British Journal of General Practice, October 2002 805
Differences in health status between long-
term and short-term benzodiazepine users
S M Zandstra, J W Furer, E H van de Lisdonk, J H J Bor, F G Zitman and C van Weel
intervention (maximum of 8 to 12 weeks), nevertheless 1.7%
to 4.9% of the population receive benzodiazepine prescrip-
tions for more than six months.4-12Apparently, despite the
fact that the initial benzodiazepine prescription is the result
of carefully weighing benefits against risks, some patients
end up as long-term users.
Specific patient characteristics could be responsible for
long-term use, and this is supported by the literature. Based
on the comparison of long-term users with non-users, the
following health status factors were related with long-term
benzodiazepine use: older age, psychological problems
(more), physical disease (more), and sex (female).8,11,13-21It
is unclear however, if this refers specifically to long-term
benzodiazepine users or to benzodiazepine use in general.
Only a few studies have compared long-term with short-term
benzodiazepine users: they were older, had poorer health,
more depression, a higher daily benzodiazepine use and
got their prescription more often from a hospital physi-
An alternative explanation for long-term use that is patient-
unrelated would be prescription failures. A minimal failure to
discontinue initial benzodiazepine use (for example, in 1% of
prescriptions) would cumulate in a prevalence of long-term
use of 1 per 1000 patients per year (under a prescription rate
of 10%12,24). Prescription failure is plausible, given that most
repeat prescriptions are provided without a doctor–patient
encounter25,22As a consequence a proportion of the original
short-term benzodiazepine users become long-term users.
Further insight into benzodiazepine use is needed to devel-
op a tool for prevention of long-term use. Therefore, this
study compared the physical and mental health status of
long-term and short-term benzodiazepine users.
ONG-TERM benzodiazepine use is a precarious theme in
general practice. The guidelines1-3recommend short
The design of this study was a cross-sectional comparison
of short-term and long-term benzodiazepine users in gener-
al practice. Data were used from the practices of the
Nijmegen Health Area Project-2, a study on psychopatholo-
gy in the general population.26Benzodiazepine users were
identified from the practices’ prescription files.
Definition criteria for benzodiazepine users
Under the Dutch health insurance guidelines, benzodi-
azepines (defined according to the standardised classifica-
tion system for drugs of the WHO using the Anatomical
Therapeutic Chemical classification — index groups N05BA,
CD, CF and CG;27N = nervous system N05 = psycholep-
S M Zandstra, MD, general practitioner;; E H van de Lisdonk, MD,
PhD, general practitioner; J H Bor, BSc, statistician; and C van Weel,
MD, PhD, professor of general practice, Department of General
Practice; J W Furer, MA, PhD, social psychologist, Department of
Social Medicine; University Medical Centre, St Raboud, Nijmegen,
the Netherlands. F G Zitman, MD, PhD, professor of psychiatry,
Department of Psychiatry, University Medical Centre, Leiden, the
Address for correspondence
S M Zandstra, Internal postal code 229 HSV-SG, PO Box 9101,
6500 HB Nijmegen, The Netherlands.
Submitted: 3 September 2001; Editor’s response: 19 November
2001; final acceptance: 8 April 2002.
© British Journal of General Practice, 2002, 52, 805-808
Background: Despite generally accepted advice to keep treat-
ment short, benzodiazepines are often prescribed for more than
six months. Prevention of long-term benzodiazepine use could be
facilitated by the utilisation of risk indicators for long-term use.
However, the characteristics of long-term benzodiazepine users
described in the literature are based on studies in which long-
term users were compared with non-users. Thus these character-
istics may be imprecise.
Aim: To study the characteristics of long-term benzodiazepine
users by comparing their demographic data and health status
(mental and physical) with those of short-term users.
Design of study: Cross-sectional comparison of short-term and
long-term benzodiazepine users.
Setting: Patients from 32 GP practices of the Nijmegen Health
Area, The Netherlands.
Method: The characteristics of 164 short-term and 158 long-
term benzodiazepine users in general practice were compared,
using interview data and morbidity, referral and prescription
data from GP records.
Results: Long-term benzodiazepine users were (a) older, (b)
had a more severe history of mental health problems for which
they had received more serious treatment, (c) used more psy-
chotropic drugs, (d) had a higher hospital specialist consultation
frequency, (e) had more diagnoses of the following: diabetes,
asthma, chronic obstructive pulmonary disease, hypertension, a
serious skin disorder, and (f) reported a lower perceived general
health status. There were no sex differences.
Conclusion: Specific risk characteristics of long-term benzodi-
azepine users can be used to develop a risk profile for the man-
agement of benzodiazepines in general practice. We believe that
(somatic) secondary care also contributes to benzodiazepine use.
It may be worthwhile to coordinate care for benzodiazepine users
between GPs and hospital specialists.
Keywords: benzodiazepines; prescription failure; health status.
tics, N05B = anxiolytics, N05C = hypnotics and sedatives,
N06 = psychoanaleptics) have to be prescribed by a med-
ical practitioner. Short-term benzodiazepine users were
defined as having prescriptions for less than or equal to
90 days (the maximum prescription advice of the WHO) with
the exclusion of patients who finished taking the drug at the
beginning of the measurement period or who started at the
end of that period, because of uncertainty about their user
pattern. Long-term benzodiazepine users were defined as
having prescriptions for more than or equal to 180 days.
Patients with 91 to 179 days’ use were excluded from this
Practices and patients
General practitioners (GPs) had to be using a computerised
patient and medication registration system. In total, 64 prac-
tices were approached and 32 agreed to participate.
Practices’ main reason for non-participation was involve-
ment in other research; details of approach and representa-
tiveness of the general practices and the population sam-
ples have been described elsewhere.24,26The general prac-
tice patients’ lists comprised 80 315 persons aged between
18 and 74 years, of whom 4% were short-term and 2% long-
term benzodiazepine users.24An equal number of short-
term and long-term benzodiazepine users were recruited
from each practice to eliminate effects related to a GP’s work
style. This resulted in the selection of 164 short-term users
and 158 long-term benzodiazepine users for the study.
There were no sex differences between the groups selected
for this study or between all the short and long-term benzo-
diazepine users at the practices. With respect to age, the
short-term benzodiazepine user participants were older than
all the short-term users at the practices (participants aged
over 45 years = 63.4%, versus practices = 53.8%).24
Health status was measured using:
1. the 24-item chronic diseases list of the CBS in the past
year based on the Health Survey of Statistics
Netherlands,28the Short Form 36-item Health Survey
(GHQ30),31,32and the Four Neurotic Symptoms (4-NS)
2. a structured interview concerning the patient’s history of
mental health problems and details about treatment,
hospitalisation for psychiatric diseases, drug and alco-
hol abuse and treatment. Patients were also asked
about their use of over-the-counter (OTC) psychotropic
drugs in the past four weeks and the frequency with
which they had consulted a hospital specialist (exclud-
ing visits to an ophthalmologist). The GP records pro-
vided data on the number of consultations and the psy-
chotropics prescribed during the past year. The psy-
chotropics were defined using Anatomical Therapeutic
Chemical classification codes27: N05 and N06, (N06 =
psychoanaleptics) with the exception of the benzodi-
azepines named earlier.
The recruited sample was split in two parts so that the risk
profile found in the first part (split 1) could be validated in the
second part (split 2).34To detect differences of at least 16%
between long-term and short-term benzodiazepine users,
two groups of 95 subjects were necessary (if α = 0.05 and
β = 0.20) or the first split. All subjects were stratified for
each practice. Subsequently univariate procedures (χ2-test,
t-test) on single variables were used to select significant vari-
ables for the logistic regression, resulting in a risk profile.
The goodness-of-fit method was used,34testing observa-
tions from the second split (subsample of 69 short-term and
63 long-term benzodiazepine users) with the findings of the
first split. Analyses were done with the SAS statistical soft-
The long-term users were older (mean age = 56.8 years ver-
sus 48.5 years), but there was no sex difference between the
two groups. Many long-term and short-term users had a his-
tory of mental health problems (48% and 42%, respectively).
However, more long-term benzodiazepine users had been
treated by a psychiatrist and had a history of alcohol abuse.
As a large proportion of the patients with alcohol problems
also had psychiatric problems, we regarded these two
aspects as a history of treatment in secondary health care —
35.8% long-term users, 14.7% short-term users. In contrast,
more short-term benzodiazepine users had been treated for
psychological problems in primary care — 27.4% versus
Long-term benzodiazepine users had received more anti-
depressants and antipsychotics (32.6% versus 13.7%) in the
past year and visited a hospital specialist more often (mean
number of visits = 2.4 versus 1.3). Both groups reported one
or more chronic disease in more than 80% of cases, but
among long-term benzodiazepine users there was signifi-
cantly more diabetes, asthma or chronic obstructive pul-
monary disease (COPD), hypertension or a serious skin dis-
order (combined in 50.5% versus 24.2% of short-term
806 British Journal of General Practice, October 2002
S M Zandstra, J W Furer, E H van de Lisdonk, et al
HOW THIS FITS IN
What do we know?
Long-term use of benzodiazepines is
common, but should be discouraged, given
the harmful side effects. Benzodiazepine use is
particularly common among the elderly, women patients with
psychological problems, and chronic physical diseases.
However, it is unclear whether this profile singles out the
group at risk for long-term use.
What does this paper add?
This paper reports a comparison of long-term users with short-
term users. The findings are that patients — in particular those
who use other psychotropic drugs or are under regular hospi-
tal care of specialists of physical diseases, and those with
poor quality of life and high medical consumption — are using
benzodiazepines on a long-term basis. The involvement of
secondary care is remarkable. Insight into this profile can help
GPs to forestall benzodiazepine use or wean patients off them
at an early stage.
users). Therefore ‘chronic diseases’ were taken as one cat-
egory in the logistic regression risk profile.
Long-term benzodiazepine users had poorer perceived
general health (mean GHQ score = 56.8 versus 65.9), phys-
ical functioning (mean SF-36 score = 78.3 versus 85.8) and
mental health (mean 4-NS score = 64.3 versus 71.4).
Otherwise, the scores on SF-36 GHQ30and 4-NS were ele-
vated in both groups. Also, reported OTC psychotropic use
and the GP consultation frequency were similar.
All the above-named significant variables of the first split
were entered into the logistic regression. Controlling for
redundancy in logistic regression eliminated perceived gen-
eral health and physical functioning. We controlled for con-
founding of age, but all the variables were independent.
Because the outcome of the second split logistic regression
was comparable with that of the first (AUC1 = 0.77 ± 0.7;
AUC2 = 0.80 ± 0.8) Table 1 gives the logistic regressions of
both splits combined — older age, a history of psychiatric
treatment, and chronic illness were independently related to
long-term benzodiazepine use.
The study provided the profile of long-term benzodiazepine
users; they were older, had a more severe psychiatric histo-
ry, had frequent prescriptions of psychotropics, and poorer
mental health status, with more common chronic diseases
and more visits to medical specialists.
Not included in the analysis were socioeconomic status,
coping behaviour or GPs’ work style. In fact the long-term
users we studied were long-term users despite the efforts of
their GP to use benzodiazepines in a more appropriate,
short-term or intermittent way. As we did not study the
GP–patient interaction in benzodiazepine use and prescrib-
ing we are not able to comment on this in depth.
Our findings differ from others with respect to sex as well
as showing some unexpected similarities,8,11,14-21in all prob-
ability due to our comparison with short-term benzodi-
azepine users rather than non-users. In particular this was
the case for the elevated scores on SF-36 and GHQ30,
indicating that long- and short-term benzodiazepine users
have many psychiatric symptoms and a broad range of dys-
functions. The relationship between long-term benzodi-
azepine use and common chronic diseases was interesting.
A possible explanation for this is that these are all diseases
with highly protocolised35-38treatments that require frequent
visits to primary and secondary care physicians. As the sec-
ond split confirmed the findings of the first split, we are con-
fident that these characteristics are indeed specific for long-
term benzodiazepine users.
Frequent visits to a physician in themselves increase the
chance of starting prescriptions for drugs like benzodi-
azepine. Moreover, treatment started in secondary care may
be continued in primary care. This is in line with the report-
ed role of (somatic) hospital specialists23as well as the
described role of follow-up prescriptions in general prac-
tice.22,25. Given the disadvantages of long-term benzodi-
azepine use (cognitive39and sedative3,20-23effects) and their
consequences40and given that stopping benzodiazepine
therapy is a problem for many patients,41-43it is important to
prevent long-term use.
Recapitulating, we found specific risk indicators for long-
term benzodiazepine use, and so patient-unrelated pre-
scription failures are an improbable cause of long-term use.
Many of the risk indicators we found suggest the involve-
ment of secondary care, which can only be dealt with by
accurate communication and coordination of the various
disciplines by the GP . More research will provide greater
insight into the role of these indicators in creating long-term
benzodiazepine users. Further exploration of differences
between short- and long-term benzodiazepine users is
desirable (for example, coping). The findings of this study
will be used in developing GP support to counter long-term
benzodiazepine use in a more effective way.
1. Committee on the Safety of Medicines (CSM). Benzodiazepines,
dependence and withdrawal symptoms. Curr Probl 1988; 21: 1-2.
2. Salzman C. The APA Task Force report on benzodiazepine depen-
dence, toxicity, and abuse. [Editorial.] Am J Psychiatry 1991; 148:
3. Gabe J. Promoting benzodiazepine withdrawal. Addiction 1994;
4. Balter MB, Manheimer DI, Mellinger GD, Uhlenhuth EH. A cross-
national comparison of anti-anxiety/sedative drug use. Curr Med
Res Opin 1984; 8 (Suppl 4): 5S-20S.
5. Barbui C, Gregis M, Zappa M. A cross-sectional audit of benzodi-
azepine use among general practice patients. Acta Psychiatr
Scand 1998; 97:153-156.
6. Magrini N, Vaccheri A, Parma, et al. Use of benzodiazepines in the
Italian general population: prevalence, pattern of use and risk fac-
tors for use. [See comments.] Eur J Clin Pharmacol 1996; 50:19-
7. Mellinger GD, Balter MB, Uhlenhuth EH. Anti-anxiety agents: dura-
tion of use and characteristics of users in the USA. Curr Med Res
Opin 1984; 8 (Suppl 4): 21S-36S.
8. Mellinger GD, Balter MB, Uhlenhuth EH. Prevalence and corre-
British Journal of General Practice, October 2002807
Table 1. Logistic regression of the health status factors of long-term benzodiazepine users compared with short-term users (first split and
second split combined, missing n = 3).
interval (CI)Logistic regression risk profile
Age >50 years
Frequency of consulting specialist in past half year >1
History of treatment in secondary care (psychiatrist) and/or treatment for alcohol abuse
Diabetes mellitus or asthma or COPD or hypertension or a serious skin disorder
Mental Health (SF-36)
Prescribed psychotropics (antidepressives and antipsychotics)
lates of the long-term regular use of anxiolytics. JAMA 1984; 251:
9. Ohayon MM, Caulet M, Priest RG, Guilleminault C. Psychotropic
medication consumption patterns in the UK general population. J
Clin Epidemiol 1998; 51: 273-283.
10. Rodrigo EK, King MB, Williams P . Health of long-term benzodi-
azepine users. BMJ 1988; 296: 603-606.
11. Salinsky JV, Dore CJ. Characteristics of long-term benzodiazepine
users in general practice. J R Coll Gen Pract 1987; 37: 202-204.
12. Van Hulten R, Leufkens H, Bakker A. Usage patterns of benzodi-
azepines in a Dutch community: a 10-year follow-up. Pharm World
Sci 1998; 20: 78-82.
13. Wells K, Kamberg C, Brook R, et al. Health status, sociodemo-
graphic factors, and the use of prescribed psychotropic drugs.
Med Care 1985; 23: 1295-1306.
14. Ashton H, Golding JF. Tranquillisers: prevalence, predictors and
possible consequences. Data from a large United Kingdom sur-
vey. Br J Addict 1989; 84: 541-546.
15. Blijenberg-Ruis B, Garretsen HFL, Schuurman JH, Verdonk ALT.
The use of sleep medication and tranquilizers. Tijdschrift voor
Sociale Gezondheidszorg 1986; 64: 491-497.
16. Catalan J, Gath DH, Bond A, et al. General practice patients on
long-term psychotropic drugs. A controlled investigation. Br J
Psychiatr 1988; 152: 399-405.
17. Gene BJ, Blay PC, Soler VM. Risk factors in the use of benzodi-
azepines. Fam Pract 1988; 5: 283-288.
18. Lagro-Janssen ALM, Liberton ILW. Profiles of regular consumers
of benzodiazepines in a general practice. Nederlands Tijdschrift
voor Geneeskunde 1993; 137: 1969-1973.
19. Mellinger GD, Balter MB, Manheimer DI, et al. Psychic distress, life
crisis, and use of psychotherapeutic medications: national house-
hold survey data. Arch Gen Psychiatr 1978; 35: 1045-1052.
20. Simpson RJ, Power KG, Wallace LA, et al. Controlled comparison
of the characteristics of long-term benzodiazepine users in gener-
al practice. [See comments.] Br J Gen Pract 1990; 40: 22-26.
21. Uhlenhuth EH, Balter MB, Lipman RS. Minor tranquilizers: clinical
correlates of use in an urban population. Arch Gen Psychiatr
1978; 35: 650-655.
22. Vissers FHJA, Jongbloet AH, Knottnerus JA. Benzodiazepinen: de
gebruiker en zijn huisarts. [Benzodiazepines, the user and his
general practitioner.] Huisarts en Wetenschap 1998; 41: 323-327.
23. Isacson D. Long-term benzodiazepine use: factors of importance
and the development of individual use patterns over time — a 13-
year follow-up in a Swedish community. Soc Sci Med 1997; 44:
24. Zandstra SM, Furer JW, Lisdonk van de EH, et al. Different study
criteria affect the prevalence of benzodiazepine use. Soc Psychiatr
Psychiatr Epidemiol 2002; 37: 139-144.
25. Van der Waals, FW. Sex differences among recipients of benzodi-
azepines in Dutch General Practice. Amsterdam: Academic
Medical Centre, University of Amsterdam, Department of Women’s
Health. [Thesis.] 1995.
26. Konig-Zahn C, Furer JW, Tax B, et al. Regioproject Nijmegen 2:
Psychiatrische morbiditeit in de regio.[The Nijmegen Health Area
Project No 2: psychiatric morbidity in NHA2.] Nijmegen: UMC
Nijmegen, Department of Social Medicine, General Practice and
27. WHO Collaborating Centre for Drugs. Statistics Methodology
Guidelines ATC Classification and DDD Assignment. Oslo:
28. van den Berg J, van den Bos GAM. Het (meten van het)
voorkomen van chronische aandoeningen 1974-1987. [The (mea-
surement of the) presence of chronic disorders.) Maandblad
Gezondheid (CBS) 1989; 4-21.
29. Ware JE, Snow KK, Kosinski M., Gandeli B. SF-36 Health Survey.
Manual and interpretation guide. Boston: The Health Institute, New
England Medical Center, 1993.
30. Van der Zee K, Sanderman R. Het meten van de algemene
gezondheidstoestand met de RAND-36: een handleiding. RU
Groningen: Noordelijk Centrum voor Gezondheidsvraagstukken,
31. Goldberg DP , Williams P . A user’s guide to the General Health
Questionnaire. Windsor: NFER-Nelson, 1988.
32. Koeter MWJ, Ormel J. General Health Questionnaire Nederlandse
Bewerking (Manual). Lisse: Swets and Zeitlinger, 1991.
33. Jorm AF, Duncan-Jones P . Neurotic symptoms and subjective
well-being in a community sample: different sides of the same
coin? Psychol.Med. 1990; 20: 647-654.
34. Hosmer DW, Lemeshow S. Applied Logistic Regression. New
York: Wiley Interscience, 1989; 171-175.
35. Geijer RMM, Thiadens HA, Smeele IJM, et al. NHG-standaard
COPD en astma bij volwassenen: diagnostiek. [Guidelines of the
Dutch College of General Practitioners on asthma in adults —
diagnosis.] Huisarts en Wetenschap 2001; 44: 107-117.
36. Geijer RMM, Van Schayck CP , Van Weel C, et al. NHG-standaard
COPD: behandeling. [Guidelines of the Dutch College of General
Practitioners on COPD.] Huisarts en Wetenschap 2001; 44: 207-
37. Rutten GEHM, Verhoeven S, Heine RJ, et al. NHG-standaard dia-
betes mellitus type 2. Huisarts en Wetenschap 1999; 42: 67-84.
38. Walma EP , Grundmeijer HGLM, Thomas S, et al.(eds). NHG-
Standaarden voor de huisarts I. Maarssen: Elsevier/Bunge, 1999:
39. Tata PR, Rollings J, Collins M, et al. Lack of cognitive recovery fol-
lowing withdrawal from long-term benzodiazepine use. Psychol
Med 1994; 24: 203-213.
40. Herings RM, Stricker BH, de Boer A, et al. Benzodiazepines and
the risk of falling leading to femur fractures. Dosage more impor-
tant than elimination half-life. Arch Intern Med 1995; 155: 1801-
41. Rickels K, Schweizer E, Case WG, Greenblatt DJ. Long-term ther-
apeutic use of benzodiazepines. I. Effects of abrupt discontinua-
tion. Arch Gen Psychiatr 1990; 47: 899-907.
42. Power KG, Jerrom DW, Simpson RJ, Mitchell M. Controlled study
of withdrawal symptoms and rebound anxiety after six-week
course of diazepam for generalised anxiety. BMJ 1985; 290: 1246-
43. Lader MH. Limitations on the use of benzodiazepines in anxiety
and insomnia: are they justified? Eur Neuropsychopharmacol
1999; 9 (Suppl 6): 399S-405S.
We would like to thank the participating practices and patients for their
cooperation during this study.
This work was funded by a grant of the Prevention Fund, now known
as the Council for Medical and Health Research (ZonMW).
S M Zandstra, J W Furer, E H van de Lisdonk, et al
808British Journal of General Practice, October 2002