Warnings Unheeded:The Risks of Co-Prescribing Opioids and Benzodiazepines

“Opioid analgesics should be used with caution when combined with CNS depressant drugs.”¹

Despite this common warning found in virtually all opioid product package inserts, reports on prescribing practices suggest that this warning may be too often ignored with grave consequences. Not only are opioids and benzodiazepines (the most commonly prescribed sedative class) widely prescribed in the United States, they are commonly co-prescribed for patients. Analysis of a commercial database on prescription drugs by the CDC indicated that, in 2012, health care professionals wrote 82.5 opioid prescriptions and 37.6 benzodiazepine prescriptions for every 100 people in the U.S.² Moreover, of the top 25 medications dispensed from U.S. pharmacies, three opioids and one benzodiazepine made the list (hydrocodone/acetaminophen [#1], tramadol [#20], and oxycodone/acetaminophen [#23] alprazolam [#13]).³

Recently released draft guidelines by the Centers for Disease Control⁴ for the treatment of chronic pain recommend that “providers should avoid prescribing of opioid pain medication and benzodiazepines concurrently whenever possible.” This recommendation is in response to the finding that these drugs are frequently prescribed in combination to patients with pain conditions. For example, one study examined data from a U.S. Veterans health care system and reported that among patients receiving long-term (i.e., ≥ 3 months) benzodiazepine treatment, those patients receiving the highest dose equivalents (typically exceeding the recommended guidelines) were more likely to be simultaneously receiving a prescription for oxycodone.⁵ A large insurance claims analysis reported that patients who are co-prescribed opioids and benzodiazepines for chronic pain are also prescribed opioids at higher doses and for longer.⁶ Co-prescribing is not only a problem in patients with pain. High rates of co-prescribing have also been documented for those enrolled in opioid maintenance therapy for opioid dependence in the VA Health Care System, with an estimated 13% and 20% of patients receiving methadone and buprenorphine (respectively) also receiving prescriptions for benzodiazepines.⁷

It is important to note that when discussing co-prescribing, this does not mean that the prescriptions necessarily are coming from the same physician. Patients often legitimately see multiple doctors for different conditions, and may have multiple prescribers. In some practice organizations, prescribers may be able to review the full list of prescribed medications, but not always. Moreover, some patients illegitimately engage in doctor shopping in order to obtain controlled drugs by prescription. The broader availability of functional state prescription monitoring programs (PMPs) has provided a new tool for physicians to circumvent both of these hurdles. Regular review of prescribing records from the PMP (whether mandated by the government or not) is recommended when treating patients with controlled substances.

Fatal Consequences: Concomitant Opioid and Benzodiazepine Use

Each year since 2000, overdose deaths from licit and illicit drug use in the general population in the United States have increased.⁸ In the 12-year span between 2001–2013, there was a three-fold increase in deaths involving prescription opioids.⁹ Although prescription opioids contribute to the highest rates of overdose deaths, benzodiazepines have also been implicated in overdose death. Data from 2010 suggest that nearly 60% of overdose deaths were due to a prescription drug (rather than illicit drugs), with prescription opioids contributing to the greatest number of deaths at 75%, while benzodiazepines contributed to approximately 29%.¹⁰

In populations of patients who receive prescription opioids for the treatment of pain (particularly chronic pain), the risk of overdose is a serious concern.¹¹ Although several factors may contribute to this risk (e.g., opioid dose, co-morbid respiratory conditions), one of the greatest risks is from concomitant use of benzodiazepines. In the United States, deaths from co-prescribed opioids and benzodiazepines increased 14% per year between 2006 and 2011.¹² In one study from West Virginia, filling an opioid prescription in the 6 months prior to death increased the risk of drug overdose death three-fold, compared to those who filled a controlled drug (non-opioid/non-benzodiazepine) prescription; those filling a prescription for a benzodiazepine had a 7-fold increase in death. Critically, patients filling both opioid and benzodiazepine prescriptions had a 15-fold increase in the risk of death compared to those filling neither prescription.¹³ One alarming study from Canada reported that, in a cohort of chronic pain patients who died of drug overdose with opioids, 85% were co-prescribed a benzodiazepine that was detected in body fluids at death.¹¹ Similarly, in a population of VA patients being treated with opioids for acute, chronic or non-terminal cancer pain who died from an opioid overdose, 49% died while they had active prescriptions for both opioids and benzodiazepines, translating to a 3.86 odds ratio increase of death compared to those who were only prescribed opioids. When controlling for opioid dose (which independently and significantly contributes to overdose death rates), the authors found that benzodiazepine dose contributed to risk, with a three-fold increase in risk of death in those receiving the highest benzodiazepines doses (> 40 mg diazepam equivalents) compared to those receiving opioids with low dose benzodiazepines (>0–10 mg).¹⁴ Finally, national data collected between 2003 and 2009 identified oxycodone and alprazolam as the two prescription drugs with the greatest increases in associated death rates (i.e., 265% and 234% increases, respectively).¹⁵

Nature of the Pharmacological Interaction between Opioids and Benzodiazepines: Potential Mechanisms of Action

Because so few controlled studies have directly examined the nature of the interaction between opioids and benzodiazepines (or other sedatives for that matter), the available data do not clearly define the specific mechanisms of action whereby opioids and benzodiazepines, despite being safe when given alone, lead to untoward outcomes including death when taken in combination. As there are many different drugs within both the opioid and benzodiazepine classes, each with a unique structure, pharmacological action (i.e., intrinsic efficacy, potency, receptor binding characteristics) and metabolic pathway, it is likely that multiple factors are at play, both pharmacodynamic and pharmacokinetic, when the drugs are taken in combination. While it is impossible to identify a single mechanism of action that may lead to additive or synergistic actions, it is informative to review the available data and to appreciate the numerous possible mechanisms underlying these potential interactions (see Table 1).

Table 1.

Opioid and Benzodiazepine Interactions: Potential Mechanisms of Action in Overdose

Metabolic Drug-Drug Interactions

Altering parent drug concentrations

Altering active metabolite concentrations

Genetic differences in sensitivity

Changes in Drug Transport (e.g., p-glycoprotein)

Muscle relaxant effects in throat/airway

Exacerbation of sleep apnea

Loss of tolerance (e.g., interrupted therapy)

Benzodiazepines have a remarkable safety margin and have rarely been attributed as the sole reason for overdose death (e.g., see review by Gaudreault and colleagues¹⁶). Mu opioid agonists reliably reduce respiratory function, this effect is dose-dependent, and tolerance can develop to this effect in patients who take opioids on a chronic basis. Mu opioid receptors are found in high concentrations in the respiratory control center of the brain in the medulla. These chemoreceptors actively monitor O₂ and CO₂ concentrations circulating in blood and provide a feedback mechanism to modify respiratory function (frequency, depth) to maintain homeostatic oxygen concentrations. Mu opioid agonist binding essentially dampens or dulls the sensitivity of these chemoreceptors to CO₂ reducing the responsiveness of the system. Thus, low doses of an opioid given to an opioid-naïve individual may lead to a modest reduction in respiratory markers (respiratory rate, tidal volume) but have no clinically significant effect. In contrast, supratherapeutic doses given to an opioid-naïve individual may lead to respiratory failure, which is typically the assigned cause of death in opioid overdose. The physiology of respiration is obviously more complicated than the simple system described here, and the reader is referred to a thorough review by White and Irvine¹⁷ for further discussion.

Biological Interactions: Enzymes, Transporters and Channels

One obvious potential mechanism of additive or synergistic action is through a simple pharmacokinetic interaction whereby the presence of one drug increases the concentration of a second agent. We typically think of drug-drug interactions occurring at the point of drug metabolism (usually in the liver) with at least two scenarios that may lead to enhanced drug concentrations and action. The first example would be where Drug 1 acts as an inhibitor of the enzyme responsible for metabolism of Drug 2; thus, in the presence of Drug 1, the metabolism of Drug 2 is decreased or blocked and concentrations of Drug 2 are higher than expected. This may be especially problematic when drugs with a long-half life are used (e.g., methadone) and the introduction of an enzyme inhibitor or enzyme competitor may lead to ever-increasing concentrations of Drug 1. The second example would be where Drug 1 is an inducer of the enzyme responsible for the metabolism of Drug 2 and Drug 2 exerts a significant amount of its activity through a secondary active metabolite. In that instance, the introduction of the enzyme inducer leads to greater concentrations of the active metabolite than expected. As opioids often share structural similarities (with a few exceptions, which are usually the synthetic opioids), many opioids are substrates for the same or overlapping P450 enzymes. This is also true for the benzodiazepines, whereby individual agents share structural similarities within class (and also often share the same long-acting metabolites). In addition, metabolism of several opioids and benzodiazepines leads to the formation of active metabolites that may exert pharmacodynamic effects. It is also important to recognize that, while understanding drug-drug-interactions through pharmacokinetic studies of drug concentrations in plasma, concentrations at the site of action (e.g., brain) may be quite different and be a reflection of local metabolism in the brain rather than liver (e.g., see review by Pai et al.¹⁸). However, to understand the potential risk for drug-drug metabolic interactions, it is essential to at least know the specific metabolic pathways for each of the prescribed agents. Table 2 provides an overview (but not an exhaustive list) of commonly prescribed opioids and benzodiazepines along with their known hepatic metabolic pathways and also denotes whether their metabolism leads to known active metabolites.

Table 2.

Drug Metabolism Mediated by Cytochrome P450 (CYPs)

Opioid Analgesics	Benzodiazepines
Oxycodone 3A4, 2D6	Alprazolam 3A4, 3A5
oxymorphone (2D6)	α-hydroxyalprazolam (3A4)
	4-hydroxy alprazolam (3A5)
Hydromorphone 3A4, 2D6
	Midazolam 3A4, 3A5, 3A3
Morphine 3A4, 2C8	α-hydroxymidazolam (3A4)
Codeine 3A4, 2D6	Triazolam 3A4
morphine (2D6)
Tramadol 3A4, 2D6	Diazepam 3A4, 2C19
o-desmethyltramadol (2D6)	desmethyldiazepam (2C19)
Methadone 3A4, 2B6, 2D6, 1A2	Clonazepam 3A4, 2C19
Fentanyl 3A4
Hydrocodone 3A4, 2D6
hydromorphone (2D6)
Buprenorphine 3A4, 2C8
norbuprenorphine (3A4)

Commonly prescribed opioid and benzodiazepine medications are listed along with the P450 CYP enzymes involved in hepatic metabolism of each compound. Active metabolites are listed in italicized text underneath the parent drug (when applicable). P450 CYP3A4 is highlighted in blue throughout the table to highlight this common metabolic pathway. This table was adapted from several comprehensive sources.^37–42

Another pharmacokinetic mechanism that can produce clinically relevant drug-drug interactions is through the superfamily of ATP-binding cassette (ABC) protein transporters of which P-glycoprotein (P-gp) is probably the best characterized. P-gp transporters play a critical role in drug efflux and are found in tissues that act as biological barriers, including the blood brain barrier and epithelial cells of the liver, kidney and intestine (see Silva et al.¹⁹ for a recent interesting review). Thus, P-gp plays a critical role in controlling the movement and concentrations of drugs both by limiting their absorption and by limiting their transport out of the brain. There are at least two compelling illustrative examples of clinically relevant P-gp effects on opioid responses. Loperamide, an opioid used for the treatment of diarrhea, is a substrate for P-gp and is, therefore, largely prevented from crossing the blood brain barrier (rendering it essentially a peripherally-acting opioid). In a study by Sadeque and colleagues,²⁰ loperamide was administered to healthy adults alone and produced no respiratory depression (as expected given the central mediation of opioid-induced respiratory depression). However, when given in combination with quinidine, a substrate known to inhibit P-gp transport, loperamide produced significant respiratory depression, presumably by quinidine-inhibition of P-gp allowing for loperamide entry into the brain. The second illustrative study focused on buprenorphine, a partial opioid agonist that produces a relatively flat dose response function even at very high doses;²¹ buprenorphine is recognized for its wide therapeutic window and low risk of opioid overdose when given alone. Buprenorphine also has an active metabolite, norbuprenorphine, which may have limited functional activity in humans with therapeutic dosing but has been demonstrated in preclinical studies to have greater potency as a respiratory depressant compared to the parent drug. In a study of mice, it was demonstrated that buprenorphine administration produced limited change in respiratory function when given alone. However, when given in the presence of a P-gp inhibitor or to P-gp knock-out mice (i.e., those without the transporter), respiratory depression (in some cases fatal) was observed.²² The authors demonstrated that inhibition of P-gp precluded norbuprenorphine, the more potent respiratory depressant, from exiting the brain as evidenced by substantially higher concentrations of norbuprenorphine in the brain compartment after pretreatment with the P-gp inhibitor compared to placebo. These studies demonstrate clearly that changes in drug transport may be a source of significant drug-drug interactions, but one that may be more difficult to predict in patients.

A third relevant site for potentially risky drug-drug interactions is the hERG channel, (human ether-a-go-go-related gene), a potassium ion channel, known to play an important role in QTc prolongation and risk of the rare but often fatal Torsades de Pointes. QTc prolongation has received extensive attention in pharmaceutical development as many drugs have been shown to increase the QRS complex; thus, development plans for new drugs commonly require cardiac studies. With regard to opioids, methadone is probably the most studied and widely reported opioid with documented effects on QTc prolongation²³, which led to a black box warning related to QTc prolongation and serious arrhythmias. Moreover, this has led several professional societies to develop new clinical guidelines for the use of methadone and cardiac monitoring,^23,24 although this is not without some controversy (see Krantz et al.²⁵ and related subsequent published commentary). Importantly, some studies have suggested that the QTc interval prolongation observed with methadone may be prolonged further in the presence of benzodiazepines (e.g., see Peles et al.²⁶). Thus, awareness of cardiac risk factors and ECG evaluation prior to initiating treatment or prior to adding other QTc prolonging drugs should be considered as a potential preemptive strategy for reducing patient risks.

Behavioral Risks: Another Pharmacodynamic Concern

Most opioids used for pain management produce their behavioral and pharmacological effects primarily through μ-opioid receptor activation, producing dose-dependent sedation, respiratory depression, and analgesia. Benzodiazepines, which are positive allosteric modulators of GABAα receptors (ligand-gated chloride channels), produce cognitive and motoric impairment, anxiolytic, muscle relaxant, sleep-inducing and anti-epileptic effects.

Each drug, administered alone, produces dose-dependent sedative effects; however, dose combinations, co-prescribed or co-abused, produce enhanced, often synergistic effects of psychomotor impairment, declines in cognitive abilities, increases in sedation and decreases in arousability. Several controlled studies have examined the behavioral and other interactions of these two drug classes in healthy research participants (e.g., no histories of COPD, heart disease). For example, one study²⁷ explored the pharmacodynamic effects of a single high dose of diazepam ([40 mg], a dose 20 times higher than the clinical starting dose) in participants maintained on either methadone or buprenorphine for opioid use disorder. Placebo or diazepam challenges were tested with the normal maintenance dose and 150% of the opioid maintenance dose during each of four sessions. The results indicated that diazepam produced increases in participant ratings of sedation and produced impairment on laboratory measures of psychomotor function, including reaction time and speed and accuracy tests.²⁷ These effects were largely independent of opioid dose, with high-dose diazepam impairment occurring in both methadone- and buprenorphine-maintained subjects. However, these dose combination effects do not appear to be driven by pharmacokinetic interactions. Two controlled studies have examined acute²⁸ and repeated dose combinations²⁹ of methadone and diazepam and both studies reported no changes in the plasma levels of the parent drugs or metabolites or their time course-concentration profile.²⁸ Similarly, post-mortem blood samples taken from individuals with fatal overdose have yielded the same findings.^{30, 31}

These studies have implications not only for opioid-benzodiazepine overdose, but are also a concern for increased risk of accidents from daily activities, including risk of falls, workplace accidents, driving ability and motor vehicle accidents, engaging in sports and outdoor activities, and caring for children, particularly in patients who have limited experience with the impairing effects of these medications.

Recommendations

There are rational and clinically judicious reasons why these agents are co-prescribed. Patients with chronic pain have higher incidence rates of depression, anxiety and difficulties with sleep, which may explain why opioids and benzodiazepines are so commonly co-prescribed in this population. However, some of the conditions that warrant treatment with benzodiazepines (e.g., anxiety disorders) are also independent risk factors for opioid overdose.³² Thus, it should be with caution and thoughtfulness that the decision is made to start one in the presence of the other. Starting doses should be conservative, and any dose increase should be accompanied by extra monitoring (even if by phone) to assess safety. Physicians may consider alternative agents for sleep (e.g., zolpidem or trazodone) but should be aware that most sedatives, regardless of class, have the potential to potentiate opioid sedation, and that few controlled data exist to inform their preferential use and safety. Physicians should be familiar with the risk of drug-drug interactions for the agents they prescribe and should be counseling their patients accordingly so that they and their family members are aware of the signs of potential opioid toxicity (e.g., “nodding” off or oversedation, unusually loud snoring, behavioral impairment). Patients should be monitored for misuse (i.e., taking more than prescribed, taking by a different route of administration), and calls for early refills should raise a red flag. Naloxone overdose kit distribution and overdose prevention training should be considered, as this tool becomes more readily available for prescribing physicians.³³ While the efficacy of opioids and benzodiazepines for the treatment of their respective conditions will likely encourage their continued widespread clinical use, careful thought should always be given to prescribing them in combination. Some clinicians recommend never starting one in the presence of the other, given the difficulty of weaning should adverse effects arise. Vigilance is warranted if co-prescribing opioids and benzodiazepines, and a cautious and integrated approach should be taken whenever combined therapy is needed for the treatment of co-morbid disorders.^34–36

Inside the Black Box: Identifying Patients at Risk.

• Does your patient have a history of alcohol use?

• Does your patient have a history of drug abuse or misuse?

  • If yes - have you completed a urine drug screen?

• Have you tried non-opioid analgesics?

• Have you considered non-benzodiazepine medications for anxiety and/or sleep?

• Have you read the package insert for each prescribed drug?

• Do you always start at the lowest dose (e.g., start low and go slow)?

• Have you accessed the patient’s record in a prescription monitoring database?

  • If yes - are any other providers prescribing opioids, benzodiazepines or contraindicated medications?

• Have you counseled your patient about the risks?