Liver fibrosis is the most critical factor directly associated with patient prognosis. Advanced liver fibrosis is the leading risk factor of HCC and can lead to fatal conditions, including decompensated cirrhosis and liver failure. Liver fibrosis was previously considered irreversible; however, recent reports have revealed that fibrosis can be ameliorated by therapeutic intervention that addresses the cause of chronic liver disease [24,25,26,27].

This study is the first to demonstrate an improvement in histological fibrosis over a substantially long period (a median of 60 months from EOT) by comparing histological evaluation before and after treatment with DAA. Furthermore, improvement in histological fibrosis was observed in patients with and without advanced fibrosis at baseline. Three previous studies have evaluated changes in histological fibrosis after DAA treatment [18,19,20]. However, these studies included a small number of patients (9–40) and performed histological evaluation relatively quickly after SVR, with median durations of 6 months [18], 10 months [19], and 19.2 months [20]. Two of these studies showed no improvement in histological fibrosis after DAA treatment [19, 20]; however, the remaining one reported that 38% (15/40) of patients had improved their Ishak fibrosis score by > 1 point at a median 6-month histological evaluation after SVR compared to that before SVR [18]. Conversely, improvement in fibrosis after SVR to IFN treatment has been reported over long-term intervals of approximately 5 years [8,9,10]. Furthermore, Pockros et al. suggested that better histological fibrosis improvement may be associated with a longer period of viral suppression after IFN treatment [28]. Therefore, the lack of improvement in histological fibrosis after DAA treatment in previous reports may have been due to the short observation period. Our findings showed that viral elimination after treatment with DAA improved histological fibrosis at 5 years after EOT.

In the current study, we also assessed changes in histological fibrosis according to the treatment regimen. The IFN-free and IFN-based DAA groups exhibited significant improvement in histological fibrosis. The proportion of HCV genotypes was significantly different between the two groups; however, the improvement in histological fibrosis was significant in both groups, even in patients with HCV genotype 1 (IFN-based DAA group; n = 23, IFN-free DAA group n = 39; Supplementary Figure S5). These results suggest that the anti-inflammatory effects of DAA treatment associated with HCV elimination contribute to the improvement of fibrosis, even without the antifibrotic effects of IFN [29].

Moreover, the finding that 12 patients (16.9%) had worsening Ishak scores is clinically important. Supplementary Tables S8 and S9 show the characteristics at baseline and 5 years after SVR in patients with and without histologic fibrosis progression after achieving SVR. The BMI and number of patients with habitual alcohol intake were significantly higher in patients with histological fibrosis progression than in those without histological fibrosis progression at baseline and 5 years after SVR. Moreover, AST, ALT, AFP, HA, 4COL7S, M2BPGi, and the FIB-4 index were significantly lower in patients with histological fibrosis progression than those without at baseline. At the same time, these factors showed either no significant differences or comparable levels even if they had a significant difference 5 years after the EOT. These findings suggest that a high BMI and alcohol consumption may influence the progression of histological fibrosis after SVR. Therefore, careful follow-up is desirable for patients with such background characteristics. However, since the number of patients with these factors was small in this study, further analysis involving a large cohort is needed to substantiate these findings.

Another strength of our study is the simultaneous assessment of liver fibrosis using histological evaluation and NITs at baseline and 5 years after the EOT with antiviral treatment. The use of NITs for the evaluation of fibrosis after SVR has not been established. The EASL guidelines state that noninvasive methods are unreliable after SVR and should not be used to assess the fibrosis stage after treatment [30]. The problem with evaluating the fibrosis stage after SVR using NITs is that improved NIT measurements may reflect improved inflammation due to HCV elimination. Therefore, to address this issue, NITs and histological evaluation should be performed simultaneously to confirm the diagnostic ability of NITs in assessing the fibrosis stage before and after SVR. Consistent with previous reports [15,16,17], all fibrosis biomarkers in our study had significantly decreased levels after treatment with DAA. Except for 4COL7S, all NITs could predict advanced-stage liver fibrosis and cirrhosis after SVR, with AUROCs of > 0.7 and > 0.75, respectively. Notably, M2BPGi (AUROC > 0.85) and LSM (AUROCs > 0.95) were exceptionally useful in predicting liver fibrosis after SVR. Although LSM is useful for diagnosing fibrosis after SVR, it is important to determine whether LSM at 5 years after EOT reflects only fibrosis without inflammation or includes inflammation. LSM showed no tendency to increase with increasing mHAI score 5 years after the EOT (P = 0.23) (Supplementary Figure S6). Supplementary Table S10 presents the characteristics of patients with and without high LSM (≥ 6.75 kPa). There was no difference between the two groups in terms of alcohol intake and metabolic factors, such as fatty liver and diabetes mellitus. We excluded patients with other chronic liver diseases, such as autoimmune hepatitis, primary biliary cholangitis, and drug-induced liver injury. Considering these results, LSM at 5 years after SVR mainly reflects fibrosis due to hepatitis C.

Furthermore, our results showed a greater decrease in the cut-off values of all fibrosis biomarkers for advanced-stage fibrosis after SVR than before treatment. Similar results were reported in a previous study on IFN treatment. Tachi et al. reported that fibrosis indices, namely the AST-to-platelet ratio index (APRI), FIB-4 index, and Forns index, can predict liver fibrosis even after SVR-to-IFN treatment, with an AUROC of ≥ 0.8 [31]. Significant differences in the cut-off values of these indices, especially APRI, were observed between the biopsy performed before and after treatment [31]. In our study, LSM was measured only after 5 years of EOT, and we could not evaluate changes in cut-off values before and after SVR. However, the cut-off value of 6.75 kPa for identifying advanced fibrosis after 5 years of EOT is equivalent to that for mild fibrosis (METAVIR fibrosis stage F0-1) in previous reports [32]. Therefore, serum fibrosis markers and elastography are promising for evaluating the long-term severity of liver fibrosis after SVR to DAA treatment. Notably, the underestimation of fibrosis might be attributed to the use of the conventionally reported cut-off value of NITs. Further validation is required to avoid the underestimation of fibrosis using NITs, including the establishment of appropriate cut-off values for NITs in patients with HCV after achieving SVR for DAA treatment.

The long-term course of NIT measurements after SVR is another notable finding of this study. In our results, serum HA and FIB-4 index decreased rapidly in their measurements until 1 year after the EOT, after which they remained stable or gradually decreased. Notably, several studies examining the long-term course of NITs before and after antiviral treatment for HCV have also reported a rapid decline in NITs over a short period (approximately 1 year after SVR), followed by a stable or gradual decline [17, 33, 34]. Thus, we speculate that the short-term decrease in NITs after DAA treatment reflects a reduction in inflammation, and the subsequent sustained decrease in NITs may reflect the degree of improvement in histological fibrosis. However, it is impossible to determine when NIT measurements accurately reflect histological liver fibrosis improvement after DAA-induced SVR. Additionally, previous reports have been inconsistent in their conclusions regarding the short-term improvement in histological liver fibrosis after DAA-induced SVR [18,19,20]. Therefore, repeated liver biopsies and NIT measurements after DAA-induced SVR might be necessary to elucidate these issues; however, they lack feasibility because of the invasiveness of liver biopsies.

Hepatic steatosis is a histological feature of HCV infection, and HCV is associated with the development of hepatic steatosis. However, the frequency at which it develops varies with viral genotype. In patients with HCV genotype 3, the virus is directly involved in the development of the fatty liver, and the frequency of hepatic steatosis complications is high. Moreover, achieving viral elimination with IFN treatment decreases histological steatosis in patients with HCV genotype 3 [35, 36]. However, in patients with HCV genotype 1–2, the development of hepatic steatosis is mainly associated with host factors. Therefore, viral elimination does not improve hepatic steatosis. A histological study after IFN treatment showed no change in fatty liver in patients with HCV genotype 1, regardless of treatment response [35]. Conversely, previous studies that evaluated hepatic steatosis after SVR to DAA treatment using NITs have reported inconsistent results with contradictory conclusions [37,38,39,40,41]. In the current study, hepatic steatosis did not improve over the long term after treatment with DAA. This may be due to the selection bias of our cohort, which included only patients with HCV genotype 1/2, in whom hepatic steatosis development is mainly associated with host factors. Previous studies have reported that increased BMI, concomitant diabetes mellitus, and human patatin-like phospholipase domain-containing 3 gene polymorphism were risk factors for progressive steatosis after DAA-induced SVR [42].

Unfortunately, this study lacked data on changes in glycated hemoglobin levels after SVR and host genetic factors, and we were unable to assess factors contributing to advanced steatosis.

This study has some limitations. First, given the invasiveness of liver biopsy, few patients consented to post-treatment liver biopsy compared to those treated with DAA, likely introducing patient selection bias. Second, although some authors have reported that liver specimens are adequate if they are > 15 mm and contain six or more portal tracts [19], this study included 21 liver specimens with tissue sections ≥ 10 mm and 3–5 portal tracts at histological evaluation 5 years after SVR. The size of the tissue specimen, especially 5 years after SVR, was influenced by our institution’s practice of using a single puncture with a 16-gauge needle for liver biopsy examinations to minimize the bleeding risk. The evaluation of fibrosis and inflammation in this study using specimens with a tissue length > 10 mm and 3–5 portal tracts might be less comprehensive compared to that in previous studies using larger tissue samples. Third, only a small number of patients with cirrhosis (12 cases, 16.9%) were enrolled. Therefore, we might have failed to thoroughly investigate whether HCV elimination improved fibrosis in patients with pre-treatment cirrhosis. Finally, we did not perform pre-treatment elastography-based LSM. We could not evaluate the correlation between pre- and post-treatment changes in liver stiffness and histological fibrosis. A larger prospective study is needed to confirm our findings. However, it may be difficult to achieve given the disadvantages of liver biopsy including its invasive nature.

In conclusion, long-term follow-up after SVR to treatment with DAA showed improvement in histological liver fibrosis. NITs also helped predict the extent of liver fibrosis after SVR to treatment with DAA. However, their cut-off values need to be redefined to avoid underestimating the degree of liver fibrosis after SVR.