Changes in Alcohol Consumption Following Direct-Acting Antiviral Treatment for Hepatitis C in VA Patients with Comorbid Alcohol Use Disorder and PTSD

Abstract

Objective:

To investigate whether direct acting antivirals (DAA) for hepatitis C viral infection (HCV): glecaprevir/pibrentasvir (GLE/PIB), ledipasvir/sofosbuvir (LDV/SOF) and sofosbuvir/velpatasvir (SOF/VEL) are associated with reduced alcohol consumption among veterans with alcohol use disorder (AUD) and co-occurring post-traumatic stress disorder (PTSD).

Methods:

We measured change in Alcohol Use Disorder Identification Test-Consumption Module (AUDIT-C) scores in a retrospective cohort of veterans with PTSD and AUD receiving DAAs for HCV.

Results:

1,211 patients were included (GLE/PIB n = 174, LDV/SOF n = 808, SOF/VEL n = 229). Adjusted frequencies of clinically meaningful improvement were 30.5% for GLE/PIB, 45.5% for LDV/SOF, and 40.5% for SOF/VEL. The frequency was lower for GLE/PIB than for LDV/SOF (OR = 0.59; 95% CI [0.40, 0.87]) or SOF/VEL (OR = 0.66; 95% CI [0.42, 1.04]).

Conclusions:

DAA treatment for HCV was associated with a substantial reduction in alcohol use in patients with AUD and co-occurring PTSD. Further exploration of the role of DAAs in AUD treatment is warranted.

Introduction

Alcohol use disorder (AUD) is a common substance use disorder (SUD) in the United States (US), afflicting 5% of persons over age 12 and associated with significant disability (Tucker et al., 2020). Among the US population, an estimated 42% of people with AUD have a co-occurring post-traumatic stress disorder (PTSD) diagnosis (Pietrzak et al., 2011). Among those with PTSD, the prevalence of AUD ranges from 9.8% to 61.3% (Debell et al., 2014). The association of AUD and PTSD is especially strong among US military veterans. In this group, 63% of those with SUD have co-occurring PTSD, and over 75% of veterans with PTSD have co-occurring SUD (Seal, Cohen, Bertenthal, et al., 2011). There is a 4-fold increase in the odds of a comorbid AUD diagnosis for Iraq and Afghanistan Veterans with PTSD (Seal, Cohen, Waldrop, et al., 2011). One hypothesis for this high rate of co-occurrence is self-medication, where PTSD leads to drinking in order to alleviate trauma-related symptoms (Kwako et al., 2015; McLean et al., 2015). This problem may be compounded by limited access to evidence-based treatments for PTSD (Maguen et al., 2020; Shiner et al., 2018; Watkins et al., 2018).

Hepatitis C virus (HCV) is spread through blood contact (often via needle sharing), infecting the liver and resulting in acute illness, with nearly half of patients proceeding to develop chronic infection associated with cirrhosis and liver cancers (Memon & Memon, 2002; Westbrook & Dusheiko, 2014). HCV-infected liver cells are known to activate an immune cell protein called the NLRP3 inflammasome, which can lead to chronic systemic inflammation (Chen et al., 2014; Ramachandran et al., 2021). The role of hyperactive NLRP3 is well established in many diseases including various cancers, metabolic disorders, autoimmune disorders, and Alzheimer’s disease (Fusco et al., 2020; Sharma & Kanneganti, 2021; Wang et al., 2020).

Chronic HCV is more prevalent in persons with AUD than in the general population (Marshall, 2021), and an estimated 44% of veterans with HCV have a history of AUD (Butt et al., 2007). AUD and HCV are common causes of chronic liver disease, and together they have a synergistic effect on its progression (Marshall, 2021). Prior to the advent of current direct-acting antiviral (DAA) treatment for HCV, research suggested that alcohol consumption was associated with a decreased likelihood of sustained virologic response (SVR) or HCV cure (Okazaki et al., 1994). Therefore, clinicians treating HCV would often delay therapy until patients were abstinent from alcohol (Okazaki et al., 1994). Thus, most of the available literature in this area focuses on the impact of alcohol consumption on HCV treatment efficacy, hepatic functioning and rates of hepatocellular carcinoma (Singal & Anand, 2007). Alcohol consumption negatively impacted the previous generation of treatment for HCV, pegylated interferon (Chang et al., 2005). Because of this history, the impact of alcohol consumption on SVR has been carefully monitored with the emergence of DAAs for HCV, and a growing literature now indicates that alcohol consumption has little to no impact on DAA efficacy (Tsui et al., 2016). Despite the growing evidence against withholding DAA treatment until abstinence from alcohol consumption, many payers—and consequently clinicians—continued to endorse this practice (Anand et al., 2006; Barua et al., 2015). In 2015, the Centers for Medicare & Medicaid Services (CMS) issued a notice to relax policies that required abstinence from alcohol and substances in order to obtain DAA treatment for HCV (Herink et al., 2021). In 2016, the Department of Veterans Affairs (VA) followed suit by updating Clinical Practice Guidelines to encourage HCV treatment providers to prescribe DAAs regardless of alcohol or substance use status (Owens et al., 2018). Additionally in 2016, the VA announced the expansion of treatment to all HCV patients regardless of liver disease stage, prior to which, limited resources prioritized HCV treatments to the sickest patients (VA, 2016).

Unexpected but highly promising epidemiological data suggest that two DAAs for HCV, sofosbuvir/velpatasvir (SOF/VEL), which was FDA approved in 2016 (Jackson & Everson, 2017), and glecaprevir/pibrentasvir (GLE/PIB which was FDA-approved in 2017, (Lamb, 2017), are associated with PTSD symptom improvement (Shiner, Forehand, et al., 2022). While all DAAs for HCV are associated with a greater than 90% SVR in VA patients (Belperio et al., 2017), there was no association with PTSD symptom improvement for a third commonly-prescribed DAA called ledipasvir/sofosbuvir (LDV/SOF) which was FDA-approved in 2014 (Kowdley et al., 2014). Subsequent analyses controlling for patient differences between agents indicated that GLE/PIB was the most strongly associated DAA with PTSD symptom improvement (Shiner, Huybrechts, et al., 2022). Although it was of interest to explore the mediating role of common comorbid psychiatric disorders such as AUD underlying the association of DAA treatment with PTSD symptom improvement, the sample size was too small for formal mediation analyses.

The biological mechanism by which DAAs could have psychiatric effects is unclear. However, emerging evidence has linked NLRP3 activation to various psychiatric disorders likely due to its role in neuroinflammation (Çelik et al., 2022). It has been established that NLRP3 can be activated by pathogens, such as HCV, or environmental stressors, such has fear for one’s life (Cheon et al., 2020). In animal models of PTSD and anxiety, NLRP3 activation has demonstrated association with dysregulation of fear memory and response (Dong et al., 2020). Furthermore, NLRP3 inhibition has shown improved anxiety (Yamanashi et al., 2020). Administration of alcohol in animal models has exhibited inhibition of NLRP3 (Hoyt et al., 2016), thereby supporting the PTSD/AUD self-medication hypothesis. Additionally, NLRP3 inhibition has demonstrated reduced alcohol consumption in mice (Lowe et al., 2020). Thus, theoretically, if a medication were to inhibit NLRP3 activity, it is plausible that it could simultaneously inhibit HCV replication, fear response, and alcohol consumption. While existing DAAs are only classified as acting on virus, there is some evidence of off-target effects (Giovannini et al., 2020).

The goal of this study was to identify individuals with comorbid PTSD, AUD, and HCV who completed a standardized assessment of alcohol consumption before and after DAA treatment and assess clinically meaningful change in alcohol consumption between the specific DAA treatments. To our knowledge, there has only been one study evaluating the impact of DAA treatment on alcohol consumption. A retrospective cohort study of VA patients receiving DAAs for HCV from 2013–2015 (prior to the FDA approval of GLE/PIB and SOF/VEL) found that the overall frequency of alcohol misuse declined from 9.1% to 7.7% following DAA treatment (Kim et al., 2020). Our study differs in several ways. First, we include the most recently approved DAAs for HCV, GLE/PIB and SOF/VEL, which were not included in the prior work and provide data following the directive discouraging the withholding of DAAs until alcohol abstinence. Second, we specifically focus on patients with diagnoses of AUD and comorbid PTSD with levels of alcohol consumption consistent with current AUD. In addition to the primary goal of understanding the impact of HCV treatment on alcohol use among a sample with co-occurring AUD and PTSD, we further compared DAA treatments with one another to determine if observations are tied to specific treatments, as has been observed for PTSD (Shiner, Huybrechts, et al., 2022), or if DAAs have an overall effect on change in AUD regardless of the specific treatment used.

Methods

Data sources and study sample

To conduct the present study, we utilized a subgroup of the parent cohort of Veterans with PTSD that was constructed for the initial exploratory study (Shiner, Forehand, et al., 2022). Following the parent cohort construction, we identified VA patients with comorbid diagnoses of PTSD, AUD, and HCV between October 1, 1999 and September 30, 2019 using the VA Corporate Data Warehouse (CDW), a national repository of electronic medical record data (Price et al., 2015). We used International Classification of Diseases (ICD) Versions 9 and 10 codes to ascertain diagnoses of PTSD (ICD-9: 309.81; ICD-10: F43.1x), HCV (ICD-9: 070.41, 070.44, 070.51, 070.54, 070.7, 070.70, 070.71; ICD-10: B17.1, B17.10, B17.11, B19.2, B19.20, B19.21, B18.2), and AUD (ICD-9: 303.x, 305.0x; ICD-10: F10.x). We supplemented diagnostic information in the CDW with linked Medicare files when available. We restricted the sample to patients who completed a course of DAA (see direct acting antiviral treatment, below). We further restricted the sample to patients who had standardized assessments of alcohol consumption at baseline and follow-up, as well as heavy alcohol use at baseline (see alcohol consumption measurement, below). We excluded patients whose first AUD diagnosis occurred after DAA initiation to ensure that our study population contained only persons who clearly had active AUD at the time of DAA treatment. The Veterans Institutional Review Board of Northern New England approved this study.

Direct acting antiviral treatment

We identified patients who received at least 56 continuous days of prescription drug fills for the DAA agents LDV/SOF, SOF/VEL, and GLE/PIB. A validation study of pharmacy fill records for DAA agents in the VA CDW found that 98% (95% CI [97%, 100%]) of recorded DAA fills in the VA CDW were confirmed after clinician verification (Rentsch et al., 2018). Patients were included if their DAA trial and related alcohol consumption measurement (see below) was completed by February 2020.

Alcohol Consumption Measurement

We measured alcohol consumption using the Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) Module (Bush et al., 1998), a validated instrument administered annually as part of behavioral health screening in the VA (Shiner et al., 2014). We required a baseline AUDIT-C within year prior to DAA start and a follow-up AUDIT-C between one month and one year after DAA start. We required a baseline score of 8 or higher, indicating a level of alcohol consumption consistent with current dependence (Rubinsky et al., 2013). We examined clinically meaningful improvement as a binary variable (yes/no), defined as a follow-up AUDIT-C score below the threshold (three points or less for men and two points or less for women) for alcohol misuse screening (Bradley et al., 2009). We also examined continuous change in AUDIT-C score from baseline to follow-up. For patients who had multiple AUDIT-C scores in the baseline or follow-up period, we calculated the mean values for each period.

Covariates

We measured seven groups of covariates (Table 1): HCV disease status, prior evidence-based PTSD treatment, timing of AUDIT-C scores relative to DAA initiation, concurrent mental health treatment, patient demographic characteristics at baseline, VA health services use in the year prior to DAA initiation, and comorbidities in the two years prior to DAA initiation.

Table 1:

Explanation of Covariates ^a

Hepatitis C Disease Status
Chronicity	Days from HCV diagnosis to Direct Acting Antiviral medication start b
Fibrosis	FIB-4 score rated as mild, moderate, or advanced c
Prior Evidence-Based PTSD Treatment
Evidence-Based Antidepressants	Receipt of at least 12 weeks of continuous treatment with medications recommended by VA clinical practice guidelines for depression d
Evidence-Based Psychotherapy	Receipt of at least 8 sessions prolonged exposure or cognitive processing therapy over the course of one year e
AUDIT-C Availability
Chronicity	Days from alcohol use disorder diagnosis to Direct Acting Antiviral medication start b
Timing	Days between baseline and follow-up AUDIT-C scores relative to medication start and stop dates
Concurrent Treatments
Evidence-Based Antidepressants	Number of weeks of treatment with medications recommended by VA clinical practice guidelines for depression d
Evidence-Based Psychotherapy	Number of sessions of prolonged exposure or cognitive processing therapy e
Other Medications	Categorical receipt of other antidepressants, anticonvulsants, sedative hypnotics, opioids, atypical antipsychotics, prazosin, as well as medications for alcohol use disorder and opioid agonist treatments.
Patient Characteristics at Baseline
Age	Continuous
Sex	Categorical male or female
Race	Categorical Black Non-Hispanic, Hispanic, White Non-Hispanic, Other
Military Exposures	Combat, Sexual Trauma, Service-Connected Disability
VA Health Service use Characteristics in the Year Preceding Baseline
Outpatient visits	e.g. visits to specialized PTSD clinics or to primary care clinics
Acute psychiatric care use	Emergency department and urgent care visits for psychiatric indications, psychiatric hospitalizations
Residential treatment	Stays in residential PTSD or substance abuse programs
Diagnoses in the Two Years Preceding Baseline
Psychiatric Comorbidities	Number of non-alcohol use disorder diagnoses and three most common diagnoses g
Medical Comorbidities	Number of non-HCV diagnoses and three most common diagnoses g

^aVA=US Department of Veterans Affairs, HCV=Hepatitis C Virus, PTSD=Posttraumatic Stress Disorder, PE=Prolonged Exposure, CPT=Cognitive Processing Therapy

^bVariable was assessed with a look-back period to October 1, 1999.

^cFIB-4 grades level of hepatic fibrosis based on age, platelet counts, and aminotransferase levels (Sterling et al., 2006; Vallet-Pichard et al., 2007).

^dIncluding fluoxetine, paroxetine, sertraline, and venlafaxine (Bernardy N, 2017).

^eMeasured using a natural language processing algorithm that classifies psychotherapy note text (Maguen et al., 2018).

^fCutoff of 8 based on optimal diagnostic efficiency (Rubinsky et al., 2013).

^gUsing a previously published index of mental and physical comorbidities adapted for VA medical record data (Shiner, Peltzman, et al., 2021).

Statistical Analysis

We followed an analytic strategy developed for our prior study of changes in PTSD symptoms among patients with PTSD and HCV (Shiner, Huybrechts, et al., 2022). Briefly, the first step in our analysis was to balance confounders across our DAA medication exposure groups. We considered as confounders covariates with a standardized mean difference (SMD) of greater than or equal to 0.1 for the univariate association with clinically meaningful improvement and for the bivariate relationship between exposure to one DAA versus each of the other DAAs (Austin, 2009). After identification of confounders, we calculated propensity scores representing the probability that a particular trial would be of each DAA combination (Stuart, 2010). We estimated propensity scores with multinomial logistic regression using generalized booster effects (McCaffrey et al., 2013), in which the dependent variable was an indicator for each of the three medications and variables meeting our definition of confounders (i.e., variables associated with the exposure and the outcome with an SMD ≥ 0.1) were the independent variables (McCaffrey et al., 2013; Stuart, 2010). We estimated weights based on the population average treatment effect model.

The second step in our analysis was to compare continuous and categorical AUD outcomes among the three DAA combinations with weighted regression analyses, using DAA combination received as the independent variable. These weighted medication groups were defined by the inverse of the propensity scores and adjusted for covariates that remained unbalanced at the SMD ≥ 0.1 level after propensity score weighting. For our categorical outcome of clinically meaningful improvement, we used weighted logistic regression analysis, whereby the coefficient of the variable tests the hypothesis that each of the three DAA combinations results in the same percentage of patients achieving clinically meaningful improvement. For our continuous outcome of change in AUDIT-C score, we used weighted linear regression analysis, whereby the coefficient of the variable tests the hypothesis that each of the three DAA combinations, coded as a multilevel categorical variable, has the same mean change from baseline to follow-up. We performed data management in SAS version 9.4 (SAS Institute) and statistical modeling in R version 4.0.2 (R core team).

Results

A total of 1,211 patients met our inclusion criteria, including 174 who received GLE/PIB, 808 who received LDV/SOF, and 229 who received SOF/VEL (Appendix Table 1). There were many differences between patients across groups (Table 2), and twelve covariates met our definition of confounders based on having an association with both the exposure and the outcome at the SMD ≥ 0.1 level (Appendix Table 2). Patients receiving GLE/PIB had lower frequency of advanced liver fibrosis, higher frequency of military sexual trauma, and higher frequency of non-AUD SUDs. Patients receiving LDV/SOF had fewer days from first available AUD diagnosis to DAA initiation, more weeks of evidence-based antidepressants recommended by the VA for PTSD (EBAs for PTSD) concurrent with their DAA, fewer emergency department visits for psychiatric indications and psychiatric hospitalizations in the year preceding DAA initiation, as well as higher frequency of diabetes diagnoses in the two years prior to DAA initiation. Patients receiving SOF/VEL had fewer sessions of evidence-based psychotherapy protocols recommended by the VA for PTSD (EBP for PTSD) concurrent with their DAA. Patients in all three groups differed on the timing of their baseline and follow-up AUDIT-C measurements relative to their DAA trial and on whether they received FDA-approved medications for AUD. After propensity score weighting, seven confounders remained unbalanced and were retained as covariates in all weighted outcome models (Table 3). These included lower frequency of advanced fibrosis for the GLE/PIB group, more days from the baseline AUDIT-C score to DAA initiation in the LDV/SOF group, more concurrent sessions of EBP for PTSD in the LDV/SOF group, fewer weeks of concurrent EBA for PTSD in the GLE/PIB group, continued difference across groups in receipt of FDA-approved medications for AUD, higher frequency of non-AUD SUDs in the GLE/PIB group, and higher frequency of diabetes in the LDV/SOF group.

Table 2.

Baseline Patient and Clinical Characteristics ^a

Drug Combination	GLE/PIB (N=174)		LDV/SOF (N=808)		SOF/VEL (N=229)
	n	%	n	%	n	%
Hepatitis Disease Status
Days Since HCV Diagnosis (M±SD)	3035.38±2436.47		3330.59±2044.97		3229.26±2278.13
Fibrosis (FIB-4 Score)
Mild (<1.45)	64	36.8	205	25.4	63	27.5
Moderate (≥1.45 and ≤3.25)	61	35.1	304	37.6	83	36.2
Advanced (>3.25)	26	14.9	181	22.4	55	24.0
HCV cure	157	90.2	775	95.9	208	90.8
Alcohol Use Disorder Status
Days since Alcohol Use Disorder Diagnosis	4074.65±2215.98		3677.85±2020.12		3987.87±2269.33
Prior Evidence-Based PTSD Treatment
Number of Prior EBA Trials (M±SD)	0.55±1.03		0.36±0.65		0.47±0.91
Number of Prior EBP Trials (M±SD)	0.10±0.30		0.07±0.27		0.08±0.31
Alcohol Use Disorder Timing, Version, and Severity
Days from Earliest Baseline AUDIT-C to start of DAA treatment (M±SD)	92.07±89.88		139.29±97.37		112.20±92.63
Days from Baseline AUDIT-C to Follow-Up AUDIT-C (M±SD)	277.55±138.48		329.28±116.32		297.03±137.84
Days from DAA End to Follow-Up AUDIT-C (M±SD)	124.85±102.06		118.96±91.99		113.05±94.09
Baseline AUDIT-C (M±SD)	10.42±1.44		10.25±1.50		10.21+1.48
Concurrent Treatment
Sessions of EBP for PTSD (M±SD)	0.74±2.76		0.63±2.87		0.38±2.13
Weeks of EBA for PTSD (M±SD)	4.90±9,77		6.86±12.30		5.62±10.14
Any Non-EBA Antidepressant	118	67.8	516	63.9	139	60.7
Any Anticonvulsant	41	23.6	148	18.3	35	15.3
Any Sedative/Hypnotics	17	9.8	141	17.5	31	13.5
Any Opioid	55	31.6	306	37.9	58	25.3
Any Atypical Antipsychotic	53	30.5	241	29.8	62	27.1
Any Prazosin	36	20.7	158	19.6	40	17.5
Any FDA Approved Medications for AUD	47	27.0	131	16.2	51	22.3
Any Opioid Agonist Therapy	24	13.8	65	8.0	21	9.2
Patient Characteristics at Baseline
Age (M±SD)	55.31±11.40		58.10±9.21		56.87±10.93
Women	8	4.6	26	3.2	7	3.1
Married	29	16.7	152	18.8	43	18.8
Rural	40	23.0	172	21.3	63	27.5
White Non-Hispanic	112	64.4	408	50.5	161	70.3
Black Non-Hispanic	47	27.0	331	41.0	44	19.2
Hispanic	7	4.0	38	4.7	16	7.0
Other Racial or Ethnic Group	6	3.4	13	1.6	5	2.2
Combat Exposure	41	23.6	162	20.0	57	24.9
Sexual Trauma while in Military	36	20.7	133	16.5	41	17.9
VA Disability Level 70% or Greater	49	28.2	229	28.3	65	28.4
Service Use Characteristics in the 1 Year Preceding Baseline
Number of PTSD Outpatient Clinical Team Visits (M±SD)	2.54±9.43		3.85±16.23		3.10±14.41
Number of Outpatient Mental Health Visits (M±SD)	104.53±98.11		96.58±102.96		85.89±93.07
Number of Outpatient Substance Abuse Visits (M±SD)	30.49±37.29		40.04±58.18		29.21±40.81
Number of Outpatient Primary Care Visits (M±SD)	10.20±9.14		11.23±8.91		9.85±8.38
Number of Outpatient Specialty Medical Visits (M±SD)	3.99±3.15		4.55±3.69		4.51±3.57
Number of ED Visits for Psychiatric Indication (M±SD)	2.26±3.53		1.65±3.17		2.06±3.45
Days of Acute Inpatient Mental Health (M±SD)	61.93±118.33		50.51±100.20		61.94±106.82
Days of Residential PTSD Treatment (M±SD)	9.67±33.56		8.84±33.29		9.27±34.00
Days Residential Substance Abuse Treatment (M±SD)	53.50±89.63		49.13±88.20		49.07±86.53
Comorbidities in the 2 Years Preceding Baseline
Number of Non-Alcohol Use Disorder MH Comorbidities (M±SD)	3.30±2.06		2.92±1.83		3.02±1.81
Trauma	117	67.2	544	67.3	156	68.1
Depression	125	71.8	581	71.9	163	71.2
Substance use disorder	143	82.2	620	76.7	174	76.0
Number of Non-HCV PH Comorbidities (M±SD)	2.73±2.64		2.74±2.32		2.75±2.50
Uncomplicated Hypertension	93	53.4	521	64.5	129	56.3
Diabetes	19	10.9	142	17.6	28	12.2
Liver disease	50	28.7	271	33.5	90	39.3

^aDAA=Direct-Acting Antiviral; PTSD=posttraumatic stress disorder, FY=Fiscal Year; VA=Department of Veterans Affairs; HCV=Hepatitis C Virus; EBA=Evidence-Based Antidepressant; EBP=Evidence-Based Psychotherapy PCL=PTSD Checklist; MH=Mental Health; PH=Physical Health; SL=suppressed low: cell is suppressed due to less than 3 patients; SH=suppressed high: cell is suppressed due to all but fewer than 3 patients

Table 3.

Post-Weighting Balance of Confounders ^a

Drug combination	GLE/PIB (N=174)	LDV/SOF (N=808)	SOF/VEL (N=229)
Advanced Fibrosis (>3.25) b	15.4%	22.5%	23.1%
Thousands of Days since Alcohol Use Disorder Diagnosis (M±SD)	3751±2124	3714±2050	3789±2204
Days from DAA Start to Earliest Baseline AUDIT-C (M±SD) b	121±97	131±97	119±92
Days from Baseline AUDIT-C to Follow-Up AUDIT-C (M±SD)	318±138	320±122	312±132
Sessions of EBP for PTSD (M±SD) b	0.4±2	0.6±3	0.4±2
Weeks of EBA for PTSD (M±SD) b	5±9	6±12	6±11
Any FDA Approved Medications for AUD b	19.9%	17.3%	22.2%
Sexual Trauma while in Military	17.2%	16.7%	16.6%
Number of ED Visits for Psychiatric Indication (M±SD)	2±3	2±3	2±3
Days of Acute Inpatient Mental Health (M±SD)	47±100	52±102	54±99
Substance use disorder b	81.9%	77.6%	75.2%
Diabetes b	14.6%	16.3%	11.4%

^aHCV=Hepatitis C Virus; PH=Physical health

^bVariable was not balanced (SMD > 0.10) after weighting and was retained as a covariate in outcomes models.

Consistent with our patient selection strategy, crude and adjusted AUDIT-C scores were approximately 10 at baseline (Table 4), reflecting heavy alcohol consumption. Crude frequency of clinically meaningful improvement ranged from 44.3% in the GLE/PIB group to 59.9% in the LDV/SOF group, although adjustment for differences between groups substantially lowered these proportions to 30.5% and 45.5%, respectively. This reflected a mean adjusted change in AUDIT-C scores ranging from −4.24 (SD = 6.90) in the GLE/PIB group to −5.43 (SD = 10.16) in the LDV/SOF group. In direct comparisons (Table 5), GLE/PIB performed worse than LDV/SOF, including both lower odds of clinically meaningful improvement (OR = 0.59; 95% CI [0.40, 0.87]) and a smaller mean change in AUDIT-C score (MD = −1.20; 95% CI [−2.03, −0.36]). Differences in categorical and continuous outcomes in other bivariate comparisons were smaller in magnitude.

Table 4.

Distribution of Baseline AUDIT-C and Outcomes Before and After Weighting ^a

	GLE/PIB (N=174)		LDV/SOF (N=808)		SOF/VEL (N=229)
	M±SD		M±SD		M±SD
Baseline AUDIT-C
Unweighted	10.42±1.44		10.25±1.50		10.21+1.48
IPTW	10.11±2.46		10.09±7.05		9.98±3.92
Change in AUDIT-C
Unweighted	−5.28±4.68		−6.66±4.26		−5.59±4.75
IPTW	−4.24±6.90		−5.43±19.16		−4.66±11.01
Clinically Meaningful Improvement (decrease to 3 points or less on AUDIT-C for men; decrease to 2 points or less on AUDIT-C for women)
	n	%	n	%	n	%
Unweighted	77	44.3	484	59.9	118	51.5
IPTW	77	30.5	484	45.5	118	40.5

^aAUDIT-C=Alcohol Use Disorders Identification Test, M=Mean, SD=Standard Deviation, IPTW=Inverse Probability of Treatment Weighting

Table 5.

Comparison Baseline AUDIT-C and Outcomes Across DAA Treatments ^a

	GLE/PIB versus LDV/SOF		GLE/PIB versus SOF/VEL		SOF/VEL versus LDV/SOF
	MD	95% CI	MD	95% CI	MD	95% CI
Baseline AUDIT-C
Unweighted	−0.17	−0.41, 0.07	−0.21	−0.50, 0.08	0.04	−0.18, 0.26
IPTW	−0.02	−0.34, 0.30	−0.13	−0.48, 0.22	0.11	−0.12, 0.33
Change in AUDIT-C
Unweighted	−1.38	−2.14, −0.62	−0.31	−1.24, 0.62	−1.07	−1.76, −0.39
IPTW	−1.20	−2.03, −0.36	−0.42	−1.41, 0.57	−0.66	−1.37, 0.05
Clinically Meaningful Improvement (decrease to 3 points or less on AUDIT-C for men; decrease to 2 points or less on AUDIT-C for women)
	OR	95% CI	OR	95% CI	OR	95% CI
Unweighted	0.74	0.62, 0.88	0.86	0.70, 1.06	0.86	0.75, 0.99
IPTW	0.59	0.40, 0.87	0.66	0.42, 1.04	0.89	0.65, 1.21

^aDAA=Direct Acting Antiviral, MD=Mean Difference, OR=Odds Ratio, CI=Confidence Interval, AUDIT-C=Alcohol Use Disorders Identification Test, M=Mean, SD=Standard Deviation, IPTW=Inverse Probability of Treatment Weighting

Discussion

In a cohort of VA patients with a diagnosis of HCV, AUD, and co-occurring PTSD, we observed a substantial reduction of alcohol use in all three DAA treatment groups, with mean improvements in AUDIT-C scores between 4.2 and 5.4 points. Similarly, in considering alcohol use as a categorical outcome, each DAA resulted in large proportions of patients with final AUDIT-C scores below the threshold for misuse, again ranging from 30% to 45% depending on the specific medication. After adjustment for differences between treatment groups, LDV/SOF appears to have the strongest association with clinically meaningful improvement in alcohol consumption compared to GLE/PIB and SOF/VEL. This result may be explained by LDV/SOF being the only medication in our analysis that was available prior to the 2016 HCV treatment directives around alcohol abstinence and liver health status. Furthermore, the smaller decrease in alcohol consumption observed in the GLE/PIB and SOF/VEL groups may be explained by the timing of their release falling after the directive. Notably, baseline AUDIT-C scores were obtained up to 365 days prior to DAA initiation, and LDV/SOF had the most days (139.29 ± 97.37 days) from baseline AUDIT-C to DAA start. This may have allowed a longer period to achieve abstinence prior to HCV treatment.

In contrast to findings in previous work demonstrating greater PTSD improvement on GLE/PIB compared with the other DAAs (Shiner, Huybrechts, et al., 2022), GLE/PIB was associated with the smallest improvement in drinking outcomes in the current analysis. This is notable, as an observation of greater decreases in alcohol consumption for patients prescribed GLE/PIB compared to the other agents may have indicated AUD as a potential mediator of the GLE/PIB and PTSD association. While sample size issues preclude us from conducting formal mediation tests of these associations, this finding provides some evidence that changes in alcohol consumption are unlikely to be a mechanism of the previously documented association between GLE/PIB and improvements in PTSD symptoms (Shiner, Huybrechts, et al., 2022).

There were several limitations to this analysis. First, the sample was limited to VA users with HCV, PTSD, and AUD, thus results may not be generalizable to other populations. Second, it can be challenging to draw conclusions about drinking outcomes in a population of patients who are being treated for a liver infection and are thus being counseled against alcohol use in accordance with practice guidelines (IDSA, 2021). Similarly, as HCV infection was successfully treated by the DAA in most patients, it is possible that the reduced drinking was due to the clinical improvement resulting from HCV cure rather than a direct effect of DAA. Third, it is possible that there is unmeasured confounding. This dataset lacks information about whether abstinence was a condition of DAA treatment for certain patients, thus we also cannot dismiss the possibility that some patients may have underreported alcohol use in order to obtain HCV treatment. Additionally, while advanced liver fibrosis was accounted for as a categorical variable, scores beyond the cut-point of 3.25 have meaningful differences, such as decompensated cirrhosis, which would preclude treatment with GLE/PIB (Lamb, 2017). Moreover, current prescribing guidelines as outlined by the American Association for the Study of Liver Diseases and Infectious Diseases Society of America recommend GLE/PIB or SOF/VEL for patients without cirrhosis or with compensated cirrhosis in a simplified treatment algorithm, whereas LDV/SOF appears to be reserved for prescribing by specialists in cases where patients are not eligible for simplified treatment (IDSA, 2021). Patients with such conditions may be more likely to abstain from alcohol.

The evidence regarding effectiveness for AUD treatment in HCV populations have not previously been impressive. For example, a study comparing provider delivered Screening, Brief Intervention and Referral to Treatment (SBIRT) versus intensive SBIRT and integrated alcohol counseling, in a liver clinic treating HCV, found no enhancement over standard of care, and only 19–21% of participants achieved abstinence from alcohol (Proeschold-Bell et al., 2020). In a systematic review of psychosocial interventions for alcohol abstinence in patients with chronic liver disease, no psychosocial intervention demonstrated success (Khan et al., 2016). Thus, available evidence around expected alcohol consumption reduction in liver clinic treatment settings would suggest that our observed improvement both overall and within DAA treatment group is compelling.

Furthermore, the observed drinking reductions in our cohort may be higher than that observed with standard psychopharmacology for AUD (Kranzler & Soyka, 2018). Currently, there are three FDA-approved agents for the treatment of AUD including disulfiram and the two frontline treatments, acamprosate and oral or long-acting injectable formulations of naltrexone (Kranzler & Soyka, 2018). Additionally, several off-label medications including gabapentin and topiramate are currently utilized as second-line AUD treatments (Swift & Aston, 2015). These disparate medications, unified by their indication rather than mechanism of action, have fairly limited efficacy with effect sizes in the small to medium range compared to placebo and the literature calls for investigation into more efficacious medications (Kranzler & Soyka, 2018).

Mechanistically, it is unclear why GLE/PIB would be associated with the greatest PTSD improvement and the smallest drinking reduction. Notably, GLE/PIB is the only agent in this analysis to contain a serine protease inhibitor of HCV non-structural viral protein NS3/4A (Lamb, 2017). Serine protease inhibitors are known to act in the interferon signaling pathway which plays a critical role in the innate immune response (Casella et al., 2020). Overexpression of interferon signaling in the innate immune response has been implicated in PTSD pathophysiology (Breen et al., 2015). If GLE/PIB were to inhibit the interferon signaling pathway it is plausible that PTSD symptoms could improve.

All three DAA agents in our analysis contain NS5A inhibitors. NS5A is an HCV non-structural viral protein that is known to exploit protein kinase R (PKR) activity in immune cells during the viral replication process (He et al., 2001; Suzuki et al., 2019). It has been demonstrated that PKR regulates NLRP3 activation (Lan et al., 2020; Lu et al., 2012). It is plausible that NS5A inhibitors may also inhibit their cellular counterpart, PKR, whose signaling in immune cells has a demonstrated role in neuroinflammation (Gal-Ben-Ari et al., 2019; Kapil et al., 2014), possibly due to monocyte trafficking to the brain (Wohleb et al., 2014). More specifically, PKR inhibitors have been identified as potential inhibitors of nuclear factor kappa B (NF-kB) protein complex (Gupta et al., 2010; Zhang et al., 2013), a transcription factor that has demonstrated the ability to induce genetic expression involved in addiction pathways including opioid receptors (Nennig & Schank, 2017). With the identification of NLRP3 as a common component in HCV, PTSD, and AUD, the potential for a DAA acting in this pathway with off-target addiction effects emerges, which may have been masked by expected alcohol consumption reduction in the context of HCV treatment. Replication of results in a larger, less restrictive sample of patients receiving the most recent DAAs for HCV would be beneficial to future research in this area. Furthermore, prospectively testing DAAs in a non-HCV sample of patients with AUD would be an important next step in identifying a potential AUD medication with an alternative mechanism of action.

Conclusion

Patients with co-occurring AUD and PTSD who are receiving DAA treatment for HCV experience a substantial reduction in alcohol use. Further exploration of the role of DAAs in AUD treatment is warranted.

Appendix Figure 1.

Diagram of the timing of Alcohol Use Disorders Identification Test measurements in relation to exposure to direct acting antiviral medications and HCV cure.

Appendix Table 1.

Diagram of the Timing of Alcohol Use Disorders Identification Test Measurements in Relation to Exposure to Direct Acting Antiviral Medications and HCV Cure

Question		Answer
A	How many Department of Veterans Affairs (VA) patients had any diagnosis of PTSD between FY2000 and FY2019?	2,098,389
B	How many patients from (A) also had a diagnosis of Hepatitis C Virus (HCV)?	147,174
C	How many patients from (B) also had a diagnosis of AUD?	103,588
D	How many patients from (C) received at least one prescription of a DAA for HCV?	43,070
E	How many patients from (D) received at least one prescription for GLE/PIB, LDV/SOF, or SOF/VEL?	27,606
F	How many prescriptions from (E) followed a 6-month washout of that medication?	28,556
G	How many prescriptions from (F) were continued for at least 56 days?	21,870
H	How many starts from (G) had an Alcohol Use Disorder Identification Test (AUDIT-C) within 365 days prior to 7 days after DAA start.	19,982
I	How many starts from (H) had an AUDIT-C within 365 days after the 35-day point?	16,121
J	How many person trials from (I) had their first AUD diagnosis prior to trial start?	14,619
K	How many persons had a baseline AUDIT-C score of 8 or greater?	1,211
Detailed Information from (J): Person Trials from All Potential Patients
GLE/PIB	LDV/SOF	SOF/VEL
174	808	229

^aGLE/PIB = Glecaprevir/Pibrentasvir, LDV/SOF = Ledipasvir/Sofosbuvir, SOF/VEL = Sofosbuvir/Velpatasvir

Appendix Table 2.

Assessment of Potential Confounding, as Measured by Absolute Standardized Mean Difference ^{a b}

Association	Outcome	Exposure
Variable	No misuse c	GLE/PIB versus LDV/SOF	GLE/PIB versus SOF/VEL	SOF/VEL versus LDV/SOF
Hepatitis Disease Status
Thousands of Days Since HCV Diagnosis	0.05	0.13	0.08	0.05
Fibrosis (FIB-4 Score)
Mild (<1.45)	0.05	0.25	0.20	0.05
Moderate (≥1.45 and ≤3.25)	0.06	0.05	0.03	0.03
Advanced (>3.25)	0.11	0.19	0.23	0.04
Alcohol Use Disorder Status
Thousands of Days since Alcohol Use Disorder Diagnosis	0.10	0.19	0.04	0.14
Prior Evidence-Based PTSD Treatment
Number of Prior EBA Trials	0.02	0.22	0.08	0.14
Number of Prior EBP Trials	0.06	0.11	0.06	0.05
AUDIT-C Timing, Version, and Severity
Days from DAA Start to Earliest Baseline AUDIT-C	0.26	0.50	0.22	0.29
Days from Baseline AUDIT-C to Follow-Up AUDIT-C	0.24	0.41	0.14	0.25
Days from DAA End to Follow-Up AUDIT-C	0.02	0.06	0.12	0.06
Baseline AUDIT-C	0.05	0.12	0.14	0.03
Concurrent Treatment
Sessions of EBP for PTSD	0.14	0.04	0.15	0.10
Weeks of EBA for PTSD	0.22	0.18	0.07	0.11
Any Non-EBA Antidepressant	0.06	0.08	0.15	0.07
Any Anticonvulsant	0.05	0.13	0.21	0.08
Any Sedative/Hypnotics	0.04	0.23	0.12	0.11
Any Opioid	0.08	0.13	0.14	0.27
Any Atypical Antipsychotic	0.06	0.01	0.08	0.06
Any Prazosin	0.08	0.03	0.08	0.05
Any FDA Approved Medications for AUD	0.10	0.27	0.11	0.15
Any Opioid Agonist Therapy	0.06	0.19	0.15	0.04
Patient Characteristics at Baseline
Age	0.08	0.27	0.14	0.12
Women	0.02	0.07	0.08	0.01
Married	0.08	0.06	0.06	0
Rural	0.06	0.04	0.10	0.15
White Non-Hispanic	0.08	0.28	0.13	0.41
Black Non-Hispanic	0.07	0.30	0.19	0.49
Hispanic	0.03	0.03	0.13	0.10
Other Racial or Ethnic Group	0.01	0.12	0.08	0.04
Combat Exposure	0.07	0.09	0.03	0.12
Sexual Trauma while in Military	0.12	0.11	0.07	0.04
VA Disability Level 70% or Greater	0	0	0.01	0
Service Use Characteristics in the 1 Year Preceding Baseline
Number of PTSD Outpatient Clinical Team Visits	0.02	0.10	0.05	0.05
Number of Outpatient Mental Health Visits	0.02	0.08	0.20	0.11
Number of Outpatient Substance Abuse Visits	0.04	0.20	0.03	0.22
Number of Outpatient Primary Care Visits	0.08	0.11	0.04	0.16
Number of Outpatient Specialty Medical Visits	0.02	0.17	0.16	0.01
Number of ED Visits for Psychiatric Indication	0.27	0.18	0.06	0.12
Days of Acute Inpatient Mental Health	0.24	0.10	< 0.001	0.11
Days of Residential PTSD Treatment	0.03	0.03	0.01	0.01
Days Residential Substance Abuse Treatment	0.19	0.05	0.05	0
Comorbidities in the 2 Years Preceding Baseline
Number of Non-Alcohol Use Disorder MH Comorbidities (M±SD)	0.01	0.19	0.15	0.05
Trauma	0.04	0	0.02	0.02
Depression	0.03	0	0.02	0.02
Substance use disorder	0.13	0.14	0.15	0.02
Number of Non-HCV PH Comorbidities (M±SD)	0.05	0.01	0.01	0
Uncomplicated Hypertension	0.02	0.23	0.06	0.17
Diabetes	0.15	0.19	0.04	0.15
Liver disease	0.04	0.10	0.22	0.12

^aDAA=Direct-Acting Antiviral; PTSD=posttraumatic stress disorder, FY=Fiscal Year; VA=Department of Veterans Affairs; HCV=Hepatitis C Virus; EBA=Evidence-Based Antidepressant; EBP=Evidence-Based Psychotherapy; PHQ-9=Patient Health Questionnaire-9; MH=Mental Health; PH=Physical Health

^bCovariates associated with both the outcome and exposure (SMD ≥ 0.10) are indicated by the gray rows.

^CDecrease to 3 points or less on AUDIT-C for men; decrease to 2 points or less on AUDIT-C for women.