Anxiolytics and Hypnotics

Anxiolytics and hypnotics.

Topics covered:

  • Introduction
    • Definitions
  • Anxiolytic drugs
  • The benzodiazepines
    • Pharmacokinetics
    • Basic pharmacology
    • Clinical pharmacology
  • Hypnotic drugs
  • Clinical effects of anxiolytics
    • Unwanted effects
    • Overdosage
    • Tolerance and dependence
    • Management of withdrawal
  • Other anxiolytics
    • Benzodiazepine-receptor partial agonists
    • 5-HT1A partial agonists
    • Antipsychotic drugs
    • Antidepressants
    • b-Adrenoceptor antagonists
    • Antihistamines
  • Clinical effects of hypnotics
    • Residual effects Idiosyncratic effects
    • Rebound
    • Dependence
    • Abuse
  • Conclusions 
  • References


Anxiety is a ubiquitous emotion that becomes a clinical disorder when it is too severe, too protracted, or too pervasive for the subject to bear. Insomnia is a failure to experience satisfying sleep, together with a feeling of tiredness during the day. A wide variety of compounds—the anxiolytics and hypnotics—are used to treat these conditions, but much overlap exists between the two groups of drugs.

The classical antianxiety (anxiolytic) drugs are alcohol, the opioids, and the barbiturates. For the past 40 years, the benzodiazepines, such as diazepam and lorazepam, have dominated the field. They are effective anxiolytics in the short term but their long-term efficacy remains in dispute. Their drawbacks include cognitive and psychomotor impairment, paradoxical reactions, tolerance, and dependence, and they are major drugs of abuse. The search for more selective compounds with an improved risk–benefit ratio has been largely unsuccessful.

Other anxiolytics act on the 5-hydroxytryptamine (5-HT) or serotonin systems of the brain and include buspirone and the selective serotonin reuptake inhibitors (SSRIs). The parameters of their effectiveness and risks are still being established. b-Blockers and antihistamines are enjoying a renewed lease of life as the benzodiazepines become used in an increasingly restricted way.

The use of benzodiazepine and benzodiazepine-like hypnotics, by contrast, continues unabated. Some switching to the shorter-acting benzodiazepines has occurred, together with the introduction of zopiclone, zolpidem, and (imminently) zaleplon. The drugs tend to have few, if any, residual effects the next day, and may be less likely to induce rebound and dependence than equivalent benzodiazepines. Particular care is needed in prescribing such hypnotics to the elderly.

Other groups of compounds used as hypnotics include sedative antihistamines and some herbal remedies.

The rational use of both anxiolytics and hypnotics requires minimal dosage, short durations of use, and simultaneous exploitation of non-pharmacological methods.


‘Sedative' originally meant a substance that has the property of allaying anxiety. However, it has now come to denote feelings of drowsiness or torpor. This state was originally called ‘oversedation', and was often noted with the use of barbiturates and other older drugs such as chloral. Furthermore, the term ‘tranquillizer' was introduced 30 or more years ago in an attempt to distinguish between the older sedatives and the newer drugs, supposedly non-sedative, such as the benzodiazepines. But this distinction is artificial as, apart from safety in overdosage, the benzodiazepines closely resemble the barbiturates in pharmacological and clinical properties. The term ‘anxiolytic' is often used, but is of dubious etymology.

Anxiolytic drugs

Mankind has used anxiety-allaying drugs for thousands of years, dating back to the discovery that, among its psychotropic properties, alcohol could induce sedation. The nineteenth century saw the development of inorganic and, later, organic chemistry. Bromides were introduced as sedatives and became widely used despite their poor effectiveness, toxicity, and potential abuse. Organic chemists in the second half of the nineteenth century introduced sedatives such as chloral and paraldehyde. The former is still used as an hypnotic but paraldehyde is now obsolete.

The first barbiturate was introduced a hundred years ago. This group of drugs is divisible into the ultrashort-acting (e.g. anaesthetic-induction agents such as thiopentone and methohexitone), short-acting (e.g. quinalbarbitone), medium-acting (e.g. butobarbitone), and long-acting (e.g. phenobarbitone) barbiturates. Thiopentone owes its short action to its redistribution phase in the body after intravenous administration, but methohexitone is rapidly metabolized in the liver. Most of the rest are of medium duration with half-lives of 16 h or so. The disadvantages of the barbiturates include the side-effect of drowsiness, tolerance to their effects, dangers of overdose, and possible physical and psychological dependence with severe withdrawal syndromes. (1)

Next, meprobamate was developed from mephenesin, which was a muscle relaxant with its primary action being on the spinal cord. Meprobamate was introduced as the first of the ‘tranquillizers', but its advantages over the barbiturates proved minimal. It is still used as a sedative and muscle relaxant, usually in combination. Other non-barbiturate sedatives and hypnotics were developed, including glutethimide, methyprylon, ethchlorvynol, and the ill-fated thalidomide. Each enjoyed a brief vogue.

The benzodiazepines were first synthesized in the 1930s, but not developed until 2 years later. The prototype, chlordiazepoxide, had similar actions to meprobamate in cats but with several times the potency. Further evaluation followed in the clinic and it was soon introduced into medical practice. More than a 1000 benzodiazepines and related compounds have been synthesized, including diazepam, the most widely used of all. Anxiolytic and hypnotic, as well as muscle-relaxant and anticonvulsant properties, are licensed indications. However, the distinction between anxiolytic and hypnotic uses often seems to owe more to commercial expediency than to scientific rationale. Some compounds, such as lorazepam, are marketed for both indications. 

The benzodiazepines

The use of barbiturates as anxiolytics has declined to the point of obsolescence. The main reason for this is the safety in overdose of the benzodiazepines compared with the quite marked toxicity of the barbiturates. In turn, concern has mounted concerning the benzodiazepines. (2) These drugs are widely prescribed by many physicians for patients with emotional problems, circulatory disorders, tension headaches, and pains in the chest and back as well as digestive disorders, all with the common symptom of anxiety. This widespread use, even overuse (the ‘benzodiazepine bonanza'), and the induction of dependence even at normal therapeutic dose has led to official injunctions for greater care in prescribing. Nevertheless, the drug treatment of choice for anxiety, at least in the short term, remains a benzodiazepine. (3) However, an alternative position held by some practitioners is to view these drugs as useful, safe, and underused in the management of anxiety. (4)


The two aspects of the pharmacokinetics of the benzodiazepines relevant to the prescriber are the speed of onset of action and the duration of that action. The speed of onset depends on the mode of administration and the penetration time to the brain. Given by mouth, most benzodiazepines are rapidly absorbed and exert a prompt anxiolytic effect, for instance in panic states. Diazepam and lorazepam are prime examples. Although temazepam enters the brain more rapidly than, say, oxazepam, it still takes an appreciable time to induce sleep. The redistribution phase can be pronounced and will then largely determine the duration of effect of single doses of benzodiazepines such as diazepam and flunitrazepam.

The metabolic half-lives of the benzodiazepines also vary greatly. An important compound is N-desmethyldiazepam (nordiazepam), the major and active metabolite of diazepam and several other benzodiazepines. It has a long half-life, about 60 h, and accumulates over the first month of treatment. Metabolism of these drugs is even slower in the elderly and in patients with liver damage.

Benzodiazepines with a 3-hydroxyl grouping, such as lorazepam, oxazepam, and temazepam, have half-lives averaging 12 h or less. Liver damage has to be severe before the metabolism of these drugs is affected. Alprazolam is a triazolobenzodiazepine with a half-life of 9 to 16 h and with hydroxy metabolites of low biological activity. (5) Both chlordiazepoxide and diazepam are absorbed erratically after intramuscular injection. Lorazepam, however, is well absorbed after intramuscular injection.

Several studies have searched for any relationship between plasma benzodiazepine concentrations and clinical response without finding any clinically useful correlates.

Basic pharmacology

The benzodiazepines potentiate the widespread inhibitory neurotransmitter g-aminobutyric acid ( GABA). Benzodiazepines do not act directly on GABA receptors but have their own receptors. To date, the natural transmitter acting on benzodiazepine-binding receptors has still not been identified with certainty, although several candidates (‘endozepines') have been suggested. Barbiturates bind to separate receptors, as do neurosteroids.

Because of this widespread inhibitory effect, benzodiazepines alter the turnover of neurotransmitters such as noradrenaline and 5-HT. The main sites of action of the benzodiazepines are in the spinal cord where muscle-relaxant effects are mediated, the brainstem (perhaps accounting for their anticonvulsant properties), the cerebellum (causing ataxia), and the limbic and cortical areas involved in the organization of emotional experience and behaviour.

An antagonist agent at the benzodiazepine receptor is flumazenil. (6) This is commercially available for reversing the effects of benzodiazepine-induced anaesthesia. It can also be useful in the management of benzodiazepine overdoses. Finally, inverse agonists have the opposite effects to the benzodiazepines, being anxiogenic and proconvulsant.

Clinical pharmacology

The depressant effects of single therapeutic doses of a benzodiazepine can usually be readily detected. However, lower doses may fail to impair psychological functioning and subjective effects are usually absent. In the clinical context with anxious patients and with repeated higher doses, sustained impairment of functioning is also difficult to demonstrate. Some studies have shown decrements in performance after the first dose, but improvements in functioning, in comparison to predrug levels, may become apparent by the end of a week of repeated usage. (7) This suggests that the well-known impairment of performance produced by pathologically high levels of anxiety is first worsened by the sedative effects. Then as the antianxiety effects build up, the patient's psychological functions improve.

A second mechanism concerns tolerance, which reflects several biochemical mechanisms including alteration in benzodiazepine-receptor type. Patients who have a high alcohol intake are tolerant to benzodiazepines.

The benzodiazepines have marked and selective effects on memory (8) by interfering with episodic memory, that is to say the system concerned with remembering personal experiences. (9) This effect seems independent of any sedation or attentional impairment. (10) Alcohol adds to the cognitive impairment induced by the benzodiazepines but does not necessarily potentiate it. (11)

The dependence potential of benzodiazepines is seen in drug-preference studies, but these drugs are much less preferred than say the amphetamines. Differences among benzodiazepines have been documented; for example, oxazepam seems to have less abuse liability than diazepam. The largest gap in our knowledge of these drugs is on their long-term usage, which has been evaluated in relatively few studies. Thus, it is still largely unestablished whether therapeutic effects are maintained for longer than a few weeks and when dependence supervenes in the minority of patients who encounter problems on protracted usage.

Hypnotic drugs

The main groups of drugs used in the modern treatment of insomnia are the benzodiazepines, and the newer compounds, zopiclone, zolpidem and zaleplon. The pharmacology of these benzodiazepines is essentially the same as that of the anxiolytic compounds.

Nitrazepam is a long-acting benzodiazepine with an elimination half-life ranging between 25 and 35 h, but it is longer in the elderly. Because of this, it is likely to produce residual effects and to accumulate. Flunitrazepam is more potent, but somewhat shorter acting with a half-life of 10 to 20 h. It has a rapid redistribution phase, which can result in a short duration of intense action. Flurazepam is still widely used in the United States. It has a very long-acting metabolite, N-desalkylflurazepam, which accumulates and can produce psychological impairment on regular dosage, especially in the elderly. Quazepam is also a long-acting drug. Of the intermediate-acting compounds, temazepam has a half-life of 10 to 15 h, without active metabolites. At modest dose (10–15 mg daily), it results in few residual effects and is fairly well tolerated by the elderly. Major problems with abuse have limited its popularity, but it is still widely prescribed worldwide. Lormetazepam is slightly shorter acting, estazolam somewhat longer acting. Loprazolam has a fairly short half-life, but its absorption may be slow and erratic. (12)

Triazolam is the archetypal short-acting benzodiazepine, with a mean half-life of around 3 to 4 h, and no clinically significant metabolites. Daytime sedation is seen after high doses (0.5 mg daily), but not usually with lower ones. These higher doses have also been associated with an increased incidence of anterograde amnesia and unusual behaviours, including depressive reactions and hostility. Concerns over these reactions led to the removal of this drug from the United Kingdom market, and to a reappraisal of its use elsewhere. (13)

Zopiclone is a cyclopyrrolone derivative believed to bind close to, but not exactly at, the benzodiazepine receptor. Its effects can be reversed by flumazenil. It has a half-life of about 5 h in younger subjects and about 8 h in the elderly. Its sedative and hypnotic effects are similar to those of the benzodiazepines, but its side-effect profile is generally superior with fewer central nervous system effects such as oversedation, confusion, and memory impairment. Rebound and withdrawal problems also seem to be less. (14,15) Long-term benzodiazepine users can be switched to zopiclone with later withdrawal of zopiclone usually being uneventful.

Zolpidem is an imidazopyridine compound that binds selectively to one subtype of the benzodiazepine receptor. It is rapidly absorbed and has a short elimination half-life of 0.7 to 3.5 h (mean 2.4 h). It decreases sleep-onset latency but has less consistent effects on total sleep time. (16,17) Residual effects are uncommon, as are memory disturbances. Rebound and withdrawal have not been seen in clinical trials, but monitoring is continuing. (15)

Zaleplon is also a selective compound with a very short half-life averaging only 1 h. It shortens sleep onset without usually prolonging total sleep time. Residual effects are absent, and memory is minimally disturbed.

These newer compounds appear to improve sleep without affecting sleep architecture, such as suppressing slow-wave sleep. (18)

Clinical effects of anxiolytics

Despite the usual licensed indications of generalized anxiety and panic disorder, the main practical application of the benzodiazepines is to aid in the symptomatic management of anxiety and stress-related conditions. (19) These indications are often so wide as to be difficult to define in terms of recognized disorders. Instead the symptoms of anxiety, in whatever context, are the main indication. Thousands of comparative trials among the benzodiazepines have been carried out, the earlier ones generally using chlordiazepoxide as the standard treatment, more recent ones diazepam or lorazepam, usually with placebo controls. Few differences with respect to risk–benefit ratios have been found among the benzodiazepines, although their superiority to placebo, and usually to the barbiturates, is generally demonstrable.

Difficulties abound in assessing anti anxiety medications. Anxiety disorders are very varied in their natural history; some subside rapidly over a few weeks, whereas others become chronic for no apparent reason, with subsequent acute-on-chronic exacerbations. The latter group of patients tend to be referred to psychiatric outpatient departments because of the treatment problems they present, and because specialist evaluation of the factors sustaining the anxiety is needed. Uncontrolled observations on family practice patients will give a more encouraging impression of antianxiety drugs than will assessment of the more chronic patients attending psychiatric clinics. Even in the latter type of patient, useful symptomatic relief is often obtained without complete resolution of the illness.

Because of the long elimination half-life of drugs such as diazepam, once daily or nightly dosage is sufficient. Nevertheless, many patients prefer to take a divided dosage during the day, often claiming that they can detect further antianxiety activity after each dose and are thereby reassured. For episodic anxiety, shorter-acting compounds such as lorazepam can be used, taken 30 min or so before entering the anxiety provoking situation. If the panic has already started, lorazepam can still be given and will exert a fairly prompt action. Lorazepam is also invaluable in the emergency management of the acutely anxious and disturbed psychotic patient.

More recently, benzodiazepines have been evaluated as treatment for panic disorder, acting by preventing the episodes rather than aborting them. Alprazolam is superior to placebo in preventing panic attacks, but is possibly less efficacious than tricyclic, SSRI, or monoamine oxidase inhibitor (MAOI) antidepressants. Other benzodiazepines for which antipanic actions have been claimed include diazepam and clonazepam. The latter is probably the most widely used by European psychiatrists. However, although suppression of the panic attack is often quite effective, relapse and even rebound may occur when the benzodiazepine is discontinued, even if it is tapered off. (20)

Other uses for the short-acting benzodiazepines are as adjuncts to relaxation therapy, preoperative medication, and deep sedation for minor operative procedures such as dentistry. The drugs render the patient calm, conscious, and cooperative, with often total anterograde amnesia for the operation.

Unwanted effects

The most common unwanted effects of the anxiolytic benzodiazepines are tiredness, drowsiness, and torpor—so-called ‘over-sedation'. The effects are dose and time related, being maximal within the first 2 h after large doses. Drowsiness is most common during the first week of treatment, after which it largely disappears probably due to a true tolerance effect. Smokers seem less affected than non-smokers because they metabolize the drugs more rapidly due to increased liver enzyme activity induced by constituents in the tobacco smoke. Patients should be warned of the potential side-effects of any prescribed benzodiazepine and the initial dosage should be cautious. Both psychomotor skills and intellectual and cognitive skills are affected. In particular, patients should be advised not to drive during the initial adjustment of dosage. Important decisions should be deferred during this period because judgement may be affected.

Benzodiazepines have major effects on memory in patients. (9) They differ in their ability to produce memory deficits, with lorazepam being especially powerful. However, most benzodiazepines can cause problems, especially in higher dose and in the elderly.

Psychomotor performance is also affected. Elderly drivers seem particularly at risk. In one survey of drivers over the age of 65 in the Tennessee Medicaid scheme, an increased risk (1.5 of having an injurious crash) was found for those taking a benzodiazepine. (21) However, the risk was even higher for tricyclic antidepressants and the hazard was strongly dose related.

As with other depressant drugs, potentiation of the effects of alcohol can occur. Patients must be warned not to drink alcohol when taking benzodiazepines, either chronically or intermittently.

Patients taking benzodiazepines may develop paradoxical behavioural responses such as uncontrollable weeping, increased aggression and hostility, and acute rage reactions or uncharacteristic criminal behaviour such as shoplifting. This phenomenon is by no means confined to the benzodiazepines; alcohol is a cardinal example of a drug whose use may lead to excessive violence or criminal behaviour. Paradoxical reactions, including the release of anxiety or hostility, are most common during the initial week of treatment, and usually resolve spontaneously or respond to dose adjustment. One cannot predict who will respond adversely. (22) Benzodiazepines may reduce anxiety levels in depressed patients and thereby uncover suicidal ideation. (23) However, reports of the induction of depression by the benzodiazepines in patients with apparent generalized anxiety disorder is probably the result of an initial misdiagnosis and a failure to detect the underlying depression.

Other unwanted effects include respiratory depression, excessive weight gain, skin rash, impairment of sexual function, menstrual irregularities, and, rarely, blood dyscrasias. The use of benzodiazepines in pregnancy is unestablished, but reports of birth defects among babies born of mothers given these drugs have not been confirmed. Benzodiazepines pass readily into the fetus and can produce respiratory depression in the neonate. Withdrawal effects are also documented. Finally, benzodiazepines pass into the mother's milk and can oversedate the baby, so breast feeding should be discouraged if benzodiazepines are prescribed, especially in high dose.


Overdosage with benzodiazepines is common; deaths are not. Although fatal-overdose statistics contain deaths ascribed to benzodiazepines alone, (24) many such attributions are suspect. Only in children and the physically frail, especially those with respiratory illness, are the benzodiazepines on their own hazardous. However, they can markedly potentiate other central nervous system depressant drugs such as alcohol. Typically, persons who take an overdose, say 100 mg of diazepam, become drowsy and fall deeply asleep. Some develop dysarthria, rigidity or clonus of limbs, or a bullous eruption. Sleep lasts from 24 to 48 h but patients are generally rousable. (24) Plasma benzodiazepine concentrations are often high on admission to hospital and are still very high even when the patient wakes up and is conscious; presumably this reflects acute short-term tissue tolerance.

Tolerance and dependence

If tolerance occurred regularly, then escalation of dosage would be the norm. This does occur with the benzodiazepines, but is fairly uncommon. Escalation of dose is often stepwise, with each increment following a temporary deterioration in psychosocial circumstances. Most patients later reduce the dose as the stress recedes, but others continue the higher dose to which they presumably have developed some tolerance.

Tolerance to the clinical effects in patients maintaining moderate doses of benzodiazepines is more controversial. (26) Few controlled observations concern the long-term efficacy of antianxiety compounds in chronically anxious patients. If medication is withdrawn, the original symptoms may reappear. This is taken as evidence that therapeutic benefit still continues. However, it may reflect ‘rebound' rather than long-term clinical benefit. Undoubtedly, many chronically anxious patients are helped by their treatment with benzodiazepines, but this raises the question as to the frequency of psychological and physical dependence on these drugs.(27) Dependence is easily demonstrable in those patients who have attained high doses. Rebound and withdrawal symptoms after the long-acting benzodiazepines diazepam and clorazepate are not usually apparent until about 5 to 10 days after discontinuation. It is much shorter in patients discontinuing the shorter-acting benzodiazepines (2–4 days). The mildest symptoms and signs are anxiety, tension, apprehension, dizziness, tremulousness, insomnia, and anorexia. More severe physical dependence is shown by the withdrawal symptoms of nausea and vomiting, severe tremor, muscle weakness, postural hypotension, and tachycardia. Occasionally, hyperthermia, muscle twitches, convulsions, and confusional psychoses may develop. (28)

By the 1970s, some patients taking normal doses of benzodiazepines (e.g. up to 30 mg/day of diazepam or the equivalent) were complaining of difficulty in discontinuing their drugs, even though their dosage had remained within therapeutic limits. On withdrawal, the symptoms typical of high-dose withdrawal were noted. In particular, perceptual changes were troublesome. (29,30) The proportion of patients taking benzodiazepines chronically who experience withdrawal symptoms on discontinuing medication ranges between 27 and 45 per cent, depending on the criteria used. Tyrer's group believes that persons characterized by unstable mood and impulsive behaviour are prone to abuse sedative drugs, whereas the ‘timid worrier' suffering from chronic anxiety is likely to show a withdrawal syndrome after long-term normal-dose use. (31) Sometimes the withdrawal reactions seem very prolonged (32) or depression may supervene. 

Management of withdrawal

It is widely accepted that the most appropriate way to manage patients withdrawing from benzodiazepines is to taper the dose gradually, because the severe symptoms of withdrawal, such as epileptic fits and confusional episodes, are more likely to follow abrupt than gradual withdrawal. Views differ as to the rate of withdrawal. Detailed guidelines, (33) based on a consensus view in the United Kingdom, recommend minimal intervention first, usually by a general practitioner. This may comprise a letter to the long-term user, or an interview on a routine visit, with advice to taper the medication. More active intervention involves careful assessment, education, and then establishment, with the patient's agreement, of a timetable of about 6 to 8 weeks for withdrawal. Some agencies suggest a month of tapering for every year of benzodiazepine use, but this may result in patients becoming overfocused on their symptoms. One strategy is to try a fairly brisk withdrawal, say over 6 to 8 weeks, and only resort to more gradual tapering if the symptoms become intolerable. 

Another stratagem is first to substitute a long-acting for a short-acting benzodiazepine, say 10 mg of diazepam for 1 mg of lorazepam, and then to taper off the diazepam later.

Patients must be carefully followed up as a depressive illness is not uncommon and may need vigorous treatment. Such an illness may be reactive to the stress of withdrawal or be a recurrence of an earlier affective episode.

Other drugs have been advocated, but most patients are loathe to substitute yet another medication. A b-adrenoceptor antagonist (beta-blocker) may lessen some symptoms, but clonidine is ineffective. Some anecdotal reports have suggested that prescribing carbamazepine might be helpful, but these have not been successfully followed up. Phenobarbitone has also been advocated. Based on evidence from animal studies, fairly large single doses of flumazenil have been tried with some success. (34)

Psychological support is essential, with the doctor or a practice nurse maintaining close contact with the patient during withdrawal. The physician should show clearly that he understands the problems of withdrawal in order to capture the confidence of the patient. He or she must recognize that patients frequently harbour numerous misconceptions and negative expectations about tranquillizers and withdrawal. These must be elicited, identified, discussed, and corrected.

Formal psychological help has not yet been shown to be particularly effective. Relaxation treatment and training in anxiety management skills in the framework of group therapy can boast of only moderate effectiveness. Cognitive–behavioural treatment is currently favoured.

Other anxiolytics

Benzodiazepine-receptor partial agonists

As described earlier, the disadvantages of the benzodiazepines include sedation, psychomotor and cognitive impairment, and withdrawal symptoms after long-term use. Increased understanding of benzodiazepine-receptor mechanisms suggested that compounds might be developed which are partial agonists and/or selective to some subtypes of receptor. (34) Such compounds would be less efficacious than full agonists but might have better adverse-effect profiles and less dependence potential, i.e. superior risk–benefit ratios.

One such compound, alpidem, was actually marketed in France but had to be withdrawn because of liver toxicity. Initial data suggested that it was somewhat less effective than a typical full benzodiazepine agonist but that it had a distinctly better side-effect profile. (36). Abecarnil is a b-carboline which is anxiolytic in doses of 3 to 9 mg/day. However, efficacy has been difficult to demonstrate convincingly, although the side-effect profile does appear to resemble that of a comparator benzodiazepine.

5-HT1A partial agonists

Many of these are chemically azapirones and they have a complex pharmacology. The first, buspirone, was licensed in many countries some years ago. (37) These drugs suppress activity in presynaptic serotonergic neurones, diminishing serotonin activity and leading on to down-regulation of 5-HT 2 and perhaps other 5-HT receptors. (38) Tests in animals suggested anxiolytic activity without hypnotic, anticonvulsant, or muscle-relaxant properties. Buspirone is much less sedative than the benzodiazepines and causes little or no psychomotor or cognitive impairment, (39) nor does it potentiate the effects of alcohol.

In formal clinical trials, buspirone was equi-effective and equipotent to diazepam, but patients taking buspirone improve more slowly. (40) Patients previously exposed to extensive benzodiazepine treatment do poorly on buspirone. The side-effects of buspirone include headache, dizziness, and nausea. Discontinuation is not accompanied by either rebound or withdrawal.

Other 5-HT1A partial agonists, such as gepirone, tandospirone , ipsapirone, and flesinoxan, are being developed mainly as potential antidepressants (41,42) rather than antianxiety agents, (43) but only tandospirone (used in China and Japan) has yet reached the market.

Antipsychotic drugs

Phenothiazines, such as chlorpromazine and trifluoperazine, and a range of other antipsychotic drugs have been advocated for treating anxiety. The dosage recommended is quite low, typically less than half the initial antipsychotic dose used in psychotic patients. Sometimes, even at this dosage, the antipsychotic drug is not well tolerated by the anxious patient because unwanted autonomic effects, such as dry mouth and dizziness, too closely resemble the symptoms of anxiety. Even more unwelcome are extrapyramidal symptoms such as restlessness (mild akathisia) and parkinsonism, although at the low doses advocated such unwanted effects are uncommon. There may even be a risk of tardive dyskinesia. The chief advantage of this medication is that dependence is virtually unknown, so the main indication for their use is in patients with histories of dependence on other central nervous system depressant drugs such as alcohol or barbiturates.


Several of these drugs, such as amitriptyline, doxepin, and trazodone, have useful secondary sedative properties. They are widely prescribed for depressed patients with anxiety or agitation. Although regarded as inappropriate for patients with primary anxiety disorders, some clinical trials have showed quite impressive efficacy. (44)

More recently, several SSRI antidepressants have been evaluated in the treatment of various anxiety disorders. A large series of trials has established their efficacy in a range of disorders, including obsessive– compulsive disorder, panic disorder, social anxiety disorder, and post-traumatic stress disorder. (45,46 and 47) Some compounds are now licensed for one or more of these indications. The antianxiety effect is independent of any antidepressant effect. The antianxiety dosage of the SSRIs is sometimes greater than the modal antidepressant dose. Venlafaxine, selective on both noradrenaline and serotonin mechanisms, has been extensively tested in the indications for generalized anxiety disorder and results are very encouraging. SSRIs may well be soon the treatment of choice in chronic anxiety disorders.

MAOIs have been used for many years to treat phobic states, and evidence of the efficacy for some, particularly phenelzine, is quite impressive. Unfortunately, the well-known range of unwanted effects, including hypotension, oedema, and dietary and drug interactions, preclude their widespread use. The selective reversible MAOI antidepressant, moclobemide, has been evaluated in social-anxiety disorder with some success. A similar compound, brofaromine, appeared even more successful in anxiety disorders, but its development has been discontinued. (48)

b-Adrenoceptor antagonists (beta-blockers)

Anxiety states are accompanied by many different bodily symptoms, some of which are mediated by the sympathetic nervous system. In particular, palpitations, tremor, and gastrointestinal upset are related to the overactivity of b-adrenergic pathways. Consequently, blockade of this activity by means of b-adrenoceptor antagonists may help patients with anxiety, but usually only those complaining of the symptoms mentioned above. (49) In appropriate patients, a test dose, say 20 mg of propranolol, should be administered to detect undue sensitivity of the patient, as shown by a bradycardia below 60 beats/min at rest. If no sensitivity is found, the dose can be instituted at 20 mg four times daily and increased over the course of a week or two to 40 mg four times daily.


H1 antagonists usually also have some anticholinergic effects and are used in the treatment of allergic reactions. The older compounds penetrate the brain readily and are quite sedative. In conservative dosage, more specific anxiolytic properties can be demonstrated as well. Hydroxyzine has been evaluated in at least two placebo-controlled trials and proved to be significantly better than placebo. (50,51) The effective dose is generally 50 mg/day in divided doses, and sedation generally wanes over a few days. The advantages are that paradoxical reactions are rare, cognitive function including memory is largely unaffected, and rebound and withdrawal remain rare.

Clinical effects of hypnotics

Insomnia is a very common symptom, (52) but many complaints of insomnia are unfounded as the patient has unreal expectations concerning sleep. (53) Elderly people fail to appreciate that it is normal to sleep less and less deeply as they age. Cat-napping during the day also decreases the need for sleep at night. Some old people cannot afford to heat their rooms for more than a few hours each day and take to their beds as the warmest place. Not surprisingly their sleep patterns become disrupted.

Other subjects can manage on 5 to 6 h a night as a normal procedure and yet worry that this is insufficient. Explanation and reassurance relieve their worries.

In many cases of more severe insomnia, the cause is a physical complaint such as pain, breathlessness, or pruritus. The treatment is for that of the primary complaint. In many other cases, the insomnia is either a symptom of psychiatric distress, anxiety, or depression, or it is iatrogenic, caused by the very drugs prescribed to relieve the insomnia. In the first instance, treatment is directed towards the primary condition; in the second, a careful regimen of drug withdrawal or substitution and subsequent withdrawal must be planned, as discussed earlier for anxiolytic medication. Some drugs, of which caffeine is the most common, induce insomnia. (54)

Despite this, a substantial number of patients cannot be placed into these categories and yet they persistently complain of insomnia (primary insomnia). Careful evaluation of the issues may yet reveal some relationship to stresses, both transient or persistent. It can be established that these patients are responding to unusual or protracted pressures of life: a man worries over possible redundancy, his wife is concerned about their delinquent son, their daughter is lovelorn, and grandfather is anxious over his physical decrepitude. Giving drugs may set in train a long-term process culminating in drug-related insomnia without solving the basic problems. (55)

Short-term symptomatic relief seems appropriate when the stress is undoubtedly severe but transient. (56) Even so, the hypnotic agent must be chosen carefully. (57) The elimination half-life is the most important consideration. Those with half-lives over 12 h, such as nitrazepam, are only appropriate where an anxiolytic effect is required during the day as well as sleep induction at night. Even here, diazepam 5 to 15 mg, one dose at night, may be preferred. For a more selective hypnotic effect, temazepam or triazolam (outside the United Kingdom), with their shorter half-lives, will encourage sleep onset without leaving the patient with residual sedative effects the next day. The management of chronic insomnia is much more problematic. (58) The newer compounds zopiclone and zolpidem are also short-acting agents and can help assure a good night's sleep without much risk of residual sedative effects the next day. (59) This is dependent on the dosage being kept modest, especially in the elderly.

Zolpidem, and zaleplon even more so, can be used in a different strategic way from other longer-acting drugs. Hypnotics are traditionally taken every night before going to bed to induce or maintain sleep. However, the severity of insomnia usually varies from night to night. Consequently, regular usage may be partly, or even largely, unnecessary, and increases the risk of habituation and dependence. Very-short-acting compounds are unlikely to leave residual effects the next day, even taken up to 5 h or so before the expected time of wakening. Consequently, the insomniac can refrain from regular hypnotic usage but, instead, wait up to an hour or so after going to bed to see if natural sleep supervenes before resorting to medication. This changes the regular prophylactic use of hypnotics to ‘as needed', and lessens the risk of habituation and dependence. Furthermore, the patient is gratified to feel that he or she has control of the medication instead of vice versa.

Residual effects

Residual effects can be a problem especially when long-acting drugs are used repeatedly. (60) Dosage is important here, since residual effects increase in both magnitude and duration as the dose is increased. It should be remembered that hypnotics are the only class of drugs in which the main therapeutic effect (drowsiness) is identical with the main unwanted effect; the two are merely separated by 8 h in time! A short-acting hypnotic compound will be devoid of residual effects the next day, but the patient may wake early. After taking a longer-acting compound, sleep may be prolonged but hangover effects pronounced.

Idiosyncratic effects

Adverse effects with triazolam alerted prescribers and regulators to possible major adverse effects of short-acting benzodiazepines. (61) The adverse reactions in question include daytime anxiety, amnesic effects and episodes, and morbid affects such as depression and hostility. In summary, the evidence suggests that these are class effects common to the benzodiazepines, although more likely to occur the shorter the duration of action of the drug and the higher the dose. (62,63) Alcohol is also capable of producing these effects. (64)


Discontinuation of many hypnotics is often followed by worsening of sleep compared with pretreatment levels. In practical terms, insomniac patients find that their sleep is disturbed for a night or two after abrupt discontinuation of what appeared to be effective medication. (65) Some of this rebound is subjective as patients taking sleeping pills tend to overestimate their sleeping time (compared with sleep laboratory recordings); on withdrawal, they underestimate their sleep. (66) The intensity of rebound insomnia is strongly related to dose but less clearly to the duration of use, and marked individual differences exist. The risk of rebound is greater with short-half-life compared with the long-half-life compounds. Tapering off medication lessens the likelihood of rebound. However, despite clinical impressions that rebound insomnia might lead to the resumption of medication, little evidence for this is extant. (67)


Dependence may supervene on the longer term use of hypnotics; giving a benzodiazepine drug only once in 24 h does not protect against such an eventuality. The management of the withdrawal syndrome that may occur is largely the same as with the anxiolytic benzodiazepines.


A growing problem with these drugs is abuse—non-medical use, on a regular or sporadic basis, often in a polydrug context. Worldwide, flunitrazepam is the main problem and can be taken orally, by injection, or by sniffing. In the United Kingdom, temazepam is widely abused by injection. (68) The contents of the original formulation of temazepam (a liquid-filled capsule) could be extracted and injected. Reformulation as a gel-filled capsule was unsuccessful as the gel could still be liquefied and injected, often with subsequent major vascular trauma. The injected drug has a marked sedative and/or disinhibiting effect, resulting in chaotic behaviour, carelessness, and an enhanced risk of the transmission of communicable diseases such as HIV infection and hepatitis.

Another abuse of flunitrazepam that has come to public attention is its clandestine administration for the purpose of sexual abuse and rape. Some hypnotics have been reformulated to prevent any such undetectable administration.


The drug treatment of both anxiety and insomnia still largely revolves around the use of the benzodiazepine class of drugs. Nevertheless, controversy and disagreement still rage about the risk–benefit ratio of compounds in this area. Short-term use in both indications is well established, with a favourable database as a rationale for this approach. However, long-term use is still only researched in a limited way. While both the efficacy and safety of long-term use remain unclear, acceptance of current guidelines limiting the use of benzodiazepines seems wise.

Nevertheless, in the author's opinion the most important outstanding issue is the relationship between drug and non-drug treatments. (69) The management of anxiety disorders and of insomnia is complex and is hampered by a dearth of information concerning the relative merits of various treatment modalities. Much research is also needed on the optimum strategies for combining all the therapies available to us, and on identifying predictors of response. 


1. Allgulander, C. (1986). History and current status of sedative-hypnotic drug use and abuse. Acta Psychiatrica Scandinavica, 73, 465–78.

2. Lader, M. (1994). Benzodiazepines. A risk–benefit profile. CNS Drugs, 1, 377–87.

3. Ashton, H. (1994). Guidelines for the rational use of benzodiazepines. When and what to use. Drugs, 48, 25–40.

4. Balter, M.B., Ban, T., and Uhlenhuth, E.H. (1993). International study of expert judgment on therapeutic use of benzodiazepines and other psychotherapeutic medications: I. Current concerns. Human Psychopharmacology, 8, 253–61.

5. Greenblatt, D.J. and Wright, C.E. (1993). Clinical pharmacokinetics of alprazolam. Therapeutic implications. Clinical Pharmacokinetics, 24, 453–71.

6. Brogden, R.N. and Goa, K.L. (1988). Flumazenil. A preliminary review of its benzodiazepine antagonist properties, intrinsic activity and therapeutic use. Drugs, 35, 448–67.

7. Ghoneim, M.M., Hinrichs, J.V., and Mewaldt, S.P. (1986). Comparison of two benzodiazepines with differing accumulation: behavioral changes during and after 3 weeks of dosing. Clinical Pharmacology and Therapeutics, 39, 491–500.

8. Ghoneim, M.M. and Mewaldt, S.P. (1990). Benzodiazepines and human memory: a review. Anesthesiology, 72, 926–38.

9. Curran, H.V. (1991). Benzodiazepines, memory and mood: a review. Psychopharmacology, 105, 1–8.

10. Curran, H.V. and Birch, B. (1991). Differentiating the sedative, psychomotor and amnesic effects of benzodiazepines: a study with midazolam and the benzodiazepine antagonist, flumazenil. Psychopharmacology, 103, 519–23.

11. Linnoila, M., Stapleton, J.M., Lister, R., et al. (1990). Effects of single doses of alprazolam and diazepam, alone and in combination with ethanol, on psychomotor and cognitive performance and on autonomic nervous system reactivity in healthy volunteers. European Journal of Clinical Pharmacology, 39, 21–8.

12. Maczaj, M. (1993). Pharmacological treatment of insomnia. Drugs, 45, 44–55.

13. Gelenberg, A.J. (ed.) (1992). The use of benzodiazepine hypnotics: a scientific examination of a clinical controversy. Journal of Clinical Psychiatry, 53 (Supplement 12).

14. Elie, R., Lavoie, G., Bourgouin, J., and Le Morvan, P. (1990). Zopiclone versus flurazepam in insomnia: prolonged administration and withdrawal. International Clinical Psychopharmacology, 5, 279–86.

15. Lader, M. (1997). Zopiclone: is there any dependence and abuse potential? Journal of Neurology, 244 (Supplement 1), S18–22.

16. Langtry, H.D. and Benfield, P. (1990). Zolpidem. A review of its pharmacodynamic and pharmacokinetic properties and therapeutic potential. Drugs, 40, 291–313.

17. Lahmeyer, H., Wilcox, C.S., Kann, J., and Leppik, I. (1997). Subjective efficacy of zolpidem in outpatients with chronic insomnia. A double-blind comparison with placebo. Clinical Drug Investigations, 13, 134–44.

18. Monti, J.M. (1989). Effect of zolpidem on sleep in insomniac patients. European Journal of Clinical Pharmacology, 36, 461–6.

19. Hoehn-Saric, R. (1998). Generalised anxiety disorders. Guidelines for diagnosis and treatment. CNS Drugs, 9, 85–98.

20. Otto, M.W., Pollack, M.H., and Sachs, G.S. (1993). Discontinuation of benzodiazepine treatment: efficacy of cognitive–behavioral therapy for patients with panic disorder. American Journal of Psychiatry, 150, 1485–90.

21. Ray, W.A., Fought, R.L., and Decker, M.D. (1992). Psychoactive drugs and the risk of injurious motor vehicle crashes in elderly drivers. American Journal of Epidemiology, 136, 873–83.

22. Dietch, J.T. and Jennings, R.K. (1988). Aggressive dyscontrol in patients treated with benzodiazepines. Journal of Clinical Psychiatry, 49, 184–7.

23. Tiller, J.W.G. and Schweitzer, I. (1992). Benzodiazepines. Depressants or antidepressants? Drugs, 44, 165–9.

24. Serfaty, M. and Masterton, G. (1993). Fatal poisonings attributed to benzodiazepines in Britain during the 1980s. British Journal of Psychiatry, 163, 386–93.

25. Höjer, J. (1994). Management of benzodiazepine overdose. CNS Drugs, 2, 7–17.

26. Michelini, S., Cassano, G.B., Frare, F., and Perugi, G. (1996). Long-term use of benzodiazepines: tolerance, dependence and clinical problems in anxiety and mood disorders. Pharmacopsychiatry, 29, 127–34.

27. Marriott, S. and Tyrer, P. (1993). Benzodiazepine dependence. Avoidance and withdrawal. Drug Safety, 9, 93–103.

28. Zipursky, R.B., Baker, R.W., and Zimmer, B. (1985). Alprazolam withdrawal delirium unresponsive to diazepam: case report. Journal of Clinical Psychiatry, 46, 344–5.

29. Petursson, H. and Lader, M.H. (1981). Withdrawal from long-term benzodiazepine treatment. British Medical Journal, 283, 643–5.

30. Petursson, H. and Lader, M. (1984). Benzodiazepine dependence, tolerance and withdrawal syndrome. In Advances in human psychopharmacology (ed. G.D. Burrows and J.S. Werry), pp. 89–119. JAI Press, Greenwich, CT.

31. Tyrer, P. (1989). Risks of dependence on benzodiazepine drugs: the importance of patient selection. British Medical Journal, 298, 102–5.

32. Tyrer, P. (1991). The benzodiazepine post-withdrawal syndrome. Stress Medicine, 7, 1–2.

33. Russell, J. and Lader, M. (ed.) (1993). Guidelines for the prevention and treatment of benzodiazepine dependence. Mental Health Foundation, London.

34. Lader, M.H. and Morton, S.V. (1992). A pilot study of the effects of flumazenil on symptoms persisting after benzodiazepine withdrawal. Journal of Psychopharmacology, 6, 357–63.

35. Potokar, J. and Nutt, D.J. (1994). Anxiolytic potential of benzodiazepine receptor partial agonists. CNS Drugs, 1, 305–15.

36. Frattola, L., Piolti, R., Bassi, S., et al. (1992). Effects of alpidem in anxious elderly outpatients: a double-blind placebo-controlled trial. Clinical Neuropharmacology, 15, 477–87.

37. Gelenberg, A.J. (ed.) (1994). Buspirone: seven-year update. Journal of Clinical Psychiatry, 55, 222–9.

38. Deakin, J.F.W. (1993). A review of clinical efficacy of 5-HT 1A agonists in anxiety and depression. Journal of Psychopharmacology, 7, 283–9.

39. O'Hanlon, J.F. (1991). Review of buspirone's effects on human performance and related variables. European Neuropsychopharmacology, 1, 489–501.

40. Fulton, B. and Brogden, R.N. (1997). Buspirone. An updated review of its clinical pharmacology and therapeutic applications. CNS Drugs, 7, 68–88.

41. Glitz, D.A. and Pohl, R. (1991). 5-HT 1A partial agonists. What is their future? Drugs, 41, 11–18.

42. Pecknold, J.C. (1994). Serotonin 5-HT1A agonists. A comparative review. CNS Drugs, 2, 234–51.

43. Boyer, W.F. and Feighner, J.P. (1993). A placebo-controlled double-blind multicenter trial of two doses of ipsapirone versus diazepam in generalized anxiety disorder. International Clinical Psychopharmacology, 8, 173–6.

44. Rickels, K., Downing, R., Schweizer, E., and Hassman, H. (1993). Antidepressants for the treatment of generalized anxiety disorder. A placebo-controlled comparison of imipramine, trazodone, and diazepam. Archives of General Psychiatry, 50, 884–95.

45. Lecrubier, Y., Bakker, A., Dunbar, G., Judge, R., and the Collaborative Paroxetine Panic Study Investigators (1997). A comparison of paroxetine, clomipramine and placebo in the treatment of panic disorder. Acta Psychiatrica Scandinavica, 95, 145–52.

46. Fichtner, C.G., Poddig, B.E., and deVito, R.A. (1997). Post-traumatic stress disorder. Pathophysiological aspects and pharmacological approaches to treatment. CNS Drugs, 4, 293–322.

47. Boyer, W. (1995). Serotonin uptake inhibitors are superior to imipramine and alprazolam in alleviating panic attacks: a meta-analysis. International Clinical Psychopharmacology, 10, 45–9.

48. Fahlén, T., Nilsson, H.L., Borg, K., Humble, M., and Pauli, U. (1995). Social phobia: the clinical efficacy and tolerability of the monoamine oxidase-A and serotonin uptake inhibitor brofaromine. Acta Psychiatrica Scandinavica, 92, 351–8.

49. Morgan, J. and Tyrer, P. (1994). Treating the somatic symptoms of anxiety. CNS Drugs, 1, 427–34.

50. Ferrari, M., Darcis, T., Burtin, B., and the French GP Study Group for Hydroxyzine (1995). A multicentre double-blind placebo-controlled study investigating the anxiolytic efficacy of hydroxyzine in patients with generalized anxiety. Human Psychopharmacology, 10, 181–7.

51. Lader, M. and Scotto, J.C. (1998). A multicentre double-blind comparison of hydroxyzine, buspirone and placebo in patients with generalized anxiety disorder. Psychopharmacology, 139, 402–6.

52. Üstün, T.B., Privett, M., Lecrubier, Y., et al. (1996). Form, frequency and burden of sleep problems in general health care: a report from the WHO Collaborative Study on Psychological Problems in General Health Care. European Psychiatry, 11 (Supplement 1), 5s–10s.

53. Lader, M. (ed.) (1992). The medical management of insomnia in general practice. Round Table Series No. 28. Royal Society of Medicine, London.

54. Tiffin, P., Ashton, H., Marsh, R., and Kamali, F. (1995). Pharmacokinetic and pharmacodynamic responses to caffeine in poor and normal sleepers. Psychopharmacology, 121, 494–502.

55. Vela-Bueno, A. and Kales, A. (1986). Benzodiazepine hypnotics in the multidimensional treatment of insomnia. Drugs of Today, 22, 271–81.

56. Lader, M.H. (1986). A practical guide to prescribing hypnotic benzodiazepines. British Medical Journal, 2, 1048–9.

57. Nicholson, A.N. (1986). Hypnotics—their place in therapeutics. Drugs, 31, 164–76.

58. Monti, J.M. and Monti, D. (1995). Pharmacological treatment of chronic insomnia. CNS Drugs, 4, 182–94.

59. Nowell, P.D., Mazumdar, S., Buysse, D.J., et al. (1997). Benzodiazepines and zolpidem for chronic insomnia. A meta-analysis of treatment efficacy. Journal of the American Medical Association, 278, 2170–7.

60. Greenblatt, D.J., Miller, L.G., and Shader, R.I. (1990). Neurochemical and pharmacokinetic correlates of the clinical action of benzodiazepine hypnotic drugs. American Journal of Medicine, 88 (Supplement 3A), 18–24.

61. O'Donovan, M.C. and McGuffin, P. (1993). Short acting benzodiazepines. Dream drugs or nightmare? British Medical Journal, 306, 945–6.

62. Jonas, J.M. (1992). Idiosyncratic side effects of short half-life benzodiazepine hypnotics: fact or fancy? Human Psychopharmacology, 7, 205–16.

63. Rush, C.R., Higgins, S.T., Hughes, J.R., and Bickel, W.K. (1993). A comparison of the acute behavioral effects of triazolam and temazepam in normal volunteers. Psychopharmacology, 112, 407–14.

64. Roache, J.D., Cherek, D.R., Bennett, R.H., Schenkler, J.C., and Cowan, K.A. (1993). Differential effects of triazolam and ethanol on awareness, memory, and psychomotor performance. Journal of Clinical Psychopharmacology, 13, 3–15.

65. Lader, M. and Lawson, C. (1987). Sleep studies and rebound insomnia: methodological problems, laboratory findings, and clinical implications. Clinical Neuropharmacology, 10, 291–312.

66. Schneider-Helmert, D. (1988). Why low-dose benzodiazepine-dependent insomniacs can't escape their sleeping pills. Acta Psychiatrica Scandinavica, 78, 706–11.

67. Roehrs, T., Merlotti, L., Zorick, F., and Roth, T. (1992). Rebound insomnia and hypnotic self administration. Psychopharmacology, 107, 480–4.

68. Strang, J., Seivewright, N., and Farrell, M. (1992). Intravenous and other novel abuses of benzodiazepines: the opening of Pandora's box? British Journal of Addiction, 87, 1373–5.

69. Fineberg, N. and Drummond, L.M. (1995). Anxiety disorders. Drug treatment or behavioural cognitive psychotherapy? CNS Drugs, 3, 448–66.