Longitudinal effects of Nucleos(t)ide analogue therapy in chronic hepatitis B patients and the utility of non-invasive fibrosis markers during treatment: A single-center experience for up to 17 years

Abstract

Background: Fibrosis regression has been associated with nucleoside analogue (NA) treatment in chronic hepatitis B (CHB) patients. Although non-invasive fibrosis markers have been evaluated in CHB, their utility for monitoring on-treatment histologic regression has not been evaluated.

Aims: To characterize improvements in disease severity and the utility of non-invasive biomarkers in CHB NA treated patients.

Methods: Histology, labs, AST-to-platelet ratio index, and Fibrosis-4 (Fib-4) from treatment-naïve CHB patients were evaluated at baseline and longitudinally. Relative change from baseline to various time points during treatment were evaluated. Correlative analysis of APRI and Fib-4 with histology was performed longitudinally.

Results: 80 CHB patients (84% male, median age 45 (IQR 32, 54)) with histology up to 17 years (median 6(IQR 3.9, 8.0)) years were studied. Median baseline Ishak fibrosis was 3 (IQR 2, 4), histologic activity index (HAI) inflammation was 9 (IQR 7, 11), and AUROC of fibrosis markers for detecting cirrhosis (Ishak ≥ 5) was >0.64. HAI improved at a rate of 54% during year 1 and 37% in year 2, both greater than in the remaining follow-up periods. Within the first year, fibrosis improved by 35%, greater than all other time periods. Non-invasive biomarkers began to correlate with histology beyond 4 years (APRI: 4-6 years: r = 0.33, p = 0.03; ≥6 years: r = 0.41, p = 0.009; Fib-4: ≥6 years: r = 0.35, p = 0.03).

Conclusion: Early dynamic changes in histology occur in CHB patients on NA followed by linear improvements. Non-invasive fibrosis biomarkers do not capture these dynamic changes and may demonstrate clinical utility beyond 4 years of treatment.

Keywords: Fibrosis; Hepatitis B; Inflammation; Non-invasive markers of fibrosis; Nucleoside analogues.

Figure 1.. Receiver Operating Characteristic Curves for Baseline Performance of APRI and Fib-4.

Both Fib-4 and APRI perform adequately at baseline for detecting cirrhosis (Ishak ≥ 5). Areas under the curve are indicated in parenthesis in the figure legend.

Figure 2.. Cumulative Rate of Histologic Change from Baseline.

Data points demonstrate the cumulative changes from time-category 1 to time-category 5. Slopes (shown in table) between data points indicate adjusted means of absolute changes and are in Ishak units of change or HAI units of change where appropriate. Fibrosis (red line) dramatically improves during the first year of treatment and becomes more linear from year 2 onwards. Improvements in inflammation (blue line) are more dynamic during the first two years of treatment followed by a more linear pattern in the remaining years

Figure 3.. Cumulative Time-adjusted Rate of Fibrosis Changes by Ishak and Non-invasive Markers.

Data points demonstrate the cumulative time-adjusted changes from time-category 1 to time-category 5. Slopes (shown in table) between data points indicate time-adjusted means of relative changes (TARC values). “SIG” and “NS” indicate the p-value of a repeated measures mixed model. “SIG” indicates that a difference did exist among the time categories and “NS” indicates that there was no difference. Supplemental table 1 indicates which pairwise comparisons of time categories were significantly different. While Ishak (red line) shows dynamic time-adjusted improvement in the first year of treatment, similar to APRI and Fib-4, the rates of change are not correlated. APRI and Fib-4 show greater change during both time-categories 1 (0–1 years of treatment) and 2 (1–2 years of treatment).

Figure 4.. Cumulative Time-adjusted rate of Inflammation Changes by Histologic Activity Index, ALT, and AST.

Data points demonstrate the cumulative time-adjusted changes from time-category 1 to time-category 5. Slopes (shown in table) between data points indicate time-adjusted means of relative changes (TARC values). “SIG” and “NS” indicate the p-value of a repeated measures mixed model. “SIG” indicates that a difference did exist among the time categories and “NS” indicates that there was no difference. Supplemental table 4 indicates which pairwise comparisons of time categories were significantly different. ALT and AST appear to mirror improvements in HAI during the first two years of treatment. While ALT and AST continue to decrease in later years, they do not decrease as dramatically as HAI scores.