- 现金
- 62111 元
- 精华
- 26
- 帖子
- 30437
- 注册时间
- 2009-10-5
- 最后登录
- 2022-12-28
|
Is Tenofovir Superior to Entecavir in Reducing the Risk of Hepatocellular Carcinoma in Chronic Hepatitis B? The Controversy Continues
Pietro Lampertico∗,'Correspondence information about the author Pietro LamperticoEmail the author Pietro Lampertico
CRC “A. M. and A. Migliavacca” Center for the Study of Liver Disease, Division of Gastroenterology and Hepatology, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano, Milan, Italy
George V. Papatheodoridis
Department of Gastroenterology, Medical School of National and Kapodistrian University of Athens, General Hospital of Athens “Laiko”, Athens, Greece
PlumX Metrics
DOI: https://doi.org/10.1053/j.gastro.2019.11.007 |
showArticle Info
Abstract
Full Text
References
-
See “Tenofovir is associated with lower risk of hepatocellular carcinoma than entecavir in patients with chronic HBV infection in China,” by Yip TC-F, Wond VW-S, Chan HL-Y, et al, on page 215.
The article by Yip et al1 published in this issue of Gastroenterology fuels the discussion on whether different nucleos(t)ide analogues may modulate the risk of hepatocellular carcinoma (HCC) in chronic hepatitis B (CHB). This large retrospective territory-wide cohort study that included 29,350 patients with CHB who started entecavir (ETV) or tenofovir disoproxil fumarate (TDF) in Hong Kong between 2008 and 2018 has several strengths, such as a large sample size, relevant good follow-up, sophisticated statistical analysis, baseline assessment of liver disease severity and all major virologic variables, 1-year virologic and biochemical responses and determination of several HCC risk scores. In contrast, study findings must be taken with caution for several reasons. The 2 groups differed significantly in several well-known strong baseline HCC risk factors; the TDF-treated patients were younger, more frequently hepatitis B e antigen positive, and less frequently male, cirrhotic, and diabetic. Only 4.5% of 29,350 patients and particularly only 1% of 3860 patients with cirrhosis received TDF as a likely consequence of the fact that this drug became available in Hong Kong years after ETV. In addition, the number of TDF-treated patients who developed HCC was very limited: only 8 of 1394 HCCs developed in TDF-treated patients (6 patients with and 2 without baseline cirrhosis), questioning the robustness of a comparative analysis with so few events. The cumulative HCC incidence in this study started to diverge very early, immediately after week 48 (but HCC diagnosed within the first 48 weeks were excluded), which is more in line with different baseline HCC risks between the 2 groups rather than with a usually more delayed different drug effect on the HCC risk.
To adjust for significant differences in baseline characteristics and HCC risk factors between the 2 groups, Yip et al1 used several statistical tools such as multivariable adjustment, propensity score (PS) weighting and matching, inverse probability of treatment weighting, and competitive risk analysis. In 1200 TDF-treated patients finally matched with ETV-treated patients, TDF was associated with significantly lower HCC risk after PS weighting and 1:5 matching (weighted subdistribution hazard ratio [HR], 0.36 and 0.39; P ≤ .016). These findings were confirmed by other subanalyses. However, sophisticated statistical methods cannot replace randomization, because they cannot completely adjust for all the differences in HCC risk factors. Even if an excellent balance can be achieved by PS, the matching will be only for identified and available rather than all possible confounders. In fact, the exact phase of chronic HBV infection and the appropriate treatment indication represent 2 related HCC risk factors2 that cannot be accurately assessed in such territory-wide studies using code based diagnosis extracted from large databases and therefore cannot be considered in any matching method.
The possible mechanisms explaining the lower HCC risk in patients with CHB treated with TDF are unclear. It has been suggested that TDF, but not ETV, may activate the interferon lambda 3 pathway.3 Interestingly, in the PS-matched cohort of the current study, virologic response rate at year 1 was significantly lower with ETV than TDF (69.7% vs 77.6%; P < .001), although the biochemical response rate was higher with ETV (71% vs 59%; P < .001). This finding is unexpected, because a virologic response is usually the major driver of biochemical response. The activation of the interferon lambda 3 pathway by TDF could justify, at least partly, the suboptimal biochemical response in TDF patients, as the persistence of mildly elevated alanine aminotransferase in patients with suppressed HBV replication could play some role in reducing HCC development, for example, by nonspecific killing or deaths of hepatocytes committed to neoplastic transformation. In contrast, observations even from the same group that elevated on-therapy alanine aminotransferase is associated with higher HCC risk complicate the interpretation of such findings.4
The results of the Hong Kong study are in line with the initial report from South Korea by Choi et al,5 suggesting a lower HCC risk with TDF (adjusted HR, 0.68), but 2 large well-performed independent studies again from South Korea failed to confirm these findings.6,7 Kim et al6 reported that the annual HCC incidence did not differ between 1484 ETV-treated and 1413 TDF-treated patients (1.92 vs 1.69 per 100 person-years); by multivariable (adjusted HR, 0.975; P = .852) or PS-matched and ITPW analyses. Similar findings were reported by Lee et al7 in a study enrolling 7015 consecutive patients with CHB, both in the entire cohort (PS-matching model HR, 1.03; P = .880) and in the subgroups of chronic hepatitis and cirrhotic patients. A third recently published study from a large international consortium of CHB also did not show any significant difference in the 5-year HCC risk between 520 paired matched patients treated with TDF or ETV.8
In addition to these 5 recently fully published studies, several as yet unpublished reports have addressed the same topic with contrasting results. In the PAGE-B cohort, which included approximately 2000 Caucasian patients, the 5-year HCC risk was 5.4% in ETV-treated and 6.0% TDF-treated patients (adjusted HR, 1.00; 95% confidence interval, 0.70–1.42).9 In a French cohort including 2658 patients, the annual HCC risk was also not significantly different between ETV-treated and TDF-treated patients (0.91 vs 0.88 per 100 person-years; adjusted HR, 1.41; 95% confidence interval, 0.65–3.03).10 At variance, a recent US study demonstrated that, after adjustment of baseline variables and PS weighting, TDF (6145 patients) was associated with significantly decreased risk of HCC compared with ETV therapy (4060 patients) (HR, 0.56; 95% confidence interval, 0.37–0.86).11
In summary, whether patients treated with TDF have a lower risk of HCC compared with those treated with ETV remains unsettled, because different studies, even from the same country, reached opposite conclusions.12,13 The only exception is Europe, where 2 large independent studies reported no difference in the HCC risk between the 2 agents. To improve the quality of the data and subsequently the validity of the results, carefully collected individual data from large cohorts of homogeneous and clinically relevant subpopulations, such as compensated cirrhotics or patients with different stages of disease severity or hepatitis B e antigen profiles, should be analyzed. Thus, investigators are encouraged to merge their cohorts, because only careful collaborative efforts could unravel this interesting clinical issue. |
|