Abstract: TH-PO912

Incidence of Dialysis-Related Infections in the Automated Peritoneal Dialysis Population in the United States

Session Information

  • Dialysis: Infection
    November 02, 2017 | Location: Hall H, Morial Convention Center
    Abstract Time: 10:00 AM - 10:00 AM

Category: Dialysis

  • 610 Dialysis: Infection

Authors

  • Weinhandl, Eric D., NxStage Medical, Inc., Victoria, Minnesota, United States
  • Collins, Allan J., NxStage Medical, Inc., Victoria, Minnesota, United States
Background

Infection increases risks of peritoneal dialysis technique failure and death. However, data about the magnitude of and trends in the incidence of all-setting peritoneal dialysis-related infection (DI) in the US are lacking. We estimated incidence of both hospitalized and non-hospitalized DI in the automated peritoneal dialysis (APD) population.

Methods

We analyzed United States Renal Data System records. From 2006 to 2013, we collected cohorts of end-stage renal disease patients on APD. We identified hospitalized and non-hospitalized cases of DI from Medicare claims with ICD-9-CM diagnosis codes [a] 567.x (peritonitis), [b] 996.68 (infection due to peritoneal dialysis catheter), and [c] 038.x (septicemia). We estimated incidence rates with definitions of diagnosis code [a], codes [a]-[b], and codes [a]-[c], and with only principal or with both principal and secondary diagnosis codes.

Results

In 2013, the incidence rate was 11.2 (annualized rate of change between 2006 and 2013, -6.3%), 25.1 (-4.7%), and 35.1 (-3.1%) events per 100 patient-years with diagnosis code [a], codes [a]-[b], and codes [a]-[c], respectively, as a function of only principal diagnosis codes. In contrast, as a function of both principal and secondary diagnosis codes, corresponding rates were 45.7 (-3.3%), 55.1 (-2.7%), and 66.3 (-2.1%) events per 100 patient-years. These rates corresponded to one DI case per 26, 22, and 18 patient-months, respectively. With DI defined by both principal and secondary diagnosis codes, hospitalized cases comprised between 53% and 61% of all cases.

Conclusion

Between 2006 and 2013, the incidence of DI decreased among APD patients in the US. However, the absolute magnitude of the incidence of DI was uncertain, as rates defined by an array of diagnosis code sets and diagnosis code positions varied by a factor of nearly 6 in 2013. The incidence rate associated with broader claims-based definitions indicated that DI remains a common complication in 2013; for frame of reference, the incidence of all-setting peritonitis ranged from 35 to 40 events per 100 patient-years in Australia between 2012 and 2015. In addition, more than half of DI cases involved hospitalization, a possible marker of inadequate monitoring in the outpatient setting. With continued growth of PD utilization in the US, novel tools to reduce risk of DI on APD are needed.