The new study, published in Frontiers in Nephrology, drew on ten years of data from ten Kaiser Permanente Southern California hospitals between 2014 and 2023. Researchers led by Max T. Hukill of the Kaiser Permanente Bernard J. Tyson School of Medicine, together with Andrea M. Yeung and Talha H. Imam of the Southern California Permanente Medical Group, mined the Renal Registry of the Southern California Permanente Medical Group for quarterly records of exit-site infection episodes and program sizes. Their central finding: each additional calendar year was associated with a 6.4 percent relative decrease in infection rate, a statistically significant trend that compounded into an estimated 45.1 percent decline across the study period.

What makes the result remarkable is not just its magnitude but its setting. Unlike the small, tightly controlled trials that dominate the literature on dialysis complications, this analysis captured the messy reality of routine clinical practice across a large integrated health care system. The overall mean exit-site infection rate across the decade was 0.86 episodes per 100 patient-months, with observed yearly means falling from 1.26 in 2014 to 0.72 in 2023. Nine hospitals reported data throughout the full study period; a tenth joined in 2018. After excluding seven hospital-quarters with missing records, the final analytic sample comprised 375 hospital-quarters of data.

The stakes of these infections are well documented. Peritonitis, the most frequent reason for peritoneal dialysis catheter removal and patient dropout, has long been known to follow exit-site infections. Researchers first noted the connection in the 1980s, when patients with an infected exit site were found to be more likely to develop peritonitis than those without. A strong association between exit-site infections and subsequent peritonitis persists up to 60 days after the initial infection diagnosis. The International Society of Peritoneal Dialysis now recommends that overall peritonitis rates should not exceed 0.4 episodes per patient-year at risk, a benchmark that reflects decades of improvement in prevention strategies.

Yet while peritonitis rates have been extensively tracked by major registries, exit-site infections have remained a blind spot. Large databases such as the Australia and New Zealand Dialysis and Transplant Registry and the US Renal Data System generally do not record or report exit-site infection rates. The limited studies that do exist report rates ranging from 0.06 to 0.42 episodes per year, but most were small-scale trials not representative of real-world conditions. One larger study involving 1,133 patients over one year found an exit-site infection rate of 0.13 episodes per year, while a Portuguese study reported 0.41 episodes per year with an average of 726 patients over two years. The new analysis is the largest retrospective examination of exit-site infection rates published to date.

To handle the unusual statistical character of the data, the researchers employed a Tweedie generalized linear mixed model, a technique suited to semicontinuous, right-skewed distributions with an exact point mass at zero. Nearly 18 percent of hospital-quarters recorded no infections at all, creating a spike at zero that conventional models struggle with. The model included fixed effects for calendar year and program size, along with a random intercept and slope for calendar year by hospital, allowing each center’s rate of change over time to vary. Confidence intervals for pooled rates were generated by a location-cluster bootstrap with 5,000 replicates, resampling centers with replacement.

Crucially, the decline was not explained by growth in program size. After adjustment for calendar year, program size showed no significant association with infection rate, with a rate ratio of 0.76 per 100-patient increase and a confidence interval spanning 0.52 to 1.11. Variance inflation factors for both predictors were low at 1.39, indicating that the null finding was not an artifact of collinearity between the two variables. The cohort itself grew substantially over the decade, from roughly 631 patients per quarter across nine hospitals in 2014 to 933 patients per quarter across ten hospitals in 2023, yet infection rates fell regardless.

The authors suggest that evolving infection prevention practices, staff training, and care coordination may have driven the improvement, though their data do not permit direct testing of these mechanisms. The care protocols in place across the system are extensive. All catheters were placed with advanced laparoscopic techniques with pre-insertion antibiotics. Exit sites were created with minimal trauma and optimized positioning, avoiding locations under abdominal folds, along belt lines, or facing downward. Daily exit-site care involved washing, topical antibiotics such as mupirocin or gentamicin, sometimes with additional sodium hypochlorite spray, and sterile dressings. Nurses provided continuous support and regular follow-up, monitoring patient practices and adherence with retraining as needed. A surveillance system tracking infections helped identify potential problems before they progressed.

The analysis also revealed significant heterogeneity between hospitals in their rates of decline, with a random slope variance that a likelihood ratio test confirmed was superior to a simpler model. While most centers showed declining trends broadly consistent with the population-level estimate, one center showed a trend in the opposite direction. Given the limited number of hospitals, center-specific estimates carry substantial uncertainty and should be interpreted as hypothesis-generating rather than confirmed hospital-level differences. Still, the heterogeneity indicates that a single shared temporal trend does not adequately describe hospital-level variation, possibly reflecting differences in local infection prevention practices, staffing, or patient case-mix not captured in the analysis.

The study has limitations that temper its conclusions. The dataset lacked baseline demographics of the peritoneal dialysis cohort, preventing analysis of differences in age, sex, ethnicity, and dialysis modality over the study period. As a representative approximation, the current population of 1,481 patients across the system is 61 percent male with a median age of 61 years, and 67.2 percent use continuous cyclic peritoneal dialysis. Program size, patient-months at risk, and discrete infection episode counts were not available as independent quantities, so the researchers modeled infection rate directly rather than episode counts with an offset for patient-time at risk. Missing data clustered by year rather than hospital, concentrated in the first study year, consistent with reporting ramp-up rather than informative missingness. A sensitivity analysis excluding the center with delayed reporting produced materially consistent estimates. As an observational analysis without patient-level data, the findings establish a temporal association rather than a causal mechanism, and generalizability beyond this single integrated health care system should be hypothesized cautiously. Future work will examine whether the reduction in exit-site infections parallels broader declines in peritonitis rates, and mechanism-specific studies may finally pinpoint which factors are driving these improvements.

Subject of Research: Health

Article Title: Improvement in peritoneal dialysis exit-site infection rates in an integrated health care system in the United States

Article References: Hukill, M. T., Yeung, A. M., & Imam, T. H. (2026). Improvement in peritoneal dialysis exit-site infection rates in an integrated health care system in the United States. Frontiers in Nephrology, 6, Article 1825353. https://doi.org/10.3389/fneph.2026.1825353

Image Credits: AI Generated

DOI: 10.3389/fneph.2026.1825353

Keywords: peritoneal dialysis, exit-site infection, kidney disease, infection prevention, nephrology, Improvement, peritoneal, dialysis, exit-site, infection, rates, integrated