We appreciate the insightful comments from Dr. Su and his colleagues. The age, distribution of sex, hepatitis C virus genotype, and cirrhosis status did not differ between patients who were enrolled for analysis of aspartate aminotransferase-to-platelet ratio index (APRI) at 6 months after end-of-treatment (M6) and those who were excluded because of no APRI data available at M6.1
The benefits of preventing disease progression in chronic hepatitis C (CHC) patients who did not respond to interferon, as observed in earlier studies,2 no longer existed in a longer observation period.3 Achievement of a sustained virologic response (SVR) could reduce risk of hepatic complications, regardless of treatment regimen.4, 5 Thus, the heterogeneity of treatment population would not influence the results of our study. Because we looked at the therapeutic effect of interferon-based therapy for CHC on long-term outcome, per-protocol population was selected for analysis. Patients who terminated early from prescribed interferon-based therapy usually are grouped into nonresponders in intention-to-treat analysis. Nevertheless, an SVR might be achieved in patients who terminated early from 24 weeks of peginterferon/ribavirin therapy.6 The long-term outcome of the subgroup needs further evaluation.
The major limitation of the current study was the nonrandomized controlled, retrospective-prospective design. However, it would be unethical to withhold antiviral therapy from CHC patients. Thus, definitive proof of the effectiveness of antiviral therapy in preventing disease progression is unlikely to come from prospective controlled trials. Our treated group was older and had a higher proportion of patients with cirrhosis than the untreated group did. Because age and cirrhosis are 2 major risk factors for liver disease progression,4 the effect of antiviral treatment in reducing hepatic complications would not be overestimated in this study. Using Cox proportional hazards regression analysis, we could control the confounding factors and reduce the potential bias between the 2 groups.
The annual incidence of mortality and cumulative survival rate were similar between untreated and interferon-treated patients with a 15-year follow-up, even considering that the interferon-treated patients were much older and more cirrhotic than the untreated ones.1, 4 However, the cumulative survival rate was significantly higher in sustained responders than in nonresponders and untreated patients, but did not differ between nonresponders and untreated patients.4 Preexisting cirrhosis and treatment failure, instead of antiviral therapy, were independent predictors for patient death. The results further emphasized the importance of achieving an SVR in reducing risk of hepatic complication.
There were 2 reasons why patients with hepatocellular carcinoma (HCC) in the untreated group survived more than those in the treated group at the end of follow-up. First, the outcomes of patients were determined based on the occurrence of hepatic complications at the last visit before the time of analysis in both cohorts. Search for death registry was not performed for patients lost to follow-up. Second, in the untreated controls, 2 of the 4 patients who died of hepatic decompensation were concomitant with small HCC development. Thus, we assigned their cause of death as hepatic decompensation instead of HCC.
Because regression of hepatic fibrosis was estimated about 0.28 units per year in sustained responders,7, 8 it is not surprising that fibrosis stage may not significantly change at M6 when using the scoring systems of semiquantitative assessment with the proposed numeric unit of 5-7 scales.9–11 It is hard to discriminate relatively small but significant differences of hepatic fibrosis by using current scoring systems. Among our sustained responders, the platelet count rose significantly from baseline to M6 (182 ± 58 × 109/l versus 187 ± 60 × 109/l, P < 0.0001, paired t test). By contrast, the platelet counts at baseline and M6 were similar (166 ± 64 × 109/l versus 165 ± 62 × 109/l) among nonresponders. Thus, the platelet component of APRI may more closely reflect the ongoing change in hepatic fibrosis after interferon-based therapy than the present proposed scoring systems. Taken together, the APRI determined after interferon-based therapy may not only correlate well with SVR and sustained biochemical response, but also with the slow, but significant regression of hepatic fibrosis. All the factors have been important factors associated with long-term outcome of liver disease.4, 7
Because patients with cirrhosis with and without an SVR are at risk of HCC, universal close follow-up is recommended in clinical practice and may overcome the limitation of APRI-M6 prediction for the subgroup. Nevertheless, with optimized cut-off values of APRI-M6, we could identify patients at risk for HCC development in low-risk groups (sustained responders and patients without cirrhosis). These results could provide decision-making information for scheduling follow-up intervals to detect HCC early.