Incidence and risk factors of sepsis in hemodialysis patients in the United States

Background: Sepsis is one the most serious and life-threatening complication in patients with chronic hemodialysis (HD) access. Arteriovenous fistula (AVF) use is associated with a lower risk of infection. However, several prior studies identified significantly higher number of patients initiating HD using a catheter (HC) or arteriovenous graft (AVG). The aim of this study was to use a large national renal database to report the incidence and risk factors of sepsis in patients with end-stage renal disease (ESRD) initiating HD access using AVF, AVG, or HC in the United States. Methods: All patients with ESRD initiating HD access (AVF, AVG, HC) between January 1, 2006, and December 31, 2014, in United States Renal Data System were included. International Classification of Diseases, 9th edition-Clinical Modification diagnosis code (038x, 790.7) was used to identify patients who developed first onset of sepsis during follow-up. Standard univariate (Students t-test, chi(2), and Kaplan-Meier) and multivariable (logistic/Cox regression) analyses were performed as appropriate. Results: A total of 870,571 patients were identified, of whom, 29.8% (n = 259,686) developed sepsis. HC (31.2%) and AVG (30.6%) were associated with a higher number of septic cases compared with AVF (22.9%; P < .001). The incident rate of sepsis was 12.66 episodes per 100 person-years. It was the highest among HC vs AVG vs AVF (13.86 vs 11.49 vs 8.03 per 100 person-years). Patients with sepsis were slightly older (mean age 65.09 +/- 14.49 years vs 63.24 +/- 15.17 years) and had higher number of comorbidities including obesity (40.7% vs 37.7%), congestive heart failure (36.6% vs 30.8%), peripheral arterial disease (15.6% vs 12.4%), and diabetes (59.6% vs 53.5%) (all P < .001). After adjusting for potential confounders, compared with AVF, patients with AVG (hazard ratio [HP], 1.35 [95% confidence interval [CI], 1.31-1.40) and HC (HR,1.80 [95% CI,1.77-1.84) were more likely to develop sepsis at 3 years (both P < .001). Compared with patients with no sepsis, sepsis was associated with a three-fold increase the odds of mortality (odds ratio, 3.16; 95% CI, 3.11-3.21; P < .001). Additionally, in patients who developed sepsis, AVF use was associated with significantly lower mortality compared with AVG and HC (73.7% vs 78.7% vs 78.0%; P < .001). After adjusting for significant covariates, compared with AVF, mortality at 1 year after sepsis was 21% higher in AVG (HR, 1.21; 95% CI, 1.15-1.28; P < .001) and nearly doubled in HC (HR, 1.94; 95% CI, 1.88-2.00; P < .001). Conclusions: Sepsis risk in HD patients is clearly related to access type and is associated with dramatic increase in mortality. Initiating HD access with AVF to meet the National Kidney Foundation Kidney Disease Outcomes Quality recommendations should be implemented to reduce the incidence of sepsis and improve survival in patients with ESRD.