The peritoneal equilibration test (PET) is well established as a tool for classifying patients as low (L), low average (LA), high average (HA), or high (H) peritoneal transporters. We performed this retrospective 6 year cohort survey to evaluate the impact of different types of PET results on technique survival and patient survival on automated peritoneal dialysis (APD) therapy. From March 1992 to May 1998, 50 patients (20 men, 30 women) receiving APD were enrolled. The mean follow-up period was 25.2 +/- 9.2 months. Basic data and PET results of each patient at the initiation of APD therapy were retrospectively obtained for analysis. Adequacy of dialysis was estimated by measurement of total weekly urea clearance (Kt) normalized to total body water (V) and total weekly creatinine clearance (Ccr) per 1.73 m2 body surface area. The clinical outcomes evaluated were technique survival and patient survival. For statistical analyses we used the Kruskal-Wallis test, Friedman test, Kaplan-Meier life table analysis, and Cox's proportional hazards regression model. There were no differences in age, gender, prevalence of diabetes mellitus (DM), duration of APD, or the initial value of serum albumin between the four subgroups (H, HA, LA, and L). There were 11 (22%) deaths and 8 (16%) technique failures. The 2 year patient survival probability was significantly higher (100%) in the L subgroup than in the LA (62.6%), HA (48.4%), or H (46.2%) subgroups. Patients with DM had a lower patient survival rate than patients without DM; however, there was no statistical significance in technique survival rate between them. Diabetes mellitus (RR = 2.898) and the final albumin value (RR = 0.2099 per increase of 1 gm/dl) had a significant influence on patient survival. By stepwise regression analysis of final serum albumin levels, we found that patients with lower serum albumin values (< or = 3.0 gm/dl vs. >3.0 gm/dl) had a significantly lower probability of patient survival (p = 0.0156). We conclusively demonstrate four important findings in this work: 1) patients with H peritoneal transport had a lower probability of patient survival, but not a decreased rate of technique survival; 2) patients with L peritoneal transport can tolerate APD well; 3) there was no significant difference in technique survival rate between the different PET subgroups; and 4) DM and a lower serum albumin, implicating malnutrition, may contribute to the lower probability of patient survival among H peritoneal transporters.