Comment on Domzaridou et al.: Recognising the complexities of co-prescriptions and lifestyle factors in opioid agonist treatment

We read with great interest the study by Domzaridou et al. [1], which examined the risk of non-fatal overdose (NFO) in patients with opioid use disorder who were prescribed opioid agonist treatment (OAT) alongside benzodiazepines, gabapentinoids, antipsychotics, antidepressants and Z-drugs. While applauding their valuable work, we note that differences in risky behaviour and lifestyles between the analysed subpopulations were not addressed.

For example, their finding that benzodiazepine co-prescribed patients have an increased weighted relative NFO risk of 1.3 (compared with other OAT-patients) seems unsurprising based on wider evidence of risk factors in OAT. It is well-recognised that people with opioid dependency who also misuse benzodiazepines are more difficult to engage in treatment and have increased risk profiles [2]. They are more likely to have injected recently, used borrowed needles, consumed cocaine and practised unsafe sex [3]. They also have more psychiatric morbidity, higher prevalence of alcohol use disorder [4] and more frequent contact with law enforcement [5].

Benzodiazepine dependence is associated with increased suicide risks, premature death and adverse childhood experiences [6, 7]. Notably, the drug related death-rate triples among individuals with opioid dependency if using illicit benzodiazepines [8] and all-cause mortality is increased too [9]. These cited studies all refer to illicit rather than prescribed benzodiazepines, suggesting that harm reduction programmes should not overlook benzodiazepine problems. Although illicit benzodiazepine consumption often involves chaotic use of very high doses [10], instalment-dispensed prescribed benzodiazepine seems to reduce risky behaviours [11-13].

Interestingly, the subsequent stratified analysis of the NFO risk by Domzaridou et al. [1] showed very little difference for the periods when benzodiazepine were not co-prescribed in this high-risk group, as judged by the largely overlapping confidence intervals for the periods with and without benzodiazepines (Figure 1). Regrettably, in the abstract, the stated NFO risk for co-prescribing benzodiazepines during OAT is 1.45, based on the figure from the sub-analysis, without mentioning that this did not differ significantly when benzodiazepines were not co-prescribed. This gives an exaggerated impression of the NFO risk for prescribing benzodiazepines.

That said, an intriguing result from Domzaridou et al. [1] is the increased NFO-risk for patients treated with buprenorphine in comparison with methadone when benzodiazepines or gabapentinoids were co-prescribed. This might seem counter intuitive [14], but one conceivable hypothesis for this could be that the more chaotic lifestyles associated with patients accessing these treatments are better stabilised by methadone than buprenorphine, given buprenorphine's inferior treatment retention [15]. Ultimately, in harm reduction schemes, we need to consider how best to promote healthy lifestyle changes while also recognising pharmacological interactions between the medications.

Adam Bakker: Conceptualization; writing—original draft. Alexander Smith: Writing—review editing. Michael Liebrenz: Writing—review editing.

Not applicable.

A.B. and A.S have none to declare. M.L. has represented the Swiss Federal Office of Public Health several times in the work of the Council of Europe-International Cooperation Group on Drugs and Addictions (Pompidou Group). The authors have no other competing interests to declare. Any opinions expressed are solely those of the authors.