Predicting Recompensation in Alcohol, HBV and HCV-associated Decompensated Cirrhosis.

BACKGROUND

Recompensation of cirrhosis represents sustained clinical improvement in decompensated cirrhosis following effective etiological treatment. Baseline predictors can help prognostication in hepatology practice.

METHODS

We conducted a retrospective study of patients with decompensated cirrhosis due to alcohol-associated (ALD) or viral (HBV or HCV) cirrhosis who received etiology-specific treatment following decompensation. Demographic details, laboratory parameters, and liver stiffness measurement (LSM) parameters were recorded at index decompensation, presentation, and follow-up. Recompensation was defined per Baveno-VII criteria. CT-derived sarcopenia (L3) and vertebral bone density (L1) were obtained when scans were available within 2 years of decompensation. Fine-Gray competing risk regression accounted for death, hepatocellular carcinoma (HCC), and liver transplantation as competing events.

RESULTS

Of 244 patients (median age 43 (35-50) years; 78.7% males) with ALD (n = 104), HBV (n = 81), or HCV (n = 59) decompensated cirrhosis, 70 (28.7%) achieved recompensation over a median follow-up of 38 months (IQR 18.5-69). Recompensation was more frequent in HBV(42%) and HCV(35.6%) cirrhosis vs ALD (14.4%, P < 0.001). Recompensated patients had lower MELD score (10.1 vs 14.7, P < 0.001), bilirubin (1.5 vs 2.2 mg/dL, P = 0.006), and INR (1.3 vs 1.5, P = 0.002) at baseline. Univariate competing-risk analysis showed stepwise reduction in recompensation likelihood with increasing MELD category (MELD 10-15: sHR 0.56, MELD >15: sHR 0.34, vs MELD <10). On multivariable analysis, higher platelet count, absence of jaundice at baseline, and HBV etiology were independently associated with recompensation. LSM (n = 145) was numerically lower in recompensated patients (29.3 vs 36.9 kPa, P = 0.1), with greater median decline at follow-up (-20.7% vs 0%, P = 0.007). CT-based measurements (n = 69) showed recompensated patients had lower prevalence of sarcopenia (27.3% vs 59.6%, P = 0.01), higher vertebral bone density (243.1 ± 50.5 vs 196.2 ± 62.8 HU, P = 0.003), and no osteopenia (L1 vertebral body density <150 HU, 0% vs 31.9%, P = 0.003). Three (4.3%) recompensated patients developed HCC.

CONCLUSION

Recompensation occurs in one-fourth of patients with decompensated cirrhosis and is more frequent in viral than alcohol-associated etiologies. Higher platelet count and absence of jaundice at baseline independently predicted recompensation.