Reply—Examining pregnancy outcomes in alcohol-associated hepatitis: A call for further research

We thank Dr El-Kassas and colleagues for their thoughtful comments on our recent publication exploring pregnancy outcomes in women with alcohol-associated hepatitis (AH).1 We feel that the authors both recognized the novelty of our study and raised valid points of concern regarding study design. Among these, there are 2 points we would like to address: First, we acknowledge the concerns regarding the use of ICD-10 codes to identify patients with AH, given the variability in diagnostic patterns. Our colleagues suggested using validated algorithms and/or incorporating laboratory data to improve patient identification—citing a recent study that validated coding algorithms for cirrhosis—not AH.2 Looking closer, the supplemental materials for the referenced study suggest that the AH ICD-10 code (K70.1) alone might have higher specificity for identifying AH compared to the ICD-10 code plus laboratory data. Although there is an opportunity to learn more through large-scale chart reviews and prospective data collection, we believe that our strategy minimized the risk of falsely including women without AH and that our findings accurately represent the target population. Nevertheless, we agree that future studies should focus on developing both sensitive and specific coding algorithms to identify AH for database research. Second, the authors noted the absence of detailed information on alcohol consumption patterns, suggesting these data are needed to determine AH severity. Although we agree that patterns of alcohol intake influence the risk of AH development and recurrence, there are several modifiers of this relationship, such as genetic predisposition, body composition, and environmental factors.3 We chose to use biochemical markers as objective measures of AH severity and recurrent AH as a surrogate for high-risk drinking behaviors. As the authors point out, we observed fewer pregnancies among women with higher bilirubin levels and more adverse pregnancy outcomes in women with recurrent AH. While understanding alcohol consumption patterns would further characterize this population and inform future interventions, our findings sufficiently demonstrate that both the severity of liver dysfunction and high-risk alcohol use influence pregnancy outcomes. Finally, we are very encouraged by the suggestions for future research, including biomarker identification and evaluation of medication-assisted therapies, and investigation into racial and ethnic disparities. Given the increasing incidence of AH in women of reproductive age, further research will be essential to inform management strategies and improve clinical outcomes. We hope our use of large-scale, real-world data will inspire such investigations. We thank Dr El-Kassas and colleagues again for their in-depth analysis of our article and the opportunity to advance the conversation around this important clinical issue.