Outnumbered: Control Prothrombin Time in Maddrey’s Discriminant Function Impacts Steroid Use but Not Mortality in Alcoholic Hepatitis

SIMPLE SUMMARY: Alcoholic hepatitis is associated with high morbidity and mortality. Maddrey’s discriminant function has been classically used to define the severity, with scores that are greater than 32 indicating poor outcomes and consideration for treatment with steroids. In order to use this calculation, a control prothrombin time is needed, which is not standardized across different electronic medical records or institutions. In this study, we aimed to determine if the control prothrombin time that is chosen has any effect on steroid use and mortality. We found that the higher the value of the control prothrombin time that was used, the less likely it was that individuals with alcoholic hepatitis would receive steroids. However, we found that this did not affect mortality in those who received steroids. This research is novel and important for providers to know when using Maddrey’s discriminant function, as the control PT values that they choose do have implications for treatment. Our findings also suggest that the model for end-stage liver disease (MELD) seems to be a better indicator of 30-day survival in this population. ABSTRACT: Background and aims: In alcoholic hepatitis (AH), increases in the total bilirubin (TB) and the prothrombin time (PT), which are included in the Maddrey’s discriminant function (MDF) and the model for end-stage liver disease (MELD), are associated with poor outcomes. However, the impact of which control PT in the MDF to use compared to the MELD on the outcomes in AH is unknown. Our aim is to determine whether the choice of the control PT used in the MDF calculations has any impact on steroid use and survival when compared to the MELD in those with AH. Methods: Through retrospective chart review, we analyzed 882 subjects who were admitted from 2012 to 2020 with acute AH. Their MDF was calculated [(TB + 4.6 × (PT–control)] using the following three different control PTs: 12, 13.5, and 14.8 s, and was compared to the MELD. The primary outcomes were steroid use and 30-day survival. Results: When it was stratified by the control PT, the percentage of MDF ≥ 32 (the threshold for steroids) decreased with increasing control PT (70%, 61%, and 52%, respectively), along with decreased steroid use (91%, 84%, and 75%, respectively). Those who received steroids were not shown to have improved 30-day survival compared to those who did not receive steroids (p = 0.41). The ability of the MDF for each control PT threshold to predict 30-day survival was similar (AUROC 0.735), and was lower compared to the MELD (0.767). Conclusion: While the choice of PT control in the MDF impacted the use of steroids in AH, the use of steroids and the choice of PT control used did not impact the overall survival. Regardless of which control PT was used in the MDF, the MELD was better at predicting 30-day survival. Important information: Background: Treatment with steroids is indicated in alcoholic hepatitis (AH) with Maddrey discriminant function (MDF) ≥ 32 and the model for end-stage liver disease (MELD) ≥ 20. The impact that the control prothrombin time (PT) value that is used in MDF has on steroid use and survival in AH is poorly understood. Findings: The choice of control PT that is used when calculating the MDF impacts the use of steroids but does not impact mortality. The MELD was better than the MDF at any control PT used in predicting survival in acute AH. Implications for patient care: Providers should be aware that higher control PT’s have an effect on treatment decisions but should not generally impact survival in this population. The MELD appears to better predict 30-day survival in this population.