The present study investigated the prevalence and characteristics of potential CDIs and DDIs in a sample of inpatients during treatment on an addiction-specific ward of a university hospital in Germany over a period of six years. Two different tools to detect potential drug interactions were used, namely the drugs.com classification for CDIs and UpToDate Lexicomp^® for the detection of DDIs. To the best of our knowledge, this is the first study to apply these tools for the detection of potential drug interactions in patients with CUD.

The mean age of our study population was approximately 36 years, and the most common psychiatric diagnoses beside CUD were also substance use disorders followed by depression. Thereby, our study collective showed great similarities with foregoing studies on inpatients with CUD in terms of age, sex and comorbidity burden (Ricci et al. 2021; Oladunjoye et al. 2022).

There are now several case reports and some reviews in the literature that illustrate potential clinical consequences of interactions between cannabinoids and individual drugs (Geffrey et al. 2015; Grayson et al. 2018; Nasrin et al. 2021). Evidence has been generated that the combination of cannabis use with intake of the anticoagulant drug warfarin increases the risk of bleeding complications (Damkier et al. 2019). The exact mechanisms behind this are partially unclear, although the results of a study by Bansal et al. indicate that CBD inhibits CYP2C19, but the previously postulated inhibition of CYP2C9 by THC was not verified (Bansal et al. 2023). Accordingly, the probability of interactions between cannabis and warfarin was classified as very high in a systematic review (Lopera et al. 2022). A similar level of evidence is also available for clobazam, whereby CBD can lead to the accumulation of clobazam via inhibition of CYP3A4 and CYP2C19 (Geffrey et al. 2015). However, several studies have yet investigated potentially beneficial effects of this pharmacokinetic interaction for the treatment of rare forms of epilepsy (Geffrey et al. 2015; Golub and Reddy 2021). A risk of pharmacokinetic and pharmacodynamic interactions with cannabis can also be stated for several other drugs (Lopera et al. 2022). Two recent systematic reviews of Nachnani et al. and Maldonado an Colleagues have proven that interactions of cannabinoids with concomitantly prescribed drugs are likely and the strongest evidence has been generated for warfarin, valproate, tacrolimus, and sirolimus (Maldonado et al. 2024; Nachnani et al. 2024). An evaluation which of these drugs are prescribed particularly frequently to patients with CUD is not available yet.

Within our study, the majority of patient cases was affected by potential CDIs. The most frequently prescribed drugs associated with major risk of CDIs were levomethadone, buprenorphine, and morphine. Levomethadone and buprenorphine are utilized for substitution treatment of patients with opioid use disorder, which illustrates the risk of potential CDIs in this patient group in particular.

A study conducted by Vierke et al. indicates that cannabis consumption leads to a reduction in the formation of norbuprenorphine and an elevation in the levels of buprenorphine and norbuprenorphine in the blood, likely due to the inhibition of CYP3A4 enzyme (Vierke et al. 2021). This pharmacokinetic interaction could potentially lead to heightened or modified opioid effects and an increased risk of intoxication (Vierke et al. 2021). Comparable pharmacokinetic interaction potentials have also been shown for levomethadone and methadone. This should result in clinical monitoring for opioid intoxication symptoms in patients with CUD and concomitantly prescribed opioid substitution therapy. On the other hand, positive effects of such interactions can also be utilized. So, Abrams et al. found that vaporized cannabis given to patients with chronic pain on opioid therapy (morphine or oxycodone) increased the analgesic effect of opioids (Abrams et al. 2020).

Among potentially moderate CDIs, the three most affected drugs were levetiracetam, quetiapine, and mirtazapine. The interaction potential of cannabinoids and antiepileptic drugs has repeatedly been characterized in the literature, although data on levetiracetam are sparse (Lucas et al. 2018). Hereby, a mouse study identified that CBD decreased antiseizure activity of levetiracetam against externally induced psychomotor seizures (Socała et al. 2019). Possible CDIs involving quetiapine and mirtazapine are also due to the influence on metabolization via CYP isoenzymes resulting in an increase of sedation and psychomotor slowing (Lucas et al. 2018). In general, drugs with influence on the central nervous system (CNS) were commonly involved in potential CDIs in our study population. This in turn suggests that vigilance should be monitored in patients with CUD, and that the indication for sleep-inducing and sedative drugs should be given rather cautiously.

Patients with CUD are often affected by polypharmacy due to their comorbidity profile and are therefore particularly susceptible to the development of ADRs caused by possible DDIs (Connor et al. 2021). However, a systematic recording of potential DDIs in patients with CUD has so far only been carried out for other addiction disorders, but not for CUD (Guerzoni et al. 2018; Schröder et al. 2024). In the context of AUD, benzodiazepines and disulfiram have been shown to be frequently involved in potential DDIs (Guerzoni et al. 2018). A retrospective cohort study by Schröder et al. also identified the combination of potassium supplements with renin-angiotensin-aldosterone system inhibitors to be commonly responsible for potentially severe DDIs in geriatric patients with AUD (Schröder et al. 2024).

The results of the present study have shown that DDIs are also a common phenomenon in patients with CUD. Buprenorphine and levomethadone were particularly frequently involved in the 196 potential DDIs in our collective, both in the category of combinations to be avoided and those to be critically evaluated. These drugs for opioid substitution treatment exhibited extensive interaction potential with other drugs that influence CNS functions (such as benzodiazepines and antipsychotic drugs). Therefore, these DDIs can also result in a potentiation of CNS depressive effects, analogous to the CDIs outlined above. This in turn possibly leads to states of confusion, risk of falls and reduced vigilance. Furthermore, most of the drugs involved in potential DDIs in our study bear the risk of prolongation of the QT interval (Sarganas et al. 2014). This should lead to electrocardiographic (ECG) controls in this patient population to prevent it from malignant cardiac arrythmias.

In summary, the results of the present study reveal that potential CMIs and DDIs are common among patients with CUD. The interaction potential of cannabinoids is especially due to their influence on the activity of CYP isoenzymes, which should be considered when prescribing drugs metabolized via this pathway in patients with CUD. In addition, drugs used for opioid substitution treatment were often involved in potential CMIs and DDIs. The interaction of levomethadone or buprenorphine with cannabinoids and other drugs with effects on CNS functions can in turn result in CNS depressive effects, whereby their occurrence should be clinically monitored. So, especially patients suffering from opioid use disorder in addition to CUD seem at risk for potential drug interactions and consecutive ADRs. The findings of our study indicate that a significant portion of drugs prescribed to patients with CUD should be critically evaluated in accordance with the drugs.com list and the UpToDate Lexicomp^® interaction check. UpToDate Lexicomp^® was used because of its easy availability and thus the potentially good reproducibility of our results and associated possibilities for implementation in clinical routine. On the other hand, the utilized interaction tools showed good performance in foregoing studies with regard to the detection of potentially clinically relevant drug interactions (Muhič et al. 2017; Marcath et al. 2018). However, it is important to note that none of the applied tools was specifically designed for addiction psychiatry but rather for assessing drug safety in patients with CUD. Therefore, a comprehensive assessment of prescribed drugs in patients with CUD requires thorough analysis of the benefits and risks, as well as careful consideration of alternative pharmacological options. Nevertheless, the results of our study should have some more implications for clinicians regarding the care of patients with CUD. As patients with CUD are particularly affected by CDIs and DDIs, withdrawal treatment from cannabis or other psychotropic substances should be done with caution in this patient group and should contain close monitoring of respiratory and cardiac functions as well as repeated measures of drug serum concetrations. Furthermore, since drugs used for opioid substitution treatment were frequently involved in CDIs and DDIs, complete withdrawal from opioids should be discussed in patients with opioid use disorder and comorbid CUD. In order to quantify effects of possible drug interactions, methods such as therapeutic drug monitoring (TDM) should be incorporated even more into clinical routine (Hiemke et al. 2018).

As limitations, the applied drug interaction tools do not specify pharmacological alternatives. Furthermore, our study cohort only contained recreational cannabis consumers, so no conclusions can be drawn regarding drug interactions with medicinal cannabis. Moreover, the monocentric design and its exclusive setting within a specialized unit of a university hospital has to be mentioned. Consequently, the generalizability of our findings to other healthcare settings may be limited. In addition, the retrospective nature of our study prevents us from determining whether the identified potential CDIs and DDIs resulted in adverse outcomes in our population. To overcome these limitations, future research should use a prospective design to thoroughly analyze risks of adverse outcomes associated with drug interactions in patients with CUD. Serum levels of the drugs potentially interacting with cannabis could then also be determined to enable a more detailed description of the clinical relevance of CDIs and their course during withdrawal treatment. Such studies will allow healthcare professionals to accurately stratify individuals with CUD based on their risk profile at the time of prescribing. In addition, randomized controlled trials should be conducted to prospectively assess whether addressing potential CDIs and DDIs can reduce the incidence of adverse effects in patients with CUD.