Serum cholesterol concentration and death from suicide in men: Paris prospective study I

Abstract Objective : To investigate whether low serum cholesterol concentration or changing serum cholesterol concentration is associated with risk of suicide in men. Design : Cohort study with annual repeat measurements of serum cholesterol concentration (for up to four years). Setting : Paris, France. Subjects : 6393 working men, aged 43-52 in 1967-72, who had at least three measurements of serum cholesterol concentration. Main outcome measures : Individual change over time in serum cholesterol concentration (estimated using within person linear regression method); death from suicide during average of 17 years' follow up after last examination. Results : 32 men committed suicide during follow up. After adjustment for age and other factors, relative risk of suicide for men with low average serum cholesterol concentration (<4.78 mmol/l) compared with those with average serum cholesterol concentration of 4.78-6.21 mmol/l was 3.16 (95% confidence interval 1.38 to 7.22, P = 0.007). Men whose serum cholesterol concentration decreased by more than 0.13 mmol/l a year had multivariate adjusted relative risk of 2.17 (0.97 to 4.84, P = 0.056) compared with those whose cholesterol remained stable (change of </=0.13 mmol/l a year). Conclusion : Both low serum cholesterol concentration and declining cholesterol concentration were associated with increased risk of death from suicide in men. Although there is some evidence in favour of a concomitant rather than a causal effect for interpreting these associations, long term surveillance of subjects included in trials of lipid lowering treatments seems warranted. Key messages In this cohort study of 6393 men with repeated measurements of serum cholesterol concentration, both low serum concentrations and declining con- centrations was associated with increased risk of subsequent death from suicide Our results, together with those of recent trials of lipid lowering drugs, suggest that these associa- tions might not be causal However, long term surveillance of subjects included in trials of lipid lowering treatments seems warranted