated hypertensive subjects-a random- ized controlled trial. Lancet, 1: 647-651 (1987). 28. Malhoutra, H. et al. Pressor effect of alcohol in nor- motensive and hypertensive subjects. Lancet, 2: 584-586 (1985). 29. Howes, L.G. & Reid, J.L. The effects of alcohol on local, neural and humoral cardiovascular regulation. Clin. sci., 71: 111-115 (1986). 30. Potter, J.F. & Beevers, D.G. Pressor effect of alcohol in hypertension. Lancet, 1: 119-122 (1984). 31. Arkwright, P.D. et al. Alcohol, personality and predis- position to essential hypertension. J. hypertension, 1: 365-371 (1983). 32. Arkwright, P.D. et al. The pressor effect of moderate alcohol consumption in man: a search for mechanisms. Circulation, 66: 515-519 (1982). 33. Puddey, I.B. et al. Alcohol stimulation of renin release in man-its relation to the hemodynamic, electrolyte and sympatho-adrenal responses to drinking. J. clin. endo- crinol. metabol., 61: 37-42 (1985). 34. Howes, L. & Reid, J. Changes in blood pressure and autonomic reflexes following regular, moderate alcohol consumption. J. hypertension, 4: 421-425 (1986). 35. Puddey, I.B. et al. Alcohol consumption, age and personality characteristics as important determinants of within-subject variability in blood pressure. J. hyperten- sion, 6: S617-619 (1988). 36. Abdel-Rahman, A.A. & Wooles, W.R. Ethanol-induced hypertension involves impairment of baroreceptors. Hypertension, 10: 67-73 (1987). 37. Smals, A.G. et al. Alcohol-induced Cushingoid syn- drome. Brit. med. j., 2: 1298 (1976). 38. Howes, L.G. & Reid, J.L. Decreased vascular respon- siveness to noradrenaline following regular ethanol con- sumption. Brit. j. clin. pharmacol., 20: 669-674 (1985). 39. MacMahon, S.W. et al. Obesity, alcohol consumption and BP in Australian men & women-the National Heart Foundation of Australia Risk Factor Prevalence Study. J. hypertension, 2: 85-91 (1984). 40. Mathews, J.D. Alcohol and hypertension. Austral. NZ j.med., 9: 124-128 (1979). 41. Kozararevic, D. et al. Frequency of alcohol consump- tion and morbidity and mortality: the Yugoslavia cardio- vascular disease study. Lancet, 1: 613-616 (1980). 42. Blackwelder, W.C. et al. Alcohol and mortality: the Honolulu Heart Study. Am. j. med., 68: 164 (1980). 43. Marmot, M.G. et al. Alcohol and mortality: a U-shaped curve. Lancet, 1: 580-583 (1981). 44. Stampfer, M.J. et al. A prospective study of moderate alcohol consumption and the risk of coronary disease and stroke in women. New Eng. j. med., 319: 267-273 (1988). 45. Gill, J. et al. Stroke and alcohol consumption. New Engl. j. med., 315: 1041-1046 (1986). WHO Bulletin OMS. Vol 69 1991. 381 World Hypertension League 46. Shaper, A.G. et al. Alcohol and mortality in British men: explaining the U-shaped curve. Lancet, 2: 1267-1273 (1988). 47. Alcohol and mortality: the myth of the U-shaped curve. Lancet, 2:1292-1293 (1988) (Editorial). 48. Kornhuber, H.H. et al. Alcohol and obesity: a new look at high blood pressure and stroke. Europ. arch. psy- chiatr. neurol. sci., 234: 357-362 (1985). 49. Kornhuber, H.H. et al. Alcohol, smoking and body build: obesity as a result of the toxic effect of 'social' alcohol consumption. Clin. physiol. biochem., 7: 203-216 (1989). 50. Parker, M. et al. Two-way factorial study of alcohol and salt restriction in treated hypertensive men. Hyperten- sion, 16: 398-406 (1990). 51. Cox, K.L. et al. A controlled comparison of the effects of exercise and alcohol on blood pressure and serum high-density lipoprotein cholesterol in sedentary males. Clin. exp. pharmacol. physiol., 17: 251-256 (1990). 52. Puddey, I.B. et al. Alcohol restriction and weight reduc- tion have independent and additive effects in lowering blood pressure-A randomized controlled intervention study in overweight men. J. hypertension, 8 (suppl. 3): 533 (1990) (abstract). 382 WHO Bulletin OMS. Vol 691991.