Non-cardiovascular mortality, LDL cholesterol and statins: A metaregression analysis

BACKGROUND: As of today, the effect of statins on non-cardiovascular mortality is still being debated. Single studies have not been able to provide definite answers. We performed a meta-regression analysis on randomized statin trials in order to provide evidence that non-cardiovascular mortality is related to statin treatment and low-density lipoprotein (LDL) cholesterol plasma level. METHODS: We selected 29 randomized controlled trials of statins versus placebo, a total of 90,480 patients, with a follow-up of >12 months. Baseline and follow-up LDL levels and all-cause, cardiovascular and non-cardiovascular mortality were recorded. Weighted linear regression analysis was carried out separately for placebo and treatment groups. RESULTS: LDL level was inversely related to overall mortality (p = 0.0105) and non-cardiovascular mortality (p = 0.0171) in the treatment group. By contrast, in the placebo group only non-cardiovascular mortality was inversely correlated to LDL (p = 0.0032). The regression lines have similar slopes and run almost parallel to each other, with the treatment line lying below the placebo line. To identify the threshold of risk for starting statin therapy, we analysed the relationship between baseline cardiovascular risk and overall mortality in the two groups. Both correlations are highly significant and regression lines intersect at a risk of 0.29% per year. This implies that the effects of statins are favourable when the baseline cardiovascular risk exceeds approximately 3% in 10 years. CONCLUSIONS: A trend of increased non-cardiovascular mortality with decreased LDL exists both in placebo and treatment groups. However, at each given LDL cholesterol level, non-cardiovascular mortality is lower in treated patients. Therefore, statin therapy may improve the biological impact of LDL on non-cardiovascular mortality.