Increased Risk for Hepatocellular Carcinoma Persists Up to 10 Years After HCV Eradication in Patients with Baseline Cirrhosis or High FIB-4 Scores

Abstract

Background & Aims

It is unclear if hepatocellular carcinoma (HCC) risk declines over time after hepatitis C virus (HCV) eradication. We analyzed changes in HCC annual incidence over time following HCV eradication and identified dynamic markers of HCC risk.

Methods

We identified 48,135 patients who initiated HCV antiviral treatment from 2000 through 2015 and achieved a sustained virologic response (SVR) in the Veterans Health Administration (29,033 treated with direct-acting antiviral (DAA) agents and 19,102 treated with interferon-based regimens). Patients were followed after treatment until February 14, 2019 (average 5.4 years), during which 1509 incident HCCs were identified.

Results

Among patients with cirrhosis before treatment with DAAs (n=9784), those with pre-SVR fibrosis-4 (FIB-4) scores ≥3.25 had a higher annual incidence of HCC (3.66%/year) than those with FIB-4 scores <3.25 (1.16%/year) (adjusted hazard ratio, 2.14; 95% CI, 1.66–2.75). In DAA-treated patients with cirrhosis and FIB-4 scores ≥3.25, annual HCC risk decreased from 3.8%/year in the first year after SVR to 2.4%/year by the fourth year (P=.01). In interferon-treated patients with FIB-4 scores ≥3.25, annual HCC risk remained above 2%/year, even 10 years after SVR. A decrease in FIB-4 scores from ≥3.25 pre-SVR to <3.25 post-SVR was associated with an approximately 50% lower risk of HCC, but the absolute annual risk remained above 2%/year. Patients without cirrhosis before treatment (n=38,351) had a low risk of HCC, except for those with pre-SVR FIB-4 scores ≥3.25 (HCC risk 1.22%/year) and post-SVR FIB-4 scores≥3.25 (HCC risk 2.39%/year)—risk remained high for many years after SVR.

Conclusions

Patients with cirrhosis before an SVR to treatment for HCV infection continue to have a high risk for HCC (>2%/year) for many years, even if their FIB-4 score decreases, and should continue surveillance. Patients without cirrhosis but with FIB-4 scores ≥3.25 have a high enough risk to merit HCC surveillance, especially if FIB-4 remains ≥3.25 post-SVR.

Lay Summary

Patients with cirrhosis or severe fibrosis before treatment for hepatitis C virus (HCV) infection should be monitored for hepatocellular carcinoma, even if their HCV infection is cured.

Graphical Abstract

Introduction

Eradication of hepatitis C virus (HCV) by successful antiviral treatment either with interferon-alpha (IFN) or with direct-acting antivirals (DAA) has been shown to reduce the risk of hepatocellular carcinoma (HCC) significantly^1–5. We recently developed models estimating HCC risk in patients with HCV following antiviral treatment⁶. Among patients without cirrhosis at the time of antiviral treatment, the risk of HCC after sustained virologic response (SVR) appears to be very low ⁶. However, among patients with established cirrhosis prior to antiviral treatment, a substantial absolute risk of HCC persists after SVR, even though it is significantly reduced compared to untreated patients and those who did not achieve SVR. As a result, current guidelines recommend that patients with cirrhosis should continue HCC surveillance indefinitely after SVR⁷. However, HCC risk varies widely between patients – both among those with and without cirrhosis – as highlighted by HCC risk models following antiviral treatment⁶. Further, the duration of time for which patients remain at high risk of HCC after SVR and therefore might benefit from ongoing HCC surveillance is uncertain.

Given the high effectiveness of DAAs, the majority of patients with diagnosed HCV cirrhosis have either already been treated or are expected to be treated and cured in the near future. It is critical to develop easy-to-apply methods for accurately estimating HCC incidence in patients with cirrhosis who achieve SVR, as this is one of the most important determinants of the cost-effectiveness of surveillance. The cost of surveillance per quality-adjusted life year has been shown to decrease almost exponentially with increasing HCC incidence^{8, 9}.

Although we have developed risk stratification models among post-SVR patients using pre-treatment data⁶, time that elapses since viral eradication is a complexity that has not yet been considered. In some patients with cirrhosis, resolution of liver fibrosis and liver remodeling after SVR may lead to a decline in HCC risk over time to a level at which HCC surveillance is no longer warranted. However, others, especially those with more advanced cirrhosis, may not have resolution of fibrosis and resultant reduction in HCC risk after SVR. Furthermore, as time accrues after SVR, patients become older or may acquire factors that forestall resolution of fibrosis or mitigate any decreases in HCC risk (e.g., diabetes, obesity, alcohol use) or even increase HCC risk among those without cirrhosis at baseline. Finally, fibrosis regression and persistent HCC risk may not be perfectly correlated, as prior studies suggest some patients can still develop HCC years after SVR despite documented fibrosis regression. Preneoplastic changes that occurred before SVR may persist “indefinitely” after SVR¹⁰, predisposing a patient to develop HCC many years after SVR.

The long-term trajectory of HCC risk after HCV eradication in patients with or without pre-treatment cirrhosis has not been previously reported. We aimed to describe how HCC risk changes over time after HCV eradication and to identify factors that might distinguish patients with persistently high HCC risk from those with declining risk.

Methods

Data source

We derived data from the Veterans Health Administration (VHA), the largest integrated healthcare system in the US currently serving more than 8.9 million Veterans at 168 VA Medical Centers and 1053 outpatient clinics throughout the country¹¹. The VHA has treated the largest number of HCV-infected patients of any healthcare system and was an early adopter of unrestricted access to DAAs¹² making it an ideal setting for the proposed study. Furthermore, the VHA uses a single, nationwide, comprehensive electronic healthcare information network. We obtained electronic data on all patients who initiated antiviral treatment in the VA system using the VA Corporate Data Warehouse (CDW), a national, continually updated repository of VHA electronic health records developed specifically to facilitate research¹³. The study was approved by the Institutional Review Board of the VA Puget Sound Healthcare System.

Study population

We identified 105,310 HCV antiviral regimens initiated in the VA during 16 calendar years from 1/1/2000 to 12/31/2015, of which 51,828 resulted in SVR, defined as a serum HCV RNA viral load below the lower limit of detection performed at least 12 weeks after the end of HCV treatment¹⁴. Follow-up for development of HCC after antiviral treatment extended to 02/14/2019.

Among these 51,828 patients who achieved SVR, we excluded 1317 patients who had a diagnosis of HCC recorded prior to antiviral treatment or within 180 days. We also excluded 80 patients who died within 180 days from the start-date of antiviral treatment or had fewer than 180 days of available follow-up, 826 patients who underwent liver transplantation prior to antiviral treatment and 1470 patients without baseline Fibrosis-4 (FIB-4) scores (defined below), leaving 48,135 patients in the current analysis (9784 with and 38,351 without pre-treatment cirrhosis), including 1509 who developed HCC more than 180 days after antiviral treatment initiation.

Diagnosis of cirrhosis and hepatocellular carcinoma

The diagnosis of cirrhosis prior to antiviral treatment was based on ICD-9 and ICD-10 codes for cirrhosis or complications of cirrhosis (ascites, varices [bleeding or non-bleeding], encephalopathy, spontaneous bacterial peritonitis, hepatorenal syndrome, or hepatopulmonary syndrome) listed in Supplemental Table 1 recorded at least twice in any inpatient or outpatient encounter prior to initiation of antiviral treatment—an approach validated and widely used in VA-based studies by us^15–23 and others^24–28. The diagnosis of cirrhosis using a single ICD-9 code in VA data was shown to have a 90% positive predictive value compared to chart extraction. By requiring the relevant ICD-9 codes to be recorded at least twice, we found that the positive predictive value increased to 97%²⁹.

We identified incident cases of HCC diagnosed for the first time at least 180 days after initiation of antiviral treatment based on ICD-9 code 155.0 or ICD-10 code C22.0 documented at least twice. The ICD-9 code-based definition of HCC using VA records has been shown to have a positive predictive value of 84–94% compared to chart extraction^{27, 30, 31} and has been widely used by us^{20, 21, 23, 32} and other investigators^33–35.

Antiviral treatment regimens

The regimens were divided into:

1. “DAA-only” regimens (n=29,033), which included regimens consisting only of DAAs (± ribavirin), without any interferon, the most common being ledipasvir/sofosbuvir (59.7%).

2. “IFN-based” regimens (n=19,102), which included IFN-only regimens (± ribavirin) (n=14,943) or regimens that combined IFN with first generation DAAs such a boceprevir, telaprevir, simeprevir or sofosbuvir (n=4159).

Dispensed drugs (rather than just prescribed drugs) were used to define antiviral treatment regimens, as we previously published^{12, 18, 32, 36–40}. All VA pharmacy data are included in CDW and were available to us for analysis. Analyses were performed separately for DAA-only and IFN-based regimens.

Baseline characteristics prior to antiviral treatment

We collected baseline data prior to antiviral treatment including age, sex, body mass index (BMI), and HCV genotype. We extracted all laboratory tests prior to treatment and recorded the value of each test closest to the date of treatment initiation within the preceding 6 months as the baseline value.

We defined HBV coinfection by positive HBV surface antigen or viral load. We determined the presence of type 2 diabetes mellitus, alcohol use disorders, substance use disorders, and HIV infection based on appropriate ICD-9 or ICD-10 codes recorded at least twice prior to treatment initiation in any inpatient or outpatient encounter. These ICD-based definitions of these comorbidities^{20, 22, 25–28} have been widely used and validated in studies using VA medical records. In addition to extracting alcohol-related ICD9/10 codes at baseline, we also extracted the score of the Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) questionnaire conducted closest to HCV antiviral treatment within a year prior to treatment³⁸. The AUDIT-C is a validated screening test^{41, 42} used to screen all outpatients in the VA for unhealthy alcohol use annually⁴³.

Fibrosis-4 (FIB-4) scores before treatment and after SVR

We used baseline laboratory tests ascertained within 6 months prior to the initiation of antiviral treatment to calculate the baseline FIB-4 score as follows:

$$(\text{Age}\times \text{AST})/(\text{Platelet}\text{count}\times \surd \text{ALT}).$$

Although the FIB-4 score was developed as a measure of advanced fibrosis of the liver, a FIB-4 score ≥3.25 has also been shown to be strongly associated with HCC risk^{3, 29}. This is not surprising since each component of the FIB-4 score (increasing age and AST/√ALT and decreasing platelet count) is strongly associated with HCC risk.

We also extracted all laboratory tests after SVR and used them in combination with a patient’s age to calculate all FIB-4 scores after SVR. We were interested in evaluating whether “improvements” in laboratory tests after SVR (e.g., an increase in platelet count, or a decrease in ALT relative to AST) resulting in a decrease in FIB-4 score might be associated with decreased HCC risk. We used FIB-4 scores to categorize patients into the following 4 groups.

1. Baseline FIB-4 ≥3.25 and post-SVR FIB-4 ≥3.25 (if all post-SVR FIB-4 results were ≥3.25)

2. Baseline FIB-4 ≥3.25 and post-SVR FIB-4 <3.25 (if at least one post-SVR FIB-4 result was <3.25)

3. Baseline FIB-4 <3.25 and post-SVR FIB-4 <3.25 (if all post-SVR FIB-4 results were <3.25)

4. Baseline FIB-4 <3.25 and post-SVR FIB-4 ≥3.25 (if at least one post-SVR FIB-4 results was ≥3.25)

Since a change from high to low or from low to high FIB-4 can happen at any time during follow-up, we modeled this change in two different ways:

1. Many patients experience normalization of serum AST and ALT levels during or immediately after antiviral treatment, which may result in a reduction in FIB-4 score from ≥3.25 before treatment to <3.25 immediately after. Therefore, we identified categories a-d above based in changes in FIB-4 score that occurred within 1 year of antiviral treatment initiation.

2. In addition, we wanted to capture the impact of drop in FIB-4 score occurring at any time during follow-up, including many years after SVR. To do that, we analyzed FIB-4 level (categorized as ≥3.25 or <3.25) as a time-dependent covariate. In this analysis, each patient was analyzed under the low FIB-4 category for time period when their FIB-4 was low and under the high FIB-4 category for the time periods when their FIB-4 was high. For example, a patient who drops from FIB-4 ≥3.25 before SVR to <3.25 at time t after SVR is analyzed under the ≥3.25 category from time 0 until time t and under the <3.25 category after time t. This method accounts for immortal time bias, by correctly assigning the cancer-free time before FIB-4 drops (i.e., from time 0 to time t) to the high FIB-4 category.

Statistical analysis

We identified all new HCCs diagnosed at least 180 days after initiation of antiviral treatment, as we previously described^{6, 15}, in order to exclude as much as possible HCCs that might have been present but undiagnosed or occult at the time that antiviral treatment was initiated. We developed multivariable Cox proportional hazards models to predict the time to HCC development using baseline characteristics, including FIB-4 score, AUDIT-C score, sex, race/ethnicity, body mass index, HCV genotype, HCV viral load, HIV co-infection, HBV co-infection, type 2 diabetes mellitus, alcohol use disorders, substance use disorders, cirrhosis complications, serum bilirubin, serum creatinine, serum albumin, INR and hemoglobin levels. We also performed a competing risks proportional hazards analysis with death as a competing risk to HCC development, as a sensitivity analysis.

We performed simple descriptive analyses, stratified by FIB-4 score, in which we calculated the annual HCC incidence during each year (i.e., 12-month period) of follow-up after treatment, among patients who were still alive and in care during that year, in order to evaluate whether HCC risk appears to be decreasing over time or not. We also calculated HCC risk according to whether the FIB-4 score declined to a value <3.25 or increased to a value ≥3.25 during follow-up, analyzed as a time-dependent covariate or according to the change within 1 year from treatment.

Results

Baseline characteristics of patients with pre-treatment cirrhosis

Among 9784 patients with cirrhosis who achieved SVR, the majority (n=7553, 77%) were treated with DAA-only regimens and only 23% were treated with IFN-based regimens (Table 1). The majority of patients were male (97%) and non-Hispanic White (56%), with substantial representation of racial/ethnic minority groups. Mean age was 61 years and diabetes (35%), alcohol use disorders (46%) and substance use disorders (34%) were common.

Table 1.

Selected baseline characteristics of 48,135 HCV-infected patients who achieved SVR following antiviral treatment initiated from 1999–2015 in the Veterans Affairs healthcare system, according to presence or absence of pre-treatment cirrhosis.

	CIRRHOSIS				NO CIRRHOSIS
CHARACTERISTICS	IFN-based† Regimens (N=2251)		DAA-only Regimens (N=7533)		IFN-based† Regimens (N=16,851)		DAA-only Regimens (N=21,500)
	FIB-4 <3.25 n=1208	FIB-4 ≥3.25 n=1043	FIB-4 <3.25 n=2546	FIB-4 ≥3.25 n=4987	FIB-4 <3.25 n=14,921	FIB-4 ≥3.25 n=1930	FIB-4 <3.25 n=16,818	FIB-4 ≥3.25 n=4682
Age, yrs (mean [SD])	56.1 [6.6]	57.8 [5.5]	61.6 [5.3]	62.0 [5.3]	52.6 [7.0]	56.9 [6.0]	60.3 [7.0]	62.8 [5.5]
BMI, Kg/m2 (mean [SD])	28.9 [5.2]	29.0 [5.6]	28.8 [5.5]	28.8 [5.5]	28.2 [5.1]	28.1 [5.2]	27.9 [5.3]	27.7 [5.3]
Male (%)	97.3	96.9	97.2	96.9	95.3	97.6	96.2	97.2
Race/Ethnicity (%)
White, non-Hispanic	59.4	66.6	47.7	56.8	65.1	67	51.7	51.7
Black, non-Hispanic	18.7	12.4	36.7	25.1	16.5	13.1	35.1	33.1
Hispanic	9.9	7.2	7.0	8.3	5.2	7.2	4.2	5.2
Other	2.3	2.4	1.5	1.8	1.9	1.9	1.6	1.6
Declined to answer/missing	9.6	11.4	7.1	7.9	11.4	10.8	7.5	8.3
Genotype (%)
Genotype 1	69.9	61.9	88.8	85.5	54	51.5	84.6	84.1
Genotype 2	13.7	13.8	5.9	6.7	22.4	19.3	9.7	8.6
Genotype 3	9.1	15.2	3.0	5.6	11.4	15	3.8	5.1
Genotype 4	1.2	0.7	0.9	0.6	0.6	0.6	0.8	0.7
Missing	6.0	8.3	1.4	1.6	11.6	13.7	1	1.5
HBV co-infection (%)	2.5	1.8	1.8	1.7	1	1.2	1	1.5
HIV co-infection (%)	0.8	2.8	3.8	3.1	2	2.4	4.9	5.1
Diabetes (%)	27.3	26.6	40.2	36.6	14	20.1	25.1	28.5
Substance Use Disorders (%)	41.3	44.1	46.8	47.6	29.3	24.6	38.8	36.1
History of Alcohol Use Disorders* (%)	28.6	26.7	39.0	34.8	35.6	34.4	42.1	43.7
Alcohol use based on AUDIT-C score‡ (%)
Abstinent	75.6	76.8	72.7	76.1	63.4	66.6	64	60.8
Low-level drinking	17.2	16.8	20.4	16.9	24.7	21.4	26.1	26.1
Unhealthy drinking	7.2	6.5	6.9	7.0	11.9	12	9.9	13.1
Cirrhosis complications††	18.1	27.4	12.1	31.5	N/A	N/A	N/A	N/A
MELD score (mean [SD])	7.9 [3.5]	8.8 [3.7]	8.6 [4.0]	9.6 [3.7]	N/A	N/A	N/A	N/A
Bilirubin, g/dL (mean [SD])	0.6 [0.4]	0.9 [0.5]	0.6 [0.4]	1.0 [0.7]	0.6 [0.4]	0.8 [0.6]	0.6 [0.4]	0.8 [0.5]
INR (mean [SD])	1.2 [1.4]	1.4 [1.6]	1.2 [1.2]	1.3 [1.2]	1.1 [1.0]	1.1 [0.7]	1.1 [0.9]	1.2 [0.8]
Creatinine, mg/dL (mean [SD])	1.0 [0.5]	0.9 [0.4]	1.1 [0.6]	1.0 [0.4]	1.0 [0.4]	0.9 [0.2]	1.0 [0.5]	1.0 [0.4]

^†IFN-based regimens include IFN-only regimens (± ribavirin) or regimens in which IFN was combined with boceprevir, telaprevir, sofosbuvir, or simeprevir.

^*Based on the presence of ICD9/10 diagnostic codes for alcohol use disorders recorded at least twice at any time prior to antiviral treatment initiation

^‡Based on AUDIT-C questionnaires completed within 12 months prior to antiviral treatment initiation, categorized into abstinent (score of 0), low-level drinking (score of 1–3 in men, 1–2 in women), and unhealthy drinking (score of 4–12 in men, 3–12 in women) ^{38, 51}

^††Complications of cirrhosis defined by ICD9/10 diagnostic codes for: ascites, varices (bleeding or non-bleeding), encephalopathy, spontaneous bacterial peritonitis, hepatorenal syndrome, or hepatopulmonary syndrome.

Compared to IFN- patients, those treated with DAA-only regimens were older (61.9 vs 56.9 years), more likely to have genotype 1 HCV (86.6% vs 66.2%), more likely to be Black (29% vs 16%) and more likely to have a FIB-4 score ≥3.25 (66.2% vs 46.3%).

Compared to patients with FIB-4 <3.25, those with FIB-4 ≥3.25 were more likely be white, to have had HCV genotype 3, and to have complications of cirrhosis.

Baseline characteristics of patients without pre-treatment cirrhosis

Among 38,351 patients without pre-treatment cirrhosis who achieved SVR, the majority (n=21,500, 56%) were treated with DAA-only regimens, and most were male (96%) and non-Hispanic white (56%) (Table 1). Compared to IFN-treated patients, those treated with DAA-only regimens were older (60.8 vs 53.1 years), more likely to have genotype 1 HCV (84.5% vs 53.7%), more likely to be Black (34.7% vs 16.1%) and more likely to have a FIB-4 score ≥3.25 (21.8% vs 11.5%). Compared to patients with FIB-4 <3.25, those with FIB-4 ≥3.25 were older, more likely to have had HCV genotype 3, and to have diabetes.

Baseline FIB-4 ≥3.25 is an independent risk factor for HCC in patients with pre-treatment cirrhosis

Among patients with pre-treatment cirrhosis, 850 out of 9784 patients developed HCC during a mean follow-up period of 3.9 years (incidence=2.2 per 100 patient-years). Mean follow-up was much longer in patients treated with IFN-based regimens (12.4 years) than DAA-only regimens (3.0 years), as expected since DAA-only regimens first became available in late 2013.

Among DAA-treated patients, those with a baseline FIB-4 ≥3.25 had a significantly higher incidence of HCC (3.66 per 100 patient-years) than patients with baseline FIB-4 <3.25 (1.16 per 100 patient-years) (Table 2 and Figure 1). After adjustment for other baseline characteristics by Cox proportional hazards regression, patients with baseline FIB-4 ≥3.25 had ~2-fold higher risk of HCC than patients with FIB-4 <3.25 (adjusted hazard ratio [AHR] 2.14, 95% CI 1.66–2.75) (Table 2).

Table 2.

Association between pre-treatment FIB-4 score and HCC risk among patients with or without baseline cirrhosis who achieved SVR, stratified by type of antiviral treatment.

		Number of patients (%)	Patient-years	Number who developed HCC (%)	HCC per 100 patient-years	Crude hazard ratio (95% CI)	Adjusted* hazard ratio (95%CI)
CIRRHOSIS
IFN-based REGIMENS	FIB-4 <3.25	1,208(53.7)	9,286	63(5.2)	0.68	1	1
	FIB-4 ≥3.25	1,043(46.3)	6,283	168(16.1)	2.67	3.91 (2.93–5.23)	2.78 (1.91–4.03)
DAA-only REGIMENS	FIB-4 <3.25	2,546(33.8)	7,753	90(3.5)	1.16	1	1
	FIB-4 ≥3.25	4,987(66.2)	14,448	529(10.6)	3.66	3.17 (2.54–3.96)	2.14 (1.66–2.75)
NO CIRRHOSIS
IFN-based REGIMENS	FIB-4 <3.25	14,921(88.5)	143,790	203(1.4)	0.14	1	1
	FIB-4 ≥3.25	1,930(11.5)	15,663	168(8.7)	1.07	7.79 (6.34–9.56)	5.11 (3.94–6.62)
DAA-only REGIMENS	FIB-4 <3.25	16,818(78.2)	50,029	121(0.7)	0.24	1	1
	FIB-4 ≥3.25	4,682(21.8)	13,651	167(3.6)	1.22	5.06 (4.01–6.40)	3.56 (2.74–4.63)

^*Adjusted for sex, race/ethnicity, body mass index, HCV genotype, HCV viral load, HIV co-infection, HBV co-infection, type 2 diabetes mellitus, alcohol use disorders, AUDIT-C score, substance use disorder, cirrhosis complications, serum bilirubin, serum creatinine, serum albumin, blood INR and blood hemoglobin levels. The laboratory tests were categorized into quartiles and modeled as dummy categorical variables.

Figure 1.

Cumulative HCC incidence curves following SVR, stratified by cirrhosis versus no cirrhosis and by high FIB-4 (≥3.25) versus low FIB-4 (<3.25)

a. IFN-treated patients

b. DAA-treated patients

p-value <0.001 comparing FIB-4 ≥3.25 to <3.25 groups

Similarly, among IFN-treated patients, those with a baseline FIB-4 ≥3.25 had a significantly higher incidence of HCC (2.67 per 100 patient-years) than patients with baseline FIB-4 <3.25 (0.68 per 100 patient-years), AHR 2.78, 95% CI 1.91–4.03, (Table 2).

Accounting for the competing risk of death by competing risks analysis had little effect on the associations between FIB-4 and HCC risk (Supplemental Table 2).

Baseline FIB-4 ≥3.25 identifies patients without pre-treatment cirrhosis who have high HCC risk

A substantial number (659 out of 38,351) patients without pre-treatment cirrhosis diagnosis developed HCC during a mean follow-up period of 5.8 years, although, as expected, the incidence (0.3 per 100 patient-years) was much lower than in patients with pre-treatment cirrhosis. FIB-4 level ≥3.25 was strongly associated with HCC in both IFN (AHR 5.11, 95% CI 3.94–6.62) and DAA-treated patients (AHR 3.56, 95% CI 2.74–4.63), compared to FIB-4 <3.25 (Table 2 and Figure 1b). HCC incidence in patents without pre-treatment cirrhosis who had FIB-4 ≥3.25 was greater than 1 per 100 patient-years (1.22%/yr in DAA-treated and 1.07%/yr in IFN-treated patients) and very similar to the HCC incidence in patients with pre-treatment cirrhosis but low FIB-4 <3.25.

Change in HCC risk over time after SVR in patients with pre-treatment cirrhosis

Among DAA-treated patients, the annual incidence of HCC appeared to decrease over time during the first 4 years after treatment (Figure 2a and Supplemental Table 3) both in patients with baseline FIB-4 ≥3.25 (from 3.8% in year 1 to 2.4% in year 4) and those with baseline FIB-4 <3.25 (from 1.4% in year 1 to 0.5% in year 4). This decrease was statistically significant (p=0.01) in a test of trends over time. However, only 4 years of follow-up were available for analysis since the introduction of DAA-only regimens.

Figure 2.

Annual HCC incidence after SVR in patients according to treatment type and FIB-4 score in

a. Patients with pre-treatment cirrhosis

b. Patients without pre-treatment cirrhosis

In contrast, among IFN-treated patients who had more than 10 years of follow-up, there did not appear to be any clear decrease in HCC risk over time after SVR among patients with baseline FIB-4 ≥3.25 (Figure 2a). Annual HCC risk remained consistently above 2% even 10 years after SVR and the test for trends over time was not statistically significant (p=0.7). Among IFN -treated patients with baseline FIB-4 <3.25, annual HCC risk remained low (<1%) during up to 10 years of follow-up, without any obvious reduction in HCC risk over time.

Change in HCC risk over time after SVR in patients without pre-treatment cirrhosis

Patients without pre-treatment cirrhosis who had baseline FIB-4 ≥3.25 had elevated HCC risk that persisted for many years after SVR without any decline over time (Figure 2b and Supplemental Table 4). In patients with baseline FIB-4 <3.25, HCC risk remained low over time after SVR.

Drop in FIB-4 score from ≥3.25 pre-treatment to <3.25 after SVR is associated with reduced HCC risk

Among patients with pre-treatment cirrhosis and baseline FIB-4 ≥3.25, more than half attained a FIB-4 score <3.25 within one year post-SVR (Table 3). Among patients with baseline FIB-4 ≥3.25 treated with DAA-only regimens, those whose FIB-4 score declined to a level <3.25 within one year after SVR had significantly lower annualized HCC incidence (2.47%) than those whose FIB-4 remained ≥3.25 (5.07%) (AHR 0.59, 95% CI 0.48–0.73) (Table 3 and Figure 3a). When the drop of FIB-4 was analyzed as a time-dependent variable, again patients whose FIB-4 dropped to a level < 3.25 after SVR had lower annualized HCC incidence than patients whose FIB-4 remained ≥3.25 (2.87% vs. 4.43%, AHR 0.76, 95% CI 0.62–0.94) (Table 3). Almost identical results were observed in patients treated with IFN-based regimens (Tables 3). Despite the reduction in HCC risk that we observed in association with a drop in FIB-4 score, a substantial residual absolute risk of HCC remained among patients whose FIB-4 dropped from ≥3.25 pre-treatment to <3.25 post-SVR for both DAA-only regimens (2.87% per year) and IFN-based regimens (2.21% per year).

Table 3.

Association between a drop in FIB-4 score from ≥3.25 before SVR to <3.25 after SVR and reduced HCC risk

	FIB-4 Score Pre-treatment	FIB-4 Score Post-treatment	Number of patients	Patient-years	Number who developed HCC (%)	HCC per 100 patient-years	Crude hazard ratio (95% CI)	Adjusted* hazard ratio (95% CI)
CIRRHOSIS
DROP in FIB-4 post-treatment ascertained within 1 year of SVR†
IFN-based REGIMENS	≥3.25	≥3.25	328(33.5)	1,887	76(23.2)	4.03	1	1
	≥3.25	<3.25	650(66.5)	4,000	77(11.8)	1.93	0.48 (0.35–0.66)	0.54 (0.37–0.79)
DAA-only REGIMENS	≥3.25	≥3.25	2,233(49.3)	6,256	318(14.2)	5.08	1	1
	≥3.25	<3.25	2,296(50.7)	6,888	169(7.4)	2.45	0.48 (0.40–0.58)	0.58 (0.47–0.72)
DROP in FIB-4 post-treatment ascertained at any time after SVR as a time-dependent covariate††
IFN-based REGIMENS	≥3.25	≥3.25	N/A	1,767	59	3.34	1	1
	≥3.25	<3.25	N/A	4,470	98	2.19	0.48 (0.34–0.68)	0.58 (0.38–0.88)
DAA-only REGIMENS	≥3.25	≥3.25	N/A	6,138	272	4.43	1	1
	≥3.25	<3.25	N/A	7,972	229	2.87	0.60 (0.50–0.72)	0.77 (0.62–0.94)
NO CIRRHOSIS
DROP in FIB-4 post-treatment ascertained within 1 year of SVR†
IFN-based REGIMENS	≥3.25	≥3.25	304(17.4)	2,278	44(14.5)	1.93	1	1
	≥3.25	<3.25	1,448(82.6)	11,955	104(7.2)	0.87	0.44 (0.31–0.62)	0.57 (0.36–0.89)
DAA-only REGIMENS	≥3.25	≥3.25	841(21.2)	2,348	56(6.7)	2.39	1	1
	≥3.25	<3.25	3,135(78.8)	9,296	88(2.8)	0.95	0.40 (0.28–0.55)	0.47 (0.33–0.69)
DROP in FIB-4 post-treatment ascertained at any time after SVR as a time-dependent covariate††
IFN-based REGIMENS	≥3.25	≥3.25	N/A	2,523	37	1.47	1	1
	≥3.25	<3.25	N/A	12,976	120	0.92	0.52 (0.32–0.83)	0.72 (0.40–1.27)
DAA-only REGIMENS	≥3.25	≥3.25	N/A	3,059	50	1.63	1	1
	≥3.25	<3.25	N/A	10,110	103	1.02	0.51 (0.36–0.74)	0.61 (0.41–0.91)

^*Adjusted for antiviral regimen, sex, race/ethnicity, body mass index, HCV genotype, HCV viral load, HIV co-infection, HBV co-infection, type 2 diabetes mellitus, alcohol use disorders, AUDIT-C score, substance use disorder, cirrhosis complications, serum bilirubin, serum creatinine, serum albumin, blood INR and blood hemoglobin levels. The laboratory tests were categorized into quartiles and modeled as dummy categorical variables.

^†This means at least one FIB4 <3.25 within one year after treatment

^††In this analysis, a patient who drops from FIB-4 ≥3.25 before SVR to <3.25 at time t after SVR is analyzed under the ≥3.25 category from time 0 until time t and under the <3.25 category after time t.

Figure 3. HCC incidence in patients after SVR according to change in FIB-4 score from pre-treatment to within one year after SVR, presented separately for patients with pre-treatment cirrhosis (A-D) or without pre-treatment cirrhosis (E-H).

Panels A and E show that a drop in FIB-4 score from ≥3.25 pre-treatment to <3.25 within one year after SVR identifies patients with lower HCC incidence than those with a FIB-4 score ≥3.25 both pre-treatment and after SVR, both in patients with (A) and without (E) pretreatment cirrhosis

Panels B and F show that an increase in FIB-4 score from <3.25 pre-treatment to ≥3.25 within one year after SVR identifies patients with higher HCC incidence that those with FIB-4 <3.25 both pre-treatment and after SVR, both in patients with (B) and without (F) pretreatment cirrhosis. However, the absolute risk in patients without pre-treatment cirrhosis is very low.

Panels C and D (patients with pre-treatment cirrhosis) and G and H (patients without pretreatment cirrhosis) demonstrate how HCC risk varies according to pre and post-treatment FIB-4 in DAA-treated and IFN-treated patients

Among patients without pre-treatment cirrhosis and baseline FIB-4 ≥3.25, the majority (~80%) attained a FIB-4 score <3.25 within one year post-SVR and this drop in FIB-4 was again associated with a significant reduction in HCC risk (Table 3 and Figure 3b). The minority of patients whose FIB-4 remained ≥3.25 after SVR had a very high annualized HCC risk (1.93% for IFN-treated and 2.39 for DAA-treated patients) despite the fact that they did not have pre-treatment cirrhosis.

The drop in FIB-4 score after SVR was driven primarily by a reduction/normalization in levels of serum AST and ALT, rather than by an increase in platelet count. A reduction in AST or ALT to a level <30 IU/L was also associated with decrease HCC risk (Supplemental Table 5).

Increase in FIB-score from <3.25 pre-treatment to ≥3.25 after SVR is associated with increased HCC risk

Among DAA-treated patients with pre-treatment cirrhosis and baseline FIB-4 score <3.25, only a small minority (9.5%) experienced an increase to a FIB-4 score ≥3.25 within one year post-SVR, while the majority maintained a persistently low FIB-4 below 3.25 (Table 4). DAA-treated patients whose FIB-4 increased after SVR to a level ≥3.25 had a higher annualized HCC incidence than those who remained persistently <3.25 (2.29% vs. 1.02%, AHR 2.18, 95% CI 1.16–4.09) (Table 4 and Figure 3c). The difference was even more pronounced when the increase in FIB-4 was analyzed as a time-dependent variable: patients whose FIB-4 increased to a level ≥3.25 at some point after treatment had an annualized HCC incidence after that time point of 2.75%, as compared to 0.93% for patients with FIB-4 persistently <3.25 (AHR 3.17, 95% CI 1.87–5.37) (Table 4). Similar results were noted among IFN-treated patients (Table 4).

Table 4.

Association between increase in FIB-4 score from <3.25 before SVR to ≥3.25 after SVR and increased HCC risk

	FIB-4 Score Pre-treatment	FIB-4 Score Post-treatment	Number of patients	Patient-years	Number who developed HCC (%)	HCC per 100 patient-years	Crude hazard ratio (95% CI)	Adjusted* hazard ratio (95%CI)
CIRRHOSIS
INCREASE in FIB-4 post-treatment ascertained within 1 year of SVR†
IFN-based REGIMENS	<3.25	<3.25	957(86.8)	7,397	38(4.0)	0.51	1	1
	<3.25	≥3.25	146(13.2)	1,040	14(9.6)	1.35	2.57 (1.39–4.75)	3.20 (1.49–6.88)
DAA-only REGIMENS	<3.25	<3.25	2,070(90.6)	6,360	65(3.1)	1.02	1	1
	<3.25	≥3.25	216(9.4)	609	14(6.5)	2.3	2.26 (1.27–4.02)	2.21 (1.18–4.17)
INCREASE in FIB-4 post-treatment ascertained at any time after SVR as a time-dependent covariate††
IFN-based REGIMENS	<3.25	<3.25	N/A	7,545	28	0.37	1	1
	<3.25	≥3.25	N/A	1,683	31	1.84	5.27(3.09–9.00)	7.36 (3.81–14.24)
DAA-only REGIMENS	<3.25	<3.25	N/A	6,689	62	0.93	1	1
	<3.25	≥3.25	N/A	909	25	2.75	3.19(1.98–5.13)	3.16 (1.86–5.36)
NO CIRRHOSIS
INCREASE in FIB-4 post-treatment ascertained within 1 year of SVR†
IFN-based REGIMENS	<3.25	<3.25	12,321(94.6)	118,075	157(1.3)	0.13	1	1
	<3.25	≥3.25	703(5.4)	5,967	21(3.0)	0.35	2.69 (1.70–4.24)	2.08 (1.20–3.59)
DAA-only REGIMENS	<3.25	<3.25	13,395(95.7)	40,178	98(0.7)	0.24	1	1
	<3.25	≥3.25	596(4.3)	1,699	8(1.3)	0.47	1.92 (0.93–3.94)	1.65 (0.76–3.61)
INCREASE in FIB-4 post-treatment ascertained at any time after SVR as a time-dependent covariate††
IFN-based REGIMENS	<3.25	<3.25	N/A	128,183	117	0.09	1	1
	<3.25	≥3.25	N/A	14,656	80	0.55	5.66 (4.22–7.58)	4.85 (3.43–6.84)
DAA-only REGIMENS	<3.25	<3.25	N/A	45,574	85	0.19	1	1
	<3.25	≥3.25	N/A	2,772	25	0.9	4.18 2.66–6.59)	4.10 (2.54–6.62)

^*Adjusted for antiviral regimen, sex, race/ethnicity, body mass index, HCV genotype, HCV viral load, HIV co-infection, HBV co-infection, type 2 diabetes mellitus, alcohol use disorders, AUDIT-C score, substance use disorder, cirrhosis complications, serum bilirubin, serum creatinine, serum albumin, blood INR and blood hemoglobin levels. The laboratory tests were categorized into quartiles and modeled as dummy categorical variables.

^†This means at least one post-treatment FIB4 ≥3.25 within one year after treatment

^††In this analysis, a patient whose FIB-4 rises from <3.25 before SVR to ≥3.25 at time t after SVR is analyzed under the <3.25 category from time 0 until time t and under the ≥3.25 category after time t.

Among patients without pre-treatment cirrhosis and baseline FIB-4 score <3.25, only a very small minority (4.8%) experienced an increase to a FIB-4 score ≥3.25 within one year post-SVR, and this was associated with a significantly higher risk of HCC (Table 4 and Figure 3d). However, the absolute risk of HCC even in the patients whose FIB-4 increased was low (0.35%/yr in IFN-treated and 0.47%/yr in DAA-treated patients).

Patients with FIB-4≥3.25 both before and after SVR had the highest risk and those with FIB-4<3.25 both before and after SVR had the lowest risk

Among patients with pre-treatment cirrhosis, only the group that had FIB-4 <3.25 both before and after SVR had an absolute HCC incidence <1% per year (Table 4 and Figure 3e and 3g). Conversely, patients who had persistently elevated FIB-4 ≥3.25 both before and after SVR had the highest risk, e.g., ~5% per year for DAA-treated patients. Those with dynamic FIB-4 (either decreasing below 3.25 or increasing above 3.25) have intermediate risk.

Among patients without pre-treatment cirrhosis, the same pattern of decreasing HCC risk was observed going from patients whose FIB-4 was ≥3.25 both before and after SVR to those whose FIB-4 was <3.25 both before and after SVR. The group with FIB-4 ≥3.25 both before and after SVR had a high absolute HCC risk (>2% per year) despite not having recognized cirrhosis.

Discussion

The majority of patients with HCV either already have undergone antiviral treatment with DAAs or are expected to in the near future, and almost all will experience SVR. In turn, SVR is associated with improved clinical outcomes including a reduced risk of HCC^1–5. Although SVR reduces the risk of HCC, patients with established cirrhosis prior to treatment retain a substantial residual risk of HCC after SVR. The utility of HCC surveillance depends on a precise estimate of that residual risk for which data are lacking. In this study of 48,135 patients from the VA healthcare system who achieved SVR and were followed for 5.4 years, we extend our knowledge of the post-SVR HCC risk in several ways. First, we show that both pre-treatment (baseline) FIB-4 level as well as post-treatment (after SVR) FIB-4 level can be combined for HCC risk estimation. For example, among DAA-treated patients with cirrhosis, a persistently high FIB-4 ≥3.25 both pre and post-treatment has the highest HCC risk (5.07%/yr), while persistently low FIB-4<3.25 both pre and post-treatment has the lowest HCC risk (1.0%/yr). Patients with cirrhosis and dynamic FIB-4 (either decreasing from ≥3.25 to <3.25 [HCC incidence 2.47%] or increasing from <3.25 to ≥3.25 [HCC incidence 2.29%]) have intermediate risk. Second, we show that changes in FIB-4 score after SVR correlate with changes in HCC risk. A decrease in FIB-4 after SVR from > 3.25 to < 3.25 is associated with a decrease in HCC risk, whereas an increase from <3.25 to ≥3.25 is associated with an increase in HCC risk. Third, we found that patients without a pre-treatment diagnosis of cirrhosis who have a FIB-4 ≥3.25 have a high risk of HCC (1.22%/yr) especially if FIB-4 remains ≥3.25 after SVR (2.39%/yr in DAA-treated patients) and therefore merit HCC surveillance. Finally, we demonstrate that annual HCC risk is declining over time in patients with cirrhosis during up to 4 years of follow-up after DAA-induced SVR, although it remains high in those with FIB-4>3.25. However, among IFN-treated cirrhotic patients who achieved SVR and had much longer follow-up (>10 years), there was no reduction in HCC risk over time during long-term follow-up. IFN-treated cirrhotic patients with baseline FIB-4 ≥3.25 continued to have an annual HCC risk >2% for up to 10 years after SVR. This would suggest that patients with cirrhosis and baseline FIB-4 ≥3.25 might be at high risk of HCC requiring surveillance “indefinitely” after SVR, until data with longer follow-up have accrued among DAA-treated patients.

Using FIB-4 level before and after SVR to stratify HCC Risk

Although the FIB-4 score was developed as a measure of advanced fibrosis of the liver, it is also emerging as a convenient, readily available marker of HCC risk in patients with HCV-related cirrhosis^{3, 29}. Prior studies have shown that FIB-4 is strongly associated with the risk of HCC in viremic patients with HCV, in patients with NAFLD with and without cirrhosis, and in patients with chronic hepatitis B^44–46. Our data demonstrate that by combining both pre-SVR and post-SVR FIB-4 levels (<3.25 vs. ≥3.25) we can stratify patients with pre-treatment cirrhosis and SVR into very distinct risk categories (Figure 3g). Patients with FIB-4 ≥3.25 both pre and post-SVR have extremely high HCC risk (5.08%/yr), those with FIB-4 <3.25 both pre and post-SVR have lower risk (1.02%/yr), while those with dynamic FIB-4 have intermediate HCC risk ranging from 2 to 2.5%/yr.

Patients without a pre-treatment diagnosis of cirrhosis have a much lower absolute risk of HCC after SVR. However, we showed that pre and post-SVR FIB-4 score can still stratify HCC risk in these patients (Figure 3h). In doing so we identified that patients with pre-SVR FIB-4 ≥3.25 have a relatively high HCC risk (1.22%/yr), especially if FIB-4 remains above 3.25 after SVR (HCC risk 2.39%/yr). Therefore, patients with FIB-4 ≥3.25, who represented 21.8% of all patients without pre-treatment cirrhosis diagnosis, should be considered for HCC surveillance even if they do not carry a diagnosis of cirrhosis. It is possible that some of these patients had unrecognized cirrhosis, which has been reported in up to one-fourth of patients at time of HCC presentation⁴⁷. This issue of misclassification may be particularly true of patients with compensated cirrhosis and no portal hypertension, in whom the development of cirrhosis can be clinically silent.

What are the implications of a drop or increase in FIB-4 score after SVR?

Our results demonstrate that a drop in FIB −4 (from ≥3.25 pre-SVR to <3.25 post-SVR) was consistently associated with a reduced HCC risk across all subgroups tested, both among patients with and without cirrhosis (Table 3). The vast majority of these improvements in FIB-4 occurred within the first year after SVR (as shown by other studies⁴⁸), suggesting that they reflect improvements in necroinflammation more than fibrosis stage. Alternatively, the drop in FIB-4 immediately after SVR may simply be a marker of patients who had less advanced cirrhosis or portal hypertension than the patients whose FIB-4 remained elevated. Regardless of the underlying mechanism, a drop in FIB-4 level is a useful marker of reduced HCC risk.

Conversely, an increase in FIB-4 score (from <3.25 pre-SVR to ≥3.25 post-SVR), although a relatively uncommon event, was consistently associated with a substantial increase in HCC risk (Table 4). It is unclear if this increase in FIB-4 was a marker of unrecognized cirrhosis, fibrosis progression after SVR, or a mechanism independent of fibrosis.

FIB-4 and cost-effectiveness of HCC screening following SVR

The incremental cost-effectiveness ratio of HCC surveillance following SVR in patients with cirrhosis or advanced fibrosis has been shown to increase approximately exponentially with decreasing HCC incidence^{8, 9}. This makes sense since only patients who are destined to develop HCC can potentially benefit from surveillance, whereas all others only incur costs and potential harms. Therefore, if HCC incidence drops below a certain threshold following SVR, HCC surveillance may no longer be considered cost-effective. It was estimated using a hypothetical Markov model that if HCC incidence dropped to a value <1.32% per year, then the ICER for biannual ultrasound-based surveillance would exceed the traditional cost-effectiveness benchmark of $50,000/quality adjusted life-year (QALY). Our results demonstrate that DAA-treated patients with cirrhosis and baseline FIB-4 ≥3.25 have an HCC incidence of 3.66%/yr, well above the accepted minimum threshold. The annual HCC incidence of these patients with high baseline FIB-4 is still greater than 1.32%/yr even 4 years after SVR, including among those who subsequently drop to FIB-4 <3.25. The only group of patients with cirrhosis that had HCC incidence below 1% were those with FIB-4 <3.25 both before and persistently after treatment. Also, it appeared that by year 4 after SVR, patients with baseline FIB-4 <3.25 had an HCC incidence that was <1%. Nonetheless it would be unwise for such patients to forego screening until both our results and those of the cost-effectiveness analyses are confirmed. Also, if the acceptable ICER benchmark is considered to be higher than $50,000/QALY, then the HCC incidence required to result in cost-effective screening would be much lower. For example, an annual HCC incidence >0.55%/yr was estimated to result in an ICER of approximately $100,000/QALY in cirrhotic patients with SVR⁸.

In our study, we could not identify reliably patients with stage 3 (F3) fibrosis (i.e., advanced, pre-cirrhotic fibrosis), since this requires liver biopsy, which is nowadays rarely performed especially in the VA and other healthcare systems that do not require evidence of advanced fibrosis for antiviral treatment eligibility. There are currently discrepant guidelines as to whether patients with F3 fibrosis should undergo HCC surveillance or not after SVR^{7, 49, 50}. Furthermore, it is notoriously difficult to identify F3 fibrosis without histological confirmation. Our data suggest that instead of trying to estimate fibrosis stage, a more accurate and convenient method to stratify post-SVR risk in patients who do not carry a diagnosis of cirrhosis is by using their pre and post-SVR FIB-4 level. These patients should be offered screening if their pre-SVR FIB-4 score is ≥3.25 and especially if it remains ≥3.25 after SVR.

Is there a Decline in HCC Risk over Time After SVR?

We expected to see a steady decline in HCC risk during long-term follow-up after SVR. It is tempting to conclude that HCC risk appears to be declining during the first 4 years after SVR, looking at DAA-treated patients with pre-treatment cirrhosis (Figure 2a). However, in IFN-treated patients, who had much longer follow-up, no such decline was noted, and in fact, HCC incidence remained >2% per year even 10 years after SVR in patients with a high baseline FIB-4 ≥3.25. The cause of this difference is unclear. It may be related to the fact that the IFN-treated patients were a heterogenous group, with inherently different degrees of liver dysfunction and HCC risk than DAA-treated patients, that spanned >15 years of treatment. Ultimately more years of follow-up in the DAA-treated patients will be needed to confirm whether the risk is declining, which had not accrued at the time we analyzed our data despite the fact that we extended follow-up to as recent a time as possible (02/14/2019).

Study Limitations

Our results were derived from a predominantly male, VA patient population. Although the majority of HCV-infected with cirrhosis is indeed male in most healthcare systems, our results should be validated in other populations with a higher representation of women. We derived our data retrospectively from VA electronic health records. Prospective follow-up of patients after SVR might be expected to result in even more accurate ascertainment of exposures and outcomes. However, prospective follow-up of thousands of patients with cirrhosis for many years after SVR is not feasible in practice. The diagnosis of cirrhosis was based on validated diagnostic codes recorded by healthcare providers. As such, the diagnosis was not based on uniform histological or other criteria and occult cirrhosis could have been missed. Our study benefitted from a large sample size, derived from a national comprehensive healthcare system with detailed ascertainment of all baseline characteristics known or suspected to be associated with HCC risk.

Conclusions

Pre-treatment FIB-4 score (≥3.25 vs. <3.25), together with the change in FIB-4 score after SVR, can be used as a convenient, readily available method of stratifying HCC risk in patients with HCV who achieve SVR. Changes in FIB-4 score reflect changes in HCC risk. Patients with established cirrhosis appear to have a persistently high risk of HCC even many years after SVR and should continue HCC surveillance indefinitely. Among patients with cirrhosis, only those whose FIB-4 level is <3.25 both before and persistently after SVR have an annual HCC risk <1%. Patients without a pre-treatment diagnosis of cirrhosis generally have low HCC risk after SVR, except those with pre-SVR FIB-4 ≥3.25 and especially if post-SVR FIB-4 remains ≥3.25. These patients should be offered HCC surveillance.

What you need to know.

BACKGROUND AND CONTEXT

It is unclear if hepatocellular carcinoma (HCC) risk decreases after hepatitis C virus (HCV) eradication. We analyzed changes in HCC annual incidence over time following HCV eradication and searched for markers of HCC risk

NEW FINDINGS

Patients with cirrhosis before an SVR to treatment for HCV infection continue to have a high risk for HCC (>2%/year) for many years, even if their FIB-4 score decreases. Patients without cirrhosis before treatment and FIB-4 scores ≥3.25 also have a high risk of HCC.

LIMITATIONS

This was a retrospective study of patients in the Veterans Health Administration.

IMPACT

Patients with cirrhosis before an SVR to HCV treatment should continue surveillance for HCC. Patients without cirrhosis before treatment and FIB-4 scores ≥3.25 should also undergo HCC surveillance