cardiac death (RR 0.90 for both) relative to males age 55 and older receiving in-center HD. In patients without diabetes, the pattern of results was similar with relative risk of 0.70. Hospitalization (Appendix C, Table 4) None of the registry studies reported hospitalization. A NECOSAD publication reported that 46% of PD patients and 58% of HD patients were hospitalized at least once over a follow-up period that ranged from 5 months to 7.8 years.52 A longitudinal study from the US with 181 incident patients (119 HD, 62 PD) reported higher total admissions per year at risk in the HD group (2.4) compared to the PD group (1.4) (P < .0001).56 Admissions for infection per year at risk were higher for PD patients (0.42) than HD patients (0.29) (P = .02).56 A second US study with 177 patients also reported more hospitalizations (1.5 vs 0.4, P < .01) and more hospital days (12.2 vs 2.4, P < .05) over follow-up of up to 15 months in HD patients compared to PD patients. 57 In a UK study of patients who started dialysis at age 70 years or older, hospitalization did not differ between HD (2.0 events/1 patient-year) and PD (1.9 events/1 patient-year) (RRPD vs HD 0.97 [95% CI 0.77, 1.22]).58 A cross-sectional study from Canada reported no difference in mean hospitalizations in the past year for HD (1.68) and PD (1.43) patients.24 Quality of Life, Cognition, Depression (Table 3, Appendix C, Table 4) None of the registry studies reported quality of life or related outcomes. We identified a 2011 systematic review that included published and grey literature studies (English language only) through July 2010, enrolling adults on either in-center HD or PD, and using a validated tool to assess and compare quality of life for HD and PD patients. 59 Outcomes for both generic (ie, broad aspects of quality of life, suitable for different locations and different cultures such as the SF-36) and disease-specific quality of life tools were reported. Twenty-six studies from the US, Europe/UK, and the Asia/Pacific region were included. Twenty were cross-sectional studies, 4 were cohort studies, and 2 were retrospective analyses. Of the 12 studies that used the SF-36, only 4 reported physical and mental health component summary scores. A significant difference, with better health in the PD group, was observed for both scores in one cross-sectional study from Turkey enrolling 115 patients. This study also reported significant differences, favoring PD, for the 8 individual dimensions of the SF-36. One other study, a cross-sectional study from China with 1,062 participants, reported significant differences, favoring PD, for 6 of the 8 individual dimensions (bodily pain, general health 33 Home-based versus In-center Dialysis Evidence-based Synthesis Program perception, vitality, social functioning, role limitations due to emotional functioning, and mental health). Two additional cross-sectional studies, one from the Netherlands with 1,553 participants and one from Taiwan with 244 participants, reported significant differences favoring PD for the bodily pain and role limitations due to emotional functioning dimensions. Of 5 studies reporting kidney disease-specific quality of life with the 11-item KDQOL instrument, significant differences favoring PD were found for 4 of 11 dimensions in a cross-sectional study from Denmark (N=130), 4 of 9 dimensions assessed in a cohort study from France (N=387), and 3 of 3 dimensions assessed in a cross-sectional study from the US (N=226). Other quality of life tools were used in only one or 2 studies and generally no differences were observed between HD and PD patients. This review was of average quality based on the AMSTAR criteria. 6 Although it was reported that study quality was assessed, the quality ratings were not provided nor used in developing the conclusions for the review. Eleven studies were excluded from the analysis because of either weak design or irrelevance to the topic but no additional information was provided. Little information was provided about the study populations of the included studies and the timing of the quality of life assessment was not reported. Results were provided for only some of the studies reported to have used a particular quality of life assessment tool and little information was provided about tools other than the SF-36 and KDQOL. We supplemented the information from the Boateng and East review59 with more complete data from the 3 US studies included in the review (Table 3) and with data from studies identified in our search of MEDLINE and other sources but not included in the review (Appendix C, Table 4). The 3 US studies, all rated as high risk of bias, found few differences between HD and PD patients in overall measures of physical or mental function with mixed results for individual dimensions (Table 3).60-62 Among the studies not included in the review, the small RCT (n=38) from the Netherlands found no significant difference in the quality adjusted life year scores for the PD and HD groups (54 vs 59; adjusted difference 3.1 [95% CI -9.9, 16.1], P = .63).50 Among 949 patients from the CHOICE study, higher overall functional support (assessed with the Medical Outcomes Study Social Support Survey) was higher for the PD patients (81 vs 76, P = .002).63 Higher scores were reported for the emotional support, tangible support, and positive social interaction domains but not for the affectionate support