A Predictive Model for Hyperglycemia after Liver Transplant
Presentation Number: SUN 595
Date of Presentation: April 2nd, 2017
Henry Jorge Zelada*1, Lisa VanWagner1, Teresa Pollack1, Devan Higginbotham1, Lihui Zhao1, Mark E Molitch2 and Amisha Wallia2
1Northwestern University Feinberg School of Medicine, 2Northwestern University Feinberg School of Medicine, Chicago, IL
BACKGROUND: Poor post-liver transplant glycemic control has been associated with an increased risk of transplant rejection, higher infection rates and increased length of stay. The goal of this study was to develop a predictive model to identify patients at risk for hyperglycemia 1 month post-liver transplantation.
METHODS: Adults (n=164) who underwent liver transplantation from April 2009 to December 2014 at Northwestern Medicine were included. Hyperglycemia was defined as a glucose measurement of >200mg/dl post-discharge up to 1 month following liver transplant surgery. Donor and recipient factors evaluated for model inclusion included: age, sex, race, BMI, diabetes status before transplant, model for end stage liver disease (MELD) score, transplant etiology, length of hospital stay, diabetes medications at discharge, Donor Risk Index, donor organ quality, cold ischemic-time, transplant network location and principal cause of donor death. All covariates significant a p-value <0.05 were entered into a forward selection logistic regression analysis. The final model chosen was based on additive contribution to the model based on the Bayesian-Information-Criteria (BIC).
RESULTS: Liver transplant recipients (LTR) were predominantly male (64.6%) and Caucasian (73.1%) and 29.9% had pre-transplant type 2 diabetes (T2D). Mean age (±SD) was 57.5±7.8 years and the median (interquartile-range) MELD score was 31(27.5-35). There were 84 (51.2%) patients who had at least one episode of Glucose >200 mg/dl. On univariate analysis, LTR age, sex, diabetes status before transplant, MELD score, hospital length of stay, T2D medication at discharge and donor sex and race were associated with post-LT hyperglycemia (P<0.05 for all). In multivariate analysis, LTR age (OR 1.06, 95%CI:1.01-1.11, p=0.02), female sex (OR 0.34, 95%CI:0.14-0.85, p=0.02), length of hospital-stay (OR 0.87, 95%CI:0.82-0.93, p<0.001), T2D medications at discharge (supplement insulin only OR 1.65 95%CI:0.55-4.93 p<0.001, basal insulin [(basal insulin only, basal-bolus+SSI, basal-basal bolus, basal-oral]) OR 54.93 95%CI:9.54-316.05 p<0.001, oral treatment OR 10.7 95%CI:1.76-64.88. p<0.001) and donor female sex (OR 4.97, 95%CI:2.04-12.08, p<0.001) were independently associated with post-discharge hyperglycemia. Receiver-operating characteristic (ROC) analysis showed that the area under the curve (AUC) for the model was 0.86.
CONCLUSION: Recipient older age, male sex, shorter hospital stay, use of T2D medications at discharge and donor female sex are important determinants in predicting hyperglycemia post-discharge up to 1 month following liver transplantation. This prognostic model may help identify patients at risk for hyperglycemia and thus reduce post-operative liver transplant hyperglycemia-related complications.
Disclosure: MEM: Member of advisory committees or review panels, Merck & Co., Advisory Group Member, Jansen Pharmaceuticals, Investigator, Johnson &Johnson, Investigator, Ipsen, Investigator, Novo Nordisk, Investigator, Novartis Pharmaceuticals, Investigator, Bayer, Inc., Owner, Amgen, Ad Hoc Consultant, Abbott Laboratories, Member of advisory committees or review panels, Pfizer, Inc.. AW: Investigator, Eli Lilly & Company, Consultant, Glytec, Investigator, Merck & Co.. Nothing to Disclose: HJZ, LV, TP, DH, LZ