Age-specific associations between serum cholesterol levels and suicidal behaviors in patients with depressive disorders: A naturalistic prospective observational cohort study

Abstract

Introduction: This study investigated the effects of total cholesterol levels on prevalent, and incident suicidal behaviors according to age group (<60 vs. ≥60 years) in depressed patients. Methods: Consecutive outpatients with depressive disorders who visited the Chonnam National University Hospital between March 2012 and April 2017 were recruited. Among 1,262 patients assessed at baseline, 1,094 agreed to blood sampling for measurement of serum total cholesterol levels. Among the patients, 884 completed the 12-week acute treatment phase and were followed up at least once during the 12-month continuation treatment phase. Suicidal behaviors assessed at baseline included baseline suicidal severity; behaviors assessed at the 1-year follow-up included increased suicidal severity and fatal/non-fatal suicide attempts. Associations of baseline total cholesterol levels with the above-mentioned suicidal behaviors were analyzed using logistic regression models after adjustment for relevant covariates. Results: Of 1,094 depressed patients, 753 (68.8%) were women. The mean (SD) age of patients was 57.0 (14.9) years. Lower total cholesterol levels (87–161 mg/dl) were associated with increased suicidal severity (Linear Wald = 4.478, p < 0.05) and fatal/non-fatal suicide attempt (Linear Wald = 7.490, p < 0.01) in patients <60 years of age. U-shaped associations between total cholesterol levels and 1-year follow-up suicidal outcomes (increased suicidal severity, Quadratic Wald = 6.299, p < 0.05; fatal/non-fatal suicide attempt, Quadratic Wald = 5.697, p < 0.05) were observed in patients ≥60 years of age. Conclusions: These findings suggest that differential consideration of serum total cholesterol levels according to age group may have clinical utility for predicting suicidality in patients with depressive disorders. However, because our research participants came from a single-hospital, the generalizability of our findings may be limited.