Comparison of the Charlson comorbidity index, the modified Charlson comorbidity index, and the recipient risk score in prediction of the graft and patient survival among renal graft recipients: historical cohort in a single center.

OBJECTIVE

To compare the predictive values of Charlson comorbidity index (CCI), modified Charlson comorbidity index kidney transplant (mCCI-KT) and recipient risk score (RRS) indices in prediction of patient and graft survival in kidney transplant patients.

METHODS

In this retrospective study, all patients who underwent a live-donor KT from 2006 to 2010, were included. Demographic data, comorbidities and survival time after KT were extracted and the association between above indices with patient and graft survival were compared.

RESULTS

In ROC curve analysis of 715 included patients, all three indicators were weak in predicting graft rejection with the area under curve (AUC) less than 0.6. The best models for predicting the overall survival were mCCI-KT and CCI with AUC of 0.827 and 0.780, respectively. Sensitivity and specificity of mCCI-KT at cut point of 1 were 87.2 and 75.6. Sensitivity and specificity of CCI at cut point of 3 were 84.6 and 68.3 and for RRS at cut point of 3 were 51.3 and 81.2, respectively.

CONCLUSION

The mCCI-KT index followed by the CCI index provided the best model in predicting the 10-year patient survival; however, they were poor in predicting graft survival and this model can be used for better stratifying transplant candidates prior to surgery.