Alcohol consumption and cardiovascular risk in hypertensives with left ventricular hypertrophy: the LIFE study

Transcription

1 (24) 18, & 24 Nature Publishing Group All rights reserved /4 $3. ORIGINAL ARTICLE Alcohol consumption and cardiovascular risk in hypertensives with left ventricular hypertrophy: the LIFE study HM Reims 1, SE Kjeldsen 1,2, WE Brady 3, B Dahlöf 4, RB Devereux 5, S Julius 2, G Beevers 6, U de Faire 7, F Fyhrquist 8, H Ibsen 9, K Kristianson 1, O Lederballe-Pedersen 11,LH Lindholm 12, MS Nieminen 8, P Omvik 13, S Oparil 14 and H Wedel 15, for the LIFE study group 1 Ullevaal University Hospital, Oslo, Norway; 2 University of Michigan Hospital, Ann Arbor, MI, USA; 3 Merck Research Laboratories, West Point, PA, USA; 4 Sahlgrenska University Hospital/O stra, Göteborg, Sweden; 5 Cornell Medical Center, New York, NY, USA; 6 City Hospital, Birmingham, UK; 7 Karolinska University Hospital, Stockholm, Sweden; 8 Helsinki University Central Hospital, Finland; 9 Glostrup University Hospital, Denmark; 1 Merck Research Laboratories Scandinavia, Stockholm, Sweden; 11 Viborg Hospital, Denmark; 12 Umeå University, Sweden; 13 Haukeland University Hospital, Bergen, Norway; 14 University of Alabama, Birmingham, AL, USA; 15 The Nordic School of Public Health, Göteborg, Sweden The Losartan Intervention For End point reduction in hypertension (LIFE) study showed superiority of losartan over atenolol for reduction of composite risk of cardiovascular death, stroke, and myocardial infarction in hypertensives with left ventricular hypertrophy. We compared hazard ratios (HR) in 4287 and 685 participants who reported intakes of 1 7 and 48 drinks/week at baseline, respectively, with those in 4216 abstainers, adjusting for gender, age, smoking, exercise, and race. Within categories, clinical baseline characteristics, numbers randomized to losartan and atenolol, and blood pressure (BP) lowering were similar on the drug regimens. Overall BP control (o14/9 mmhg) at end of follow-up was similar in the categories. Composite end point rate was lower with 1 7 (24/1 years; HR.87, Po.5) and 48 drinks/week (26/1 years; HR.8, NS) than in abstainers (27/1 years). Myocardial infarction risk was reduced in both drinking categories (HR.76, Po.5 and HR.29, Po.1, respectively), while stroke risk tended to increase with 48 drinks/ week (HR 1.21, NS). Composite risk was significantly reduced with losartan compared to atenolol only in abstainers (HR.81 95% confidence interval, CI (.68,.96), Po.5), while benefits for stroke risk reduction were similar among participants consuming 1 7 drinks/ week (HR.73, Po.5) and abstainers (HR.72, Po.1). Despite different treatment benefits, alcoholtreatment interactions were nonsignificant. In conclusion, moderate alcohol consumption does not change the marked stroke risk reduction with losartan compared to atenolol in high-risk hypertensives. Alcohol reduces the risk of myocardial infarction, while the risk of stroke tends to increase with high intake. (24) 18, doi:1.138/sj.jhh Published online 22 April 24 Keywords: alcohol; cardiovascular risk factors; losartan; atenolol; LIFE study Introduction Prospective studies have suggested an inverse relationship between low-to-moderate alcohol consumption and the risk of coronary heart disease, but an increase in risk at higher levels of consumption. 1,2 Although heavy alcohol consumption increases the risk of ischaemic as well as haemorrhagic stroke, moderate intake appears to reduce the risk of ischemic stroke. 3 Low-to-moderate Correspondence: Dr HM Reims, Department of Cardiology, Ullevaal University Hospital, Kirkeveien 166, N-47 Oslo, Norway. henrik.reims@ioks.uio.no Published online 22 April 24 alcohol consumption seems to have beneficial effects on some cardiovascular (CV) risk factors, 4 while heavy alcohol consumption has been shown to increase the risk of hypertension. 5 Thus, the relationship between alcohol intake and CV disease is complex, and may differ according to existing risk factors and disease history. To our knowledge, the impact of alcohol intake on CV risk has not previously been assessed specifically in high-risk hypertensive patients with left ventricular hypertrophy (LVH). The Losartan Intervention For End point reduction in hypertension (LIFE) study was designed to test the hypothesis that antagonizing the actions of angiotensin II by the selective AT 1 -receptor antago-

2 382 nist losartan would reduce CV morbidity and mortality compared to antihypertensive treatment with the beta-blocker atenolol in hypertensive patients with electrocardiographic (ECG) LVH. 6 Evaluation of the influence of alcohol consumption on CV event rates was defined as a tertiary objective of the LIFE study. 6 In order to evaluate alcohol consumption as a predictor of CV events, we performed a detailed analysis of its impact on various parameters in the study population. The main objectives of this analysis were to study (1) effects of alcohol consumption on the composite primary and individual secondary end points, (2) possible contributing factors for such effects, that is, effects of alcohol consumption on baseline risk factors (including blood pressure (BP) levels), and (3) possible differences in the effects of losartan and atenolol between alcohol consumption categories. Materials and methods Study design The LIFE study was an investigator-initiated, prospective, multinational, multicentre, double-blind, double-dummy, randomized, active-controlled, parallel-group study. The primary objective was to compare the long-term effects of the angiotensin II receptor antagonist losartan with those of the betaadrenergic receptor blocker atenolol on CV morbidity and mortality (a composite of CV mortality, stroke, and myocardial infarction) in 9193 patients with essential hypertension and ECG LVH. The rationale, objectives, design and methods, 6 baseline characteristics, 7 12-month follow-up data, 8 and primary outcome 9 have been published previously. The blinded treatment period continued until 16 September, 21, when at least 14 patients had experienced a primary endpoint, defined as CV death, nonfatal clinically evident acute myocardial infarction (AMI), or nonfatal cerebral stroke. Participants then had a follow-up clinic visit or at least a vital status check within 6 weeks. All clinical data were verified from source documents. The trial protocol was approved by all concerned ethics committees and conducted in accordance with the Declaration of Helsinki. The study was overseen by an independent Data and Safety Monitoring Board. 6 All participants gave written informed consent. Subjects Alcohol in the LIFE study The LIFE study population consisted of subjects with essential hypertension and ECG-documented LVH. Sitting BP was recorded 24 h postdose (range h). To be included, patients had to have a mean trough sitting diastolic BP of mmhg and/or a mean sitting systolic BP of 16 2 mmhg after 1 2 weeks on single-blind placebo treatment. Those with secondary hypertension, myocardial infarction, or stroke within the previous 6 months; angina pectoris requiring treatment with beta-blockers or calcium antagonists; heart failure or left ventricular ejection fraction o4%; or a disorder that, according to the treating physician s opinion, required treatment with study medication or related drugs were excluded. In the whole population, the mean age at inclusion was 66.9 years, 54% were women, and mean baseline BP was 174.4/ 97.8 mmhg. Smoking, alcohol intake, and exercise status were assessed at baseline. Subjects who reported no alcohol intake were slightly older on average than those who consumed alcohol, and there were more men than women in the higher categories of consumption (Table 1). Smoking was also more common with increasing alcohol consumption (Table 1). There was a slight tendency towards higher diastolic BP with increasing alcohol intake (Table 2). Baseline blood hemoglobin and serum uric acid levels also increased with higher alcohol intake. On the other hand, serum high-density lipoprotein (HDL) cholesterol level tended to be higher in alcohol consumers than in abstainers. The subjects medical history differed between the drinking categories, with a tendency towards more coronary, cerebral, and peripheral vascular disease and type II diabetes in subjects who did not consume alcohol than in those who drank 1 7 drinks per week (Figure 1). Other baseline characteristics of the population have been published previously. 7 Clinical end points The primary end point was a composite of CV death, stroke, and myocardial infarction. Other prespecified outcome measures were total mortality, angina pectoris or heart failure requiring hospitalization, coronary or peripheral revascularization procedures, resuscitated cardiac arrest, and new-onset diabetes mellitus. An Endpoint Classification Com- Prevalence (%) Alcohol Consumption (Drinks/wk) >1 CHD CVD PVD AF CHF Type 2 DM Figure 1 Medical history by alcohol consumption as prevalence of disease at baseline. CHD, coronary heart disease; CVD, cerebrovascular disease; PVD, peripheral vascular disease; AF, atrial fibrillation; CHF, congestive heart failure; DM, diabetes mellitus.

3 Alcohol in the LIFE study Table 1 Demographic and lifestyle characteristics by alcohol consumption 383 Alcohol consumption (n ¼ 4216) 1 4/week (n ¼ 363) 5 7/week (n ¼ 684) 8 1/week (n ¼ 314) 41/week (n ¼ 371) Age (years (s.d.)) 68.1 (6.9) 66.3 (6.9) 65.3 (6.8) 65.2 (6.9) 64.8 (6.7) X65, n (%) 2886 (68) 2125 (59) 355 (52) 156 (5) 178 (48) Female, n (%) 2865 (68) 1741 (48) 221 (32) 72 (23) 64 (17) White, n (%) 3752 (89) 3447 (96) 654 (96) 299 (95) 347 (94) Black, n (%) 357 (8) 119 (3) 24 (4) 12 (4) 2 (5) Hispanic, n (%) 67 (2) 23 (.6) 4 (.6) 3 (1) 3 (.8) Asian, n (%) 29 (.7) 11 (.3) 2 (.3) 1 (.3) Other race, n (%) 11 (.3) 3 (.1) Cigarette smoking, n (%) Never 2668 (63) 1628 (45) 213 (31) 74 (24) 71 (19) Previous smokers 136 (25) 1363 (38) 38 (45) 149 (47) 177 (48) 1 5/day 154 (4) 194 (5) 54 (8) 26 (8) 25 (7) 6 1/day 161 (4) 194 (5) 32 (5) 15 (5) 26 (7) 11 2/day 132 (3) 166 (5) 61 (9) 37 (12) 39 (11) 42/day 65 (2) 55 (2) 16 (2) 13 (4) 32 (9) Exercise, n (%) 1119 (27) 651 (18) 19 (16) 68 (22) 73 (2) p3 min twice/week 111 (26) 983 (27) 163 (24) 67 (21) 93 (25) 43 min twice/week 1994 (47) 1965 (55) 411 (6) 179 (57) 25 (55) Table 2 Clinical and biochemical baseline characteristics by alcohol consumption Alcohol consumption Variables (n ¼ 4216) 1 4/week (n ¼ 363) 5 7/week (n ¼ 684) 8 1/week (n ¼ 314) 41/week (n ¼ 371) BMI (kg/m 2 ) 28 (5) 28 (4) 28 (4) 27 (4) 28 (5) Systolic BP (mmhg) 175 (14) 174 (14) 173 (15) 175 (14) 175 (15) Diastolic BP (mmhg) 97 (9) 98 (8) 99 (9) 1 (8) 99 (9) Heart rate (bpm) 75 (11) 73 (11) 73 (11) 75 (12) 74 (12) Cornell product (mm ms) (153.2) (122.9) (119.4) (126.8) (193.) Sokolow Lyon voltage (mm) 29.6 (1.6) 3.3 (1.5) 3.8 (1.2) 3.8 (9.8) 3.6 (9.6) Haemoglobin (g/l) 14. (12.) (11.6) (11.6) (11.7) (12.) Sodium (mmol/l) 14.5 (2.5) 14.2 (2.5) 14.3 (2.6) (2.7) 14.1 (2.7) Potassium (mmol/l) 4.16 (.4) 4.16 (.38) 4.22 (.47) 4.19 (.39) 4.19 (.4) Serum glucose (mmol/l) 6.15 (2.35) 5.9 (1.99) 5.9 (2.1) 5.94 (2.41) 6.3 (2.2) Serum creatinine (mmol/l) (21.51) (19.16) 87.8 (18.9) (2.2) (17.92) Total cholesterol (mmol/l) 6.5 (1.16) 6.6 (1.1) 6.6 (1.7) 5.89 (1.7) 5.82 (1.8) HDL cholesterol (mmol/l) 1.45 (.42) 1.51 (.45) 1.58 (.46) 1.57 (.46) 1.59 (.48) Serum uric acid (mmol/l) 32.1 (76.2) 33.8 (75.8) 35.1 (79.7) (8.4) (89.1) Urine albumin (mg/l) (232.55) (223.9) 56.8 (191.49) (317.5) (243.7) Urine creatinine (mmol/l) 9.35 (5.5) 1.1 (5.97) 1.9 (5.96) 1.74 (6.7) 1.55 (6.7) Urine albumin/creatinine (mg/mmol) 8.68 (36.3) 6.41 (31.6) 6.16 (23.85) 1.23 (52.32) 7.4 (23.1) Data are presented as mean (s.d.). mittee (ECC) consisting of two experienced clinicians, blinded to therapy, reviewed clinical records concerning all CV events reported by clinical centres to determine if they met protocol criteria for end points. Adverse experiences were monitored throughout the study. The ECG core center also assessed silent and unrecognized myocardial infarctions. The product of QRS duration times Cornell voltage (with adjustment of 8 mm in women and a partition value of 4244 mm ms) and Sokolow-Lyon voltage 438 mm were chosen to recognize LVH. 6 ECG coding and LVH criteria All ECGs were centrally assessed for signs of LVH, and were Minnesota coded at one reading centre. Biochemical methods Two central laboratories measured serum and plasma concentrations of haemoglobin, creatinine,

4 384 Alcohol in the LIFE study glucose, uric acid, total cholesterol, HDL cholesterol, and urine concentrations of albumin and creatinine. Serum uric acid was measured by an enzymatic uricase method performed on a Hitachi analyzer (Roche Diagnostics, Indianapolis, IN, USA) (US laboratory) and PAP uricase method performed on a DAX96 analyzer (Roche Diagnostics, Indianapolis, IN, USA) (European laboratory). Serum creatinine was analysed using the Jaffé reaction without deproteinizing and then quantified by a photometric method by using the same analyzer. lowered slightly more in the highest drinking category (Figure 3) than in the nondrinking category (by 31 vs 29 mmhg with atenolol and 33 vs 3 mmhg with losartan). Overall, the presence of systolic BP control (o14 mmhg) at the end of follow-up was identical (48%) in those who consumed no alcohol, 1 7 drinks per week, and 48 drinks per week. Diastolic control (o9 mmhg) was also similar within the respective categories (89, 88, and 87%), as was the presence of both systolic and diastolic control (47, 46, and 47%). Statistical methods Analyses of BP and CV end points were based on the intention-to-treat principle. All randomized patients were included in their randomized treatment group, and all available follow-up data were included from randomization to the end of the study. Participants who experienced more than one end point event were counted as having had an event in all relevant end point analyses; however, only the first event in a specific category was counted in individual analyses. The effect of alcohol consumption on cardiovascular end points was assessed by Cox regression models with adjustment for gender, age, smoking status, exercise habits, and race. An unadjusted analysis was also performed to validate the adjusted results. Treatment effects were measured by hazard ratios (HR) (relative risks) and 95% confidence intervals (CI) from the Cox regression models, using three drinking categories: nondrinkers (n ¼ 4216), participants consuming 1 7 drinks per week (n ¼ 4287), and participants consuming 48 drinks per week (n ¼ 685). Changes in BP and ECG measures of LVH were assessed with three drinking categories from baseline to the end of follow-up or at the last visit before a primary end point occurred, if one did. All tests were performed at two-sided 5% significance levels. Role of the funding source All study data reside in the Merck & Co., Inc database. Merck provided the study steering committee with free access to all data. The steering committee is free to interpret data and write manuscripts. Results Alcohol consumption, treatment regimens, and BP control The numbers of patients randomized to the two study drugs were similar within the main drinking categories, but those who had eight drinks or more tended to use more antihypertensive drugs at the end of follow-up than the other two categories (Figure 2). On both regimens, systolic BP was Alcohol consumption, treatment regimens, and ECG changes Overall, there appeared to be a larger reduction in the Cornell product in subjects who consumed 48 drinks per week ( 324 (standard error (s.e.) 33) mm ms) compared to 1 7 drinks per week and no alcohol ( 215 (s.e. 12) and 181 (s.e. 13) mm ms, respectively). The same pattern was present for the Sokolow Lyon voltage, which was reduced by 4.9 (s.e..3) mm with 48 drinks per week, 3.9 (s.e..1) mm with 1 7 drinks per week, and 3.2 (s.e..1) mm without alcohol. This general pattern was present on both drug regimens (Figure 4). % Within Categories BP Change (mmhg) Losartan Atenolol n = 217 n = No Alcohol Intake Losartan Atenolol n = 213 n = 2157 Systolic Losartan Atenolol Losartan n = 366 Atenolol n = Drinks per Week >8 Drinks per Week 1 mg + 1 mg 5 mg + 5 mg Off Study Drug Figure 2 Study drug at end of follow-up or end point using three categories of alcohol consumption. Percents are given within each category of alcohol consumption. þ indicates therapy additional to given dose of study drug. Diastolic Losartan Atenolol -2 Alcohol Consumption Drinks/wk -35 >8 Drinks/wk Figure 3 Lowering of BP by treatment and alcohol consumption.

5 Alcohol in the LIFE study Cornell Product (mm ms) Losartan Atenolol a Men Women 8-1 >1 385 Sokolow-Lyon Voltage (mm) Figure 4 Regression of ECG LVH by treatment in the main categories of reported alcohol consumption, assessed by change in Cornell product (upper panel) and Sokolow Lyon voltage (lower panel). Open bars represent nondrinkers, grey bars subjects consuming 1 7 drinks per week, and black bars subjects consuming 48 drinks per week. Error bars represent standard error. b >1 Current Smokers Previous Smokers Non-Smokers Figure 5 Primary composite end point rate per 1 years of follow-up according to reported weekly alcohol consumption. Data shown by gender (a) and smoking status (b). Effects of alcohol consumption on end points The primary composite end point rate was significantly lower in subjects who consumed 1 7 drinks per week compared to abstainers (Table 3). There was no significant effect for the subcategories, except in the group consuming 8 1 drinks per week (adjusted HR.69, 95% CI (.48,.99), Po.5). The protective effect of an intake of 1 7/week on the primary end point appeared to be dependent on gender, the adjusted HR being.77 (95% CI (.63,.94), Po.1) in women and.94 (95% CI (.79, 1.11), P ¼.459) in men (P ¼.223 for interaction). There seemed to be a J-shaped relationship between alcohol intake and the primary end point in nonsmokers, which was disrupted in smokers (P ¼.41 for interaction) (Figure 5). The adjusted analysis revealed no significant effects of alcohol intake on CV mortality, although there was a strong trend (P ¼.81) for a benefit in subjects who had 1 7 drinks per week (Table 3). There was a weak trend towards lower stroke risk in subjects who had 1 7 drinks per week and a trend for increased risk with higher intake. Myocardial infarction was less common among subjects who had 1 7 or 48 drinks per week than in nondrinkers, and the benefit was significant also for the subdivided categories (Figure 6). Treatment effect by alcohol consumption In both treatment groups, end point rates were higher in abstainers than in participants consuming 1 7 drinks per week, except for identical rates of myocardial infarction in losartan-treated patients (Table 4). Among nondrinkers, the adjusted reduction in the primary end point rate in losartan-treated patients compared to atenolol treatment was 19.2%, (95% CI (4.1, 31.9), P ¼.15), while there was a weak trend (7% risk reduction) in favour of losartan treatment among subjects consuming 1 7 drinks per week, and a stronger trend (32.1% risk reduction) in those consuming 48 drinks per week (Figure 7). The hazard ratios were only slightly different in the unadjusted analysis (Table 4). The adjusted risk reduction for stroke in subjects assigned to losartan compared to atenolol was similar in nondrinkers (28.4%, 95% CI (8.2, 44.2), P ¼.9) and subjects consuming 1 7 drinks per week (26.6%, 95% CI (5.1, 43.3), P ¼.18), but did not reach statistical significance in the high drinking category (Figure 7). There were no significant differences between the drug regimens in the risk of myocardial infarction or CV death, despite a trend in favour of atenolol (P ¼.157) for myocardial infarction in those consuming 1 7 drinks per week. The results were

6 Alcohol in the LIFE study 386 Table 3 End points and hazard ratios by alcohol consumption HR (95% CI) Consumption Rate a Events (%) Unadjusted P Adjusted b P b Primary composite end point (12.6) 1 7 drinks/week (11.3).89 (.79, 1.1) NS.87 (.76,.99) o.5 48 drinks/week (11.7).95 (.75, 1.21) NS.8 (.63, 1.2) NS Cardiovascular death (5.3) 1 7 drinks/week (4.2).79 (.65,.96) o.5.83 (.68, 1.2) NS 48 drinks/week (4.7).91 (.63, 1.31) NS.81 (.55, 1.2) NS Stroke (6.) 1 7 drinks/week (5.6).92 (.77, 1.9) NS.96 (.8, 1.16) NS 48 drinks/week (7.2) 1.23 (.91, 1.67) NS 1.21 (.88, 1.67) NS Myocardial infarction (4.6) 1 7 drinks/week (4.2).89 (.73, 1.1) NS.76 (.62,.95) o.5 48 drinks/week (1.9).42 (.24,.73) o.1.29 (.16,.52) o.1 a Rate per 1 years of follow-up. HR vs nondrinking category. b Adjusted for smoking, exercise, gender, age, and race (white, black, and other). CV Death (1/1 yrs) Stroke (1/1 yrs) MI (1/1 yrs) P <.5 P <.5 P <.1 P < >1 Figure 6 Individual end point rates according to reported weekly alcohol consumption. P given for hazard ratios vs nondrinkers based on adjusted Cox model. similar when race was excluded from the adjusted analyses (data not shown). The alcohol-treatment interactions from adjusted models were not significant (P ¼.314 for the composite end point, P ¼.668 for CV death, P ¼.98 for stroke, and P ¼.225 for myocardial infarction). Discussion The present study showed that hypertensive patients with LVH with a low-moderate consumption of alcohol (1 7 drinks/week), as well as nondrinkers, had a marked benefit with regards to stroke prevention from treatment with losartan compared to atenolol. The results suggest that low-moderate consumers benefit less than nondrinkers from losartan compared to atenolol with regards to composite end point risk reduction. However, none of the alcohol-treatment interactions were statistically significant. Even in this high-risk population, alcohol intake reduced the overall risk of myocardial infarction in a dose-dependent manner, and low moderate intake reduced the risk of the primary composite end point. On the other hand, the adjusted risk of stroke tended to increase with high alcohol intake. Losartan reduced the risk of stroke by 28.4 and 26.6% compared to atenolol in nondrinkers and participants with a weekly consumption of 1 7 drinks, respectively. Thus, even with less than half of the LIFE population in each of these two subgroups, there were significant benefits comparable to that in the entire LIFE population (26%), 9 while the subgroup consuming 48 drinks weekly was too small to show a significant treatment benefit. The subgroup with a weekly consumption of 1 7 drinks differed from the abstainers by

7 Alcohol in the LIFE study Table 4 End point rates and treatment effect by alcohol consumption 387 Losartan Atenolol Consumption Rate a Events (%) Rate a Events (%) HR (95% CI) b P b Primary composite end point (11.4) (13.7).82 (.69,.97) o drinks/week (1.8) (11.8).92 (.77, 1.1) NS 48 drinks/week (9.8) (13.8).69 (.45, 1.8) NS Cardiovascular death (4.9) (5.7).86 (.66, 1.11) NS 1 7 drinks/week (4.) (4.5).9 (.68, 1.21) NS 48 drinks/week (3.8) (5.6).67 (.34, 1.35) NS Stroke (5.1) (6.9).73 (.57,.94) o.5 1 7/week 1. 1 (4.7) (6.4).73 (.57,.95) o.5 48/week (6.6) (7.8).82 (.47, 1.43) NS Myocardial infarction (4.6) (4.6).97 (.73, 1.28) NS 1 7 drinks/week (4.6) (3.8) 1.22 (.91, 1.63) NS 48 drinks/week 3. 5 (1.4) (2.5).54 (.18, 1.66) NS a Per 1 years of follow-up. b Losartan vs atenolol, unadjusted. Endpoint Alcohol Composite 1-7 >8 CV Death 1-7 >8 Stroke MI 1-7 >8 1-7 >8 Events (%) 53 (12.6) 485 (11.3) 8 (11.7) 224 (5.3) 181 (4.2) 32 (4.7) 253 (6.) 239 (5.6) 49 (7.2) 194 (4.6) 179 (4.2) 13 (1.9) Favours Losartan Favours Atenolol Figure 7 Total event rates for individual end points and adjusted HR for treatment with losartan vs atenolol. Participants are categorized as nondrinkers (n ¼ 4216), and consuming 1 7 (n ¼ 4287), and 48 (n ¼ 685) drinks per week. Bars represent 95% CI for hazard ratios. CV, cardiovascular; MI, myocardial infarction. showing no significant difference between the drug regimens for the primary composite end point. This contrast was largely due to the trend favouring atenolol for prevention of myocardial infarction in this category. The opposite trend in the high category of alcohol intake may be a chance finding, in view of the very large CI. Alternatively, as will be discussed, it may be explained by a particular benefit of angiotensin II antagonism in this group. Overall, the apparent effect of alcohol intake on CV disease in the present analysis parallels findings from previous studies, with an inverse relation between moderate alcohol intake and the risk of coronary heart disease, 1 and a tendency towards increased stroke risk with high intake. 3 The beneficial effect of low moderate intake on atherosclerotic disease was apparent already from the medical history of the study participants. HDL cholesterol levels tended to increase with increasing alcohol intake, but although alcohol intake is associated with higher HDL levels, the protective effect of alcohol intake on the risk of ischaemic heart disease is strongly dependent on low-density lipoprotein (LDL) cholesterol, 11 which was not measured in the present study. The incidence of myocardial infarction was reduced with increasing alcohol intake, even when adjusted for smoking. However, the protective effect for the composite end point was absent when smokers were viewed separately. Although the prevalence of smoking was relatively low in the present study, 46.7% of men were previous smokers, while only one-third had never smoked. 7 Among women, on the other hand, twothirds have never smoked. Therefore, men had a much greater total exposure to tobacco smoke than women. But even after the adjustment for smoking, the benefit of alcohol for the composite end point was not significant in men. This is somewhat surprising, as a larger range of alcohol intake appears to be protective, at least of coronary heart disease, in men than in women. 2 Recent heavy alcohol consumption has been found to significantly increase the risk of cardioembolic stroke, even after adjustment for hypertension, smoking, hyperlipidaemia, and diabetes. 12 Nevertheless, light-to-moderate alcohol consumption has been shown to reduce the overall risk of stroke and the risk of ischaemic stroke in men. 13 Even in male smokers, light alcohol consumption may reduce the risk of cerebral infarction and haemorrhage, while heavy drinking increases the risk of all stroke

8 388 Alcohol in the LIFE study subtypes. 14 The beneficial effect of alcohol was attributed partly to increased HDL concentrations, and the harmful effect to increased BP. 14 In a prospective study, it was shown that among men with an initially low consumption, the risk of CV disease was reduced in those who moderately increased their consumption compared to those who maintained a low consumption. 15 While alcohol may also act protectively by decreasing platelet aggregation, lowering fibrinogen levels, interacting with aspirin in prolonging bleeding time, and increasing release of plasminogen activator, 16 these antithrombotic effects may conversely increase the risk of haemorrhagic stroke. 3 Alcohol may also cause cardiac arrhythmias and cardiomyopathy, in addition to hypertension, 1 and thereby further increase the CV risk with high consumption. Endothelial dysfunction may play a role in the progression of coronary atherosclerosis. 17 The renin angiotensin system, through the actions of angiotensin II, is likely to be involved by enhancement of endothelial dysfunction, vascular LDL oxidation, adhesion molecule expression, and upregulation of cytokine expression. 18 Losartan has been shown to correct human small-artery structural changes and normalize endothelial function, while atenolol had no such effects despite similar lowering of BP during treatment of hypertension. 19 In patients with atherosclerosis, losartan improves endothelium-dependent epicardial coronary vasomotion. 2 Accordingly, this drug prevents angiotensin II-induced superoxide production, inhibition of nitric oxide (NO) production, and inhibition of endothelium-dependent NO-mediated dilation in coronary arterioles. 21 Losartan apparently also has anti-inflammatory properties by inhibition of the production and release of inflammatory mediators. 22,23 Animal studies have suggested favourable effects of low-dose alcohol exposure on endothelial function, but induction of endothelial dysfunction with higher doses. 24 Studies of the effects of alcohol on endothelial function in humans are limited, and have not established any beneficial or harmful effects of direct relevance to effects on CV outcomes. 22 However, recent studies have suggested beneficial effects of long-term moderate consumption 25 as well as acute intake 26 of alcohol on endothelial function. This is, however, probably antagonized by the increasing prevalence of cigarette smoking with increasing alcohol intake, because smoking is dosedependently associated with endothelial dysfunction, 27 facilitating inflammation, thrombosis, vasoconstriction, and atherosclerosis. 28 In addition, smoking is dose-dependently associated with higher serum concentrations of cholesterol, triglycerides, very-lowdensity lipoprotein cholesterol and LDL cholesterol, and lower HDL. 29 This probably explains the disruption of the J-shaped relationship between alcohol intake and the primary end point in smokers. Although alcohol intake is directly associated with hypertension, 1 there were essentially no differences in BP at baseline or in BP reduction between categories of alcohol intake. The effect of alcohol intake on the risk of developing hypertension has been shown to be dose-dependent, starting at lowto-moderate levels. 3 However, while high levels of consumption have been shown to independently increase the risk of hypertension, low moderate consumption was associated with an increase in BP in black, but not in white men. 5 Thus, because the vast majority of participants in the LIFE study were white men, our conclusion may be restricted to these. The mechanisms of alcohol-induced hypertension have remained unclear, 31 but hypotheses include involvement of the renin angiotensin system. 4 Thus, alcohol intoxication and hangover have been reported to increase plasma renin activity, possibly due to dehydration. 32 An acute increase in renin activity was later confirmed with more moderate alcohol ingestion, and suggested to be secondary to alcohol-induced changes in fluid and electrolyte balance and BP, but not sympathetic activation. 33 With this in mind, it is interesting to note that losartan decreased LVH by ECG criteria most markedly in patients with high alcohol intake (Figure 4). However, this trend was present among atenolol-treated patients as well. In this respect, one also has to remember that beta-adrenergic receptor blockers inhibit renin release 34 and suppress plasma angiotensin II levels. 35 Nevertheless, there seems to be a trend favouring losartan in preventing composite end points, CV death, and myocardial infarction in the subgroup of highest alcohol intake (Figure 7). Moderate alcohol consumption may also reduce the risk of type II diabetes, 36,37 as suggested by the somewhat lower diabetes prevalence in moderate drinkers in the present study. We did not adjust the results for diabetes at baseline, but the risk reduction for coronary heart disease with low moderate intake appears at least as large in diabetic men as that in men without diabetes. 38 In conclusion, the angiotensin II receptor antagonist losartan markedly reduces the risk of stroke compared to atenolol in hypertensive patients with LVH with a low moderate consumption of alcohol, as well as in abstainers. Although the risk reduction with losartan for the composite end point of CV death, stroke, and myocardial infarction could not be verified in moderate drinkers, no interaction between alcohol and treatment was observed. Even in this high-risk population, alcohol intake independently reduces the risk of myocardial infarction, while the risk of stroke tends to increase with high alcohol intake. Acknowledgements We thank Paulette A Lyle for her valuable administrative assistance. The trial was supported by an unrestricted grant from Merck & Co., Inc.

9 References 1 Goldberg IJ, Mosca L, Piano MR, Fisher EA. AHA Science Advisory. Wine and your heart: a science advisory for healthcare professionals from the Nutrition Committee, Council on Epidemiology and Prevention, and Council on Cardiovascular Nursing of the American Heart Association. Circulation 21; 13: Corrao G et al. Alcohol and coronary heart disease: a meta-analysis. Addiction 2; 95: Reynolds K et al. Alcohol consumption and risk of stroke: a meta-analysis. JAMA 23; 289: Klatsky AL. Alcohol and hypertension. Clin Chim Acta 1996; 246: Fuchs FD et al. Alcohol consumption and the incidence of hypertension: The Atherosclerosis Risk in Communities Study. Hypertension 21; 37: Dahlöf Bet al. The Losartan Intervention For Endpoint reduction (LIFE) in Hypertension study: rationale, design, and methods. The LIFE Study Group. Am J Hypertens 1997; 1: Dahlöf Bet al. Characteristics of 9194 patients with left ventricular hypertrophy: the LIFE study. Losartan Intervention For Endpoint Reduction in Hypertension. Hypertension 1998; 32: Kjeldsen SE et al. Lowering of blood pressure and predictors of response in patients with left ventricular hypertrophy: the LIFE study. Losartan Intervention For Endpoint. Am J Hypertens 2; 13: DahlöfBet al. Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomised trial against atenolol. Lancet 22; 359: Rimm EB, Klatsky A, Grobbee D, Stampfer MJ. Review of moderate alcohol consumption and reduced risk of coronary heart disease: is the effect due to beer, wine, or spirits. Br Med J 1996; 312: Hein HO, Suadicani P, Gyntelberg F. Alcohol consumption, serum low density lipoprotein cholesterol concentration, and risk of ischaemic heart disease: six year follow up in the Copenhagen male study. Br Med J 1996; 312: Hillbom M, Numminen H, Juvela S. Recent heavy drinking of alcohol and embolic stroke. Stroke 1999; 3: Berger K et al. Light-to-moderate alcohol consumption and risk of stroke among US male physicians. N Engl J Med 1999; 341: Leppälä JM et al. Alcohol consumption and stroke incidence in male smokers. Circulation 1999; 1: Sesso HD et al. Seven-year changes in alcohol consumption and subsequent risk of cardiovascular disease in men. Arch Intern Med 2; 16: Klatsky AL. Alcohol, coronary disease, and hypertension. Annu Rev Med 1996; 47: Al Suwaidi J et al. Long-term follow-up of patients with mild coronary artery disease and endothelial dysfunction. Circulation 2; 11: Brasier AR, Recinos III A, Eledrisi MS. Vascular inflammation and the renin angiotensin system. Arterioscler Thromb Vasc Biol 22; 22: Schiffrin EL, Park JB, Intengan HD, Touyz RM. Correction of arterial structure and endothelial dysfunction in human essential hypertension by the Alcohol in the LIFE study angiotensin receptor antagonist losartan. Circulation 2; 11: Prasad A, Halcox JP, Waclawiw MA, Quyyumi AA. Angiotensin type 1 receptor antagonism reverses abnormal coronary vasomotion in atherosclerosis. J Am Coll Cardiol 21; 38: Zhang C, Hein TW, Wang W, Kuo L. Divergent roles of angiotensin II AT1 and AT2 receptors in modulating coronary microvascular function. Circ Res 23; 92: Schieffer B et al. Expression of angiotensin II and interleukin 6 in human coronary atherosclerotic plaques: potential implications for inflammation and plaque instability. Circulation 2; 11: Krämer C et al. Angiotensin II receptor-independent antiinflammatory and antiaggregatory properties of losartan: role of the active metabolite EXP3179. Circ Res 22; 9: Puddey IB, Zilkens RR, Croft KD, Beilin LJ. Alcohol and endothelial function: a brief review. Clin Exp Pharm Physiol 21; 28: Teragawa H et al. Effect of alcohol consumption on endothelial function in men with coronary artery disease. Atherosclerosis 22; 165: Matsuo S et al. Effect of red wine and ethanol on production of nitric oxide in healthy subjects. Am J Cardiol 21; 87: Celermajer DS et al. Cigarette smoking is associated with dose-related and potentially reversible impairment of endothelium-dependent dilation in healthy young adults. Circulation 1993; 88: Puranik R, Celermajer DS. Smoking and endothelial function. Prog Cardiovasc Dis 23; 45: Craig WY, Palomäki GE, Haddow JE. Cigarette smoking and serum lipid and lipoprotein concentrations: an analysis of published data. Br Med J 1989; 298: Nakanishi N et al. Alcohol consumption and risk for hypertension in middle-aged Japanese men. J Hypertens 21; 19: Klatsky AL. Alcohol and cardiovascular diseases: a historical overview. Ann NY Acad Sci 22; 957: Linkola J et al. Renin aldosterone axis in ethanol intoxication and hangover. Eur J Clin Invest 1976; 6: Puddey IB, Vandongen R, Beilin LJ, Rouse IL. Alcohol stimulation of renin release in man: its relation to the hemodynamic, electrolyte, and sympatho-adrenal responses to drinking. J Clin Endocrinol Metab 1985; 61: Prichard BN, Tomlinson B. The additional properties of beta adrenoceptor blocking drugs. J Cardiovasc Pharmacol 1986; 8(Suppl 4): S1 S Blumenfeld JD et al. Beta-adrenergic receptor blockade as a therapeutic approach for suppressing the renin angiotensin aldosterone system in normotensive and hypertensive subjects. Am J Hypertens 1999; 12: Rimm EB et al. Prospective study of cigarette smoking, alcohol use, and the risk of diabetes in men. Br Med J 1995; 31: Perry IJ et al. Prospective study of risk factors for development of non-insulin dependent diabetes in middle aged British men. Br Med J 1995; 31: Ajani UA et al. Alcohol consumption and risk of coronary heart disease by diabetes status. Circulation 2; 12: