« back · www.benzo.org.uk »


The Medical Journal of Australia, Vol. 144, May 26, 1986

K. Jean Lennane, FRACP, DPM,
Director, Drug and Alcohol Services,
Rozelle and Gladesville Hospitals,
The Rozelle Hospital, Church Street,
Leichhardt, NSW 2040.

[Note: Tapering and withdrawal advice, duration of withdrawal, equivalencies etc indicated in this document are not necessarily now considered correct or appropriate. The information in this document is therefore for reference purpose only. – benzo.org.uk]

The first benzodiazepine agent was introduced in 1960. The first report of probable dependence appeared in 1961(1) and the introduction of newer and "better" benzodiazepine drugs over the next 24 years has been followed equally promptly by reports of dependence problems.(2-5) Because of the weight of evidence that now implicates all known benzodiazepine drugs, the UN Commission on Narcotic Drugs last year imposed the same international regulations on them as apply to narcotic and other known dependence-producing drugs.(6) However, most doctors in Australia seem unaware that benzodiazepine agents can cause dependence at all, and are continuing to prescribe them at the rate of 6 million PBS prescriptions (not including hospital and repatriation benefits) per year (Assistant Director-General, Pharmaceutical Branch, Commonwealth Department of Health, personal communication) often for doubtful indications such as dysmenorrhoea, tinnitus, dizzy spells, postoperative shock, backache, neck pain and headaches.(7)


As were all new sedative/tranquillizer drugs, benzodiazepine agents were initially launched as "non-habit forming". However, they appeared rapidly on the illegal drug scene. Benzodiazepine abuse has been reported extensively among heroin and methadone users.(8.9) and also among persons with alcohol dependence.(10) There were regular reports of individual cases who developed dependence problems.(10,12) Reports were published of withdrawal problems in neonates who had been exposed to benzodiazepine agents in utero.(13,14) More recently, a substantial body of work in animals has been collected; no difficulty seems to exist in creating benzodiazepine dependence in rats, cats and baboons. (15-17) Finally, in the last three to four years, there has been an increasing number of reports of dependence in groups of humans, indicating a widespread problem among normal users.(7,18-20) This has many features in common with dependence on other sedative/tranquillizer drugs, and some distinctive features.

One distinctive feature is the frequent development of dependence without classic tolerance and, therefore, without increasing dosage. While classic tolerance can and does occur, particularly in users with a history of dependence on other drugs, it is relatively uncommon.(21,22) However, dependence while taking normal doses – which is reported only spasmodically in "drug-scene" users,(9,10) neonates(13,1) and individuals(23) – seems from studies of groups of patients to be very common. A high proportion of long-term normal-dose users are physically dependent and will suffer significant symptoms if the drug is withdrawn. Two studies by Tyrer et al. suggest that the frequency of significant dependence is between 27% and 45%.'(18,19) Owen and Tyrer suggest that the true figure, after correcting for "pseudo-withdrawal" reactions, is around 30% in unselected normal dose users.(24)

However, it is suggested that the shorter-acting benzodiazepine agents (for example, oxazepam, lorazepam and temazepam) have a greater potential to cause dependence than do the longer-acting ones (such as diazepam and nitrazepam).(18,24) Since the study which led Owen and Tyrer to the prevalence figure of 30% was of patients who were withdrawing from diazepam, the figure for the Australian population, where short-acting benzodiazepine drugs make up half the market (Assistant Director-General, Pharmaceutical Branch, Commonwealth Department of Health, personal communication), may well be nearer to 45%.

"Long-term" usage in Tyrer et al.'s studies was considered to be usage for three months or longer. The Committee on the Review of Medicines states that the tranquillizing effects of benzodiazepine drugs do not persist beyond three to four months.(21) This suggests that tolerance of the limited type that occurs with benzodiazepine agents (that is, loss of the effect of the usual dose) must be a usual occurrence at about that point. The frequency of dependence may increase up to about 12 months of use, but there is no evidence that longer usage increases it further.(22)

However, dependence can occur in less than three months. Occasional patients particularly those who have been dependent on other drugs – give anecdotal reports of almost instantaneous dependence on benzodiazepine drugs, and many such patients develop obvious dependence in five to 10 days. Withdrawal symptoms have been noted in patients who have received benzodiazepine agents for three weeks.(25) Animal studies show that the acute administration of, for example, diazepam leads to a rapid increase in brain benzodiazepine receptor density,(26) and the development of withdrawal symptoms after as few as seven days.(17)

Dependence on the hypnotic effect occurs even more rapidly. Studies in humans consistently show withdrawal symptoms of disturbed steep and daytime dysphoria after seven days of benzodiazepine administration.(27,28) Rebound insomnia may also occur acutely, even during the first night of administration of very short-acting benzodiazepine agents. Similar rebound anxiety may also occur during the day.29,30)

Therefore, it seems that the occurrence of withdrawal symptoms if patients try to decrease or stop benzodiazepine agents is keeping up to half the patients taking the tablets. Major features of the withdrawal syndrome are anxiety and insomnia. Since these are the usual reasons for the prescription of a benzodiazepine drug in the first place, patients and their doctors may see the emergence of such symptoms as justification for continuing therapy, without considering withdrawal as a possibility.

In fact, the only way to differentiate initial withdrawal symptoms from a relapse into anxiety is by persisting with the withdrawal. Withdrawal symptoms will have largely subsided by three to four weeks, whereas anxiety will persist or worsen.(24)

The withdrawal syndrome

Benzodiazepine withdrawal has been described in detail by many authors.(5,18,19,24) Many symptoms are typical of the withdrawal of any sedative or tranquillizer drug. Grand mal seizures(1,4) and/or delirium(12) may occur if the drug is stopped suddenly. In 1982, eight of 200 patients who were admitted to the McKinnon Unit for detoxification, primarily from benzodiazepine drugs, had one or more seizures, an four had delirium – an incidence of 4% and 2%, respectively. This appears to indicate a slightly higher risk of seizures on abrupt withdrawal than with alcohol, and a somewhat lower risk of delirium. In these, as in other reported cases, the dosage has varied from a normal dosage (that is, 45-60 mg a day of oxazepam) to very high (2000 mg a day).

Delirium differs from that which occurs with alcohol in that patients do not appear toxic, are pale rather than flushed, and are normotensive. They tend to look blank rather than perplexed, and while they may be extremely paranoid (often experiencing interesting hallucinations by way of nonexistent radios and television screens), they will often give no indication of this in their outward behaviour. The first indication of a problem may be an attempt to jump out of a window to escape.

Delirium and seizures, which are uncommon even with abrupt withdrawal, do not occur when benzodiazepine drugs are withdrawn gradually. However, a long list of other symptoms do. The most detailed description is that given by Ashton.(7) Anxiety, insomnia, irritability, tremulousness, gastrointestinal disturbances and dysphoria occur much as in the withdrawal of any tranquillizer or sedative drug. Other symptoms appear to be more specific. Perceptual distortions occur in all modalities. Sounds may be unduly loud and patients may hear non-existent thumps or tunes. Sights are distorted and may be misinterpreted, with occasional brief visual hallucinations. The most common complaint is of feelings of unreality and depersonalization, and of seeing "through a veil". Paraesthesiae are common, as are distortions of smell and taste. Paranoid thoughts and feelings occur frequently. Pain and stiffness in various parts of the body, especially the face, are common, with muscular spasms which may appear as myoclonic jerks, or as local tremors and fasciculation. There may be difficulty in walking. Ashton also reports a 'flu-like illness in 10 of her 12 patients that was reminiscent of narcotic withdrawal, and also menorrhagia and breast pain. Marked weight loss is common.

The symptoms can be exceedingly unpleasant and distressing.(7,23) What makes withdrawal so bad, apart from the perceptual disturbances which make patients feel they are going mad, is that it goes on or so long. With other sedative or tranquillizing drugs, return to a normal sleeping pattern is likely to take some weeks after withdrawal, and return to full normal health takes some months, but specific withdrawal symptoms are over within a few days. Benzodiazepine withdrawal symptoms typically last at least four weeks.(24) Ashton's study suggests that many symptoms, though improved by four weeks, may continue intermittently for months.(7) It is clear that patients who are dependent and want to withdraw may face a very unpleasant illness – in most cases, very much worse than the condition for which they were prescribed benzodiazepine drugs in the first place.

Mechanism of dependence

The mechanism that causes dependence on benzodiazepine drugs resembles that of other sedative/tranquillizer drugs.(31) The effect appears to be mediated by specific receptors(15,26) which enhance the transmission of GABA, a central nervous system inhibitory neurotransmitter substance that acts as a tranquillizer agent.(32) Experience of the long succession of drugs that have been introduced as "non-addictive" should already have told us that which the work on opiate and benzodiazepine receptors now makes clear: any drug which relieves pain or anxiety centrally does so by occupying a specific receptor site. The site exists for an endogenous substance which resembles the drug. If the drug occupies the receptor site for more than a few days, the body will reduce or stop production of the endogenous substance which is being consistently displaced. The number of receptor sites may also increase. If the drug is then withdrawn, there will be nothing available to fill the receptor site until the body can return to its normal level of production – a process that with most drugs takes several days, but with benzodiazepine agents, for reasons not yet explained, takes much longer. Uninhibited firing from empty receptor sites causes the withdrawal symptoms. (The perceptual disturbances may have a more complex explanation.(33)

Treatment of dependence

Treatment of benzodiazepine dependence at present is unsatisfactory. There is no way of relieving the withdrawal symptoms to a degree that is acceptable to many patients.(7,24,34) Therefore, they may be stuck with their problem, the only solution being to keep on taking the tablets. This can hardly be considered good medicine, and there are other difficulties. One is the risk of accumulation in elderly patients, who may develop apathy and confusion, that are attributed erroneously to dementia.(35,36) Another is the risk that withdrawal symptoms may develop inconveniently and dangerously after admission to hospital for other reasons (for example, surgery). Yet a third is the risk of permanent brain damage, analogous to alcohol-related damage, which it seems may occur with long-term usage.(37)

However, the most pressing reason for trying to cease the administration of benzodiazepine drugs is that highlighted by Ashton.(7) There is a period of a few days to a few months during which the benzodiazepine drug has an active effect.(21) A period of months to years then occurs when there is no longer any active effect, but the drug in normal dosage prevents the occurrence of withdrawal symptoms. Some patients then progress to the "problem" phase, when withdrawal symptoms start to occur although they are still taking the medication. The great majority of such patients do not attempt to compensate for this tolerance by increasing their dosage, and this phase may continue for months or years – until someone realizes that the benzodiazepine drug is actually making the patient ill. Any or all of the withdrawal symptoms that have been described already may occur, daily or many times a day, as long as benzodiazepine use continues. Three are particularly common: disturbed sleep (nightmares, frequent waking and insomnia); anxiety or panic attacks that occur when the next dose is nearly due; and agoraphobia. Eleven of Ashton's 12 patients had developed agoraphobia, and this resolved with no treatment but benzodiazepine withdrawal in all but two. Patients who keep on taking the tablets must be carefully monitored for the appearance of such "neurotic" symptoms, and the doctor must aware, if they appear, that the answer is not more medication, but benzodiazepine withdrawal.

How should one treat patients who are dependent? This really means: What should we do about patients who have been receiving benzodiazepine drugs for more than three months? Since the benzodiazepine agent will no longer be effective,(21) and those who are not already dependent may in due course become so, the drugs should be ceased in all such patients. Over half will not be dependent and will be able to stop without difficulty. However, to minimize the risk of fits or delirium, withdrawal should be gradual.(18)

Management of withdrawal

If patients are drug dependent and develop significant symptoms, treatment becomes complicated and unsatisfactory, since attempts to alleviate the symptoms have in general been unimpressive.(7,18,19,22,23) The following suggestions are an amalgam of these experiences.

Discuss with the patient the need to withdraw, and offer continued support throughout the withdrawal period. If the patient refuses, bring up the topic with every repeat prescription thereafter.

If the patient agrees, decide on a mutually acceptable timetable. This will often be the reduction by one-sixth of the total dose every week, starting with half the lunch-time dose, then half the breakfast dose, half the night dose, then the remaining lunch, breakfast and night doses in that order.

See the patient at least weekly to reinforce the dosage reduction and check for the occurrence of withdrawal symptoms. Over 50% will succeed without problems of any note, but should be seen at least weekly for four weeks after the last dose, as withdrawal symptoms may appear for the first time at that stage.

If significant symptoms occur, or if the patient is already suffering from withdrawal symptoms when first seen, review the benzodiazepine drug that is involved. Withdrawal in dependent patients tends to be more difficult and unpleasant with short-acting benzodiazepine agents. Therefore, patients on short-acting drugs should be changed to long-acting drugs (usually diazepam) substituting an equivalent dose over three to four days. (The approximate equivalent of 5 mg of diazepam is 15 mg of oxazepam, 0.5 mg of lorazepam and 10 mg of temazepam.)

Once you know the patient is going to develop symptoms, he or she has to decide whether they want to experience more severe symptoms for a shorter time, or less severe symptoms for a longer time, which seems to be the choice. Reduction of the dose by one sixth every three days will achieve the former course, by one-sixth every two weeks the latter course.

Continue on the timetable chosen and see the patient twice a week during withdrawal and for four weeks after it is completed. Offer support, explanation and reassurance throughout. Refer the patient to a support group if this is available. Drug therapy is not very helpful, but propranolol (40-160 mg a day) may relieve some symptoms. If insomnia is persistent and distressing, non-benzodiazepine hypnotic agents can be used on alternate nights for one to two weeks (for example, promethazine, 75-100 mg; chlormethiazole, 384 mg; butobarbitone, 100-200 mg). Clonidine might be expected to help, as it does in narcotic withdrawal, but the evidence so far is unimpressive.(7) Frank paranoia may be controlled by small doses of antipsychotic agents, such as haloperidol; however, this may make other symptoms worse.

If outpatient withdrawal is intolerable, the patient should be referred for inpatient withdrawal.

If both outpatient and inpatient withdrawal fail because the patient is unable to tolerate it, there is no option but to return to full dosage of a long-acting benzodiazepine agent and to increase the dose if this does not control symptoms. In due course the dose may have to be increased again and again. Withdrawal will have to be advised and discussed again with each new prescription.

Patients who withdraw without problems should be advised against any but very occasional use of benzodiazepine agents in future; patients who develop withdrawal symptoms should be advised strongly against benzodiazepine use of any kind.


Since withdrawal may be so distressing and difficult, prevention is all-important. There are, in my opinion, very few indications for benzodiazepine use; they may be used in insomnia caused by acute stress, night-shift work change, or air travel. However, we must recognize even here there may be a price to pay in drowsiness the next day, with adverse effects on cognitive performance and driving,(28) and possibly an increased mortality in susceptible patients.(38) Their use even in the treatment of anxiety is problematical. They will make the often-associated underlying depression worse and, according to at least one expert, are unnecessary.(39) A particular difficulty is that the patients most likely to develop dependence problems are those with the personality characteristics that are associated with anxiety and insomnia.(22) That is, the patients with the clearest psychological indications for benzodiazepine therapy are those who are most likely to get into trouble with them. Therefore, other methods of anxiety management, such as relaxation and hypnosis, should be the treatment of choice.

There seems to be little therapeutic justification for the long-term administration of benzodiazepine drugs, apart from their use in some forms of epilepsy. It is my personal opinion that it would be helpful if the profession as a whole could accept the evidence and alter their prescribing habits.

Suggestions for achieving this are:

  • Never prescribe benzodiazepine drugs for anyone with a history of dependence on any drug, including alcohol. (This does not refer to their emergency use, for example, the intravenous administration of diazepam in status epilepticus.)

  • Be very cautious about prescribing benzodiazepine drugs for anyone with a family history of drug or alcohol dependence.

  • Do not prescribe regular benzodiazepine therapy for periods of longer than five days. (This means prescribing fewer than the 50 tablets allowed at present by the PBS.)

  • Warn the patient of the likelihood of rebound insomnia.

  • Warn the patient of the possibility of dependence.

  • Include this warning in the instructions on the bottle.

  • Primum non nocere is an excellent precept, and is often the best we can do. It is unfortunate that we sometimes fail to achieve even that.


  1. Hollister LE, Motzembecker FP, Degan RO. Withdrawal reactions from chlordiazepoxide (Librium). Psychopharmacologia 1961; 2: 63-68.

  2. De la Fuente JR, Rosenbaum AH, Martin HR, Niven RG. Lorazepam-related withdrawal seizures. Mayo Clin Proc 1980; 55: 190-192.

  3. Gordon EB. Addiction to diazepam (Valium). Br Med J 1967; 1: 112.

  4. Ratna L. Addiction to temazepam. Br Med J 1981; 282: 1837-1838.

  5. Selig J. A possible oxazepam abstinence syndrome. JAMA 1966; 198: 951-952.

  6. Anonymous. UN Commission on Narcotic Drugs. Benzodiazepine and pentazocine scheduled under convention on psychotropic substances. Lancet 1984; 1: 637.

  7. Ashton H. Benzodiazepine withdrawal: an unfinished story. Br Med J 1984; 288: 1135-1140.

  8. Woody GE, O'Brien CP, Greenstein R. Misuse and abuse of diazepam: an increasingly common medical problem. Int J Addict 1975; 10: 843-848.

  9. Stitzer MC, Griffiths RR, McLellan AT, Grabowski J. Diazepam use among methadone maintenance patients: patterns and dosages. Drug Alcohol Depend 1981; 8: 189-199.

  10. Schuster CC, Humphries RH. Benzodiazepine dependency in alcoholics. Conn Med 1981; 45:11-13.

  11. Stewart RB, Salem RB, Springer PK. A case report of lorazepam withdrawal. Am J Psychiatry 1980; 137: 1113-1114.

  12. Chandora DB. Delayed diazepam withdrawal syndrome: a case of auditory and visual hallucinations and seizures. J Med Assoc Ga 1980; 69: 769-770.

  13. Athinarayanan P, Pierog SH, Nigam SK, Glass L. Chlordiazepoxide withdrawal in the neonate. Am J Obstet Gynecol 1976; 124: 212-213.

  14. Rementeria JL, Bhatt K. Withdrawal symptoms in neonates from intrauterine exposure to diazepam. J Pediatr 1977; 90: 123-126.

  15. Martin WR, McNicholas LF, Cherian S. Diazepam and phenobarbital dependence in the rat. Life Sci 1982; 31: 721-730.

  16. Rosenberg HC, Chin TH. An antagonist induced benzodiazepine abstinence syndrome. Eur J Pharmacol 1982; 81: 153-157.

  17. Lukas SE, Griffiths RR. Precipitated withdrawal by a benzodiazepine receptor antagonist after 7 days of diazepam. Science 1982; 217: 1161-I163.

  18. Tyrer P, Owen R, Dawlin S. Gradual withdrawal of diazepam after long-term therapy. Lancet 1983; 1: 1402-1406.

  19. Tyrer P, Rutherford D, Huggett T. Benzodiazepine withdrawal symptoms and propranolol. Lancet 1981; 1: 520-522.

  20. Winokur A, Rickiels K, Snyder PJ, Schatz NJ. Withdrawal reaction from long-term, low-dosage administration of diazepam. A double-blind, placebo-controlled case study. Arch Gen Psychiatry 1980; 37: 101-105.

  21. Committee on the Review of Medicines. Systematic review of the benzodiazepine. Br Med J 1980; 280: 910-912.

  22. Tyrer P, Sievewright N. Identification and management of benzodiazepine dependence. Postgrad Med J 1984; 60(suppl 2): 41-46.

  23. Khan A, Joyce P, Jones AV. Benzodiazepine withdrawal syndromes. NZ Med J 1980; 92: 94-96.

« back · top · www.benzo.org.uk »