Dialysis and mortality: does it matter where you live?

A 70-year-old man with diabetic nephropathy and a history of multiple admissions for congestive heart failure is approaching the need for renal replacement therapy. He is struggling with his decision regarding the optimal dialysis modality. He lives alone in rural Washington, 60 miles from the nearest dialysis unit. He is a man of modest means and cannot afford the transportation costs associated with thrice weekly incenter hemodialysis. Given his less than optimal prior compliance with his complex medical regimen, he is a risky candidate for peritoneal or home hemodialysis. Due to transportation issues, he decides that peritoneal dialysis is his only viable option. Does geography impact on patient survival? Clinicians usually do not consider a patient’s residence when making clinical decisions, but epidemiologic evidence has shown that patient characteristics, health services, and patient outcomes vary significantly according to rural versus urban residence. These geographic differences also exist within urban areas depending on residential characteristics. Rural America represents approximately 20% of the US population and is comprised of diverse cultures, landscapes, and economic and social characteristics. In 2005 the Institute ofMedicine report,Quality through Collaboration: The Future of Rural Health Care, was releasedwith the goal of providing an independent assessment of health and health care in rural America and formulating an action plan for quality-focused rural community health systems (1). The report highlighted a number of important issues in rural America including reduced access to quality health services, poor access to quality emergency services, and adverse patient characteristics such as higher rates of obesity, suicide, alcohol and tobacco use, poverty, and inferior treatment for acute myocardial infarction. These factors are of importance to rural patientswithCKDand could represent major barriers to quality renal and general medical care. In this issue of theCJASN,Maripuri et al. (2) present a retrospective cohort study showing that rural residence is associated with higher mortality in those rural patients receiving peritoneal dialysis. This study is a nice extension to our previous work, which had not stratified survival results by dialysis modality and therefore had not shown an increase in mortality in rural patients on peritoneal dialysis. Our study did hint at a problem by showing that facilities in rural areas were less likely to offer peritoneal dialysis than those units in urban areas (3). Other studies in the United States and Canada have also have also shown an association between poor patient outcomes and rural residence (4,5). When reading this study, it is important to note the difficulty in comparing survival between peritoneal and hemodialysis patients. Not only do the therapies differ tremendously, but the patients themselves are quite different, with peritoneal dialysis patients tending to be youngerwith less comorbidities. The findings of numerous studies comparing survival between the modalities have not been consistent (6). In an attempt to better match patients by important baseline characteristics, a recent study compared survival between the modalities by using a matched-pair cohort with the matching based on propensity of initial peritoneal dialysis (7). The study found that survival was modestly longer for patients initiating renal replacement therapy with peritoneal dialysis compared with matched hemodialysis patients. However, similar to other studies, they found this modality survival effect is modified by age and the presence of diabetes. In general, peritoneal dialysis patients .65 years of age and those with diabetes have a higher risk of death than those on hemodialysis. Adding to the complexity of comparing dialysis modalities, the rural population tends to be older with more comorbidities, and this is true in this study by Maripuri et al., in which the rural and micropolitan population is older and all medical comorbid conditions with the exception for alcohol and drug use aremore prevalent in the non-urban population. Therefore, despite controlling for the baseline characteristics, the results of the study could be confounded by the unmeasured differences and/or the differences in severity of the comorbid conditions in the rural population. Despite this possibility, the results of the study are important and should not be ignored by clinicians. The results of this study are consistent with other studies and call attention to the fact that many rural patients are high-risk candidates for peritoneal dialysis on the basis of age and comorbid conditions. It is also important to note that there is no accepted definition of rural and that the level of isolation varies dramatically among the rural population in the United States. As clinicians, we worry about patients who live University of Washington, Seattle and VA Puget Sound Healthcare System, Seattle, Washington