Direct–acting Antiviral Treatment for Hepatitis C Virus Infection and Risk of Incident Liver Cancer

A Retrospective Cohort Study

A. W. Singer; K. R. Reddy; L. E. Telep; A. O. Osinusi; D. M. Brainard; M. Buti; A. P. Chokkalingam

Aliment Pharmacol Ther. 2018;47(9):1278-1287.

Abstract and Introduction

Abstract

Background Eradication of hepatitis C virus (HCV) infection via interferon–based treatment lowers hepatocellular carcinoma risk; some research suggests this effect extends to interferon–free treatment.

Aims The objective of this retrospective cohort study was to examine the association of direct–acting antiviral (DAA) exposure with risk of incident liver cancer in real–world data.

Methods From United States administrative claims data through March 31, 2017, we identified 30 183 adult HCV patients exposed to DAAs. For comparison, we identified contemporary adult HCV patients without evidence of HCV treatment (N = 137 502), and historical HCV patients treated with interferon prior to the introduction of DAAs (N = 12 948). Included patients had at least 12 months of prior enrolment and no evidence of prior liver cancer at baseline. Hazard ratios (HRs) estimating risk of incident liver cancer associated with DAA treatment were calculated using Cox proportional hazards methods.

Results Relative to untreated HCV patients, DAA–treated patients were older, more likely to be male, and more likely to have cirrhosis at baseline. After adjustment, DAA treatment was associated with a significantly reduced risk of liver cancer relative to no treatment (adjusted HR = 0.84, 95% CI: 0.73–0.96), and relative to interferon–based treatment in the pre–DAA era (HR = 0.69, 95% CI: 0.59–0.81).

Conclusions In this large, population–based study, DAA–based treatment was associated with a reduced risk of incident liver cancer relative to both no HCV treatment and to interferon–based treatment in the pre–DAA era. As additional follow–up time of DAA–treated patients accrues, we anticipate that the long–term benefits of DAA treatment will become more apparent.

Introduction

Chronic infection with hepatitis C virus (HCV) is a leading cause of hepatocellular carcinoma (HCC) globally.^[1] Although the presence of fibrosis and cirrhosis are the primary risk factors for the development of HCC among HCV–infected persons, a number of other patient characteristics, including older age, male sex, diabetes, obesity, smoking, HCV genotype 3 infection, heavy alcohol use and hepatitis B virus co–infection, have been demonstrated to influence HCC risk.^[2–8]

A substantial body of evidence indicates that sustained virological response (SVR) following treatment with interferon–based therapy dramatically reduces, but does not eliminate, the risk of developing HCC.^[2,9–13] A recent meta–analysis found an approximate 75% reduction in risk following SVR among HCV patients at all stages of fibrosis,^[14] though HCV patients still remain at increased risk of HCC for many years following attainment of SVR.^[11,15]

Based on the evidence of a reduced risk of HCC following SVR with interferon–based therapy, it has been postulated that the uptake of direct–acting antiviral (DAA) therapy, with its associated high SVR rates and ability to treat populations at the highest risk of HCC, will lead to reductions in HCC incidence among HCV–infected populations.^[16,17] In contrast, some recent studies have reported unexpected high incidence^[18,19] and recurrence^[20] rates of HCC following treatment with DAA therapy, while other studies have found no increased rates or risk of HCC incidence or recurrence^[21–26] following DAA treatment. Conclusions from these studies have been complicated by methodological limitations, such as the lack of appropriate control groups, potential confounding and limited follow–up time.^[22,27,28] The seriousness of HCC as a potential treatment–related outcome, contrasted with the serious implications of withholding HCV treatment from sick patients due to a potentially unfounded concern of an increased risk of HCC development after DAA therapy,^[27] warrants further investigation.

The objective of this study was to examine whether DAA therapy in HCV patients has an impact on the development of liver cancer when compared to untreated and interferon–treated HCV patients in a large, population–based administrative claims database from the United States.

1 2 3 4

Next Section

Abstract and Introduction
Materials and Methods
Results
Discussion
References

Table 1. Baseline characteristics of untreated HCV patients, pre–DAA interferon–treated patients and DAA–treated patients
	Untreated HCV patients	Pre–DAA IFN–treated patients	DAA–treated patients
	N (%)	N (%)	N (%)
Total	137 502 (100)	12 948 (100)	30 183 (100)
Age, in years
18–34	19 583 (14.24)	1000 (7.72)	2061 (6.83)
35–44	15 822 (11.51)	1704 (13.16)	2086 (6.91)
45–54	33 514 (24.37)	5973 (46.13)	6655 (22.05)
55+	68 583 (49.88)	4271 (32.99)	19 381 (64.21)
Sex
Female	60 710 (44.15)	4701 (36.31)	10 873 (36.02)
Male	76 792 (55.85)	8247 (63.69)	19 310 (63.98)
Baseline comorbidities
Chronic obstructive pulmonary disease	42 728 (31.07)	2433 (18.79)	8147 (26.99)
Smoking	61 301 (44.58)	3060 (23.63)	12 514 (41.46)
Prior nonhepatic cancer	28 887 (21.01)	2096 (16.19)	7812 (25.88)
Cirrhosis/fibrosis	15 611 (11.35)	2946 (22.75)	9405 (31.16)
Anaemia	34 911 (25.39)	2267 (17.51)	7421 (24.59)
Arrhythmia	23 175 (16.85)	1065 (8.23)	4890 (16.20)
Overweight/obese	25 193 (18.32)	1007 (7.78)	5445 (18.04)
Autoimmune disease	8437 (6.14)	575 (4.44)	1716 (5.69)
Coronary atherosclerosis	14 996 (10.91)	932 (7.20)	2994 (9.92)
Hyperlipidaemia	60 326 (43.87)	4334 (33.47)	12 440 (41.22)
Hepatitis B infection	10 010 (7.28)	740 (5.72)	1587 (5.26)
Mental illness	63 821 (46.41)	3112 (24.03)	12 547 (41.57)
Alcoholic fatty liver disease	764 (0.56)	65 (0.50)	198 (0.66)
Alcoholic hepatitis	1874 (1.36)	102 (0.79)	341 (1.13)
Alcoholic liver damage	5125 (3.73)	449 (3.47)	1728 (5.73)
Portal hypertension	4255 (3.09)	428 (3.31)	2535 (8.40)
Hepatorenal syndrome	306 (0.22)	21 (0.16)	88 (0.29)
End–stage liver disease	5191 (3.78)	500 (3.86)	1756 (5.82)
Thrombocytopenia	9683 (7.04)	862 (6.66)	3954 (13.10)
Diabetes	32 517 (23.65)	2287 (17.66)	7073 (23.43)
Immunomodulator use	2206 (1.60)	168 (1.30)	697 (2.31)
Betablocker use	27 477 (19.98)	2141 (16.54)	7240 (23.99)
Anti–hypertensive use	47 027 (34.20)	3574 (27.60)	11 807 (39.12)
Statin use	25 252 (18.36)	910 (7.03)	3 949 (13.08)
Prior radiological procedure	1571 (1.14)	71 (0.55)	353 (1.17)

DAA, direct–acting antiviral; IFN, interferon.

Table 2. Risk estimates for liver cancer associated with DAA treatment vs no HCV treatment
Exposure group	Patients (N)	Person–years (N)	Events (N)	Unadjusted incidence rate per 100 PY (95% CI)	Unadjusted HR (95% CI)	Adjusted, IPTW corrected HR^a (95% CI)
Untreated	137 502	192 669	1232	0.64 (0.60–0.68)	1.00 (ref)	1.00 (ref)
DAA–treated (total)	30 183	36 599	433	1.18 (1.07–1.30)	1.84 (1.65–2.05)	0.84 (0.73–0.96)
DAA–treated, IFN–experienced^b	1579	2654	46	1.73 (1.27–2.31)	2.74 (2.04–3.68)	0.76 (0.50–1.16)
DAA–treated, IFN–naive^b	28 604	33 946	387	1.14 (1.03–1.26)	1.77 (1.58–1.98)	0.85 (0.74–0.97)
DAA–treated, IFN–containing^c	1918	3633	25	0.69 (0.45–1.02)	1.10 (0.74–1.63)	0.68 (0.43–1.09)
DAA–treated, IFN–free^c	28 265	32 967	408	1.24 (1.12–1.36)	1.92 (1.72–2.15)	0.85 (0.74–0.98)

CI: confidence interval; DAA: direct–acting antiviral; HR: Hazard ratio; IFN: interferon; IPTW: Inverse probability of treatment weighting; PY: person–years.
^aAdjusted for age (18–35, 35–44, 45–54, 55 + years), sex, prior diagnoses of COPD, tobacco use, liver cirrhosis/fibrosis, arrhythmia, overweight/obesity, hyperlipidaemia, hepatitis B infection, mental illness, alcoholic hepatitis, alcoholic liver damage, portal hypertension, end–stage liver disease, thrombocytopenia and diabetes; and prior use of immunomodulators, beta blockers, anti–hypertensives, and statins. Inverse probability of treatment weights calculated as the inverse of propensity of DAA treatment vs no treatment on basis of baseline conditions.
^bIFN–experienced: evidence of prior IFN use; IFN–naive: no evidence of prior IFN use.
^cIFN–containing: IFN dispensed ± 14 days of DAA start; IFN–free: no evidence of IFN dispensed ± 14 days of DAA start

Table 3. Risk estimates (HRs) for liver cancer associated with DAA treatment and other risk factors
Variable	HR (95% CI)	P–value
DAA (vs untreated)	0.84 (0.73–0.96)	0.0088
Age 55 + (vs 18–34)	9.88 (5.93–16.45)	<0.0001
45–54 (vs 18–34)	5.81 (3.47–9.75)
35–44 (vs 18–34)	2.67 (1.50–4.76)
Sex (M)	1.84 (1.64–2.06)	<0.0001
Chronic obstructive pulmonary disease	0.87 (0.78–0.97)	0.0099
Tobacco use	1.30 (1.17–1.43)	<0.0001
Cirrhosis/fibrosis	3.89 (3.45–4.37)	<0.0001
Arrhythmia	0.83 (0.73–0.94)	0.0029
Overweight/obese	0.85 (0.75–0.97)	0.0149
Hyperlipidaemia	0.82 (0.73–0.91)	0.0002
Hepatitis B infection	1.16 (0.98–1.37)	0.0765
Mental illness	0.74 (0.67–0.83)	<0.0001
Alcoholic hepatitis	0.77 (0.58–1.01)	0.0608
Alcoholic liver damage	1.17 (1.01–1.35)	0.0383
Portal hypertension	1.48 (1.28–1.70)	<0.0001
End–stage liver disease	1.16 (1.00–1.35)	0.0513
Thrombocytopenia	1.96 (1.74–2.21)	<0.0001
Diabetes	1.20 (1.08–1.34)	0.0009
Use of immunomodulators	0.65 (0.42–1.02)	0.0611
Use of beta blockers	1.32 (1.18–1.47)	<0.0001
Use of anti–hypertensives	1.12 (1.01–1.25)	0.0336
Use of statins	0.66 (0.56–0.77)	<0.0001

CI, confidence interval; DAA, direct–acting antiviral; HR, Hazard ratio.

Table 4. Risk estimates (HRs) for liver cancer associated with DAA treatment vs pre–DAA interferon treatment
Exposure group	Patients (N)	Person– years (N)	Events (N)	Unadjusted incidence rate per 100 PY (95% CI)	Unadjusted HR (95% CI)	Adjusted, IPTW corrected HR^a (95% CI)
Pre–DAA IFN–treated	12 948	39 488	388	0.98 (0.89–1.09)	1.00 (ref)	1.00 (ref)
DAA–treated (Total)	30 183	36 599	433	1.18 (1.07–1.30)	1.22 (1.04–1.42)	0.69 (0.59–0.81)
DAA–treated, IFN–experienced^b	1579	2654	46	1.73 (1.27–2.31)	1.76 (1.29–2.41)	0.64 (0.46–0.91)
DAA–treated, IFN–naive^b	28 604	33 946	387	1.14 (1.03–1.26)	1.17 (1.00–1.37)	0.70 (0.60–0.82)
DAA–treated, IFN–containing^c	1918	3633	25	0.69 (0.45–1.02)	0.70 (0.47–1.06)	0.48 (0.31–0.73)
DAA–treated, IFN–free^c	28 265	32 967	408	1.24 (1.12–1.36)	1.29 (1.10–1.52)	0.72 (0.61–0.84)

CI: confidence interval; DAA: Direct–acting antiviral; HR: Hazard ratio; IFN: interferon; IPTW: Inverse probability of treatment weighting; PY: person–years
^aAdjusted for age (18–35, 35–44, 45–54, 55 + years), sex, prior diagnoses of cancer, liver cirrhosis/fibrosis, anaemia, arrhythmia, autoimmune disease, coronary atherosclerosis, mental illness, alcoholic fatty liver disease, alcoholic liver damage, portal hypertension, hepatorenal syndrome and thrombocytopenia; prior use of beta blockers and statins; and prior radiological procedures. Inverse probability of treatment weights calculated as the inverse of propensity of DAA treatment vs pre–DAA IFN on basis of baseline conditions.
^bIFN–experienced: evidence of prior IFN use; IFN–naive: no evidence of prior IFN use.
^cIFN–containing: IFN dispensed ± 14 days of DAA start; IFN–free: no evidence of IFN dispensed ± 14 days of DAA start.