Prediction of inpatient survival and graft loss in rehospitalized kidney recipients.

INTRODUCTION

Despite a sizeable amount of research conducted hitherto into predictors of renal transplantation outcomes, there are scarce, data on predictors of in-hospital outcomes of post-kidney transplant rehospitalization. This study sought to provide a user-friendly prediction model for inpatient mortality and graft loss among rehospitalized kidney recipients.

METHOD

This retrospective review of 424 consecutive kidney recipients rehospitalized after kidney transplantation between the years 2000 and 2005 used multiple logistic regression analysis to evaluate predictors of hospitalization outcomes.

RESULTS

Multivariate analysis showed that age at admission, diabetes mellitus as the cause of end-stage renal disease (ESRD), admission due to cerebrovascular accident (CVA), surgical complications were predictors of in-hospital death; age at transplantation, surgical complications, and rejection were predictors of graft loss. Equation for prediction of in-hospital death was Logit(death) -0.304 * age at transplantation (year) + 0.284 age at admission (year) + 1.621 admission for surgical complication + 4.001 admission for CVA-ischemic heart disease + 2.312 diabetes as cause of ESRD. Equation for prediction of in-hospital death was Logit(graft loss) = 0.041 age at transplantation (year) + 1.184 admission for graft rejection + 1.798 admission for surgical complication.

CONCLUSIONS

Our prediction equations, using simple demographic and clinical variables, estimated the probability of inpatient mortality and graft loss among re-hospitalized kidney recipients.