In this study, men and women aged between 40 years and 80 years with nonfasting blood total cholesterol concentrations of at least 3.5mmol/L (135mg/dL) were eligible if they had diabetes mellitus, coronary disease, occlusive disease of noncoronary arteries, or treated hypertension (also if a man and aged at least 65 years). Before randomisation, patients received four weeks of placebo followed by four to six weeks of 40mg simvastatin daily. Only those compliant patients who didn’t have a major problem during the run-in were randomly allocated to receive 40mg simvastatin daily (SG, n=10,269) or matching placebo (PG, n=10,267). Although the paper doesn’t mention if allocation was masked, visiting the study website reveals that the allocation occurred via a central randomisation service. They saw the participants intermittently for five years. The main outcome, evaluated by intention to treat, was the time to the first major coronary event (defined as nonfatal myocardial infarction or death from coronary disease), and the secondary outcome was the time to the first major vascular event (defined as a major coronary event, stroke of any type, and a coronary or noncoronary revascularisation). They randomised 5,963 patients with diabetes mellitus and 14,573 high-risk patients without diagnosed diabetes. Among the diabetic patients, 279 (9.4%) SG patients had major coronary events compared with 377 PG (12.6%, number needed to treat [NNT]=31). Among the high-risk patients without diabetes, major coronary events occurred in 619 (8.5%) SG patients and in 835 (11.5%; NNT=33) PG patients. Among the diabetic patients, 149 (5%) SG patients had a stroke compared with 193 (6.5%; NNT=67) PG patients. Among the high-risk patients without diabetes, 295 (4%) SG patients and 392 (5.4%; NNT=72) PG patients had a stroke. Also significant, 260 (8.7%) SG patients with diabetes underwent revascularisations compared with 309 (10.4%) PG patients (NNT=59). Finally, 679 (9.3%) nondiabetic SG patients and 896 (12.3%; NNT=34) PG patients had revascularisations. When all these numbers are pooled, a common way to look at all bad things (and to make the data look even more impressive), 601 (20.2%) diabetic SG patients and 748 (25.1%; NNT=21) PG patients had a major vascular event. Among those without diabetes, 1,432 (19.6%) SG patients and 1,837 (25.2%; NNT=18) PG patients had major events. All the NNTs are unlikely to have been caused by random error and are for five years of treatment. The authors focus on the relative risk reductions rather than the absolute risk reductions. These are always much more impressive sounding and are misleading.
In this well-done study on a narrow subset of patients with diabetes or at high risk of coronary artery disease, simvastatin modestly reduces the frequency of coronary and vascular events. It is unclear whether these same effects would be seen in the real world.
(Level of evidence = 1b)
MRC/BHF Heart Protection Study of cholesterol-lowering with simvastatin in 5963 people with diabetes: a randomised placebo-controlled trial. Lancet 2003;361:2005-16.