Benzodiazepines: A Perspective

A week after the mandate for all to shelter at home in Massachusetts, I received this message from a severely ill disease-phobic patient (who was working remotely with her behavioral medicine psychologist to reduce her disabling fears): “I am very thankful I have lorazepam at home with me these days! I am trying hard not to take it and am putting forth great effort on what I am thankful for each day.”

I confess that some of the most gratifying memories in my career had to do with benzodiazepines. In the early 1980s, there was a woman with agoraphobia in Beacon Hill who, after being given alprazolam and the support of a visiting research assistant, declared she had walked on grass for the first time in many years (having missed her son’s wedding, among other events over that time.) There was the physician with a history of panic attacks, ongoing social anxiety, and irritable bowel syndrome who achieved remission for decades on clonazepam and as-needed alprazolam for emergencies, eventually tapering to a low dose of clonazepam at bedtime. At his 60th college reunion, he encountered a classmate he had once admired and, knowing she had also held affection for him, he later mused to me that “if I had met you back then, I wonder how my life might have unfolded differently.” There was the college freshman who had onset of panic attacks in the first days of school, with family then preparing to fly him back home, who returned to class, rather than home, protected by a benzodiazepine. Admittedly, these anecdotes derive from a time before much of what we have come to know about the efficacy of cognitive and behavioral therapies and before we talked of concepts like anxiety sensitivity and exposure to bodily symptoms.

The anxiety about benzodiazepine prescribing is an old story (1). The efficacy of benzodiazepines in anxiety is well established, but based on the current practice of many more recently trained physicians, these medications reduce anxiety for the physicians themselves by their refusing to prescribe them. My own son, a first-year resident in psychiatry, looks at me as if I served on the wrong side in the Spanish Civil War when I speak of benzodiazepines.

During the early years of my career, I studied mainly higher-potency benzodiazepines in a variety of conditions, but principally in panic disorder, and I helped lead pivotal studies for U.S. Food and Drug Administration approval of clonazepam for panic disorder (2). I perceived that popular opinion and research funding agencies trivialized anxiety distress, even as we have come to appreciate the critical role of anxiety and insomnia in risk for suicidal behavior (3). My work turned to focus on the antecedents of anxiety and the potential for early recognition and interventions to limit lifelong disability (4). As others have also reported, roughly one in six individuals appear to be born predisposed to develop uncomfortable physiological and behavioral responses that predispose to fearful behaviors, and in some cases this predisposition evolves into recurrent or lifelong social anxiety and other mood and anxiety disorders (5). Behavioral interventions help many young individuals (6). For others, as adults, when given an adequate prescription of a benzodiazepine in a regimen avoiding interdose rebound, their symptoms will respond well, and they will ask if this is how normal people feel. Indeed, patients with panic disorder may be subsensitive to benzodiazepines and require higher doses than those without severe anxiety (7).

This commentary is not meant to be a call for a benzodiazepine renaissance but rather an attempt to offer a perspective. Beyond their established efficacy in anxiety distress and insomnia and fueling the debate between “pharmacological Calvinism and psychotropic hedonism” (8), these medications can also offer transient relief and comfort from stress; in a world replete with distress, it may be difficult for people to refrain from seeking a comforting remedy. Thus, these medications’ potential for nonmedical use, and their associated risks and side effects, is a concern but is not of itself a reason always to deny comfort and relief. As the aphorism guides us: “To cure sometimes, to relieve often, and to comfort always.” There may be times when the best or only way to comfort will be to use or add a benzodiazepine.

That said, if one can do as well or better with agents or methods less fraught with concerns, it would be a most reasonable decision. Now, 40 years from my first perspective publication on benzodiazepines (1), I work at a center for the treatment of anxiety disorders, the majority of which is staffed by brilliant and highly skilled psychologists trained in cognitive-behavioral therapy (CBT). In 1976, our department hired its first behaviorist, and now there are hundreds on our staff. I have seen that light and am passionate that this is the approach of first choice. But still, there are far too few of these skilled experts to meet the need in this country, and, even at our center, the wait for treatment can be long. There is evidence that other approaches (9), perhaps any caring and thoughtful psychotherapy, can be helpful whatever the underlying theoretical model. But even with the most established evidence-based psychotherapy, some patients’ symptoms do not respond to treatment, or they suffer residual symptoms.

Some benzodiazepine prescribing is straightforward, for example, as a brief intervention for acute distress or as-needed use for a phobic anxiety (e.g., airplanes) or transient insomnia. Some prescribing situations are, however, to be avoided if possible, like prescribing to manage persisting distress resulting from a personality disorder or for patients with known current or past substance use disorders. But prescribing is never carefree. An important guideline is to avoid chronic administration for acute problems and to set a goal for those on maintenance therapy of gradually working to find the lowest effective dose, which over time might become less or none, especially in older patients, in whom increasing sensitivity to the medication, the likely presence of more drugs interacting, memory concerns, and fall risk are important clinical issues to be assessed. This approach respects both patients’ symptoms and the physiological dependence that results from sustained use. Past pharmacoepidemiological studies (10) have also indicated that individuals chronically using benzodiazepines, compared with others, were more ill, had higher levels of psychic distress, and generally met criteria for disorder. The majority of this population had discussed their treatment with a physician in the previous 4 months. Although many patients on chronic treatment can taper off, assessment at follow-up of these patients indicates that ongoing treatment was for chronic or recurrent anxiety and not just for physiological dependence (11). Rates of successful taper can be enhanced by CBT approaches (12).

The question of the risk of prescribing a benzodiazepine and inadvertently inducing a substance use disorder is a complicated issue. The older literature on the addiction propensity of benzodiazepines seemed reassuring (13) in suggesting that these drugs were not strong reinforcers and were less likely than others to be the preferred drug of misuse. The voice of the substance use disorder community strongly counters this notion, citing the high rates of benzodiazepine misuse, especially in the substance-using population, and, recently and most alarmingly, given the adverse synergistic risk with overdose, compounding the opioid use epidemic. Although benzodiazepine overdoses are of a lower lethality compared with other drugs ingested in overdose (14), given their availability, they are frequently used in overdose and increase risk when combined with agents with a narrower therapeutic margin and greater lethality in overdose. Thus, responsible prescribing, beyond considering alternatives, requires knowing the patient and monitoring drug consumption. Concern about one population, however, need not proscribe the use of benzodiazepines in other populations who might benefit with a chance to live their lives with greater freedom from anxiety.

In 1984, Carl Salzman (then the chair of the APA Task Force on Benzodiazepine Prescribing) and I debated, at an American College of Psychiatrists conference, the issue of whether benzodiazepines were overprescribed. Then, as now, one of the key issues was alternatives. A remark I made about azapirones as alternatives (e.g., buspirone) went viral: “Everything you want in an anxiolytic except efficacy.” Maybe that was not fair, but it was a debate, and inasmuch as for the next four decades I continually encountered that quote (even from those who did not know its origins), the concern must have resonated with colleagues.

These days, we have more than tricyclic antidepressants, monoamine oxidase inhibitors, and buspirone as treatment alternatives. Other GABA-ergic drugs, including gabapentin and tiagabine, are viewed by some as safer to prescribe, although they are not side-effect free and are of uncertain efficacy relative to the much better studied benzodiazepines in treating anxiety. New GABA-ergic drugs are being developed for mood disorders, and it is not yet clear how they compare with benzodiazepines, which also reduce many symptoms that occur with mood disorders. In an effort to avoid prescribing benzodiazepines, many clinicians default to low doses of second-generation antipsychotics such as quetiapine. Antipsychotics had long been used acutely for their ataractic effects, but ongoing use was less endorsed because of concerns about tardive dyskinesia, metabolic effects, and even increased risk of sudden death from different mechanisms.

I do favor antidepressants, particularly selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs), over benzodiazepines for treatment of persisting anxiety disorders. The “standard advice” for these classes is interestingly opposite to that for benzodiazepines. In the case of antidepressants, we advocate ensuring adequate dose and duration and maintaining treatment for the long term or indefinitely for persisting or recurrent disorders, while for benzodiazepines, we insist on keeping doses low and discontinuing early. With both classes, we encounter distressing symptoms with abrupt withdrawal that can obscure the difference between discontinuation effects and relapse (15). Early nonadherence, however, often follows prescribing of SSRIs or SNRIs in part because of induction of anxiety or agitation, although benzodiazepines rarely worsen or induce those symptoms. Indeed, benzodiazepines are efficacious in helping some patients with anxiety tolerate SSRI initiation (16). With both classes of medication, some patients’ symptoms respond very well and some less well, and response may fade for some over time. Although there are no known irreversible effects from long-term administration, for both SSRIs and benzodiazepines there is a small percentage of those treated who attribute persisting postdiscontinuation symptoms to treatment. The main controversy around prescribing of these medications centers on the issues of abuse liability, cognitive impairment, and physical dependence. As with most of our therapeutics, we lack long-term studies comparing outcomes of patients on benzodiazepines with those on alternative pharmacologic, behavioral, or even no treatment. The use of “big data” and new tools, such as artificial intelligence and machine learning for studying longitudinal electronic health record data, should give us insight into the true risks and benefits of benzodiazepine use.

As with the article in this issue of the Journal by Osler and Jørgensen (17), what we learn with well-designed studies may run counter to long-held beliefs. In this case, the notion that benzodiazepines contributed to cognitive decline was not proven when outcomes were accounted for use by indication. In fact, the results of the study suggest that cumulative use might even be neuroprotective. This finding is plausible especially in light of recent findings that depression, often associated with anxiety and managed in part with benzodiazepines, is a risk for or precursor of dementia (18), that insomnia risks preventing the brain’s glymphatic system from clearing β-amyloid (19), and that stress over time increases risk for neurodegenerative diseases (20).

To close with a banality, benzodiazepines are medications. As such, they are intended to be prescribed by those trained in how and when to use them and in how to weigh risks and benefits. They are also intended to be prescribed to address suffering, distress, and functional loss from psychiatric symptoms and disorders. As with many tools available to clinicians, they are imperfect but are potentially beneficial.