Prediction of Renal Allograft Function at POD 7 With Pre/Peri-Operative Clinical Data.

Abstract

(Background) Serum creatinine (s-Cr) level reaches its nadir around POD 7 after renal transplant. Although it is an important marker of allograft function, systematic formula to predict it is lacking. (Method) We retrospectively analyzed a correlation of recipient's s-Cr with clinical data of a donor and a recipient through symbolic regression. We used 176 pairs of live renal transplantations (LRTx) that showed eventless peri/post-operative courses as a training set to construct a prediction formula of recipient's s-Cr at POD 7. Then we took other 70 pairs of LRTx that also had uncomplicated courses as a future set to validate the efficacy of the formula. (Results) In the course of constructing a formula, body weight, gender, age were significant factors in the prediction. A donor's s-Cr/2hr-Cr clearance, ischemic time and graft weight were also significant. ABO-incompatibility, HLA mismatch, dialysis period, immunosuppression protocol and blood pressure were not. Average error of absolute values between observed s-Cr and predicted one using the training set was 0.18±0.16mg/dl. 61.9% of such errors were <0.2mg/dl and 91.4% were <0.4mg/dl. The same formula was applied to the future set and average error was 0.22±0.19mg/dl while in 78.6% of them the predicted level was higher than the observed one.Figure: No Caption available.Then 11 cases (15.7%) showed prediction errors of >0.4mg/dl and their background revealed significantly lighter body/graft weight of a donor and heavier body weight of a recipient than the rest. (Conclusion) We conclude that it was possible to predict the POD7 s-Cr of a recipient with an error of 0.22±0.19mg/dl as long as the clinical course was uneventful. It was also revealed that s-Cr was not affected so much in uneventful cases by ABO-incompatibility, HLA mismatch, regimen and blood pressure as by age, body weigh, graft weight, donor's renal function and ischemic time. Since in 78.6% of the future cases predicted s-Cr was nearly equal or higher than the observed one, it could be used as an upper threshold of s-Cr for standard renal graft function.