Optimisation of the use of APRI and FIB-4 to rule out cirrhosis in patients with chronic hepatitis B: results from the SONIC-B study

Abstract

Background: Ruling out the presence of cirrhosis is important for the management of chronic hepatitis B. We aimed to study and optimise the performance of two non-invasive indices for ruling out cirrhosis: the aspartate aminotransferase-platelet ratio index (APRI) and fibrosis score based on four factors (FIB-4).

Methods: We applied established cutoffs to rule in (APRI >2·00; FIB-4 >3·25) or rule out (APRI <1·00; FIB-4 <1·45) cirrhosis to data from eight global randomised trials that required baseline biopsy, and identified new cutoffs aiming for a sensitivity for detection of cirrhosis greater than 90% and a negative predictive value (NPV) of greater than 95% in the same dataset. We externally validated the new cutoffs using data from all consecutive biopsied patients from two tertiary referral hospitals in the Netherlands and Canada.

Findings: In the derivation dataset (n=2926; of whom 1750 were Asian); 340 (12%) individuals had cirrhosis. The validation cohort consisted of 1034 individuals (of whom 575 were Asian), with 155 (15%) individuals with cirrhosis. Application of conventional cutoffs for FIB-4 in the derivation dataset yielded unclassifiable results in 686 (23%) individuals, and 139 (41%) of the 340 patients with cirrhosis were misclassified as having no cirrhosis. Similarly, conventional cutoffs for APRI in the derivation dataset yielded unclassifiable results in 706 (24%) individuals, and 153 (45%) were misclassified as having no cirrhosis. An APRI of 0·45 or less had sensitivity of 91·5%, an NPV of 95·4%, and misclassified 29 (9%) of 340 individuals with cirrhosis in the derivation dataset, but performance was reduced in the validation set (22 [14%] of 155 individuals with cirrhosis misclassified). A FIB-4 score of 0·70 had a sensitivity of 90·9%, an NPV of 96·6%, and misclassified 31 (9%) of individuals with cirrhosis in the derivation dataset. In the validation cohort, the same score gave a sensitivity of 94·2%, an NPV of 97·3%, and misclassified nine (6%) of the individuals with cirrhosis. Subgroup analysis indicated that the new FIB-4 cutoff performed acceptably in all subgroups except for individuals aged 30 years or younger.

Interpretation: Conventional cutoffs for APRI and FIB-4 should not be used to guide management of patients with chronic hepatitis B due to high rates of misclassification. A newly identified and externally validated cutoff for FIB-4 (≤0·70) can be used to exclude cirrhosis in patients over 30 years of age.

Funding: Foundation for Liver and Gastrointestinal Research, Rotterdam, Netherlands.