Center-Level Characteristics and Variations in Peritoneal Dialysis Outcomes

One of the most common reasons for discontinuation of peritoneal dialysis therapy is peritonitis. A significant proportion of patients receiving peritoneal dialysis who develop peritonitis experience serious adverse outcomes, including hospitalization, relapsed or recurrent peritonitis, removal of the peritoneal dialysis catheter, permanent transfer to hemodialysis therapy, and/or death.

Studies have documented marked variation in peritonitis rates and technique survival across peritoneal dialysis centers. In addition, there is an association between a substantial proportion of this variation and center-level characteristics. There are few data related to center variation in peritoneal dialysis-related outcomes of peritonitis and the relative contributions of center and patient effect.

Htay Htay, MBBS, MRCP, and colleagues recently conducted a retrospective cohort study designed to examine the associations of key peritonitis outcomes with center-level characteristics after adjustment for patient-level characteristics. A secondary aim was to assess changes in peritonitis outcomes over time. The researchers reported results of the study in the American Journal of Kidney Diseases [2018;71;(6):814-821].

The primary outcome of interest was cure with antibiotic therapy alone, defined as an episode not complicated by relapse or recurrent peritonitis. Secondary outcomes were peritonitis-related catheter removal, transfer to hemodialysis therapy for ³30 days, relapse or recurrent peritonitis, hospitalization, and mortality. Multilevel mixed logistic regression was used to analyze outcomes.

The study population included all incident peritoneal dialysis patients in Australia who developed peritonitis form January 1, 2004, through December 31, 2014, utilizing deidentified data from the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA). Patient-level characteristics examined were age at peritoneal dialysis therapy initiation, sex, race, body mass index (BMI), smoking status, primary kidney disease, comorbid conditions (diabetes, chronic lung disease, and cardiovascular disease), late referral to nephrology care, modality of initial renal replacement therapy, initial peritoneal dialysis modality, and socioeconomic position.

During the study period, 9452 episodes of peritoneal dialysis-related peritonitis occurred in 4624 incident peritoneal patients. Of those patients, 167 were excluded from the analysis, resulting in 9100 episodes of peritonitis in 4428 patients from 51 centers in the final analysis.

Among the final cohort, 59% (n=2591) were male; mean age was 59.6 years; 72% (n=3191) were white; mean BMI was 27.3 kg/m2; 45% (n=2015) were nonsmokers; 45% (n=1993) had diabetes mellitus, 48% (n=2144) had cardiovascular disease, 15% (n=664) had chronic lung disease; primary kidney disease cause was glomerulonephritis in 25% (n=1120), diabetes mellitus in 35% (n=1565), hypertension in 14% (n=626), polycystic kidney disease in 6% (n=239), and other causes in 20% (n=878); 21% (n=921) had late nephrology referral; 79% (n=3358) had continuous ambulatory peritoneal dialysis as initial peritoneal dialysis modality; 14% (n=616) had peritoneal dialysis as initial renal replacement therapy; and the mean IRSAD (Index of Relative Socioeconomic Advantage and Disadvantage) score was 974.

Of the 9100 peritonitis episodes included in the analyses, 69% (n=6285) were cured by antibiotic therapy alone. Rates of peritonitis cure for individual centers varied between 38% and 86%. Independent and significant center-level characteristics associated with achievement of cure were a higher proportion of dialysis patients treated with peritoneal dialysis (>29% peritoneal dialysis patients; adjusted odds ratio [OR], 1.21; 95% confidence interval [CI], 1.04-1.40) and a higher proportion of peritonitis episodes receiving complete empirical antibiotic cover (OR, 1.22; 95% CI, 1.06-1.42). There were no other center-level characteristics significantly associated with peritonitis cure, including center size, transplantation center status, automated peritoneal dialysis exposure, peritoneal equilibration test performance at initiation of peritoneal dialysis therapy, icodextrin exposure, proportion of culture-negative peritonitis, and proportion of patients co-prescribed antifungal prophylaxis with antibiotic treatment.

Patient-level characteristics significantly associated with higher odds of cure with antibiotic therapy were younger age (OR, 0.96; 95% CI, 0.93-1.00) and low-risk causative organisms (OR, 0.96; 95% CI, 0.93-1.00). Following adjustment for all patient-level characteristics, variation in the odds of cure with antibiotic therapy across centers was increased by 9%, but reduced by 66% after adjustment for center-level characteristics, compared with the model adjusted for causative organisms only.

Of the 9100 peritonitis episodes, 1739 required removal of the peritoneal dialysis catheter. Rates of catheter removal at individual centers varied from 12% to 50%. In centers with higher proportions of dialysis patients treated with peritoneal dialysis (>29%), odds of peritonitis-related catheter removal were lower (OR, 0.78; 95% CI, 0.62-0.97). There were no significant associations between center-level characteristics and catheter removal.

There was an association between episodes of peritonitis with organisms categorized as high and moderate risk for catheter removal and higher odds of catheter removal compared with peritonitis episodes from organisms categorized as low risk (OR, 9.64; 95% CI, 8.34-11.1 and OR, 2.63; 95% CI, 2.29-3.02, respectively). Following adjustment for all patient-level characteristics, variation in the odds of catheter removal across centers was increased by 6%, but reduced by 42% following adjustment for center-level characteristics, compared with the model adjusted for causative organisms only.

There were 1667 peritonitis episodes that culminated in patients transferring to hemodialysis therapy for ³30 days. Transfer rates for individual centers varied between 10% and 50%. Centers with higher proportions of patients receiving peritoneal dialysis had lower odds of transfer to hemodialysis therapy (OR, 0.78; 95% CI, 0.62-0.97).

Of the 9100 peritonitis episodes, 12% (n=1126) resulted in relapse or recurrence, with variation between centers between 0% and 23%. Centers with higher (>29% peritoneal dialysis patients; OR, 0.68; 95% CI, 0.48-0.98) and lower proportions of patients receiving peritoneal dialysis (<18% patients receiving peritoneal dialysis; OR, 0.68; 95% CI, 0.51-0.90) had lower risks for relapsed/recurrent peritonitis compared with centers with an average proportion of peritoneal dialysis patients. Smoking status was the only patient-level characteristic significantly associated with the odds of relapsed or recurrent peritonitis.

Nine percent of the patients (n=406/4428) had peritonitis-related mortality. Individual center rates varied between 0% and 25%; there were no significant associations between center-level characteristics and the odds of peritonitis-related mortality. Patient-level characteristics that were significantly and independently associated with peritonitis-related mortality were older age, presence of diabetes mellitus, presence of chronic lung disease, lower socioeconomic status, and causative organism.

Hospital admission was required in 68% of peritonitis episodes (n=6222), with individual center rates varying between 37% and 100%. Lower proportion of peritonitis patients was the main center-level characteristic associated with higher odds of hospital admission. Patient-level characteristics associated with higher odds of hospital admission were lower socioeconomic status and moderate- or high-risk causative organism.

Finally, compared with an earlier study period (2004-2009), the contemporary period (2010-2014) was significantly associated with higher odds of peritonitis cure with antibiotic therapy (OR, 1.17; 95% CI, 1.04-1.30) and lower odds of relapsed or recurrent peritonitis (OR, 0.66; 95% CI, 0.55-0.80).

Limitations to the study cited by the authors included the retrospective design and the limited depth of data available in the ANZDATA database.

“These results suggest that center effects contribute substantially to the appreciable variation in peritoneal dialysis peritonitis outcomes that exist across peritoneal dialysis centers within Australia,” the researchers said.

Takeaway Points

  1. Researchers in Australia conduced a retrospective cohort study to examine whether center-level characteristics influence rates of peritoneal dialysis peritonitis independent of patient-level characteristics.
  2. The primary outcome of interest was cure of peritonitis with antibiotics, Secondary outcomes included catheter removal, transfer to hemodialysis, relapse/recurrence, hospitalization, and death.
  3. Rates of cure with antibiotics varied between 38% and 86% across centers; center-level characteristics had substantial influence on the variations in rates of the primary and secondary outcomes.