FDA urges caution about withholding opioid addiction medications

We conducted a drug utilization analysis of concomitant use of benzodiazepines with buprenorphine products being used for medication-assisted treatment (MAT), using the Prescription Behavior Surveillance System’s (PBSS) data from 1/1/2013 to 12/31/2013. The PBSS includes de-identified, longitudinal data from eight participating state prescription drug monitoring programs: California, Ohio, Louisiana, Kentucky, West Virginia, Idaho, Maine, and Delaware.

Of 190,907 patients prescribed at least 7 consecutive days of buprenorphine therapy, 17.7 percent had at least one overlapping benzodiazepine prescription period of 7 or more days, and another 2.5 percent had at least one overlapping prescription period of less than 7 days. The median length of overlap in therapy across the eight states ranged from 29 to 41 days. The overlapping benzodiazepine and buprenorphine treatments were commonly prescribed by the same prescriber (ranging from 33.1 percent to 65.2 percent across the eight states) and dispensed by the same pharmacy (from 71.5 percent to 93.3 percent). PBSS data capture prescriptions filled for any payment source (e.g., cash, public insurance); however, prescriptions filled outside of the participating PBSS states may not have been captured, resulting in an underestimate of the number of patients who had concomitant buprenorphine and benzodiazepine prescriptions.

We also reviewed the published medical literature. Using prescription data from the Veterans Health Administration for fiscal year 2010, Park et al.3 noted that patients receiving MAT (n=5,409 for methadone; n=5,690 for buprenorphine) were commonly dispensed a benzodiazepine prescription. The median percent of patients who received concomitant methadone or buprenorphine prescriptions was 13.3 percent and 20.2 percent across multiple U.S. regions, respectively. These results from the Veterans Health Administration may not be generalizable to the entire U.S. population, but the finding of frequent overlapping prescriptions for benzodiazepines and buprenorphine was similar to findings from PBSS 2013 data described above.

Abrahamsson et al.4 analyzed data from a nationwide cohort of Swedish residents 18-50 years old who were dispensed methadone- or buprenorphine-based MAT between July 1, 2005 and December 31, 2012. Within the cohort of patients prescribed MAT, the authors compared the risk of fatal overdose, non-overdose-related mortality, and all-cause mortality during therapy episodes of prescribed CNS depressants to periods of time without CNS depressant therapy. CNS depressants included benzodiazepines (diazepam, oxazepam, lorazepam, alprazolam, nitrazepam, flunitrazepam, triazolam, midazolam, and clonazepam) and three non-benzodiazepine hypnotic drugs (zopiclone, zolpidem, and zaleplon). Of 4,501 patients in the MAT cohort, 32.4 percent filled a prescription for a benzodiazepine and 40.8 percent filled a prescription for one of three non-benzodiazepine hypnotic drugs. During periods of cotreatment with MAT and benzodiazepines, there was an elevated risk of non-overdose-related and all-cause mortality, although the associations were not considered statistically significant (Table 3). During periods of cotreatment with MAT and the non-benzodiazepine hypnotic drugs, there was an elevated risk of fatal overdose, non-overdose mortality, and all-cause mortality, but the associations were considered statistically significant only for fatal overdose and all-cause mortality. The hazard ratios for the non-benzodiazepine hypnotic drugs and benzodiazepines are not directly comparable because the authors did not adjust for the indication for use of these drugs. The authors estimated periods of active MAT based on prescription fill date and assumptions about how long MAT typically lasts, so periods of active MAT during follow-up may have been misclassified. Periods of active MAT may have also been missed if patients in the cohort received in-clinic MAT during follow-up. In addition, while the authors attempted to adjust for time periods when patients were using more than one concomitant CNS depressant, it is not clear whether the adjustment was sufficient to isolate the effects of a single CNS depressant when used concomitantly with MAT. Further, the authors’ adjustments for sex, age, and time-dependent variables representing previous nonfatal overdose, psychiatric inpatient treatment, and suicide attempt, may not have been sufficient to control for the greater severity of comorbid psychiatric conditions likely during periods of concomitant use of CNS depressant drugs. Other potential confounders, including socioeconomic status, were not available in this data source.

Table 3. Risk of Death Associated with Periods of Concomitant Therapy with MAT and Benzodiazepines or Non-benzodiazepine Hypnotic Drugs in Sweden, July 2005-December 2012*

*The table is based on the paper by Abrahamsson et al.4 Benzodiazepines included in the study were diazepam, oxazepam, lorazepam, alprazolam, nitrazepam, flunitrazepam, triazolam, midazolam, and clonazepam. The non-benzodiazepine hypnotic drugs included in the study were zopiclone, zolpidem, and zaleplon.

A study by Warner et al.7 examined trends in drug overdose deaths from 2010 through 2014 using data from the National Vital Statistics System, which aggregates nationwide death certificate data. In 2014, 3,495 drug overdose deaths involved methadone, and other CNS depressant drugs were frequently involved in those deaths (Table 4). The benzodiazepines alprazolam and diazepam were the first and fifth most frequently involved concomitant drugs.

New FDA analyses using the same data source found 322 drug overdose deaths in 2014 that involved buprenorphine. Concomitant use of CNS depressant drugs was also frequently observed in deaths involving buprenorphine. Alprazolam, clonazepam, and diazepam were the most frequently involved concomitant drugs at death. Although the absolute number of methadone-involved deaths in 2014 was ten times the number of deaths involving buprenorphine, the data source did not have information to adjust for confounding or differences in drug utilization between methadone and buprenorphine. Therefore, we could not assess whether the risk of overdose death with concomitant CNS depressant drugs differed between methadone and buprenorphine. A limitation is that the quality of data extracted from death certificates depends on the level of detail provided by medical certifiers, which can vary by jurisdictions and over time. We also do not know whether the drugs were being used to treat pain or opioid use disorder, just that they were involved in the death.

Table 4: Most Frequent Concomitant Drugs for Drug Overdose Deaths Involving Selected Opioids: United States, 2014

Selected results using National Center for Health Statistics (NCHS), National Vital Statistics System. Mortality files linked with death certificate literal text, constructed for analysis on October 6, 2016.

Source