Untitled

Age and Ageing Advance Access published February 24, 2013
The Author 2013. Published by Oxford University Press on behalf of the British Geriatrics Society.
All rights reserved. For Permissions, please email: [email protected] Increased risk of hip fracture among olderpeople using antidepressant drugs: datafrom the Norwegian Prescription Databaseand the Norwegian Hip Fracture Registry MARIT STORDAL BAKKEN1,2, ANDERS ENGELAND2,3, LARS B. ENGESÆTER4,5, ANETTE HYLEN RANHOFF1,6,STEINAR HUNSKAAR2, SABINE RUTHS2,7 1Kavli Research Centre for Ageing and Dementia, Haraldsplass Deaconess Hospital, PO Box 6165, 5892 Bergen, Norway2Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway3Department of Pharmacoepidemiology, Norwegian Institute of Public Health, Oslo, Norway4Institute of Surgical Sciences, University of Bergen, Bergen, Norway5The Norwegian Arthroplasty Register, Department of Orthopaedic Surgery, Haukeland University Hospital, Bergen, Norway 6Institute of Medicine, University of Bergen, Bergen, Norway7Research Unit for General Practice, Uni Health, Uni Research, Bergen, Norway Address correspondence to: M. S. Bakken. Tel: (+47) 55978500; Fax: (+47) 55978555. Email: [email protected] Background: hip fractures are usually caused by a combination of reduced bone mineral density and falls; using antidepres- sant drugs may affect both of these.
Objective: we aimed to examine associations between exposure to antidepressant drugs and the risk of hip fracture amongolder people, and, provided associations found, to estimate the attributable risk of hip fracture.
Design: we conducted a nationwide prospective cohort study of the 906,422 people in Norway born before 1945.
Methods: information on all prescriptions of antidepressants dispensed in 2004–10 and all primary hip fractures in 2005–10 was obtained from the Norwegian Prescription Database, and the Norwegian Hip Fracture Registry, respectively. The in-cidence rates of hip fracture during the time people were exposed and unexposed to antidepressant drugs were comparedby calculating the standardised incidence ratio (SIR).
Results: altogether 39,938 people (4.4%) experienced a primary hip fracture. The risk of hip fracture was increased forpeople exposed to any antidepressant [SIR = 1.7, 95% confidence interval (CI) 1.7–1.8]; tricyclic antidepressants (SIR = 1.4,95% CI: 1.3–1.5); selective serotonin reuptake inhibitors (SSRIs) (SIR = 1.8, 95% CI: 1.7–1.8) and other antidepressants(SIR = 1.6, 95% CI: 1.5–1.7). The risk of hip fracture attributable to exposure to antidepressant drugs was 4.7%.
Conclusions: this study indicated an increased risk of hip fracture among people exposed to antidepressants, especiallythose with serotonergic properties such as SSRIs. This association needs to be explored further in clinical studies.
Keywords: antidepressants, hip fractures, pharmacoepidemiology, population registers, older people Antidepressants are prescribed to 10–25% of women and 5–20% of men 60 years and older in Europe and the The risk of hip fracture increases with age. The estimated USA, mostly selective serotonin reuptake inhibitors (SSRIs) lifetime risk is 25% for women and 7% for men ]. Hip []. Clinical guidelines recommend SSRIs for depression fractures represent critical events, with great implications for because they have fewer sedative, anticholinergic and car- morbidity and mortality . Most hip fractures result from a diovascular side effects than tricyclic antidepressants combination of reduced bone mineral density and a fall ]; (TCAs). Observational studies indicate associations between using antidepressants may affect both of these [, ].
the use of antidepressants and hip fracture, especially recently initiated drug treatment ]. Some studies TCAs (N06AA), SSRIs (N06AB) and other antidepressants suggest that SSRIs may be associated with a greater risk of (N06AG, N06AX); (ii) the drugs’ serotonergic effects: high hip fracture than TCAs , ], postulating a specific sero- or intermediate serotonergic properties, and those with low We aimed to examine associations between exposure to antidepressant drugs and the risk of hip fracture amongolder people. If we found associations, we aimed to esti- mate the attributable risk of hip fracture.
Since the NorPD does not include information on whetheror when the purchasers consumed the dispensed drugs, wehad to make assumptions on drug exposure. Antidepressants are usually prescribed in 30 or 100 day lots, and we assumed This was a nationwide prospective study based on merged people to start using them the day they were purchased.
data from the Norwegian Prescription Database (NorPD) Antidepressants are predominantly used on a regular daily the Norwegian Hip Fracture Registry and the basis. The DDD is the assumed average maintenance dose per day for a drug used for its main indication among adults, as defined by the WHO [Because the actually prescribeddose may diverge from the DDD, we calculated the risk of hip fracture for various assumed total exposure times (3 and The NorPD, starting from January 2004, contains informa- 14 days and the number of days corresponding to the tion on all prescription drugs purchased at all pharmacies number of DDDs prescribed, respectively). The latter was in Norway []. The data extracted for this study comprises considered the best proxy for the number of person-days all prescriptions of antidepressants, Anatomical Therapeutic Chemical (ATC) system code ] N06A, dispensed from We defined overall use as any exposure to antidepres- January 2004 until December 2010, by the items’ ATC sants within the study period, including all exposure code, drug volume by defined daily dose (DDD) [and periods. Recently started treatment was defined as the first date of dispensing. The NorPD lacks individual informa- 14 days of first-time exposure to antidepressants after a tion on medication dispensed to people living in nursing The Norwegian Hip Fracture Registry, starting from January 2005, contains national data (i.e. injury, fracture and surgery) on people operated on for hip fracture at all 55 We compared the incidence of hip fracture during the hospitals in Norway performing such surgery For the person-days exposed and unexposed to antidepressants in purpose of this study, we extracted the date of hip fracture, the study period by calculating standardised incidence ratios or the date of surgery in case of missing information, for (SIRs) ]. An SIR >1 indicates an increased risk of hip the period January 2005 until December 2010. Although fracture associated with exposure. We adjusted the SIRs for the registry comprises fractures in nursing home patients, sex, birth year and time period (in 2-month intervals).
these individuals cannot be identified.
Results based on overall use of antidepressants in the entire The Central Population Registry contains demographic study population are presented throughout the article. We information on the entire population of Norway. The data performed subanalysis for recently started drug use.
extracted for this study comprise birth year, sex and date of For the SIRs, 95% CIs were calculated assuming a Poisson distribution of the observed number of hip frac- The variables selected from these three registries were tures among exposed people, estimating the mean by the linked, using the unique 11-digit personal identity number expected number of hip fractures among the exposed assigned to everyone living in Norway after 1960.
To calculate the attributable risk of exposure to anti- depressant drugs on hip fracture, we divided the observed minus the expected number of fractures during the number The study population included everyone born before 1945 of person-days exposed to antidepressant drugs by the and living in Norway on 1 January 2005. They were fol- observed number of fractures in the study population.
lowed up until the day of any first hip fracture, emigrationor death or until the end of the study period on 31December 2010.
Antidepressant medication was divided in two different people with a mean age of 72.8 years [standard deviation ways, according to; (i) therapeutic subgroups (ATC group): (SD) 8.9 years] on 1 January 2005 (56% women) and mean Hip fracture risk and antidepressant drugs follow-up 5.2 (SD 1.7) years. Altogether 218,775 persons using antidepressants, the excess risk of hip fracture died (53% women) and 4,949 emigrated (44% women).
decreased with increasing age (Table ).
A total of 153,301 (17%) people purchased at least one Within all antidepressant subgroups, the SIR increased prescription of an antidepressant drug during the study with increasing numbers of assumed exposed person-days period; 69% were women. Antidepressants with serotoner- from 3 to 14 days, and remained largely stable when SIR gic properties and SSRIs were most frequently used (Table ). More women than men used TCAs (31 versus The risk of hip fracture was elevated among people 26% of antidepressant users) and SSRIs (63 versus 58%), exposed to any antidepressant drug (SIR = 1.7, 95% CI: with minor sex differences for other subgroups. The use of 1.7–1.8). The excess risk of hip fracture was higher among antidepressants in various subgroups was similar across dif- people exposed to SSRIs (1.8, 1.7–1.8) than among those ferent birth cohorts, except for TCAs, which decreased exposed to TCAs (1.4, 1.3–1.5) and other antidepressants with increasing age, from 33% (among those born in (1.6, 1.5–1.7). The risk of hip fracture was higher for drugs 1935–44) to 21% (born before 1915) (not shown).
with high or intermediate serotonergic properties (1.7, 1.7– During the study period, 39,938 individuals experienced 1.8) than for drugs with no or low serotonergic properties a primary hip fracture. Most fractures among the people exposed to antidepressants occurred among those born in The sex differences in SIR were greatest for SSRIs 1915–24 (40%) or in 1925–34 (41%). The mean age at the and drugs with high or intermediate serotonergic effects time of fracture was 83 years, and the mean number of days of exposure prior to fracture was 116. Of all hip frac-tures, 72% occurred in women. Within all birth year cohorts, the prevalence of hip fracture was higher among exposed women than exposed men (Table However, the The risk of hip fracture was higher during recently started excess risk of hip fracture was more pronounced among treatment with any antidepressant drug (SIR, all 2.0, exposed men (SIR = 1.9; 95% CI: 1.8–2.0) than among 1.5–2.4; women 1.8, 1.3–2.3; men 2.7, 1.7–4.0) and SSRIs exposed women (1.7; 1.6–1.7). Sex differences were most (all 2.7, 2.1–3.4; women 2.5, 1.9–3.3; men 3.3, 2.0–5.2) prominent in the youngest cohort born in 1935–44 (men than during overall use (not shown in table). The total 2.9; 2.6–3.4, women 2.5; 2.3–2.7). Generally, for the people Table 1. Number of people in Norway born before 1945 who purchased antidepressant drugs during 2005–10 andproportions by various antidepressant subgroups Antidepressant subgroups (% of people who purchased) (%). . . . . . . . .T .C .A . . . . . . . . . .
The totals in each of the subgroups do not add up to 100%, because individuals may have purchased more than one antidepressant.
TCA (ATC N06AA): tricyclic antidepressants (clomipramine, trimipramine, amitriptyline, nortriptyline and doxepin).
SSRI (N06AB): selective serotonin reuptake inhibitors (fluoxetine, citalopram, paroxetine, sertraline, fluvoxamine and escitalopram).
Others (N06AG, N06AX): moclobemide, mianserin, mirtazapine, bupropion, venlafaxine, reboxetine and duloxetine.
High or intermediate serotonergic properties: All SSRIs, TCAs (clomipramine and amitriptyline) and others (mianserin, mirtazapine, venlafaxine and duloxetine).
Low or no serotonergic properties: TCAs (nortriptyline, doxepin and trimipramine), moclobemide, bupropion and reboxetine.
Table 2. Number of hip fractures (n) and excess risk of hip fracture (SIR, 95% CI) in Norway’s population born before1945 exposed to antidepressant drugs in 2005–10, by birth cohort and sex SIR, standardised incidence ratio. CI, confidence interval.
Table 3. Observed number of hip fractures (n) and excess risk of hip fracture (SIR, 95% CI) in Norway’s population bornbefore 1945 exposed to various antidepressant drug subgroups in 2005–10 (Exposed person days, DDD) period) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SIR, standardised incidence ratio. CI, confidence interval. DDD, defined daily dose.
TCA (ATC N06AA): tricyclic antidepressants (clomipramine, trimipramine, amitriptyline, nortriptyline and doxepin).
SSRI (N06AB): selective serotonin reuptake inhibitors (fluoxetine, citalopram, paroxetine, sertraline, fluvoxamine and escitalopram).
Others (ATC N06AG, N06AX): moclobemide, mianserin, mirtazapine, bupropion, venlafaxine, reboxetine and duloxetine.
High or intermediate serotonergic properties: all SSRIs, TCAs (clomipramine and amitriptyline) and others (mianserin, mirtazapine, venlafaxine and duloxetine).
Low or no serotonergic properties: TCAs (nortriptyline, doxepin and trimipramine), moclobemide, bupropion and reboxetine.
antidepressant drug use (<100) was too small for analysis It is reasonable to suspect that SSRIs adversely affect regarding TCAs and antidepressants with low or no seroto- bone strength. A meta-analysis of epidemiological studies nergic properties to yield representative results.
indicates that using SSRIs is associated with an increasedrisk of fracture, although when SSRI use was adjusted fordepression and bone mineral density at baseline, it was non- significantly associated with bone loss ]. A recent case–control study demonstrated associations between using The percentage of hip fractures attributable to exposure to SSRIs and osteoporosis, independent of depression and con- any antidepressant drug was estimated at 4.7%, with TCAs comitant use of medication affecting bone metabolism, while comprising 0.3%, SSRIs 3.6% and other antidepressants using TCAs was associated with higher bone mineral density 1.0%. Antidepressants with high or intermediate serotoner- [Although functional 5-hydroxytryptamine (5-HT) recep- tors and 5-HT transporters have been localised to osteo-blasts and osteocytes, and 5-HT seems to modulate theskeletal effects of parathyroid hormone and mechanical stimulation , the detailed effects of serotonin on bonemetabolism remain unknown . SSRIs and other drugs Excess risk of hip fracture among exposed people with serotonergic properties may possibly contribute to an Our results showing excess risk of hip fracture in persons increased risk of hip fracture through falls in recently initiated using antidepressants, especially SSRIs and other antide- treatment, due to hyponatraemia and haemodynamic distur- pressants with high or intermediate serotonergic properties bances ], and through effects on bone physiology, (e.g. venlafaxine or mirtazapine), are in accordance with assumedly impairing bone architecture and bone strength, in previous studies [, , , About 5% of the hip fractures in the study period were attributable to antidepres- The excess risk of hip fracture in our study was higher among exposed men than exposed women, supporting the We found a higher excess risk of hip fracture among findings of a previous study []. As hip fracture and anti- persons on recently started treatment with antidepressants depressant drug use are less prevalent among men than when compared with all users of SSRIs and other antide- among women, exposure to antidepressants will affect the pressants with high or intermediate serotonergic effects.
hip fracture risk among men relatively more. Attenuation of The numbers of hip fractures are small and these findings the impact of antidepressant drug use with advancing age must be interpreted with caution, but they support previous is possibly due to ‘dilution’ of the effect by other factors research []. The relationships between the dose [, giving rise to an increased risk of hip fracture, such as co- and duration of SSRI use and the risk of hip morbidities, frailty and concomitant drug use.
Other studies suggest that tolerance for TCAs evolves after a few weeks of use, while the risk of hip fracture remains elevated over time among older people using The national health registries provided us a unique oppor- tunity to link complete data on antidepressant drugs Hip fracture risk and antidepressant drugs purchased by a nationwide unselected community-dwelling hip fractures. SSRIs are the most widely prescribed antide- older population with all primary hip fractures registered pressants for older people. We found a marked over- in Norway. The 6-year follow-up period revealed high representation of hip fractures among people using SSRIs numbers of cases, and the nationwide prospective study and other antidepressants with serotonergic properties design prevented selection and information bias. In add- compared both with people not using antidepressants and ition, the conservative definition of exposed person-days, with people using TCAs. At the population level, the excess not allowing for such factors as non-adherence, yields con- risk of hip fracture corresponds to 1,900 fractures during the study period. These associations need to be explored The databases have some limitations. The NorPD lacks further in clinical studies. The growing evidence of SSRIs individual information on medication dispensed to nursing and other antidepressants not being necessarily safer than home residents, leading to systematic misclassification of TCAs requires conscientious evaluation of the potential 40,000 people at any time as drug non-users. The preva- risks and benefits when prescribing antidepressants for lence of both antidepressant use and hip fracture is high among nursing home residents, and the excess risk ofhip fracture has been underestimated in exposed people.
The Norwegian Hip Fracture Registry comprises >80% of all hip fracture operations in Norway ], with somewhatlower completeness during the first years.
• The risk of hip fracture was markedly increased among The most important limitation of our study is the lack older people exposed to antidepressant drugs.
of clinical information such as depression, comorbidities, • Individuals exposed to SSRIs and other drugs with sero- frailty, concomitant drug use and life style. Thus, we do not tonergic properties were at greatest excess risk.
know whether people purchasing prescriptions for antide- • About 5% of hip fractures were attributable to antidepres- pressants were actually diagnosed with depression or not.
Low-dosage TCAs are also prescribed for chronic pain andSSRIs for neuropsychiatric symptoms among people withdementia. Confounding by indication must be taken intoaccount when interpreting the results. Most importantly, both depression itself and antidepressants may influence S.R., L.B.E., A.E. and M.S.B.: study concept and design.
muscle strength, psychomotor function, activity level and A.E., L.B.E. and S.R.: acquiring the data. M.S.B., A.E., bone mineral density and thus interfere with the risk of S.R., A.H.R., S.H. and L.B.E.: analysing and interpreting falls and fractures. However, previous studies have shown the data. M.S.B.: drafting the manuscript. S.R., A.E., that associations between an increased risk of hip fracture A.H.R., L.B.E. and S.H.: critically revising the manuscript among individuals using antidepressants remain even when for important intellectual content. A.E. and M.S.B.: statistic- adjusted for depression [, Further, a large population- al analysis. A.E. had full access to all of the data in the based cohort study among older people diagnosed with de- study and takes responsibility for the integrity of the data pression in the UK adjusted for patient characteristics and the accuracy of the data analysis. S.R.: administrative, revealed no evidence that using SSRIs is safer than using technical or material support and study supervision.
Since antidepressants are predominantly used on a daily basis, we considered the number of days corresponding tothe number of DDDs dispensed the best proxy for the number of drug-exposed person-days. Studies of psycho- tropic drug prescribing for older people in Sweden in 2006and the UK 1996–2007 [showed that the actuallyprescribed mean daily dose of SSRI was nearly 1.0 DDD; the latter study also demonstrated that the mean prescribeddaily dose of TCAs was <0.5 DDD [Thus, our study The Regional Committee for Medical and Health Research may have underestimated the assumed number of person- Ethics (138/07) and the Norwegian Data Inspectorate (08/ days exposed to TCAs and the effect on the risk of hip 00133) approved the study. The Norwegian Directorate of fracture. However, the risk of hip fracture remained largely Health granted an exemption from the duty of confidential- stable when the SIR was calculated for different assumed Since the risk of hip fracture increases with age, demo- graphic changes will contribute to an increasing number of 1. Kanis JA, Johnell O, De Laet C, Jonsson B, Oden A, 17. Furu K. Establishment of the nationwide Norwegian Ogelsby AK. International variations in hip fracture probabil- Prescription Database. Nor J Epidemiol 2008; 18: 129–36.
ities: implications for risk assessment. J Bone Miner Res 18. The WHO Collaborating Centre for Drug Statistics Methodology.
2. Leibson CL, Tosteson AN, Gabriel SE, Ransom JE, Melton 19. Lingjærde O. Psykofarmaka. Medikamentell behandling av LJ. Mortality, disability, and nursing home use for persons psykiske lidelser. 5. utg. (Psychotropic drugs: pharmaceutical with and without hip fracture: a population-based study.
treatment of mental disorders) Oslo: Høyskoleforlaget, 2006.
J Am Geriatr Soc 2002; 50: 1644–50.
20. Engeland A, Skurtveit S, Morland J. Risk of road traffic acci- 3. Cumming RG, Nevitt MC, Cummings SR. Epidemiology of dents associated with the prescription of drugs: a registry- hip fractures. Epidemiol Rev 1997; 19: 244–57.
based cohort study. Ann Epidemiol 2007; 17: 597–602.
4. Leipzig RM, Cumming RG, Tinetti ME. Drugs and falls in 21. Vestergaard P, Rejnmark L, Mosekilde L. Anxiolytics, seda- tives, antidepressants, neuroleptics and the risk of fracture.
I. Psychotropic drugs. J Am Geriatr Soc 1999; 47: 30–9.
5. Wu Q, Liu J, Gallegos-Orozco JF, Hentz JG. Depression, 22. Bolton JM, Metge C, Lix L, Prior H, Sareen J, Leslie WD.
fracture risk, and bone loss: a meta-analysis of cohort Fracture risk from psychotropic medications: a population- studies. Osteoporos Int 2010; 21: 1627–35.
based analysis. J Clin Psychopharmacol 2008; 28: 384–91.
6. Department of pharmacoepidemiology, Norwegian Institute 23. Wu Q, Bencaz AF, Hentz JG, Crowell MD. Selective sero- of Public Health. Drug Consumption in Norway 2006–2010.
tonin reuptake inhibitor treatment and risk of fractures: a Oslo: Norwegian Institute of Public Health, 2011.
meta-analysis of cohort and case-control studies. Osteoporos 7. Hansen DG, Rosholm JU, Gichangi A, Vach W. Increased use of antidepressants at the end of life: population-based 24. Bolton JM, Targownik LE, Leung S, Sareen J, Leslie WD.
study among people aged 65 years and above. Age Ageing Risk of low bone mineral density associated with psychotrop- ic medications and mental disorders in postmenopausal 8. Pratt LA, Brody DJ, Gu Q. Antidepressant use in persons women. J Clin Psychopharmacol 2011; 31: 56–60.
aged 12 and over: United States, 2005–2008. Hyattsville, 25. Haney EM, Warden SJ, Bliziotes MM. Effects of selective MD: National Center for Health Statistics, 2011 (NCHS serotonin reuptake inhibitors on bone health in adults: time for recommendations about screening, prevention and man- 9. Coupland C, Dhiman P, Morriss R, Arthur A, Barton G, Hippisley-Cox J. Antidepressant use and risk of adverse out- 26. Darowski A, Chambers SA, Chambers DJ. Antidepressants comes in older people: population based cohort study. BMJ and falls in the elderly. Drugs Aging 2009; 26: 381–94.
27. Kannegaard PN, van der Mark S, Eiken P, Abrahamsen B.
10. van den Brand MW, Pouwels S, Samson MM et al. Use of Excess mortality in men compared with women following a anti-depressants and the risk of fracture of the hip or femur.
hip fracture. National analysis of comedications, comorbidity and survival. Age Ageing 2010; 39: 203–9.
11. Hubbard R, Farrington P, Smith C, Smeeth L, Tattersfield A.
28. Ruths S, Sørensen PH, Kirkevold O et al. Trends in psycho- Exposure to tricyclic and selective serotonin reuptake inhibi- tropic drug prescribing in Norwegian nursing homes from tor antidepressants and the risk of hip fracture. Am J 1997 to 2009: a comparison of six cohorts. Int J Geriatr Psychiatry 2012 Nov (Epub ahead of print, November 2012).
12. Coupland CA, Dhiman P, Barton G et al. A study of the 29. Schneeweiss S, Wang PS. Association between SSRI use and safety and harms of antidepressant drugs for older people: a hip fractures and the effect of residual confounding bias in cohort study using a large primary care database. Health claims database studies. J Clin Psychopharmacol 2004; 24: Technol Assess 2011; 15: 1–202. , iii–iv.
13. Richards JB, Papaioannou A, Adachi JD et al. Effect of se- 30. Lesen E, Petzold M, Andersson K, Carlsten A. To what lective serotonin reuptake inhibitors on the risk of fracture.
extent does the indicator “concurrent use of three or more Arch Intern Med 2007; 167: 188–94.
psychotropic drugs” capture use of potentially inappropriate psychotropics among the elderly? Eur J Clin Pharmacol 15. Gjertsen JE, Engesaeter LB, Furnes O et al. The Norwegian Hip Fracture Register: experiences after the first 2 years and Received 28 August 2012; accepted in revised form 15,576 reported operations. Acta Orthop 2008; 79: 583–93.

Source: http://nrlweb.ihelse.net/Forskning/Publikasjoner/Age%20and%20Ageing_Bakken%20MS%20-%20increased%20risk%20of%20hip%20fracture%202013.pdf