Comparing Catheters with Fistulas in Older Patients Starting Hemodialysis: A Pilot Randomized Controlled Trial

We assessed the feasibility of randomizing hemodialysis patients to an attempt at creating an arteriovenous fistulas with central venous catheters for vascular access. There was no signal of harm associated with randomization to the catheter strategy, and most outcomes favored the catheter group. Patients had strong preferences regarding choice of vascular access, reinforcing the central importance of shared decision making.

Fistulas are the preferred form of vascular access and historically have been recommended by guidelines because they were associated with improved clinical outcomes and costs. However, there has never been a randomized comparison of catheters with fistulas, and the evidence on which recommendations were based is observational with potential for bias.

A pilot randomized controlled trial was conducted in 12 hemodialysis centers in Canada and Australia. Patients aged 55 years and older who started hemodialysis using a tunneled or nontunneled central venous catheter for vascular access were randomized to an attempt at fistula creation or continued use of a tunneled central venous hemodialysis catheter. The primary outcome was feasibility, as measured by the proportion of eligible patients who consented to randomization and the proportion of patients randomized to the fistula group who underwent fistula placement within 90 days. Secondary outcomes included invasive access-related procedures, hospitalizations, access-related infections, tissue plasminogen activator usage, and death.

A total of 1287 patients were screened, and 67 patients underwent randomization. A total of 25% of eligible patients agreed to randomization, and 71% randomized to the fistula arm underwent fistula placement within 90 days. The most common reason patients declined to participate in the trial was that they preferred to continue to dialyze with their catheter. Patients randomized to fistulas had nominally more access-related interventions, hospitalizations, and bacteremia (most of which were catheter related) while the catheter group experienced greater reliance on thrombolytic therapy.

Findings from this trial suggest that conducting a definitive randomized trial of fistulas versus catheters is not feasible in the current climate. There was no signal of harm associated with randomization to the catheter strategy, and most outcomes favored the catheter group. Patients exhibited strong preferences regarding choice of vascular access.

ClinicalTrials.gov, NCT02675569 .