Hemodialysis versus Peritoneal Dialysis: A Comparison of Survival Outcomes in South-East Asian Patients with End-Stage Renal Disease

Abstract

Background: Studies comparing patient survival of hemodialysis (HD) and peritoneal dialysis (PD) have yielded conflicting results and no such study was from South-East Asia. This study aimed to compare the survival outcomes of patients with end-stage renal disease (ESRD) who started dialysis with HD and PD in Singapore.

Methods: Survival data for a maximum of 5 years from a single-center cohort of 871 ESRD patients starting dialysis with HD (n = 641) or PD (n = 230) from 2005-2010 was analyzed using the flexible Royston-Parmar (RP) model. The model was also applied to a subsample of 225 propensity-score-matched patient pairs and subgroups defined by age, diabetes mellitus, and cardiovascular disease.

Results: After adjusting for the effect of socio-demographic and clinical characteristics, the risk of death was higher in patients initiating dialysis with PD than those initiating dialysis with HD (hazard ratio [HR]: 2.08; 95% confidence interval [CI]: 1.67-2.59; p<0.001), although there was no significant difference in mortality between the two modalities in the first 12 months of treatment. Consistently, in the matched subsample, patients starting PD had a higher risk of death than those starting HD (HR: 1.73, 95% CI: 1.30-2.28, p<0.001). Subgroup analysis showed that PD may be similar to or better than HD in survival outcomes among young patients (≤65 years old) without diabetes or cardiovascular disease.

Conclusion: ESRD patients who initiated dialysis with HD experienced better survival outcomes than those who initiated dialysis with PD in Singapore, although survival outcomes may not differ between the two dialysis modalities in young and healthier patients. These findings are potentially confounded by selection bias, as patients were not randomized to the two dialysis modalities in this cohort study.

Fig 1. Kaplan-Meier survival curves by initial modality from day 90, using the entire cohort (Fig 1a, HD: 641, PD: 230) and the propensity score matched subsample (Fig 1b, HD: 225, PD: 225).

Fig 2. Hazard ratios for risk of death for patients initiating PD compared with those initiating HD and 95% confidence intervals using flexible RP model after adjusting for the effect of socio-demographic and clinical characteristics.

Fig 3. Hazard ratio for risk of death during 5-year follow-up for patients initiating PD compared with those initiating HD and 95% confidence intervals, stratified by age and the presence of DM.

Fig 4. Hazard ratio for risk of death during 5-year follow-up for patients initiating PD compared with those initiating HD and 95% confidence intervals, stratified by age and the presence of CVD.