« back · www.benzo.org.uk »

Long-Term Effects of Benzodiazepine Usage:
Research Proposals, 1995-96

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

Long-term effects of benzodiazepine usage: Assessment of cognitive performance, brain electrophysiological activity and magnetic resonance imaging in chronic benzodiazepine users after withdrawal.


The question of whether long-term use of benzodiazepines may cause persisting cognitive impairment and structural brain changes has long been debated. The aim of this study is to investigate whether long-term prescribed benzodiazepine users who have withdrawn from the drugs for at least one year show persisting impairment of brain function and/or structure. The hypotheses are:

  1. That, compared with controls, such patients will show impairments in cognitive performance, especially visuospatial and attentional ability.

  2. These deficits will be accompanied by changes in EEG evoked potentials including decreased contingent negative variation (CNV) magnitude and increased P300, latency, and increased background slow wave activity.

  3. Magnetic resonance imaging (MRI) may also show structural changes in hippocampal and temporal lobe areas.

The study will detect both structural and functional impairments and show the relationship between them.

In the process of obtaining volunteers from a large bank of known benzodiazepine users who have sought help in withdrawal (see subjects, below), the study will also provide important information on the long-term clinical outcome in this population.


Studies of cognitive function: Several studies have shown that chronic benzodiazepine use is associated with specific cognitive impairments which are different from those seen on acute administration (Hendler et al.1980; Petursson et al. 1983; Lucki et al. 1986; Brosan et al. 1986; Golombok et al. 1988; Lader 1987; Tata et al. 1994). For example, Golombok et al. (1988) tested the performance in a battery of psychomotor tests of 50 patients who had been taking benzodiazepines (up to 30mg diazepam equivalent daily) for at least once year and by control subjects who had not taken benzodiazepines regularly. The cognitive performance of the chronic benzodiazepine users was specifically impaired in two main areas: (a) visuospatial ability and (b) ability to sustain attention in a repetitive task under time pressure. The pattern of impairment was consistent with deficits in posterior cortical (parietal, posterior temporal and occipital) rather than frontal lobe function.

More recently Tata et al. (1994) found a wider range of cognitive impairments in 21 patients who had taken larger doses (10-100mg diazepam equivalent, daily) for a mean of 13.2 years compared with 21 normal control subjects matched for sex, age and I.Q. The benzodiazepine users showed significant impairments in verbal learning and memory, and in psychomotor, visuomotor and visuo-conceptual abilities. The main adverse effects of benzodiazepines on memory and psychomotor performance in this study appeared to implicate functions of the hippocampus and diencephalon/recticular formation. In neither of the studies (Golombok et al. 1988 and Tata et al. 1994) were differences in anxiety levels between benzodiazepine users and controls considered likely to account for the rather specific differences in performance.

While the weight of evidence agrees that chronic prescribed dose benzodiazepine users show cognitive impairments, it has been difficult to ascertain whether or not such impairments persist after benzodiazepine withdrawal. The difficulty has been largely due to methodological problems, such as the use of different benzodiazepines in different doses for different periods of time, and the special difficulty of finding adequate groups for comparison. Tata et al. (1994) addressed some of these problems in a prospective study in which their 21 chronic benzodiazepine users underwent a standard withdrawal regimen from diazepam. They found that the cognitive impairments noted pre-withdrawal were still present (though slightly improved) 6 months after withdrawal. Borg (1987) and Bergman et al. (1989) found slight but incomplete neuropsychological improvement in patients 4-6 years after cessation of benzodiazepine use, but their patients were mainly high dose abusers of benzodiazepines. There are no controlled studies of cognitive function in long-term therapeutic dose benzodiazepine users which extend for more than 6 months after withdrawal, although clinical observations suggest that psychological status in such patients continues to improve only slowly for 12 or more months after withdrawal (Ashton 1991,1994).

Electrophysiological studies: Benzodiazepines cause changes in brain electroencephalographic (EEG) activity. In acute doses they produce a slowing of alpha activity and an increase in fast activity especially in the beta range (14-40 + Hz)(Fink 1969; Itil and Soldatos 1980). In addition they reduce the amplitude of cortical evoked potentials including the contingent negative variation (CNV) and auditory, visual and somatosensory evoked potentials (Saletu et al. 1972; Ashton et al. 1974,1976,1978) and prolong the latency of the auditory P300 potential (Shinotoh et al. 1989).

The changes reverse after withdrawal from short-term medication (Hallström & Lader,1981) but there are no systematic studies on EEG changes during or after withdrawal from long-term use. However, changes in brain electrophysiological activity would be expected in long-term benzodiazepine users with impaired cognitive performance. For example, the magnitude of the CNV is related to the degree of arousal and attention (Tecce et al. 1978) and has been found to correlate with attentional measures on the Cambridge Automated Neuropsychological Test Battery (CANTAB) and also with performance on the Digit Symbol Substitution Test (DSST) (Bahrainian et al. 1994). Similarly, the P300 is thought to reflect stimulus evaluation time (speed of information processing) and its latency correlates with performance in various tests of cognition (Verleger et al. 1991). Increased slow wave EEG activity, in the delta, theta and alpha range, has been found to correlate with reduced cortical evoked potentials and impaired psychomotor performance in patients clinically recovered from depression (Ashton, unpublished observations).

Electrophysiological studies in long-term benzodiazepine users after withdrawal would provide information on the extent and duration of changes in brain activity.

Brain imaging studies: Lader et al. (1984) in a computerised tomography (CT) study reported that some long-term benzodiazepine users had enlarged cerebral ventricles. This finding was not confirmed in two other studies (Poser et al. 1983 and Perera et al. 1987). However, Schmauss and Kreig (1987) found increased ventricular/brain ratios (VBRs) both for high dose (above 50mg diazepam daily) and lower dose chronic benzodiazepine users compared with controls, the greatest increase being found in the high dose patients. Uhde and Kellner (1987) studied patients with panic disorder and found a significant positive correlation between VBR and duration of benzodiazepine use, although the mean VBR of patients fell within the published normal range. Bergman et al. (1989) noted ventricular dilatation on CT in some patients who were sedative/hypnotic abusers, but had not abused alcohol. There are methodological problems, including small sample sizes, in these studies, and the question of whether chronic benzodiazepine use is associated with cortical atrophy or other brain lesions, and if so whether the changes are reversible, remains unclear.

Magnetic resonance imaging (MRI) of the brain offers advantages over CT scanning. The greater resolution of MRI, along with its ability to differentiate white and grey matter, allows better identification and localisation of brain lesions as well as measurements of specific cortical regions and subcortical structures. An MRI study of chronic benzodiazepine users who have stopped taking the drugs after long-term use (compared with controls) would help to show whether or not structural changes occur or persist. No such study has yet been reported.

Importance of study: It is important to establish whether benzodiazepine use leads to long-term functional and/or structural brain changes, and if so to what degree such changes are reversible, for a number of reasons. First, there is a large population of long-term benzodiazepine users in the UK, estimated as at least 1.2 million (Taylor, 1987; Ashton and Golding,1989). It is not known how many of these have subsequently withdrawn. Secondly, as pointed out by Golombok et al. (1988), the cognitive impairments found in long-term users may not only be debilitating but also dangerous with adverse effects on motor car driving and safety at work involving machinery operation. Thirdly, the effects of long-term benzodiazepine use raise clinical problems. Experience with chronic benzodiazepine users both in psychiatric clinics and in general practice suggests that such patients are difficult to treat and their prognosis uncertain. Clinical studies have shown that even after benzodiazepine withdrawal psychological, cognitive and sometimes neurological impairment may persist for long periods in some patients (Tyrer,1991; Ashton 1991,1994) but it is not clear to what extent these symptoms eventually resolve.

The combined use of EEG, CANTAB and MRI will examine the relationship between any structural and functional impairments related to benzodiazepine use.


Subjects: Two large groups of long-term benzodiazepine users on whom accurate records have been kept are available to the principal applicant: (a) patients who attended the applicant's benzodiazepine withdrawal clinic between 1982 and 1994 (total approx. 300 patients), (b) clients who attended the Newcastle Tranquilliser Advice and Support Project between 1984 and 1994 (total approx. 300 clients). Approximately 80% of these patients/clients are known to have withdrawn from benzodiazepines.

As many as possible of these individuals will be contacted and asked about their present health and benzodiazepine use. A pool of suitable subjects will be obtained from this group; subjects for study will be randomly selected from this pool and will consist of:

  1. 50 subjects (both sexes, aged 20-55 years) who have been prescribed benzodiazepines regularly for at least two years in doses of 10-100mg diazepam daily or equivalent and have stopped using benzodiazepines for 1-5 years. (The mean daily dosage in these groups is estimated from previous surveys of the groups mentioned above [Ashton,1984,1987] to have been approximately 30mg diazepam equivalent and mean duration of use over 10 years.).

  2. 50 control subjects, drawn from benzodiazepine users' siblings and spouses, who have never used benzodiazepines regularly, matched for age, sex and (premorbid) I.Q.

Exclusion criteria: History of major depression prior to first benzodiazepine prescription; past or present functional psychosis; history of alcohol or other drug abuse (DSM IV criteria), epilepsy, brain injury, ischaemic heart disease, uncontrolled hypertension.


Psychiatric: Psychiatric assessment using DSM IV criteria, Hospital Anxiety Depression Scale, Spielberger Trait and State Anxiety Scale.

Psychological: I.Q. (National Adult Reading Test), CANTAB, DSST, Trail Making Tests.

EEG: Evoked potentials (CNV and P300), power frequency spectrum.

MRI: Measurements of VBR, hippocampus, parahippocampal gyrus, temporal and frontal lobes, and white matter hyperintensities. All volumetric measurements carried out by coronal inversion recovery using a 0.5 or 1 Tesla machine. White matter intensities measured on axial T2 - weighted spin-echo sequences and graded on a standard scale. Scans examined blindly and independently by the investigator and a consultant neuroradiologist (Dr. V. McAllister).

Urine screen: for presence of benzodiazepines.

Benzodiazepine exposure: A global measure of exposure, taking into account dose and duration of use will be calculated (Golombok et al.1988).

Procedure: Following psychiatric screening, patients would attend on two occasions. The first attendance would be for the EEG recordings and psychological tests at the Department of Psychiatry, Royal Victoria Infirmary. The second attendance would be for the MRI at Newcastle General Hospital.

Analysis of results: Many of the tests, e.g. CANTAB and EEG give automatic computerised results or direct scores at the time of testing. The MRI scans will be analysed independently (see MRI measurements above) by investigators blind to the subject's benzodiazepine status.

The scores for the variables will be compared between subject groups. Correlations will also be sought within the benzodiazepine group between degree of benzodiazepine exposure and cognitive performance, EEG and MRI results. Multiple regression analysis will be performed for the purpose and advice has been taken from the Department of Medical Statistics.

It is estimated from previous studies of evoked potentials and CANTAB scores that the number of subjects in the present investigation (50 in each group) is sufficient to give a statistical power of 85-90% to detect meaningful differences between the groups at the 1% level.


  1. Ashton, H. (1984) Br.med.J. 288: 1135-40.

  2. Ashton, H. (1987) Br.J.Addiction 82: 665-71.

  3. Ashton, H. (1991) J.Sutbstance Abuse Treatment 8: 19-28.

  4. Ashton, H. (1995) Psychiatric Annals 25: 174-179.

  5. Ashton, H., Millman, J.E., Telford, R., Thompson, J.W. (1974) Electroenceph Clin. Neurophysiol. 37: 59-71.

  6. Ashton, H., Millman, J. E., Telford, R., Thompson, J.W. (1976) Brit J Clin Pharmacol 3: 551-559.

  7. Ashton, H., Millman, J.E., Telford, R., Thompson, J.W. (1978) Brit J.Clin Pharmacol 5: 141-147.

  8. Ashton, H., Golding, J.F. (1989) Brit J. Addict 84: 541-546.

  9. Bahrainian, S.A., Ashton, H., Britton, P.G., Ferner, I.N. (1994) J. Psychopharmacol. Abstract 14 of Joint Meeting of British Association for Psychopharmacology and Interdisciplinary Society for Biological Psychiatry, Cambridge 10-13 July 1994.

  10. Bergman, H., Borg, S., Engelbrektson, K., Vikander, B. (1989) Brit J. Addict. 84: 547-553.

  11. Borg, S. (1987) Nord Psykiatr Tidsskr 41: 17-19.

  12. Brosan, L., Broadbent, D., Nutt, D., Broadbent, M. (1986) Psychological Medicine 16: 561-571.

  13. Fink, M. (1969) Ann Rev Pharmacol 9: 241-251.

  14. Golombok, S., Moodley, P., Lader, M. (1988) Psychological Medicine 18: 365-374.

  15. Hallström, C. Lader, M. (1981) Int. Pharmacopsychiat. 16: 235-244.

  16. Hendler, N., Cimini, C., Ma, T., Long, D. (1980) Am J. Psychiatry 137: 828-830.

  17. Itil, T.M., Soldatos, C. (1980) Psychotropic Agents Part I. (eds. Hoffmeister, F., Stille, G.) 437-469.

  18. Lader, M.H., Ron, M., Petursson, H. (1984) Psychological Medicine 14: 203-206.

  19. Lader, M. (1987) The Benzodiazepines in Current Clinical Practice (ed. Freeman, H., Rue, Y.) RSM Int. Congr. & Symp. Series pp.55-70.

  20. Lucki, I., Rickels, K. (1986) Psychopharmacol Bull. 22: 424-433.

  21. Perera, K.M.H., Powell, T., Jenner, F.A. (1987) Psychological Medicine 17: 775-777.

  22. Petursson, H., Gudjonsson, G.H., Lader, M.H. (1983) Psychopharmacology 81: 345-349.

  23. Poser, W., Poser, S., Roscher, D., Argyrakis, A. (1983) Lancet i: 715.

  24. Saletu, B., Saletu, M., Itil, T. (1972) Psychopharmacologia 24: 347-358.

  25. Schmauss, C., Krieg, J-C. (1987) Psychological Medicine 17: 869-873.

  26. Shintoh, H., Iyo, M., Yamada, T., Inoue, 0., Suzuki, K., Itoh, T., Fukuda, H., Yamasaki, T., Tateno, Y., Hirayama, K. (I989) Psychopharmacology 99: 202-207.

  27. Tata, P.R., Rollings, J., Collins, M., Pickering, A., Jacobson, R.R. (1994)Psychological Medicine 24: 203-213.

  28. Taylor, D. (1987) The Benzodiazepines in Current Clinical Practice RSM Int. Congr. Symp. Series. (ed. Freeman, H. & Rue, Y.) pp.13-18.

  29. Tecce, J.J., Savignano-Bowman, J., Meinbresse, D. (1976) Electroenceph Clin. Nettrophysiol. 41: 277- 286.

  30. Tyrer, P. (1991) Stress Medicine 7: 1-2.

  31. Uhde, T.W., Kellner, C.H. (1987) J. Affect. Dis. 12: 175-178.

  32. Verleger, R., Neukäter, Kömpf, D., Viertegge, P. (1991) Electroenceph Clin. Neurophysiol. 79: 488- 502.

The Ashton Manual · Professor Ashton's Main Page

« back · top · www.benzo.org.uk »