Arteriovenous Access Type and Risk of Mortality, Hospitalization, and Sepsis Among Elderly Hemodialysis Patients: A Target Trial Emulation Approach.

Abstract

RATIONALE & OBJECTIVE\nEvidence is mixed regarding the optimal choice of the first permanent vascular access for elderly patients on hemodialysis (HD). Lacking data from randomized controlled trials (RCT), we used a target trial emulation approach to compare arteriovenous fistula (AVF) versus arteriovenous graft (AVG) placement among elderly patients on HD.\n\n\nSTUDY DESIGN\nRetrospective cohort study.\n\n\nSETTING & PARTICIPANTS\nElderly patients included in the US Renal Data System (USRDS) who initiated HD with a catheter and had an AVF or AVG placed within six months of starting hemodialysis.\n\n\nEXPOSURE(S)\nPlacement of an AVF versus AVG as the incident arteriovenous access.\n\n\nOUTCOME(S)\nAll-cause mortality, all-cause and cause-specific hospitalization, and sepsis.\n\n\nANALYTICAL APPROACH\nTarget trial emulation approach, high-dimensional propensity score and inverse probability of treatment weighting (IPTW), and instrumental variable (IV) analysis using the proclivity of the operating physician to place fistulas as the instrumental variable.\n\n\nRESULTS\nA total of 19,867 patients were included, with 80.1% receiving an AVF and 19.9% an AVG. In unweighted analysis, AVF placement was associated with significantly lower risk of mortality and hospitalization, especially within 6 months after vascular access placement. In IPTW analysis, AVF placement was associated with lower incidence of mortality and hospitalization within 6 months after placement (HR [95% confidence interval] 0.82 [0.75, 0.91] for mortality; HR 0.82 [0.78, 0.87] for all-cause hospitalization) but not between 6 months and 3 years after access placement. No association between AVF placement and mortality, sepsis, nor all-cause, cardiovascular disease (CVD)-related, or infection-related hospitalization was found in IV analyses. However, AVF placement was associated with lower risk of access-related hospitalization not due to infection.\n\n\nLIMITATIONS\nPotential for unmeasured confounding, analyses limited to elderly patients, and absence of data on actual access use during follow-up.\n\n\nCONCLUSIONS\nUsing observational data to emulate a target RCT, type of initial arteriovenous access placement was not associated with the risk of mortality, sepsis, or all-cause, CVD-, or infection-related hospitalization among elderly patients who initiated hemodialysis with a catheter.