Clinical characteristics of study population

The 227 patients had a median antiviral prophylaxis duration of 52 weeks. The clinical characteristics of the entecavir and TAF groups are presented in Table 1. Of the 27 patients with hematologic malignancy, 23 were lymphoma and 20 received rituximab therapy. There was no significant difference in clinical features between entecavir and TAF groups.

Table 1 Clinical characteristics of study populationIncidence and predictors of virological relapse and clinical relapse after NA cessation

During the antiviral prophylaxis conducted with NA therapy, none of the patients experienced HBVr. Of the 227 patients with CHB, 57 in the entecavir group and 35 in the TAF group experienced virological relapse after NA cessation. The cumulative incidences of virological relapse at 13, 26, 52, and 104 weeks were 2.1%, 16.7%, 31.3%, and 37%, respectively, in the entecavir group, and 20.5%, 28.9%, 41.9%, and 46.7%, respectively, in the TAF group. There was borderline significant difference in virological relapse rate was detected between the entecavir and TAF groups (P = 0.053, Fig. 1A).

Fig. 1

Cumulative incidences of HBV relapse after cessation of antiviral prophylactic entecavir and tenofovir alafenamide (TAF) therapy for all patients. A Virological relapse B clinical relapse

Of the 227 patients, 16 in the entecavir group and 15 in the TAF group experienced clinical relapse after NA cessation. The cumulative incidences of clinical relapse at 13, 26, 52, and 104 weeks were 0%, 1.4%, 6.3%, and 10.4%, respectively, in the entecavir group, and 6%, 13.3%, 17.4%, and 19.5%, respectively, in the TAF group. There was a significant difference in clinical relapse between the entecavir and TDF groups (P = 0.021, Fig. 1B).

The univariate and multivariate analysis results for virological and clinical relapse are presented in Tables 2 and 3, respectively. The multivariate analysis revealed that the use of hematologic malignancy, TAF use, and high-viremia group at baseline were independent risk factors for virological relapse, and use of rituximab, TAF use, higher FIB-4 index and high-viremia group at baseline were independent risk factors for clinical relapse.

Table 2 Factors predictive of virological relapse after entecavir or TAF cessation for all patientsTable 3 Factors predictive of clinical relapse after entecavir or TAF cessation for all patientsComparison of virological and clinical relapse rates between entecavir and TAF groups based on baseline HBV viral load

Because HBV DNA level at baseline was a significant factor for virological or clinical relapse, the patients were divided into 3 subgroups on the basis of their baseline HBV DNA levels; Groups I (high-viremia group), II (moderate-viremia group), and III (low-viremia group) comprised the patients with HBV DNA levels of > 2000 IU/mL, ≧ 20 and < 2000 IU/mL, and < 20 IU/mL, respectively.

In the Group I, the cumulative incidences of virological relapse at 26 and 52 weeks were 27.7% and 57.4%, respectively, for the entecavir group, and 56.7% and 75.6%, respectively, for the TAF group. A significant difference was detected in the virological relapse rates between the entecavir and TAF groups in Group I (P = 0.006, Fig. 2A). The cumulative incidences of clinical relapse at 26 and 52 weeks were 2.1% and 12.9%, respectively, for the entecavir group, and 23.5% and 30.4%, respectively, for the TAF group. A borderline significant difference was detected in the clinical relapse rates between the entecavir and TAF groups (P = 0.061, Fig. 2B). The multivariate analysis revealed that being in the TAF group was an independent factor for virological relapse (Hazard ratio [HR] 1.450; 95% confidence interval [CI]. 1.103–1.906; P = 0.008) and clinical relapse (HR, 2.095; 95% CI 1.264–3.471; P = 0.004) after adjustment for other factors.

Fig. 2

Cumulative incidences of HBV relapse after cessation of antiviral prophylactic entecavir and TAF therapy in high-viremia group (HBV DNA > 2000 IU/mL) at initial treatment (A) virological and (B) clinical relapse

In the Group II, the cumulative incidences of virologic relapse at 26, 52 and 104 weeks were 14.8%, 21.3% and 23.3%, respectively, for the entecavir group, and 13.8%, 28.2% and 33.0%, respectively, for the TAF group; the clinical relapse rates at 26, 52 and 104 weeks were 1.6%, 1.6% and 3.7%, respectively, for the entecavir group, and 10.3%, 13.9% and 13.9%, respectively, for the TAF group. In the Group II, no significant difference in virological relapse rate was detected between the entecavir and TAF groups (P = 0.383). However, there was a significant difference in clinical relapse rate between the two groups (P = 0.044). The multivariate analysis revealed that being in the TAF group was an independent factor for clinical relapse (HR, 4.972; 95% CI 1.399–17.673; P = 0.013) after adjustment for other factors.

In the Group III, the cumulative incidences of virological relapse at 26, 52 and 104 weeks were 5.6%, 13.9% and 17.1%, respectively, in the entecavir group, and 12.5%, 12.5% and 12.5%, respectively, in the TAF group. For clinical relapse, the cumulative incidences at 26, 52 and 104 weeks were 0%, 5.6% and 5.6%, respectively, in the entecavir group, and 4.2%, 4.2% and 4.2%, respectively, in the TAF group. No significance difference between the entecavir and TAF groups was detected with respect to their virological and clinical relapse rates (P = 0.964 and 0.966 for virological and clinical relapse rates, respectively).

Comparisons of virological and clinical relapse rates of entecavir and TAF groups after propensity score matching

Propensity score matching yielded 115 and 83 matched patients in the entecavir and TAF groups, respectively. No significant differences in clinical features were detected between the groups (Table 4). The patients who discontinued TAF therapy exhibited significantly higher virological (P = 0.031) and clinical (P = 0.012) relapse rates than did those who discontinued entecavir therapy (Fig. 3).

Table 4 Clinical characteristics of study population after propensity score matching for all patientsFig. 3

Cumulative incidences of HBV relapse after cessation of antiviral prophylactic entecavir and TAF Groups after propensity score matching for all patients. A virological and B clinical relapse

In the Group I, propensity score matching yielded 37 and 30 matched patients in the entecavir and TAF groups, respectively. No significant difference in clinical features was detected between the groups. The patients who discontinued TAF therapy exhibited a significantly higher rates of virological relapse (P = 0.014) and clinical relapse (P = 0.037) than did those who discontinued entecavir therapy.

The role of HBsAg level at the time of stopping entecavir or TAF therapy in HBV relapse

In this study, HBsAg data at the time of stopping entecavir or TAF therapy (end of treatment (EOT)) was available in 133 of 144 patients who received entecavir therapy and 74 of 83 patients who received TAF therapy. Of these 207 patients, there was no significant difference in HBsAg levels at EOT between entecavir or TAF groups (2.02 ± 1.16 versus 1.84 ± 1.10 log10 IU/mL, P = 0.265). HBsAg level at EOT was an independent factor for virological relapse (HR: 1.915, 95% CI 1.749–2.480, P < 0.001) and clinical relapse (HR: 2.218, 95% CI 1.182–4.163, P = 0.013) after adjusting for other factors. Furthermore, HBsAg level at EOT was an independent factor for virological relapse in entecavir or TDF subgroups, but not clinical relapse. In the high-viremia group, HBsAg at EOT was not a significant factor for virological and clinical relapse.

Hepatic decompensation upon clinical relapse

In accordance with the 2015 Asian Pacific Association for the Study of the Liver guidelines, we defined hepatic decompensation as a total bilirubin level of > 3.5 mg/dL (> 2.5 times the ULN [1.4 mg/dL]) and an increase in prothrombin time of > 3 s, or an INR of > 1.5 [7]. A patient in the entecavir group and two patients in the TAF group experienced hepatic decompensation upon clinical relapse. Three patients had lymphoma and underwent rituximab therapy (two patients had HBV DNA < 2000 IU/mL and one patient had HBV DNA > 2000 IU/mL at baseline). Of the three patients, one developed acute on chronic liver failure upon clinical relapse and died even though timely retreatment.