BACKGROUND: Though cholesterol-lowering therapy can reduce cardiovascular morbidity and mortality, earlier studies raised concerns that reducing cholesterol concentrations might increase the risk of cancer and deaths from suicides, accidents, and violence (ie, nonillness mortality).
POPULATION STUDIED: This meta-analysis included clinical trials of cholesterol-lowering treatments in which participants were randomly assigned to a cholesterol-lowering intervention group or a control group. The investigators only included trials designed to measure effects of treatment on clinical events and mortality. Most participants were men aged between 40 and 70 years.
STUDY DESIGN AND VALIDITY: The studies were identified using an ancestry approach (locating previous studies cited in reference lists of already identified studies cited in reference articles) and a MEDLINE computer literature search from 1966 to March 2000. A total of 21 trials met inclusion criteria, though only 15 contained data on nonillness mortality. Some investigators provided previously unpublished data for 4 trials, yielding data from 19 of the 21 trials. The included studies investigated cholesterol-lowering either by drug therapy, diet modification, or both. The most common reasons for exclusion were the use of multifactorial risk interventions and studies not designed to monitor clinical events and cause-specific mortality.
OUTCOMES MEASURED: The outcomes were nonillness mortality, including suicides, accidents, and violence.
RESULTS: The studies that met inclusion criteria generated approximately 338,000 patient-years of randomized clinical trial data. Overall, treatments with the goal of lower total cholesterol did not affect the rate of nonillness mortality (odd ratio [OR]=1.18; 95% confidence interval [CI], 0.91-1.52). There was no effect in studies of primary prevention, secondary prevention, or studies of the “statin” drugs. Trials of diet and non-statin drugs (13 trials including 39,260 patients) also did not show a difference, although there was a trend toward increased nonillness mortality (OR=1.32; 95% CI, 0.98-1.77; P=.06). The absolute risk increase in these trials was 4.7 per 1000, which translates to a number needed to harm of 213.The rate of nonillness mortality was not related to the degree of cholesterol reduction.
This meta-analysis did not show a statistically significant relationship between cholesterol lowering and increased risk of nonillness mortality. We would not be benefiting our patients, especially those at highest risk for cardiovascular disease, by limiting our use of lipid-lowering therapy because of this theoretical concern.