Letter: Amino Acid Imbalance and Mortality in Cirrhosis—Interpretation and Methodological Considerations. Authors' Reply

We thank the authors for their thoughtful comments on our study examining the association between serum amino acid imbalance and mortality in patients with cirrhosis [1, 2]. We appreciate their interest in the clinical interpretation of our findings and in whether amino acid measures provide prognostic information beyond established severity scores. Below, we address each point in turn. First, to improve clinical interpretation, we performed additional analyses by rescaling serum branched-chain amino acid (BCAA) and tyrosine concentrations according to a 50 μmol/L increase. In the multivariable Cox models adjusted for age, sex, aetiology of cirrhosis, ammonia, and the model for end-stage liver disease–sodium (MELD-Na) score, higher BCAA concentrations were independently associated with a lower risk of mortality (hazard ratio [HR] per 50 μmol/L increase, 0.78; 95% confidence interval [CI], 0.71–0.86; p < 0.001), whereas higher tyrosine concentrations were associated with an increased risk of mortality (HR per 50 μmol/L increase, 1.79; 95% CI, 1.39–2.29; p < 0.001). In the model including the same covariates and the branched-chain amino acid to tyrosine ratio (BTR) instead of BCAA and tyrosine levels, BTR remained independently associated with mortality (HR, 0.69; 95% CI, 0.60–0.79; p < 0.001). These findings support that amino acid imbalance, whether assessed by individual amino acids or by BTR, is not merely a surrogate marker of liver functional reserve but an independent factor for mortality in patients with cirrhosis. Second, the restricted cubic spline analyses were further used to visualise the dose–response relationships between amino acid measures and mortality risk (Figure 1) [3]. Lower BTR and BCAA levels were associated with progressively higher HRs without significant evidence of non-linearity, whereas tyrosine demonstrated a significant non-linear association, with increased risk observed at both lower and higher concentrations. Robust linear associations between BTR, BCAA and mortality have been revealed from the figures; however, the non-linear association in tyrosine requires further research. To assess whether amino acid profiling improved model discrimination, we examined changes in the C-index after adding amino acid measures to a baseline multivariable model including age, sex, body mass index, cirrhosis aetiology, MELD-Na score, and serum ammonia levels. The addition of BCAA and tyrosine increased the C-index from 0.74 to 0.78 (ΔC-index, 0.04; 95% CI, 0.01–0.07). Similarly, adding BTR increased the C-index to 0.766 (ΔC-index, 0.03; 95% CI, 0.01–0.06). These findings suggest that evaluating amino acid imbalance can improve the existing prognostic frameworks by modestly enhancing risk discrimination, suggesting further study into its potential role in risk stratification and nutritional assessment. In conclusion, we thank the authors for their insightful comments, which have allowed us to clarify the clinical interpretation and robustness of our findings. Our additional analyses suggest that assessment of serum amino acid imbalance provides prognostic information in addition to routinely used clinical variables and severity scores, indicating an important role in risk stratification and management for patients with cirrhosis. Masashi Aiba: investigation. Shinji Unome: validation. Takao Miwa: writing – review and editing, conceptualization, methodology, supervision. Masahito Shimizu: project administration. This work was supported by the JSPS KAKENHI grant (grant number JP24K18908). This work was supported by the JSPS KAKENHI (grant number JP24K18908). The authors' declarations of personal and financial interests are unchanged from those in the original article [1]. The study protocol was reviewed and approved by the Institutional Review Board of the Graduate School of Medicine, Gifu University (approval number: 2025–038). Informed consent was obtained from all the participants using an opt-out method due to the retrospective nature of the study. The authors declare no conflicts of interest. This article is linked to Utakata et al. papers. To view these articles, visit https://doi.org/10.1111/apt.70525 and https://doi.org/10.1111/apt.70557. The datasets generated and/or analysed data during the current study are available from the corresponding author upon reasonable request.