Peritoneal dialysis catheter implantation by nephrologists is associated with higher rates of peritoneal dialysis utilization: a population-based study

Background — The likelihood of peritoneal dialysis (PD) utilization following a PD catheter insertion attempt is poorly described. We explored the risk factors for PD nonuse, focusing on the method of PD catheter implantation.

Methods — This population-based retrospective cohort study employed Ontario administrative health data to identify 3886 predialysis adults who had an incident PD catheter implantation between 2002 and 2010. The impact of the method of catheter implantation including open-surgical (open, n = 1884), surgical-laparoscopic (laparoscopic, n = 1154), nephrology-percutaneous (nephrology, n = 498) and radiology-percutaneous (radiology, n = 350) on rates of PD utilization (defined as four consecutive weeks of PD) was examined.

Results — Eighty-three percent of study patients received PD. After adjustment, relative to patients with openly inserted catheters, PD utilization was greater for those with nephrology-inserted catheters [adjusted hazard ratio (aHR) 1.59, 95% confidence interval (CI) 1.29-1.95] and similar for radiology-inserted catheters [aHR 1.16, 95% CI 0.94-1.43] or laparoscopic-inserted catheters [aHR 0.97 (95% CI 0.86-1.09)]. Among PD nonusers, death occurred in 10% of the open group, 6% of the laparoscopic group, 27% of the radiology group and in fewer than 3% of the nephrology group. Sixty-nine percent received hemodialysis in the open group, 63% in the laparoscopic group, 61% in the radiology group and 88% in the nephrology group. Those remaining predialysis comprised 12% of the open group, 22% of the laparoscopic group, 11% of the radiology group and <3% of the nephrology group.

Conclusions — Nephrology insertion resulted in lower overall rates of PD nonuse, particularly due to death or remaining predialysis. Greater use may be related to insertion timing, technique or greater commitment on the part of nephrologists to the success of PD.