Use of Antidepressants and Rates of Hip Bone Loss in Older Women the Study of Osteoporotic Fractures

Home , Bisphosphonate

ORIGINAL INVESTIGATION Use of Antidepressants and Rates of Hip Bone Loss in Older Women The Study of Osteoporotic Fractures

Susan J. Diem, MD, MPH; Terri L. Blackwell, MA; Katie L. Stone, PhD; Kristine Yaffe, MD; Elizabeth M. Haney, MD; Michael M. Bliziotes, MD; Kristine E. Ensrud, MD, MPH

Background: Serotonin transporters have recently been toms were identified using a cutoff score of at least 6 on described in bone, raising the possibility that medica- the Geriatric Depression Scale. tions that block serotonin reuptake could affect bone metabolism. Results: After adjustment for potential confounders, including the Geriatric Depression Scale score, mean total Methods: We assessed current use of selective seroto- hip BMD decreased 0.47% per year in nonusers com- nin reuptake inhibitors (SSRIs) and tricyclic antidepres- pared with 0.82% in SSRI users (PϽ.001) and 0.47% in sants (TCAs) and obtained serial bone mineral density TCA users (P=.99). Higher rates of bone loss were also (BMD) measurements in a cohort of 2722 older women observed at the 2 hip subregions for SSRI users. Results (mean age, 78.5 years) participating in the Study of Os- were not substantially altered when women who scored teoporotic Fractures, a prospective cohort study of com- at least 6 on the Geriatric Depression Scale were ex- munity-dwelling women. Hip BMD was measured at the cluded from the analysis. sixth examination and an average of 4.9 years later at the eighth examination. We categorized women as nonus- Conclusion: Use of SSRIs but not TCAs is associated with ers (used no SSRIs or TCAs at either examination; an increased rate of bone loss at the hip in this cohort of n=2406), SSRI users (used SSRIs but no TCAs at either older women. examination; n=198), or TCA users (used TCAs but no SSRIs at either examination; n=118). Depressive symp- Arch Intern Med. 2007;167:1240-1245

NTIDEPRESSANTS ARE ONE metabolism and that medications that of the most commonly affect these transporter systems may also prescribed classes of phar- affect bone metabolism. Author Affiliations: macologic agents in the Department of Medicine and United States; in 2002, Division of Epidemiology, CME course available at University of Minnesota 8.5% of Americans used antidepres- A1 www.archinternmed.com (Drs Diem and Ensrud), and sants. With the development of the selective serotonin reuptake inhibitors Center for Chronic Disease To determine whether SSRI and TCA (SSRIs), prescriptions for antidepres- Outcomes Research, use among older women is associated with Minneapolis Veterans Affairs sants for the elderly have increased sub- increased rates of hip bone loss, we ascer- Medical Center (Dr Ensrud), stantially during the past 2 decades.2 Minneapolis, Minn; California tained use of antidepressants, assessed evi- Pacific Medical Center Research See also pages dence of depressive symptoms, and per- Institute, San Francisco formed hip bone mineral density (BMD) (Ms Blackwell and Dr Stone); 1231 and 1246 measurements at 2 examinations in a co- Departments of Psychiatry, hort of 2722 women 65 years and older Neurology, and Epidemiology Selective serotonin reuptake inhibi- who were enrolled in the Study of Osteo- and Biostatistics, University of tors function by inhibiting the serotonin porotic Fractures. California–San Francisco transporter.3 An older class of antidepres- (Dr Yaffe); Department of sants, the tricyclic antidepressants (TCAs), METHODS Psychiatry, San Francisco inhibit uptake of norepinephrine and sero- Veterans Affairs Medical Center, 3 San Francisco (Dr Yaffe); and tonin to varying degrees. The recent de- PARTICIPANTS Department of Medicine, scription of functional serotonin trans- Oregon Health Sciences Center, porters in osteoblasts, osteoclasts, and From September 10, 1986, through January 5, Portland (Drs Haney and osteocytes4,5 raises the possibility that sero- 1989, 9704 women 65 years or older were re- Bliziotes). tonin transporters may play a role in bone cruited for participation in the prospective

(REPRINTED) ARCH INTERN MED/ VOL 167, JUNE 25, 2007 WWW.ARCHINTERNMED.COM 1240

Downloaded From: https://jamanetwork.com/ on 09/26/2021 Study of Osteoporotic Fractures. Women were recruited from DEPRESSIVE SYMPTOMS population-based listings in Baltimore County, Maryland; Min- neapolis, Minn; Portland, Ore; and the Monongahela Valley, Depressive symptoms were evaluated at visit 6 using the 15- Pennsylvania.6 We initially excluded African American women item GDS,14 a self-report scale consisting of 15 yes/no ques- because of their low incidence of hip fracture, women who were tions regarding symptoms of depression. A standard cutoff of unable to walk without help, and women with a history of bi- 6 or more symptoms was used to define evidence of depres- lateral hip replacement. sion; the cutoff of 6 or more symptoms has a sensitivity of 88% From January 3, 1997, through February 8, 1999, 7008 of and a specificity of 62% compared with a structured clinical the original cohort (91.6% of survivors) and an additional 662 interview for depression.15 African American women attended a sixth clinic examination (visit 6).7 From January 15, 2002, through April 30, 2004, 4727 women (74.4% of survivors) completed at least the question- OTHER MEASUREMENTS naire component of the eighth clinic examination (visit 8). Of these, 2844 women completed a medication inventory at both Participants completed a questionnaire and were interviewed visits, completed the Geriatric Depression Scale (GDS) at visit at visits 6 and 8 about self-reported health, physical activity, 6, and had technically adequate hip BMD measurements at both and smoking status. Current use of oral estrogen, thiazides, bis- visits. We excluded 122 women from our analysis who were phosphonates, vitamin D supplements, and calcium supple- taking antidepressants other than SSRIs or TCAs (n=91) or who ments were determined using the method described for ascer- reported use of both an SSRI and a TCA (n=31). The remain- tainment of antidepressant use. Dietary calcium intake was ing 2722 women are included in these analyses of SSRI use, TCA estimated by using the validated 60-item Block semiquantita- use, and rate of change in hip BMD. tive food frequency questionnaire, which was developed from the Second National Health and Nutrition Examination Because depression is associated with lower BMD in some 16 studies,8-10 we also performed a secondary analysis in which we Survey. excluded 336 women who had a GDS score of at least 6 at visit To assess function, women were asked whether they had difficulty performing any of 6 independent activities of daily 6or8. 17-19 The appropriate institutional review boards approved the living. A composite functional impairment score (range, 0-6) study, and written informed consent was obtained from all par- expressed the total number of activities that a participant re- ticipants. ported difficulty performing. Cognitive function was assessed with the Mini-Mental State Examination (maximum score, 30).20 Body weight and height were measured by using a balance beam SSRI AND TCA USE scale and wall-mounted Harpenden stadiometer (Holtain Ltd, Crymych, Wales).21 Body mass index was calculated as weight Participants attending visits 6 and 8 were asked to bring all cur- in kilograms divided by height in meters squared. Neuromus- rent (any use within past 2 weeks) prescription and nonpre- cular function was assessed by measuring the time in seconds scription medications. Interviewers completed a medication his- needed to walk 12 m and by determining whether the partici- tory for each participant, including the name of the medication pant could rise up from a chair (without using the chair’s arms) and frequency of use. A computerized dictionary was used to 5 times.22 categorize the type of medication from product brand and ge- neric names obtained from containers.11 Subsequently, a physician (S.J.D.) blinded to outcome status reviewed the com- STATISTICAL ANALYSIS puterized drug data for SSRI and TCA use and verified ␹2 classification of medications. We used tests for categorical variables; unpaired, 2-tailed t tests Nonusers were defined as women not taking an SSRI, a TCA, for normally distributed continuous data; and Wilcoxon rank or any other type of antidepressant at visit 6 or 8. The SSRI us- sum tests for skewed continuous data to compare characteris- ers were defined as women reporting SSRI use but not TCA use tics at visit 6 by category of antidepressant use (nonusers vs at either visit; TCA users, those reporting TCA use but not SSRI SSRI users and nonusers vs TCA users). use at either visit. Women reporting SSRI or TCA use at only 1 To examine the association between SSRI use and the rate of these visits were classified as partial users of that medica- of change in BMD at the total hip, femoral neck, and trochan- tion; women reporting use at both visits were classified as con- ter, we calculated the annualized mean change in BMD and its tinuous users. 95% confidence interval (CI) at these sites for nonusers and SSRI users using the least squares method. Similar analyses were performed to examine the association between TCA use and MEASUREMENT OF BMD the rate of change of hip BMD. Known risk factors for bone loss in our cohort and characteristics related to antidepressant use Bone mineral density of the total hip and 2 subregions (fem- were examined for inclusion in multivariable models for the oral neck and trochanter) were measured at visits 6 and 8 associations between SSRI and TCA use and the rate of change (mean±SD, 4.9±0.6 years between examinations) using dual- in BMD. We included in our multivariable models age and those energy x-ray absorptiometry with bone densitometry scan- variables (race, health status, number of independent activi- ners (QDR-1000 or QDR-2000; Hologic, Bedford, Mass). The ties of daily living impairments, walking speed, ability to rise second measurements of hip BMD were performed using the from a chair, Mini-Mental State Examination score, smoking same instruments as used for the initial measurements. Fur- status, calcium supplement use, vitamin D supplement use, es- ther details of the measurement method, densitometry quality trogen use, thiazide use, bisphosphonate use, body mass in- control procedures, and precision of the measurements in our dex, weight change, total hip BMD at visit 6, and GDS score) cohort have been published elsewhere.12,13 For the primary analy- that were related to SSRI or TCA use at PՅ.10 or a rate of change sis, the rate of change in BMD was expressed as an annualized in BMD at PՅ.10 independent of age. percentage of the difference between the follow-up BMD and Because depression is associated with lower BMD in some the initial BMD divided by the initial BMD. Secondary analyses studies,8-10 we performed secondary analyses in which we ex- were performed with absolute rate of change in BMD (reported cluded women who had a GDS score of at least 6 at visit 6 or 8. as grams per square centimeter) as the outcome measure. We also performed secondary analyses that examined the as-

(REPRINTED) ARCH INTERN MED/ VOL 167, JUNE 25, 2007 WWW.ARCHINTERNMED.COM 1241

Downloaded From: https://jamanetwork.com/ on 09/26/2021 Table 1. Use of Antidepressants* Table 2. Baseline Characteristics at Visit 6 by Category of Antidepressant Use* No. (%) of Subjects Using Antidepressants Nonusers TCA Users SSRI Users Characteristic (n = 2406) (n = 118) (n = 198) Antidepressant Visit 6 Visit 8 Age, mean ± SD, y 78.4 ± 3.7 78.6 ± 3.3 78.6 ± 3.3 SSRIs 65 (2.4) 178 (6.5) African American 12.2 7.6 8.1† Fluoxetine hydrochloride 9 (0.3) 21 (0.8) GDS score (range, 0-15), 1.6 ± 2.1 2.0 ± 2.0† 2.7 ± 2.7† Paroxetine hydrochloride 25 (0.9) 63 (2.3) mean ± SD Sertraline hydrochloride 30 (1.1) 63 (2.3) GDS score Ն6 5.2 5.1 13.1† Citalopram hydrobromide 0 28 (0.8) Self-reported health Escitalopram oxalate 0 3 (0.1) status Fluvoxamine maleate 1 (0.04) 0 Excellent or good 86.6 75.4† 82.3 TCAs 86 (3.2) 74 (2.7) Fair 12.5 23.7 15.7 Amitriptyline hydrochloride 55 (2.0) 47 (1.7) Poor or very poor 0.9 0.9 2.0 Nortriptyline hydrochloride 10 (0.4) 11 (0.4) BMI, mean ± SD 27.3 ± 5.1 27.5 ± 4.6 26.9 ± 4.9 Imipramine hydrochloride 12 (0.4) 8 (0.3) Weight change from −2.0 ± 5.1 −1.7 ± 6.9 −1.9 ± 6.0 Desipramine hydrochloride 1 (0.04) 1 (0.04) visits 6 to 8, Doxepin hydrochloride 8 (0.3) 7 (0.3) mean ± SD, kg Amoxapine 0 0 No. of IADL impairments 0.7 ± 1.2 1.0 ± 1.4† 1.0 ± 1.5† Protriptyline hydrochloride 0 0 (range, 0-6), mean ± SD Abbreviations: SSRIs, selective serotonin reuptake inhibitors; TCAs, Walks for exercise 45.3 42.4 47.2 tricyclic antidepressants. Walking speed, 0.96 ± 0.2 0.89 ± 0.2† 0.94 ± 0.2† *Use of antidepressants among the 2722 subjects in the primary analysis. mean ± SD, m/s Forty-three participants were taking an SSRI at visits 6 and 8. One participant Inability to rise from chair 8.4 17.0† 10.6 was taking 2 SSRIs at visit 6; another participant was taking 2 SSRIs at visit 5 times 8. Forty-two participants were taking a TCA at visits 6 and 8. Current smoker 3.7 2.5 3.6 Dietary calcium intake, 763.5 ± 378.5 752.0 ± 389.2 742.0 ± 335.2 mean ± SD, mg/d sociation between antidepressant (SSRI or TCA) use at visit 6 Current calcium 50.9 47.5 48.5 and change in BMD. To explore whether there was an effect of supplement user duration of treatment, we also calculated the annualized mean Current vitamin D 54.5 62.7† 58.1 supplement user change in BMD and its 95% CI for nonusers, partial SSRI us- Current oral estrogen use 20.9 31.4† 26.3 ers, and continuous SSRI users. Current thiazide use 21.2 21.2 19.7 All analyses were performed using SAS statistical software Bisphosphonate use 5.0 5.1 9.1† (version 9.1, SAS Institute Inc, Cary, NC). MMSE score, mean ± SD 28.3 ± 1.7 28.4 ± 1.4 28.3 ± 1.8 Total hip BMD, 0.76 ± 0.14 0.76 ± 0.13 0.74 ± 0.15 2 RESULTS mean ± SD, g/cm

Abbreviations: IADL, independent activities of daily living; BMD, bone mineral CHARACTERISTICS OF density; BMI, body mass index (calculated as weight in kilograms divided by THE STUDY POPULATION height in meters squared); GDS, Geriatric Depression Scale; IADL, independent activities of daily living; MMSE, Mini-Mental State Examination; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant. The cohort included 2722 women, of whom 198 (7.3%) *Unless otherwise specified, data are expressed as percentage of subjects. were SSRI users and 118 (4.3%) were TCA users. The †PϽ.10. remaining 2406 women (88.4%) reported no use of an SSRI, a TCA, or other antidepressant at either visit. Spe- cific drug use among women taking SSRIs and TCAs is use, are shown in Table 3. On average, women taking listed in Table 1. SSRIs experienced a higher age-adjusted rate of bone loss Characteristics of the 2722 participants at visit 6 ac- at the total hip than nonusers (−0.77% vs −0.49% per year; cording to use of an antidepressant are shown in Table 2. P=.005). Results were not substantially altered after ad- Compared with nonusers, SSRI users were more likely justing for multiple potential confounding factors in- to have a GDS score of at least 6 (13.1% vs 5.2%; PϽ.001); cluding age, race, health status, functional status, walk- similar proportions of TCA users and nonusers of anti- ing speed, ability to rise from a chair, Mini-Mental State depressants had a GDS score of at least 6. The SSRI and Examination score, smoking status, calcium supple- TCA users had more impairment of independent activi- ment use, vitamin D supplement use, estrogen use, thia- ties of daily living than did nonusers. The TCA users had zide use, bisphosphonate use, body mass index, weight evidence of poorer physical functioning, as reflected by change, total hip BMD at visit 6, and GDS score (−0.82% slower walking speeds (PϽ.01) and more difficulty aris- vs −0.47% per year; PϽ.001). At any of the sites, the ad- ing from a chair (P=.002). justed rate of loss among SSRI users was at least 1.6-fold higher than that among nonusers of antidepressants. SSRI USE AND RATE OF HIP BONE LOSS In secondary analyses that examined rates of change in BMD for nonusers (n=2406), partial SSRI users Mean rates of bone loss at the total hip, femoral neck, (n=155), and continuous SSRI users (n=43), there was and trochanter, according to category of antidepressant no difference in the rate of bone loss between partial and

(REPRINTED) ARCH INTERN MED/ VOL 167, JUNE 25, 2007 WWW.ARCHINTERNMED.COM 1242

Downloaded From: https://jamanetwork.com/ on 09/26/2021 Table 3. Mean Annualized Rate of Bone Loss at Total Hip, Femoral Neck, and Trochanter by Category of Antidepressant Use

Antidepressant Use Category Mean Annualized Change in Bone Mineral Density Nonusers TCA Users SSRI Users by Location, % (95% CI) (n = 2406) (n = 118) (n = 198) Total hip Age-adjusted model −0.49 (−0.54 to −0.43) −0.44 (−0.68 to −0.20) −0.77 (−0.96 to −0.58)† Multivariable model* −0.47 (−0.53 to −0.42) −0.47 (−0.70 to −0.24) −0.82 (−1.00 to −0.64)‡ Femoral neck Age-adjusted model −0.24 (−0.31 to −0.17) −0.31 (−0.62 to −0.01) −0.57 (−0.82 to −0.03)† Multivariable model* −0.23 (−0.29 to −0.16) −0.32 (−0.62 to −0.02) −0.60 (−0.84 to −0.36)† Trochanter Age-adjusted model −0.49 (−0.57 to −0.42) −0.47 (−0.80 to −0.14) −0.89 (−1.16 to −0.63)† Multivariable model* −0.48 (−0.55 to −0.41) −0.47 (−0.79 to −0.15) −0.93 (−1.18 to −0.68)‡

Abbreviations: CI, confidence interval; SSRIs, selective serotonin reuptake inhibitors; TCAs, tricyclic antidepressants. *Adjusted for age, race, health status, functional status, walking speed, ability to rise from a chair, Mini-Mental State Examination score, smoking status, calcium supplement use, vitamin D supplement use, estrogen use, thiazide use, bisphosphonate use, body mass index, change in weight, total hip bone mineral density at visit 6 examination, and Geriatric Depression Scale score. †PϽ.01 for comparison between SSRI users and nonusers. ‡PϽ.001 for comparison between SSRI users and nonusers.

continuous users. The mean adjusted rate of total hip bone CI, −0.93% to −0.31%) per year for SSRI users com- loss was −0.47% (95% CI, −0.54% to −0.42%) per year pared with −0.48% (95% CI, −0.53% to −0.43%) per year for nonusers, −0.83% (95% CI, −1.03% to −0.63%) per for nonusers (P=.36). For TCAs, results of this second- year for partial users, and −0.76% (95% CI, −1.14% to ary analysis were unchanged from those of the primary −0.38%) per year for continuous users. analysis.

TCA USE AND RATE OF HIP BONE LOSS COMMENT Average rates of hip bone loss among TCA users and antidepressant nonusers were similar in age- and multi- We found that use of SSRIs in our cohort of older women variable-adjusted analyses (Table 3). Findings were simi- was independently associated with an increased rate of lar at the femoral neck and trochanter. hip bone loss. Use of a TCA was not similarly associated with increased rates of hip bone loss in our cohort. SUBJECTS WITH GDS SCORE One potential explanation for our findings is that SSRI LESS THAN 6 use may have a direct deleterious effect on bone. This theory is supported by findings of in vitro and in vivo When 336 women with a GDS score of at least 6 at visit laboratory investigations. Functional receptors for sero- 6 or 8 were excluded, the mean rate of hip bone loss for tonin and serotonin transporter systems have been iden- SSRI users remained higher than that for nonusers, al- tified in osteoblasts, osteoclasts, and osteocytes.4,5,23 Sero- though the magnitude of the difference between the tonin has been shown to induce murine osteoblast groups was less pronounced. In the multivariable model, proliferation and human osteoclast differentiation in the mean rate of bone loss at the total hip among SSRI vitro.24 Other in vitro data suggest that fluoxetine hy- users with a GDS score less than 6 was −0.68% (95% CI, drochloride inhibits osteoblast formation24 and reduces −0.89% to −0.47%) per year compared with −0.43% (95% osteoclast differentiation.4,24 These findings suggest that CI, −0.48% to −0.38%) per year among nonusers (P=.02). a reduction in osteoblast activity or a reduction in coupled For TCA users with a GDS score less than 6, the mean osteoclast/osteoblast activity owing to serotonin trans- rate of bone loss at the total hip was −0.42% vs −0.43% porter inhibition could be a potential mechanism by which per year for nonusers (P=.99). Similar results were ob- SSRIs may influence BMD. served at the other subregions of the hip (results not In vivo, both young and adult mice with a null mu- shown). tation in the gene encoding for the serotonin transporter had reduced bone mass, altered skeletal architec- USE OF SSRIs AND TCAs AT VISIT 6 ture, and inferior mechanical properties,25 suggesting a role for the serotonin transporter in bone metabolism. We also performed analyses limited to subjects who were Growing mice treated with an SSRI demonstrated re- taking an SSRI or a TCA at visit 6 compared with non- duced bone mineral accrual.25 In both models (genetic users. Use of an SSRI was less common at visit 6 than at disruption of the serotonin transporter and pharmaco- visit 8. In this analysis, the association between SSRI use logic inhibition of it), bone formation rates were re- and the rate of change in hip BMD was weaker and did duced, indicating that osteoblast function is signifi- not reach significance. In the multivariable model, the cantly reduced in vivo, with inhibition of serotonin mean rate of bone loss at the total hip was −0.62% (95% transporter function.

(REPRINTED) ARCH INTERN MED/ VOL 167, JUNE 25, 2007 WWW.ARCHINTERNMED.COM 1243

Downloaded From: https://jamanetwork.com/ on 09/26/2021 On the other hand, evidence in humans that blockade sequent rates of bone loss. We observed an increase in of serotonin reuptake has a negative effect on bone me- the prevalence of SSRI use in our cohort between visit 6, tabolism is limited. Recently, Haney et al26 reported an as- which occurred in 1997 through 1998, and visit 8, which sociation between SSRI use and decreased BMD in older occurred in 2002 through 2004, consistent with other men. In other cross-sectional analyses,8,9,27,28 use of anti- data.1 Future research with larger numbers of SSRI us- depressants has not been associated with reduced BMD, ers will be important to determine whether SSRI use is although these other analyses have generally not sepa- prospectively associated with increased rates of bone loss. rated TCA use from SSRI use. Use of antidepressants has We did not find that use of TCAs, which also have an been linked to an increased risk of fractures,27,29,30 al- effect on serotonin reuptake, was associated with an in- though the mechanisms underlying this association re- creased rate of hip bone loss. There are several potential main unclear. Owing to their limitations, cross-sectional explanations for this finding. Tricyclic antidepressants studies might underestimate or overestimate any associa- are often prescribed for reasons other than depression, tion between antidepressant use and bone density. such as sleep disorders or chronic pain. As a result, fewer Observational studies examining possible associa- subjects receiving TCAs than the number receiving SSRIs tions among antidepressant use, bone density, and frac- may have underlying depression. When prescribed for tures are subject to confounding, which may also ex- treatment of insomnia or chronic pain, TCAs are often plain our findings. In particular, confounding by prescribed at lower doses than those used to treat de- indication may be an important issue. Antidepressants pression; as a result, the degree of serotonin blockade may are often prescribed for depressive symptoms, and de- be lower than that associated with SSRI use. Alterna- pression itself has been associated with a lower BMD.8-10 tively, we may not have observed an association be- Depression has been postulated to have a direct effect on tween TCA use and rates of bone loss because the de- bone density, through such pathways as alterations in the gree of serotonin transporter inhibition differs among the hypothalamic-pituitary-adrenal system and up- many TCAs; for example, desipramine’s potency at the regulation of the proinflammatory cytokines interleu- serotonin transporter is lower than that of fluoxetine in kin 6 and tumor necrosis factor ␣.9,31,32 In our study, when osteoblasts.5 women with a GDS score of at least 6 were excluded, the Our findings suggest that, in this cohort, use of SSRIs magnitude of the difference in rates of hip bone loss be- is associated with increased rates of hip bone loss. Al- tween SSRI users and nonusers was attenuated, suggest- though some of this association is likely due to con- ing that confounding by indication at least partially ex- founding by indication, further investigation of SSRI use plains our findings. The GDS measurement was only and rates of change in BMD in other populations with available at the clinic examinations; thus, we cannot ac- longer follow-up is warranted given the recent descrip- count for depressive symptoms that may have been pres- tion of serotonin transporters in bone. ent between visits. In addition, although the GDS is a use- ful screening tool for depression, it cannot substitute for Accepted for Publication: January 22, 2007. a clinical diagnosis of depression based on established Correspondence: Susan J. Diem, MD, MPH, Epidemiol- diagnostic criteria. ogy Clinical Research Center, University of Minnesota, Medical conditions associated with increased loss of 1100 Washington Ave S, Suite 201, Minneapolis, MN bone density, such as chronic obstructive pulmonary dis- 55415 ([email protected]). ease, liver disease, and diabetes mellitus,33 may also pre- Author Contributions: Dr Diem had full access to all the dispose to depression; thus, patients with these condi- data in the study and takes responsibility for the integ- tions may be more likely to be prescribed antidepressants. rity of the data and the accuracy of the data analysis. Study In addition, depression is associated with decreased mo- concept and design: Diem, Stone, and Ensrud. Acquisi- bility and weight loss, both of which have effects on bone. tion of data: Diem, Stone, and Ensrud. Analysis and in- Owing to concerns about the adverse effects of TCAs, terpretation of data: Diem, Blackwell, Stone, Yaffe, Haney, SSRIs may be preferentially prescribed to participants per- Bliziotes, and Ensrud. Drafting of the manuscript: Diem. ceived to be at higher risk for falls because of comorbidi- Critical revision of the manuscript for important intellec- ties; these comorbidities may also predispose them to tual content: Blackwell, Stone, Yaffe, Haney, Bliziotes, and higher rates of bone loss, another potential source of con- Ensrud. Statistical analysis: Blackwell, Stone, and Yaffe. founding. To address these potential confounders, we ad- Obtained funding: Stone and Ensrud. Study supervision: justed for health status, functional status, and weight Stone. change, although unmeasured factors might explain our Financial Disclosure: Dr Diem has participated in trials results. funded by Pfizer Inc, Eli Lilly and Company, and Merck Because our cohort consists of elderly women, we can- & Co Inc in the area of osteoporosis treatment and pre- not extrapolate our findings to other populations. Be- vention. Dr Ensrud reports research funding from Pfizer cause we have limited information on dose and dura- Inc, Eli Lilly and Company, and Bionovo. Dr Haney has tion of use of antidepressants in the cohort, we were participated in trials funded by Sanofi-Synthelabo and limited in our ability to look for evidence of a dose effect Pfizer Inc that did not involve treatments for depression of antidepressants on the rate of change of BMD. We did or osteoporosis. Dr Bliziotes has acted as a consultant for not find evidence that continuous users had higher rates and received research grants and honoraria from Merck of bone loss than partial users. In addition, the small num- & Co Inc and Proctor & Gamble. ber of SSRI users at visit 6 was insufficient to determine Funding/Support: This study was supported by grants whether SSRI use was prospectively associated with sub- AG05407, AG05394, and AG08415 from the National In-

(REPRINTED) ARCH INTERN MED/ VOL 167, JUNE 25, 2007 WWW.ARCHINTERNMED.COM 1244

Downloaded From: https://jamanetwork.com/ on 09/26/2021 stitute on Aging and by grants AR35582, AR35583, 16. Cummings SR, Block G, McHenry K, et al. Evaluation of two food frequency meth- AR35584, and K23 AR051926 from the National Insti- ods of measuring dietary calcium intake. Am J Epidemiol. 1987;126:796-802. 17. Ensrud KE, Nevitt MC, Yunis C, et al. Correlates of impaired function in older tute of Arthritis and Musculoskeletal and Skin Diseases. women. J Am Geriatr Soc. 1994;42:481-489. 18. Fitti JE, Kovar MG. The supplement on aging to the 1984 National Health Inter- REFERENCES view Survey. Vital Health Stat 1. 1987:1-115. 19. Pincus T, Summey JA, Soraci SA, et al. Assessment of patient satisfaction in activities of daily living using a modified Stanford Health Assessment Questionnaire. 1. Trends in antidepressant use. AHRQ News and Numbers, May 16, 2005. Rock- Arthritis Rheum. 1983;26:1346-1353. ville, Md: Agency for Healthcare Research and Quality. http://www.ahrq.gov /news/nn/nn051605.htm. Accessed February 3, 2006. 20. Folstein MF, Robins LN, Helzer JE. The Mini-Mental State Examination. Arch Gen 2. Harman JS, Crystal S, Walkup J, Olfson M. Trends in elderly patients’ office vis- Psychiatry. 1983;40:812. its for the treatment of depression according to physician specialty: 1985-1999. 21. Lohman TG, Roche AF, Mavtorell R. Anthropometric Standardization Reference J Behav Health Serv Res. 2003;30:332-341. Manual. Champaign, Ill: Human Kinetics Books; 1988:177. 3. Stahl SM. Basic psychopharmacology of antidepressants, I: antidepressants have 22. Ensrud KE, Walczak TS, Blackwell T, Ensrud ER, Bowman PJ, Stone KL. Anti- seven distinct mechanisms of action. J Clin Psychiatry. 1998;59(suppl 4):5-14. epileptic drug use increases rates of bone loss in older women: a prospective 4. Battaglino R, Fu J, Spate U, et al. Serotonin regulates osteoclast differentiation study. Neurology. 2004;62:2051-2057. through its transporter. J Bone Miner Res. 2004;19:1420-1431. 23. Bliziotes M, Eshleman A, Burt-Pichat B, et al. Serotonin transporter and recep- 5. Bliziotes MM, Eshleman A, Zhang X, Wiren K. Neurotransmitter action in osteo- tor expression in osteocytic MLO-Y4 cells. Bone. 2006;39:1313-1321. blasts: expression of a functional system for serotonin receptor activation and doi:10.1016/j.bone.2006.06.009. reuptake. Bone. 2001;29:477-486. 24. Gustafsson BI, Thommesen L, Stunes AK, et al. Serotonin and fluoxetine modu- 6. Cummings SR, Black DM, Nevitt MC, et al; Study of Osteoporotic Fractures Re- late bone cell function in vitro. J Cell Biochem. 2006;98:139-151. search Group. Appendicular bone density and age predict hip fracture in women. 25. Warden SJ, Robling AG, Sanders MS, Bliziotes MM, Turner CH. Inhibition of the JAMA. 1990;263:665-668. serotonin (5-hydroxytryptamine) transporter reduces bone accrual during growth. 7. Vogt MT, Rubin DA, Palermo L, et al. Lumbar spine listhesis in older African Ameri- Endocrinology. 2005;146:685-693. can women. Spine J. 2003;3:255-261. 26. Haney EM, Chan BKS, Diem SJ, et al; Osteoporotic Fractures in Men Study Group. 8. Robbins J, Hirsch C, Whitmer R, Cauley J, Harris T. The association of bone min- Association of low bone mineral density with selective serotonin reuptake in- eral density and depression in an older population. J Am Geriatr Soc. 2001; hibitor use by older men. Arch Intern Med. 2007;167:1246-1251. 49:732-736. 27. Ensrud KE, Blackwell T, Mangione CM, et al. Central nervous system active medi- 9. Michelson D, Stratakis C, Hill L, et al. Bone mineral density in women with cations and risk for fractures in older women. Arch Intern Med. 2003;163: depression. N Engl J Med. 1996;335:1176-1181. 949-957. 10. Schweiger U, Weber B, Deuschle M, Heuser I. Lumbar bone mineral density in 28. Kinjo M, Setoguchi S, Schneeweiss S, Solomon D. Bone mineral density in sub- patients with major depression: evidence of increased bone loss at follow-up. jects using central nervous system–active medications. Am J Med. 2005;118: Am J Psychiatry. 2000;157:118-120. 1414.e7-1414.e12. doi:10.1016/j.amjmed.2005.07.033. 11. Pahor M, Chrischilles EA, Guralnik JM, et al. Drug data coding and analysis in 29. Liu B, Anderson G, Mittman N, To T, Axcell T, Shear N. Use of selective serotonin- epidemiologic studies. Eur J Epidemiol. 1994;10:405-411. reupate inhibitors or tricyclic antidepressants and risk of hip fractures in elderly 12. Steiger P, Cummings SR, Black DM, Spencer NE, Genant HK. Age-related dec- people. Lancet. 1998;351:1303-1307. rements in bone mineral density in women over 65. J Bone Miner Res. 1992; 7:625-632. 30. Hubbard R, Farrington P, Smith C, Smeeth L, Tattersfield A. Exposure to tricy- 13. Ensrud KE, Palermo L, Black DM, et al. Hip and calcaneal bone loss increase clic and selective serotonin reuptake inhibitor antidepressants and the risk of hip with advancing age: longitudinal results from the study of osteoporotic frac- fracture. Am J Epidemiol. 2003;158:77-84. tures. J Bone Miner Res. 1995;10:1778-1787. 31. Furlan PM, Ten Have T, Cary M, et al. The role of stress-induced cortisol in the 14. Sheikh JI, Yesavage JA. Geriatric Depression Scale (GDS): recent evidence and relationship between depression and decreased bone mineral density. Biol development of a shorter version. In: Brink TL, ed. Clinical Gerontology: A Guide Psychiatry. 2005;57:911-917. to Assessment and Intervention. Binghamton, NY: Haworth Press Inc; 1986: 32. Kahl KG, Rudolf S, Stoeckelhuber BM, et al. Bone mineral density, markers of 165-173. bone turnover, and cytokines in young women with borderline personality dis- 15. Gerety MB, Williams JW Jr, Mulrow CD, et al. Performance of case-finding tools order with and without comorbid major depressive disorder. Am J Psychiatry. for depression in the nursing home: influence of clinical and functional charac- 2005;162:168-174. teristics and selection of optimal threshold scores. J Am Geriatr Soc. 1994; 33. Fitzpatrick LA. Secondary causes of osteoporosis. Mayo Clin Proc. 2002;77:453- 42:1103-1109. 468.

(REPRINTED) ARCH INTERN MED/ VOL 167, JUNE 25, 2007 WWW.ARCHINTERNMED.COM 1245

Downloaded From: https://jamanetwork.com/ on 09/26/2021