This was the first study that utilized multiple data sources to systematically examine methadone-drug interactions since the MMTP was implemented in Taiwan and to determine demographic and clinical correlates of comedication and MDIs among MMTP patients. Owing to the high coverage rate of the NHI (99%) in Taiwan, we were able to use the comprehensively collected national clinical data to characterize MDIs. Because the probability of MDIs and associated morbidities increases with polydrug use and some MDIs (e.g., an enhanced risk of overdose and poorer retention) may intensify poor outcomes among MMTP patients, it is clinically important to understand factors affecting MDIs in order to inform early detection and prevention of MDIs and to reduce adverse effects.
The majority of identified MDIs were associated with benzodiazepines (38.1%), the most-prescribed therapeutic drugs found in our participants. By comparison, higher proportions of patients (51.5% and 73.0%) were prescribed benzodiazepines in Switzerland  and Germany , respectively. Previous studies have shown that anxiety disorders were highly prevalent among opioid-dependent MMTP patients [47–49], and therefore many patients might have used benzodiazepines. Furthermore, MMTP patients who use or abuse benzodiazepines can have a negative clinical effect because they may have higher levels of addiction, use more illicit substances, be affected by other mental and social problems, and have difficulty in retaining in treatment [49, 50]. Co-administration of benzodiazepines and methadone may increase the risk of methadone overdose or even death [18, 51, 52]. In this study, the most commonly prescribed benzodiazepine was alprazolam (14.2%), which is also frequently reported to be involved in methadone-related deaths [53, 54].
Co-usage of opiate agonists and partial agonists (29.7%) with methadone was the second common pattern of use in this study. Combinations of methadone with buprenorphine, pethidine, and tramadol were also found in a study conducted in China , which are consistent with the results of this study. In particular, we found that morphine and pethidine were the most frequently prescribed opiate agonists among participants. The reason for co-administration may be a result of MMTP patients' surgical use or seeking of additional opiate analgesics to alleviate their craving. Patients in an opioid maintenance treatment program can receive buprenorphine or methadone, but normally they would not be prescribed simultaneously. However, a MMTP patient was prescribed with buprenorphine to relieve pain after surgery (Table 4). In addition, Manchikanti et al. reported that a significantly increasing proportion of patients receiving controlled substances were revealed to use illicit drugs or other prescription opioids for nonmedical use. Therefore, clinicians should be cautious when MMTP patients have received opiate agonists, as they may enhance the risk of addictive effects or opioid toxicity .
Clinically significant MDIs may share one of the three common drug interaction mechanisms, which include (i) increasing the QTc prolonging effect, the most serious ADR of MDIs, which may consequently develop into fatal TdP [19–21]; (ii) enhancing addictive CNS and respiratory depression, which cause methadone-related deaths ; (iii) triggering opioid withdrawal symptoms, which may cause an increase in catecholamines plasma concentrations, leading to the development of stress cardiomyopathy  that, in turn, can cause death. Potential ADRs were also observed among patients who had MDIs in this study. In addition to the three mechanisms stated above, MDIs  caused by co-administering CYP3A4-inducing or CYP3A4-inhibiting agents should be carefully monitored and addressed. Moreover, studies [27, 59, 60] have suggested that CYP2B6 plays an important role in mediating methadone metabolism. Thus, caution should be taken when combining methadone with CYP2B6-inducing agents, such as carbamazepine, phenytoin, rifampin, and phenobarbital, or with CYP2B6-inhibiting agents, such as paroxetine, sertraline, and desipramine. However, because of the difference in the genetic polymorphism of CYP 450 between Oriental and Caucasian populations, the results may not be generalized to those of the US and Western Europe.
The identified correlates for co-medication included use of amphetamines (e.g., methamphetamine) and co-infection with HCV. Poly-pharmacy or poly- substance use has been found to be common among MMTP patients, especially use of amphetamines and benzodiazepines . On the other hand, chronic HCV infection may exhibit some symptoms, such as fatigue, nausea, loss of appetite, muscle ache, flu-like symptoms, and depression, or may even result in the development of cirrhosis or liver cancer , which may increase use of other illicit or non-prescribed drugs to relieve symptoms (e.g., self-medication). For instance, we found that 89.9% (160/178) of patients were infected with HCV, but only 3.4% (6/178) patients received HCV therapy (ribavirin plus peginterferon alpha 2A or 2B). A very low proportion of HCV treatment has been observed in MMTP patients .
The prevalence of cigarette use among MMTP patients was threefold the rate than that in the general population, and cigarette use increases the risk of morbidity and mortality . Smoking is also associated with current mood or anxiety disorders among MMTP patients, especially in women . In line with the literature, some sedatives or antidepressants were found to be frequently used in our MMTP patients, such as sertraline, paroxetine , fluoxetine, amitriptyline , and zolpidem , which may affect the serum methadone concentration due to the involvement of the same CYP450 enzymes with methadone metabolism. Further, methadone interacts not only with the above drugs, but also with nicotine, which may increase euphoria and decrease restlessness, irritability, and depression .
Identification of clinical factors associated with MDIs will assist physicians in eliminating or preventing potential risks. Male patients were strongly associated with having MDIs (AOR = 4.88, P = 0.02). Moreover, the odds of MDIs in patients with co-medications per MMTP participation months were 1.41 times greater than those without co-medications. Younger age was weakly associated with decreased odds of MDIs by 0.94. This can be explained in part by their lower rate of cardiovascular diseases, which are associated with serious ADRs (i.e., TdP) and are more prevalent among older patients . In this study, some common ADRs, such as sweating, constipation, and insomnia, were noticed during the treatment, though two patients had been diagnosed with arrhythmia before this study, and one patient had tarchycardia related to a suspected methadone- chlorpromazine interaction in this study. Therefore, the prescription of medications that could induce QT interval prolongation should be avoided in these patients.
These results should be interpreted within the context of the following limitations. First, this study was conducted in southern Taiwan, and the results may not be generalized to other regions in Taiwan. Because MMTP patients may have different attitudes towards or habits related to the treatment of diseases, some may use over-the-counter (OTC) medications or Chinese herbal medicines to alleviate symptoms. Second, because OTC medications or Chinese herbal medicines are not covered by the national health insurance program, the interaction events resulting from OTC drugs-methadone or Chinese herbal medicines-methadone may be underestimated among MMTP patients who used additional OTC medications or Chinese herbal medicines. Moreover, flunitrazepam, a benzodiazepines drug, was not included in Micromedex® or Lexi-Interact™. Hence, we have not included this potential MDI in our list, and such MDIs may be underestimated. However, it should be noted that this kind of MDI may often be missed by clinicians. Finally, our definitions for
MDIs depend on the level of completeness of the available data banks (Micromedex® and Lexi-Interact™). To minimize bias, several well-documented databases were also integrated in this study.
In conclusion, to improve clinical care and prevent MDIs and MDI-related deaths, it is important that frequent MDIs and their characteristics among MMTP patients are identified and incorporated into care management. The present study utilized multiple data sources (a national medical database, face-to-face interviews, medical records, and methadone computer databases) to systematically characterize MDIs to help inform clinical care. The results demonstrate clinical evidence of significant MDIs (quinolones, benzodiazepines, opiate agonists and partial agonists) among MMTP patients in Taiwan. Given the potential adverse effects from co-usage of methadone with benzodiazepines or opiate partial agonists, clinicians should be cautious in prescribing these medications to MMTP patients and incorporate clinical monitoring of potential ADRs of MDIs into treatment plans.
To protect the privacy of MMTP patients, it is difficult to interconnect medical data from different sources. However, the incumbent agency should be responsible for developing pharmacovigilant measures to improve the quality of essential MMTP care services and to prevent the MDIs from occurring. Meanwhile, it is recommended that physicians should check the past and current medication history of MMTP patients, especially who have ever used quinolones, benzodiazepines, opiate agonists and partial agonists. From the findings in this paper, it is desperately needed to provide proper training to prescribing clinicians/providers in MMTPs and generalists in primary care to pay special attention when prescribing methadone and concomitant medications for co-occurring disorders. At the same time, it is also imperative to educate patients and their families of the danger of methadone drug interactions and methadone overdose.
In the future, it will be of interest to look into the association between the patients' doses and genetic polymorphism that affects methadone metabolism, as well as their impacts on the frequency or severity of MDIs.