Utility of scores to predict alcohol use after liver transplant: Take them with a grain of salt

Kevin Houston, Nikki Duong, Richard K. Sterling, Amon Asgharpour, Sheila Bullock, Stephan Weinland, Nicole Keller, Ekaterina Smirnova, Hiba Khan, Scott Matherly, Joel Wedd, Hannah Lee, Mohammad Siddiqui, Vaishali Patel, Albert Arias, Vinay Kumaran, Seung Lee, Amit Sharma, Aamir Khan, Daisuke ImaiMarlon Levy, David Bruno

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

The Sustained Alcohol use post-Liver Transplant (SALT) and the High-Risk Alcohol Relapse (HRAR) scores were developed to predict a return to alcohol use after a liver transplant (LT) for alcohol-associated liver disease. A retrospective analysis of deceased donor LT from October 2018 to April 2022 was performed. All patients underwent careful pre-LT psychosocial evaluation. Data on alcohol use, substance abuse, prior rehabilitation, and legal issues were collected. After LT, all were encouraged to participate in rehabilitation programs and underwent interval phosphatidylethanol testing. Patients with alcohol-associated liver disease were stratified by < or > 6 months of sobriety before listing. Those with <6 months were further stratified as acute alcoholic hepatitis (AH) by NIAAA criteria and non-AH. The primary outcome was the utility of the SALT (<5 vs. ≥5) and HRAR (<3 vs. ≥3) scores to predict a return to alcohol use (+phosphatidylethanol) within 1 year after LT. Of the 365 LT, 86 had > 6 months of sobriety, and 85 had <6 months of sobriety; 41 with AH and 44 non-AH. In those with AH, the mean time of abstinence to LT was 58 days, and 71% failed prior rehabilitation. Following LT, the return to drinking was similar in the AH (24%) compared to <6-month non-AH (15%) and >6-month alcohol-associated liver disease (22%). Only 4% had returned to heavy drinking. The accuracy of both the SALT and HRAR scores to predict a return to alcohol was low (accuracy 61%-63%) with poor sensitivity (46% and 37%), specificity (67%-68%), positive predictive value (22%-26%) with moderate negative predictive value (81%-83%), respectively with higher negative predictive values (95%) in predicting a return to heavy drinking. Both SALT and HRAR scores had good negative predictive value in identifying patients at low risk for recidivism.

Original languageEnglish
JournalLiver Transplantation
DOIs
StateAccepted/In press - 2024
Externally publishedYes

Cite this