P1182PROGNOSTIC VALUE OF CHARLSON COMORBIDITY INDEX IN PERITONEAL DIALYSIS PATIENTS

Abstract

Abstract Background and Aims Mortality risk is very high in the dialysis population and the Charlson Comorbidities Index (CCI) is considered as an useful risk stratification tool in these patients. The purpose of this study was to examine the accuracy of CCI for predicting mortality in peritoneal dialysis (PD) patients and to compare the prognostic power of CCI with that of a risk prediction model jointly including a clinical score of malnutrition and the NYHA classes. Method We analyzed this problem in a series of 66 consecutive PD patients on follow-up in our unit. Their mean age was 69±14 years, 64% were male, 36% were diabetic. The median dialysis vintage was 39 months (Interquartile range 23-67 months). Results During follow-up period, 37 patients died. On univariate Cox regression analysis, CCI largely failed to predict mortality [Hazard ratio (HR): 1.08, 95% CI: 0.94-1.24, P=0.30) and this relationship did not improve (HR: 0.93, 95% CI: 0.79-1.08, P=0.32) also after data adjustment for malnutrition and NYHA classes. In the same model, malnutrition (HR: 1.98, 95% CI 1.20-3.27, P=0.007) and NYHA classes (HR: 3.15, 95% CI 1.67-5.94, P<0.001) were strongly and significantly related to the risk of death. Of note, the prognostic accuracy of the model based on malnutrition and NYHA classes (ROC curve area: 74%) was higher than that provided by CCI alone (ROC curve area: 54%) and did not materially differ from that of an expanded model including CCI, malnutrition and NYHA classes (ROC curve area: 78%) Conclusion In this study, CCI did not predict survival in PD patients whereas malnutrition and NYHA score displayed a relevant prognostic accuracy for death in the same patient-population. These results generate the hypothesis, to be confirmed in a larger PD population, that CCI solely for risk stratification is unwarranted in PD patients.