Cardiovascular Morbidity and Mortality in Hypertensive Patients With Lower Versus Higher Risk
A LIFE Substudy
We hypothesized that losartan was superior to atenolol in reducing cardiovascular events in a lower-risk group (LRG) versus a higher-risk group (HRG) of patients in a Losartan Intervention For Endpoint reduction (LIFE) substudy, independently of blood pressure (BP) reduction. In a post hoc analysis, we designated 4282 patients as LRG on the basis of: (1) no previous cardiovascular disease (coronary, cerebral, peripheral vascular disease); (2) no diabetes; (3) no isolated systolic hypertension; and (4) inclusion of the lowest 3 quartiles of electrocardiographically documented left ventricular hypertrophy. The HRG consisted of 4911 remaining patients who did not qualify for the LRG. In the LRG, losartan was superior to atenolol in reducing stroke: hazard ratio (HR), 0.72 (95% confidence interval [CI], 0.53 to 0.98); new-onset diabetes (HR, 0.74 [95% CI, 0.58 to 0.93]; and new-onset atrial fibrillation: HR, 0.69 (95% CI, 0.51 to 0.92), all P<0.05 but not composite end points or cardiovascular mortality (both P=NS). In the HRG, losartan was superior to atenolol in reducing composite end points: HR, 0.82 (95% CI, 0.71 to 0.94), P<0.01; cardiovascular mortality: HR, 0.77 (95% CI, 0.62 to 0.95), P<0.05; stroke: HR, 0.75 (95% CI, 0.61 to 0.92), P<0.01; new-onset diabetes: HR, 0.76 (95% CI, 0.60 to 0.96), P<0.05; and new-onset atrial fibrillation: HR, 0.71 (95% CI, 0.58 to 88), P<0.05. Test for interaction of treatment with LRG versus HRG was not significant for composite end point, stroke, or atrial fibrillation, but was for cardiovascular mortality (P=0.018). Achieved systolic BP reduction favored losartan over atenolol by −1.8 mm Hg in LRG (P=NS) and −0.7 mm Hg (P=0.001) in HRG, but no significant differences occurred in diastolic or mean BP in either group. In conclusion, losartan compared with atenolol reduces the risk of stroke, new-onset diabetes, and new-onset atrial fibrillation in the LRG and the HRG.
The Losartan Intervention For Endpoint reduction (LIFE) study demonstrated that the angiotensin II type I receptor blocker losartan was more effective than the β-blocker atenolol in preventing cardiovascular morbidity and death, predominantly stroke, for a similar reduction in blood pressure (BP) in patients with hypertension and ECG-confirmed left ventricular hypertrophy (LVH).1 Because the LIFE protocol recruited patients with ECG-LVH, all subjects had considerable cardiovascular risk before being randomized into the 2 treatment arms. Nevertheless, subsequent LIFE publications have shown similar or greater advantages of losartan over atenolol for reducing cardiovascular events in even higher-risk groups (HRGs) with diabetes,2 isolated systolic hypertension,3 or history of atrial fibrillation4 at baseline. In contrast, another LIFE study report confirmed the advantage of losartan over atenolol in primary prevention of composite end points, primarily fatal and nonfatal stroke reduction, in patients without clinically evident vascular disease.5 In general, effective antihypertensive treatment is beneficial in direct proportion to absolute cardiovascular risk.
Accordingly, the present study represented a post hoc subgroup analysis in which the double-blind, randomized LIFE study patients were designated as a lower-risk group (LRG) on the basis of no previous cardiovascular disease, diabetes, isolated systolic hypertension, or LVH in the highest quartile of Cornell voltage-duration product. The HRG consisted of the remaining patients not qualifying for LRG. The present LIFE investigation was designed to test the hypothesis that losartan, when compared with atenolol, has preventive effects beyond BP control in patients with much less cardiovascular risk than was present in the overall LIFE study population.
The LIFE study design, patient baseline characteristics, and results have been published previously.1–8 The total population consisted of 9193 hypertensive patients with ECG-defined LVH,6,7 55 to 80 years of age, who were randomly assigned after 1 to 2 weeks of placebo use to losartan or atenolol if the trough sitting BP was 160 to 200 mm Hg (systolic BP [SBP]), 95 to 115 mm Hg diastolic BP (DBP), or both. Patients were followed for ≥4 years with regular visits and upward titration of medication to reach a goal SBP <140 and DBP <90 mm Hg. Sitting BP was recorded 24 hours after last medication dose (range 22 to 26 hours).6
The present analysis divided the entire study population into 2 groups: the LRG consisted of 4282 patients with: (1) no previous cardiovascular disease (coronary heart, cerebral, or peripheral vascular disease); (2) no diabetes mellitus; (3) no isolated systolic hypertension (SBP ≥160 mm Hg and DBP <90 mm Hg); and (4) lowest 3 quartiles of LVH by Cornell voltage-duration criteria. The HRG consisted of the 4911 remaining patients with ECG-LVH in the highest quartile who did not meet the above criteria for lower risk.
ECG Coding and LVH Criteria
All ECGs were evaluated at a central core reading center for LVH criteria and Minnesota code. The product of QRS duration×Cornell voltage (with adjustment of 6 mV in women and a partition value of >2440 mV×ms) and Sokolow–Lyon voltage >38 mV were chosen to define LVH.6,9,10
The primary end point was a composite of the first occurrence of cardiovascular death, stroke, and myocardial infarction. Other prespecified outcome measures were components of primary end point (cardiovascular death, strokes, and myocardial infarctions), total mortality, hospitalization for angina pectoris, hospitalization for heart failure, coronary or peripheral revascularization, resuscitated cardiac arrest, new-onset diabetes mellitus, and new-onset atrial fibrillation. Laboratory tests were performed in 2 central laboratories. New-onset diabetes, defined according to 1985 World Health Organization criteria,11 was evaluated by a subcommittee of the steering committee.
Cardiovascular end points and BP were analyzed using the intention-to-treat approach. Only end points confirmed by the end point committee were analyzed. Analysis of the primary composite end point was confirmed with an on-treatment approach that censored end points from patients 14 days after the study drug was permanently stopped (data not shown). Only the first event in a specific category counted in any individual analysis.
Cox regression models were used to assess differences between treatment groups with respect to clinical events with adjustments made for degree of LVH and the Framingham risk score12 at baseline. Secondary unadjusted analyses were performed to validate the adjusted results. The risk reduction for losartan versus atenolol was calculated as: 100×(1−relative risk). Treatment effects were measured by hazard ratios (HRs) and their 95% confidence intervals (CIs) from Cox models. Event rates over time are presented as Kaplan–Meier curves. Adjustment for BP was based on Cox regression models with BP values throughout the trial as time-varying covariates. To confirm that the randomization was successfully achieved in the high- and low-risk subset, we performed 3 different supplementary analyses. First, we compared baseline characteristics between treatment groups using the Wilcoxon rank-sum or χ2 test as appropriate. Second, we performed a logistic regression analysis in which treatment assignment was the dependent variable to assess whether there was a systematic difference in distribution between the 2 treatment groups. Third, we used the model to generate a propensity score for the drug assignment and added that score to the Cox regression analyses. All tests were performed at 2-sided 5% significance levels using the statistical package SPSS version 12.0.1 (SPSS Inc.).
There were 9139 patients randomized to either losartan or atenolol; 4282 were assigned to LRG, of whom 2098 received losartan and 2184 atenolol, and 4911 were assigned to HRG, of whom 2507 received losartan and 2404 atenolol (Figure 1). No participants were lost to follow-up for survival status before the occurrence of an end point.
Patients assigned to losartan or atenolol in the LRG and HRG, respectively, were closely matched for demographic characteristics, severity of hypertension, prevalences of coexisting medical conditions, Framingham risk scores, and electrocardiographic LVH (Table 1). As expected, the LRG was younger and had much less abnormal baseline values for serum glucose, blood lipids, serum creatinine, urine albumin/creatinine ratio, and body mass index. In addition, the LRG had a lower SBP (1.1 mm Hg for the losartan and 2.2 mm Hg for the atenolol arm) and a higher DBP (5.1 mm Hg for the losartan and 5.4 mm Hg for the atenolol arm) compared with the HRG. There were no significant differences in cardiovascular risk factors between the 2 drug arms in the LRG. However, patients in the HRG treated with atenolol had slightly, but significantly, higher prevalence of smoking and Framingham risk scores than those treated with losartan.
Follow-Up and BP Control
Mean follow-up from randomization to death, loss to follow-up, or end of study was 4.8 years. Losartan- and atenolol-treated patients received primary study therapy in the LRG for 85% and 82% of follow-up time and received hydrochlorothiazide for 74% and 72% of follow-up time, respectively. Similarly, losartan- and atenolol-treated patients received study therapy in the HRG for 80% and 74% of follow-up time and hydrochlorothiazide 70% and 67% of follow-up time, respectively. Table 2 shows the distribution of blinded study treatments for patients at the end of follow-up or at occurrence of the first primary end point, if earlier. Mean losartan and atenolol values at the end point or end of randomly assigned therapy were 79 and 77 mg for LRG and 80 and 76 mg for HRG, respectively.
SBP, DBP, pulse pressure, and mean BPs were reduced substantially and almost equally in both treatment groups of the LRG and HRG, respectively (Table 3). At end of follow-up or last visit before a primary end point, there was very little difference in sitting BP reduction between the 2 treatment arms of the LRG and HRG. In the LRG, SBP was reduced by 29.1±18.3 mm Hg in the losartan arm and 27.3±18.4 mm Hg in the atenolol arm (difference 1.8 mm Hg; P=0.001). In the HRG, SBP was reduced by 28.0±19.4 mm Hg in the losartan arm and 27.3±20.3 mm Hg in the atenolol arm (difference 0.7 mm Hg; P=NS). There were no significant differences in DBP or mean arterial pressure reduction between losartan and atenolol arms of either the LRG or HRG.
BP ≤140/90 mm Hg was achieved in 68% and 66% of the losartan and atenolol groups, respectively, in the LRG, and 64% and 60%, respectively, in the HRG. Heart rate decreased more with atenolol than losartan treatment in the LRG (−7.9±12 versus 2.2±12 bpm) and the HRG (−7.7±13 versus −2.2±12 bpm; both P<0.001), respectively. Furthermore, losartan was more effective than atenolol in reducing ECG LVH in the LRG (Cornell voltage-duration −210±652 versus −78±725 mV×ms; Sokolow–Lyon voltage −4.8±6.6 versus −0.1±7.1) and the HRG (Cornell voltage-duration −314±943 versus −69±999 mV×ms; Sokolow–Lyon voltage −4.5±7.6 versus −2.8±7.7; all P<0.001).
Occurrence of End Points
The prespecified primary composite and individual end point event rates were approximately twice as high in the HRG as in the LRG (Table 4). As shown in Table 5 and in the Kaplan–Meier curves for HRG (Figure 2), there was a highly significant reduction with losartan versus atenolol for (Figure 2A) composite end points, (HR, 0.80; 95% CI, 0.69 to 0.92; P<0.01) and for (Figure 2B) cardiovascular mortality, (HR, 0.74; 95% CI, 0.60 to 0.92; P<0.01). However, in the LRG, there were no significant reductions with losartan versus atenolol for (Figure 2A) composite end points (HR, 0.95; 95% CI, 0.76 to 1.18) or for (Figure 2B) cardiovascular mortality (HR, 1.26; 95% CI, 0.86 to 1.84).
As expected, more strokes occurred in the HRG (n=368) than in the LRG (n=173; HR, 1.6 [95% CI, 1.3 to 2.5]; P<0.001; Table 4). Losartan was superior to atenolol (Table 6 and Figure 2C) for stroke reduction in the LRG (HR, 0.72; 95% CI, 0.54 to 0.98; P<0.05) and in the HRG (HR, 0.75; 95% CI, 0.61 to 0.92; P<0.01). This benefit favoring losartan began within the first few months of beginning therapy and continued to increase throughout the duration of the study in the LRG and HRG. The losartan-treated patients in the HRG had significantly lower incidence of ischemic, hemorrhagic, and embolic strokes (χ2=12.9; P=0.012), whereas there was a nonsignificant (χ2=6.2; P=0.187) lower incidence of all stroke types in the LRG. This treatment benefit was significant for atherothrombotic strokes (HR, 0.70; P=0.005) in the HRG with nonsignificant parallel trends for embolic and hemorrhagic strokes in the HRG (HRs, 0.88 and 0.75; P=0.74 and P=0.27) and parallel nonsignificant trends for the same 3 groups (HR, 0.55 to 0.76; P=0.11 to 0.79) in the LRG. Tests for interaction between treatment and stroke type with regard to LRG and HRG, using a general linear model, showed a nonsignificant interaction (P=0.984), whereas stroke types (P<0.001) and treatment (P=0.023) remained significant in the model.
The patients in the LRG had similar benefit to the HRG with regard to stroke severity, as judged by the presence or absence of ataxia, central cranial nerve dysfunction, depression of consciousness, disturbance of vision, memory defect, movement disorder, paresis or paralysis, sensory impairment, or speech impairment (data not shown); tests for interaction between treatment and all stroke manifestations were nonsignificant (P=0.152 to 0.901).
Reduction in new-onset diabetes favored losartan over atenolol in the HRG (HR, 0.76; 95% CI, 0.60 to 0.96; Table 7) and the LRG (HR, 0.74; 95% CI, 0.58 to 0.93; Table 6). Losartan reduced new-onset atrial fibrillation similarly in the LRG and the HRG (Tables 5 and 6⇑). Furthermore, the incidence of stroke in the LRG was strongly dependent on either a history of atrial fibrillation (HR, 3.8 [95% CI, 2.04 to 6.92]; P<0.001) or incident atrial fibrillation (HR, 4.6 [95% CI, 3.10 to 6.89]; P<0.001). However, the rate of stroke in LRG members with neither prevalent nor incident atrial fibrillation also tended to be lower in losartan-treated patients (HR, 0.75 [95% CI, 0.55 to 1.02]; P=0.064). Adjustment for Framingham risk score, differences in Cornell voltage-duration, and Sokolow–Lyon voltage criteria covariates had minimal effect on the advantage of losartan over atenolol in reduction in stroke, new-onset diabetes, or new-onset atrial fibrillation in either the LRG or HRG. Time-varying adjustment for differences in SBP and DBP had minimal effect on the estimates and did not change any conclusions (data not shown). Tests for interactions between the LRG and HRG status and the effects of losartan compared with atenolol (Table 7) were not significant for the composite end points (P=0.56), stroke (P=0.93), new-onset diabetes (P=0.98), or new-onset atrial fibrillation (P=0.72), but was significant for cardiovascular mortality (P=0.018).
The main finding of this study was the superiority of losartan-based compared with atenolol-based therapy for reduction of the end point of fatal and nonfatal stroke by 27% in the LRG and by 25% in the HRG. Treatment benefit was significant for atherothrombotic strokes in the HRG; in addition, nonsignificant trends were noted for embolic and hemorrhagic strokes in the HRG and for parallel nonsignificant trends for the same 3 stroke groups in the LRG. The patients in the LRG versus HRG had similar reductions in stroke severity, as judged by a decrease in a variety of stroke manifestations.
In addition, losartan was superior to atenolol in reducing the secondary end point of new-onset diabetes by 26% and the tertiary end point of new-onset of atrial fibrillation by 31% in the LRG; these reductions were comparable to those seen in the HRG. Losartan was superior to atenolol in reducing composite end points by 18% and cardiovascular mortality by 23% in the HRG, but there was no losartan advantage over atenolol in reducing cardiovascular mortality in the LRG. Finally, the nonsignificant interaction term of LRG versus HRG for composite end points makes it likely that losartan has similar efficacy for reducing composite end points in the LRG. In contrast, the significant interaction term of LRG versus HRG for cardiovascular mortality suggests that only the HRG group showed superior efficacy of losartan over atenolol in reducing cardiovascular mortality.
Although there was a difference in SBP reduction between losartan and atenolol, this difference was smaller in the HRG (0.7 mm Hg; P=NS) than in LRG (1.8 mm Hg; P=0.001). In-treatment analysis of SBP, as a time-varying covariate, did not show a meaningful contribution of small differences in pressure reduction between the 2 arms to the superiority of losartan over atenolol in decreasing events in either the LRG or HRG.
Patient selection in the present study resulted in event rates in the LRG that were about half those in the HRG. Therefore, it is possible that diminished statistical power may explain, at least in part, the lack of superiority of losartan over atenolol in reducing either composite end points or cardiovascular mortality in the LRG. Furthermore, the comparable reduction in HRs for fatal and nonfatal stroke, new-onset diabetes, and new-onset atrial fibrillation suggests equal benefit of losartan over atenolol in the LRG and HRG. However, keep in mind that the LRG still qualifies as a high-risk group (>20% chance of developing a cardiovascular event over 10 years) on the basis of a Framingham risk score of 20, a mean Cornell voltage duration of 2400 as an indication of significant LVH, and a cardiovascular mortality rate of 6.1%.13
The primary end point results of this investigation, as well as the main LIFE study,1 confirm the value of blockade of the renin-angiotensin system in protecting against composite end points in general14 and stroke reduction in particular in a manner that appears independent of BP reduction. These results contrast with recent meta-analyses that showed no inherent cardiovascular protective effect of angiotensin II receptor blockers beyond their lowering of BP15,16; however, these studies did not compare angiotensin II receptor antagonists with β-blockers. Furthermore, in the entire LIFE study population, there was a nonsignificant reduction in cardiovascular mortality of 11.4% (HR, 0.87; P=0.136) favoring losartan over atenolol. The present study showed a 23% reduction in cardiovascular mortality (HR, 0.74; P=0.007) favoring losartan over atenolol in the HRG, confirming the thesis that a greater absolute risk population will show greater benefit from effective antihypertensive drug therapy.17
A major hypothesis of the LIFE study, that losartan is more effective than atenolol in regressing LVH, proven in the original study,1 was confirmed in the LRG and HRG. Moreover, in the present study, we were able to show comparable reduction of LVH in the LRG and HRG, suggesting that regression of LVH may be an important mechanism that protects against stroke events and the new onset of atrial fibrillation. In addition, patients with LVH also exhibit increased left atrial size. Therefore, it is possible that the parallel effects of losartan on regression of atrial and ventricular hypertrophy may have contributed to the reduction in new-onset atrial fibrillation equally in LRG and HRG.
This was a post hoc subgroup analysis of the original LIFE study, and therefore, LRG and HRG were not preselected patient populations. Furthermore, in the LIFE study, there was no stratification as part of the randomization process, either in the main study or in the LRG and HRG of the present investigation. However, the losartan and atenolol arms of the LRG and HRG showed comparable baseline characteristics. Despite the importance of elevated BP in causing cardiovascular complications, additional adjustment of primary outcomes for observed differences in SBP and DBP had little effect on HRs that showed benefit of losartan over atenolol therapy. Twenty-four–hour ambulatory BP, central arterial pressure, and waveforms were not measured in this study; therefore, we cannot rule out that central BP measurements may be more adversely affected by β-blockade compared with losartan.18 Finally, there is the possibility that β-blockers may have greater protective powers in younger patients with increased sympathetic activity or in an older-aged population preselected on the basis of ischemic heart disease rather than on the basis of hypertension-related LVH, as was present in the LIFE study.19
In this post hoc subgroup analysis of the LIFE study, in which most subjects received diuretic therapy, losartan was superior to atenolol in regression of LVH and in reduction of fatal and nonfatal stroke, new-onset diabetes, and new-onset atrial fibrillation in the LRG; this superiority was largely independent of BP reduction. The prevention of stroke and new-onset atrial fibrillation with losartan therapy in the absence of previous cardiovascular events, diabetes, isolated systolic hypertension, or the most severe LVH suggests an even greater indication for its use than was shown in the original LIFE study.
- Received January 28, 2005.
- Revision received February 21, 2005.
- Accepted July 19, 2005.
Dahlöf B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U, Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Wedel H; LIFE Study Group. Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomized trial against atenolol. Lancet. 2002; 359: 995–1003.
Lindholm LH, Ibsen H, Dahlöf B, Devereux RB, Beevers G, de Faire U, Fyhrquist F, Julius S, Kjeldsen SE, Kristiansson K, Lederballe-Pedersen O, Nieminen MS, Omvik P, Oparil S, Wedel H, Aurup P, Edelman J, Snapinn S; LIFE Study Group. Cardiovascular morbidity and mortality in patients with diabetes in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomized trial against atenolol. Lancet. 2002; 359: 1004–1010.
Kjeldsen SE, Dahlöf B, Devereux RB, Julius S, Aurup P, Edelman J, Beevers G, de Faire U, Fyhrquist F, Ibsen H, Kristianson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Snapinn S, Wedel H; LIFE (Losartan Intervention for Endpoint Reduction) Study Group. Effects of losartan on cardiovascular morbidity and mortality in patients with isolated systolic hypertension and left ventricular hypertrophy: a Losartan Intervention for Endpoint Reduction (LIFE) substudy. J Am Med Assoc. 2002; 288: 1491–1498.
Wachtell K, Hornestam B, Lehto M, Slotwiner DJ, Gerdts E, Olsen MH, Aurup P, Dahlof B, Ibsen H, Julius S, Kjeldsen SE, Lindholm LH, Nieminen MS, Rokkedal J, Devereux RB. Cardiovascular morbidity and mortality in hypertensive patients with a history of atrial fibrillation: the LIFE study. J Am Coll Cardiol. 2005; 45: 705–711.
Devereux RB, Dahlöf B, Kjeldsen SE, Julius S, Aurup P, Beevers G, Edelman JM, de Faire U, Fyhrquist F, Helle Berg S, Ibsen H, Kristianson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Snapinn S, Wedel H; LIFE Study Group. Effects of losartan or atenolol in hypertensive patients without clinically evident vascular disease: a substudy of the LIFE randomized trial. Ann Intern Med. 2003; 139: 169–177.
Dahlöf B, Devereux RB, de Faire U, Fyhrquist F, Hedner T, Ibsen H, Julius S, Kjeldsen S, Kristianson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Wedel H. The Losartan Intervention For Endpoint reduction (LIFE) in Hypertension Study: rationale, design, and methods. The LIFE Study Group. Am J Hypertens. 1997; 10: 705–713.
Dahlöf B, Devereux RB, Julius S, Kjeldsen SE, Beevers G, de Faire U, Fyhrquist F, Hedner T, Ibsen H, Kristianson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Wedel H. Characteristics of 9194 patients with left ventricular hypertrophy: the LIFE study. Hypertension. 1998; 32: 989–997.
WHO Technical Report Series 727. 1995; Report.
Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile. A statement for health professionals. Circulation. 1991; 83: 356–362.