Peritoneal dialysis-associated peritonitis outcomes reported in trials and observational studies: A systematic review.

Background: Peritoneal dialysis (PD)-associated peritonitis carries significant morbidity, mortality, and is a leading cause of PD technique failure. This study aimed to assess the scope and variability of PD-associated peritonitis reported in randomized trials and observational studies. Methods: Cochrane Controlled Register of Trials, MEDLINE, and Embase were searched from 2007 to June 2018 for randomized trials and observational studies in adult and pediatric patients on PD that reported PD-associated peritonitis as a primary outcome or as a part of composite primary outcome. We assessed the peritonitis definitions used, characteristics of peritonitis, and outcome reporting and analysis. Results: Seventy-seven studies were included, three were randomized trials. Thirty-eight (49%) of the included studies were registry-based observational studies. Twenty-nine percent (n = 22) of the studies did not specify how PD-associated peritonitis was defined. Among those providing a definition of peritonitis, three components were reported: effluent cell count (n = 42, 54%), clinical features consistent with peritonitis (e.g. abdominal pain and/or cloudy dialysis effluent) (n = 35, 45%), and positive effluent culture (n = 19, 25%). Of those components, 1 was required to make the diagnosis in 6 studies (8%), 2 out of 2 were required in 22 studies (29%), 2 out of 3 in 11 studies (14%), and 3 out of 3 in 4 studies (5%). Peritonitis characteristics and outcomes reported across studies included culture-negative peritonitis (n = 47, 61%), refractory peritonitis (n = 42, 55%), repeat peritonitis (n = 9, 12%), relapsing peritonitis (n = 5, 7%), concomitant exit site (n = 16, 21%), and tunnel infections (n = 8, 10%). Peritonitis-related hospitalization was reported in 38% of the studies (n = 29), and peritonitis-related mortality was variably defined and reported in 55% of the studies (n = 42). Peritonitis rate was most frequently reported as episodes per patient year (n = 40, 52%). Conclusion: Large variability exists in the definitions, methods of reporting, and analysis of PD-associated peritonitis across trials and observational studies. Standardizing definitions for reporting of peritonitis and associated outcomes will better enable assessment of the comparative effect of interventions on peritonitis. This will facilitate continuous quality improvement measures through reliable benchmarking of this patient-important outcome across centers and countries.