Abstract
BACKGROUND: New-onset diabetes after kidney transplant (NODAT) is a severe metabolic complication that frequently occurs in recipients following transplantation.
AIM: The study aims to verify NODAT, compare cases and non-cases of this entity, and explore potential predictors in recipients within 1 year following kidney transplantation.
METHODS: The research is a retrospective study of 90 renal transplant recipients (n = 90). Demographic factors and clinical aspects were analyzed using non-Bayesian statistics and machine learning (ML). The clinical aspects included the glycated hemoglobin (HbA1c) level, associated viral infections (hepatitis B virus [HBV], hepatitis C virus [HCV], and cytomegalovirus [CMV]), prior kidney transplant, hemodialysis status, body mass index (BMI) at transplant time, and 3 months later, primary causes of renal failure, and post-transplant therapeutics. All individuals were on cyclosporine and prednisolone treatment.
RESULTS: The mean age was 39 (±1.5) years; recipients included 27 females (30%) and 63 males (70%). Donor type was live related (16, 17.8%) or live unrelated (74, 82.2%); 27 recipients (30%) had O+ blood group, while 70% belonged to other groups. Thirteen recipients (14.4%) were not on dialysis. Only 32 individuals (35.6%) developed NODAT. Concerning virology, confirmed by real-time polymerase chain reaction before transplantation, 19 recipients (21.1%) were CMV positive, 9 (10%) were HCV positive, and 2 (2.2%) had HBV.
CONCLUSIONS: In reconciliation with frequentist statistics, the dual ML model validated several predictors that either negatively (protective) or positively (harmful) influenced HbA1c level, the majority of which were significant at 95% confidence interval. Individuals who are HCV and CMV positive are predicted to develop NODAT. Further, older individuals, with blood group O+ve, prior history of hemodialysis, a relatively high BMI before the transplant, and receiving higher doses of prednisolone following the transplant are more likely to develop NODAT. The current study represents the first research from Iraq to explore NODAT predictors among kidney transplant recipients using frequentist statistics and artificial intelligence models.
Publisher
Scientific Foundation SPIROSKI