Longitudinal real-world study on estimated glomerular filtration rate (eGFR) changes in entecavir (ETV) versus tenofovir disoproxil fumarate (TDF)-treated chronic hepatitis B (CHB) patients: A REAL-B study

Abstract

Background & aims: Decreased renal function is sometimes observed in CHB patients maintained on long-term oral antiviral therapy, and it is controversial whether TDF is more renal toxic than ETV. We aimed to compare the longitudinal eGFR changes of ETV vs. TDF patients and to identify factors associated with eGFR changes. Methods: We performed a retrospective study of 4983 adult treatment-naïve CHB patients who were initiated on TDF (n=1790) or ETV (n=3193) then observed for ≥12 months at 22 centers from the US, HK, Korea, Taiwan, Japan and Mainland China. We assessed eGFR (mL/min/1.73m2) using the CKD-EPI formula at baseline and about every 6 months. ETV and TDF patients were balanced via propensity score matching (PSM) on age, gender, DM, HTN, cirrhosis, baseline eGFR, and follow-up duration. We used multivariable generalized linear modeling (GLM) to adjust mean eGFR during follow up for cirrhosis, HTN, and DM, and multivariable Cox regression to identify factors associated with decreased renal function by at least one CKD stage as per KDIGO staging (eGFR >90 for stage 1, 60-89 stage 2, 45-59 stage 3a, 30-44 stage 3b, and <30 stage 4/5). Results: In the total cohort (mean age 48.9 years, 66.7% male), DM, HTN, and cirrhosis were respectively present in 13.1%, 21.2%, and 43.9% of the ETV group as compared to 10.7%, 17%, 29.7% for the TDF group (P<0.05). The baseline eGFR was higher for the TDF vs. ETV group (76.9 vs. 74.1, P=0.018). PSM yielded 1406 pairs of ETV or TDF patients with baseline eGFR ≥60 and 357 pairs for the eGFR<60 group. Multivariable GLM analysis of the total (unmatched) cohort as well as the matched eGFR ≥60 and matched eGFR <60 cohorts revealed lower adjusted mean eGFRs in TDF (vs. ETV) patients (all P<0.01; Figures 1A-C), though the mean eGFR difference was small in all comparisons (1.2-3.1 mL/min/1.73m2). Among PSM eGFR≥60 patients, the 5-year cumulative incidence of renal impairment were 45.08% for ETV and 49.8% for TDF (P=0.0039; Figure 1D). In multivariable Cox regression, TDF (HR 1.41), higher baseline eGFR (HR 0.99), older age (HR 1.02), male (HR 2.16), presence of DM or HTN (HR 1.4), and higher FIB-4 score (HR 1.02) were all associated with worsening renal function by at least one CKD stage. Conclusion: TDF was associated with lower mean eGFR throughout the 5-year on-treatment follow-up and independently associated with incident renal impairment, suggesting cautious use especially in those at high risk for renal injury