LAUSR

BACKGROUND Elevated serum high-density lipoprotein cholesterol (HDL-C) has not been associated with better cardiovascular (CV) and all-cause mortality in hemodialysis patients. However, the association between change in HDL over time and mortality has not been fully examined. OBJECTIVE In a nationally representative cohort of incident hemodialysis patients who had available HDL data at baseline and 6 months after dialysis initiation, we studied the association of change in HDL-C during the first 6 months of dialysis with all-cause and CV mortality. METHODS Associations between HDL-C change and mortality were determined in Cox proportional hazard regression models with adjustment for multiple variables. RESULTS In case-mix models, there was a J-shaped association between change in HDL-C and mortality, such that quartiles 1 (<-5 mg/dL) and 4 (≥7 mg/dL) were each associated with higher all-cause (hazard ratio, 1.32 [95% confidence interval, 1.21-1.45] and 1.09 [1.01-1.18]) and CV (1.28 [1.06-1.55] and 1.23 [1.04-1.45]) death risk, respectively. In fully adjusted models that included indices of malnutrition and inflammation, the higher death risk observed in the lowest quartile was attenuated, whereas the highest quartile continued to demonstrate significantly higher all-cause (1.11 [1.02-1.20]) and CV mortality (1.15 [1.00-1.32]). These associations persisted across various subgroups. CONCLUSIONS Although malnutrition and inflammation may explain the increased risk for mortality in patients with decreasing serum HDL-C concentrations over time, these indices do not mitigate the elevated risk in patients with rising serum HDL-C. We found that increasing serum HDL-C over time is paradoxically associated with worse outcomes in incident hemodialysis patients.