The goal of the Dialysis Outcomes in Colombia (DOC) study was to compare the survival of patients on hemodialysis (HD) vs peritoneal dialysis (PD) in a network of renal units in Colombia. The DOC study examined a historical cohort of incident patients starting dialysis therapy between 1 January 2001 and 1 December 2003 and followed until 1 December 2005, measuring demographic, socioeconomic, and clinical variables. Only patients older than 18 years were included. As-treated and intention-to-treat statistical analyses were performed using the Kaplan-Meier method and Cox proportional hazard model. There were 1094 eligible patients in total and 923 were actually enrolled: 47.3% started HD therapy and 52.7% started PD therapy. Of the patients studied, 751 (81.3%) remained in their initial therapy until the end of the follow-up period, death, or censorship. Age, sex, weight, height, body mass index, creatinine, calcium, and Subjective Global Assessment (SGA) variables did not show statistically significant differences between the two treatment groups. Diabetes, socioeconomic level, educational level, phosphorus, Charlson Co-morbidity Index, and cardiovascular history did show a difference, and were less favorable for patients on PD. Residual renal function was greater for PD patients. Also, there were differences in the median survival time between groups: 27.2 months for PD vs 23.1 months for HD (Pandlt;0.001) by the intention-to-treat approach; and 24.5 months for PD vs 16.7 months for HD (Pandlt;0.001) by the as-treated approach. When performing univariate Cox analyses using the intention-to-treat approach, associations were with age ≥65 years (hazard ratio (HR)=2.21; confidence interval (CI) 95% (1.77-2.755); Pandlt;0.001); history of cardiovascular disease (HR=1.96; CI 95% (1.58-2.90); Pandlt;0.001); diabetes (HR=2.34; CI 95% (1.88-2.90); Pandlt;0.001); and SGA (mild or moderate-severe malnutrition) (HR=1.47; CI 95% (1.17-1.79); P=0.001); but no association was found with gender (HR=1.03, CI 95% 0.83-1.27; P=0.786). Similar results were found with the as-treated approach, with additional associations found with Charlson Index (0-2) (HR=0.29; Cl 95% (0.22-0.38); Pandlt;0.001); Charlson Index (3-4) (HR=0.61; Cl 95% (0.48-0.79); Pandlt;0.001); and SGA (mild-severe malnutrition) (HR=1.43; Cl 95% (1.15-1.77); Pandlt;0.001). Similarly, the multivariate Cox model was run with the variables that had shown association in previous analyses, and it was found that the variables explaining the survival of patients with end-stage renal disease in our study were age, SGA, Charlson Comorbidity Index 5 and above, diabetes, healthcare regimes I and II, and socioeconomic level 2. The results of Cox proportional risk model in both the as-treated and intention-to-treat analyses showed that there were no statistically significant differences in survival of PD and HD patients: intention-to-treat HD/PD (HR 1.127; CI 95%: 0.855-1.484) and as-treated HD/PD (HR 1.231; CI 95%: 0.976-1.553). In this historical cohort of incident patients, there was a trend, although not statistically significant, for a higher (12.7%) adjusted mortality risk associated with HD when compared to PD, even though the PD patients were poorer, were more likely to be diabetic, and had higher co-morbidity scores than the HD patients. The variables that most influenced survival were age, diabetes, comorbidity, healthcare regime, socioeconomic level, nutrition, and education.
