INTRODUCTION

In the United States (U.S.), it is estimated that only 4% of people with alcohol use disorder (AUD) receive a Food and Drug Administration (FDA)–approved medication for treatment, such as disulfiram, naltrexone, or acamprosate.1 There are six additional medications available in the U.S., not FDA-approved for the treatment of AUD, for which there is clinical trial evidence suggestive of safety and efficacy when used for AUD treatment: baclofen, gabapentin, ondansetron, topiramate, varenicline, and zonisamide.2 Physicians can legally prescribe these medications for AUD or another clinical indication, based on their individual judgment of patients’ likelihood of benefit and risk, even without formal FDA approval, a practice known as off-label prescribing. However, little is known about the extent to which both on-and-off-label medications for AUD are used in outpatient care. Using nationally representative prescribing data from ambulatory visits in the U.S., we estimated the prevalence of and factors associated with on-and-off-label medication prescription for AUD.

METHODS

We used data from the 2014–2016 National Ambulatory Medical Care Survey (NAMCS), an annual nationally representative survey of office-based physician visits administered by the Centers for Disease Control and Prevention.3 We identified all visits made by patients with AUD, based on whether patients had an AUD-related diagnostic code and/or were categorized as having “alcohol misuse, abuse, or dependence.” Next, we determined whether these visits by patients with AUD were associated with an active prescription for oral naltrexone, acamprosate, and/or disulfiram, which were categorized as on-label treatment, or associated with baclofen, gabapentin, ondansetron, topiramate, varenicline, and/or zonisamide, which were categorized as off-label treatment. We did not consider nalmefene because the oral formulation, which is used for AUD in Europe, is not available in the U.S.

We used a multivariable-adjusted logistic regression analysis to examine associations between any AUD medication use with age, sex, race/ethnicity, and patient/prescriber characteristics (with p<0.05 from bivariate analyses). Analyses were conducted in Stata (v.15.1, StataCorp.) and accounted for the complex survey sampling design in NAMCS using svy commands. This study was deemed exempt from review by Yale School of Medicine’s Institutional Review Board because we used de-identified, publicly available data.

RESULTS

In 2014–2016, there were 687 outpatient visits, among adults aged ≥18 with a known diagnosis of AUD, nationally representative of 7,197,178 visits (Table 1). Of these, 15.4% (95% CI, 11.6–20.2%) were associated with an active prescription for any AUD medication, including 3.6% (1.8–7.2%) who received on-label treatment and 11.8% (8.7–15.9%) who received only off-label treatment.

Table 1 Characteristics of Adults (Weighted %) According to AUD Medications Received in Office-Based Outpatient Visits: 2014–2016 National Ambulatory Medical Care Survey
Full size table

Several factors were independently associated with greater likelihood of receiving any AUD medication, including current tobacco use (adjusted odds ratio = 2.32; 95% CI, 1.09–4.93), receiving ≥ 6 concomitant medications (2.99; 1.05–8.42), being prescribed antidepressants (2.92; 1.19–7.15), and being prescribed opioids (2.48; 1.70–8.54) (Table 2).

Table 2 Prevalence of AUD Medication Prescriptions (Weighted %) in Office-Based Outpatient Visits According to Subgroups of U.S. Adults: 2014–2016 National Ambulatory Medical Care Survey
Full size table

DISCUSSION

From 2014 to 2016, fewer than one in seven outpatient visits to physicians in the U.S. made by patients with AUD were associated with an active prescription for AUD treatments with clinical trial evidence suggestive of safety and efficacy, three-quarters of which were for off-label treatments that had not been formally approved by the FDA as safe and effective for use for AUD. While this evaluation indicates low rates of on-label treatment for AUD, similar to what has been observed among commercially insured and Veterans Health Administration patients,4, 5 it also suggests that most AUD treatment is off-label. It is possible that off-label medications are being prescribed to treat multiple indications and reduce unnecessary polypharmacy (e.g., gabapentin for AUD and anxiety). However, our findings raise concerns about the underutilization of FDA-approved medications in outpatient clinical practice.

Limitations include small number of eligible records; limited generalizability outside of outpatient physician visits; lack of information on medication dosing (e.g., strengths, duration, and route of administration); inability to determine whether on-and-off-label medications may have been prescribed for reasons other than AUD; and non-accounting for other prescribed medications that may have been used for off-label AUD treatment. While efforts to increase AUD treatment in ambulatory care settings might be improved by targeting certain patients and physicians,6 further research is necessary to confirm the effectiveness and safety of off-label treatments. In addition, more work should be done to determine reasons for low AUD medication prescribing, including the impact of treatment goals, uncertainties about whether and for whom different medications may work, and patient preferences.1