Position paper on the use of benzodiazepines in psychiatry

The marketing material widely available when benzodiazepines (BZDs) were first introduced in 1960 (chlordiazepoxide) and 1965 (diazepam) emphasised that BZDs were not addictive, did not cause respiratory depression and so were safe in overdose. They quickly replaced barbiturates on the basis of this advertising. Comparisons of per capita prescribing suggest that BZDs are more commonly prescribed in Ireland than in the UK. There may be cultural reasons for this that are specific to Ireland, though the greater influence in Ireland of pharmaceutical advertising and sponsorship may have been relevant also. 

Various authorities, including NICE, Cochrane and the Maudsley Handbook, give conflicting advice about the use of BZDs, particularly in acute inpatient settings. In general, European/UK authorities are more critical of the use of BZDs,^1-3 while US authorities, which often rely on uncontrolled and industry-sponsored research, are more agnostic. Accordingly, this short position paper is based on clinical experience, pharmacological principles and recognition of the need to take account of international best practice rather than relying on custom and practice in Ireland.   

Pharmacology

BZDs act as agonists or modulators at γ-aminobutyric acid (GABA) receptors. GABA receptors are inhibitory, causing hyper-polarisation. BZDs have essentially the same mechanism of action as alcohol at GABA receptors. There is cross-tolerance between BZDs and alcohol and cross-addiction is common. Recently introduced ‘z-drugs’ are sometimes called ‘non-benzodiazepines’ but their mode of action is the same as BZDs and alcohol, with all the same problems to some degree. Advertised claims of advantages for the ‘z-drugs’ over BZDs should be treated with scepticism.

Although many different BZDs are licensed, and some BZDs are more active at some GABA receptor sub-types than others, at more than minimal doses the differences in receptor affinities are marginal. BZDs have a wide range of half-lives, from the very short to the extremely long, so that some lead to rapid swings in affective stability while others accumulate and cause persistent sedation. The different uses of BZDs in practice, some for alcohol detoxification, some for sleep, some for tranquillisation or anxiety, has more to do with advertising and marketing than receptor specificity. The prescribing of more than one BZD or related ‘z-drug’ is always pointless.

Licensing

Almost all BZDs are licensed only for short-term use. Anything other than short-term use is usually prescribing off-licence and therefore has to be justified as an exception. In practice, however, BZDs are commonly used off-licence as long-term prescriptions in primary care and in psychiatry.

Tolerance and withdrawal

Like alcohol, all BZDs lead to the development of tolerance and all give rise to a withdrawal syndrome.⁴ The withdrawal syndrome can be characterised as an exaggeration or rebound of symptoms for which the BZD was prescribed with severe anxiety (panic, derealisation), particularly in those for whom anxiety was the indication, worsened sleep initiation and maintenance for those in whom sleep problems were the indication, etc. This is because the pharmacological phenomenon of tolerance arises from a variety of molecular mechanisms to compensate for the enhanced neuronal inhibition. These include compensatory up-regulation of excitatory glutamate pathways to counter-balance the increased GABA inhibition. When the BZD (or alcohol, barbiturate or ‘z-drug’) is withdrawn, the glutamate pathway is unopposed so that when the BZD clears there is ‘rebound’ of the symptoms normally inhibited by GABA. Breakthrough symptoms can emerge even during continued steady state prescribing of BZDs, as tolerance develops due to down-regulation of GABA receptor numbers and receptor sensitivity. Those BZDs with the shortest half-lives are most likely to give rise to withdrawal symptoms, sometimes with a daily cycle of mood swings. Increasing the dose in response to these effects is a common error, reinforcing addiction when gradual detoxification with the longest-acting BZDs is the correct course of action. 

Benzodiazepines and the therapeutic relationship

The urge to prescribe ‘anything’ should be resisted over the urge to rely instead on a more conservative approach with explanation and practical recovery-oriented advice. In Ireland, there is a widespread folk belief that people ‘need something for their nerves’. This may explain the apparent direct transition from alcohol dependence to BZD dependence in many cases. Because BZDs have a strong potential for dependence, doctors often come under pressure from patients and their relatives to prescribe BZDs, eg. for grief, for back pain, for sleep and for ‘nerves’.  

Doctors working in addiction clinics and prison settings (where substance misuse problems occur in 65-70% of inmates) come under particular pressure to prescribe BZDs. Experience shows that much of the medication prescribed in these settings is diverted, sold on to others in exchange for other drugs, endangering the unintended end users. In prisons, inmates may be coerced to seek BZDs, which they are then forced to give up to others. The most common cause of death in prison now is accidental overdose, typically of smuggled opiates combined with diverted BZDs. Deaths due to opiate overdose in the community are due to cocktails of opiates with alcohol or BZDs.

Therapeutic relationships forged around the prescribing of BZDs are seldom genuinely recovery-orientated and the prescriber may be unwittingly colluding in denial and demotivation or may simply be duped. Therapeutic relationships are more effective when prescribing BZDs and other sedatives (other than detox courses where there is clear evidence of withdrawal) is ruled out at the outset, as part of the preliminary discussion with the new patient about the nature of the consultation and the goals of treatment. 

Respiratory suppression

Use of BZDs can exacerbate or give rise to sleep apnoea, either because of central depressant/inhibitory effects, or muscle relaxation of the pharynx, or both. Giving BZDs by bolus intravenous (IV) injection can cause respiratory arrest. The IV antagonist flumazenil should always be available when BZDs are given intravenously but because flumazenil has a shorter duration of action than most BZDs, patients may wake up then lapse back into an obtunded state (so-called ‘re-sedation’) unless vigilance is sustained.

Disinhibition (so-called paradoxical reactions)

GABA mediates frontal lobe inhibition. Alcohol and benzodiazepines lead to a disinhibition which increases the risk of impulsive behaviour including violence⁵ and self-harm.⁶ BZDs, like alcohol, are associated with suicide, possibly as unforeseen disinhibition, or as Dutch courage.⁷ A large population-based study of prescriptions in Canada showed that those prescribed BZDs had an odds ratio (OR) for attempted suicide of 6.2, while a past history of drugs and alcohol problems had an OR of 13.4. By contrast, the use of antidepressants was associated with attempted self-harm only in association with past drugs and alcohol problems (OR = 5.8).⁸

In psychiatric settings, particularly in intensive care settings, there is an increased risk of violence and self-harm due to this direct effect. There is also an indirect effect caused by alcohol withdrawal and BZD withdrawal giving rise to dysphoria, irritability and anger. While these tendencies towards self-harm and violence are sometimes described as ‘paradoxical reactions’ and are said to be more common in those with personality disorders and comorbid substance misuse problems, it is more accurate to regard these not as paradoxical but as the expected effects of BZD and/or alcohol intoxication and withdrawal in those with established tolerance and addiction.⁹

The dependence syndrome can also lead to a cycle of dysfunctional behaviour when such behaviour leads to injections or increased doses of BZDs. Patients with a history of alcohol or BZD misuse will regard this as rewarding and reinforcing the problem behaviour. As a general rule, the use of BZDs for aggression, irritability or excitement leads to a short-term relief but is followed by a rebound exacerbation and may instil a pattern of violence for drug-seeking rewards. The overall ‘cost-benefit’ is negative.  

Cognitive impairment

BZDs impair short-term memory and also impair judgement associated with impulsivity. There are two important adverse effects of this. The first is that all BZDs in their licensing information carry warnings against driving or using heavy machinery or other risk-responsible activity. There is evidence that many road traffic accidents and fatalities are associated with use of BZDs.^10,11   

The second is that, because of the interference with memory and the blunting of affective reactions, the ability to learn from experience is greatly impaired by the use of alcohol and BZDs. Isaac Marks taught that the cognitive behavioural treatment of phobias could not work in patients treated with BZDs or who continued to misuse any other intoxicant.¹² This was because memory impairment prevented the retention of new information and because exposure therapy requires the patient to experience and habituate to somatic anxiety and thereby gain confidence and a sense of mastery over the fear. The same is not necessarily true of antidepressants, where the patient is also depressed.

Mental illness and comorbid addiction

It is now well established that depression and anxiety, bipolar affective disorder¹³ and schizophrenia¹⁴ are all strongly associated with increased rates of dependence on alcohol and other substances. There may also be a causal link in the reverse direction, with alcohol causing affective illnesses.¹⁵ There is therefore a substantially increased vulnerability to addiction for BZDs also.   

Indications for benzodiazepine treatment (licensed and off-licence)

Not all of the indications listed here are licensed. For those that are licensed, the overall cost-benefit may not be advantageous.

Benzodiazepines as antiepileptics

BZD use can be lifesaving in status epilepticus because BZDs cause rapid increase in the seizure threshold in those who are BZD-naive. However, the evidence that BZDs cause a sustained increase in the seizure threshold is weak and largely at odds with the evidence for tolerance. BZD withdrawal, however, like alcohol withdrawal, is very likely to lower the seizure threshold increasing the risk of seizures including status epilepticus which can be life-threatening. Seizures due to abrupt alcohol or BZD withdrawal will be stopped by giving a BZD. However, BZD tolerance will make BZDs less effective as an emergency treatment for status epilepticus due to other causes. This should be regarded as a reason for avoiding the use of BZDs in maintenance treatment of epilepsy.

Benzodiazepines for alcohol or benzodiazepine withdrawal

When prescribing a course of BZDs for withdrawal/detox, it is important to explain to the patient that the BZD course will help to alleviate the physical effects of withdrawal, which is unlikely to take more than a few weeks and in most cases can be accomplished in less than two weeks (as for alcohol dependence). 

The psychological craving may continue for much longer (as for alcohol) but is only prolonged by very prolonged tapering courses (courses over months) which seldom succeed since the correct approach to psychological dependence is psychological (education, motivation, cycle of change etc) not pharmacological treatment. 

Benzodiazepines and acute psychosis

Neuroleptic malignant syndrome is sometimes managed by using BZDs as tranquillisers, as is the similar acute excited state with hyperpyrexia. Catatonia may be treated with BZDs, though there is no clear evidence for this.¹⁶

BZD withdrawal can itself cause an acute psychosis, which is poorly responsive to antipsychotics.¹⁷ This is a reason for avoiding the use of BZDs, particularly longer-term BZDs, in psychoses generally. 

The psychoses, like the affective disorders and anxiety disorders, carry with them an enhanced vulnerability to substance misuse and dependence. Persons with any of the psychoses (or any other mental illness) are therefore more likely to develop dependence, withdrawal and rebound and less likely to benefit from the use of BZDs when a cost-benefit analysis is considered. 

Acute psychosis (other than BZD withdrawal states) is best treated with antipsychotics and/or mood stabilisers. Modern antipsychotics carry less extrapyramidal side-effects than the older antipsychotics, cause less akathisia and have greater inherent mood-stabilising effects (particularly olanzapine). Even the most severe acute psychosis with irritability and anger can be successfully managed without the use of BZDs. BZDs and other tranquillisers should not be used as a substitute for adequate nurse-to-patient ratios and special nursing observations. 

Schizophrenia and other psychoses are associated with violence and homicide, with most of the risk attributable to substance misuse and dependence,^18,19 including BZDs.

Benzodiazepines and akathisia

BZDs are sometimes prescribed for akathisia. Akathisia is more often a consequence of the use of older antipsychotics in higher doses and is much less common when using more modern antipsychotics in areas with higher nurse-to-patient ratios. As with all use of BZDs, the risks of tolerance, withdrawal and rebound exacerbation with irritability mean that the overall cost-benefit from use of BZDs for this indication is negative.  

Benzodiazepines for sleep disturbance

BZDs cause an abnormal sleep architecture not equivalent to normal sleep. Tolerance develops rapidly. BZDs invariably lead to a deterioration in sleep hygiene and rebound disturbance of sleep patterns. Sleep is further worsened by sleep apnoea due to over-sedation, respiratory depression and/or relaxation of the pharynx. Because of the high propensity for dependence, the overall cost-benefit balance is negative. There is no indication for the use of BZDs for sleep, even in the short term.  

Benzodiazepines for anxiety disorder including panic and phobia

BZDs cause short-term relief of symptoms during the first weeks of use only. Even during the first weeks, rebound and withdrawal are frequently seen when short-acting or medium-acting BZDs are used. As tolerance develops, only the adverse effects remain. This can be mistaken for affective instability/borderline personality disorder. As outlined above, the use of BZDs can undermine the benefits of curative modern psychotherapies of all sorts, including exposure therapy and cognitive behavioural therapy (CBT).¹² Anxiety disorders, like all affective disorders and psychoses, increase the vulnerability to substance misuse and dependence, including BZD dependence. The overall cost-benefit analysis for the use of BZDs and their congeners in these disorders is negative.

Depression

There is no indication for the use of BZDs in depression and the risk that suicidal impulses may be disinhibited cannot be ignored.^6-8

Grief and nervous shock

The natural history of grief and the normal process of resolution may be arrested by the affective blunting and memory-impairing effects of BZDs. The use of BZDs for grief, ‘nervous shock’ and adjustment reactions may actually worsen the long-term prognosis for post-traumatic stress disorder (PTSD) by interfering with and disrupting the normal psychological processes of remembering, reasoning, coping and achieving an internal locus of control. The standard psychological treatment of exposure in imagination for PTSD would be disrupted by use of BZDs in the same way that exposure treatment for phobias is disrupted.

Benzodiazepines and affective instability/repetitive self-harm

Many patients who repetitively self-harm by cutting do so to relieve tension. In prison and personality-disordered populations this is strongly associated with BZD misuse or dependence, often with comorbid use of alcohol and other intoxicants. The tension complained of is related to ‘breakthrough’ tolerance and short-term withdrawal effects, particularly when short-acting BZDs are used. Full detoxification from BZDs and related drugs enables the person to desist from repetitive cutting. It is worth noting that many of the features of affective instability/borderline personality disorder, including mood swings (tearful, angry, anxious, tense, fatigued) and repetitive self-harm, are exacerbated by, or in some cases caused by, BZD dependence and misuse, often along with alcohol.

Benzodiazepines as muscle relaxants

BZDs are often prescribed by orthopaedic surgeons and GPs to relieve supposed muscle spasm in back pain. The evidence that back pain is caused by muscle spasm or relieved by muscle relaxants is poor. The potential for dependence, tolerance and rebound on withdrawal is very high. Baclofen and other specific muscle relaxants would be more appropriate but are less often used, probably because muscle relaxation is ineffective – the subjective ‘benefit’ of BZDs reported by many patients who complain of back pain is most likely related to the central actions of BZDs, leading to tolerance and addiction.

Benzodiazepines as a cause of anxiety and depression, mood swings and irritability

When BZD treatment is initiated for some other reason such as muscle spasm, the development of tolerance and withdrawal can give rise to rebound anxiety, panic and dysphoria on withdrawal. In those prone to psychosis, a withdrawal psychosis can occur. 

Summary

This is an intentionally didactic paper, because the model of review adopted by consensus panels tends to produce inconclusive recommendations. No benefit can be shown for the use of BZDs other than for rapid relief of status epilepticus, detoxification for the physical withdrawal symptoms of alcohol dependence (and BZD detoxification) and in rare cases of neuroleptic malignant syndrome. All other ‘indications’ do not satisfy a rational cost-benefit analysis and many are unfounded. 

A rational cost-benefit analysis for the various licensed and ‘off-label’ clinical indications for BZDs consistently finds strong short-term subjective benefits, but the concurrent development of side-effects, such as cognitive and motor impairments followed by tolerance with ‘breakthrough’ symptoms, physiological withdrawal and psychological dependence, all lead to negative cost-benefit balances. A simple but rigorous avoidance of prescribing BZDs from the outset is the only reliable way of eliminating these problems. Making this clear to newly referred patients enhances the therapeutic relationship. Public education programmes should support this modernisation of practice.

In our experience, eliminating the use of all BZDs in prisons and in hospital settings improves therapeutic engagement and motivation, eliminates drug-seeking from the therapeutic process and reduces violence and repetitive self-harm when combined with effective nursing and multidisciplinary care. Unlike many other psychiatric settings, our patients are not visibly over-sedated, slowed and slurred – they are alert, capable and interested in recovery. Since we have eliminated the prescribing of BZDs in our prison clinics, the suicide rate in prisons has fallen. In the Central Mental Hospital the most violent patients in the state are safely managed without any BZDs at all. We believe this is in part because of the elimination of BZDs. Much of the recent controversy regarding reported ‘paradoxical’ effects of antidepressants may be due to the concurrent use of alcohol or BZDs. No-one would recommend the use of alcohol with an antidepressant or in anyone with a depressive illness. Why prescribe a BZD to anyone with a mental illness?

Declaration of interest: none.

References 

1. Priest RG, Montgomery SA. Benzodiazepines and depression: a College statement. Psychiatric Bulletin 1988, 12: 107-109
2. Chief Medical Officer. Benzodiazepines warning. A communication to all doctors from the Chief Medical Officer. CMO’s Update 37. January 2004. Available at: www.benzo.org.uk/cmo.htm 
3. Committee on Safety of Medicines. Current problems (no. 21). Guidance on benzodiazepines. London (UK): Committee on Safety of Medicines and the Medicines and Healthcare Products Regulatory Agency; Jan. 1988
4. Schweizer E, Rickels K. Benzodiazepine dependence and withdrawal: a review of the syndrome and its clinical management. Acta Psychiatr Scand Suppl. 1998; 393: 95–101
5. Paton C. Benzodiazepines and disinhibition: a review. Psychiatr Bull. 2002; 26: 460–462
6. Neutel CI, Patten SB. Risk of suicide attempts after benzodiazepine and/or antidepressant use. Ann Epidemiol. 1997; 7(8): 568–574
7. Carlsten, A; Waern M, Holmgren P, Allebeck P (2003). “The role of benzodiazepines in elderly suicides”. Scand J Public Health 31 (3): 224–8
8. Neutel CI, Patten  SB. Risk of suicide attempts and benzodiazepine and / or antidepressant use.  Ann Epidemiol, 1997 Nov; 7(8): 568-74
9. Dietch JT, Jennings RK. Aggressive dyscontrol in patients treated with benzodiazepines. J Clin Psychiatry. 1998; 49: 184-188
10. Ericsson HR; Holmgren P, Jakobsson SW, Lafolie P, De Rees B (November 10, 1993). Benzodiazepine findings in autopsy material. A study shows interacting factors in fatal cases. Läkartidningen; 90 (45): 3954–7
11. Thomas RE. (1998) Benzodiazepine use and motor vehicle accidents. Systematic review of reported association, Can Fam Physician. 1998 Apr; 44: 799-808 
12. Marks IM. Living with fear: understanding and coping with anxiety. 2nd edition, 2001. McGraw Hill, Maidenhead: 162
13. Liand W, Chikritzhs T. Affective disorders, anxiety disorders and the risk of alcohol dependence and misuse. British Journal of Psychiatry 2011; 199: 219-224
14. Dixon L. Dual diagnosis of substance abuse in schizophrenia: prevalence and impact on outcomes. Schizophrenia Research 199, 35 Suppl 1 (8): 161-3
15. Gilman SE, Abraham HD. A longitudinal study of the order of onset of alcohol dependence and major depression. Drug Alcohol Depend 2001; 63: 277-286
16. Gibson RC, Walcott G. Benzodiazepines for catatonia in people with schizophrenia and other serious mental illnesses. Cochrane Database of Systematic Reviews 2008, Issue 4
17. Wolkowitz, O. M., Turetsy, N., Reus, V. I., et al (1992) Benzodiazepine augmentation of neuroleptics in treatment-resistant schizophrenia. Psychopharmacology Bulletin; 28: 291-295
18. Fazel S, Langstrom N, Grann M, Lichtenstein P. Schizophrenia, Substance Abuse and Violent Crime JAMA. 2009; 301(19): 2016-2023
19. Fazel S, Gulati G, Linsell L, Geddes JR, Grann M, 2009 Schizophrenia and Violence: Systematic Review and Meta-Analysis. PLoS Med; 6(8)