SOFA como predictor de mortalidad en terapias extracorporeas en insuficiencia hepática aguda

INTRODUCTION

Therapeutic options for critically ill patients with acute liver failure are limited. In addition to standard therapy and intensive monitoring, extracorporeal liver support is an alternative, especially in cases progressing to multiorgan failure. Organ dysfunction secondary to liver damage hampers prognostic assessment and timely decision-making. In this context, various prognostic models have been proposed, including the Model for End-stage Liver Disease (MELD) and the Sequential Organ Failure Assessment (SOFA). Identifying a SOFA cutoff associated with mortality could provide a useful clinical tool for selecting patients who would benefit from timely initiation of extracorporeal therapies or early evaluation for liver transplantation.

The objective of this study is to evaluate the prognostic value of the SOFA score in predicting mortality in patients with acute liver failure treated with extracorporeal therapies (MARS or SPAD) and to identify a cutoff associated with a higher risk of mortality.

An observational, retrospective, and descriptive study was conducted. The medical records and institutional database of patients who received extracorporeal therapy using the Molecular Adsorbent Recirculating System (MARS) or Single Pass Albumin Dialysis (SPAD) for acute liver failure between January 2017 and July 2025 were reviewed. Demographic, clinical, biochemical, and outcome variables, including prognostic scores (MELD and SOFA), were collected, with the primary outcome being 3-month in-hospital mortality.

Thirty-five patients with acute liver failure who received extracorporeal support using MARS or SPAD during the study period were included. Of the total, 19 patients (54.3%) survived and 16 (45.7%) died during hospitalization.

Demographic and clinical variables: No significant differences were observed in age (41.2 ± 20.1 vs. 44.7 ± 20.3 years; p = 0.61) or in the proportion of men (63.2% vs. 36.8%; p = 0.30) between the non-mortality and mortality groups, respectively. The most frequent etiologic diagnoses were acute viral hepatitis, primary biliary cholangitis, and autoimmune hepatitis, with no differences between groups. The presence of acute kidney injury was similar (50% in both groups; p = 0.45).

Prognostic scales: The MELD score showed no significant differences (31 ± 13.3 vs. 29.7 ± 6.0; p = 0.76). In contrast, the SOFA score tended to be higher in the mortality group (17 [12–18]) compared to the survivors (9 [6.2–19]; p = 0.05).

Biochemical variables: No significant differences were found in creatinine, total bilirubin, direct bilirubin, albumin, ammonia, or INR. Hemoglobin (11.3 ± 3.3 vs. 9.2 ± 2.3 g/dL; p = 0.09) and platelets (130 ± 77 vs. 92 ± 70 ×10³/µL; p = 0.14) tended to be lower in the mortality group. Post-treatment creatinine reductions trended toward a lower magnitude in deceased patients (23.6% vs. 1.3%; p = 0.06).

Extracorporeal therapies: There were no significant differences in the type of therapy received (MARS 36.8% vs. 18.8%; SPAD 63.2% vs. 81.2%; p = 0.28).

ROC curve: ROC analysis for the SOFA score showed an AUC of 0.71 (95% CI: 0.52–0.90; p = 0.052). The optimal cutoff was SOFA ≥13.5, with a sensitivity of 76.9% and a specificity of 70.6% for predicting mortality.

In this study, the SOFA score demonstrated acceptable mortality prediction ability in patients with acute liver failure treated with extracorporeal support using MARS or SPAD. Although most demographic, clinical, and biochemical variables did not differ significantly between groups, the SOFA score was consistently higher in deceased patients, and ROC analysis revealed an AUC of 0.71, indicating moderate discriminatory ability. The identified optimal cutoff (SOFA ≥13.5) was associated with a sensitivity of 76.9% and a specificity of 70.6%, making it a clinically useful threshold.

Our results are consistent with previous research that has indicated the usefulness of the SOFA score as a predictor of adverse outcomes in severe liver disease and in critically ill patients in general. Unlike the MELD, which showed no significant differences in this cohort, the SOFA better reflected the impact of multiorgan dysfunction on clinical outcome, which is particularly relevant in settings where liver failure is accompanied by renal, hemodynamic, or respiratory compromise.

Although hemoglobin, platelets, and creatinine reduction showed trends toward significance, none reached statistical association, likely due to the limited sample size. However, these findings suggest that integrating hematological and renal parameters into predictive models could provide additional value.

This study has limitations, primarily its retrospective design and small number of patients, which limits the statistical power and generalizability of the findings. However, it provides local evidence on the usefulness of the SOFA as a prognostic tool in patients with acute liver failure undergoing extracorporeal support therapies and raises the need for prospective studies with larger sample sizes to validate the proposed cutoff point.

The SOFA score showed acceptable ability to predict mortality in patients with acute liver failure treated with MARS or SPAD. A cutoff score ≥13.5 was associated with a higher risk of mortality, suggesting its usefulness as a prognostic tool to guide the timely initiation of extracorporeal therapies and the prioritization of liver transplantation.