Medication Treatment for Opioid Use Disorder Plus Benzodiazepine May Increase Overdose Risk

When medication treatment for opioid use disorder (OUD) is not started, discontinued, or when patients also use benzodiazepines at baseline, patients experience an increased overdose risk, according to findings published in the American Journal of Psychiatry.

Investigators from City College of New York and Columbia University Irving Medical Center pooled data for this study from 3 large randomized controlled trials of medication treatment for OUD: START ( Identifier: NCT01592461), POATS ( Identifier: NCT00316277), and X:BOT ( Identifier: NCT02032433). The studies had open-label designs and randomized patients (N=2199) to receive buprenorphine (n=1387), methadone (n=529), or extended-release naltrexone (n=283) MOUD. In this analysis, risk for overdose events through 24 weeks were evaluated.

The pooled study population included 1471 men, 1653 were White, 1725 used heroin at baseline, 1000 used stimulants at baseline, and 999 had a lifetime diagnosis of anxiety, panic disorder, or major depression.

Overall, 57 likely overdose events occurred in the studies, of which 35 occurred during X:BOT, 17 during START, and 5 during POATS. The most common culprit substance was opioids (n=28), specifically benzodiazepines (n=7), cocaine or a stimulant (n=4), and polysubstances (n=1).

Among overdose events, 51 patients were involved, in which 4 individuals had 2 overdoses and one had 3.

For the overdoses that occurred in the first 24 weeks (n=45), they occurred among 1.78% of the pooled study population. Compared with the entire group, rates of overdose differed significantly on the basis of medication treatment for OUD (c2, 21.62; P =.0002), benzodiazepine and heroin use at baseline (c2, 17.28; P <.0001), and benzodiazepine use at baseline (c2, 13.84; P =.0002).

The overdose rates were highest for extended-release naltrexone (5.30%) compared with methadone (1.51%) or buprenorphine (1.15%). However, among benzodiazepine users, the overdose rates were 8.96%, 7.41%, and 1.85% for naltrexone, methadone, and buprenorphine, respectively.

Some of the observed overdose events attributed to naltrexone may be due to the fact that 27.9% of those assigned to naltrexone never initiated medication. The rate of non-initiation was higher for naltrexone than for buprenorphine (2.2%) and methadone (1.7%).

Among participants who stopped taking their assigned medication treatment for OUD, 15 patients experienced an overdose event.

The strongest message from these data is that patients who fail to initiate medication, or stop their medication, are at greater risk of experiencing an overdose event.

Overdose risk was associated with not starting medication treatment for OUD (hazard ratio [HR], 6.44; 95% CI, 2.12-19.54), stopping MOUD (HR, 4.04; 95% CI, 1.54-10.65), and using benzodiazepines at baseline (HR, 3.36; 95% CI, 1.76-6.42).

Lower overdose risk was associated with buprenorphine use in 2 of the 3 buprenorphine trial arms compared with naltrexone (HR range, 0.12-0.36) but not in the third trial arm (HR, 0.68; 95% CI, 0.23-2.01). In addition, lower overdose risk tended to be associated with age (HR, 0.96; 95% CI, 0.93-1.00).

The major limitation of this study was that data were sourced from a mixture of clinician report, hospital records, and patient self-report.

Patients undergoing medication treatment for OUD remain at risk of overdose events in the first 24 weeks after seeking treatment, according to the study authors. “The strongest message from these data is that patients who fail to initiate medication, or stop their medication, are at greater risk of experiencing an overdose event.”

Disclosure: One study author declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures.


Brandt L, Hu M-C, Liu Y, et al. Risk of experiencing an overdose event for patients undergoing treatment with medication for opioid use disorder. Am J Psychiatry. 2023;appiajp20220312. doi:10.1176/appi.ajp.20220312