Hepatocellular Carcinoma Risk After Direct‐Acting Antiviral Therapy

Watch a video presentation of this article

Abbreviations

CI

confidence interval

DAA

direct‐acting antiviral

HCC

hepatocellular carcinoma

HCV

hepatitis C virus

IFN

interferon

SVR

sustained virological response

VA

US Department of Veterans Affairs

Chronic hepatitis C virus (HCV) infection is a leading cause of hepatocellular carcinoma (HCC) worldwide. In the United States, HCV accounts for more than 60% of HCC cases.1, 2 The most important risk factor for HCC among patients with HCV is cirrhosis. Viral eradication is thought to reduce cancer risk by preventing cirrhosis and causing regression of fibrosis. Eliminating HCV also reverses other processes implicated in HCC pathogenesis, such as chronic inflammation and direct carcinogenic effects of the virus.

Direct‐acting antivirals (DAAs) are well‐tolerated agents that cure HCV in more than 95% of treated patients. A growing body of evidence suggests that DAA‐induced sustained virological response (SVR) improves portal hypertension and fibrosis in patients with chronic HCV and may reduce mortality.3 However, the question of whether DAAs prevent HCC has generated substantial controversy. In this review, we examine the impact of DAA therapy on the risk of de novo and recurrent HCC.

The Impact of DAAs on DE NOVO HCC

A primary objective of antiviral therapy is to reduce the risk for HCC. This was indeed the case after interferon (IFN)‐induced SVR. In two meta‐analyses, patients with HCV who achieved SVR with IFN had significantly lower rates of HCC compared with nonresponders (relative risk [RR], 0.26; 95% confidence interval [CI], 0.18‐0.314 versus RR, 0.34; 95% CI, 0.26‐0.465).

In contrast, initial studies in the DAA era produced less promising results. Conti et al.6 described an Italian cohort of 285 patients with HCV‐related cirrhosis who completed a course of DAAs, of whom 9 (3.16%) experienced development of de novo HCC in the first 24 weeks after treatment completion. Additional small studies reported even higher rates of HCC after DAA treatment. These rates compared unfavorably with historic HCC rates in untreated or IFN‐treated patients. It was hypothesized that DAAs cause rapid viral suppression that disrupts cancer immune surveillance, thereby allowing malignant cells to escape immune regulation.

Comparison of crude HCC rates in DAA‐treated patients with that of historic controls, however, fails to account for the higher baseline HCC risk of DAA‐eligible patients because of older age and more advanced liver disease. In a retrospective cohort study of more than 60,000 US Department of Veterans Affairs (VA) patients with chronic HCV, DAA recipients tended to be older than IFN recipients, were more likely to have cirrhosis, and were more likely to have comorbid conditions, such as diabetes and alcohol abuse, that elevate the risk for HCC.7 After adjusting for these confounders in multivariate analysis, the risks for HCC after DAA and IFN treatments were no different (adjusted hazard ratio [aHR], 1.12; 95% CI, 0.95‐1.32). Other studies comparing DAA‐ and IFN‐treated patients have come to similar conclusions.8, 9, 10, 11 (Table 1)

Table 1.

Cohort Studies Evaluating the Association Between DAA Therapy and De Novo HCC

First Author	Year	Study Design	Data Source	Patients	Start of Follow‐up	Crude HCC Incidence Rate per 100 Person‐Years (Total No. of Patients)		Adjusted HR of HCC* (95% CI)
DAA treatment with SVR versus DAA treatment without SVR						SVR	No SVR
Kanwal12	2017	Retrospective cohort	VA	HCV	Antiviral completion	0.90 (19,518)	3.45 (2982)	0.28 (0.22‐0.36)
Ioannou7	2018	Retrospective cohort	VA	HCV	180 days after antiviral initiation	0.92 (19,909)	5.19 (2039)	0.29 (0.23‐0.37)
Calvaruso13	2018	Prospective cohort	Italian multicenter cohort	HCV cirrhosis	Antiviral initiation	2.6% at 1 year (2140)	8.0% at 1 year (109)	0.35 (0.19‐0.64)†
DAA treatment versus IFN treatment						DAA	IFN
Ioannou7	2018	Retrospective cohort	VA	HCV	180 days after antiviral initiation	1.32 (21,948)	0.81 (35,871)	1.12 (0.95‐1.32)
Singer8	2018	Retrospective cohort	US administrative claims database	HCV	Antiviral initiation	1.18 (30,183)	0.98 (12,948)	0.69 (0.59‐0.81)
Li9	2018	Retrospective cohort	VA	HCV cirrhosis	Antiviral initiation excluding HCC diagnosed during first 3 months	2.52 (1,161)	3.47 (463)	1.07 (0.55‐2.08)
						DAA + SVR	IFN + SVR
Innes10	2018	Retrospective cohort	Scottish HCV database	HCV cirrhosis	Antiviral initiation	2.53 (272)	1.26 (585)	1.15 (0.49‐2.71)
Nahon11	2018	Prospective cohort	French multicenter cohort	HCV cirrhosis	Antiviral initiation	3.5% at 3 years (274)	3.1% at 3 years (495)	0.70 (0.28‐1.74)
DAA treatment versus no treatment						DAA	Untreated
Singer8	2018	Retrospective cohort	US administrative claims database	HCV	Antiviral initiation, index date in untreated	1.18 (30,183)	0.64 (137,502)	0.84 (0.73‐0.96)
Li9	2018	Retrospective cohort	VA	HCV cirrhosis	Antiviral initiation excluding HCC diagnosed during first 3 months, index date in untreated	2.52 (1161)	4.53 (1236)	Not calculated

^*DAA therapy with or without SVR relative to control group.

^†Inverse of reported aHR of 2.88.

Large cohort studies also confirm that patients with DAA‐induced SVR have a significantly lower risk for HCC compared with treatment failure or no treatment (Table 1). A study of more than 20,000 VA patients found that DAA recipients who achieved SVR had a more than 70% lower risk for HCC compared with nonresponders (aHR, 0.28; 95% CI, 0.22‐0.36).12 An Italian study of 2249 patients treated with DAAs reported a 2.6% cumulative incidence rate of HCC at 1 year among patients who achieved SVR and 8% among nonresponders, with treatment failure identified as an independent risk factor for HCC.13 Using an administrative claims database, Singer et al.8 demonstrated that DAA recipients have a lower risk for HCC compared with untreated controls (aHR, 0.84; 95% CI, 0.73‐0.96).

Collectively, observational studies support the effectiveness of DAA therapy for primary prevention of HCC. However, achievement of SVR does not obviate the need for continued HCC surveillance in patients with cirrhosis. The residual incidence rate of HCC after SVR ranges from 1.4% to 1.96% per year in the aforementioned studies, which is roughly the incidence rate at which HCC surveillance is considered cost‐effective. As such, we agree with current guideline recommendations that post‐SVR patients with advanced fibrosis receive routine HCC surveillance.14 In contrast, patients with no or minimal fibrosis at the start of antiviral therapy have low residual HCC incidence after SVR and do not warrant surveillance.

Impact of DAAs on HCC Recurrence

Meta‐analyses show that IFN‐based antiviral therapy after curative HCC treatment reduces the likelihood of cancer recurrence and improves survival.15 Whether the same is true of DAAs is an area of ongoing debate due to small, single‐arm studies that suggested that DAAs might precipitate HCC recurrence. The first such study included 58 patients with previously treated HCC, of whom 16 (27.8%) experienced recurrent tumor after a median follow‐up of 6 months from DAA initiation.16 A simultaneous publication reported HCC recurrence in 17/59 patients (28.8%).6 These recurrence rates were alarmingly high compared with historic controls, and some recurrences appeared to be temporally associated with DAA exposure.

Other studies have arrived at different conclusions (Table 2). An analysis of two French cohorts reported lower recurrence rates of 7.7% and 12.7% in patients treated with DAAs followed for a median of ~20 months after antiviral initiation.17 Moreover, there was no difference in recurrence rates between DAA‐treated and untreated patients in multivariate analysis (aHR, 1.09; 95% CI, 0.55‐2.16 and aHR, 0.40; 95% CI, 0.05‐3.03, respectively). A single‐center study of patients awaiting liver transplant compared patients treated with DAA with those who remained untreated and found no difference in HCC recurrence rates after locoregional therapy (aHR, 0.91; 95% CI, 0.58‐1.42).18 Two Japanese cohort studies and a French single‐center study suggested that DAA‐treated patients may in fact have lower rates of HCC recurrence than patients who do not achieve SVR or remain untreated.19, 20, 21

Table 2.

Studies Evaluating the Association Between DAA Therapy and Recurrent HCC

First Author	Year	Data Source	HCC Treatments	Median Time From HCC Treatment to DAA Initiation (Months)	Start of Follow‐up	Median Duration of Follow‐up in Patients Treated with DAA (Months)	Number of Patients Treated With DAA	Number of Controls	HCC Recurrence (Proportion)	Comparison of HCC Recurrence in Patients Treated With DAA Versus Controls
Reig16	2016	Spanish multicenter cohort	Resection	11.2	Antiviral initiation	5.7	58	—	27.6%	—
			Radiofrequency ablation
			Chemoembolization
Conti6	2016	Italian regional database	Resection	12.4	Antiviral completion	6	59	—	28.8%	—
			Radiofrequency ablation
			Chemoembolization
			Percutaneous ethanol
Bielen23	2017	Belgian multicenter cohort	Liver transplant	12	Antiviral completion	—	41	—	15%	—
			Resection
			Radiofrequency ablation
			Chemoembolization
Cabibbo24	2017	Italian regional database	Resection	11	Antiviral initiation	8.7	143	—	20.3%	—
			Radiofrequency ablation
			Chemoembolization
Ogawa25	2018	Japanese multicenter cohort	Resection	14.4	Antiviral initiation	17	152	—	17.1%	—
			Radiofrequency ablation
			Chemoembolization
			Radioembolization
ANRS17	2016	French multicenter cohort	—	21.6*	Antiviral initiation	20.2	189	78 (untreated)	12.7%	aHR, 1.09 (95% CI, 0.55‐2.16)
ANRS17	2016	French multicenter cohort	Resection	__	Antiviral initiation	21.3	13	66 (untreated)	7.7%	HR, 0.40 (95% CI, 0.05‐3.03)
			Radiofrequency ablation
Ikeda26	2017	Japanese single center	Resection	10.7	Antiviral initiation	21.5	89	89 (untreated)	19.3%	aHR, 0.35 (95% CI, 0.19‐0.65)
			Radiofrequency ablation
			Chemoembolization
			Particle radiation therapy
Nagata19	2017	Japanese multicenter cohort	Resection	—	HCC curative treatment	27.6	83	60 (IFN treated)	22.9% DAA + SVR	P = 0.022
			Radiofrequency ablation
									40.0% DAA − SVR
									(3‐year incidence)
									45.1% DAA	P = 0.54
									54.2% IFN
									(5‐year incidence)
Virlogeux21	2017	French single center	Resection	7.2	Antiviral initiation	—	23	45 (untreated)	47.8%	HR, 0.24 (95% CI, 0.10‐0.55)
			Radiofrequency ablation
			Chemoembolization
Huang18	2018	US single center	Resection	11.9†	Complete response	—	61	51 (untreated)	47% DAA	aHR, 0.91 (95% CI, 0.58‐1.42)
			Radiofrequency ablation						49.8% no DAA
									(1‐year incidence)
			Cryotherapy
			Percutaneous ethanol
Mashiba20	2018	Japanese multicenter cohort	—	10.9	Antiviral completion	7.7	368	148 (IFN treated)	—	HR, 0.95 (95% CI, 0.61‐1.45)‡

^*From HCC diagnosis to inclusion in cohort.

^†From HCC diagnosis to DAA initiation.

^‡IFN versus DAA therapy, restricted to SVR patients.

Various methodological limitations hamper the interpretation of existing studies and prevent firm conclusions from being drawn. These limitations are detailed in a recent meta‐analysis.22 (Fig. 1) First, sample sizes of studies are typically small, and many do not include control arms. Second, studies are inconsistent in excluding HCC or suspicious nodules prior to DAA treatment. Consequently, some “recurrences” may actually represent prevalent tumors that were present at the start of antiviral therapy. Third, most studies do not specify a standard surveillance protocol after curative treatment, leading to heterogeneity in ascertainment of recurrence. Fourth, studies variably include patients with non–early‐stage HCC, patients with past recurrences, and patients who received noncurative locoregional therapies, all of which would be expected to increase recurrence rates. Lastly, determining when to start the clock in terms of follow‐up varies. Follow‐up may be calculated from the time of HCC treatment, antiviral therapy initiation, or antiviral completion, each of which produces different estimates of recurrence rates. The delay between HCC treatment and antiviral initiation also varies widely. In a meta‐regression analysis, the most common factor associated with recurrence is a short interval between HCC cure and DAA initiation, suggesting that many cases of “recurrence” are in fact patients who have not achieved a durable complete response.

Figure 1.

Sources of heterogeneity in studies examining the association between DAAs and HCC recurrence.

Given the inconclusiveness of existing evidence, some clinicians are hesitant to initiate DAA therapy in patients with previously treated HCC. Yet withholding treatment risks further liver decompensation, and the possibility remains that treatment may actually lower the risk for recurrence. Hence we propose DAA therapy be considered in patients with prior HCC provided a durable complete response has been demonstrated. It seems reasonable to wait at least 6 months after the first demonstration of complete response and to obtain two multiphase computed tomography or magnetic resonance imaging studies to document absence of HCC before initiating antivirals.

Conclusion

DAAs dramatically enhanced our ability to cure chronic HCV and prevent its downstream complications. Several large cohort studies demonstrate that DAA‐induced SVR reduces the risk for de novo HCC. However, the residual risk for HCC after SVR means that surveillance is still required for patients with advanced fibrosis and cirrhosis. Additional well‐designed studies are needed to determine the consequences of DAA therapy in patients with previously treated HCC, although the initial concern that DAAs increase recurrence risk has generally not been borne out in subsequent studies.