Background and Objective

1. Table 1. Demographic characteristics of study subjects under maintenance hemodialysis (HD) or peritoneal dialysis (PD), before (A) and after (B) matching with age, gender, and co-morbid diabetes mellitus (DM). Note: HD, hemodialysis; PD, peritoneal dialysis. Comparison of Life Expectancy, Expected Years of Life Lost and Survival between Hemodialysis and Peritoneal Dialysis Patients Tze-Wah Kao1,2, Jenq-Wen Huang2,Yu-Yin Chang1, Pau-Chung Chen1, Tun-Jun Tsai2, Kwan-Dun Wu2, Jung-Der Wang1,21Institute of Occupational Medicine and Industrial Hygiene, National Taiwan University College of Public Health;2Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine Under the policy of cost containment of the National Health Insurance of Taiwan, peritoneal dialysis (PD) has been promoted mainly due to its lower expenditure compared with hemodialysis (HD). PD may also be superior to HD in other aspects. PD preserves residual renal function better and is associated with higher scores of health-related quality of life. When survival is compared, however, the results are still controversial. At present, there is still no formal literature from Taiwan comparing the long-term outcome of HD and PD patients. Moreover, comparison of expected years of life lost (EYLL) between HD and PD patients has not been reported before. Thus, the purpose of this study was to compare the life expectancy, EYLL, and survival between HD and PD patients in Taiwan. Background and Objective Results and Discussion The underlying demographic characteristics of all eligible subjects before matching were summarized in Table 1(A). Before matching, life expectancy of HD patients was significantly shorter (8.8 ± 0.8 years) than that of PD patients (19.9 ± 2.7 years) (p<0.0001), though EYLL between the two groups were similar. Life expectancies of both HD and PD patients were markedly shorter than their age- and gender-matched referents [Table 2(A)]. After matching with age, gender and diabetic status, 236 comparable pairs of HD and PD patients were generated [Table 1(B)]. No statistical difference of life expectancy (p=0.790) or EYLL (p=0.793) was found between the two groups after matching [Table 2(B)]. Table 3 shows the impact of age, gender, DM, and dialysis modality on survival of the dialysis patients. Age (adjusted hazard ratio, AHR=1.07, 95% CI 1.05-1.09) and DM (AHR=3.81,95% CI 2.28-6.36) were independent significant predictors of mortality. For patients having DM, survival was significantly better if the patients were treated with HD rather than PD (AHR=0.24, 95% CI 0.11-0.53).  Material and Methods All adult patients who underwent maintenance dialysis at the National Taiwan University Hospital from 1995 to 2006 were included. All were followed-up until the end of 2007 and linked with the National mortality database to verify whether the patients were still alive. Baseline characteristics between HD and PD patients were compared using independent t-test and chi-square test. The follow-up data were analyzed by the Kaplan–Meier method to yield the estimated survival function of both groups. A constant excess hazard model was used to project the long-term survival of these patients by utilizing linear extrapolation of a logit-transformed curve of survival ratio between dialysis patients and their age- and gender-matched reference populations. Life expectancy was estimated by extrapolating the survival curves to 50 years after initiation of dialysis, while EYLL was calculated by subtracting the life expectancies of the patients from those of their referents. HD patients were then matched with PD patients on age, gender, and diabetic status. Life expectancy, EYLL, and survival for the two groups were re-compared. The Cox Model was also applied to determine the risk factors of mortality. Table 2. Comparison of life expectancy and expected years of life lost (EYLL) between patients under maintenance hemodialysis (HD) and patients under peritoneal dialysis (PD), before (A) and after (B) matching. Table 1. Demographic characteristics of study subjects under hemodialysis (HD) or peritoneal dialysis (PD), before (A) and after (B) matching with age, gender, and diabetes mellitus (DM). Table 3. Impact of age, gender, diabetes mellitus (DM), and dialysis modality on survival of our dialysis patients. Corresponding author’s e-mail: jdwang@ntu.edu.tw; First author’s e-mail: twkao2@ntuh.gov.tw OMIH