The review article by Lai and Yuen provided an overall conclusion consistent with that of various meta-analyses or review articles. However, some relevant studies included in two meta-analyses[2, 3] were not mentioned or discussed. In addition, several issues require clarification and discussion.
The authors commented on the two studies of interferon-alpha (IFN-α) therapy by Lin et al.[4, 5] Actually, all 116 patients in the earlier randomized, control trial (RCT) were included in the later matched control study involving a total of 466 patients. The earlier study was therefore correctly excluded from the meta-analyses.[2, 3] It is not surprising that data may differ as the patients increased from a single center to three centers with a much greater number of patient and greater statistical power. This may explain the discrepancies in data between the two studies and why the reduction of cirrhosis became significant in the large matched control study. For hepatocellular carcinoma (HCC) development, it was clearly described in the RCT that 4 of the 29 untreated patients without cirrhosis developed cirrhosis and 2 of them progressed to HCC, that is, 2 of 29 developed HCC, whereas none of the treated counterparts did. Because patients with advanced fibrosis or cirrhosis are more prone to develop HCC, it requires shorter duration to demonstrate HCC development in this patient population. Conceivably, the effect of reducing HCC will not be demonstrated in patients without cirrhosis if the follow-up duration is not long enough.
Of note is the marked difference between the two large studies from Taiwan and Hong Kong. As compared in Table 1, all patients in the Taiwan study had active hepatitis, whereas the majority in the Hong Kong study were immune tolerant patients deemed to respond poorly to IFN therapy, and significantly (P = 0.004) more females were included in the controls. For this reason, this Hong Kong study was excluded from a meta-analysis, which demonstrated significant reduction of both cirrhosis and HCC development in IFN-treated patients. Strangely, two earlier reports from the same department showed very different data.[7, 8] It is not clear whether the patients in these two earlier reports were included in the large Hong Kong study conducted by the authors of this review article, and why the rate of HCC development in untreated patients was so different between the studies (Table 1). It is also hard to understand why IFN therapy resulted in a significantly higher rate of HCC development. To explain this intriguing data, perhaps the authors should check their database to find out whether very few of their controls had cirrhosis at baseline, in contrast to 14.3% in their treated patients.
|Source Year (Reference)||No. of Patients||Male (%)||Age (Years)||ALT (U/L)||Cirrhosis (%)||Follow-up (Years)||HBeAg Seroconversion (%)c||HCC (%)||P Value|
|Lin et al. 2007||233/233||94/94||32/31a||175/187||8.1/10.7||6.8/6.1||75/52 (P = 0.031)||2.1/6.9||<0.025|
|Yuen et al. 2001||208/203||71/57 (P = 0.004)||27/28b||46/38||14.3/NA||8.9/9.0||47/46||2.9/0||<0.015e|
|Yuen et al. 1997||48/10||60||20||NA||NA||8.8/8.3||60/50||0/0||—|
|Lau et al. 1997||138/40||65/60||30/29b||NA||NA||8.8/NA||80/60 (P = 0.02)||3.6/5.0d||NSe|
The authors erroneously indicated that 31% and 39% of the treated and untreated patients in the lamivudine RTC had cirrhosis. The correct figures are 60% and 65%, respectively. For entecavir, an online HEPATOLOGY article has clearly shown that entecavir is superior to lamivudine in reducing HCC in patients with cirrhosis.