Altered cognition and patient-reported outcomes predict 1-year hospitalization risk in a multi-center outpatient...

The clinical characteristics of cirrhosis are changing, and many continue to have complications despite having a controlled etiology. In a contemporary cohort, patient reported outcomes (PROs) and cognitive function were evaluated to assess their impact on risk for future hospitalization.

1007 outpatients with cirrhosis, median MELD 11, 39% compensated, 61% decompensated, all controlled etiology where possible (no alcohol, HCV SVR, suppressed HBV), were prospectively enrolled at 11 NACSELD-3 (North American Consortium for the Study of End-Stage Liver Disease) centers. Clinical, comorbidity and demographic data were captured. All patients underwent baseline and 6-month testing for SF-36 and Stroop and were systematically followed for 1-year. Time-to-event regression with competing risks for death/transplant for all-cause, liver-related, and HE-related hospitalizations was performed. By 1-year, one-third required hospitalization; two-thirds were liver-related, and one quarter of liver-related admissions were hepatic encephalopathy (HE)-driven. On multivariable modeling, higher Stroop time and lower SF-36 general health score independently predicted all-cause (Stroop/SF-36,p<0.001 each), liver-related (Stroop, p=0.01, SF-36, p<0.001) and HE-related (Stroop/SF-36, p=0.02 each) hospitalizations. Adding Stroop and SF-36 to models with traditional risk factors significantly improved C-index across all outcomes (all p<0.001).

In an outpatient cohort with etiology-controlled cirrhosis with a median MELD of 11, one-third required non-elective hospitalization by 1-year. PROs using SF-36 and cognitive performance measured by Stroop were independently additive to clinical parameters to predict future risk of all-cause, liver-related and HE-related hospitalization. Routine assessment of cognition and PROs could improve holistic outpatient management and prognostication in cirrhosis.