Unexpected viral relapses in hepatitis C virus–infected patients diagnosed with hepatocellular carcinoma during treatment with direct-acting antivirals

Current treatment of chronic hepatitis C has shown very high success rates, both in clinical trials and in real-life settings. The few relapses are mainly seen in patients with advanced cirrhosis, but predicting factors are still poorly identified.

Recently, there has been a lot of interest in the unexpected high rate of early hepatocellular carcinoma (HCC) recurrence in patients undergoing treatment with direct-acting antivirals (DAAs) soon after having treated HCC.1, 2 Other groups questioned these findings, presenting consistent cohort data with no evidence of this association.3 Besides the speculation on possible mechanisms underpinning this phenomenon, there have been no reports so far on “the other side of the moon,” i.e., the possible effect of HCC emerging during treatment in hindering sustained virologic response (SVR).

Here, we report five cases of patients who were diagnosed with HCC while receiving DAAs or soon after treatment completion and had a viral relapse after the end of treatment. Possible mechanisms underlying this clinical observation are then suggested.

PROREDAA (Prospective Retrospective DAA) is an ongoing single-center cohort study, approved by our institutional review board, enrolling hepatitis C virus (HCV) patients who have been treated with DAAs since January 2015. Among patients with cirrhosis treated with DAAs, we compared the relapse rate between those who developed HCC within 6 months from treatment start (a timing consistent with HCC surveillance) and those who did not.

Case Series

Among 155 HCV-infected subjects with cirrhosis undergoing treatment with DAAs for the first time between January 1, 2015, and May 31, 2016, 5 were diagnosed with HCC within 6 months from treatment initiation. Patients with or without HCC were comparable in terms of age, gender, and pretreatment HCV RNA (Welch t test P > 0.05); but, as expected, those with HCC had more advanced liver disease than those without it. In particular, they were more likely to have Child B cirrhosis (2/5 = 40% versus 8/150 = 5.3%, respectively; Fisher exact test P = 0.034) and had higher liver stiffness (32.8 [interquartile range 28.9-35.5] versus 20 [14.5-26.3] kPa; P = 0.01) and lower levels of platelets (62 [62-71] versus 105 [75-147] × 10³/mm³, P = 0.04).

Of note, all patients diagnosed with active HCC experienced viral relapse after the end of treatment: 3 patients relapsed early at week 4, while 2 relapsed despite SVR at week 12 (1 relapse at week 24 after treatment completion and 1 after 36 weeks) (Table 1). Among the remaining 150 patients with cirrhosis but without HCC, we observed 12 relapses, 9 of which were in patients receiving suboptimal therapy (i.e., sofosbuvir and ribavirin for genotype 3 in patients starting treatment before the availability of daclatasvir) or interrupting treatment early. Among the remaining 3 patients who had received proper treatment, 2 relapsed early at week 4 and 1 experienced a late relapse, diagnosed at week 24 posttreatment, although the result at week 12 was not available. The genotype of all relapsing HCVs was consistent with the genotype determined before treatment.

Despite the low numbers, the relapse rate among patients with HCC was significantly higher than that among those without HCC (100% versus 8%, respectively; Fisher exact test P < 0.0001). Using exact logistic regression analysis, a diagnosis of HCC within 6 months from treatment initiation was associated with significantly lower chances of obtaining SVR (odds ratio = 0.01, 95% confidence interval 0-0.10; P < 0.001). Adjustment for other predictors of SVR unequally distributed in the two groups, such as Child stage, liver stiffness, or platelet count, slightly reduced the magnitude of the association (adjusted odds ratio = 0.02, 95% confidence interval 0-0.2 for all adjustments), which remained highly statistically significant (all P < 0.001).

Interestingly enough, among patients without HCC, 5 had a past diagnosis of HCC that had been successfully treated with surgical resection, local ablation (radiofrequency), or chemoembolization before DAA treatment. All of these five patients achieved SVR after treatment with DAAs, though they were comparable, in terms of liver disease staging, to the patients with active HCC.

Discussion

We observed an unexpectedly high rate of viral relapse after DAA treatment in patients with cirrhosis and a concurrent HCC diagnosis, the reason of which is unclear.

One hypothesis could be that subverted cellular architecture and vascularization of HCC foci can impair the penetration of DAAs, thus creating sanctuary sites for HCV replication and favoring relapse after treatment. Another possible explanation for the lower DAA treatment efficacy could be the reduced expression of membrane transporters, such as OATP1B1, which is critical for DAA uptake in liver cells4; and this has been shown to be down-regulated in HCC cells.5 Impaired immunological control might also be involved in favoring viral relapse.

These complex mechanisms should be further investigated, and our association should be tested and verified on a larger real-life cohort of patients treated with DAAs. If confirmed, this observation would have immediate practical application: it could help clinical decision-making in the timing and duration of treatment, particularly in patients with HCC who are on a waiting list for liver transplantation, suggesting better chances of HCV eradication after HCC cure or the need of a prolonged treatment among those diagnosed with HCC during treatment. Moreover, given the fact that two out of five relapses in patients with HCC occurred beyond week 12 posttreatment, prolonged serial HCV RNA assessment may be needed. On the other hand, the high prevalence of HCC among HCV relapsers (29% in our small cohort) may suggest that intensive workup, in order to exclude undiagnosed HCC, could be indicated when viral relapses occur.