Sex-specific and metabolic subgroup heterogeneity in high-density lipoprotein cholesterol associations with diabetic kidney disease risk: a retrospective cohort study.

Abstract

Background: The role of high-density lipoprotein cholesterol (HDL-C) in diabetic kidney disease (DKD) remains controversial. This study aimed to delineate the subgroup-specific relationships between the two by exploring cumulative and threshold effects.

Methods: 3,040 patients with type 2 diabetes and no baseline evidence of DKD were included. Cox proportional hazards regression models were performed to investigate the potential relationship between HDL-C level and DKD risk. To address subgroup heterogeneity, sex-stratified restricted cubic splines (RCS) were employed to model nonlinear relationships. The optimal threshold was identified through the maximum selected statistics and validated via 1,000 bootstrap iterations. Subgroup analyses stratified by sex, diabetes duration, and metabolic status were performed to evaluate heterogeneity. Survival analysis using Kaplan-Meier curves further validated these threshold effects.

Results: During a median follow-up of 3.13 years, 665 subjects (21.9%) progressed to DKD. Overall, each 1 mmol/L increase in HDL-C level independently reduced DKD risk by 43%. RCS analysis demonstrated an inverse correlation between HDL-C and DKD risk (P for overall = 0.025, P for nonlinear = 0.317), with increased risk reduction at lower concentrations, plateauing at higher levels. A robust threshold of 0.93 mmol/L was identified, showing significantly stronger protection against DKD progression (hazard ratio (HR) = 0.69, P < 0.001) compared to the traditional cutoff (HR = 0.86, P = 0.109). Females showed continuous protection (HR = 0.41, P = 0.009) without threshold dependency. The male and diabetes duration < 10 years subgroups exhibited threshold effects at > 0.93 mmol/L without continuous protection. The metabolically unstable (hypertension, poorly controlled glycemia, body mass index (BMI) > 28 kg/m²) and BMI < 24 kg/m² subgroups displayed dual effects (P < 0.05). Survival analysis confirmed lower cumulative DKD incidence with HDL-C > 0.93 mmol/L (P = 0.007).

Conclusions: This study reveals sex- and metabolic context-dependent heterogeneity in HDL-C-DKD associations: males and short-duration diabetes exhibited threshold effects (0.93 mmol/L), females showed continuous protection, and subgroups with hypertension, poorly controlled glycemia, or obesity (BMI > 28 kg/m²) exhibited both continuous protection and threshold effects. These findings may inform individualized risk stratification in specific populations.