Background: Most comparisons of hemodialysis (HD) and peritoneal dialysis (PD) have used mortality as an outcome. Relatively few studies have directly compared the hospitalization rates, an outcome of perhaps equal importance, of patients using these different dialysis modalities.
Methods: Eight hundred twenty-two consecutive patients at 11 Canadian institutions with irreversible renal failure had an extensive assessment of comorbid illness and initial mode of dialysis collected prospectively immediately prior to starting dialysis therapy. The cohort was assembled between March 1993 and November 1994. The mean follow-up was 24 months. Admission data were used to compare hospitalization rates in HD and PD.
Results: Thirty-four percent of patients at baseline and 50% at three months used PD. Twenty-five percent of HD and 32% of PD patients switched dialysis modality at least once after their first treatment (P = NS). Nine percent of HD patients and 30% of PD patients switched modality after three months (P < 0. 001). Total comorbidity was higher in HD patients at baseline (P < 0. 001) and at three months (P = 0.001). The overall hospitalization rate was 40.2 days per 1000 patient days after baseline and 38.0 days per 1000 patient days after three months. When an adjustment was made for baseline comorbid conditions, patients on PD had a lower rate of hospitalization in intention-to-treat analysis according to the type of dialysis in use at baseline (RR 0.85, 95% CI, 0.82 to 0.87, P < 0.001), but a higher rate according to the type of dialysis in use three months after study entry (RR 1.31, 95% CI, 1.27 to 1.34, P < 0.001). In analyses based on the amount of time actually spent on each treatment modality, PD was associated with a higher rate of hospitalization when analyzed according to the type of dialysis in use at baseline (RR 1.10, 95% CI, 1.07 to 1.13, P < 0.001) and according to the type of dialysis in use three months after study entry (RR 1.26, 95% CI, 1.23 to 1.30, P < 0.001).
Conclusions: Conclusions regarding comparative hospitalization rates are heavily dependent on the analytic starting point and on whether intention-to-treat or treatment-received analyses are used. When early treatment switches are accounted for, HD is associated with a lower rate of hospitalization than PD, but the effect is modest.