Purpose
To explore whether glycated albumin (GA) or fasting plasma glucose (FPG), both routinely monitored during patients’ hospital stay, can be used to predict post-transplantation diabetes mellitus (PTDM).
Methods
All kidney transplantation recipients (KTRs) from January 2017 to December 2018 were followed-up for 1 year. PTDM was diagnosed from day 45 post-operation to 1 year. When the completeness was above 80%, FPG or GA data on the day was selected, analyzed, and presented as range parameters and standard deviation (SD) and compared between PTDM and non-PTDM groups in fluctuation and stable periods. The predictive cut-off values were determined via receiver operating characteristic (ROC) analysis. The PTDM combined predictive mode, formed by the independent risk factors derived from logistic regression analyses, was compared with each independent risk factor with the independent ROC curve test.
Results
Among 536 KTRs, 38 patients developed PTDM up to 1 year post-operatively. The family history diabetes mellitus (DM; OR, 3.21; P = 0.035), the FPG SD in fluctuation period > 2.09 mmol/L (OR, 3.06; P = 0.002), and the FPG maximum in stable period > 5.08 mmol/L (OR, 6.85; P < 0.001) were the PTDM independent risk factors. The discrimination of the combined mode (area under the curve = 0.81, sensitivity = 73.68%, and specificity = 76.31%) was higher than each prediction (P < 0.05).
Conclusions
The FPG SD during the fluctuation period, FPG maximum during the stable period, and family history DM predicted PTDM with good discrimination and potential routine clinical use.