Trajectory of total cholesterol in the last years of life over age 80 years: Cohort study of 99,758 participants

Abstract

Background: Epidemiological studies suggest that lower total cholesterol (TC) may be associated with higher mortality. This study aimed to evaluate whether a decline in TC before death might account for the association of TC with mortality over age 80 years. Methods: Cohort study using primary care electronic health records of 99,758 participants aged 80-105 years from the UK Clinical Practice Research Datalink. Hazard ratios for all-cause mortality were adjusted for age, gender, frailty, comorbidity, blood pressure, and smoking. Fractional polynomial models were fitted to evaluate longitudinal trends in TC before death or end-of-study. Adjusted odds ratios were estimated using generalized estimating equations. Results: There were 63,630 women and 36,128 men, mean age 86 years, with 29,200 deaths. There were 41,164 treated with statins at cohort entry. Compared with TC values of 4.5-5.4 mmol/L, TC values <3.0 mmol/L were associated with higher mortality (statin treated hazard ratios 1.53, 95% confidence interval 1.43-1.64, p < .001; not treated, 1.41, 1.29-1.54, p < .001). A secular decline in TC values accelerated in the last 2 years of life. In the last quarter of follow-up, the adjusted odds of TC < 3.0 mmol/L for those who died, compared with surviving participants, were 3.33 (2.84-3.91, p < .001) for untreated and 1.88 (1.68-2.11, p < .001) for statin-treated participants. Conclusions: TC values show a terminal decline in the last years of life. Reverse causation may contribute to the association of lower TC with higher mortality in nonrandomized studies.