Korean J Transplant.  2022 Nov;36(Supple 1):S306. 10.4285/ATW2022.F-4443.

Determination of tacrolimus dosage using machine learning in kidney transplantation

Affiliations
  • 1Department of Surgery, The Catholic University of Korea, Uijeongbu St. Mary's Hospital, Uijeongbu, Korea
  • 2Department of Surgery, The Catholic University of Korea, Yeouido St. Mary's Hospital, Seoul, Korea
  • 3Department of Surgery, The Catholic University of Korea, Seoul St. Mary's Hospital, Seoul, Korea

Abstract

Background
Maintaining tacrolimus trough levels in kidney transplantation is very important. The purpose of this study is to analyze the determination of tacrolimus dosage to maintain tacrolimus trough levels using machine learning.
Methods
This retrospective study included 801 consecutive patients from a prospectively registered database who underwent kidney transplantation at Seoul St. Mary’s Hospital, South Korea, between January 1, 2015, and December 30, 2019. After kid-ney transplantation, supervised learning was performed based on individual tacrolimus trough levels and tacrolimus dosage during hospitalization.
Results
A total of 771 patients were enrolled in the study with a mean age 48.7±11.5 years (range, 16–75 years). Four hundred forty-five patients (57.7%) were male. Three hundred twenty-six patients (42.3%) were female. One hundred fifty-seven patients (20.4%) were ABO incompatible kidney transplantation and 196 (25.4%) were deceased donor kidney transplantation. Signifi-cant results of tacrolimus trough levels and tacrolimus dosage during hospitalization were confirmed through machine learn-ing. It was analyzed that weight had a significant effect.
Conclusions
Determination of tacrolimus dosage to maintain appropriate tacrolimus trough levels through machine learning during hospitalization after kidney transplantation should be considered as a useful tool.

Full Text Links
  • KJT
Actions
Cited
CITED
export Copy
Close
Share
  • Twitter
  • Facebook
Similar articles
Copyright © 2024 by Korean Association of Medical Journal Editors. All rights reserved.     E-mail: koreamed@kamje.or.kr