A Mediterranean age-comorbidity prognostic index for survival in dialysis populations.

BACKGROUND: The most widely used prognostic indices for estimation of survival, including for dialysis patients, were described by Charlson, and an adaptation was proposed by Hemmelgarn for dialysis patients. We present the first age-comorbidity prognostic index (ACPI) designed in a Mediterranean incident dialysis population and examine its concordance with other prognostic indices. METHODS: Incident dialysis patients were scored in relation to age and 11 diseases. Cox regression analysis was performed to construct multiple regression models, and diseases with a hazard ratio (HR) higher than 1.2 were included in the index. The impact of age was assessed by including it in a separate multivariate model. Scores were categorized in 3 levels of risk: low (0-1 points), medium (2-4 points) and high levels (5 or more points). The probability of survival of each group was calculated according to the Kaplan-Meier method, and receiver operating characteristic (ROC) curves were plotted to examine the concordance with other prognostic indices. RESULTS: A cohort of 304 patients on hemodialysis (80%) and peritoneal dialysis was analyzed. Global mortality rate was 31% (93/304). The mean score was 4.41 +/- 2.84. Diseases that received the highest scores were ischemic heart disease (IHD) with chronic heart failure (CHF), and malignancies of less than 5 years of evolution. With regard to age, the maximum score was received by patients over 60 years old. The probability of survival at 3 years was 89%, 77% and 54% for low-, medium- and high-risk groups, respectively (log-rank test, 19; p=0.0001). The ROC curves showed similar areas for our index (0.749), the Charlson index (0.758) and Hemmelgarn index (0.708), but our index scored higher than Charlson in older patients, IHD with CHF, CHF, peripheral vascular disease and systemic diseases. CONCLUSIONS: Although prospective external validation of this new index is required, this index adequately estimates the probability of survival at 3 years. The prognostic power of ACPI is similar to that of the Charlson index; however, relevant differences were found, concerning the weight of factors age, cardiovascular diseases and myocardial dysfunction. In end-stage renal disease we recommend estimating survival by indices established in incident dialysis patients, due to the particular comorbid conditions of this population.