Inpatient Frailty Assessment Is Feasible and Predicts Nonhome Discharge and Mortality in Decompensated Cirrhosis

Abstract

Objective inpatient frailty assessments in decompensated cirrhosis are understudied. We examined the feasibility of inpatient frailty measurements and associations with nonhome discharge, readmission, and all-cause mortality among patients admitted for cirrhosis complications. We conducted a prospective study at 3 liver transplantation (LT) centers. Frailty was assessed using the liver frailty index (LFI). Multivariable logistic and competing risk models evaluated associations between frailty and clinical outcomes. We included 211 patients with median MELD-Na score 21 (interquartile range [IQR],15-27); 96 (45%) were women, and 102 (48%) were on the LT waiting list. At a median follow-up of 8.3 months, 29 patients (14%) were nonhome discharged, 144 (68%) were readmitted, 70 (33%) underwent LT, and 44 (21%) died. A total of 124 patients (59%) were frail, with a median LFI of 4.71 (IQR, 4.07-5.54). Frail patients were older (mean, 59 versus 54 years) and more likely to have chronic kidney disease (40% versus 20%; P = 0.002) and coronary artery disease (17% versus 7%; P = 0.03). Frailty was associated with hospital-acquired infections (8% versus 1%; P = 0.02). In multivariable models, LFI was associated with nonhome discharge (odds ratio, 1.81 per 1-point increase; 95% confidence interval [CI], 1.14-2.86). Frailty (LFI≥4.5) was associated with all-cause mortality in models accounting for LT as competing risk (subhazard ratio [sHR], 2.4; 95% CI, 1.13-5.11); results were similar with LFI as a continuous variable (sHR, 1.62 per 1-point increase; 95% CI, 1.15-2.28). A brief, objective inpatient frailty assessment was feasible and predicted nonhome discharge and mortality in decompensated cirrhosis. Inpatient point-of-care frailty assessment prior to hospital discharge can be useful for risk stratification and targeted interventions to improve physical fitness and reduce adverse outcomes.

FIG. 1.

(A) Primary admission reason by frailty category (n = 211). “Other” admission reason includes splanchnic thrombosis, abdominal pain, worsening liver tests, ileus, symptomatic anemia, lightheadedness, chest pain, choledocholithiasis, acute coronary syndrome, post–transjugular intrahepatic portosystemic shunt monitoring, and alcoholic hepatitis. (B) Primary readmission reason by frailty category (n = 144). “Other” readmission reasons include arrhythmia, fall, alcoholic hepatitis, cholecystitis, worsening liver function, anemia, small bowel obstruction, ventral hernia repair, potential/cancelled liver transplantation, depression, diabetic ketoacidosis, respiratory failure.

FIG. 2.

Box-and-whisker plots of inpatient LFI scores by liver disease etiology. The line through each box indicates the median (50th percentile) of LFI values with the outside borders showing the 25th (lower border) and 75th (upper border) percentiles. LFI threshold for frailty (4.5) is denoted with a dashed line. *P < 0.05 for LFI with NAFLD/NASH versus other liver disease etiologies.

FIG. 3.

Box-and-whisker plots of inpatient LFI scores stratified by presence of medical comorbidities. The line through each box indicates the median (50th percentile) of LFI values with the outside borders showing the 25th (lower border) and 75th (upper border) percentiles. LFI threshold for frailty (4.5) is denoted with a dashed line. *P < 0.05 in bivariate comparisons.

FIG. 4.

Kaplan-Meier curves of readmissions after index hospitalization based on frailty status (nonfrail with LFI <4.5, frail with LFI ≥4.5). The y axis indicates the percentage free from readmission. The x axis indicates the time in days after index hospital discharge.

FIG. 5.

Kaplan-Meier curves of all-cause mortality after index hospitalization based on frailty status (nonfrail with LFI <4.5, frail with LFI ≥4.5 The y axis indicates the percentage free from mortality. The x axis indicates the time in days after index hospital discharge.