Background. Comparisons of mortality outcomes between peritoneal dialysis (PD) and hemodialysis (HD) patients have shown varying results, which may be caused by the unequally distributed clinical conditions of patients at initiation. To address this issue, we evaluated the clinical characteristics of 105,954 patients at the initiation of PD and HD, using the U.S. national incidence data on treated end-stage renal disease from the Medical Evidence Form, 1995 to 1997. Methods. A general linear model was used to analyze differences of age, albumin, creatinine, blood urea nitrogen (BUN), and hematocrit; categorical data analysis to evaluate body mass index (BMI), grouped into four categories: <19, 19-25 (<25), 25-30 (<30), and 30+; and logistic regression to assess the likelihood of initiating PD versus HD. Diabetics (DM) were analyzed separately from non-diabetics (NDM). Explanatory variables in the logistic regression included incidence year, race, gender, age, BMI, albumin, creatinine, BUN, and hematocrit. Race included white and black. Age was categorized into four groups: 20-44, 45-64, 65-74, and 75+. Results. At the initiation of dialysis PD patients were approximately 6 years younger (P < 0.0001) than HD patients. PD patients also had higher (P < 0.0001) albumin (+0.35 g/dL for DM and +0.23 g/dL for NDM) and hematocrit (+1.64% for DM and +1.71% for NDM) levels, and lower (P < 0.04) BUN (-8.75 mg/dL for DM and -5.24 mg/dL for NDM) and creatinine (-0.51 mg/dL for DM and -0.23 mg/dL for NDM) levels than HD patients. Whites had a higher (P < 0.0001) likelihood of starting PD than blacks, and patients with BMI <19 had a lower (P < 0.0001) chance of beginning on PD. Conclusion. PD patients had favorable clinical conditions at the initiation of dialysis, which should be taken into consideration when comparing dialysis outcomes between the two modalities.
Bibliographical noteFunding Information:
This project has been funded in whole or in part with federal funds from the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, under Contract No. N01-DK-9-2343. The authors wish to thank Mr. Tom Arnold, Mr. Mike Hadad, and Mr. Roger Milam from the Health Care Financing Administration for their assistance with the HCFA data. We also thank Ms. Beth Forrest for her help in manuscript preparation and regulatory assistance. Dataset construction was provided by Roger Johnson and C. Daniel Sheets.
- Data analysis
- Dialysis modality
- End-stage renal disease
- Health Care Financing Administration
- Medical Evidence Form
- Mortality outcome