Validating and expanding Baveno VII criteria of recompensation in patients with decompensated cirrhosis

Background and aims Baveno-VII consensus defines recompensation in cirrhotic patients achieving: 1) removal/suppression/cure of cirrhosis's etiology of cirrhosis; 2) resolution of ascites (off diuretics), hepatic encephalopathy (off lactulose/rifaximin) and bleeding for>12 months; c) stable improvement of liver function. This study aims to evaluate the incidence and prognostic impact of recompensation in patients with decompensated cirrhosis. Methods Outpatients with cirrhosis and curable etiologies (alcohol, HCV, HBV) were consecutively included and followed up (median time=35 months). Demographic, clinical, laboratory and endoscopic data were collected at enrolment and follow-up visits. Recompensation, defined by Baveno VII criteria, was assessed. Considering the subjectivity of treatment withdrawal, we evaluated expanded recompensation criteria for patients meeting all criteria, but still on decompensation treatment (diuretics/lactulose/rifaximin). Recompensation was considered a time-varying covariate in survival analysis. In 160 patients (62 compensated, 60 decompensated, 38 recompensated), plasma samples were analyzed for inflammatory cytokines (IL-6,IL-10,IL1beta). Results 691 patients were enrolled (mean age 57±11 years, men 72.5%; alcohol=55%). Among decompensated patients (n=525), 298 achieved an effective etiological treatment and 22 (4.2%) achieved recompensation (Baveno-VII criteria), while 115 patients achieved expanded recompensation criteria (22.3%). MELD score was the only independent predictor of recompensation (aHR=0.90;p=0.002). In multivariable analysis (adjusted for age, sex, MELD, albumin, varices and further decompensation), mortality risk showed no significant difference between patients achieving recompensation and compensated patients (aHR=2.53; p=0.107), while decompensated patients had the highest risk (aHR=4.74;p<0.001). Mortality risk showed no significant difference between patients meeting expanded recompensation criteria and Baveno-VII criteria (HR=1.02; p=0.961). IL-6 and IL-10 were significantly higher in decompensated patients than in compensated ones (Figure). Following recompensation, inflammatory cytokines significantly decreased and no difference was found vs compensated patients. Conclusions Baveno-VII criteria identify cirrhotic patients with a prognosis similar to compensated patients, but<5% achieve recompensation. Expanding the criteria to include patients on medical treatment for decompensation identifies patients at low risk of mortality. Baveno-VII consensus defines recompensation in cirrhotic patients achieving: 1) removal/suppression/cure of cirrhosis's etiology of cirrhosis; 2) resolution of ascites (off diuretics), hepatic encephalopathy (off lactulose/rifaximin) and bleeding for>12 months; c) stable improvement of liver function. This study aims to evaluate the incidence and prognostic impact of recompensation in patients with decompensated cirrhosis. Outpatients with cirrhosis and curable etiologies (alcohol, HCV, HBV) were consecutively included and followed up (median time=35 months). Demographic, clinical, laboratory and endoscopic data were collected at enrolment and follow-up visits. Recompensation, defined by Baveno VII criteria, was assessed. Considering the subjectivity of treatment withdrawal, we evaluated expanded recompensation criteria for patients meeting all criteria, but still on decompensation treatment (diuretics/lactulose/rifaximin). Recompensation was considered a time-varying covariate in survival analysis. In 160 patients (62 compensated, 60 decompensated, 38 recompensated), plasma samples were analyzed for inflammatory cytokines (IL-6,IL-10,IL1beta). 691 patients were enrolled (mean age 57±11 years, men 72.5%; alcohol=55%). Among decompensated patients (n=525), 298 achieved an effective etiological treatment and 22 (4.2%) achieved recompensation (Baveno-VII criteria), while 115 patients achieved expanded recompensation criteria (22.3%). MELD score was the only independent predictor of recompensation (aHR=0.90;p=0.002). In multivariable analysis (adjusted for age, sex, MELD, albumin, varices and further decompensation), mortality risk showed no significant difference between patients achieving recompensation and compensated patients (aHR=2.53; p=0.107), while decompensated patients had the highest risk (aHR=4.74;p<0.001). Mortality risk showed no significant difference between patients meeting expanded recompensation criteria and Baveno-VII criteria (HR=1.02; p=0.961). IL-6 and IL-10 were significantly higher in decompensated patients than in compensated ones (Figure). Following recompensation, inflammatory cytokines significantly decreased and no difference was found vs compensated patients. Baveno-VII criteria identify cirrhotic patients with a prognosis similar to compensated patients, but<5% achieve recompensation. Expanding the criteria to include patients on medical treatment for decompensation identifies patients at low risk of mortality.