For personal use. Only reproduce with permission from The Lancet Publishing Group. Summary

Transcription

1 Cardiovascular morbidity and mortality in patients with diabetes in the Intervention For Endpoint reduction in hypertension study (LIFE): a randomised trial against atenolol Lars H Lindholm, Hans Ibsen, Björn Dahlöf, Richard B Devereux, Gareth Beevers, Ulf de Faire, Frej Fyhrquist, Stevo Julius, Sverre E Kjeldsen, Krister Kristiansson, Ole Lederballe-Pedersen, Markku S Nieminen, Per Omvik, Suzanne Oparil, Hans Wedel, Peter Aurup, Jonathan Edelman, Steven Snapinn, for the LIFE study group* Summary Background The most suitable antihypertensive drug to reduce the risk of cardiovascular disease in patients with hypertension and diabetes is unclear. In prespecified analyses, we compared the effects of losartan and atenolol on cardiovascular morbidity and mortality in diabetic patients. Methods As part of the LIFE study, in a double-masked, randomised, parallel-group trial, we assigned a group of 1195 patients with diabetes, hypertension, and signs of leftventricular hypertrophy (LVH) on electrocardiograms losartanbased or atenolol-based treatment. Mean age of patients was 7 years (SD 7) and mean blood pressure 177/9 mm Hg (1/1) after placebo run-in. We followed up patients for at least years (mean 7 years [1 1]). We used Cox regression analysis with baseline Framingham risk score and electrocardiogram-lvh as covariates to compare the effects of the drugs on the primary composite endpoint of cardiovascular morbidity and mortality (cardiovascular death, stroke, or myocardial infarction). Findings Mean blood pressure fell to 1/79 mm Hg (17/11) in losartan patients and 1/79 mm Hg (19/11) in atenolol patients. The primary endpoint occurred in 13 patients assigned losartan (n=5) and 139 assigned atenolol (n=9); relative risk 7 (95% CI 5 9), p= and 1 patients in the losartan and atenolol groups, respectively, died from cardiovascular disease; 3 ( 95), p=. Mortality from all causes was 3 and 1 in losartan and atenolol groups, respectively; 1 ( 5 ), p=. *Members listed at end of reference 1 Umeå University Hospital, Sweden (Prof L H Lindholm MD); Glostrup University Hospital, Denmark (H Ibsen MD); Sahlgrenska University Hospital/Östra, Gothenburg, Sweden (B Dahlöf MD); Cornell Medical Center, New York, NY, USA (Prof R B Devereux MD); City Hospital, Birmingham, UK (Prof G Beevers MD); Karolinska University Hospital, Stockholm, Sweden (Prof U de Faire MD); Helsinki University Central Hospital, Finland (Prof F Fyhrquist MD, Prof M S Nieminen MD); University of Michigan, Ann Arbor, Michigan, MI, USA (Prof S Julius MD); Ullevaal University Hospital, Oslo, Norway (S E Kjeldsen MD); Merck Research Laboratories Scandinavia, Stockholm (K Kristiansson PhD); Viborg Hospital, Denmark (O Lederballe-Pedersen MD); Haukeland University Hospital, Bergen, Norway (Prof P Omvik MD); University of Alabama, Birmingham, Alabama, AL, USA (Prof S Oparil MD); The Nordic School of Public Health, Gothenburg (Prof H Wedel PhD); Merck Research Laboratories, West Point, PA, USA (P Aurup MD, J Edelman MD, S Snapinn PhD) Correspondence to: Prof Lars H Lindholm, Department of Public Health and Clinical Medicine, Umeå University Hospital, SE-91 5 Umeå, Sweden ( larsh.lindholm@fammed.umu.se) Interpretation was more effective than atenolol in reducing cardiovascular morbidity and mortality as well as mortality from all causes in patients with hypertension, diabetes, and LVH. seems to have benefits beyond blood pressure reduction. Lancet ; 359: 1 1 See Commentary page 99 Introduction The Intervention For Endpoint reduction (LIFE) study was designed in the early 199s 1 when no drug class for the treatment of essential hypertension had been shown to be more effective in prevention of cardiovascular morbidity and mortality than -blockers and diuretics. The main hypothesis of LIFE was that selective angiotensin-ii type-1-receptor antagonism with losartan would be more effective than -blockade with atenolol in reducing cardiovascular morbidity and mortality in patients with essential hypertension and signs of left ventricular hypertrophy (LVH) on electrocardiograms. Diabetes mellitus doubles the risk of cardiovascular disease, 5 even in patients with hypertension who are already at high risk because of their high blood pressure. Since many patients with hypertension develop diabetes, this combination of risk factors will account for a large proportion of cardiovascular morbidity and mortality. The frequency of diabetes mellitus is increasing rapidly worldwide. 7 In the LIFE study, 1 we compared the long-term effects of once-daily losartan-based with atenolol-based antihypertensive treatment in patients with hypertension and LVH on the frequency of cardiovascular morbidity and mortality (a composite of cardiovascular mortality, stroke and myocardial infarction). The most suitable drug to prevent premature cardiovascular disease in diabetic patients with high blood pressure is unclear. Therefore, we analysed the outcome in the prespecified subgroup of patients who had diabetes mellitus at the start of the LIFE study. Methods Patients and procedures The LIFE study 1 was a double-masked, randomised, parallel-group study with double dummy drugs. The main outcome 1 and the complete study protocol with study design, organisation, clinical measures, endpoint definitions, exclusion criteria, reasons for choice of comparative agents, statistical power calculations, and baseline characteristics have been published. We included patients aged 55 years with hypertension (either treated or untreated) and signs of LVH on electrocardiograms. We gave patients placebo drugs for 1 weeks after which we assigned them to treatment groups if they had a sitting systolic blood pressure of 1 mm Hg, a diastolic pressure of mm Hg, or both. In both groups, we added hydrochlorothiazide and 1 THE LANCET Vol 359 March 3,

2 5 assigned losartan 5 available for intention-totreat analyses Figure 1: Trial profile 177 patients assessed for eligibility 9 randomised 1195 with diabetes at baseline 3 dropped out 11 withdrew consent 11 incomplete follow-up 1 vital status only 155 ineligible 131 did not meet protocol criteria 15 unwilling to participate 9 excluded for irregularites 799 without diabetes at baseline 9 assigned atenolol 9 available for intention-totreat analyses 3 dropped out 7 withdrew consent 11 incomplete follow-up 1 vital status only lost to follow-up other agents (but not -blockers, angiotensin-convertingenzyme inhibitors, or angiotensin-ii antagonists) if blood pressure remained high during follow-up (13%) of 9193 LIFE participants had a diagnosis of diabetes mellitus (most likely of type-) at baseline. We randomly assigned 5 participants to losartan and 9 to atenolol. Figure 1 shows the trial profile and table 1 shows baseline characteristics. 77 (%) participants came from the Nordic countries, 35 (31%) from the USA, and 3 (5%) from the UK. % were white. Before the start of the study, 95 patients had been treated with antihypertensive drugs, 7 (1%) in the losartan group and (%) in the atenolol group. At baseline, antidiabetic drugs, insulin, or both, had been given to 9 (5%) patients: 33 in the losartan and 3 in the atenolol group. 5 (%) had received oral drugs (sulphonylureas, biguanides, or both) and 1 (1%) insulin (table ). We gave all patients who were not on these drugs non-pharmacological treatment during the study; additionally, 139 of the remaining 3 losartan patients and 13 of the remaining 3 atenolol patients started taking antidiabetic drugs during follow-up. (1%) losartan patients and 7 (1%) atenolol patients were not treated with antidiabetic drugs or insulin during the trial. Groups were also well balanced for treatment with lipid-lowering drugs and aspirin during the study (table ). Treatment for diabetes was at the discretion of patients physicians. We enrolled patients from June, 1995, to May, 1997, if a screening electrocardiogram fulfilled criteria for LVH. 1 We followed up patients for at least years with regular visits and upward-titration of study drugs to reduce blood pressure to below 1/9 mm Hg. We defined diabetes mellitus according to the 195 WHO criteria. After the study end date in September, 1, patients had a followup clinic visit or at least a vital status check within weeks. Tests were done at two laboratories that assured comparability of measurements by crossvalidation, and all electrocardiograms were coded at the same electrocardiogram core centre. 1 We recorded sitting blood pressure h after study drug dose (range ). Statistical analysis We assessed all endpoints by intention-to-treat analysis. We included all randomised patients and all available follow-up data from randomisation to the study end (n=5) (n=9) All with diabetes (n=1195) All without diabetes (n=799) Demographic and clinical characteristics Age (years)* 7 ( ) 7 (7) 7 (7) 9 (7) Women 3 (5%) 33 (55%) 3 (53%) 39 (5%) Ethnic origin White 5 (%) 519 (5%) 5 (%) 77 (9%) Black (11%) 71 (%) 133 (11%) (5%) Hispanic (%) 13 (%) 5 (%) 75 ( 9%) Asian 5 ( 9%) 5 ( %) 1 ( %) 33 ( %) Other 1 ( %) 1 ( %) ( %) ( %) Blood pressure (mm Hg)* 17/97 (1/9) 177/9 (1/1) 177/9 (1/1) 17/9 (1/9) Heart rate (bpm)* 7 () 7 (11) 7 () 7 (11) BMI (kg/m )* 3 (5 3) 3 (5 ) 3 (5 5) 7 7 ( ) Cornell voltage-duration product (mm msec)* 5 (9) 9 (995) 9 (99) 17 (157) Sokolow-Lyon (mm)* (1 3) 3 (1 ) 5 (1 3) 3 (1 5) Framingham risk score* 3 ( ) 3 ( ) 3 ( 5) ( 9) Current smokers 7 (%) 9 (15%) 1 (1%) 1337 (17%) Medical history Any vascular diasease (35%) 1 (35%) (35%) 1 (%) Coronary heart disease 13 (%) 15 (%) 3 (%) 11 (15%) Cerebrovascular disease 9 (%) 71 (%) 1 (%) 5 (7%) Peripheral vascular disease (7%) 9 (%) 9 (7%) 31 (5%) Atrial fibrillation 3 (5%) (%) (7%) (3%) Isolated systolic hypertension 13 (1%) 13 (%) 35 (%) 191 (1%) Bpm=beats per minute. BMI=body-mass index. Data are number (%) unless otherwise indicated. *Data are mean (SD). Definition >1/ 9 mm Hg. Table 1: Baseline characteristics THE LANCET Vol 359 March 3, 15

3 Baseline During study Baseline During study Treatment Antidiabetic Sulfonylureas 3 (3%) 3 (5%) (3%) 39 (1%) Biguanides 71 (%) 1 (37%) 79 (13%) 3 (%) Insulin 9 (1%) 17 (9%) 9 (15%) 15 (7%) Lipid lowering Statins* 5 (9%) 11 (31%) (%) 195 (3%) Fibrates 9 (%) 19 (3%) 17 (3%) (5%) Aspirin 15 (7%) 7 (7%) 1 (7%) 9 (7%) Data are number (%). *Statins=3-hydroxy-3-methylglutaryl coenzyme A (HMGC-A) reductases. Table : Antidiabetic, lipid lowering, and aspirin treatment date. We included endpoints in analyses only if confirmed by the endpoint committee. We classed patients with more than one endpoint as having had an event in all relevant endpoint analyses; however, we counted only the first event in a specific category in individual analyses. We assessed the difference between treatment groups with respect to clinical events with a Cox regression model with degree of LVH (measured by Cornell voltage-duration product and Sokolow-Lyon voltage) and the Framingham risk score 9 at baseline as covariates. We chose this adjusted analysis before the trial to account for any difference in key risk predictors at baseline. We also did a secondary unadjusted analysis. We measured treatment effects by hazard ratios (relative risks) and 95% CIs derived from Cox regression models. We calculated the risk reduction for losartan versus atenolol as 1 (1 relative risk). We adjusted results for blood pressure according to Cox regression models with blood pressures throughout the trial as time-varying covariates. We analysed differences between groups in changes in electrocardiogram measures of LVH with the Wilcoxon rank-sum test, and the frequency of adverse experiences with Fisher s exact test. The level of two-sided significance was 5%. Role of the funding source Study data are in a Merck database. Merck provided the study steering committee with free access to all data. The steering committee was free to interpret data and write the paper and the outcome was validated independently by the steeering committee. Merck reviewed the paper. Results Groups were closely matched in demographic characteristics, severity of hypertension, prevalence of coexisting cardiovascular conditions, Framingham risk score, and electrocardiogram-based LVH criteria (table 1). Compared with the remaining LIFE Proportion of patients with first event (%) 1 Primary composite endpoint ARR: 5%, p= 31 URR: 7%, p= Figure : Kaplan-Meier curves for primary composite endpoint ARR=adjusted risk ratio. URR=unadjusted risk ratio. participants without diabetes, 1 patients with the disease had higher body-mass index, higher Framingham risk score, a higher prevalence of cardiovascular disease at baseline (table 1), higher systolic blood pressure (difference 3 mm Hg), lower diastolic pressure ( mm Hg), higher pulse pressure (5 mm Hg), and higher serum glucose concentration (9 [SD 3 ] vs 5 5 [1 ] mmol/l). Fewer patients with diabetes smoked than other LIFE participants. 1 Primary endpoints occurred in (%) patients during 559 patient-years of follow-up (table 3): 13 in the losartan group (39 per 1 patient-years) and 139 in the atenolol group (53 ), relative risk 7 (95% CI 5 9), p= losartan patients ( 5 per 1 patient-years) and 1 in the atenolol group (37 ) died from all causes ( 1, 5, p= ). 3 (13 per 1 patient-years) and 1 patients (1 ) in the losartan and atenolol groups, respectively, died from cardiovascular disease ( 3, 95, p=, table 3). Stroke occurred in 51 losartan and 5 atenolol patients (p= 5), and myocardial infarction in 1 and 5 patients, respectively (p= 373). 3 patients were admitted to hospital for heart failure (11 per 1 patient-years) in the losartan group and 55 ( 7) in the atenolol group ( 59, 3 9, p= 19). Figure shows Kaplan-Meier curves for the primary composite endpoint, and figures 3 and show Kaplan-Meier curves for individual endpoints and total mortality, respectively. Adjustment for baseline Framingham risk score and LVH had little effect on results (table 3). There was no evidence of interaction between treatment and sex for any of the endpoints. (n=5) (n=9) Adjusted p Unadjusted p Rate* n Rate* n hazard ratio (95% CI) hazard ratio (95% CI) Endpoint Primary composite endpoint (1%) (3%) 7 ( 5 9) ( 57 95) 17 Cardiovascular mortality 13 3 (%) 1 1 (1%) 3 ( 95) ( 1 9) 19 Stroke (all) (9%) 5 5 (11%) 79 ( ) 7 ( ) 19 Myocardial infarction (all) 15 1 (7%) (%) 3 ( ) ( 5 1 ) 31 Other prespecified endpoints Total mortality 5 3 (11%) 37 1 (17%) 1 ( 5 ) ( ) 1 Admitted to hospital for Angina pectoris (5%) (5%) 1 ( 1 7) 1 ( 1 ) 99 Heart failure 11 3 (5%) 7 55 (9%) 59 ( 3 9) ( 37 9) 13 Revascularisation 3 5 (11%) 7 (11%) 9 ( 1 ) 533 ( 3 1 ) 7 *Per 1 patient-years of follow-up. For degree of left-ventricular hypertrophy and Framingham risk score at randomisation. Table 3: Endpoints 1 THE LANCET Vol 359 March 3,

4 Proporrtion of patients (%) Proportion of patients with first event (%) Proportion of patients with first event (%) Cardiovascular mortality ARR: 3 %, p= URR: 3 %, p= Myocardial infarction (fatal and non-fatal) ARR: 17 1%, p= 373 URR: 19 %, p= Stroke (fatal and non-fatal) ARR: 1 %, p= URR: 1 %, p= Figure 3: Kaplan-Meier curves for individual endpoints ARR=adjusted risk ratio. URR=unadjusted risk ratio. We assigned patients with diabetes and untreated hypertension at baseline (n=37) to losartan (n=11) or atenolol (n=3). In this subgroup, the primary composite endpoint occurred in 17 (33 3 per 1 patient-years) losartan and 3 atenolol patients ( 7), four (7 ) and 15 ( ) died from cardiovascular disease, and five (9 1) and ( ) died from all causes, respectively. Mean follow-up from randomisation to death, loss to follow-up, or end of the study was 7 years (1 1). The proportion of patients who were given additional drugs Proportion of patients (%) 1 Total mortality ARR: 3 7%, p= URR: 1%, p= Figure : Kaplan-Meier curves for total mortality ARR=adjusted risk ratio. URR=unadjusted risk ratio. to the study drug and hydrochlorothiazide did not differ between groups. Table shows that at the end of the trial, more patients in the atenolol group than in the losartan group had stopped taking the study drug (p= 7). Systolic, diastolic, and mean blood pressures fell substantially in both groups (figure 5). Mean blood pressure at the last visit before a primary endpoint, or at the end of follow-up, was 1/79 (17/11) and 1/79 mm Hg (19/11) in the losartan and atenolol groups, reductions of 31/17 (19/11) and /17 mm Hg (1/11), respectively. 99 (5%) losartan and 97 (%) atenolol patients had diastolic blood pressure less than 9 mm Hg at the end of the study. (3%) losartan and 5 (3%) atenolol patients had systolic blood pressure below 1 mm Hg; these results were not much different in LIFE patients without diabetes. Mean arterial blood pressure at the end of the study was 11 (11) and mm Hg () in losartan and atenolol groups, respectively. Adjustment for blood pressure during follow-up had little effect on the endpoint results (data not shown). Fewer patients in the losartan group ( [ 3%]) stopped taking the study drug because of serious drug-related adverse events than in the atenolol group (9 [%], p= 5). Tables 5 and show results for selected adverse events and biochemical measurements, respectively. Albuminuria was reported less frequently (p= ) as an adverse event in the losartan than in the atenolol group (table 5). Serum glucose concentrations remained high throughout the study (figure, table ), Drug doses 5 mg only (%) 3 (5%) 5 mg plus additional drugs* 1 (1%) 9 (1%) 1 mg with or without additional 3 (51%) 5 (7%) drugs* Alone (1%) 7 (1%) With HCTZ only 9 (1%) 1 (13%) With other drugs only (%) 3 (%) With HCTZ and other drugs 17 (3%) 11 (7%) Off study drugs 159 (7%) 19 (3%) *Including hydrochlorothiazide (HCTZ). Table : Number of participants on study drug at endpoint or end of follow-up THE LANCET Vol 359 March 3, 17

5 mm Hg Systolic Mean arterial Diastolic Figure 5: Blood pressure during follow-up Data are derived from intention-to-treat analysis. and did not differ significantly between the groups (p= 7). Table 7 shows serum creatinine concentrations and the number (%) of diabetic patients with clinical albuminuria. Mean Cornell voltage-duration product fell more (p< 1) in the losartan group ( % reduction) than in the atenolol group ( %), and Sokolow-Lyon voltage decreased by 13 % and 5 %, respectively (p< 1, figure 7). p Prespecified adverse events Angio-oedema 1 ( %) 3 ( 5%) 5 Bradycardia (1%) 5 (%) < 1 Cancer 9 (%) 37 (%) 1 Cold extremities 5 (%) (%) Cough (%) 15 (3%) 1 Dizziness 95 (1%) 7 (1%) 37 Hypotension (%) 7 (1%) 51 Sexual dysfunction (%) 3 (%) 1 Sleep disturbance ( 7%) (1%) 3 Additional common* adverse events Nausea (7%) (%) 5 Anaemia 1 (7%) 31 (5%) 1 Hypokalaemia (%) (7%) Headache 9 (1%) 7 (1%) 517 Vertigo (%) (7%) 1 Dyspnoea 7 (11%) (1%) Pneumonia 3 (%) 7 (%) 13 Albuminuria 3 (7%) 79 (13%) Urinary tract infection 5 (%) 59 (1%) Abdominal pain (7%) 37 (%) Asthenia/fatigue 9 (1%) 15 (17%) 5 Chest pain 71 (%) 51 (%) 3 Influenza-like disease (%) 35 (%) 17 *An incidence >5% in one of the treatment groups and a difference between treatment groups >1%. Table 5: Adverse events Serum glucose (mmol/l) Number of patients Figure : Non-fasting serum glucose concentrations Data are mean (SD). Discussion Our results show that losartan was better than atenolol in reducing the risk of cardiovascular morbidity and mortality in patients with diabetes and hypertension. Results were especially marked in the small group (%) of patients who had not been treated for hypertension before the study. We emphasise that we decided before the start of the study to adjust results for the Framingham risk score and degree of LVH to account for any difference in key risk predictors at baseline and thus accounted for the baseline differences between treatment groups. The LIFE study, although designed as a trial in patients with hypertension and LVH, was also a correctly randomised study with respect to the prespecified subgroup of patients with diabetes. In terms of achieving reductions in cardiovascular morbidity and mortality, the benefits of treating hypertension in middle-aged and elderly patients with diabetes have been investigated mostly in subgroups of patients in prospective trials. 1 1 Although the importance of effective blood pressure control in patients with hypertension and diabetes has been shown in controlled trials, 1,15,1 the relative benefits of different antihypertensive drugs on the frequency of cardiovascular disease in this group of patients was not certain. Effective lowering of blood pressure may be even more important than glucose control in these patients. 13 Angiotensin-II antagonists have beneficial renal effects in patients with diabetes and nephropathy.,3 In our patients, albuminuria was reported significantly less often in the losartan than in the atenolol group. Our results accord with those of CAPPP 1 and to some extent HOPE, 17 but differ from those of UKPDS, 13 STOP Hypertension-, 1 NORDIL, 19 and INSIGHT. (n=5) (n=9) Baseline* Year Change Baseline* Year Change Haemoglobin (g/l) 13 7 ( 9) (1 ) (11 ) 13 (11 ) (1 1) (1 ) Sodium (mmol/l) 139 ( 7) 139 ( 9) (3 ) 139 ( ) 139 ( 9) (3 ) Potassium (mmol/l) 1 ( 3) ( 1) 5 ( 3) 15 ( 3) 15 ( ) ( 9) ALAT (IU/L) 31 9 ( ) 7 (19 ) 7 (1 3) 31 ( 3) 7 ( ) 3 (19 ) Glucose (mmol/l) 9 3 (3 5) 9 1 (3 71) 5 ( ) 9 7 (3 73) 9 5 (3 31) 5 ( 3) Total cholesterol (mmol/l) 5 79 (1 11) 5 3 (1 ) 1 (1 5) 5 3 (1 ) 5 3 (1 ) (1 11) HDL (mmol/l) 1 3 ( 35) 1 3 ( 3) 1 ( ) 1 9 ( 37) 1 5 ( 3) ( 3) Uric acid ( mol/l) 3 (7 1) 35 (99 7) (5 ) 33 (79 9) 33 ( 7) 7 (9 ) Creatinine (mmol/l) 9 5 ( ) (3 1) 11 ( ) (19 3) 1 (3 ) 1 (5 9) ALAT=alanine aminotransferase. Data are mean (SD). *Among patients who also had a value at year. Table : Biochemical variables 1 THE LANCET Vol 359 March 3,

6 (n=5) (n=9) Clinical Creatinine Clinical Creatinine albuminuria* (mmol/l) albuminuria* (mmol/l) Time Baseline 59 (11%) 9 3 ( ) 3 (%) 9 5 ( ) Year 1 9 (%) 97 ( ) (9%) 95 7 ( ) Year 35 (7%) 99 ( 5) 3 (9%) 97 1 (3 ) Year 3 7 (%) 1 7 ( 5) 3 (%) 9 5 ( 9) Year 3 (%) (3 1) 9 (11%) 1 (3 3) Year 5 7 (%) (9 ) 31 (11%) 13 3 (3 7) *Urine albumin to creatinine ratio mg/mmol ( 3 mg/g). 3 Data are mean (SD). Table 7: Albuminuria and serum creatinine concentrations As expected, the incidences of cardiovascular morbidity and mortality in patients with diabetes were higher than in the entire LIFE population. 1 However, incidences of these events in our study were lower than in patients with diabetes in STOP Hypertension- 1 (n=719, mean age 7 years), which shows the effect of age on cardiovascular mortality risk. Some limitations of the study need to be mentioned. The study population was mainly white and was from western countries. Moreover, the diabetic patients were derived from a high-risk population of hypertensive patients with LVH. The outcome should be interpreted in this context. Diabetes is an independent stimulus for LVH, an effect that is magnified by hypertension. 5, Furthermore, preliminary evidence suggests that LVH has predictive value for cardiovascular events in patients with diabetes; 7 thus, reversal of hypertensive LVH may be associated with improved cardiovascular outcome in these patients. More than half all diabetics have hypertension and LVH is very frequent in this group. Despite the fact that more of our patients in the losartan than in the atenolol group (p= 7) remained on masked treatment until the end of the study, this difference did not contribute substantially to the effects of the drugs on the composite and individual endpoints (confirmed by an ondrug analysis, data not shown). Since patients were treated without restriction after study drug discontinuation, an open-labelled angiotensin-ii antagonist or an angiotensionconverting-enzyme inhibitor may have reduced the difference between groups. Therefore, our estimation of treatment differences is conservative. Serum glucose was high at baseline and remained so in both groups (figure ). Total cholesterol fell by mmol/l in both groups (table ). Other changes in biochemical measurements were as expected with the drugs involved. Fewer than % of all patients attained a systolic blood Change from baseline (%) Cornell product p< 1 Sokolow-Lyon p< 1 Figure 7: Percentage reduction of Cornell voltage-duration product and Sokolow-Lyon voltage pressure below 1 mm Hg. The goal for systolic pressure in patients with diabetes was set to below 13 mm Hg in the 1999 WHO/International Society of Hypertension guidelines. Thus, the potential for more aggressive treatment to lower blood pressure remains for this group of patients as well as the need for better metabolic control (figure, table 7). Systolic blood pressure during the trial was not associated with any change in risk of the primary composite endpoint and therefore adjustment for this factor had little effect on the results. Thus, the greater cardiovascular protective effect of losartan than atenolol could result from more pronounced blockade of the detrimental effects of angiotensin II. was more effective than atenolol in reversing LVH (figure 7), which is likely to result from more complete protection against angiotensin II with losartan, whether generated by the circulating reninangiotensin system or other mechanisms, especially since angiotensin II is a myocardial growth factor and an independent risk factor for cardiovascular disease. 9 Hence, the general message to the practising physician is that hypertensive diabetic patients with LVH benefit more from losartan than atenolol. Contributors L H Lindholm and H Ibsen were the subcommittee on diabetes. P Aurup, S Snapinn, and J Edelman are researchers at Merck. All other authors were members of the LIFE steering committee and have commented on the manuscript. A full list of the LIFE investigators and committees is given in reference 1. Conflict of interest statement K Kristiansson is a Merck employee and was a non-voting member of the steering committee. P Aurup, S Snapinn, and J Edelman are employees of Merck. Acknowledgments We thank Sigrid Helle Berg for her dedicated work with the LIFE Study. The trial was supported by an unrestricted grant from Merck. References 1 Dahlöf B, Devereux RB, Kjeldsen SE, et al. Cardiovascular morbidity and mortality in the Intervention For Endpoint reduction in hypertension study (LIFE): a randomised trial against atenolol. Lancet ; 359: Dahlöf B, Devereux RB, de Faire U, et al. The Intervention For Endpoint reduction (LIFE) in Hypertension Study: rationale, design, and methods. Am J Hypertens 1997; 1: Dahlöf B, Devereux RB, Julius S, et al. Characteristics of 9,19 patients with left ventricular hypertrophy: the LIFE Study. Hypertension 199; 3: Kjeldsen SE, Dahlöf B, Devereux RB, et al. Lowering of blood pressure and predictors of response in patients with left ventricular hypertrophy: the LIFE Study. Am J Hypertens ; 13: Kannel WB, McGee DL. Diabetes and cardiovascular risk factors in the Framingham study. Circulation 1979; 59: 13. Stamler J, Vaccaro O, Neaton JD, Wentworth D. Diabetes, other risk factors and -year cardiovascular mortality for men screened in Multiple Risk Factor Intervention Trial. Diabetes Care 1993; 1: 3. 7 Zimmet P, Alberti KGMM, Shaw J. Global and societal implications of the diabetic epidemic. Nature 1; 1: 7 7. WHO Study Group. Diabetes mellitus (technical report series 77). Geneva: WHO, Andersson KM, Wilson PWF, Odell PM, Kannel WB. An updated coronary risk profile: a statement for health professionals. Circulation 1991; 3: Curb JD, Pressel SL, Cutler JA, et al. Effect of diuretic based antihypertensive treatment on cardiovascular disease risk in older patients with isolated systolic hypertension. JAMA 199; 7: Tatti P, Pahor M, Byington RP, et al. Outcome results of the fosinopril versus amlodipine cardiovascular events randomized trial (FACET) in patients with hypertension and NIDDM. Diabetes Care 199; 1: Estacio RO, Jeffers BW, Hiatt WR, Biggerstaff SL, Gifford N, Schrier RW. The effect of nisoldipine as compared with enalapril on THE LANCET Vol 359 March 3, 19

7 cardiovascular outcomes in patients with non-insulin-dependent diabetes and hypertension. N Engl J Med 199; 33: UK Prospective Diabetes Study Group. Efficacy of atenolol and captopril in reducing risk of macrovascular and microvascular complications in type diabetes: UKPDS 39. BMJ 199; 317: Hansson L, Lindholm LH, Niskanen L, et al. Effect of angiotensinconverting-enzyme inhibition compared with conventional therapy on cardiovascular morbidity and mortality in hypertension: the Captopril Prevention Project (CAPPP) randomised trial. Lancet 1999; 353: Tuomilehto J, Rastenyte D, Birkenhäger WH, et al. Effects of calciumchannel blockade in older patients with diabetes and systolic hypertension. N Engl J Med 1999; 3: Hansson L, Zanchetti A, Carruthers SG, et al. Effects of intensive blood-pressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. Lancet 199; 351: Heart Outcomes Prevention Evaluation (HOPE) Study Investigators. Effects of ramipril on cardiovascular and microvascular outcomes in people with diabetes mellitus: results of the HOPE study and MICRO- HOPE substudy. Lancet ; 355: Lindholm LH, Hansson L, Ekbom T, et al. Comparison of antihypertensive treatments in preventing cardiovascular events in elderly diabetic patients: results from the Swedish Trial in Old Patients with Hypertension-. J Hypertens ; 1: Hansson L, Hedner T, Lund-Johansen P, et al. Randomised trial of effects of calcium antagonists compared with diuretics and betablockers on cardiovascular morbidity and mortality in hypertension: the Nordic Diltiazem (NORDIL) study. Lancet ; 35: Brown MJ, Palmer CR, Castaigne A, et al. Morbidity and mortality in patients randomised to double-blind treatment with a long-acting calcium-channel blocker or diuretic in the International Nifedipine GITS study: Intervention as a Goal in Hypertension Treatment (INSIGHT). Lancet ; 35: Yusuf S, Sleight P, Pogue J, Bosch J, Davies R, Dagenais G. Effects of an angiotensin-converting-enzyme inhibitor, ramipril, on cardiovascular events in high-risk patients: the Heart Outcomes Prevention Evaluation Study Investigators. N Engl J Med ; 3: Brenner BM, Cooper ME, de Zeeuw D, et al. Effects of losartan on renal and cardiovascular outcomes in patients with type diabetes and nephropathy. N Engl J Med 1; 35: Lewis EJ, Hunsicker LG, Clarke WR, et al. Renoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type diabetes. N Engl J Med 1; 35: 51. Devereux RB, Roman MJ, Paranicas M, et al. Impact of diabetes on cardiac structure and function: the Strong Heart Study. Circulation ; 11: Palmieri V, Bella JN, Arnett DK, et al. Impact of type II diabetes on left ventricular geometry and function: the Hypertension Genetic Epidemiology Network (HyperGEN) Study. Circulation 1; 13: 7. Bella JN, Devereux RB, Roman MJ, et al. Separate and joint effects of hypertension and diabetes mellitus on left ventricular structure and function: the Strong Heart Study. Am J Cardiol 1; 7: 5. 7 Liu JE, Palmieri V, Roman MJ, et al. Cardiovascular disease and prognosis in adults with glucose disorders: the Strong Heart Study. J Am Coll Cardiol ; 35: 3A (abstr). Guidelines Sub-Committee World Health Organisation- International Society of Hypertension guidelines for the management of hypertension. J Hypertens 1999; 17: Weir MR, Dzau VJ. The renin-angiotensin-aldosterone system: a specific target for hypertension management. Am J Hypertension 1999; (suppl): 5S 13S. 3 American Diabetes Association. Diabetic nephropathy. Diabetes Care ; 5 (suppl): S THE LANCET Vol 359 March 3,