On-Target Low-Density Lipoprotein Cholesterol in Adults with Diabetes Not at High Cardiovascular Disease Risk Predicts Greater Mortality, Independent of Early Deaths or Frailty

Abstract

Background/Objectives: Lowering low-density lipoprotein cholesterol (LDL-C) to <70 mg/dL is recommended for most patients with diabetes. However, clinical trials investigating subjects with diabetes who are not at high cardiovascular risk are inconclusive regarding the all-cause mortality benefit of the current target, and real-world studies suggest greater mortality. We aimed to assess the all-cause mortality at different LDL-C levels among subjects with diabetes not at high risk and to examine the potential roles of early deaths and frailty for this greater mortality. Methods: We followed 2098 such participants of the ELSA-Brasil cohort between 2008 and 2019. Results: Over 10.3 (1.4) years of follow-up, 204 (9.7%) individuals died. In the proportional hazards models, participants with LDL-C values < 100 mg/dL and <70 mg/dL had greater adjusted mortality compared to those with LDL-C 100-129 mg/dL (HR = 1.67; 95%CI 1.21-2.30 and HR = 2.27; 95%CI 1.51-3.41, respectively). Increased risk when LDL-C was <100 mg/dL was higher in those >60 years (HR = 2.12; 95%CI 1.35-3.34) and greatest for deaths due to cancer (HR = 2.55; 95%CI 1.10-5.91). Further analyses for those with LDL-C < 100 mg/dL that excluded early deaths and adjusted for the frailty phenotype (HR = 2.01; 1.19-3.41) or frailty index (HR = 1.92; 1.17-3.16) did not materially alter the results. The risk of death across the spectrum of LDL-C was U-shaped, with a nadir at 112.2 mg/dL. Conclusions: The higher risk of all-cause mortality in these subjects with LDL-C within currently recommended levels was not explained by early deaths or frailty. Given the recent decline in cardiovascular mortality and the increased risk of cancer and infections in persons with diabetes, the clinical significance of low LDL-C in diabetes requires reconsideration and the definition of LDL-C treatment targets in diabetes warrants further trial evaluation.

Keywords: LDL; cardiovascular risk; cholesterol; diabetes mellitus; mortality; type 2.

Figure 1

Risk of death (gray area represents the zone of 95% confidence) according to LDL-cholesterol levels in low CVD risk participants. Associations were obtained through restricted cubic spline analyses using Cox proportional hazards models adjusted for age, sex, race, per capita income, private health insurance, smoke, statin use, waist–hip ratio, HbA1c, albumin–creatinine ratio, estimated glomerular filtration rate, and systolic blood pressure. The y-axis to the left indicates the hazard ratio of death. The y-axis to the right indicates the relative frequency (%) of LDL-cholesterol levels displayed in the superimposed histogram at the bottom of the figure.