Peritonitis After Dialysis Falling, Captured by Claims-Based Approach

Pam Harrison

September 22, 2022

Rates of peritonitis associated with dialysis have dropped substantially over time, although there are still consistent differences in risk by race and age, a new retrospective cohort study indicates.

The work involved a claims-based approach using Medicare data, providing "a wider population perspective" than the usual approach to peritonitis research, which normally employs clinical data from a relatively small patient cohort, lead author Eric Young, MD, Arbor Research Collaborative for Health, Ann Arbor, Michigan, told Medscape Medial News in an email.

The method also led to the conclusion that coding of peritonitis across different institutions and healthcare systems is inconsistent, which means that the full extent of peritoneal dialysis-associated peritonitis is not appreciated.

"In summary, we find that the claims-based approach offers a promising framework for the study of PD-associated peritonitis," Young and colleagues conclude in their article, published in the American Journal of Kidney Disease.

Coding of Peritonitis Needs to Be Standardized

United States Renal Data System (USRDS) files were used to identify claims, eligibility, modality, and demographic information for the study. The sample consisted of patients who received peritoneal dialysis (PD) between 2013 and 2017 and were covered by Medicare fee-for-service (FFS) insurance, with paid claims for dialysis or hospital services. The traditional Medicare FFS program covers the largest pool of PD patients, as the authors point out.

From that claims database, investigators identified 70,271 peritonitis episodes from 396,289 peritonitis claims. "Peritonitis rates were calculated as peritonitis episodes derived by PD patient-years," the investigators note. The estimated peritonitis rate using their own reference rules was 0.54 episodes per patient-year (EPPY) over the 5 study years, but this rate declined each year by an average of 5.3% annually.

Including only those episodes with claims from a dialysis facility or a nephrologist resulted in a peritonitis rate of 0.35 EPPY, which is 35% lower than the reference rate, the authors note.

In unadjusted analyses, investigators found that peritonitis rates were lower among older patients; higher in Black patients and lower in Asian compared to White patients; lower again in Hispanic patients but were higher in patients with longer end-stage kidney disease (ESKD) duration. They were also higher in patients who had diabetes as a cause of their ESKD.

"We also found that each increment in the number of prior peritonitis episodes predicted a substantially higher risk of peritonitis," they add. Indeed, the ESKD vintage effect was reversed when prior peritonitis episodes were added to the model.

As Young elaborated, there are multiple diagnostic codes that physicians can use to indicate a diagnosis of peritonitis.

Indeed, in their study, "there was considerable variability among providers in the utilization of codes used to indicate peritonitis," he stressed, and there were concerns about the accuracy and specificity of some of the codes in use.

"We felt that the available codes do not fully capture or classify certain aspects of PD-associated peritonitis," Young added, "and standardization of peritonitis coding would assist in understanding the epidemiology of PD-associated peritonitis."

For example, the catheter code category presents a major challenge as these codes do not clearly distinguish between infection and non-infectious inflammation or between peritonitis from a catheter-related event, such as exit site and tunnel infections. Moreover, a subset of dialysis facilities and nephrologists preferentially use catheter codes; consequently, "the peritonitis rate may be overstated when catheter codes are included and understated when they are not," the authors caution.

The authors thus recommend that CMS or professional dialysis stakeholder groups address the apparent lack of authoritative coding guidance and training. The fact that the available ICD-10 codes do not optimally capture and differentiate various PD-related infection syndromes could also be addressed in future ICD releases, they suggest.

In summary, the claims-based "approach yields plausible rates and reveals potentially important risk factors but our findings also highlight the need for uniform coding standards and modernized diagnostic coding options," they conclude.

The study was supported by a grant to the Arbor Research Collaborative for Health, of which Young is an employee.                                                                 

Am J Kidney Dis. Published online September 5, 2022. Abstract

Follow Medscape on Facebook, Twitter, Instagram, and YouTube.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.