Liver frailty index improves mortality prediction of subjective clinician assessment
Addition of the liver frailty index (LFI) to the subjective clinician assessment significantly improves mortality risk prediction, according to a recent study.
The researchers recruited 529 cirrhosis outpatients (median age 58 years; 42 percent female) who were awaiting liver transplantation. Participants underwent both objective LFI and subjective clinician assessments. Concordance statistics were used to determine the discriminative ability of the methods.
The median LFI score was 3.8 (3.4–4.3), while the median subjective clinician assessment score was 2 (1–3). Eight trained professionals administered the LFI to the participants, and the scores were statistically similar (p=0.28). In both cases, higher scores indicated greater patient frailty.
Spearman’s correlation analysis showed that scores in the LFI and subjective clinician assessments were significantly and positively associated with each other (p<0.001).
A univariable Cox regression analysis showed that each unit increase in the LFI was significantly associated with waitlist mortality (hazard ratio [HR], 2.9; 95 percent CI, 2.2–3.7; p<0.001). The same trend was found with the subjective clinician assessment (HR, 1.9; 1.6–2.3; p<0.001).
Multivariable adjustments did not significantly attenuate relationship of LFI (HR, 2.21; 1.68–2.91; p<0.001) and subjective clinician assessment (HR, 1.59; 1.30–1.94; p<0.001) with waitlist mortality.
Individually, the LFI and subjective clinician assessment had acceptable abilities to correctly rank patients according to their risk of death (C-statistic, 0.71 vs 0.68; p=0.41). The combination of the two significantly improved mortality prediction, resulting in a significantly better C-statistic relative to subjective assessments alone (p<0.02) but not to LFI alone (p=0.31).