Earlier and more frequent virological and clinical relapse after cessation of tenofovir disoproxil fumarate versus entecavir therapy: Results from a large, global, multi-ethnic cohort of chronic hepatitis B patients (RETRACT-B study)

Abstract

Background and aims: Whether tenofovir disoproxil fumarate (TDF) and entecavir (ETV) differentially affect relapse across different patient subpopulations following treatment discontinuation remains unclear. We aimed to study the association between drug type and relapse in a global multi-centre cohort of chronic hepatitis B (CHB) patients who discontinued either TDF or ETV therapy. Method: We investigated outcomes in CHB patients who stopped either TDF or ETV therapy between 2001 and 2020 from 12 participating centres across North America, Europe, and Asia. All included patients were HBeAg-negative at discontinuation. Rates and risk of virological (VR; HBV DNA ≥2000 IU/ml) and clinical (CR; HBV DNA ≥2000 IU/ml and ALT ≥2 × ULN) relapse were analyzed using multivariable Cox regression stratified by site, adjusting for age, sex, race, total treatment duration, treatment history, HBeAg status at the start of therapy (SOT), and HBsAg level at the end of therapy (EOT). Results: Of 1386 CHB patients included, 416 had been treated with TDF and 970 with ETV. No significant differences in age, sex, and EOT HBsAg levels and no proportional differences in cirrhosis and HBeAg status at SOT were observed between the two groups. The TDF group had fewer Asian patients (81% vs 94%), more patients previously treated with other NAs (27% vs 11%) or (peg-)interferon (10% vs 6%), and higher median ALT at EOT compared with the ETV group (0.7 vs 0.5 × ULN); all p < 0.05. Overall, 995/1386 (72%) and 561/1386 (41%) patients experienced VR and CR, respectively, during a median offtreatment follow-up of 17 months; 838/995 (84%) VRs and 354/561 (63%) CRs occurred within 12 months of cessation. TDF vs ETV was associated with a higher rate of VR in the first 12 months overall (hazard ratio [HR]: 2.0, p < 0.01) and across different races and HBeAg status at SOT; the difference became negligible over time (HR: 0.79, p = 0.37) (Fig.). Among Asian patients, the risk of VR associated with TDF was further increased in males (p < 0.05) and those with HBsAg ≥1000 IU/ml (p < 0.01). TDF was also associated with consistently higher rates of CR compared with ETV throughout follow-up (Fig.). TDF was an independent predictor of CR (HR: 1.63, p < 0.01), adjusting for the factors aforementioned. Conclusion: VR was very frequent after ETV or TDF treatment discontinuation. VR and CR were observed earlier and more frequently in individuals treated with TDF vs ETV, adjusting for race, age, sex, treatment history, HBeAg status at SOT, and HBsAg levels at EOT. Only among TDF-treated Asian patients, the risk of VR was further increased with male sex and/or HBsAg ≥1000 IU/ml at EOT