HDD cannot be advocated for all patients.

Sir, Saner and colleagues have made a good case for home haemodialysis (HHD) as an effective renal replacement therapy in Switzerland [1]. Yet, in the last paragraph of their case–cohort study of 58 patients, they admit that patient selection may have influenced their results. We agree. Our analysis of a much larger cohort, that of the United States Renal Data System, used the standardized mortality ratio (SMR) to compare HHD patients with in-centre haemodialysis patients. The SMR takes age, race, gender and diabetes into account. Contrary to older reports, we found that HHD had a higher SMR compared with in-centre dialysis (see Table 1) [2]. In addition, the cost of HHD was not cheap, while somewhat less than in-centre dialysis, which was $54 917 per year over the same time of study. Our data run counter to accepted wisdom about HHD. Increasing patient co-morbidity, decreasing family support and waning of doctor and nurse expertise may explain the inferior HHD outcomes in the USA. It is also likely that in-centre dialysis has improved its dialysis delivery to a greater extent than has HHD in recent years. HHD cannot be advocated for all patients. Even the stillinfrequent newer daily HHD therapies will require scrutiny and proper outcome analysis.