Reduction in Cardiovascular Events During Pravastatin Therapy
Pooled Analysis of Clinical Events of the Pravastatin Atherosclerosis Intervention Program
Background It has been documented that the HMG coenzyme A reductase inhibitors, or statins, can decrease cardiovascular events and mortality in patients with clinical coronary disease and moderately to severely elevated lipid levels. Additional data are required to demonstrate a reduction of vascular events in coronary patients with less than severely elevated lipid levels and in subgroups of this population.
Methods and Results Clinical data from four atherosclerosis regression trials that evaluated pravastatin were pooled for a predetermined analysis of the effect of that agent on the risk of coronary events. All trials were double-masked, placebo-controlled designs that used pravastatin as monotherapy for 2 to 3 years. The 1891 participants in the trials had evidence of atherosclerosis and mildly to moderately elevated lipid levels. For fatal or nonfatal myocardial infarction, there was a 62% reduction in events attributable to pravastatin (P=.001). This effect was evident in younger and older patients, men and women, and patients with and without histories of hypertension and prior infarction. There was a 46% reduction in all-cause mortality (P=.17), which, although not statistically significant, is consistent with the results of other statin trials. There also was a 62% reduction in the risk of fatal or nonfatal stroke (P=.054).
Conclusions These pooled results provide strong evidence that pravastatin reduces the risk of cardiovascular events in patients with atherosclerotic disease and mildly to moderately elevated lipid levels. The benefit for reducing myocardial infarction is evident in older and younger patients, men and women, and patients with and without histories of hypertension and prior infarction.
Elevated blood cholesterol has been conclusively established in epidemiological studies as a major independent risk factor for CHD and other atherosclerotic vascular diseases. However, the effects of lipid-lowering interventions for reducing the risk of CHD morbidity and mortality are less well established, and the role of lipid-lowering therapy in the primary and secondary prevention of CHD is still being debated.
In the last 30 years, more than 50 clinical trials have evaluated nearly 60 lipid-lowering regimens over 85 000 person-years.1 The average reduction in serum total cholesterol in these trials generally ranged from 0% to 20%. Despite this large body of data, uncertainty remains regarding the net effects of lipid-lowering therapy on coronary morbidity and mortality, noncardiovascular mortality, and overall survival. An important observation from meta-analyses of trial findings is that the benefit of lipid-lowering intervention is linked to the magnitude of the reduction in total cholesterol level1 2 and duration of therapy. Clear benefit in these older trials was generally not observed until after 2 years of therapy; full benefit apparently was not achieved until after 5 years.3
The introduction of a new class of agents, the HMG coenzyme A reductase inhibitors, or statins, raised expectations that these issues would be resolved. As monotherapy, the statins are more efficacious in their lipid-lowering effect than the older agents, which suggests that the statins would be more likely to significantly reduce mortality and clinical events. Supporting this supposition, the large, long-term 4S recently reported that all-cause mortality and coronary events were substantially reduced by the use of a statin in patients with clinical coronary disease and moderately to severely elevated lipid levels.4
This report presents the results of a predetermined pooled analysis of the cardiovascular event data from four recently completed atherosclerosis regression trials that used a specific statin, pravastatin. Most of the participants in these trials had evidence of atherosclerosis and less than severely elevated lipid levels. The purpose of the pooling was to gain additional information on the effects of lipid reduction on clinical events in a population such as this; subgroups of this population, such as older and younger people and men and women; and people with and without histories of MI. These trials, which were conducted in Europe and the United States, had design features facilitating a pooled analysis and a total sample size of 1891 patients followed for an average of 2.3 years. They are PLAC-I,5 PLAC-II,6 7 , REGRESS,8 and KAPS.9
The designs of each of the four trials were described previously.5 6 7 8 9 All were double-masked, placebo-controlled designs with pravastatin as monotherapy for 2 to 3 years (Table 1⇓). Collectively, these trials had 4343 patient-years of follow-up. Briefly, PLAC-I was designed to evaluate the effect of pravastatin on the angiographic 3-year progression of coronary artery disease in 408 coronary patients with at least one stenosis ≥50%.5 PLAC-II evaluated the ability of pravastatin to retard the ultrasonographic 3-year progression of extracranial carotid atherosclerosis in 151 patients with coronary artery disease, documented as either angiographic evidence of at least one 50% stenosis or a clinical MI.6 7 Both PLAC-I and PLAC-II recruited men and women. REGRESS was a 2-year angiographic study of pravastatin in 885 men with at least one coronary artery stenosis ≥50%.8 KAPS was a 3-year ultrasonographic trial that evaluated the effects of pravastatin on the progression of carotid and femoral atherosclerosis in 447 hypercholesterolemic men selected from the general population.9
To qualify for PLAC-I or PLAC-II, LDL cholesterol levels were required to be ≥3.36 but <4.91 mmol/L (≥130 but <190 mg/dL) after diet stabilization. For REGRESS, total cholesterol levels had to be ≥4.0 but <8.0 mmol/L (≥155 but <310 mg/dL), and triglyceride levels had to be ≤4.0 mmol/L (≤355 mg/dL). For KAPS, LDL cholesterol levels had to be >4.0 mmol/L (>155 mg/dL); total cholesterol had to be <7.5 mmol/L (<290 mg/dL). Patients with hypertriglyceridemia were ineligible in all four trials.
Pravastatin (40 mg) or placebo was given once at bedtime for 2 to 3 years in PLAC-I, REGRESS, and KAPS. The pravastatin dose in PLAC-II was 20 to 40 mg/d, titrated to reduce LDL cholesterol to at least 2.84 mmol/L (110 mg/dL). The average pravastatin dose taken during PLAC-II was 30 mg/d.
The independent Policy Advisory Board common to PLAC-I, PLAC-II, and KAPS recommended early during the conduct of these concurrent trials that the clinical event data from the individual trials be pooled and analyzed when the last trial was completed. Similarities in study designs (treatment with pravastatin monotherapy for 36 months) and participant populations (subjects with documented atherosclerotic disease and mildly to moderately elevated LDL cholesterol levels) among the trials provided a rationale for pooling the events data. REGRESS was invited to participate because it was also a placebo-controlled trial of pravastatin used as monotherapy.
Events from the four regression trials were pooled to form five outcome measures established before the pooled data were examined: nonfatal or fatal MI; nonfatal MI or CHD death; all-cause mortality; nonfatal MI or all-cause mortality; and the combination of nonfatal MI, all-cause mortality, nonfatal stroke, PTCA, and CABG. The combination of nonfatal MI and all-cause mortality was selected as an outcome to encompass the possibility of competing mortality from noncardiovascular causes.
Time-to-event analyses, stratified by and adjusted for clinical trial, were performed using all first events when there was more than one event for the same subject. The Mantel-Haenszel log-rank statistic was computed to test for treatment group differences using the sas lifetest procedures with the Kaplan-Meier method.12 Event rates per 100 treated participants were also estimated over the follow-up period by use of the Kaplan-Meier method. For subgroup analyses, treatment group differences were stratified by age, sex, smoking status, history of hypertension, prior MI, and lipid levels. Tests of treatment interaction were then conducted. All statistical tests of significance were two-sided. Ninety-five percent CIs for risk reduction were estimated with Fieller’s theorem for the ratio.13
To determine whether any effect of pravastatin on nonfatal or fatal MI may be due to or independent of any effect on LDL cholesterol reduction, the effects of baseline LDL cholesterol, follow-up LDL cholesterol, LDL cholesterol change (over 3 years of follow-up), and treatment group assignment on this outcome were examined bivariately (with adjustment for clinical trial) and multivariately. Results for follow-up LDL cholesterol and LDL cholesterol change were almost identical, so only LDL cholesterol change is described.
Table 2⇓ gives baseline descriptions of the four trials. Combining the four trials created a pooled database of 1891 participants, with 955 people in the pravastatin groups and 936 in the placebo groups. There were no treatment group differences in any of the pooled characteristics, either within a trial or in the pooled database.
The mean ages ranged from 56 years (REGRESS) to 62 years (PLAC-II), with an overall average of 57 years. The proportion of women in the pooled database was 6%. The proportion of participants who were smokers at the time of their baseline examinations ranged from 13% in PLAC-II to 28% in REGRESS, with an overall average of 24%. The proportion of participants with prior MI ranged from 8% (KAPS) to 64% (PLAC-II).
Effect of Pravastatin on LDL Cholesterol
KAPS had the greatest mean LDL cholesterol value at baseline, 4.9 mmol/L (189 mg/dL). The overall mean LDL cholesterol value at baseline for the four trials combined was 4.4 mmol/L (171 mg/dL). For the actively treated groups, the mean LDL cholesterol reduction during the treatment period was 28.2% (4.43 mmol/L [171 mg/dL] at baseline to a mean of 3.18 mmol/L [123 mg/dL] during follow-up, P<.001). There was essentially no change for the placebo groups (4.42 mmol/L [171 mg/dL] at baseline to a mean of 4.45 mmol/L [172 mg/dL] during follow-up). There was a 27.5% LDL cholesterol reduction among the pravastatin participants in PLAC-I and PLAC-II, a 28.2% reduction in REGRESS, and a 28.7% reduction in KAPS.
Effect of Pravastatin on Clinical Outcome Measures
Table 3⇓ and Figs 1⇓ and 2⇓ show the comparisons of treatment effects on the combined clinical events. For each outcome measure, the pravastatin groups had fewer events, both overall and within each trial. Tests for heterogeneity did not suggest that any trial had a different treatment group effect.
For the combined outcome of nonfatal or fatal MI, there were 46 events among the placebo participants (6.4% Kaplan-Meier estimate over the follow-up period) compared with 21 among the pravastatin participants (2.4%). This is a 62% reduction in event rates attributable to pravastatin (38% to 80%, 95% CI; P=.001). It is shown in Fig 1⇑ that the event curves for this outcome measure began to separate fully after a lag time of about 12 months of therapy. During the first year of treatment, there were 15 events in the pravastatin groups and 18 events in the placebo groups; during the second year, 5 events in the pravastatin groups and 17 events in the placebo groups; and during the third year, 1 event in the pravastatin groups and 11 events in the placebo groups.
For the combined end point of nonfatal MI or CHD death, the reduction in events attributable to pravastatin is similar to the reduction observed for fatal or nonfatal MI alone, 51% risk reduction (24% to 71%, P=.006).
There were 38 deaths among the 1891 participants over the average of 2.3 years of follow-up, 23 (3.3%) in the placebo groups and 15 (1.8%) in the pravastatin groups. This is a 46% reduction in risk ((9% to 76%, P=.168). Fig 2⇑ shows the cumulative event curves for all-cause mortality. Although the treatment group difference in mortality did not meet statistical significance, the pravastatin group rates are consistently lower than the placebo group rates over the follow-up period.
Table 4⇓ lists the causes of death. Overall, the 46% reduction in all-cause mortality that was attributable to pravastatin is a function of the reduction in vascular deaths: there were 16 vascular deaths among the placebo participants compared with 10 among the pravastatin participants. There was essentially no difference in the number of cancer deaths (4 in the pravastatin and 5 in the placebo groups) or in the number of other deaths. There was 1 violent death, a homicide, in the placebo groups.
When nonfatal MI and all-cause mortality are combined, there were 61 events in the placebo group compared with 34 in the pravastatin group (Table 3⇑). This is a 53% reduction in events (29% to 71%, P=.003).
When strokes and cardiovascular procedures (PTCA and CABG) are combined with MI and deaths, the overall 3-year event rate in the placebo groups climbs to 19.5%. This contrasts with the 13.7% 3-year event rate in the pravastatin groups, which is a 30% difference in risk (12% to 45%, P=.002). When strokes are examined separately, which was not a prior hypothesis, it is noted that there were 13 fatal or nonfatal strokes among the 936 placebo participants compared with 5 among the 955 pravastatin participants. This 61.9% reduction in risk (−9.1% to 97.4%, P=.054) may be attributable to pravastatin. The cumulative event curves for this outcome also begin to separate fully after about 12 months of therapy (Fig 3⇓).
Concomitant Effects of Pravastatin on LDL Cholesterol and MI
In two bivariate analyses (with adjustment for clinical trial) of the entire pooled population, baseline LDL cholesterol was not predictive of subsequent MI (P=.79), although LDL cholesterol reduction over the follow-up period was predictive of a decreased rate of MI (P=.05). When treatment group assignment, baseline LDL cholesterol, and LDL cholesterol change were incorporated into a single multivariate model, the probability value for the treatment effect increased from P=.001 noted in Table 3⇑ to P=.01. This suggests that whereas pravastatin exerts at least some of its effect on MI reduction through its effect on LDL cholesterol reduction, the clinical effect also may result in part from some ancillary property of the drug.
Subgroup Analyses of Treatment Effect
The five prespecified outcomes were stratified by nine baseline characteristics to create 95 subgroup analyses. None of the resultant 45 tests of interaction between subgroup and treatment group assignment produced a value of P<.05. In 94 of the 95 subgroups, pravastatin was shown to have a favorable effect.
Table 5⇓ presents the results of nonfatal or fatal MI stratified by age, sex, history of hypertension, and prior MI. As with the overall treatment effects in Table 3⇑, these subgroup results for MI are generally paralleled by the results for nonfatal MI or CHD death, nonfatal MI or all-cause mortality, and the combination of any vascular event, any death, and any cardiovascular procedure. Although the number of events in some subgroups is small, there was a statistically significant reduction in the risk of MI attributable to pravastatin in almost every subgroup (P<.05). There was an 85% reduction in events for patients ≥65 years of age and a 55% reduction for patients <65 years; a 60% reduction for men and a 100% reduction for women (based on only four events among the women); a 69% reduction for patients with a history of hypertension and a 57% reduction among patients without a history; and an identical 63% reduction for patients with and without a history of a clinical MI.
When treatment effects on nonfatal or fatal MI were initially stratified by baseline LDL cholesterol level, it was observed that placebo participants below the median LDL cholesterol level (4.42 mmol/L [171 mg/dL]) had higher event rates than placebo participants above the median LDL cholesterol level. Further examination of the data showed that participants below the LDL cholesterol median were significantly older and had a higher prevalence of prior MI. For this reason, the LDL cholesterol strata were subdivided by age group and prior MI. Table 5⇑ gives the LDL cholesterol–prior MI strata. For patients with a prior MI, the relative risk reductions attributable to pravastatin are comparable for those below and above 4.42 mmol/L LDL cholesterol; for patients without a prior MI, the relative risk reduction appears to be greater for those below the median. This apparent greater benefit among patients with lower LDL cholesterol levels appears again in the bottom of Table 5⇑ when the risk reductions are adjusted for age and prior MI. However, risk reductions were observed for both levels of LDL cholesterol.
Although there was a 46% overall risk reduction for the all-cause mortality outcome, there were no statistically significant treatment group differences in any subgroup. However, the trend for all subgroups was to show reduced mortality.
In this predetermined pooled analysis of almost 2000 patients with evidence of atherosclerosis and less than severely elevated lipid levels, pravastatin was shown to significantly reduce the incidence of MI, by nearly two thirds. Participants assigned to the pravastatin groups in the four constituent trials also experienced lower rates of fatal or nonfatal coronary events and all-cause mortality, although the differences were not always statistically significant.
Analyses suggested that these results are at least partly related to the average 28% reduction in LDL observed among the pravastatin participants. We observed that after adjustment for LDL cholesterol reduction, the treatment group effect was still statistically significant. This finding suggests that this agent may have an effect beyond simple lipid lowering. Mechanisms such as plaque stabilization,14 restoration of endothelial function,15 and a decrease in platelet activation16 are possible explanations for this additional benefit.
It is noteworthy that for each outcome measure, the pravastatin groups consistently had fewer events. This was true overall, within each trial, and in all subgroups. For example, for the combined outcome measure of nonfatal MI or CHD death, although there was an overall 51% reduction attributable to pravastatin for the pooled data, there was a 53% reduction in events in PLAC-I, a 60% reduction in PLAC-II, a 38% reduction in REGRESS, and a 39% reduction in KAPS. The consistency of these results across the four trials for all outcome measures, including all-cause mortality, supports the hypothesis that these events can be reduced by pravastatin.
It is also noteworthy that although previous trials of the older, less efficacious lipid-lowering agents generally produced a reduction in events only after at least 2 years of therapy,3 this pooled analysis suggests that the effect of pravastatin may occur after only 1 year (Fig 1⇑). This observation of an earlier effect for a statin has also been made by the 4S investigators.4
A growing body of evidence supports the hypothesis that these findings are a result of a class effect of the statins. For example, many regression trials of statins reported fewer events in the treated groups than in the control groups, although few of these trials reached statistical significance for clinical outcomes.17 The observed effect of pravastatin on mortality in this pooled analysis is consistent with the statistically significant reductions in all-cause mortality reported by 4S4 and ACAPS,18 both of which also studied the effects of statins. With respect to reductions in major cardiovascular events, the 4S trial, designed as a 6-year trial of 4444 coronary patients with moderately to severely elevated lipid levels (10% higher LDL cholesterol levels than in this pooled pravastatin database), reported a 34% reduction of events associated with simvastatin use.
The reductions in clinical events observed in the age and sex subgroups of this pooled database are similar to the reductions noted in the 4S trial.4 These analyses, however, also suggest clinical benefit in people with and without histories of prior MI and hypertension and in people with LDL cholesterol levels as low as 3.36 mmol/L (130 mg/dL).
These large reductions in coronary events in patients treated with statins are substantially larger than the reductions achieved in trials of the older lipid-lowering agents or diets with smaller average reductions of LDL cholesterol. For example, Rossouw and coworkers19 reported a reduction of only 23% in fatal and nonfatal reinfarctions in their summary of the results of secondary-prevention lipid trials. The most likely explanation for the observed greater benefit in this pooled analysis of pravastatin and other statin trials is the more pronounced cholesterol lowering associated with the statins. Given the 28% to 29% reduction in LDL cholesterol, the observed reduction in events is consistent with reductions predicted by Law et al3 and Holme.20
In addition to the effect on coronary events, a 62% reduction in stroke attributable to pravastatin also was noted in our analyses (P=.054). Although not a prespecified outcome, this observation of an effect on cerebrovascular events is consistent with other post hoc analyses of statins. Again, the 4S trial reported a 30% reduction in fatal and nonfatal cerebrovascular events attributable to simvastatin (P=.024),4 and ACAPS reported that 0 of the 5 fatal or nonfatal strokes that occurred during follow-up were among lovastatin participants.21
The reductions in stroke in the three statin analyses are larger than the previously reported effects of the older lipid-lowering therapies in which no reduction in stroke is evident, as noted in a recent meta-analysis involving more than 36 000 participants from 11 clinical trials.22 In contrast, however, the observed reductions in stroke in the three statin analyses are generally comparable with the reported effects of antiplatelet and anti-hypertensive therapies.23 24
For the clinician, the reduction in clinical events associated with the use of a statin is encouraging. Our data, coupled with the reports of other trials, suggest that the clinical benefits are evident over a wide spectrum of patient subgroups. For example, we observed beneficial effects in people older than and younger than 65 years of age, men and women, patients with and without histories of hypertension or prior MI, and patients with LDL cholesterol levels as low as 3.36 mmol/L (130 mg/dL).
The ultimate role of statins in the primary and secondary prevention of coronary and noncardiac vascular diseases is becoming more defined. Within the next 2 to 3 years, the results of ongoing large morbidity and mortality trials, conducted in a variety of population settings, will be announced. Until that time, results such as this pooling analysis of pravastatin contribute to the growing body of strong evidence for a major clinical benefit of lipid lowering with statins.
Selected Abbreviations and Acronyms
|ACAPS||=||Asymptomatic Carotid Artery Plaque Study|
|CABG||=||coronary artery bypass graft|
|CHD||=||coronary heart disease|
|KAPS||=||Kuopio Atherosclerosis Prevention Study|
|PLAC-I||=||Pravastatin Limitation of Atherosclerosis in the Coronary Arteries Trial|
|PLAC-II||=||Pravastatin, Lipids, and Atherosclerosis in the Carotid Arteries Trial|
|PTCA||=||percutaneous transluminal coronary angioplasty|
|REGRESS||=||Regression Growth Evaluation Statin Study|
|4S||=||Scandinavian Simvastatin Survival Study|
This study was sponsored by Bristol-Myers Squibb Co.
- Received March 2, 1995.
- Revision received May 15, 1995.
- Accepted May 30, 1995.
- Copyright © 1995 by American Heart Association
Gordon DJ. Cholesterol lowering and total mortality. In: Rifkind BM, ed. Contemporary Issues in Cholesterol Lowering: Clinical and Population Aspects. New York, NY: Marcel Dekker Inc; 1994:33-48.
Holme I. An analysis of randomized trials evaluating the effect of cholesterol reduction on total mortality and coronary heart disease incidence. Circulation. 1990;82:1916-1924.
Law MR, Wald NJ, Thompson SG. By how much and how quickly does reduction in serum cholesterol concentration lower risk of ischaemic heart disease? BMJ. 1994;308:367-373.
Pitt B, Ellis SG, Mancini GBJ, Rosman HS, McGovern ME, for the PLAC I Investigators. Design and recruitment in the United States of a multicenter quantitative angiographic trial of pravastatin to limit atherosclerosis in the coronary arteries (PLAC I). Am J Cardiol. 1993;72:31-35.
Jukema JW, Bruschke AVG, van Boven AJ, Reiber JHC, Bal ET, Zwinderman AH, Jansen H, Boerma GJM, van Rappard FM, Lie KI, for the REGRESS Study Group. Effects of lipid lowering by pravastatin on progression and regression of coronary artery disease in symptomatic men with normal to moderately elevated cholesterol levels: the Regression Growth Evaluation Statin Study (REGRESS). Circulation. 1995;91:2528-2540.
Salonen R, Nyyssönen K, Porkkala E, Rummukainen J, Belder R, Park J-S, Salonen JT. Kuopio Atherosclerosis Prevention Study (KAPS): A population-based primary prevention trial of the effect of LDL lowering on atherosclerotic progression in carotid and femoral arteries. Circulation. In press.
Pitt B, Furberg C, McGovern M, for the PLAC-I and PLAC-II Investigators. Reduction in cardiovascular events during treatment with pravastatin: pooled analysis from coronary and carotid atherosclerosis intervention trials. Eur Heart J. 1994;15(suppl):P2565. Abstract.
sas/stat User’s Guide, Version 6. 4th ed. Cary, NC: SAS Institute Inc; 1990;2:1027-1070.
Finney DJ. Statistical Methods in Biological Assay. London, England: Griffen; 1978:80-82.
Brown GB, Zhao XQ, Sacco DE, Albers JJ. Arteriographic view of treatment to achieve regression of coronary atherosclerosis and to prevent plaque disruption and clinical cardiovascular events. Br Heart J. 1993;69:S48-S53.
Superko HR, Krauss RM. Coronary artery disease regression: convincing evidence for the benefit of aggressive lipoprotein management. Circulation. 1994;90:1056-1069.
Furberg CD, Adams HP Jr, Applegate WB, Byington RP, Espeland MA, Hartwell T, Hunninghake DB, Lefkowitz DS, Probstfield J, Riley WA, Young B, for the Asymptomatic Carotid Artery Progression Study (ACAPS) Research Group. Effects of lovastatin on early carotid atherosclerosis and cardiovascular events. Circulation. 1994;90:1679-1687.
Holme I. Relation of coronary heart disease incidence and total mortality to plasma cholesterol reduction in randomized trials: use of meta-analysis. Br Heart J. 1993;69(suppl):S42-S47.
Adams HP, Byington RP, Hoen H, Dempsey R, Furberg CD. Effect of cholesterol-lowering medications on progression of mild atherosclerotic lesions of the carotid arteries and on the risk of stroke. Cerebrovasc Dis. 1995;5:171-177.
Antiplatelet Trialists’ Collaboration. Collaborative overview of randomised trials of antiplatelet therapy, I: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ. 1994;308:81-106.
Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, Eberlein KA, Godwin J, Qizilbash N, Taylor JO, Hennekens CH. Blood pressure, stroke, and coronary heart disease, part 2: short-term reduction in blood pressure: overview of randomised drug trials in their epidemiological context. Lancet. 1990;335:827-838.