#580 Vascular access-associated excess mortality

Abstract Background and Aims Central venous catheters (CVC) are one of three vascular access types used for hemodialysis. While they are suitable as the primary options for select patient groups, they are generally considered inferior to arteriovenous access based on associations with impaired outcomes. However, it is not conclusively understood whether such associations are confounded by overall state of health and if so, by what extent. Method Patient data was extracted from electronic health records of a large US-based dialysis organization. In total, 160,753 patients initiating hemodialysis between 2016 and 2019 were included in the study population. Kaplan–Meier survival analyses were conducted stratifying for type of vascular access. Propensity score matching was further applied to mitigate the potential confounding of health status at time of initialization using demographic, clinical, and laboratory variables. Results In summary, 115,204 (78.4%) patients started on a CVC, 26,478 (18.0%) on an arteriovenous fistula (AVF), and 5,258 (3.6%) on an AV graft (AVG). CVC patients had higher rates of ischemic heart disease, heart failure, and history of myocardial infarction at baseline than controls. Patients who initiated hemodialysis on a catheter also had a shorter median survival than controls (1,106 vs. 1,290 days). Later conversion from a catheter to a fistula or graft increased the survival penalty compared to controls. Propensity score matching reduced survival differences, indicating confounding by baseline health. Receiving hemodialysis exclusively via AVF/AVG was not associated with a survival advantage compared to intermittent CVC use. (all Table 1) Cause-specific mortality was higher in CVC patients for both infections and cardiovascular causes. However, the proportion of infection-associated deaths in CVC patients is comparable to controls. CVC patients also showed signs of chronic inflammation in blood samples but had similar serum potassium, phosphorus, and calcium levels compared to AVF/AVG users. Conclusion While limitations in the present analysis such as selection bias should be taken into consideration, the survival disadvantage historically associated with CVCs in hemodialysis patients might not be based on a causal relationship. There is an increasing body of evidence suggesting that differences in vascular access-associated mortality are confounded by overall patient health in general and cardiovascular health in specific.