Letter: Statins for Cirrhosis: Are we Ready to Move From Cardiovascular Prophylaxis to Portal Pressure Management? Authors' Reply

We read with great interest the correspondence by Dr. Ismaiel [1] regarding our systematic review and meta-analysis [2]. We appreciate his thoughtful appraisal of the evolving role of statins in cirrhosis and his emphasis on the apparent discrepancy between the robust mortality benefit observed in our analysis, supported mechanistically by hepatic venous pressure gradient reduction, and the more conservative recommendations of current major guidelines [3, 4]. Dr. Ismaiel correctly highlights a key nuance in our findings: Although statin therapy significantly reduced all-cause mortality in randomised controlled trials (RCTs), the reduction in hepatic decompensation did not reach statistical significance within the RCT subgroup (odds ratio 0.75; 95% CI 0.52–1.09). He further raises the possibility that ‘healthy user bias’ in observational studies and differences in baseline disease severity may contribute to the apparent liver-specific benefits observed outside the randomised setting. To address this issue, we conducted a post hoc exploratory meta-regression restricted to RCTs, thereby minimising confounding related to treatment selection in observational data. We examined whether baseline disease severity, assessed using the study-level mean or median Model for End-Stage Liver Disease (MELD) score for the overall population, modified the treatment effect of statins. This analysis included six RCTs [5-10] reporting MELD scores, spanning a baseline severity range of approximately 6–15. For all-cause mortality, we observed no significant interaction between baseline MELD score and statin efficacy (Figure 1A; slope = 0.02; p = 0.851). This finding suggests the survival benefit observed in RCTs was consistent across the range of disease severity represented in these trials and was not driven solely by trials with lower MELD scores. Similarly, for hepatic decompensation, meta-regression showed no significant interaction with baseline MELD score (Figure 1B; slope = −0.09; p = 0.226), indicating the lack of statistical significance for this outcome in RCTs was consistent regardless of the average disease severity of the study populations. We acknowledge important limitations. This analysis is exploratory, includes a small number of trials and is inherently vulnerable to ecological bias and limited power; therefore, it cannot define an optimal ‘sweet spot’ for statin therapy across the full spectrum of cirrhosis. An individual patient data meta-analysis would be valuable to better delineate effect modification by disease severity, to account for competing risks and to more precisely characterise safety thresholds, particularly in patients with decompensated disease. Finally, we agree that safety remains central to cautious guideline recommendations as current evidence suggests that higher-dose vastatin in decompensated cirrhosis may increase rhabdomyolysis risk, supporting careful patient selection and conservative dosing regimens. In conclusion, although statins have not yet demonstrated a statistically significant reduction in hepatic decompensation in RCTs, the mortality benefit observed in randomised trials appears robust and not materially modified by baseline MELD score within the severity ranges studied. These findings reinforce the biological and clinical plausibility of statins as disease-modifying agents in cirrhosis, while underscoring the need for larger, long-term randomised studies to define their optimal and safest use. The authors' declarations of personal and financial interests are unchanged from those in the original article [2]. Bernardo de Faria Moraes: conceptualization, writing – original draft, formal analysis, methodology, data curation. Gustavo André Pedral Diniz Leite: investigation, writing – review and editing. Igor Boechat Silveira: investigation, writing – review and editing. Gabriel André Pedral Diniz Leite: investigation, writing – review and editing. Maria Luisa Motta Fonseca: investigation, writing – review and editing. Leonardo Corrêa Suffert: investigation, writing – review and editing. Luisa Medeiros Visentini: investigation, writing – review and editing. Luis Pedro Possapp Beis: investigation, writing – review and editing. Guilherme Grossi Lopes Cançado: conceptualization, supervision, data curation, writing – review and editing, methodology, writing – original draft, project administration. The authors have nothing to report. The authors nothing to report. Guarantor of the article: BFM and GGLC. The authors declare no conflicts of interest. This article is linked to Moraes et al. papers. To view these articles, visit https://doi.org/10.1111/apt.70526 and https://doi.org/10.1111/apt.70591. The data that support the findings of this study are available from the corresponding author upon reasonable request.