Background & aims: Non-cardioselective beta-blocker (NSBB) effects on mortality in cirrhosis are controversial. We evaluated the impact of NSBBs on mortality according to liver severity and mortality cause.
Methods: Two hundred and fifty-eight patients with alcoholic cirrhosis were included in a retroprospective cohort: 129 NSBB-treated and 129 controls. The NSBB group had the following significant baseline differences: higher MELD, more frequent previous gastrointestinal bleeding, large oesophageal varices (OV) and lower heart rate. Propranolol dose was 160 mg/d in 81% of NSBB patients.
Results: (i) Liver function: during 5.3 ± 2.6 years of follow-up, MELD progression was higher in NSBB patients: 1 (-1-4) than in controls: 0 (-1-1) (P = .017). (ii) Overall survival: no significant differences were observed between NSBBs and controls (Kaplan-Meier curves: P = .291). In multivariate Cox analysis, baseline MELD interacted with NSBB (P = .011). Thus, the NSBB hazard ratio (HR) was 0.99 (0.50-1.98) in MELD < 12 vs 3.17 (1.19-8.42) in MELD ≥ 12. (iii) Liver survival: NSBB decreased liver survival (Kaplan-Meier: P = .031). In multivariate Cox analysis, baseline MELD interacted with NSBB (P < .001). The NSBB HR was 0.81 (0.30-2.19) in MELD < 12 vs 6.23 (1.94-20.0) in MELD ≥ 12. In competing risk multivariate analysis for liver mortality, the MELD-NSBB interaction was significant (P < .001): the NSBB HR was 1.02 (0.36-2.91) in MELD < 12 vs 9.24 (3.18-26.9) in MELD ≥ 12. 4) Non-liver survival: contrastingly, non-liver survival was increased by NSBBs, especially in MELD ≥ 12 (competing Kaplan-Meier: P = .044). These results were confirmed in propensity risk score (PRS)-matched patients.
Conclusion: In alcoholic cirrhosis with rather high propranolol doses, overall and liver survival are significantly aggravated when MELD is ≥12.
Keywords: observational cohort; portal hypertension; prognosis; propranolol.
© 2020 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.