Quantifying vascular access-associated excess mortality in maintenance hemodialysis patients

Central venous catheters (CVCs) are commonly linked with higher mortality in hemodialysis (HD) patients compared with arteriovenous accesses (AVAs). However, patients with CVCs often have greater comorbidities, complicating causal interpretation. This study aimed to assess the association between vascular access type and survival adjusting for relevant confounders.

In this retrospective cohort study, data from 146,967 incident HD patients treated between 2016 and 2019 at a large North American dialysis organization (Fresenius Medical Care North America) were analyzed. Multiple analytic strategies were conducted including inverse probability treatment weighted and time-dependent survival analyses.

Among 146,967 incident HD patients, median survival was 1106 days for those initiating with a CVC compared with 1290 days for patients with an AVA, corresponding to a 184-day difference and an 88% restricted mean survival time (RMST) ratio. In the sustained access analysis, median survival was 448 days for CVC-only vs 1226 days for AVA-only patients (RMST difference = 778 days, RMST ratio = 52%). After inverse probability treatment weighting, AVA initiation was associated with a 25% lower mortality risk (hazard ratio: 0.75, 95% confidence interval: 0.73-0.76) and sustained AVA use with a 62% lower risk (hazard ratio: 0.38, 95% confidence interval: 0.36-0.40). Differences in infection-related deaths between the groups were small (8.6%-10.6% of deaths in all comparison groups).

CVC use was associated with higher mortality compared with AVA. Although AVA use remained linked with better survival across analyses, the precise magnitude of any access-related benefit cannot be determined within the constraints of observational data. There are strong indications that the excess risk at least partially reflects differences in baseline health and patient selection rather than a direct causal effect.