- Top of page
- SURVEY DATA
- PRESCRIPTIONS WRITTEN
- POISON CONTROL CENTER CALLS FOR HUMAN EXPOSURE TO DRUGS
- EMERGENCY ROOM CASES
- TOXICOLOGY FINDINGS
- DISTRIBUTION OF PRESCRIPTION DRUGS
- Declaration of Interest
Death data from the National Center for Health Statistics (NCHS) are classified according to the ICD-10 codes, which group drugs together into nonspecific codes that may not identify the drugs separately. “Methadone” has its own classification, but “Other Opioids” include morphine, oxycodone, hydrocodone, codeine, and percodan, as well as other opioid formulations. “Other Synthetic Narcotics” include buprenorphine, fentanyl, meperidine, and propoxyphene, and “Other and Unspecified Narcotics” includes narcotics and opiates not otherwise categorized. Thus, it is not currently possible to identify buprenorphine deaths from other deaths in the “Other Synthetic Narcotics” category.
Of the illicit substances in poisoning deaths across the United States, cocaine is the drug most likely to be mentioned, followed by “other opioids” (Fig. 3). Between 1999 and 2006, cocaine deaths increased from 3,832 to 7,475, other opioid deaths increased from 2,757 to 7,035, methadone deaths increased from 786 to 5,416, and heroin deaths increased from 1,964 to 2,090.20Table 2 shows the combinations of drugs reported nationally by NCHS in deaths mentioning methadone.
Figure 3. U.S. narcotic substances mentioned in poisoning deaths: NCHS 1999–2006. Source: Warner et al., Methadone Deaths, 1999–2005, Findings from the National Vital Statistics, PowerPoint presentation.
Download figure to PowerPoint
Of the poisoning deaths with a mention of methadone by age group in 2005, 13% were ages 15–24 years, 22% were ages 25–34 years, 27% were ages 35–44 years, 29% were ages 45–54, 7% were ages 55–64 years, and 1% was 65 and older. The death rate in 2004–2005 for males was higher than for females, with the highest rate (5.8) for males at age 41 and the highest rate for females (3.9) at age 43.21
The National Poison Data System of the American Association of Poison Control Centers reports the outcomes of calls concerning human exposure to buprenorphine or methadone. Table 1 shows that the number of methadone exposures known to have resulted in a death increased from 26 in 2000 to 103 in 2008, while the number of known buprenorphine deaths increased from 0 in 2000 to 2 in 2008. It should be noted that the deaths reported by the poison centers are only a small portion of all the deaths, as the centers only report deaths that were followed up after a call to the center. Hence, if the center does not get a call, it will not know about a death.
Of the 213 methadone deaths reported to RADARS® from 2003 through 2008, 8% were ages 13–19, 29% were in their twenties, 17% were in their thirties, 25% were in their forties, and 14% were aged 50 and older. The source of the drug was unknown in 75% of the cases, with 16% being the decedent's own medication, 8% from family member or friend, 2% bought on the street, and 0.5% stolen. Of the formulations involved in the deaths, 72% were unknown, 24% were tablets, 3% were liquid, and 0.5% were diskettes. The exposure reasons were unknown for 57% of the deaths, with abuse/misuse as the reason in 32%, suicide as the motive in 10%, and therapeutic error as the reason in 2% of the deaths.22Table 2 shows the combinations of drugs involved with methadone as reported by the poison centers.
Of the 2007 Texas deaths with a mention of methadone, 34% were female, 73% were White, 5% were Black, and 22% were Hispanic. Average age was 39 years, with 7% being age 19 and younger, 21% in their twenties, 18% in the thirties, 30% in their forties, and 25% age 50 and older. The 2007 methadone deaths reported by Texas poison control centers and those 2005 poisoning deaths with methadone as reported by NCHS were older than those reported by RADARS®.
The number of overdose deaths of clients in Texas narcotic treatment programs has remained fairly level over time (Fig. 4), although the proportion of these overdose deaths has decreased from 20% of all methadone deaths in 2000 to 6% in 2007.23 In most cases, the toxicological findings on these methadone treatment clients did not indicate if the methadone serum concentrations were within the therapeutic range.
The combinations of drugs being used with methadone are changing. The proportion of Texas deaths involving combinations with prescription or licit drugs (which may or may not have been prescribed for the individual) increased from 17% in 1992 to 46% in 2007, while combinations involving illicit drugs such as cocaine decreased from 25% to 15%.
- Top of page
- SURVEY DATA
- PRESCRIPTIONS WRITTEN
- POISON CONTROL CENTER CALLS FOR HUMAN EXPOSURE TO DRUGS
- EMERGENCY ROOM CASES
- TOXICOLOGY FINDINGS
- DISTRIBUTION OF PRESCRIPTION DRUGS
- Declaration of Interest
This paper has reviewed publicly available data relating to use, abuse, misuse, and distribution of methadone and buprenorphine. These individual data sets are constrained by significant limitations. Data are neither available on buprenorphine admissions to treatment nor on deaths involving the drug. Information is inconsistent on formulation of the drugs used, sources of the drugs, age groups of users, available years of data (2006 vs. 2008), whether the drugs were prescribed for the individual, and the frequency of combinations of these two drugs with other drugs. Yet, when these pieces of information are drawn together, a clearer picture emerges.
Between 2000 and 2008, the number of methadone tablets (5, 10, and 40 mg) sold through retail outlets increased from 110,670,209 to 728,344,627, the number of human exposure calls to poison control centers increased from 1,387 to 4,765, the number of methadone items identified by NFLIS laboratories increased from 503 to 7,525, and grams of methadone shipped as reported to ARCOS increased from 1,826 to 5,283. The number of methadone cases seen in DAWN emergency rooms increased from 48,864 to 69,506 between 2004 and 2007, and the number of deaths in NCHS with a mention of methadone increased from 786 to 5,416 between 1999 and 2006.
The proportional increases in buprenorphine indicators are much larger, but this is due to the fact the noninjectable formulation did not come to market until 2002. Between 2000 and 2008, the number of poison control center human exposure calls increased from 13 to 2,607, and between 2004 and 2008, the number of Suboxone® tablets sold at the retail level increased from 7,707,099 to 106,537,708, and the number of Subutex® tablets increased from 1,397,294 to 10,119,599. The number of NFLIS items identified as buprenorphine increased from 15 to 4,034 between 2000 and 2008 and the number of buprenorphine dosage units reported by ARCOS increased from 1,210 to 44,698 between 2003 and 2008. The number of DAWN ED cases increased from 1,001 to 10,229 between 2004 and 2007. These increases are indications of the exposures to and diversion of the drug since 70% of the DAWN cases were reported to be related to nonmedical use of the drug.
Depending on the data source, anywhere from 38% to 55% of the poison control center and DAWN ED methadone cases involved only methadone, as did 31–33% of the deaths involving methadone. However, these data indicate that the majority of these adverse events and deaths involved methadone in combination with other drugs. According to the Texas death data, combinations of methadone involving only illicit drugs are decreasing, while combinations involving prescription opioids (whether prescribed or not), are increasing.
The use of methadone for pain, as compared to treatment for opioid dependence, can be tracked through data reporting the formulation of the drug. According to the 2008 ARCOS data, distribution of the 5 mg and 10 mg methadone tablets has now surpassed distribution of the liquid used in narcotic treatment programs and Texas poison control center statistics confirm the increases in calls involving the tablets. Further, RADARS® data reported that formulation of methadone was unknown for 72% of deaths, but 24% of deaths could be attributed to methadone tablets and 3% of deaths were traced to methadone liquid. These data, when considered with the ARCOS data and Texas data, appear to point to the increasing role of methadone tablets used in the treatment of pain as a more prevalent source of methadone toxicity than methadone from narcotic treatment programs where the drug is dosed in a liquid form.
The data showed the methadone user was more likely to be male (depending on the dataset, between 53% and 67%), White (74% and 77%), with the highest prevalence of use among those of ages 21–24, according to NSDUH, DAWN, and NCHS. However, their use of methadone continued over the years with adverse events noted in Texas poison center calls at an average age of 36, the highest death rates in NCHS being 41 years for males and 43 years for females, and DAWN ED visits continuing at high levels through ages 45–54.
Buprenorphine numbers are scarcer due to ICD-10 coding. However, the rapid increases in buprenorphine as documented in the number of prescriptions written, poison center exposures, DAWN, NFLIS, and ARCOS provide evidence of the need to monitor the potential abuse of this drug. Buprenorphine cases are less likely to occur among males (47–53% females) and less likely to occur in persons of color (87% White in DAWN). The age group seen most often in DAWN data was 21–24 years and the average age for the Texas poison center calls was 35 years.
Buprenorphine is also a more difficult drug to track, since it can be prescribed for pain, for opioid treatment by office-based practitioners, as well as for use in opioid treatment programs. Thus, there is no reliable way to determine the level of buprenorphine distributed for pain as compared to office-based narcotic treatment, although ARCOS reported that only 1% of the buprenorphine tablets were distributed to narcotic treatment programs.
As of 2008, there is even less data on Subutex® in comparison to Suboxone®. Because of the small numbers of Subutex® cases, DAWN suppression rules limited analysis of the data and the numbers of different combinations could not be determined. The data do not tell us if Subutex® users differ in terms of sociodemographic characteristics. However, the increases in the number of Subutex® prescriptions should be recognized. The buprenorphine/naloxone combination tablet is recommended for use in the treatment of opioid addiction,26 with limited use for Subutex®, which is reserved for pregnant, opioid-dependent women. However, methadone is still considered the standard of care for the treatment of these women27 and the limited use for pregnant women seems unlikely to be the source of the large increase in Subutex® prescriptions. One possibility is that physicians are prescribing this formulation for the treatment of opioid dependence when patients complain that the combination tablet is not well-tolerated. Another possibility is that physicians prescribe Subutex® for treatment of pain (an indication that is not FDA-approved). The greater abuse liability for the buprenorphine-alone formulation is also confirmed by the data in this review and underscores the need for physicians to carefully consider appropriate use of buprenorphine-alone formulations, given the literature on injection of Subutex®.28–30
Additional data are needed on reasons for use of these two drugs. DAWN reported 7% of the methadone cases were adverse reactions to the patient's own drugs and 16% of the buprenorphine cases also sought help for adverse reactions. The RADARS® data reported 16% of the methadone decedents took their own medications and therapeutic error was the exposure reason in 2% of the deaths. Determining the role of drugs taken by the patient as prescribed, compared to using nonprescribed drugs or drugs taken in a manner not prescribed, is a major concern. The fact that the number of overdose deaths of clients in Texas narcotic treatment programs remained stable while the total number of methadone deaths in the state increased is evidence that the problem with methadone-associated deaths is not related to diversion of methadone from narcotic treatment programs.
Analysis of the NFLIS, N-SSATS, and ARCOS data found that availability of the two drugs differed by census region, with more buprenorphine used in treatment, distributed, and found in toxicology examinations in the northeast. More methadone was examined in toxicology laboratories in the south, with more methadone used in treatment and distributed in the northeast. Because these three datasets can be specific as to location (ARCOS and N-SSATS are available at the zip code level), tracking “hot spots” where levels of distribution and seizures are high, but with no opioid treatment programs present to distribute methadone or buprenorphine can be a way to identify areas warranting additional attention by law enforcement and licensing boards for diversion or inappropriate prescribing.
Methadone and buprenorphine must also be considered in the context of heroin and other opiates. Heroin indicators including treatment admissions, DAWN ED visits, and NCHS death data have remained fairly constant over time while indicators for other opiates have increased. Since 1992, the gender and racial characteristics of the clients having problems with illicit methadone have changed so that by 2007, they more closely resembled other opiate clients. All treatment clients, whether heroin-addicted, illicit methadone-addicted, or other opiates-addicted had substantial rates of psychiatric illness (range: 25–37%) (Table 4).
Another issue of significant concern in these data is the high rate of use of other drugs, both licit and illicit in combination with methadone and buprenorphine. DAWN ED cases show significant increases in cases involving methadone and other pharmaceuticals as well as methadone in combination with alcohol, other illicit drugs, and other pharmaceuticals (Table 3). Similarly, a myriad of other drugs in combination with methadone and with buprenorphine have been reported in data collected through RADARS®, NCHS, Texas deaths, DAWN ED, and the Texas poison centers (Table 2).
Consistently, co-occurring use of benzodiazepines, particularly alprazolam, has been reported in both methadone and buprenorphine DAWN ED, poison center, and death mentions (Table 2). These findings might be expected, given recent findings from the NSDUH, which show a significant increase (p = .05) in misuse of tranquilizers from 2002 to 2006.31 Further, the most commonly misused tranquilizers among lifetime nonmedical users were Valium® (diazepam), Xanax® (alprazolam), Ativan® (lorazepam), and Klonopin® (clonazepam), in that order.32 The rate of benzodiazepine use in the population of methadone- and buprenorphine-treated patients is high. Lifetime use of benzodiazepines in methadone-maintained patients has been reported to be as high as 94%.33 In addition, Iguchi and colleagues reported that diazepam, lorazepam, and alprazolam were frequently used by methadone clients for their “high” producing effects and for selling to produce income. They also reported that, in contrast, chlordiazepoxide and oxazepam had much lower ratings of “high” and were less likely to be obtained for getting “high” or for resale.33 Similar findings for buprenorphine with benzodiazepines have also been reported.34–36
Not all methadone- and buprenorphine-treated individuals seek benzodiazepines with the intent to abuse these drugs. Many of these patients experience significant anxiety, insomnia, and withdrawal symptoms that lead to requests for ancillary medications;37,38 benzodiazepines are often the drugs prescribed for these symptoms. One of the clinical dilemmas in treating such patients is the differentiation of true psychiatric disorders from drug-seeking behavior. However, with exceptions to be discussed below, a sound clinical approach to psychiatric complaints in this population is to monitor for a period of time rather than prescribing immediately; select nonbenzodiazepine interventions whenever possible (e.g., cognitive-behavioral therapy); and, if necessary, utilize certain of the selective serotonin reuptake inhibitors (below) for anxiety and depressive symptoms; address other co-occurring drug and alcohol abuse that may be contributing to the symptoms presented; and provide education on sleep hygiene.
Close monitoring of the use of benzodiazepines and use of long-acting medications such as clonazepam are preferred and must be undertaken if they are to be prescribed. These medications should be prescribed for a short (e.g., 2–4 months) period to prevent development of tolerance and dependence.39 Further, frequent re-assessment to determine positive benefit (e.g., increase in level of function) should be documented. If improvement does not occur, it should lead the clinician to question the benefit of the benzodiazepine treatment, given the risks that are apparent from the data in this paper.
Aside from misuse/abuse/overuse of these drugs with methadone and buprenorphine that can lead to toxicity, there are other important pharmacological reasons to expect that benzodiazepines in combination with opioids might produce adverse events and even deaths. Benzodiazepines, like opioids, depress the central nervous system, albeit via different pharmacological mechanisms. One of many effects of the binding of an opioid agonist to the mu opioid receptor is sedation and altered mental status.40 Benzodiazepines produce similar effects, but via binding to the gamma-aminobutyric acid (GABA)-A receptor, producing increased chloride ion flow into the neuron and resulting in decreased membrane excitability.41 A combined pharmacodynamic effect of these drugs can be to depress the central nervous system to the point of toxicity. Indeed, what little information exists on pharmacokinetic interactions between methadone and benzodiazepines indicates little in the way of significant interactions.42 Pharmacokinetics studies between buprenorphine and benzodiazepines have not been reported to date, despite several reports of buprenorphine toxicity in humans co-abusing these drugs.43,44 The paucity of studies available in this area of research underscores the need for it to be undertaken.
Another area of concern identified from the data in this paper is the toxicity data related to methadone and buprenorphine in combination with licit medications likely prescribed for medical and psychiatric conditions. Of greatest concern is the consistent reporting of adverse events and deaths across data sets (Table 2), in which methadone or buprenorphine occur in combination with psychotropics such as selective serotonin reuptake inhibitors (in particular, fluoxetine, paroxetine, fluvoxamine, sertraline, citalopram, and escitalopram). While these drugs are not known to produce mood-altering effects (e.g., “high,” sedation), they share a common property of inhibition of cytochrome (CYP)P450 2D6,45 an enzyme that plays a role in methadone metabolism.46,47 Further, norfluoxetine, a metabolite of fluoxetine, and fluvoxamine potently inhibit CYP P450 3A4, which is important to the metabolism of methadone and buprenorphine. CYP 450 2C19 is also important to the metabolism of buprenorphine and its function is inhibited by fluoxetine and fluvoxamine.48 Inhibition of the metabolism of long-acting opioids such as methadone and buprenorphine could lead to toxicity, which appears to be the case based on the data in this study. Clinicians need to be aware of the potential for such toxicities and judiciously prescribe these medications.
If an SSRI must be used in methadone-maintained patients, selecting those that do not affect both CYP 450 2D6 and 3A4 would be preferable (choices would include paroxetine, sertraline, and citalopram, escitalopram). If there is a clinical need for use of these medications in methadone-maintained patients, monitoring should be considered to include a baseline cardiogram with a repeated cardiogram after 2 weeks at a stable dose of antidepressant medication in order to follow the cardiac QT interval that can increase should methadone metabolism be inhibited, leading to greater exposure.49 Patients receiving these medications should also be monitored for any increased opioid effects that they may be experiencing. Evidence of toxicity should trigger re-evaluation and alteration of the treatment plan. Another consideration is the use of buprenorphine in patients needing an opioid therapy, but who would also benefit from an SSRI treatment of a mental disorder due to its ceiling effect at which increasing exposure is not associated with increasing opioid effects.50 Further, buprenorphine metabolism is less likely to be affected by SSRIs that primarily effect CYP 450 2D6, but not CYP 450 3A4 (e.g., paroxetine, sertraline, citalopram, escitalopram).51
Other medications mentioned as being present in cases of toxicity or deaths related to methadone or buprenorphine use and that inhibit the function of CYP 2D6 include duloxetine, clomipramine, doxepin, perphenazine, and hydroxyzine. These medications are commonly used for treatment of psychiatric disorders including depression and anxiety as well as psychosis (perphenazine). Amitriptyline, frequently used for the treatment of depression as well as chronic pain, is mentioned in RADARS® data for methadone. This medication is not only itself sedating with potential for pharmacodynamic interactions with methadone and, potentially, buprenorphine, but it is an inhibitor of both CYP 450 3A4 and 2D6.52 Similarly, atypical antipsychotics, specifically quetiapine, olanzapine, and risperidone are mentioned in association with methadone and buprenorphine use. Quetiapine has been reported to be abused by injection and nasal insufflation, particularly with cocaine in a combination referred to as a “Q-ball.”53 It is an inhibitor of CYP 450 2D6 and is an inhibitor of the efflux transporter p-glycoprotein.54 Quetiapine should not be considered a first-line treatment of anxiety or psychosis in methadone or buprenorphine-maintained patients. Interestingly, risperidone and olanzepine do not inhibit metabolic enzymes related to methadone or buprenorphine metabolism, but have been cited in toxicity reports.55,56
It is beyond the scope of this paper to discuss the potential interactions of every drug mentioned in these adverse events related to methadone and buprenorphine use. However, it is useful for clinicians to be aware of the wide variety of substances that may be associated with toxicities in patients treated with these opioid therapies or who are abusing these medications. What are the implications? Do these data indicate that physicians should avoid the use of many psychotropic and analgesic drugs in patients with opioid dependence? We would suggest that this paper only provides information that a clinician can use in considering treatment plans. When possible, clinicians might want to avoid medications known to have the potential to produce pharmacokinetic or pharmacodynamic interactions in patients. Another possible consideration is to undertake a genetic test to determine the level of activity of a patient's CYP2D6 enzyme, since there are well-documented genetic polymorphisms in this enzyme that include both rapid and poor metabolism of drugs.57
Clinicians should become accustomed to considering the possibility of drug interactions with medications they prescribe. These data underscore the need to educate patients about the potential for adverse drug interactions when these medications are used together—or when they are misused or abused together. Patients should ideally be receiving medications from one prescriber; when this is not possible, it is imperative that prescribing clinicians work together to avoid such toxicities. Patients must be educated about the need not to share medications and to use medications only as prescribed. Further, they must be told about possible side effects that should be reported to the treatment provider so that rapid intervention(s) can be made if needed. Patient treatment agreements outlining these parameters of treatment may be useful in this population.
This confusing picture underscores the need for significant improvement of the indicators, particularly as new abuse-resistant opioid drugs come to market. New drugs are being introduced, but they are placed in ICD-10 codes such as Other Synthetic Narcotics, which are nonspecific. However, since 2003, the narrative text on the death certificate (referred to as “death certificate literals”) has been retained by NCHS for analysis and work is underway to translate details on the death certificates into data that can expand the nonspecific ICD-10 code categories. In addition, some states with centralized medical examiner systems, such as New Mexico,58 Florida,59 and Utah,60 have information on the specific drugs and formulations, but they cannot be used to provide national information, since drugs and drug use patterns, particularly use of the opioids, vary across the states. The 2007 SAMHSA Methadone Mortality Meeting recommended that uniform nomenclature, case definitions, and standards for toxicological testing be used to encourage uniform reporting by medical examiners and coroners and to support analysis.61 This recommendation for standardization was expanded to include all the drug-related deaths in a SAMHSA Expert Panel meeting in August, 2009.
Improvement in the collection and reporting of indicator data is needed if FDA and the manufacturers are to meet the requirements of the Risk Evaluation and Mitigation Strategies (REMS).62 The key elements of the REMS will be education and certification for prescribers, pharmacists, other healthcare providers, or institutions that dispense or directly administer covered opioid products, as well as patient education. The REMS will also use metrics to assess the success of the REMS program in decreasing abuse and misuse of opioids while ensuring they remain available for patients who suffer daily from chronic pain. The available data systems discussed in this paper need to be fully supported and the data reported by manufacturers to FDA under REMS must be freely available to researchers and public health officials who must monitor the effects in their areas.
While improvements to data collection systems are essential to improving public health related to prescription opioid abuse, the data collected and presented in this paper make several important points. First, they underscore the rapid increases in prescribing of methadone and buprenorphine, the substantial and sobering statistics showing large increases in toxicities and deaths related to misuse of these drugs, and the increases in narcotic trafficking with these medications. Second, the toxicities reported with these drugs in combination with other drugs, particularly other licit medications, underline the lack of understanding or consideration by physicians of the possible toxic effects related to drug interactions. While only a few of these data show which toxicities occurred in those prescribed these medications versus those who illicitly obtained and abused these drugs, it is still incumbent on physicians to be aware of potential toxicities and to educate their patients regarding these issues. The data in this manuscript may help to assist with that process.