reported that the mortality risk was significantly higher for PD compared to HD for both males and females but the risk was accentuated for females (RR 1.30 for females vs RR 1.11 for males).36 One reported an interaction between dialysis modality and gender for patients with ischemic heart disease or peripheral vascular disease with the survival benefit of PD only observed for male patients. 44 Two studies reported that the interaction was not significant. 32,45 31 Home-based versus In-center Dialysis Evidence-based Synthesis Program Race. Interactions of modality and race were assessed in 5 studies.11,32,36,38,54 Two reported no effect of race. 36,38 Another reported a mortality benefit for PD in white patients with BMI greater than 30 but not non-white patients. 54 The significance was not reported. One study reported an interaction effect that was significant for Asian and other categories (relative to white)32 while the fifth study reported different patterns of risk over time (less than 3 years from inception of dialysis vs more than 3 years) for different ethnicity groups. 11 BMI. Five studies assessed interactions with BMI. 11,31,32,38,54 The most recent study reported no significant interaction with between modality and BMI. 11 One study found that a BMI of 30 or higher was associated with improved survival for HD patients (HR 0.89) but not PD patients (HR 0.99).31 Another study found significantly increased mortality risk for PD in the 3 highest BMI groups (BMI of 23.5 or higher) in patients with diabetes, while for patients with no diabetes the mortality risk was significantly higher only in the highest BMI group (BMI greater than 30).54 One study reported a non-significant interaction of BMI and mortality between 90 and 365 days of treatment but a significant interaction after 365 days.38 The effect sizes were clinically similar across all BMI categories, however. The fourth study found significant interactions between treatment modality and overweight (BMI 25.1 to 30) and obese (BMI greater than 30) but not underweight (BMI less than 18.5).32 Diabetes. Interactions between modality and diabetes were assessed in 12 reports from 10 datasets. 12,26,29,32,33,36,38,42,43,45,54,55 The interaction was not significant in 4 studies26,38,42,45 while 5 studies reported higher mortality risk (PD vs HD) in patients with diabetes. 12,29,32,36,43 Another study reported that across levels of BMI, patients with diabetes tended to have increased risk of mortality with PD compared to patients without diabetes. 54 Additional analyses of this data focused on patients with coronary artery disease55 or congestive heart failure. 33 Patients with CAD or CHF and diabetes had higher mortality with PD compared to patients without diabetes. Cardiovascular Disease. Six reports (2 from the same dataset) reported on interactions between cardiovascular disease and modality. 11,29,32,33,46,55 As noted above, 2 analyses from one study54 focused on coronary artery disease (CAD)55 and congestive heart failure (CHF). 33 For patients with diabetes, the mortality risk was greater for PD regardless of CAD status55 or CHF status.33 For patients without diabetes, mortality was elevated in the CAD group but not the no-CAD group55 and in the CHF group but not the no-CHF group. 33 Significant interactions with cardiovascular disease were reported in 2 other US studies29,32 while a French study and a New Zealand study reported non-significant interactions.11,46 Duration of ESRD Therapy. One study reported on duration of ESRD therapy finding a significant statistical interaction between ESRD vintage and mortality risk. It was noted that there was little clinical significance. 38 Another study reported a non-significant interaction. 11 Other Outcomes Cardiovascular Events (Appendix C, Table 1) Five registry studies reported on cardiovascular events.12,35,46-48 One study from Italy focused on the development of de novo cardiovascular disease. 47 At baseline, there were no significant differences between the HD and PD groups in the percentages of patients with a history of coronary heart disease, myocardial infarction, or chronic heart failure. During the study period (maximum follow-up of 4 years), 11.4% of the deaths in the PD group and 21.1% of the deaths 32 Home-based versus In-center Dialysis Evidence-based Synthesis Program in the HD group were due to cardiac causes although it was noted that these numbers did not take into consideration patients who switched dialysis modalities. The relative risk of developing de novo cardiovascular disease (PD vs HD) was 1.06 [95% CI 0.79, 1.43]). Similar risks were reported for ischemic heart disease and congestive heart failure. Three studies, one from Australia/New Zealand, 12 one from Romania,48 and one from France,46 reported cardiovascular mortality. In the first study, 54% of deaths in the PD group and 47% of deaths in the HD group were due to cardiovascular causes (significance not reported). 12 In the second study, the difference in cardiovascular mortality was not statistically significant (PD 47%, HD 49%, P = .70)48 while in the third study, a significant difference was reported (PD 40%, HD 35%, P = .04).46 A study from the US reported risk of cardiovascular mortality (PD vs HD) for patients age 55 and older.35 In patients with diabetes, both males and females receiving PD had a reduced risk of