Nutritional-metabolic indicators do not independently predict AVF thrombosis after adjustment for surgical frequency:...

Arteriovenous fistula (AVF) thrombosis is a leading cause of vascular access failure in hemodialysis patients. Although nutritional status is often hypothesized to influence thrombotic risk, most previous studies have not accounted for surgical burdens such as repeated surgical procedures. This study assessed whether common nutritional and metabolic markers are independently associated with AVF thrombosis after adjustment for surgical frequency.

This retrospective cohort study included 783 patients who underwent AVF creation between January 2017 and October 2024. Nutritional and metabolic markers were defined using standard formulas. Missing data were handled using multiple imputation. Candidate predictors were identified through univariate analysis (p < 0.20). Multivariate logistic regression was initially performed, followed by Cox proportional hazards regression as the primary analysis to properly account for variable follow-up durations. Model discrimination was assessed by the area under the receiver operating characteristic curve (AUC), calibration by the Hosmer-Lemeshow test, and clinical utility by decision curve analysis. Subgroup and sensitivity analyses were performed.

Among 783 patients (median follow-up 34 months), 96 (12.3%) developed AVF thrombosis. Cumulative surgical burden was the strongest independent risk marker (hazard ratio [HR] = 1.44, 95% CI: 1.28-1.62, p < 0.001). Hypertension showed a borderline association (HR = 2.24, 95% CI: 0.92-5.43, p = 0.080), while diabetes was not independently associated with thrombosis risk (HR = 0.63, 95% CI: 0.38-1.04, p = 0.091). No traditional nutritional indices-including prognostic nutritional index (PNI), albumin-to-globulin ratio (AGR), albumin, high-density lipoprotein (HDL), low-density lipoprotein (LDL), vitamin B12, or calcium-phosphate parameters-retained significance after adjustment. The Cox model demonstrated good discriminatory ability (Harrell's C-index = 0.74). In complete-case sensitivity analysis (n = 165), HDL and vitamin B12 showed associations with thrombosis, though these findings were likely influenced by the substantially reduced sample size and potential selection bias, whereas the primary predictors remained consistent.

After accounting for surgical frequency, conventional nutritional and metabolic indices did not demonstrate independent associations with AVF thrombosis in the present study. Patients with high cumulative surgical burden should be recognized as a high-risk population warranting enhanced surveillance, and optimizing blood pressure control should remain a clinical priority.