Background: The UK Renal Registry quotes a 1-year death rate for patients established on dialysis of 19.4 per 100 patient years. Clinical experience, reflected in the UK Renal Association Standards Document (RASD), recognizes qualitatively that age, diabetes, and other comorbidities increase the risk of death. The aim of this paper is to provide quantitative estimates of the relative risk of death associated with particular patient characteristics.
Methods: Quantitative techniques were used to estimate relative risk of death in the seven studies quoted in the RASD document and 17 other papers identified in a systematic literature search. Relative risk data from each study were pooled using a fixed effects model (f). A random effects model (r) was applied to pool relative risks if heterogeneity was found to exist between studies. A meta-regression analysis was also carried out to investigate whether study covariates substantially explained the heterogeneity between studies.
Results: Pooling the papers identified in the systematic literature search with those from the RASD gave rise to a relative risk of death of 1.029 (95% CI 1.013-1.045) (r) associated with each year's increase in age. The relative risk associated with the presence of diabetes was 1.91 (95% CI 1.67-2.17) (r), whilst that associated with heart disease was 1.59 (95% CI 1.49-1.69) (f), and with peripheral vascular disease 1.58 (95% CI 1.29-1.93) (r). Heterogeneity was found in the estimates of risk associated with age, diabetes, and peripheral vascular disease. Important study covariates included the use of incident or prevalent cases, the use of routine data sources or data collected specifically for a particular study, the country in which the study was located, the use of a P value to infer the standard error of a relative risk estimate in a particular study, and the method of classifying diabetes.
Conclusions: Published studies can be used to quantify the relative risk of death for dialysis patients with various comorbidities. This information is important if attempts are to be made to set standards for the performance of dialysis units, and to compare the performance of one dialysis unit with that of another.