Validation of Chronic Liver Disease Questionnaire for Nonalcoholic Steatohepatitis in Patients With Biopsy-Proven Nonalcoholic Steatohepatitis

BACKGROUND & AIMS: The chronic liver disease questionnaire for nonalcoholic steatohepatitis (CLDQ-NASH) was developed in a systematic manner for assessment of patient-reported outcomes. This instrument collects data on 36 items grouped into 6 domains: abdominal symptoms, activity/energy, emotional health, fatigue, systemic symptoms, and worry. We aimed to validate the CLDQ-NASH in a large group of patients with NASH. METHODS: We collected data from patients with biopsy-proven NASH enrolled in 2 international phase 3 trials of selonsertib (NCT03053050 and NCT03053063). Our final analysis comprised 1667 patients who completed the CLDQ-NASH (age, 58 +/- 9 y; 40% male; 52% with cirrhosis; and 69% with type 2 diabetes). The CLDQ-NASH was administered before treatment initiation. A standard patient-reported outcome instrument validation pipeline with internal consistency and validity assessment was applied. RESULTS: The domains of CLDQ-NASH showed good to excellent internal consistency: the Cronbach's alpha values were 0.80 to 0.94 and item-to-own-domain correlations were greater than 0.50 for 33 of 36 items. All items correlated to the greatest extent with their own domains (discriminant validity). Known-group validity tests indicated that the instrument consistently discriminated between patients with NASH based on the presence of cirrhosis (vs bridging fibrosis; all but 1 P value < .02), obesity (all but 1 P value < .001), psychiatric comorbidities (all P values < .0001), fatigue (all P values < .001), and type 2 diabetes (all but 1 P value < .01). Of the CLDQ-NASH domains, the highest correlated domains with the Short Form-36 were as follows: physical functioning for activity (rho = 0.70), mental health for emotional (rho = 0.72), vitality for fatigue (rho = 0.75), and body pain for systemic (rho = 0.72) (all P values < .0001). In contrast, the domains of abdominal and worry, which are disease-specific, did not correlate with the domains in the Short Form-36 (all rho <= 0.50). CONCLUSIONS: We validated the CLDQ-NASH by an analysis of data from 1667 patients with biopsy-proven NASH enrolled in phase 3 trials, observing excellent psychometric characteristics of the instrument.