A novel model based on qAnti-HBc and conventional biomarkers for identifying significant liver injury among CHB patients with ALT ≤ ULN.

Background: Antiviral therapy is not routinely recommended for CHB patients with ALT ≤ ULN (CHB-NALT), based on current international guidelines. However, it is debatable if antiviral treatment should be offered for CHB-NALT patients, because significant liver injury is observed from liver biopsy of some CHB-NALT patients. Quantification of anti-HBc (qAnti-HBc) can predict antiviral response in CHB patients, while its role in CHB-NALT patients remains to be explored.

Aim: To determine if it is reliable that the novel non-invasive model based mainly on qAnti-HBc and other conventional biomarkers for providing objective value among CHB-NALT patients with antiviral therapy, in direct comparison with liver biopsy.

Methods: 542 or 110 liver biopsied CHB-NALT patients from 2015 to 2020 or in 2021 were included in training set or validation set. Circulating IL-1β, IL-2, IL-4, IL-12p70, IL-17, TNF and IFNα were determined in the training set. A non-invasive model was developed based on qAnti-HBc and other conventional biomarkers.

Results: Among 423/542 (78%) patients with significant liver injury in the training set, 47% were in grey-zone. Circulating IL-1β, IL-12p70, IL-17 in the CHB-NALT patients with liver injury was significantly higher than these without liver injury in the training set (p < 0.01). No significant difference of IL-1β, IL-12p70, IL-17 was observed between CHB-NALT patients with significant liver injury and active CHB with elevated ALT in the training set. There was inverse correlation between liver injury grades and IFNα, IL-4, or IL-2 in these patients (p < 0.05). Serum qAnti-HBc level was significantly higher with CHB-NALT patients with liver injury than these without in the training set (P < 0.01). ALT/ULN, AST, PLT and qAnti-HBc were identified as independent predictors for significant liver injury. Furthermore, our current model demonstrated a good performance in predicting significant liver injury, i.e. AUROCs of 0.95 or 0.86 in training set or validation set. The model cut-off value for anti-viral therapy at ≥1.471.

Conclusions: qAnti-HBc appears to be well correlated with the hepatic damage, in direct comparison with liver biopsy from CHB-NALT patients. The novel model developed seems to be reliable for predicting liver injury in CHB-NALT patients. Such model also provides objective value for decision making of antiviral therapy.