New-Onset Diabetes and Preexisting Diabetes Are Associated With Comparable Reduction in Long-Term Survival After Liver Transplant: A Machine Learning Approach

Mayo Clin Proc. 2018 Dec;93(12):1794-1802. doi: 10.1016/j.mayocp.2018.06.020.

Abstract

Objective: To identify key predictors and survival outcomes of new-onset diabetes after transplant (NODAT) in liver transplant (LT) recipients by using the Scientific Registry of Transplant Recipients.

Patients and methods: Data of all adult LT recipients between October 1, 1987, and March 31, 2016, were analyzed using various machine learning methods. These data were divided into training (70%) and validation (30%) data sets to robustly determine predictors of NODAT. The long-term survival of patients with NODAT relative to transplant recipients with preexisting diabetes and those without diabetes was assessed.

Results: Increasing age (odds ratio [OR], 1.01; 95% CI, 1.00-1.02; P≤.001), male sex (OR, 1.09; 95% CI, 1.05-1.13; P=.03), and obesity (OR, 1.13; 95% CI, 1.08-1.18; P<.001) were significantly associated with NODAT. Sirolimus as a primary immunosuppressant carried a 33% higher risk of NODAT than did tacrolimus (OR, 1.33; 95% CI, 1.22-1.45; P<.001) at 1 year after LT. Patients with NODAT had significantly decreased 10-year survival than did those without diabetes (63.0% vs 74.9%; P<.001), similar to survival in patients with diabetes before LT (58.9%).

Conclusion: Using a machine learning approach, we found that older, male, and obese recipients are at especially higher risk of NODAT. Donor features do not affect risk. In addition, sirolimus-based immunosuppression is associated with a significantly higher risk of NODAT than other immunosuppressants. Most importantly, NODAT adversely affects long-term survival after LT in a manner similar to preexisting diabetes, indicating the need for more aggressive care and closer follow-up.

Publication types

  • Validation Study

MeSH terms

  • Adult
  • Age Factors
  • Algorithms
  • Cross-Sectional Studies
  • Diabetes Mellitus / epidemiology*
  • Diabetes Mellitus / etiology
  • Female
  • Humans
  • Immunosuppressive Agents / adverse effects*
  • Liver Transplantation / adverse effects
  • Liver Transplantation / mortality*
  • Machine Learning
  • Male
  • Middle Aged
  • Obesity / complications
  • Risk Factors
  • Sex Factors
  • Sirolimus / adverse effects*
  • Tissue Donors / statistics & numerical data
  • United States / epidemiology

Substances

  • Immunosuppressive Agents
  • Sirolimus