CHOLESTEROL REDUCTION AND ITS IMPACT ON CORONARY-ARTERY DISEASE AND TOTAL MORTALITY

A sample of 42 randomized cholesterol-lowering trials represented by 32 data points were subjected to a metaregression analysis. The logarithmic odds ratio (InOR) of total mortality outcome or incidence of coronary artery disease (CAD) were used as dependent variables and regressed against the net amount of total cholesterol reduction between treatment groups in each trial. Analyses were weighted by variance of OR. Adjustments were also made for single and multifactor trials and for total mortality risk level in the control/placebo group as well as for treatment modality. Analyses were performed both with and without inclusion of trials using 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors as treatment modality. A statistically significant dose-response relationship was found between InOR and the amount of cholesterol reduction for both endpoints and in both groups of trials. These findings were essentially unaltered by various adjustments for potential confounders or by various sensitivity analyses. The results of the Scandinavian Simvastatin Survival Study (4S) fell close to the regression lines for both outcome measures, and inclusion of other statin trials strengthened the significance of this dose-response compared with previous metaregression analyses without inclusion of statin trials. The type of treatment was significantly associated with both endpoints. Fibrate trials did significantly worse on all-cause mortality than statin trials and other drug trials (except hormone). The baseline cholesterol level was also significantly predictive of CAD incidence as a trial outcome; efficacy increased with increased level of baseline cholesterol. However, because the baseline cholesterol level and the percent cholesterol reduction are interrelated (r = 0.40), the relation between baseline cholesterol level and CAD incidence outcome could be confounded with the percent cholesterol reduction. By adjustment for the latter (highly significant), no association could be demonstrated between the baseline cholesterol level and InOR of CAD (p > 0.20), i.e., the effect on CAD outcome is about equal whether the baseline cholesterol level is low or high, as long as the degree of cholesterol reduction (percent) is kept constant. This metaregression analysis showed that appreciable reduction of total mortality risk should be expected with massive cholesterol reduction, at least in populations at moderate and higher CAD risk levels.