Abstract
Background and Aims
Alcohol use disorder (AUD) pharmacotherapy reduces hepatic decompensation and mortality in patients with alcohol-associated liver disease (ALD). We aimed to characterize the use of AUD pharmacotherapy in a large national cohort.
Methods
Adults with AUD were identified in PharMetrics Plus for Academics—a nationally representative claims database of commercially insured Americans. We examined receipt of AUD therapy and associated treatment patterns.
Results
A total of 28,625 patients with AUD were identified: 1640 (5.7%) had ALD, among whom 439 (1.5%) had acute alcohol-associated hepatitis (AAH). Pharmacotherapy was prescribed in 14.5% of patients without ALD, 2.3% with ALD cirrhosis, and 9.8% with AAH. The most prescribed medication was gabapentin (9.4%). Among those receiving pharmacotherapy, one-time prescriptions were observed in 28.4% of patients without ALD, 10.7% with ALD cirrhosis, and 18.6% with AAH. The median time from diagnosis to receipt of pharmacotherapy was 10 months (interquartile range [IQR] 1.5-29.2). On regression analysis, nonsubstance use psychiatric diagnosis (adjusted odds ratio [aOR]: 2.71; 95% confidence interval [CI] [2.51, 2.93]) and female sex (aOR: 1.32; 95% CI [1.22, 1.42]) were associated with receipt of pharmacotherapy, while ALD cirrhosis (aOR: 0.22; 95% CI [0.14, 0.32]) and hepatic decompensation (aOR: 0.07; 95% CI [0.02, 0.16]) were associated with not receiving pharmacotherapy.
Conclusion
AUD pharmacotherapy appeared underutilized in a large cohort of commercially insured U.S. adults, particularly among patients with ALD and hepatic decompensation. Providers caring for patients with AUD across the spectrum of liver disease should be aware of potential underutilization of these medications, especially among males and patients with cirrhosis.
Keywords: Addiction, Alcohol-Associated Hepatitis, Alcohol-Associated Liver Disease, Alcohol Use Disorder, Liver Transplantation
Graphical abstract
Introduction
Alcohol use disorder (AUD) is estimated to affect 10.6% of the U.S. population1 and 8.6% of the global population.2 Alcohol consumption is implicated in nearly half of deaths from chronic liver disease and alcohol-associated liver disease (ALD) resulted in 21.5 million disability-adjusted life years lost globally in 2016.3 ALD is the leading cause of cirrhosis worldwide4 and has become a leading indication for liver transplantation in the United States.5 Effective treatment of AUD, including behavioral and pharmacologic therapies, is therefore essential to preventing significant morbidity and mortality across the spectrum of liver disease.
There are several medications with U.S. Food and Drug Administration (FDA) approval for treatment of AUD (eg, naltrexone, acamprosate, and disulfiram) and several others for which off-label use has been shown to be efficacious (eg, baclofen, gabapentin, and topiramate).6 Indeed, receipt of pharmacotherapy is a cost-effective7 method to decrease the progression from AUD to ALD,8 as well as reduce mortality9 and hepatic decompensation among those with ALD.8,10 Consideration of pharmacotherapy for patients with ALD is therefore recommended by the American Association for the Study of Liver Diseases11 and the American Psychiatric Association,12 although there are limited data in populations with advanced liver disease.11 Despite this evidence, pharmacologic treatment of AUD is suspected to be underutilized in the general population13 and particularly among patients with ALD10,13,14 as a result of patient, provider, and systems barriers.15
Existing studies examining pharmacotherapy utilization in ALD have been limited by lack of generalizability to the U.S. population due to either single-health system design,8,14 or evaluation in a Veterans Affairs (VA)–based population.9,10 Moreover, prior studies focus on the use of FDA-approved medications alone and may therefore underestimate rates of AUD pharmacotherapy by excluding off-label prescription practices.9,10,13,14 Herein, we provide a comprehensive investigation of AUD treatment using a large pharmacoepidemiology database that is nationally representative of commercially insured adults. We included FDA-approved and off-label medications among patients with and without liver disease to identify factors associated with AUD pharmacotherapy and characterize pharmacologic treatment patterns. These data can identify AUD pharmacotherapy treatment patterns and identify vulnerable populations that may benefit from targeted efforts to increase utilization.
Methods
Database
The IQVIA PharMetrics Plus for Academics database includes fully adjudicated pharmacy and medical claims from primarily commercial health plans. The database includes >100 million unique enrollees since 2006, and previous studies suggest it is nationally representative of the commercially insured U.S. population.16 This retrospective analysis of deidentified claims data was deemed exempt from full review by the University of Texas Southwestern Institutional Review Board.
Study Populations
A random sample of the adult (ages 18–64 years) population was used to identify populations of interest using International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10) codes (Figure 1; Table A1). Inclusion criteria included continuous medical and pharmacy claims data for at least 6 months pre- and 12 months post-AUD diagnosis. The index date corresponded to the first code for AUD, ALD cirrhosis, or acute alcohol-associated hepatitis (AAH) as defined below. A validated definition of AUD using at least 2 outpatient and/or at least 1 inpatient ICD-9/10 code was used to identify patients with AUD.10,17 The population was further divided into three cohorts: AUD without liver disease, ALD cirrhosis,18 and AAH19 (Figure 1; Table A1).
Figure 1.
Identification of patients with alcohol use disorder from a random sample of adult patients within PharMetrics Plus for Academics. Patients were stratified based on the presence or absence of liver disease including those with and without acute alcohol-associated hepatitis. Receipt of pharmacotherapy was assessed any time after diagnosis for as long as the individual remained in the database.
Pharmacologic Treatment of AUD
Our primary outcome was receipt of AUD pharmacotherapy, including medications with FDA indications for AUD (naltrexone, disulfiram, and acamprosate) and those for which off-label use has been described (baclofen, topiramate, and gabapentin). Patients were deemed to have initiated AUD treatment if they had at least one outpatient pharmacy claim and prescription fill for a medication used to treat AUD at any point after the index diagnosis date.
We further characterized AUD medication use by examining the frequency of one-time prescriptions and time from index diagnosis to receipt of pharmacotherapy. Prescriber subspecialty was considered on a claims level for each cohort. Utilization of nonpharmacologic psychiatric care (ie, psychotherapy and outpatient psychiatric evaluation) was evaluated using Current Procedural Terminology codes as previously described.20 We also evaluated rates of hospitalizations stratified by receipt of pharmacotherapy in the 12-month period following index diagnosis.
Statistical Analyses
Descriptive statistics in each population were stratified by type of liver disease. Characteristics were reported as frequency (%) for categorical variables and median (Q1 – Q3) for continuous variables. Groups were compared using chi-square tests for categorical variables and either Fisher’s exact or Kruskal-Wallis tests for continuous variables. A multivariable regression analysis was used to identify factors associated with receipt of AUD pharmacotherapy among all patients with AUD while adjusting for the following covariates: age, sex, U.S. region, modified Charlson Comorbidity Index (CCI)21 excluding liver disease, non-substance use disorder (SUD) psychiatric diagnosis, non-AUD SUD, chronic kidney disease, diabetes, neuropathy, hepatocellular carcinoma, hepatic decompensation (ascites, hepatic encephalopathy [HE], and/or variceal hemorrhage), and presence of liver disease (none, ALD cirrhosis, and AAH). For our analysis of hospitalizations, we used a Cox proportional hazards model in which receipt of AUD pharmacotherapy was treated as a time-varying covariate, adjusting for age, sex, non-liver CCI, presence of diabetes, and presence of hepatic decompensation (6 months preindex through 6 months postindex) to identify factors associated with hospitalization. A sensitivity analysis was conducted investigating receipt of pharmacotherapy for only FDA-approved medications to address the possibility that off-label medications were prescribed for indications other than AUD. All statistical analyses were conducted using R version 4.2.2, with P < .05 considered statistically significant.
Results
Patient Characteristics
Of 28,625 patients with AUD, 26,985 (94.3%) had no known liver disease, 1201 (4.2%) had ALD cirrhosis, and 439 (1.5%) had AAH (Figure 1). Overall, 64.5% of the total population was male with other patient characteristics described in Table 1. Patients without liver disease were significantly younger (median 41.0 years interquartile range [IQR] [27.0–52.0]) than those with ALD cirrhosis (53.0 years IQR [48.0–55.0]) or AAH (49.0 years IQR [48.0–59.0]; P < .001). A total of 13,128 (45.9%) patients had a non-SUD psychiatric diagnosis, and 10,227 (35.7%) had SUD other than AUD. Among patients with ALD (including AAH), 699 (42.6%) had decompensated liver disease, with 550 (33.5%) having ascites, 236 (14.4%) hepatic encephalopathy, and 167 (10.2%) variceal hemorrhage; 49 (3.0%) had hepatocellular carcinoma (Table 1). Two-thirds (68.4%) of patients with decompensated cirrhosis had one type of decompensation, with one-third having two or more decompensations.
Table 1.
Demographics and Clinical Characteristics of Patients With AUD in Groups With No ALD (N = 26,985), ALD Cirrhosis (N = 1201), and AAH (N = 439)
| Variable | AUD population |
||||
|---|---|---|---|---|---|
| Total (N = 28,625) | No ALD (N = 26,985) | ALD cirrhosis (N = 1201) | AAH (N = 439) | P value | |
| Age at diagnosis, y (median/IQR) | 42.0 (28.0–52.0) | 41.0 (27.0–52.0) | 53.0 (48.0–59.0) | 49.0 (40.5–55.0) | <.001 |
| Male (%) | 18,472 (64.5%) | 17,366 (64.3%) | 831 (69.1%) | 275 (62.6%) | <.001 |
| U.S. region | |||||
| Northeast | 7009 (24.5%) | 6699 (24.8%) | 229 (19.1%) | 81 (18.5%) | <.001 |
| Midwest | 8934 (31.2%) | 8496 (31.5%) | 307 (25.5%) | 131 (29.8%) | |
| South | 5370 (18.8%) | 5002 (18.5%) | 270 (22.5%) | 98 (22.3%) | |
| West | 7312 (25.5%) | 6788 (25.2%) | 395 (32.9%) | 129 (29.4%) | |
| Year of AUD diagnosis | |||||
| 2006–2009 | 9129 (31.9%) | 8655 (32.1%) | 347 (28.9%) | 127 (28.9%) | <.001 |
| 2010–2013 | 8784 (30.7%) | 8288 (30.7%) | 382 (31.8%) | 114 (26.0%) | |
| 2014–2017 | 7901 (27.6%) | 7372 (27.3%) | 391 (32.6%) | 138 (31.4%) | |
| 2018–2021 | 2811 (9.8%) | 2670 (9.9%) | 81 (6.7%) | 60 (13.7%) | |
| Primary insurance type | |||||
| HMO | 8902 (31.1%) | 8194 (30.4%) | 563 (46.9%) | 145 (33.0%) | <.001 |
| PPO | 15,547 (54.3%) | 14,774 (54.7%) | 527 (43.9%) | 246 (56.0%) | |
| Othera | 4176 (14.6%) | 4017 (14.9%) | 111 (9.2%) | 48 (11.0%) | |
| Primary enrollment type | |||||
| Commercial | 22,434 (78.4%) | 21,389 (79.3%) | 716 (59.6%) | 329 (74.9%) | <.001 |
| Supplemental Medicareb | 661 (2.3%) | 569 (2.1%) | 80 (6.7%) | 12 (2.7%) | |
| Otherc | 5530 (19.3%) | 5027 (18.6%) | 405 (33.7%) | 98 (22.3%) | |
| Modified CCId | |||||
| 0–3 | 28,050 (98.0%) | 26,451 (98.0%) | 1168 (97.3%) | 431 (98.2%) | .40 |
| 4–6 | 491 (1.7%) | 455 (1.7%) | 29 (2.4%) | 7 (1.6%) | |
| ≥7 | 84 (0.3%) | 79 (0.3%) | 4 (0.3%) | 1 (0.2%) | |
| Diabetes | 2322 (8.1%) | 2083 (7.7%) | 190 (15.8%) | 49 (11.1%) | <.001 |
| Neuropathy | 819 (2.9%) | 721 (2.7%) | 74 (6.2%) | 24 (5.5%) | <.001 |
| CKD | 340 (1.2%) | 257 (1.0%) | 55 (4.6%) | 28 (6.4%) | <.001 |
| ESRD | 33 (0.1%) | 23 (0.1%) | 4 (0.3%) | 6 (1.4%) | <.001 |
| Ascites | 583 (2.0%) | 33 (0.1%) | 408 (34.0%) | 142 (32.3%) | <.001 |
| HE | 241 (0.8%) | 5 (0.0%) | 174 (14.5%) | 62 (14.1%) | <.001 |
| Variceal bleeding | 167 (0.6%) | 1 (0.0%) | 135 (11.2%) | 31 (7.1%) | <.001 |
| Hepatic decompensatione | 734 (2.6%) | 35 (0.1%) | 536 (44.6%) | 163 (37.1%) | <.001 |
| HCC | 51 (0.2%) | 3 (0.0%) | 45 (3.7%) | 3 (0.7%) | <.001 |
| Liver transplant | 9 (0.0%) | 0 (0%) | 8 (0.7%) | 1 (0.2%) | <.001 |
| Non-SUD psychiatric diagnosis | 13,128 (45.9%) | 12,615 (46.7%) | 323 (26.9%) | 190 (43.2%) | <.001 |
| Non-alcohol SUD diagnosis | 10,227 (35.7%) | 9784 (36.3%) | 304 (25.3%) | 139 (31.7%) | <.001 |
| Receipt of pharmacotherapy | 3995 (14.0%) | 3924 (14.5%) | 28 (2.3%) | 43 (9.8%) | <.001 |
| Follow-up after diagnosis, months (median/IQR) | 30.0 (18.0–52.0) | 31.0 (19.0–53.0) | 20.0 (12.0–37.0) | 24.0 (13.0–41.0) | <.001 |
| Hospitalizationsf | <.001 | ||||
| 0 | 20,924 (73.1%) | 20,288 (75.2%) | 534 (44.5%) | 102 (23.2%) | |
| ≥1 | 7701 (26.9%) | 6697 (24.8%) | 667 (55.5%) | 337 (76.8%) | |
| ≥3 | 1011 (3.5%) | 746 (2.8%) | 171 (14.2%) | 94 (21.4%) | |
Bolded values indicated statistical significance defined as P < .05.
CKD, chronic kidney disease; ESRD, end-stage renal disease; HMO, Health Maintenance Organization; HE, hepatic encephalopathy; HCC, hepatocellular carcinoma; IQR, interquartile range; PPO, Preferred Provider Organization.
“Other” includes indemnity/traditional, point of service, consumer directed health care, or health savings accounts (HSA).
Advantage or supplemental plan.
“Other” includes State Children’s Health Insurance Program (SCHIP), Medicaid, or self-insured.
Excluding liver disease or severe liver disease in CCI calculation.
Defined as either hepatic encephalopathy, ascites, or variceal hemorrhage.
In the 12-month period following index date of diagnosis in each cohort.
Treatment of AUD
Over a median follow-up of 30.0 months (IQR: 18.0–52.0) after diagnosis, 3995 patients (14.0%) with AUD received pharmacologic therapy. Pharmacologic therapy was prescribed in 14.5% of patients without ALD, 2.3% of those with ALD cirrhosis, and 9.8% with AAH (P < .001). Focusing on FDA-approved medications specifically, 1054 (4.2%) patients received at least one prescription, with higher proportions of prescribed FDA-approved therapies in the AUD population without liver disease (3.8%) compared to ALD cirrhosis (0.8%) and AAH (5.0%) (P < .001). Patients receiving any pharmacotherapy were more likely to be female (44.2% vs 34.1%; P < .001) and were older at the time of diagnosis (median age 45.0 IQR [33.0–53.0] vs 42.0 years IQR [27.0–52.0], P < .001) than those who did not receive pharmacotherapy. Patients receiving medications for AUD were more likely to have non-SUD psychiatric diagnosis (70.3% vs 41.9%; P < .001), SUD (50.3% vs 33.4%; P < .001), diabetes (12.3% vs 7.4%; P < .001), chronic kidney disease (2.1% vs 1.0%; P < .010), and neuropathy (8.3% vs 2.0%; P < .001) than patients who did not receive therapy (Table A2). Among patients with ALD cirrhosis, patients receiving pharmacotherapy had lower proportions of ascites (10.7% vs 34.5%; P = .01) and hepatic encephalopathy (0% vs 14.8%; P = .03) than patients not receiving AUD medications.
The most prescribed medication was gabapentin (9.4%) followed by oral naltrexone (2.6%) and topiramate (2.0%); 3.0% of patients received multiple medications (Table 2). One-time prescriptions (no subsequent fills after initial prescription) were observed in 28.4% of patients without ALD, 10.7% with ALD cirrhosis, and 18.6% with AAH. The median time from diagnosis of AUD to receipt of pharmacotherapy was 10 months (IQR: 1.5–29.2), with no significant difference across the three groups. For FDA-approved AUD medications, the median time to receipt of pharmacotherapy was 8.0 months with no significant difference between cohorts (P = .31). One-time prescriptions of FDA-approved therapies with no subsequent fills were observed in 34.9% of the total population.
Table 2.
Class of Medication Prescribed for Treatment of Alcohol Use Disorder Among Patients With No ALD, ALD Cirrhosis, and AAH
| Medication | Total N = 3995 |
AUD N = 3924 |
ALD cirrhosis N = 28 |
AAH N = 43 |
P valuea |
|---|---|---|---|---|---|
| Oral naltrexoneb | 756 (2.6%) | 732 (2.7%) | 9 (0.7%) | 15 (3.4%) | <.01 |
| Injectable naltrexoneb | 124 (0.4%) | 121 (0.4%) | 1 (0.1%) | 2 (0.5%) | .48 |
| Disulfiramb | 198 (0.7%) | 191 (0.7%) | 1 (0.1%) | 6 (1.4%) | .09 |
| Acamprosateb | 185 (0.6%) | 180 (0.7%) | 1 (0.1%) | 4 (0.9%) | .38 |
| Topiramate | 579 (2.0%) | 571 (2.1%) | 5 (0.4%) | 3 (0.7%) | .61 |
| Baclofen | 549 (1.9%) | 546 (2.0%) | 1 (0.1%) | 2 (0.5%) | .01 |
| Gabapentin | 2695 (9.4%) | 2646 (9.8%) | 20 (1.7%) | 29 (6.6%) | .90 |
| Multiplec | 866 (3.0%) | 843 (3.1%) | 8 (0.7%) | 15 (3.4%) | .04 |
Bolded values indicated statistical significance defined as P < .05.
Comparing no ALD to any patient with ALD (ALD cirrhosis and AAH groups combined).
Food and Drug Administration approval for treatment of alcohol use disorder.
Patients who received more than one of the above medications.
Among all patients with AUD, 9473 of 28,625 (33.1%) received nonpharmacologic psychiatric care (ie, outpatient visit or psychotherapy). A higher proportion of patients without liver disease received nonpharmacologic psychiatric therapy (34.6%) than those with ALD cirrhosis (3.7%) and AAH (18.9%) (P < .001) in the 12 months after diagnosis.
Factors Associated With Receipt of AUD Pharmacotherapy
Factors associated with receipt of any AUD pharmacotherapy included age ≥40 years (adjusted odds ratio [aOR]: 1.39; 95% confidence interval [CI] [1.29, 1.50]), female sex (aOR: 1.32; 95% CI [1.22, 1.42]), non-liver CCI between 1 and 6 (aOR: 1.32; 95% CI [1.21, 1.44] for CCI 1–3; aOR: 1.84; 95% CI [1.44, 2.35] for CCI 4–6), non-SUD psychiatric diagnosis (aOR: 2.71; 95% CI [2.51, 2.93]), and neuropathy (aOR: 3.16; 95% CI [2.67, 3.74]) (Table 3). Year of diagnosis, non-AUD SUD, non-commercial insurance enrollment, and location in the U.S. South and West regions (relative to the Northeast) were also significantly associated with receipt of AUD pharmacotherapy (Table 3). Hepatic decompensation in the 6 months before through 6 months after AUD diagnosis was associated with decreased odds of receiving AUD pharmacotherapy (aOR: 0.07; 95% CI [0.02, 0.16]).
Table 3.
Factors Associated With Receipt of AUD Pharmacotherapy Among Patients With AUD (N = 28,625) and ALD (N = 1640): A Logistic Regression Multivariable Analysis
| Variable | Adjusted odds ratio (95% CI) |
|
|---|---|---|
| All patients (N = 28,625) | Patients with ALDa (N = 1640) | |
| Ageb | ||
| <40 y | Reference | Reference |
| ≥40 y | 1.39 (1.29, 1.50) | 0.97 (0.51, 1.96) |
| Sex | ||
| Male | Reference | Reference |
| Female | 1.32 (1.22, 1.42) | 0.93 (0.53, 1.61) |
| U.S. region | ||
| Northeast | Reference | Reference |
| Midwest | 1.04 (0.94, 1.15) | 1.12 (0.50, 2.60) |
| South | 1.35 (1.21, 1.51) | 1.68 (0.75, 3.95) |
| West | 1.35 (1.22, 1.50) | 0.93 (0.40, 2.23) |
| Non-liver CCIb | ||
| 0 | Reference | Reference |
| 1–3 | 1.32 (1.21, 1.44) | 1.49 (0.79, 2.77) |
| 4–6 | 1.84 (1.44, 2.35) | 3.01 (0.56, 12.32) |
| ≥7 | 1.18 (0.67, 2.03) | - |
| Diabetes | ||
| No | Reference | Reference |
| Yes | 1.08 (0.94, 1.23) | 0.61 (0.23, 1.50) |
| Neuropathy | ||
| No | Reference | Reference |
| Yes | 3.16 (2.67, 3.74) | 2.48 (1.03, 5.53) |
| Non-SUD psychiatric disorder | ||
| No | Reference | Reference |
| Yes | 2.71 (2.51, 2.93) | 5.94 (3.35, 10.97) |
| Non-alcohol SUD diagnosis | ||
| No | Reference | Reference |
| Yes | 1.55 (1.44, 1.67) | 1.75 (1.01, 3.02) |
| Insurance type | ||
| HMO | Reference | Reference |
| PPO/Other | 1.02 (0.93, 1.11) | 1.03 (0.54, 2.02) |
| Enrollment type | ||
| Commercial | Reference | Reference |
| Noncommercial | 1.12 (1.01, 1.23) | 0.29 (0.13, 0.63) |
| Index y | ||
| 2006–2009 | Reference | Reference |
| 2010–2013 | 1.65 (1.49, 1.82) | 0.89 (0.40, 1.99) |
| 2014–2017 | 2.27 (2.06, 2.51) | 2.17 (1.05, 4.60) |
| 2018–2021 | 2.35 (2.07, 2.66) | 4.32 (1.90, 9.91) |
| CKD/ESRD | ||
| No | Reference | Reference |
| Yes | 1.33 (0.98, 1.78) | 1.44 (0.30, 4.85) |
| HCC | ||
| No | Reference | Reference |
| Yes | 0.70 (0.04, 3.60) | 0.87 (0.05, 4.71) |
| Hepatic decompensationc | ||
| No | Reference | Reference |
| Yes | 0.07 (0.02, 0.16) | 0.06 (0.02, 0.14) |
| Liver disease category | ||
| Non-ALD | Reference | - |
| ALD cirrhosis | 0.22 (0.14, 0.32) | - |
| AAH | 0.89 (0.62, 1.24) | - |
Bolded values denote statistically significant associations defined as P < .05.
CI, confidence interval; CKD, chronic kidney disease; ESRD, end-stage renal disease; HMO, Health Maintenance Organization; HCC, hepatocellular carcinoma; PPO, Preferred Provider Organization.
All patients with ALD combining both non-AAH ALD and AAH subgroups.
Excluding liver disease or severe liver disease in CCI calculation; per point increase in CCI.
Defined by the presence of either ascites, hepatic encephalopathy, or variceal hemorrhage.
Considering all patients with ALD (combining ALD cirrhosis and AAH subgroups; N = 1640), non-SUD psychiatric diagnosis (aOR: 5.94; 95% CI [3.35, 10.97]), and non-AUD SUD (aOR: 1.75; 95% CI [1.01, 3.02]) were associated with receipt of AUD pharmacotherapy. Non-commercial insurance enrollment (aOR: 0.29; 95% CI [0.13, 0.63]) and hepatic decompensation (aOR: 0.06; 95% CI [0.02, 0.14]) were associated with decreased odds of receiving AUD pharmacotherapy (Table 3).
Factors associated with receipt of FDA-approved AUD pharmacotherapy included non-SUD psychiatric diagnosis (aOR: 3.16; 95% CI [2.74, 3.66]), non-AUD SUD (aOR: 1.45; 95% CI [1.28, 1.65]), and AAH relative to those with no liver disease (aOR: 2.11; 95% CI [1.29, 3.27]) (Table A3). ALD cirrhosis (aOR: 0.53; 95% CI [0.25, 0.97]) and hepatic decompensation (aOR: 0.04; 95% CI [<0.01, 0.21]) were associated with decreased odds of receiving FDA-approved AUD pharmacotherapy.
Provider Prescription Patterns for AUD Pharmacotherapy
Among all AUD medication claims, a primary care provider prescribed AUD pharmacotherapy in 13.5% of patients without liver disease, 14.6% in ALD cirrhosis, and 11.5% among those with AAH (Figure 2). Psychiatrists provided 12.9% of all prescriptions among those without liver disease, 14.6% with ALD cirrhosis, and 11.3% in AAH. There were no prescriptions provided by gastroenterologists or hepatologists across all three cohorts. Hospital medicine providers prescribed a higher proportion of AUD pharmacotherapy for patients with AAH (21.0%) than those with ALD cirrhosis (6.9%) and no liver disease (7.2%). There was a statistically significant increase in the proportion of patients receiving AUD pharmacotherapy over time among patients without liver disease (8.7% from 2006 to 2009 vs 20.6% from 2018 to 2021; P < .001) and among those with ALD cirrhosis (0.8% from 2006 to 2009 vs 11.4% from 2018 to 2021; P < .001); no statistically significant difference in the proportion of patients receiving AUD pharmacotherapy was observed over time among those with AAH (Figure 3).
Figure 2.
Proportion of total AUD pharmacotherapy claims prescribed by providers in different subspecialties among those who received treatment (N = 3995). PCP, primary care provider.
Figure 3.
Proportion of patients receiving AUD pharmacotherapy over time among patients with no ALD, ALD cirrhosis, and AAH. ∗ Denotes statistically significant difference after multiplicity adjustments for multiple comparisons with P < .001 for the following comparisons: 2006–2009 vs 2010–2013, 2006–2009 vs 2014–2017, 2006–2009 vs 2018–2021, 2010–2013 vs 2014–2017, 2010–2013 vs 2018–2021. ∗∗ Denotes statistically significant difference after multiplicity adjustments for multiple comparisons with P < .001 for the following comparisons: 2006–2009 vs 2018–2021, 2010–2013 vs 2018–2021, 2014–2017 vs 2018–2021.
Healthcare Utilization Among Patients With AUD
In patients with AUD, 26.9% had at least one hospitalization during the 12-month period following index diagnosis (Table 1). Among those with ALD cirrhosis, 55.5% had at least one hospitalization with 9.2% having 3 or more hospitalizations 12 months postdiagnosis. Among patients with AAH, 76.7% had at least one hospitalization, and 21.4% had 3 or more hospitalizations. The top 10 diagnoses for hospitalizations are included in Table A4 with most diagnoses related to hepatic decompensations for the ALD cirrhosis group, and alcohol dependence or withdrawal more commonly seen among the AAH group.
In a Cox proportional hazards model where receipt of AUD medication was treated as a time-varying covariate, while adjusting for age, sex, non-liver CCI, history of diabetes, decompensation history, and ALD vs non-ALD status, receipt of AUD medication was associated with a higher hazard of hospitalization (adjusted hazard ratio: 1.48; 95% CI [1.35, 1.62]) (Table A5).
Discussion
In a large nationally representative cohort of commercially insured U.S. adults, including over 28,000 patients with AUD, pharmacotherapy for AUD was prescribed in just over 1 in 8 patients. Patients with ALD, particularly those with hepatic decompensation, were less likely to receive pharmacologic therapy than those without ALD. Indeed, 1 in 50 patients with ALD received AUD medications. Among those who received medications, over one quarter received a one-time prescription with no refill. Men were less likely to receive AUD pharmacotherapy, and the median time to initiation of pharmacotherapy was nearly 1 year after AUD diagnosis. These findings suggest potential underutilization of AUD pharmacotherapy and opportunities to target populations including males with advanced liver disease.
Consistent with our findings, Rittenberg and colleagues identified <10% of a large cohort of insured adults with AUD who received AUD pharmacotherapy with men being less likely to receive therapy.13 A retrospective analysis of a VA-based cohort found a slightly higher (9.7%), albeit low, proportion of patients with ALD cirrhosis receiving AUD pharmacotherapy with hepatic decompensation being a strong negative predictor of medication receipt.9 Another VA-based study identified very low rates of pharmacotherapy among patients with AUD and associated liver disease similar to the findings described herein.10
An important distinction between our study and prior reports in VA-based populations9,10 is our inclusion of off-label medications with evidence demonstrating safety and efficacy in ALD cirrhosis (eg, gabapentin and baclofen).11 In restricting our primary outcome to just medications with FDA-approved indications, we observed an even lower rate of AUD pharmacotherapy usage—<5% among patients with no liver disease, and <1% among those with ALD cirrhosis. This finding suggests that compared to VA-based cohorts, commercially insured Americans may be even more prone to underutilization of AUD pharmacotherapy, potentially because of heterogeneity of healthcare plans and variable access to mental health services. Indeed, a prior comparative analysis between VA and privately insured populations suggested higher quality measures for medication management of mental health disorders (including substance use) in the VA population.22 These findings may be further magnified in populations facing additional barriers to high-quality care including those without commercial insurance (eg, Medicaid, uninsured) and require further investigation. Among all medications for treatment of AUD, naltrexone and acamprosate have the most robust evidence supporting their use with limited evidence for baclofen and gabapentin.6 Our findings therefore also suggest low utilization of medications with the strongest evidence for efficacy, particularly among those with liver disease.
Nearly 30% of patients received just a one-time prescription for AUD medications. Consistent with our findings, a prior study revealed duration of therapy ≤3 months in nearly two-thirds of patients.9 Alarmingly, we also observed a latency period of nearly 1 year from diagnosis to initiation of pharmacotherapy in this cohort, despite frequent contact with the healthcare system representing opportunities to initiate pharmacotherapy: three-quarters of patients with ALD were hospitalized in the year after diagnosis, and over 20% had 3 or more hospitalizations. This delay in pharmacotherapy may result from provider discomfort in prescribing these medications or result from some individuals with an AUD diagnosis no longer being candidates for therapy (eg, have a history of AUD with enduring sobriety). Indeed, Louvet and colleagues reported alcohol relapse after hospitalization for AAH in just over 25% of patients at 1 year, suggesting that pharmacologic treatment may be of limited long-term benefit for many patients with AAH.23
Interestingly, receipt of pharmacotherapy for AUD was associated with subsequent hospitalization despite adjusting for liver disease severity, decompensation, and nonliver-related comorbidity. This is in contrast to the findings of a large Swedish cohort study wherein receipt of AUD pharmacotherapy was associated with a lower risk of hospitalization.24 This finding may reflect adverse effects of the medications themselves. Alternatively, these findings may be due to unmeasured confounding wherein those receiving pharmacotherapy were also more prone to hospitalization at baseline. In either case, it remains unclear whether AUD pharmacotherapy prescription caused hospitalization and future prospective studies should explore this relationship.
Overall, the lower rate of AUD pharmacotherapy prescriptions among patients with liver disease is likely related to a myriad of factors, including patient, provider, and systems-based factors.15 One likely contributor is provider concerns about the safety of these medications in severe liver disease. This safety concern may explain our finding that patients with decompensated liver disease were less likely to receive AUD pharmacotherapy. A bidirectional relationship may also exist wherein patients receiving pharmacotherapy are less likely to develop decompensated liver disease, which is supported by existing literature.8,10 Nevertheless, there is evidence to suggest the efficacy and safety of several medications (eg, baclofen, acamprosate) in patients with liver disease and consideration of these medications is recommended by societal guidelines.11 While our results suggest increasing utilization of AUD pharmacotherapy over time, targeted interventions to address barriers to receipt of pharmacotherapy may further enhance availability of these medications among patients with liver disease.
Encouragingly, over one-quarter of prescriptions among patients with AAH were provided by hospital medicine providers, suggesting that hospitalizations for AAH may provide pivotal moments to intervene and initiate AUD treatment. However, nearly one-fifth of patients in this population received only a one-time prescription, highlighting an opportunity to improve structured follow-up and ongoing pharmacotherapy after hospital discharge. Gastroenterologists and hepatologists may be uniquely positioned to prescribe these medications to patients with ALD, and prior studies indicate that prescription of AUD medications is common practice among hepatologists.25 However, gaps in knowledge related to AUD medications have been identified among these providers, highlighting an area for enhanced quality of care.25 In our cohort, we identified no prescriptions provided by gastroenterologists or hepatologists; however, we were limited by missing data in a large portion of medication claims as provider specialty is not required for prescription claims in this database.
Patients with a non-SUD psychiatric diagnosis were more likely to receive AUD pharmacotherapy in our cohort, which may be the result of access to existing psychiatric services for these patients. Psychiatric referral for patients with AUD, even in the absence of concomitant psychiatric diagnosis, may therefore improve uptake of pharmacologic therapy. Although we were unable to comprehensively characterize nonpharmacologic therapies for AUD, one-third of patients with AUD received outpatient psychiatric care (eg, clinic visits and psychotherapy) including just 1 in 25 with ALD cirrhosis. This finding is consistent with existing literature assessing the use of behavioral therapies for AUD in patients with liver disease and may be the result of access issues, physical debility, and neurocognitive changes in advanced liver disease limiting participation in this treatment modality.10
Our study has several limitations. The PharMetrics Plus for Academics database does not include several important clinical and demographic variables including race, ethnicity, mortality, and laboratory data including alcohol metabolites. Additionally, the use of claims data may result in misclassification bias even with the use of previously validated definitions. Several validated definitions for AAH, for example, include laboratory data,19 which was not available in our database. This may have resulted in misclassification, as the definition of AAH using diagnostic codes alone employed in our study has previously demonstrated a positive predictive value of 78.2% for AAH.19 Furthermore, this definition was not externally validated outside a VA-based cohort. Although we were able to categorize patients with ALD cirrhosis as compensated or decompensated using diagnostic codes, the absence of laboratory data limited our ability to further stratify our analysis by disease severity using clinical prediction scores (eg, Model for End-Stage Liver Disease). Importantly, our findings are specific to the commercially insured adult population and may not pertain to populations with other forms of insurance or without insurance, therefore limiting generalizability. Prescriber specialty is not required for pharmacy claims, and our conclusions related to these data are limited by missing data approaching 50% of all analyzed claims. Our study population was defined using retrospective analysis of claims data and may therefore include patients with a history of AUD for whom pharmacotherapy may no longer be indicated (eg, those with sustained remission). We are also unable to determine indications for prescribed medications. This is likely most relevant in the case of gabapentin, which was the most prescribed medication in our study and may be used for alternative conditions such as neuropathy. A sensitivity analysis restricting our outcome to just medications with FDA-approved indications for AUD revealed no association between receipt of pharmacotherapy and neuropathy, an association that was noted in our primary analysis including gabapentin. While this finding suggests the possibility of misclassification bias, it may also reflect clinician preference for pharmacologic AUD therapy to treat comorbid conditions such as neuropathy. The remaining factors associated with receipt of pharmacotherapy in this sensitivity analysis remained unchanged from our primary analysis, indicating that the impact of potential misclassification bias on our conclusions is likely minor. Although our study provides an in-depth characterization of the use of pharmacotherapy, we are unable to comment on subsequent clinical outcomes (eg, hepatic decompensation, mortality). Finally, the use of administrative data precludes consideration of important nuances in treatment decisions and shared decision-making including patient and provider factors. Despite these limitations, our findings support the existing literature by providing a contemporary analysis of AUD medication usage (including off-label usage) among a large nationally representative cohort of commercially insured adults and a detailed characterization of treatment patterns and trends for these patients over time. This study can be considered a “best case scenario” of AUD treatment given the population of well-insured individuals who should have access to appropriate treatment; if more socioeconomically disadvantaged populations were included, we may expect even lower rates of AUD treatment.
In conclusion, medications for treatment of AUD are likely underutilized among commercially insured U.S. adults, particularly among younger male patients with liver disease. Providers caring for patients with AUD across the spectrum of liver disease should be aware of the potential underutilization of these medications and consider early initiation or subspecialty referral when appropriate. Future studies are needed to assess how targeted interventions may address patient, provider, and system barriers and improve accessibility of these medications to populations in need.
Acknowledgments
The authors want to acknowledge IQVIA for the provision of the IQVIA PharMetrics Plus for Academics data and support working with the data.
Footnotes
Authors’ Contributions: Study concept and design: Sarah R. Lieber, Alex R. Jones, Mack C. Mitchell, Amit G. Singal, Lisa B. VanWagner. Data acquisition: Meena M. Tadros, Prajwal Gowda, Olgert Bardhi. Data and statistical analysis: Yue Jiang. Initial drafting of the manuscript: Sarah R. Lieber, Alex R. Jones. Critical revision of the manuscript for important intellectual content: All authors.
Sarah R. Lieber and Alex R. Jones are co-first authors. Amit G. Singal and Lisa B. VanWagner are co-senior authors.
Conflicts of Interest: These authors disclose the following: Lisa B. VanWagner receives research support from W.L. Gore & Associates and Madrigal Pharmaceuticals, advises Madrigal pharmaceuticals, consults for Gerson Lehrman Group, AlphaSights, and Noble Insights and provides medicolegal services outside the submitted work. Thomas G. Cotter consults for Madrigal Pharmaceuticals. The remaining authors disclose no conflicts.
Funding: Support for this research includes: 1. The National Center for Advancing Translational Sciences of the National Institutes of HealthUL1 TR003163 (PI Toto). 2. American College of Gastroenterology Junior Faculty Development Award (2022–2025) (PI: Lieber). TGC is supported by the American Association for the Study of Liver Diseases (AASLD) Clinical, Translational and Outcomes Research Award (CTORA) and National Institute on Alcohol Abuse and Alcoholism (NIAAA) K23AA031310 grant. The views expressed do not necessarily represent the views of the NIAAA.
Ethical Statement: The data employed in this retrospective analysis were deidentified and deemed exempt from full institutional review board review.
Data Transparency Statement: The original deidentified data supporting this research is the property of IQVIA. IQVIA has restrictions prohibiting the authors from making the data set publicly available. Interested researchers may contact IQVIA to apply to gain access to the study’s data.
Reporting Guidelines: STROBE.
Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.gastha.2025.100799.
Supplementary Data
References
- 1.SAMHSA, Center for Behavioral Health Statistics and Quality National survey on Drug use and health. 2021. https://www.samhsa.gov/data/sites/default/files/reports/rpt39441/NSDUHDetailedTabs2021/NSDUHDetailedTabs2021/NSDUHDetTabsSect5pe2021.htm#tab5.6b
- 2.Glantz M.D., Bharat C., Degenhardt L., et al. The epidemiology of alcohol use disorders cross-nationally: findings from the World Mental Health Surveys. Addict Behav. 2020;102 doi: 10.1016/j.addbeh.2019.106128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Rehm J., Shield K.D. Global burden of alcohol use disorders and alcohol liver disease. Biomedicines. 2019;7(4):99. doi: 10.3390/biomedicines7040099. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Moon A.M., Singal A.G., Tapper E.B. Contemporary epidemiology of chronic liver disease and cirrhosis. Clin Gastroenterol Hepatol Off Clin Pract J Am Gastroenterol Assoc. 2020;18(12):2650–2666. doi: 10.1016/j.cgh.2019.07.060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kwong A.J., Ebel N.H., Kim W.R., et al. OPTN/SRTR 2020 annual data report: liver. Am J Transpl. 2022;22:204–309. doi: 10.1111/ajt.16978. [DOI] [PubMed] [Google Scholar]
- 6.Jonas D.E., Amick H.R., Feltner C., et al. Pharmacotherapy for adults with alcohol use disorders in outpatient settings: a systematic review and meta-analysis. JAMA. 2014;311(18):1889. doi: 10.1001/jama.2014.3628. [DOI] [PubMed] [Google Scholar]
- 7.Avanceña A.L.V., Miller N., Uttal S.E., et al. Cost-effectiveness of alcohol use treatments in patients with alcohol-related cirrhosis. J Hepatol. 2021;74(6):1286–1294. doi: 10.1016/j.jhep.2020.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Vannier A.G.L., Shay J.E.S., Fomin V., et al. Incidence and progression of alcohol-associated liver disease after medical therapy for alcohol use disorder. JAMA Netw Open. 2022;5(5) doi: 10.1001/jamanetworkopen.2022.13014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Rabiee A., Mahmud N., Falker C., et al. Medications for alcohol use disorder improve survival in patients with hazardous drinking and alcohol-associated cirrhosis. Hepatol Commun. 2023;7(4) doi: 10.1097/HC9.0000000000000093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rogal S., Youk A., Zhang H., et al. Impact of alcohol use disorder treatment on clinical outcomes among patients with cirrhosis. Hepatology. 2020;71(6):2080–2092. doi: 10.1002/hep.31042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Crabb D.W., Im G.Y., Szabo G., et al. Diagnosis and treatment of alcohol-associated liver diseases: 2019 practice guidance from the American association for the study of liver diseases. Hepatology. 2020;71(1):306–333. doi: 10.1002/hep.30866. [DOI] [PubMed] [Google Scholar]
- 12.Reus V.I., Fochtmann L.J., Bukstein O., et al. The American psychiatric association practice guideline for the pharmacological treatment of patients with alcohol use disorder. Am J Psychiatry. 2018;175(1):86–90. doi: 10.1176/appi.ajp.2017.1750101. [DOI] [PubMed] [Google Scholar]
- 13.Rittenberg A., Hines A.L., Alvanzo A.A.H., et al. Correlates of alcohol use disorder pharmacotherapy receipt in medically insured patients. Drug Alcohol Depend. 2020;214 doi: 10.1016/j.drugalcdep.2020.108174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Winters A.C., May F.P., Wang Y., et al. Alcohol use disorder treatment and outcomes among hospitalized adults with alcoholic hepatitis. Drug Alcohol Depend Rep. 2021;1 doi: 10.1016/j.dadr.2021.100004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.DiMartini A.F., Leggio L., Singal A.K. Barriers to the management of alcohol use disorder and alcohol-associated liver disease: strategies to implement integrated care models. Lancet Gastroenterol Hepatol. 2022;7(2):186–195. doi: 10.1016/S2468-1253(21)00191-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lieber S.R., Jiang Y., Moon A., et al. Antiplatelet medications are associated with bleeding and decompensation events among patients with cirrhosis. J Clin Gastroenterol. 2022;56(7):627–634. doi: 10.1097/MCG.0000000000001558. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Li M., Peterson C., Xu L., et al. Medical costs of substance use disorders in the US employer-sponsored insurance population. JAMA Netw Open. 2023;6(1) doi: 10.1001/jamanetworkopen.2022.52378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kramer J.R., Davila J.A., Miller E.D., et al. The validity of viral hepatitis and chronic liver disease diagnoses in Veterans Affairs administrative databases. Aliment Pharmacol Ther. 2008;27(3):274–282. doi: 10.1111/j.1365-2036.2007.03572.x. [DOI] [PubMed] [Google Scholar]
- 19.Panchal S.A., Kaplan D.E., Goldberg D.S., et al. Algorithms to identify alcoholic hepatitis hospitalizations in patients with cirrhosis. Dig Dis Sci. 2022;67(9):4395–4402. doi: 10.1007/s10620-021-07321-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lieber S.R., Jones A.R., Jiang Y., et al. Psychiatric diagnoses are common after liver transplantation and are associated with increased health care utilization and patient financial burden. Liver Transpl. 2024;30(11):1145–1158. doi: 10.1097/LVT.0000000000000390. [DOI] [PubMed] [Google Scholar]
- 21.Deyo R. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613–619. doi: 10.1016/0895-4356(92)90133-8. [DOI] [PubMed] [Google Scholar]
- 22.Watkins K.E., Smith B., Akincigil A., et al. The quality of medication treatment for mental disorders in the department of Veterans Affairs and in private-sector plans. Psychiatr Serv Wash DC. 2016;67(4):391–396. doi: 10.1176/appi.ps.201400537. [DOI] [PubMed] [Google Scholar]
- 23.Louvet A., Labreuche J., Artru F., et al. Main drivers of outcome differ between short term and long term in severe alcoholic hepatitis: a prospective study. Hepatol Baltim Md. 2017;66(5):1464–1473. doi: 10.1002/hep.29240. [DOI] [PubMed] [Google Scholar]
- 24.Heikkinen M., Taipale H., Tanskanen A., et al. Real-world effectiveness of pharmacological treatments of alcohol use disorders in a Swedish nation-wide cohort of 125 556 patients. Addiction. 2021;116(8):1990–1998. doi: 10.1111/add.15384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Cotter T.G., Ayoub F., King A.C., et al. Practice habits, knowledge, and attitudes of hepatologists to alcohol use disorder medication: sobering gaps and opportunities. Transpl Direct. 2020;6(10) doi: 10.1097/TXD.0000000000001054. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.