CHB患者长期口服核苷（酸）类似物（NUCs）中HCC的残留风险 -

StephenW

Residual risk of HCC during long-term oral nucleos(t)ide analogues (NUCs) in patients with CHB – Is one NUC better than the other?
Grace Lai-Hung Wong1,2, Pietro Lampertico3,⇑
1Institute of Digestive Disease, The Chinese University of Hong Kong, Hong Kong; 2Department of Medicine, The Chinese University of Hong
Kong, Hong Kong; 3CRC ‘‘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
Antiviral therapy with oral nucleos(t)ide analogues (NUCs) is recommended
by international guidelines for patients with chronic
hepatitis B (CHB) who have treatment indications,1–3 as NUCs are
effective in suppressing HBVDNAand reducing the risk of hepatic
events and hepatocellular carcinoma (HCC).4,5 Entecavir (ETV),
tenofovir disoproxil fumarate (TDF) and tenofovir alafenamide
(TAF) are the 3 recommended NUCs, with high genetic barriers
to resistance as well as favorable safety profiles.1,3,6
In a recent issue of Journal of Hepatology, Kim and co-workers
reported the findings of a retrospective cohort study of 2,897
patients with CHB recruited from 4 academic teaching hospitals
in South Korea.7 The authors compared the 5-year cumulative
probabilities of HCC and death or liver transplantation in patients
with or without compensated cirrhosis who had received ETV
and TDF as the first antiviral agent. With several statistical
approaches, including propensity score (PS)-matched and
inverse probability of treatment weighting analyses, the authors
did not demonstrate any difference in HCC incidence nor in other
clinical outcomes.7 The yearly HCC rates were approximately
4.3% and 3.4% for ETV- and TDF-treated patients with compensated
cirrhosis (adjusted hazard ratio [aHR] 0.83; 95% CI 0.60–
1.13; p = 0.25) and 0.58% and 0.76% for the corresponding
patients without cirrhosis (aHR 1.60; 95% CI 0.95–2.71, p = 0.07).
This study adds fuel to the heated debate: is TDF better than
ETV in terms of risk reduction of HCC? An earlier report from
the same country reported different observations.8 Choi and
co-workers showed a significantly lower HCC risk in TDF-treated
patients, compared to ETV-treated patients, in a nationwide
historical population cohort study of 24,156 previously treatment-
naïve patients with CHB (11,464 started with ETV,
12,692 started with TDF), as well as a hospital cohort of 2,701
patients with CHB (1,560 ETV, 1,141 TDF).8 The authors
reported a consistent HCC risk reduction with TDF treatment,
[size=10.3591px]with HRs ranging from 0.61–0.68 in different statistical models[size=10.3591px]and cohorts. The initial publication also reported a lower risk of[size=10.3591px]all-cause mortality or transplant in TDF-treated patients (HR[size=10.3591px]0.77–0.79).[size=6.90535px]8[size=10.3591px]In their updated analysis, this difference in all-[size=10.3591px]cause mortality or transplant was no longer observed; whereas[size=10.3591px]HRs for HCC in the full TDF cohort were changed from 0.61–0.68[size=10.3591px]relative to entecavir, which remained statistically significant.[size=6.90535px]9[size=10.3591px]One may ask, why have data from the same country (and the[size=10.3591px]patients in the report by Kim[size=10.3591px]et al.[size=10.3591px]should be a subset of those[size=10.3591px]studied by Choi[size=10.3591px]et al.[size=10.3591px]) yielded such a big difference? One possi-[size=10.3591px]bility would be difference in patient selection. First, Kim[size=10.3591px]et al.[size=10.3591px]excluded HCC in the first 6 months, whereas Choi[size=10.3591px]et al.[size=10.3591px]excluded[size=10.3591px]those in the first 12 months for the hospital cohort.[size=10.3591px]Second, in Choi[size=10.3591px]et al[size=10.3591px].’s study, but not in the Kim[size=10.3591px]et al.[size=10.3591px]’s study,[size=10.3591px]patients with decompensated cirrhosis were included, a very[size=10.3591px]important limitation given the well-known high risk of HCC[size=10.3591px]for this population. Third, the nationwide cohort from Choi[size=10.3591px]did not match ETV- and TDF-treated patients for very important[size=10.3591px]baseline variables such as HBV DNA and alanine aminotrans-[size=10.3591px]ferase (ALT), well-known predictors of HCC. Fourth, the cumula-[size=10.3591px]tive HCC curves in the Choi’s paper have a specific pattern:[size=10.3591px]curves start to diverge after 2 years of therapy, which would[size=10.3591px]be biologically sound, but between year 2, 3 to year 4 not a sin-[size=10.3591px]gle HCC occurs in the TDF-treated nationwide cohort, a very sur-[size=10.3591px]prising finding for any HBV cohort. Additionally, in the hospital[size=10.3591px]cohort where all the HCC diagnosed within the first 12 months[size=10.3591px]were excluded, the cumulative incidence of HCC starts to[size=10.3591px]increase immediately after 12 months, as expected, in the ETV[size=10.3591px]group but only after approximately 16 months in the TDF[size=10.3591px]cohort. This 3 to 4-month delay cannot be easily explained.[size=10.3591px]Indeed, the 2 HCC curves remain almost parallel from this time[size=10.3591px]point to the last observation time point, at year 4. These pat-[size=10.3591px]terns were not observed in Kim’s study.[size=10.3591px]This not-yet-settled controversy was recently fueled by addi-[size=10.3591px]tional studies presented at the International Liver Congress[size=5.59288px]TM[size=10.3591px],[size=10.3591px]Vienna, Austria in April 2019. In a territory-wide cohort study[size=10.3591px]in Hong Kong,[size=6.90535px]10[size=10.3591px]Yip and colleagues compared 28,041 ETV-trea-[size=10.3591px]ted patients and 1,309 TDF-treated patients. TDF was superior[size=10.3591px]for preventing HCC as shown in multivariable Cox regression[size=10.3591px]analysis and PS-weighted with HR ranging from 0.32–0.36 but
[size=10.3591px][size=10.3591px]not in PS-matched analysis ([size=10.359068927277093px]p[size=10.3591px]= 0.06).[size=6.90535px]10[size=10.3591px]In order to demonstrate[size=10.3591px]the minimal residual bias from unmeasured confounding, 2[size=10.3591px]negative control outcomes,[size=6.90535px]11[size=10.3591px]lung cancer and acute myocardial[size=10.3591px]infarction, were adopted and illustrated the robustness of the[size=10.3591px]observations.[size=6.90535px]10[size=10.3591px]However, the fact that only 8 cases of HCC[size=10.3591px]occurred among the 1,309 TDF-treated patients compared to[size=10.3591px]1,386 HCC in the 28,041 ETV-treated patients, highlights the[size=10.3591px]fact that the 2 populations significantly differed in terms of[size=10.3591px]baseline predictors of HCC. Additionally, after PS matching the[size=10.3591px]5-year cumulative incidence of HCC was lower in the TDF com-[size=10.3591px]pared to the ETV group (1.2%[size=10.3591px]vs[size=10.3591px]. 2.3%), but the 2 cumulative[size=10.3591px]curves started to diverge immediately after week 24. This is[size=10.3591px]an unexpected finding given that all the HCC diagnosed in the[size=10.3591px]first 24 weeks were excluded per protocol and that liver car-[size=10.3591px]cinogenesis is a complex process lasting for months to years.[size=10.3591px]Different results were shown by studies performed in the US[size=10.3591px]and Europe. In a multicenter US study[size=6.90535px]12[size=10.3591px]enrolling 822 patients[size=10.3591px]with CHB, 407 treated with ETV and 415 treated with TDF, the[size=10.3591px]aHR for HCC was similar between groups (aHR 0.70; 95% CI[size=10.3591px]0.29–1.68) in the overall population, as well as in the non-Asian[size=10.3591px]and in the NUC naïve populations. The 5- and 10-year cumula-[size=10.3591px]tive incidences of HCC did not differ between ETV- and TDF-[size=10.3591px]treated patients enrolled in the PAGE-B cohort that included[size=10.3591px]approximately 2,000 Caucasian patients with CHB on long-term[size=10.3591px]oral therapy: 5% and 8% for ETV-treated patients and 6% and 9%[size=10.3591px]for TDF-treated patients (HR for TDF[size=10.3591px]vs.[size=10.3591px]ETV adjusted for PAGE-[size=10.3591px]B: 1.116; 95% CI 0.780–1.598;[size=10.359068927277093px]p[size=10.3591px]= 0.548) (G. Papatheodoridis,[size=10.3591px]personal communication).[size=10.3591px]When the rates of HCC during long-term ETV and TDF are[size=10.3591px]compared, additional important issues should be considered.[size=10.3591px]As ETV became commercially available many years before TDF[size=10.3591px]in many countries, including Asia, patients with the most severe[size=10.3591px]liver disease were preferentially treated with this drug. In addi-[size=10.3591px]tion, when TDF became available in these countries a few years[size=10.3591px]later, physicians may have prioritized patients with obesity, dia-[size=10.3591px]betes and older age to ETV rather than TDF treatment owing to[size=10.3591px]the renal and bone safety issues of the latter drug. This strategy[size=10.3591px]may have concentrated patients with additional risk factors for[size=10.3591px]HCC in the ETV cohorts. As sophisticated statistical methods can[size=10.3591px]only minimize these important baseline confounders, the inclu-[size=10.3591px]sion and analysis of homogeneous patient populations also[size=10.3591px]stratified for presence/absence of cirrhosis remain of paramount[size=10.3591px]importance.[size=10.3591px]Provided that the available studies are not concordant on[size=10.3591px]this topic, is there any biological plausibility to explain why[size=10.3591px]TDF may further reduce the risk of HCC in patients with CHB?[size=10.3591px]Virologically, Choi[size=10.3591px]et al.[size=10.3591px]reported better virological response in[size=10.3591px]the TDF-treated cohort.[size=6.905352753546971px]8[size=10.3591px]TDF may also lead to better reduction[size=10.3591px]in serum hepatitis B surface antigen (HBsAg) levels, compared[size=10.3591px]to ETV.[size=6.90535px]13[size=10.3591px]Biochemically, more TDF-patients achieved ALT nor-[size=10.3591px]malization at 1 year in Choi’s cohort.[size=6.905352753546971px]8[size=10.3591px]ALT normalization during[size=10.3591px]the first year of antiviral therapy is associated with lower risk of[size=10.3591px]HCC and other hepatic events in patients with CHB.[size=6.90535px]14[size=10.3591px]Immuno-[size=10.3591px]logically, TDF and adefovir dipivoxil, both as nucleotide ana-[size=10.3591px]logues, induced higher serum interferon lambda-3 (IFN-[size=10.359068927277093px]k[size=10.3591px]3)[size=10.3591px]levels than nucleoside analogues (lamivudine and entecavir).[size=6.90535px]15[size=10.3591px]IFN-[size=10.359068927277093px]k[size=10.3591px]has potent antitumor activity in murine models of hep-[size=10.3591px]atoma,[size=6.90535px]16[size=10.3591px]which may partly explain the lower reported HCC risk[size=10.3591px]in TDF-treated patients.[size=10.3591px]In our opinion, the observations so far are contradictory and[size=10.3591px]thereby not sufficient to lead to a widespread paradigm shift in[size=10.3591px]selecting TDF over ETV for treatment naïve patients. However,[size=10.3591px]since the residual risk of HCC during long-term oral therapy[size=10.3591px]for HBV remains the only complication affecting liver-related[size=10.3591px]survival, additional observational cohorts in homogenous[size=10.3591px]patients with data on liver disease severity, virologic response,[size=10.3591px]and adherence to surveillance are needed.[size=6.90535px]17[size=11.578px]Conflict of interest[size=10.3591px]Grace Wong has served as an advisory committee member for[size=10.3591px]Gilead Sciences and Janssen, and as a speaker for Abbott, Abbvie,[size=10.3591px]Bristol-Myers Squibb, Echosens, Furui, Gilead Sciences, Janssen[size=10.3591px]and Roche. Pietro Lampertico has served as advisor and/or[size=10.3591px]speaker for BMS, Roche, Gilead Sciences, GSK, MSD, Abbvie,[size=10.3591px]Janssen, Arrowhead, Alnylam, Eiger, Myr Pharma.

StephenW

CHB患者长期口服核苷（酸）类似物（NUCs）中HCC的残留风险 - 一个NUC比另一个好吗？
Grace Lai-Hung Wong1,2，Pietro Lampertico3，⇑
1香港中文大学消化病研究所，香港; 2中国大学医学系
香港，香港; 3CRC''A。 M.和A. Migliavacca“肝病研究中心，消化内科和肝病学，
FondazioneIRCCSCàGrandaOspedale Maggiore Policlinico，UniversitàdegliStudi di Milano，米兰，意大利
建议使用口服核苷（酸）类似物（NUC）进行抗病毒治疗
根据慢性病患者的国际指南
乙型肝炎（CHB）有治疗适应症，1-3为NUCs
有效抑制HBVDNA，降低肝脏风险
事件和肝细胞癌（HCC）.4,5恩替卡韦（ETV），
替诺福韦地索普西富马酸盐（TDF）和替诺福韦艾拉酚胺
（TAF）是3个推荐的NUC，具有高遗传障碍
阻力以及有利的安全性.1,3,6
在最近一期的“肝病学杂志”上，Kim和同事们
报道了2,897项回顾性队列研究的结果
CHB患者从4所学术教学医院招募
在韩国.7作者比较了5年的累积
HCC和死亡或肝移植患者的概率
有或没有接受ETV的代偿性肝硬化患者
和TDF作为第一种抗病毒药物。有几个统计
方法，包括倾向得分（PS） - 匹配和
作者，加权分析的反向概率
没有表现出HCC发病率和其他发病率的任何差异
临床结果.7年度HCC发病率约为
ETV和TDF治疗患者的补偿率为4.3％和3.4％
肝硬化（校正风险比[aHR] 0.83; 95％CI 0.60-
1.13;对应的p = 0.25）和0.58％和0.76％
没有肝硬化的患者（aHR 1.60; 95％CI 0.95-2.71，p = 0.07）。
这项研究为激烈的辩论增添了动力：TDF是否优于
ETV在降低HCC风险方面？早先的一份报告来自
同一个国家报告了不同的观察结果.8 Choi和
同事在TDF治疗中显示出显着较低的HCC风险
在全国范围内，与ETV治疗的患者相比，患者
历史人口队列研究24,156先前的治疗 -
幼稚的CHB患者（11,464名患有ETV，
12,692起始于TDF），以及一个2,701的医院队列
CHB患者（1,560 ETV，1,141 TDF）.8作者
报道了TDF治疗可持续降低HCC风险，在不同的统计模型中，HR为0.61-0.68
和同伙。最初的出版物也报道了较低的风险
TDF治疗患者的全因死亡率或移植率（HR
0.77-0.79）.8在他们更新的分析中，这种差异在所有情况下都有
死亡或移植不再被观察到;而
完整TDF队列中HCC的HR从0.61-0.68变为
相对于恩替卡韦，其仍具有统计学意义.9
有人可能会问，为什么有来自同一个国家的数据（和
Kim等人的报告中的患者。应该是那些的子集
由Choi等人研究过）产生了如此大的差异？一种可能性
将是患者选择的差异。首先，Kim等人。
在前6个月排除了HCC，而Choi等人。排除
那些在医院队列的前12个月。
第二，在Choi等人的研究中，但没有在Kim等人的研究中，
包括失代偿期肝硬化患者，非常
鉴于众所周知的HCC高风险，这是一个重要的局限
对于这个人口。第三，来自Choi的全国性队列
与ETV和TDF治疗的患者不匹配非常重要
基线变量如HBV DNA和丙氨酸氨基转移酶
（ALT），众所周知的HCC预测因子。四，累计
Choi论文中的HCC曲线有一个特定的模式：
经过2年的治疗，曲线开始出现分歧
生物学上是合理的，但是在第2年，第3年到第4年之间没有一个
HCC发生在TDF治疗的全国性队列中，非常令人惊讶
寻找任何HBV队列。另外，在医院里
所有HCC在前12个月内确诊的队列
被排除在外，HCC的累积发病率开始下降
如预期的那样，在ETV中12个月后立即增加
但是只有在TDF中大约16个月之后才会出现这种情况
队列。这3到4个月的延迟不容易解释。
实际上，从这时起，2个HCC曲线几乎保持平行
指向最后一个观察时间点，在第4年。这些模式
在Kim的研究中没有观察到。
这个尚未解决的争议最近得到了额外的推动
在国际肝病大会上提交的研究报告，
2019年4月在奥地利维也纳举行的一项全港性队列研究
在香港，10叶和同事比较了28,041例ETV治疗
患者和1,309例TDF治疗的患者。 TDF优越
用于预防HCC，如多变量Cox回归所示
分析和PS加权，HR范围从0.32-0.36但是不是在PS匹配分析中（p = 0.06）.10为了证明来自未测量混杂的最小残余偏差，采用了2个阴性对照结果，11个肺癌和急性心肌梗塞，并说明了观察的稳健性.10然而，仅有8例1,309例TDF治疗患者中HCC发生率与28,041例ETV治疗患者中的1,386例HCC相比，突出显示2例人群在HCC基线预测因子方面存在显着差异。此外，PS匹配后，与ETV组相比，TDF的5年累积发生率在TDF中较低（1.2％对2.3％），但2周累积曲线在第24周后立即开始出现分歧。这是一个意想不到的发现所有在第24周诊断出的HCC都被排除在每个方案之外，并且肝癌发生是一个持续数月到数年的复杂过程。在美国和欧洲进行的研究显示了不同的结果。在一项多中心美国研究中，报告了822例CHB患者，407例接受ETV治疗，415例接受TDF治疗，HCC的aHR在整体人群中均为相似（aHR 0.70; 95％CI0.29-1.68），以及非亚洲人和NUC天真的人口。 HCC的5年和10年累积发生率在参加PAGE-B队列的ETV和TDF治疗患者之间没有差异，其中包括大约2,000名白种人CHB患者进行长期治疗：5％和8％ ETV治疗的患者和TDF治疗患者的6％和9％（TDFvs.ETV的HR调整为PAGE-B：1.116; 95％CI 0.780-1.598; p = 0.548）（G。Papatheodoridis，个人通讯）。当在长期ETV和TDF期间HCC的比率进行比较，应考虑其他重要问题。由于ETV在TDF之前多年商业化，在包括亚洲在内的许多国家，患有最严重疾病的患者优先用该药物治疗。另外，当TDF在几年后在这些国家上市时，由于后一种药物的肾脏和骨骼安全问题，医生可能优先考虑肥胖，糖尿病和年龄较大的患者，而不是TDF治疗。该策略可能使患者在ETV队列中具有额外的HCC风险因素。由于复杂的统计方法可以最大限度地减少这些重要的基线混杂因素，对肝硬化存在/不存在进行分类的同质患者群体的包含和分析仍然具有重要性。提供的可用研究与本主题不一致，是否有任何生物学合理性来解释为什么TDF可能进一步降低CHB患者的HCC风险？病毒学上，Choiet报道了TDF治疗组的更好的病毒学应答.8TDF也可以导致更好的降低血清乙型肝炎表面抗原（HBsAg）水平，与ETV相比.13生物化学，更多TDF患者在Choi's队列中1年达到ALT正常化.8抗病毒治疗第一年的PRALT正常化与CHB患者的肝癌和其他肝脏事件风险降低相关.14免疫，TDF和阿德福韦酯，两者均为核苷酸分析，比核苷类似物诱导更高的血清干扰素λ-3（IFN-k3）水平es（拉米夫定和恩替卡韦）.15IFN-k在小鼠肝脏模型中具有有效的抗肿瘤活性[16]，这可能部分解释了TDF治疗患者报告的HCC风险较低。我们认为，迄今为止的观察结果是矛盾的，并且不足以引导对于治疗初治患者而言，TDV优于ETV的广泛范式转变。然而，由于HBV长期口服治疗期间HCC的残余风险仍然是影响肝脏相关生存的唯一并发症，因此需要在肝病患者严重程度，病毒学应答和监测依从性的同源患者中进行额外的观察性队列.17感兴趣的Grace Wong已服务 作为Gilead Sciences和Janssen的顾问委员会成员，以及Abbott，Abbvie，Bristol-Myers Squibb，Echosens，Furui，Gilead Sciences，Janssenand Roche的演讲嘉宾。 Pietro Lampertico曾担任BMS，Roche，Gilead Sciences，GSK，MSD，Abbvie，Janssen，Arrowhead，Alnylam，Eiger，Myr Pharma的顾问和/或顾问。