« back · www.benzo.org.uk »
Outcome in 50 Patients
British Journal of Addiction
(1987) · 82, 655-671
Professor C Heather Ashton DM, FRCP
School of Neurosciences
Division of Psychiatry
The Royal Victoria Infirmary
Queen Victoria Road
Newcastle upon Tyne NE1 4LP
The Ashton Manual · Professor Ashton's Main Page
Clinical outcome was assessed in 50 consecutive patients who completed a course of supervised benzodiazepine withdrawal following referral to a Clinical Pharmacology Unit. The patients had been taking prescribed benzodiazepines regularly for 1-22 years and all wished to stop. On presentation, all had many symptoms which they attributed to benzodiazepines. The outcome 10 months to 3.5 years later was judged as excellent (fully recovered) in 48%, good (much better) in 22%, moderate (better) in 16% and poor (no better) in 6%. One patients failed to withdraw and three relapsed onto benzodiazepine use after withdrawal. Younger age was significantly associated with a favourable outcome, but outcome was not related to duration or dosage of benzodiazepines, type of benzodiazepine, rate of withdrawal, symptom severity, psychiatric history, marital status or sex.
Dependence upon prescribed benzodiazepines is now recognized as a major clinical problem[1-4] involving perhaps half a million patients in Britain.[5-8]. There is growing agreement with the view that "most patients currently taking benzodiazepines should stop them". However, little information is available on the outcome of withdrawal, and no large-scale, long-term results have been published.
The reported outcome of benzodiazepine withdrawal in patients referred to psychiatric clinics appears discouraging. Higgitt et al., from a review of the literature and personal experience of 60 patients over 7 years, concluded that only one third of patients are free of problems after withdrawal. Of the remaining two thirds, about 50% need antidepressants and many return to benzodiazepine use. These results may not be representative of the whole population concerned. It seems likely that the majority manage to stop benzodiazepine on their own, or with help from their general practitioners and/or tranquilliser support groups.. However, there have been no investigations on how this group has fared.
Yet is it clearly important to know the outcome of benzodiazepine withdrawal. Is it worth spending time and trouble on benzodiazepine withdrawal if most patients relapse back onto benzodiazepines, develop psychiatric problems, or need treatment with other more expensive and more toxic drugs?
In view of the dearth of information, it is relevant to report the intermediate-term outcome of a third population benzodiazepine-dependent patients who were referred to a Clinical Pharmacology Unit. This report is not intended as a controlled study of withdrawal symptoms or methods; indeed, these questions are largely irrelevant to the issue of outcome, since in practice patients withdraw in a variety of ways and suffer a range of transient or prolonged withdrawal symptoms[1,3,4,5,10]. Instead, it records the clinical outcome in 50 consecutive patients who attempted benzodiazepine withdrawal and were followed up for 10 months to 3.5 years and examines some variables which might be expected to influence outcome.
Some details of the patients are given in Table 1 (Patients 1-12 are the same as those described previously 1-6 months after benzodiazepine withdrawal).. There were 40 females and 10 males, aged 20-72 years (mean age 45.92 years). The patients had been taking prescribed benzodiazepine for 1-22 years (mean duration 9.76 years). Benzodiazepine dosages at the start of withdrawal varied from 4mg diazepam daily to 5mg lorazepam plus 40mg diazepam daily (approximately equivalent to 80mg diazepam daily). The patients themselves all wished to stop taking benzodiazepines, to which they attributed adverse effects. None abused alcohol or other drugs. All except one were successful in withdrawing completely. Three patients relapsed after withdrawal. One patient committed suicide and two died of unrelated causes following withdrawal. The remaining 43 patients have been off benzodiazepines for over 3 years (10 patients), 2-3 years (7 patients), 1-2 years (20 patients) and 10-12 months (6 patients).
History and Symptoms on Presentation
All the patients had symptoms on presentation. These symptoms were the reason for referral and included the full range of psychological and somatic symptoms described previously in relation to chronic benzodiazepine use and withdrawal.[3,4]. Most had made previous attempts at dosage reduction or complete withdrawal and had found that this exacerbated their symptoms.
Several features of interest confirm and extend the findings of a previous report. Ten of the present series of patients, while on chronic benzodiazepine medication, had taken drug overdoses requiring hospital admission (sometimes on several occasions); only two of these had a definite history of depressive illness before they were prescribed benzodiazepines, although several had been prescribed antidepressant and other psychotropic drugs while on benzodiazepines. Ten had incapacitating agoraphobia which first developed after several years on benzodiazepines. Nine patients had undergone gastroenterological investigations for gastrointestinal symptoms which had ultimately been attributed to "irritable bowel", diverticulitis or hiatus hernia. Three had been referred for neurological investigations and had been told that they had multiple sclerosis (not subsequently confirmed). Although most patients complained of paraesthesiae in the extremities associated with panic attacks, two women aged 64 and 63 (Nos. 29, 44, Table 1) had constant severe burning pain in the hands and feet which regressed slowly after benzodiazepine withdrawal. It was also of interest that triazolam taken chronically in doses of only 0.25mg nightly provoked typical symptoms (including aggressiveness, hallucinations, poor memory and concentration, paraesthesiae, panic attacks and headaches) in one patient (No. 25, Table 1). In this case these symptoms subsided almost immediately on stopping the triazolam.
Management of withdrawal was individually tailored for each patient. The duration of the withdrawal period ranged from one week in a patient taking triazolam (0.25mg nocte) to 15 months in a patient taking a combination of lorazepam 5mg and ketazolam 30mg daily. The general procedure was to change all patients taking other benzodiazepines to an approximately equivalent dose of diazepam and then to reduce diazepam dosage gradually. One patient (No. 34, Table 1) found it impossible after three attempts, to change from lorazepam (1.5mg daily) to diazepam. "Equivalent" doses of diazepam[8,11] whether introduced slowly or rapidly, paradoxically induced a state of acute anxiety and panic combined with extreme drowsiness. This patient was finally withdrawn directly from lorazepam by 0.125mg decrements.
Various drugs were used temporarily in individual patients for symptomatic control, as described previously.[3,8]. These included propranolol, non-benzodiazepine hypnotics, tricyclic antidepressants, clonidine, and analgesics. These drugs helped to control certain individual symptoms but were not successful in alleviating the withdrawal syndrome as a whole. Many patients (19.38%) took no additional drugs during the course of withdrawal.
Apart from pharmacological advice, the mainstay of management was the provision of frequent consultations and repeated encouragement. About 60% of the patients also attended a tranquilliser support group.
In general, the period of benzodiazepine dosage reduction caused only slight exacerbation of symptoms already present before withdrawal. After withdrawal, the clinical course was protracted; symptoms persisted for over a year in some patients, though diminishing in intensity. Depression was common during this period. One patient committed suicide; three were diagnosed by psychiatrists as having major depressive disorders, and 17 were prescribed antidepressant drugs (usually for not more than 3-6 months). However, the depression lifted in most patients within a year and none have taken drug overdoses since withdrawal. Gastrointestinal complaints were also frequent but declined gradually in most patients including those who had "irritable bowel" for years. Similarly, agoraphobic symptoms abated dramatically within a year of withdrawal, even in patients who had previously been housebound, and none were incapacitated by agoraphobia at the time of follow-up.
As an interim evaluation of the results of benzodiazepine withdrawal, the outcome was assessed at an arbitrary time in the first 50 patients attending the clinic. Outcome was graded into categories, using the following criteria: Excellent (minimal symptoms, leading a normal life, in full-time employment where applicable, taking no regular medication); Good (some symptoms but able to lead a normal life and/or cope with a full-time job without regular medication); Moderate (better, but symptoms which interfere with life or require other drugs, such as beta-blockers or antidepressants, still present); Poor (off benzodiazepine but not improved, polysymptomatic and/or needing other psychotropic medication); Failed (relapsed or unable to withdraw benzodiazepines).
The results are shown in Table 2. The outcome was graded as excellent in 24 patients (48%), good in 11 (22%), moderate in 8 (16%), poor in 3 (6%) and failed in 4 (8%). Thus, 70% of this group did excellently or well after withdrawal and a further 16% were moderately improved. All these patients claimed to feel better after withdrawal than when they were taking benzodiazepines and were glad they had withdrawn. One patient failed to achieve complete withdrawal and three relapsed onto regular benzodiazepine usage, one after a year and two after a month of abstinence. Five other patients (Nos. 11, 16, 38, 46, 50, Table 1) temporarily took benzodiazepines 1-3 months after withdrawal, but all were successful at a second attempt and achieved final outcomes rated as excellent (1), Good (3), and moderate (1).
Relationships Between Outcome and Other Variables
In the 46 Patients who were successful in withdrawal, younger age was highly significantly associated with a favourable outcome. The mean age of the 35 patients whose outcome was judged good or excellent was 43.4 years (SD 11.1) while the mean age of the 11 with a moderate or poor outcome was 58.0 years (SD 13.0) (d.f.=44, t=3.65, p<.001, two tailed t-test). Nevertheless, the four patients who failed withdrawal or relapsed were all in the younger age group (mean age 34.5 years), while two patients aged 64 achieved a good outcome and one aged 69, rated as excellent, was able to continue a full-time career as a distinguished author.
Somewhat surprisingly, outcome of withdrawal did not appear to be related to the duration or age of onset of benzodiazepine usage, dosage at the time of withdrawal, type of benzodiazepine, rate of withdrawal, severity of symptoms, marital status, or sex. No clear relationship was found between outcome and psychological factors. The four patients rated as failed and two of the three rated as poor had histories of psychiatric disorder but this was also true of nearly 30% of patients who achieved a good or excellent outcome. Since most of the patients had been on benzodiazepines for long periods, it was difficult to obtain a reliable estimate of pre-benzodiazepine personality. The failures had all been labelled 'inadequate personality' during the course of benzodiazepine use, but so had several of the patients who achieved an excellent result.
A clear result of the present study is that a substantial majority (70%) of these patients were doing well 10 months to 3.5 years after benzodiazepine withdrawal and the rate of relapse and of failure to withdraw was low (8%). The more favourable outcome compared to that of Higgitt et al., who found that only a third of patients were free of problems after withdrawal, may have been due to patient selection. Patients with psychiatric problems are presumably referred mainly to psychiatric departments, and most of the available data has been obtained from such patients. Patients referred to clinical pharmacology departments may have a better prognosis because they have fewer underlying psychiatric problems. (It is worth noting, however, that some patients in the present series were referred by psychiatrists.)
Secondly, the study shows that long-term benzodiazepine use is associated with a considerable morbidity. While on benzodiazepines, 10 of the 50 patients had taken drug overdoses requiring hospital admission, sometimes on several occasions. Ten patients had become agoraphobic, for which several had received (unsuccessful) behavioural and other therapy. Twelve had undergone extensive investigations in gastroenterology or neurology departments and treatment had been ineffective. None of these symptoms or behaviours were the original indication for starting on benzodiazepines but developed during chronic use. It is arguable whether the patient would have developed the symptoms over time in the absence of benzodiazepines, but the fact that they were not present before benzodiazepine use, were not amenable to treatment during benzodiazepine use, yet largely disappeared when the drugs were stopped, suggests that benzodiazepines may actually cause or aggravate a variety of psychological and psychosomatic problems.
It was difficult to quantify psychological factors leading to the initiation and continuation of benzodiazepine use. As shown in Table 1, the original indication for benzodiazepines was usually, though not exclusively, anxiety or depression. Nevertheless all patients had a variety of anxiety/depressive symptoms on presentation, and these had been gradually increasing over the years despite continuous benzodiazepine use. In the large majority of patients, these symptoms greatly improved after withdrawal. Although many patients, even those whose outcome was classed as excellent after 2-3 years, still have occasional anxiety symptoms, especially during periods of stress, they appear now to have learned to cope with stress better than when on chronic benzodiazepine medication. A large-scale, long-term comparison of anxious patients who were or were not prescribed benzodiazepines would be needed to clarify this issue. The study of Catalan & Gath suggested that benzodiazepines were not more effective than brief counselling for patients with minor affective disorders seen in general practice and followed for over 7 months.
It seems likely that the mechanisms for the worsening of symptoms during chronic benzodiazepine use include the development of tolerance to the anxiolytic effects, so that withdrawal symptoms emerge even in the continued presence of the drugs, and the onset of long-term adverse effects. These effects and mechanisms have been discussed elsewhere.[3,4]
Finally, the results show that it is worthwhile withdrawing benzodiazepines in motivated patients such as those in the present study. Depression may be a problem after withdrawal as previously noted[3,12] and as evidenced by one suicide and three cases of major depression in this group. Antidepressant drugs may be indicated temporarily in such cases. In general, the outcome appears to be better in younger patients, but age is no absolute bar to success. For the future, prevention of benzodiazepine dependence is clearly a preferable strategy and may be achieved by more thoughtful prescribing of benzodiazepines and by reserving them for short-term use.
I thank Professor M.D. Rawlins for helpful advice and clinical facilities.
TABLE 1. Details of Patients
Sex Age Original
at presentation (mg)
Outcome* 1 F 34 Menorrhagia 14 Nitrazepam 10
- Relapsed 2 M 31 Tinnitus 3 Lorazepam 3 42 Excellent 3 F 43 Anx/depression 10 Lorazepam 2 39 Excellent 4 F 39 Hyperactive child 12 Diazepam 4 42 Good 5 F 50 Dizzy turns 10 Diazepam 10 39 Excellent 6 F 46 Post-op shock 10 Clobazam 20
39 Excellent 7 M 72 Backache 5 Lorazepam 7.5 36 Moderate 8 F 54 Anx/depression 22 Diazepam 20 36 Good 9 F 42 Neck pain 18 Prazepam 5 36 Excellent 10 F 67 Depression 4 Lorazepam 1.5 36 Moderate 11 F 36 Influenza 3 Lorazepam 3
36 Excellent 12 F 46 Headaches 14 Lorazepam 7 - Relapsed 13 F 43 Anxiety 6 Diazepam 7.5 30 Moderate 14 F 38 Anx/depression 4 Chlordiazepoxide 20 30 Excellent 15 F 38 Panic attacks 3 Diazepam 30
30 Excellent 16 F 64 Nerves 16 Chlordiazepoxide 30 30 Good 17 M 38 Headaches 5 Medazepam 5 28 Excellent 18 F 56 Nervous breakdown 10 Lorazepam 2.5 24 Excellent 19 M 40 Anxiety 1 Triazolam 0.5 24 Excellent 20 F 45 Nerves 2 Lorazepam 1.5 23 Excellent 21 F 47 Anx/depression 19 Diazepam 6 19 Excellent 22 F 56 Anx/depression 7 Diazepam 10 18 Moderate 23 M 34 Nerves 16 Diazepam 8 22 Excellent 24 F 20 Compound tic 11 Lorazepam 7.5 29 Excellent 25 F 59 Bereavement 10 Triazolam 0.25 22 Excellent 26 F 34 Nerves 12 Lorazepam 7.5 21 Poor 27 M 28 Phobic Anxiety 13 Lorazepam 4.5 20 Excellent 28 M 31 Palpitations 10 Diazepam 12 21 Good 29 F 64 "Run down" 12 Diazepam 10
17 Good 30 F 40 Bereavement 60 Diazepam 15 16 Excellent 31 F 55 Depression 3 Temazepam 60 11 Good 32 F 28 Bereavement 13 Diazepam 10
- Failed to
33 F 38 Headaches 3 Lorazepam 1 16 Good 34 M 30 Nervous breakdown 10 Lorazepam 2 11 Excellent 35 F 46 Nervous breakdown 12 Diazepam 6
11 Moderate 36 M 70 Bereavement 5 Lorazepam 3 14 Moderate 37 F 33 Anxiety 21 Lorazepam 5
12 Good 38 F 40 S/e anorectics 20 Diazepam 7 16 Good 39 F 36 Bereavement 18 Diazepam 10
12 Good 40 F 30 Depression 14 Diazepam 25 - Relapsed 41 F 56 Bereavement 14 Diazepam 4.5
12 Excellent 42 F 71 Anx/depression 5 Lorazepam 1 13 Moderate 43 F 49 Bereavement 3 Diazepam 15
12 Excellent 44 F 63 Nervous breakdown 12 Medazepam 15
11 Moderate 45 F 44 Anxiety 7 Lorazepam 3 14 Excellent 46 M 69 Anxiety 8 Lorazepam 2 3 Excellent** 47 F 66 Anx/depression 2 Alprazolam 0.5
2 Poor** 48 F 50 Nerves 3 Lorazepam 3 - Poor*** 49 F 43 Depression 12 Diazepam 30 10 Excellent 50 F 44 Backache 15 Lorazepam 3
11 Good Total/
- - Excellent-24
* See Table 2 for definitions of outcome
** Died of unrelated illness after withdrawal
TABLE 2. Outcome after Benzodiazepine Withdrawal
Grading Definition Number of
(%) Excellent Fully recovered - Minimal symptoms, leading normal life,
full-time job, no regular medication
(may still be 'highly strung').
24 (48%) Good Much better - Some symptoms but able to lead normal
life, full-time job.
11 (22%) Moderate Better - Coping but symptoms which interfere with
life or require other drugs (e.g. betablockers,
antidepressants) still present.
8 (16%) Poor No Better - Off benzodiazepines but still polysymptomatic
and/or needing other psychotropic medication (e.g. antidepressants, sedative/hypnotics).
3 (6%) Failed Relapsed, unable to withdraw - Started benzodiazepines again after
withdrawal, unable to withdraw, multiple
symptoms still present.
Petursson, H. & Lader, M.H. (1981) Withdrawal from long-term benzodiazepine treatment, British Medical Journal, 283, pp. 643-5.
Owen, R.T. & Tyrer, P. (1983) Benzodiazepine dependence: a review of the evidence, Drugs, 25, pp. 385-98.
Ashton H. (1984) Benzodiazepine withdrawal: an unfinished story, British Medical Journal, 288, pp. 1135-40.
Ashton, H. (1986), Adverse Effects of Prolonged Benzodiazepine Use, Adverse Drug Reaction Bulletin, 118, pp. 440-3.
Tyrer, P., Owen, R. & Dawlings, S. (1983) Gradual withdrawal of diazepam after long-term therapy, Lancet, i, pp. 1402-6.
Hallstrom, C & Lader, M.H. (1982) The incidence of benzodiazepine dependence in long-term users, Journal of Psychiatric Treatment and Evaluation, 4, pp. 293-6.
Balter, M.B., Manheimer, D.I., Melinger, G.D. & Uhlenhuth, E.H. (1984) A cross-national comparison of antianxiety/sedative drug use, Current Medical Research Opinion, 8 (Suppl. 4), pp. 5-20.
Higgitt, A.C. Lader, M.H., & Fonagy, P. (1985) Clinical Management of benzodiazepine dependence, British Medical Journal, 291, pp. 688-90.
Trickett, S. (1986) Coming off Tranquillisers (Northampton, Thorsonsís).
Busto, U., Sellers, E.M., Naranjo, C.A., Cappell, J., Sanchez-Craig, M. & Sykora, K.(1986) Withdrawal reaction after long-term therapeutic use of benzodiazepines, The New England Journal of Medicine, 315, pp. 854-9.
Benzodiazepine dependence and withdrawal - an update, (1985) Drug Newsletter, 31, pp. 125-8.
Catalan, J. & Gath, D.H. (1985) Benzodiazepines in general practice: time for decision, British Medical Journal, 290, pp. 1374-6.
Lader, M.H. & Higgitt, A.C. (1986) Management of benzodiazepine dependence - update 1986, British Journal of Addiction, 81, pp. 7-10.
The Ashton Manual · Professor Ashton's Main Page« back · top · www.benzo.org.uk »