BENZODIAZEPINE USE AMONG THE ELDERLYIN THE COMMUNITY

MICHAEL KIRBY1*{, AISLING DENIHAN1, IRENE BRUCE1, ALICJA RADIC2, DAVIS COAKLEY1 AND BRIAN A. LAWLOR1

1Mercer's Institute for Research on Ageing and Department of Psychiatry for the Elderly,St. James's Hospital, Dublin 8, Ireland

2Health Research Board, 73 Lower Baggot Street, Dublin 2, Ireland

SUMMARY

Benzodiazepines are the most commonly prescribed psychotropic drug in the elderly. Benzodiazepines with a longduration of action can produce marked sedation and psychomotor impairment in older people, and are associatedwith an increased risk of hip fracture and of motor vehicle crash. One thousand seven hundred and one individuals of65 years and over, identi®ed from General Practitioner lists, were interviewed using the Geriatric Mental State-AGECAT package and current psychotropic drug use was recorded. Benzodiazepines were classi®ed as having a shortor long elimination half-life. Two hundred and ninety-®ve (17.3%) individuals were taking a benzodiazepine, with usein females being twice that in males. Of the 295, 152 (51.5%) were taking a long acting benzodiazepine and the use oflong acting anxiolytic type benzodiazepines was particularly common. Fifty-two (17.6%) benzodiazepine users weretaking one or more other psychotropic drugs. A benzodiazepine was used by eight of 18 (44.4%) subjects with ananxiety disorder, 62 of 180 (34.4%) individuals with depression, and seven of 71 (9.9%) people with dementia. Four-®fths of older people on a psychotropic drug were taking a benzodiazepine, highlighting the importance of this classof drug in the elderly population. The choice of a benzodiazepine with a long duration of action, which have beenshown to be associated with serious adverse events in the elderly in over one half of benzodiazepine users, is ofconcern. The potential for adverse e�ects was further accentuated by polypharmacy practices. The choice ofbenzodiazepine for an older person has important consequences and should be addressed in greater detail withprimary care. Copyright # 1999 John Wiley & Sons, Ltd.

KEY WORDSÐbenzodiazepine; psychotropic; elderly; community; primary care

Benzodiazepines are the most commonly usedpsychotropic drugs among older people (Skooget al., 1993; Fichter et al., 1989). The adverse e�ectsof benzodiazepine drugs include drowsiness, dizzi-ness, ataxia, impairment of psychomotor function,and these e�ects appear to increase with age (Rayet al., 1992; Greenblatt et al., 1991, 1989). Thegreater sedation and degree of psychomotorimpairment in the elderly may be due to reducedclearance of the drug, resulting in higher plasmaconcentrations (Greenblatt et al., 1991). Benzo-diazepines with a long duration of action areparticularly likely to accumulate and thereforehave a greater potential for sedative e�ects and

psychomotor impairment. Older people who aretaking benzodiazepines with a long eliminationhalf-life have been shown to have an increased riskof falls (Tinetti et al., 1988; Sorock and Shimkin,1988), a greater risk of hip fracture (Ray et al.,1989) and an increased chance of being involved ina motor vehicle crash (Hemmelgarn et al., 1997).

METHODS

The data presented in this paper were derived fromthe Dublin study on mental disorders among thecommunity dwelling elderly (Kirby et al., 1997),and was collected during the years 1993±1997.Individuals of 65 years and over on the practice listsof ®ve general practices located within the catch-ment area of St. James's Hospital, Dublin wereidenti®ed. Elderly people in residential care were

CCC 0885±6230/99/040280±05$17.50 Received 20 May 1998Copyright # 1999 John Wiley & Sons, Ltd. Accepted 9 September 1998

INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY

Int. J. Geriat. Psychiatry 14, 280±284 (1999)

*Correspondence to: Dr M. Kirby, Mercer's Institute forResearch on Ageing, St. James's Hospital, Dublin 8, Ireland.Tel: �353-1-4537941 ext. 2640/1. Fax: �353-1-4541796.{Currently Senior Registrar, Professional Psychiatric Unit,Cork University Hospital, Ireland.

not included. The lists were reviewed with the GPsin order to maximize their accuracy, particularly toremove any subjects who were known to have diedor to have moved away, but still recorded on thelist. The area is an old and settled part of Dublin,with most elderly being long term residents, andconsequently we believe that the practice lists arerepresentative of the elderly population in the area.All subjects were contacted by letter and invited totake part in the study. If contact was not made aftertwo visits the individual was deemed to haverefused. Interviews were conducted in eachperson's home using the Geriatric Mental State(GMS)-AGECAT system (Copeland et al., 1986).The GMS (Copeland et al., 1976) is a semi-structured interview developed to assess organicand functional mental illness in the elderly. Thedata collected by the GMS is applied to theAGECAT computerized diagnostic system (Cope-land et al., 1986) to generate standardized diag-noses. Scores of 3 or above on a diagnostic cluster(organic brain syndrome, schizophrenia, mania,obsessional neurosis, hypochondriasis, phobia,and anxiety) have been shown to equate well withwhat psychiatrists would usually rate as a case(Copeland et al., 1986). Subjects with scores of1 and 2 on a diagnostic cluster are referred to assubcases and represent a collection of symptomsnot reaching case criteria. GMS-AGECAT gener-ated diagnoses have been compared to thosederived from DSM-111 criteria (American Psychi-atric Association, 1980) and have demonstratedgood agreement for case level organic disorderagainst dementia and for case level depressionagainst combined major depression and dysthymia(Copeland et al., 1990). Psychotropic drug use wasrecorded by direct inspection of the medicationsbeing currently used by each person. Medicationsthat had been discontinued at any time prior to theinterview were not included. A psychotropic drugwhich was being taken on an `as required' basis wasincluded, if the person had a supply of the drug (orawaiting a repeat prescription) and told the inter-viewer that he/she continued to take it if required.Each benzodiazepine was classi®ed as having ashort elimination (424 h) or long elimination(424 h) half-life (Hemmelgarn et al., 1997). Theshort acting agents were temazepam, triazolam,alprazolam, ¯unitrazepam, bromazepam, brotizo-lam, lormetazepam and lorazepam. The longacting agents were diazepam, ¯urazepam, nitraze-pam, chlordiazepoxide, clorazepate, clobazam andprazepam.

RESULTS

One thousand seven hundred and one individualswere interviewed, representing 80% of thoseidenti®ed from the practice lists. One thousandone hundred and seventeen (65.7%) females and584 (34.3%) males were interviewed. The mean agewas 74.2 years (S.D. 6.8), with 981 (57.7%) personsunder 75 years (the `young old') and 720 (42.3%)persons of 75 years and older (the `older old').There was no signi®cant di�erence in age andgender distribution between those who were inter-viewed and those who refused assessment.

Three hundred and seventy-three (21.9%) indi-viduals were taking psychotropic medication. Thecommonest class of psychotropic drug was thebenzodiazepines which were used by 295 (17.3%)individuals. Antidepressant drugs were taken by 85(5%) subjects and neuroleptic drugs by 22 (1.3%).Eight people were taking a non-benzodiazepineanxiolytic/hypnotic ( four chlormethiazole, fourzopiclone) and nine individuals said they wereusing a `sleeping tablet' but the identity of the drugwas not ascertained. Females (239/1117, 21.4%)were twice as likely to be taking a benzodiazepineas males (56/584, 9.6%; w2 � 37.3, df � 1,p4 0.0001), and this was consistent across bothage groups. The `older old' (127/720, 17.6%) wereusing benzodiazepines with a similar frequency tothe `younger old' (168/981, 17.1%).

Table 1 lists the individual benzodiazepines interms of frequency of use among older people inthe community. Of the 295 persons on a benzo-diazepine drug, 152 (51.5%) were taking one with along duration of action and the `older old' (52.8%)were as likely to be taking a long acting agent as the`younger old' (50.6%). The choice of a long half-life drug was particularly prominent among thegroup on benzodiazepines which are typically usedas anxiolytics (diazepam, alprazolam, chlordiazep-oxide, bromazepam, lorazepam, clorazepate,clobazam, prazepam) with 99 (77.3%) of 128individuals taking a long acting benzodiazepine.In contrast, only 67 (35.6%) of the 188 older peopleon a hypnotic type benzodiazepine (temazepam,¯urazepam, nitrazepam, ¯unitrazepam, lormetaze-pam, brotizolam) were using a long acting drug.

Fifty-two (17.6%) of the 295 individuals on abenzodiazepine were also taking one or more otherpsychotropic drugs, 24 using another benzo-diazepine and 33 using another psychotropicfrom a di�erent class. With regard to the 152 indi-viduals on long acting benzodiazepines, 37 (24.3%)

Copyright # 1999 John Wiley & Sons, Ltd. Int. J. Geriat. Psychiatry 14, 280±284 (1999)

BENZODIAZEPINE IN COMMUNITY ELDERLY 281

were taking at least one other psychotropic drug,18 on more than one benzodiazepine and 24 usinganother class of psychotropic drug.

In our population of 1701 elderly persons therewere 269 (15.8%) GMS-AGECAT cases of mentaldisorder, consisting of 180 (10.6%) depressioncases, 71 (4.2%) organic cases and 18 (1.1%)anxiety related disorders (anxiety, phobic andobsessional neuroses). Table 2 shows the frequencyof benzodiazepine use according to diagnosticcategory. Predictably there was a signi®cantlyhigher use of benzodiazepines among those witha current mental disorder (77/269, 28.6%) thanthose without (218/1432, 15.2%; w2 � 28.37,df � 1, p4 0.0001). Among those with a mentaldisorder there was a higher use of benzodiazepinesby females (61/196, 31.1%) than males (16/73,21.9%), though the di�erence was not signi®cant(w2 � 2.2, df � 1, p � 0.14). Females without amental disorder were signi®cantly more likely to beon a benzodiazepine than males without a mentaldisorder (178/921, 19.3% versus 40/511, 7.8%;w2 � 33.7, df � 1, p4 0.0001). Benzodiazepineswere the most commonly used psychotropic drug indepression, with antidepressants being taken byonly 31 of the 180 (17.2%) depressed older people.Those without a mental disorder included bothindividuals with symptoms short of criteria for case

level illness (subcases) and those with no symptom-atology (the `well'). Ninety-three of the 883(10.5%) `well' individuals were taking a benzo-diazepine, which was predominantly a hypnotictype benzodiazepine (71/93, 76.4%). There was aparticularly high use of long acting benzo-diazepines among individuals with case levelmental disorder. Of those taking benzodiazepines,there was a higher use of long acting agents amongsubjects with a mental disorder (50/77, 64.9%)when compared the rest of the population onbenzodiazepines (102/218, 46.8%; w2 � 7.502,df � 1, p � 0.0062) and also when compared tothose with symptoms at subcase level (59/125,47.2%; w2 � 6.033, df � 1, p � 0.0140).

DISCUSSION

There are two main strengths to this study: ®rstly,the relatively large population size and its natural-istic design, in that we report the actual use ofbenzodiazepines by older people in the community(rather than relying, for instance, on prescriptiondata) and secondly, our facility to relate drug usewith current mental status. The limitations includean absence of data on the duration of use of thedrugs by each individual and the lack of accurateinformation as to whether the benzodiazepineswere being used for anxiolytic or hypnoticpurposes. Where appropriate, we have tried tomake an approximate distinction between anxio-lytics and hypnotics on the basis of the particularbenzodiazepine drug used. While the availability ofnon-benzodiazepine anxiolytic/hypnotics, such aszopiclone, in recent years would be expected toimpact on prescribing patterns, benzodiazepinescontinue to constitute the vast bulk of anxiolytic/hypnotic drugs used. The exclusion of people in

Table 1. The distribution of benzodiazepine drugs usedin the elderly population (n � 1701)

Benzodiazepine drug No. of persons taking the drug

(% of those on a benzodiazepine,

n � 295)

Long acting benzodiazepines

Diazepam 71 (24.1)

Nitrazepam 35 (11.9)

Flurazepam 25 (8.5)

Chlordiazepoxide 17 (5.8)

Clorazepate 5 (1.7)

Clobazam 2 (0.7)

Prazepam 2 (0.7)

Short acting benzodiazepines

Temazepam 72 (24.4)

Triazolam 33 (11.2)

Alprazolam 20 (6.8)

Flunitrazepam 14 (4.7)

Bromazepam 9 (3.1)

Lormetazepam 5 (1.7)

Brotizolam 5 (1.7)

Lorazepam 5 (1.7)

Table 2. Benzodiazepine use among the elderly with (i)case level depression, (ii) case level anxiety disorders(anxiety, phobic and obsessional neuroses), (iii) caselevel organic disorder (dementia), and (iv) no mentaldisorder

Diagnostic category No. (%) on a

benzodiazepine

Depression (n � 180) 62 (34.4)

Anxiety disorders (n � 18) 8 (44.4)

Organic disorder (n � 71) 7 (9.9)

No mental disorder (n � 1432) 218 (15.2)

Copyright # 1999 John Wiley & Sons, Ltd. Int. J. Geriat. Psychiatry 14, 280±284 (1999)

282 M. KIRBY ET AL.

residential care should not greatly in¯uence our®ndings as only 5% of older people in Ireland arein residential care (National Council for theElderly, 1996). Furthermore, our emphasis is onthe pattern of benzodiazepine use by the elderly inthe community who may continue to lead anindependent life ( for example, they may be stilldriving) and whose activities may be particularlya�ected by adverse drug e�ects.

In this Dublin population almost four-®fths ofthose using a psychotropic drug were taking abenzodiazepine, making this class by far the mostfrequently used psychotropic and therefore deserv-ing of close attention. The ®gure of 17.3% forbenzodiazepine use was close to the 15.8% ®gurefound by Fichter et al. (1989) in a similar age groupin Germany, though lower than the rate of 34.2%for anxiolytic/hypnotic use shown by Skoog et al.(1993) in an older (85 year old) population inSweden. The higher rate of benzodiazepine useamong females is consistent with ®ndings fromother studies of older people (Skoog et al., 1993;Fichter et al., 1989), and is particularly markedamong those older people without a mentaldisorder. This may be due to a greater willingnessamong females to disclose psychological symp-toms, including less severe psychological symp-toms, to their doctor.

In spite of clear evidence that altered pharmaco-kinetics in the elderly increases the half-life ofbenzodiazepines (Greenblatt et al., 1991, 1989)and, therefore, the potential for adverse side-e�ects, most reports of psychotropic drug use inthe elderly deal with benzodiazepines or anxiolytic/hypnotics as a class (Leveille et al., 1994; Skooget al., 1993; Ray et al., 1992; Fichter et al., 1989)and do not report the relative use of short and longacting agents. Ray et al. (1989) found that therelative risk of hip fracture was 1.7 for current usersof long half-life benzodiazepines, in contrast tothat of 1.1 in current users of short half-life benzo-diazepines, and Hemmelgarn et al. (1997) reporteda signi®cantly increased risk of motor vehicle crashassociated with the use of long half-life benzo-diazepines but not short acting agents, both follow-ing initiation of the drug and during continuoususe. Over half of the benzodiazepine users in theDublin population were taking a long acting agent,and this proportion remained unaltered among the`older old' who may be particularly susceptible toadverse e�ects. There is little other published dataon the speci®c drug use of long acting benzo-diazepines by the community dwelling elderly in

Ireland or the United Kingdom, for comparativepurposes. However, North American data from1985 showed that one third of elderly benzo-diazepine users were receiving long half-life drugs(Ray et al., 1989), and nursing home data reporteda similar ®gure (Beardsley et al., 1989). Therefore,the proportion of elderly benzodiazepine users inDublin on long acting agents, and consequentlyexposed to potentially serious side e�ects, is high,especially in view of the increasing availability ofshort acting benzodiazepines in recent years. Itwould appear that the main problem lies in theprescribing of anxiolytic, as opposed to hypnotic,benzodiazepines. The undesirable `hang-over'e�ects of long acting hypnotics is frequentlyhighlighted, but it may be that the same emphasishas not been given to this issue in the context ofdaytime benzodiazepine use.

The potential for medication related adversee�ects is accentuated through psychotropic poly-pharmacy. The use of more than one psychotropicdrug in almost one-quarter of the group taking along acting benzodiazepine, who are already at ahigh risk of psychomotor impairment and seda-tion, is of concern.

We examined benzodiazepine use in the commondiagnostic categories and it was not surprising thatthey were used with greatest frequency in anxietydisorders. Consistent with the ®ndings from otherstudies on the community dwelling elderly (Wellset al., 1994; Skoog et al., 1993) benzodiazepineswere the most commonly used psychotropic drugamong the depressed elderly in Dublin. Whilejudicious use of benzodiazepines may be appro-priate in some patients with dementia, it wasreassuring that they were not being used frequently.It may be interesting to speculate as to why one inten `well' individuals were taking a benzo-diazepineÐdoes this represent successfully treatedanxiety disorders or treatment for isolated sleepdisturbance? The ®nding that over three-quarterswere taking a hypnotic type benzodiazepine lendssupport to the latter explanation.

Long acting benzodiazepines appear to be morefrequently chosen over short half-life agents forthose with case level mental disorders than for therest of the population. Does this re¯ect a percep-tion in primary care that long half-life benzo-diazepines are more e�ective or `stronger' thanshort acting agents, possibly due to their greatersedation and psychomotor impairment, and con-sequently prescribed with greater frequency wheresymptoms are more severe?

Copyright # 1999 John Wiley & Sons, Ltd. Int. J. Geriat. Psychiatry 14, 280±284 (1999)

BENZODIAZEPINE IN COMMUNITY ELDERLY 283

Studies which have shown the greater risk of hipfracture (Ray et al., 1989) and motor vehicle crash(Hemmelgarn et al., 1997) associated with longhalf-life benzodiazepines highlight the medical,social and economic consequences of the frequentuse of long half-life benzodiazepines among theelderly. This paper demonstrates the continuedwidespread use of long-acting agents, in spite of theavailability of many short half-life alternatives.There may be a number of reasons for thisprescribing practice, including a greater familiaritywith the older long acting drugs, a lack ofawareness of the greater side e�ects associatedwith long acting benzodiazepines in the elderlyand, possibly, a perception that short half-lifedrugs are not as e�ective. It will be important toaddress these issues with primary care physicians,who treat the vast majority of the elderly popula-tion, if we are to succeed in maximizing the bene®tsand minimizing the risks of psychotropic drug usein older people.

ACKNOWLEDGEMENTS

The authors would like to thank our GeneralPractitioners' colleagues for their cooperation, andthe Health Research Board for contributing to thefunding of this project.

REFERENCES

American Psychiatric Association (1980) Diagnostic andStatistical Manual of Mental Disorders (DSM-III)(3rd edn). APA, Washington DC.

Beardsley, R. S., Larson, D. B., Burns, B. J., Thompson,J. W. and Kamerow, D. B. (1989) Prescribing ofpsychotropics in elderly nursing home patients. J. Am.Geriatr. Soc. 37, 327±330.

Copeland, J. R.M., Dewey, M. E., Gri�ths-Jones, H.M.(1986) A computerized psychiatric diagnostic systemand case nomenclature for elderly subjects: GMS andAGECAT. Psychol. Med. 16, 89±99.

Copeland, J. R. M., Dewey, M. E. and Gri�ths-Jones,H. M. (1990) Dementia and depression in elderlypersons: AGECAT compared with DSM-III andpervasive illness. Int. J. Geriatr. Psychiat. 5, 47±51.

Copeland, J. R. M., Kelleher, M. J., Kellett, J. M.,Gourlay, A. J., Gurland, B. J., Fleiss, J. L. and Sharpe,

L. (1976) A semi-structured clinical interview for theassessment of diagnosis and mental state in the elderly:the Geriatric Mental State, 1: development andreliability. Psychol. Med. 6, 439±449.

Fichter, M. M., Witzke, W., Leibl, K. and Hippius, H.(1989) Psychotropic drug use in a representativecommunity sample: the Upper Bavarian study. ActaPsychiatr. Scand. 80, 68±77.

Greenblatt, D. J., Harmatz, J. S., Shapiro, L., Engle-hardt, N., Gouthro, T. A. and Shader, R. I. (1991)Sensitivity to triazolam in the elderly. N. Engl. J. Med.324, 1691±1698.

Greenblatt, D. J., Shader, R. I. and Harmatz, J. S. (1989)Implications of altered drug disposition in the elderly:studies of benzodiazepines. J. Clin. Pharmacol. 29,866±872.

Hemmelgarn, B., Suissa, S. and Huang, A. (1997)Benzodiazepine use and the risk of motor vehiclecrash in the elderly. JAMA 278, 27±31.

Kirby, M., Radic, A., Bruce, I., Coakley, D. and Lawlor,B. A. (1997) Mental disorders among the community-dwelling elderly. Brit. J. Psychiat. 171, 369±372.

Leveille, S. G., Buchnew, D. M., Koepsell, T. D.,McCloskey, L. W., Wolf, M. E. and Wagner, E. H.(1994) Psychoactive medications and injurious motorvehicle collisions involving older drivers. Epidemiology5, 591±598.

National Council for the Elderly (1996)Mental Disordersin Older Irish People: Incidence, Prevalence andTreatment. Report No. 45, National Council for theElderly, Dublin.

Ray, W. A., Gri�n, M. R. and Downey, W. (1989)Benzodiazepines of long and short elimination half-lifeand the risk of hip fracture. JAMA 262, 3303±3307.

Ray, W., Fought, R. and Decker, M. (1992) Psychoactivedrugs and the risk of injurious motor vehicle crashes inelderly drivers. Am. J. Epidemiol. 136, 873±883.

Skoog, I., Nilsson, L., Landahl, S. and Steen, B. (1993)Mental disorders and the use of psychotropic drugs inan 85-year-old urban population. Int. Psychogeriatr. 5,33±48.

Sorock, G. S. and Shimkin, E. E. (1988) Use ofbenzodiazepine sedatives and risk of falling in acommunity-dwelling elderly cohort. Arch. Intern.Med. 148, 2441±2444.

Tinetti, M. E., Speechley, M. and Ginter, S. F. (1988)Risk factors for falls among elderly persons living inthe community. N. Engl. J. Med. 319, 1701±1707.

Wells, K. B., Katon, W., Rogers, B. and Camp, P. (1994)Use of minor tranquilizers and antidepressant medi-cations by depressed outpatients: results from theMedical Outcomes Study. Am. J. Psychiat. 15,694±700.

Copyright # 1999 John Wiley & Sons, Ltd. Int. J. Geriat. Psychiatry 14, 280±284 (1999)

284 M. KIRBY ET AL.