The interactive effects of blood cholesterol levels and other risk factors, such as age and blood pressure (BP), on vascular mortality are uncertain.OBJECTIVE
To determine whether patient age, sex, and BP modify the relationship between blood cholesterol level and vascular mortality.DESIGN
This meta-analysis by the Prospective Studies Collaboration (PSC) included prospective studies that recorded BP, blood cholesterol levels, and date of birth at enrollment, had more then 500 person-years of follow-up, and cited vascular mortality as a primary end point. Studies were identified using the EMBASE and MEDLINE databases, searches of meeting abstracts, and through personal communication with study authors. Studies were not included if they enrolled patients on the basis of a history of stroke or heart disease. Individual patients with a total blood cholesterol level exceeding 12.0 mmol/l, an HDL cholesterol level 3.5 mmol/l or higher, or a total cholesterol: HDL cholesterol ratio greater than 12 were excluded from the analyses, as were patients who were not followed up between the ages of 40 and 80 years, and those with a history of stroke or heart disease at enrollment.OUTCOME MEASURES
The outcomes were death from ischemic heart disease (IHD), stroke (both ischemic and hemorrhagic), and other vascular causes.RESULTS
The meta-analysis included data from 892,337 patients enrolled in 61 studies, with a total of 11.6 million person-years of follow-up. Overall, there were 55,262 deaths; 33,744 were caused by IHD, 11,663 resulted from stroke, and 9,855 had other vascular causes. The mean follow-up was 13 years. In both sexes, a 1.0 mmol/l decrease in total cholesterol was associated with a 56% reduction in IHD mortality in patients aged 40-49 years (hazard ratio [HR] 0.44, 95% CI 0.42-0.48), a 34% reduction in patients aged 50-69 years (HR 0.66, 95% CI 0.65-0.68), and a 17% reduction in patients aged 70-89 years (HR 0.83, 95% CI 0.81-0.85; P <0.0001 for trend by age). This risk reduction diminished, however, at increasing levels of systolic BP. The effect of total cholesterol level on IHD mortality was most pronounced among individuals who were obese or who were smokers. A strong negative correlation between HDL cholesterol level and IHD mortality was observed for each of the age-groups studied. By contrast, IHD mortality had strong positive associations with the non-HDL cholesterol level and the total cholesterol: HDL cholesterol ratio in both sexes. The most robust predictive indicator of IHD mortality was the total cholesterol: HDL cholesterol ratio (χ2 = 627). In terms of stroke mortality, there was a positive, but non significant, association with total cholesterol in patients aged 40-59 years (HR 0.89, 95% CI 0.77-1.03), and a marginally significant positive association with total cholesterol: HDL cholesterol ratio in patients aged 40-69 years (HR 0.86, 95% CI 0.74-0.99); these associations were not evident in the older age-groups.CONCLUSION
A decrease in total cholesterol level was associated with a reduction in IHD mortality in patients of all ages and systolic BPs.