To examine the effects of different types of cholesterol lowering interventions on mortality outcomes.

Searches were conducted of MEDLINE and EMBASE (1966 through October 1996).

Study designs of evaluations included in the review

Randomised controlled trials (RCTs) of any cholesterol-lowering intervention irrespective of duration were included. Angiographic studies meeting inclusion criteria were also included. Study duration ranged from 6 months to 10 years.

Specific interventions included in the review

The following cholesterol-lowering interventions were included: HMG-CoA reductase inhibitors (statins) including lovastatin, pravastatin, and simvastatin; n-3 Long chain fatty acids and precursors; fibrates including clofibrate, gemfibrozil and bezafibrate; resins including colestipol and cholestyramine; niacin; hormones including estrogen and thyroxin; dietary interventions including low fat, usual diet, olive oil, soya bean, diet drugs, and strict diet; probucol; partial ileum bypass surgery; and placebo drug therapy. Trials with multiple drug interventions were excluded.

Participants included in the review

Patients being treated for primary or secondary prevention of coronary artery disease were included. Baseline cholesterol, where stated, ranged from 4.3 mmol/L to 8.4 mmol/L.

Outcomes assessed in the review

Outcomes include total mortality, mortality from coronary heart disease, and mortality from causes other than coronary heart disease.

How were decisions on the relevance of primary studies made?

The authors do not state how the papers were selected for the review, or how many of the authors performed the selection.

The authors do not state that they assessed validity.

Missing information on mortality data were requested from original authors.

How were the studies combined?

A weighted average risk ratio (RR) and 95%confidence interval was calculated using a random-effects model. The death rate from coronary heart disease per 1000 person-years was calculated for the control group in each trial (reflects the risk for participants). Coronary heart disease rates were then combined for each intervention category with each trial being weighted by the inverse of the variance. Any treatment effects for a specific drug category was compared with the total of all other cholesterol-lowering interventions.

How were differences between studies investigated?

Heterogeneity was assessed using the Breslow-Day test.

The relationship between the type of cholesterol-lowering interventions with coronary artery disease and overall mortality was explored using meta-regression with a weighted least-squares linear regression model. The following independent variables were entered in the meta-regression: type of intervention (HMG- CoA reductase inhibitors vs all other interventions); trial setting (primary vs secondary prevention); trial design (unifactorial vs multifactorial intervention); baseline risk for coronary artery disease; and absolute and relative reductions of cholesterol.

A total of 66 RCTs were included (173160 patients). One trial which did not report cause specific mortality data and 6 trials reporting on multiple drug interventions were excluded.

Statins assessed in 13 RCTs (27507 patients).

n-3 long chain fatty acids assessed in 3 RCTs (2697 patients).

Fibrates assessed in 12 RCTs (21408 patients from table one).

Resins assessed in 8 RCTs (6537 patient).

Hormones assessed in 8 RCTs (13554 patients from table 1).

Niacin acids assessed in 2 RCTs (4128 patients).

Probucol assessed in 1 RCT (118 patients).

Partial ileum bypass surgery assessed in 1 RCT (838 patient).

Diet assessed in 16 RCTs (107523 patients).

Cholesterol reduction rates ranged from 22.86% for statins (range 12.8% to 32%) to 3.07% for n-3 long chain fatty acids and precursers (range 0.1% to 5.1%).

Coronary heart disease (CAD) mortality according to intervention was as follows:

HMG-CoA reductase inhibitors: Baseline risk of death from CAD was 2%. Statistically significant reduction in coronary heart disease mortality: RR = 0.69 (95% CI: 0.59, 0.80). No heterogeneity detected. Result was statistically significantly different from summary risk estimated for all other interventions for which RR = 1.03 (95% CI: 0.86, 1.24). Difference P = 0.02.

Resins: risk ratio borderline. RR 0.71 (95% CI: 0.51, 0.99). Significant heterogeneity.

All other interventions: none reached statistical significance.

Overall mortality according to intervention was as follows:

HMG-CoA: statistically significant reduction. RR = 0.79 (95% CI: 0.71, 0.89). No statistical heterogeneity shown. Result was statistically significantly different from summary risk estimated for all other interventions for which RR = 1.03 (95% CI: 0.86, 1.25). Difference P = 0.02.

N-3 fatty acids: statistically significant reduction. RR = 0.68 (95% CI: 0.53, 0.88). No heterogeneity detected. Result was statistically significantly different from summary risk estimated for all other interventions for which RR = 1.00 (95% CI: 0.88, 1.13). Difference P = 0.007.

Treatment with hormones was associated with a borderline increase in overall mortality. RR = 1.09 (95% CI: 1.00, 1.20). No heterogeneity was detected.

All other interventions: none reached statistical significance.

Mortality from causes other than coronary heart disease:

HMG-CoA reductase inhibitors were not associated with any significant change in other-cause mortality with RR = 0.97 (95% CI: 0.81, 1.16).

All other cholesterol-lowering interventions were associated with a borderline increase in other-cause mortality with RR = 1.51 (95% CI: 1.07, 2.13).

Hormones: Treatment with hormones were associated with a borderline increase in other-cause significantly with RR = 1.29 (95% CI: 1.06, 1.57) and a borderline increase in mortality from all causes with RR = 1.09 (95%CI: 1.00, 1.20).

Results from meta-regression:

Univariate analysis: type of intervention was highly significantly related to coronary heart disease and overall mortality (HMG-CoA reductase inhibitors vs all other interventions; P = 0.006). Other factors significantly associated with overall mortality were: trial setting (lower mortality in primary vs secondary prevention; P = 0.0001); trial design (lower mortality in unifactorial vs multifactorial intervention; P = 0.0001); baseline risk for coronary artery disease (higher mortality reductions in trials in which patients were at higher risk; P = 0.0006); and absolute and relative reductions of cholesterol (higher mortality reduction when cholesterol reduction was greater; P = 0.0001 in both cases).

Factors additionally related to coronary heart disease mortality were absolute and relative reductions of cholesterol (higher mortality reduction when cholesterol reduction was greater; P = 0.001 and P = 0.002 respectively).

Multi variable analysis: after entering trial setting, trial design, and baseline risk for coronary artery disease, the type of intervention still explained a statistically significant degree of variability for both cardiovascular and all-cause mortality (P = 0.004 for overall mortality and P = 0.0001 for coronary heart disease mortality). When the degree of cholesterol reduction was introduced, no other variable explained a statistically significant degree of the remaining variability of either end point.

Statins have the largest effect on the reduction of cardiovascular and all-cause mortality and this result recommends their use in preference to other antilipidemic agents. The greater effect of statins is likely due to the larger reduction in cholesterol.

The aims and inclusion criteria were stated. Relevant details of the primary studies and the results were clearly presented in tabular format. Heterogeneity was assessed and the influence of various factors on the outcome was evaluated.

Fuller details of the literature search such as keywords used would have been welcome. No details were given of methods used to select primary studies or extract data. Validity was not assessed. Adverse reactions to therapies other than mortality were not reported. The numbers in the tables were sometimes inconsistent (fibrates and hormones give different numbers of subjects depending on the table used to calculate the total sample size). Some trials were mentioned more than once in the tables which could have resulted in the control group being counted twice, leading to problems where one class of treatment is compared with all the rest.

The authors' conclusions were supported by the evidence presented.

Practice: The authors consider that statins (HMG-CoA reductase inhibitors) are currently the only cholesterol-lowering drugs that should be used in primary and secondary prevention of heart disease. However their use in hypercholesterolaemic patients at very low risk of coronary events remains questionable due to the limited data on long term safety.

Research: The authors consider that favourable results from trials of n-3 fatty acids may warrant investigation in a larger trial.

Swiss National Research Foundation.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.