ABSTRACT
PURPOSES: Low HDL-C is associated with an increased risk of sepsis-associated AKI and subsequent decline in eGFR. HDL-C possesses anti-inflammatory, antioxidant, and endothelial repair-promoting properties. The use of Apo A-I mimetic peptides, which are the main structural components of HDL-C, has been shown to improve renal function in animal models of sepsis. However, the diagnostic value of low HDL-C in persistent sepsis-associated AKI remains unclear. METHODS: This is a retrospective cohort study based on MIMIC IV (V 2.2). The study population consisted of all adult septic patients admitted to the Beth Israel Deaconess Medical Center Intensive Care Unit from 2008 to 2019, with plasma HDL-C measured within 24 h of ICU admission. The primary endpoint was persistent severe sepsis-associated acute kidney injury (SA-AKI) and the secondary endpoint is kidney replacement therapy (KRT). Logistic regression was used to assess the correlation between HDL-C and persistent severe SA-AKI and KRT, and receiver operating characteristic (ROC) curve analysis was performed to evaluate predictive ability. RESULTS: A total of 604 cases of SA-AKI patients were included in the analysis, among which 88 cases (14.5%) experienced persistent severe SA-AKI. The median (IQR) HDL-C level in the group with persistent severe SA-AKI was lower (33.0 [24.0-45.5]) compared to the non-persistent severe SA-AKI group (42.0 [31.0-53.0]). However, HDL-C showed poor discriminatory ability with an AUROC [95%CI] of 0.62 [0.56-0.69]. Clinical prediction models based on serum creatinine concentration, 24-h creatinine change, APSIIIscore, lactate levels, APTT, and heart rate performed well in predicting persistent severe SA-AKI with an AUROC [95%CI] of 0.876 [0.84-0.91]. However, adding HDL-C to this model did not improve predictive performance. CONCLUSIONS: The plasma HDL-C measured within 24 h after admission to the ICU does not provide a good prediction for persistent severe SA-AKI, and it does not improve the clinical predictive ability compared to conventional variables.
