Effects of antidepressants on postmenopausal bone loss — A 5-year longitudinal study from the OSTPRE cohort

    Effects of antidepressants on postmenopausal bone loss – A 5-year longitudinal study from the OSTPRE cohort P¨aivi H. Rauma, Risto J. Honkanen, Lana J. Williams, Marjo T. Tuppurainen, Heikki P. Kr¨oger, Heli Koivumaa-Honkanen PII: DOI: Reference:

S8756-3282(16)30130-2 doi: 10.1016/j.bone.2016.05.003 BON 11040

To appear in:

Bone

Received date: Revised date: Accepted date:

15 December 2015 16 April 2016 10 May 2016

Please cite this article as: Rauma P¨ aivi H., Honkanen Risto J., Williams Lana J., Tuppurainen Marjo T., Kr¨ oger Heikki P., Koivumaa-Honkanen Heli, Effects of antidepressants on postmenopausal bone loss – A 5-year longitudinal study from the OSTPRE cohort, Bone (2016), doi: 10.1016/j.bone.2016.05.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

1

Effects of antidepressants on postmenopausal bone loss – A 5-year longitudinal study from the OSTPRE cohort

T

Running title: Antidepressants and bone

NU

SC R

IP

Päivi H. Rauma1,2*, MPharm, PharmD; [email protected] Risto J. Honkanen2, MD, PhD; [email protected] Lana J. Williams3, BPsych, GradDipAppPsych, MPsych (Clinical), PhD; [email protected] Marjo T. Tuppurainen2,4, MD, PhD; [email protected] Heikki P. Kröger2,5, MD, PhD; [email protected] Heli Koivumaa-Honkanen6-8, MD, PhD; [email protected] 1

CE P

TE

D

MA

Social Pharmacy, School of Pharmacy, Faculty of Health Sciences, University of Eastern Finland (UEF), Kuopio, Finland 2 Bone and Cartilage Research Unit, Surgery, Institute of Clinical Medicine, UEF, Kuopio, Finland 3 School of Medicine, Deakin University, Geelong, Australia 4 Department of Obstetrics and Gynaecology, KUH, Kuopio, Finland 5 Department of Orthopaedics, Traumatology and Handsurgery, KUH, Kuopio, Finland 6 Institute of Clinical Medicine, Psychiatry, UEF, Kuopio, Finland 7 Clinic of Child Psychiatry, Oulu University Hospital, Oulu, Finland 8 Departments of Psychiatry: Kuopio University Hospital (KUH), South-Savonia Hospital District, Mikkeli; North Karelia Central Hospital, Joensuu; SOSTERI, Savonlinna; SOTE, Iisalmi; Lapland Hospital District, Rovaniemi, Finland; University of Oulu, Finland

AC

*Correspondence to: Päivi Rauma, University of Eastern Finland, Bone and Cartilage Research Unit, P.O. Box 1627, 70211 Kuopio, Finland,+358408425823, [email protected] Number of Tables: 4, Number of Pictures: 1 Conflict of Interest All authors state that they have no conflicts of interest. Source of Funding Research has been supported by the strategic of the University of Eastern Finland, by the Academy of Finland (grants 115969 and 250707), Ministry of Education and Culture and Kuopio University Hospital Funding (VTR). PR received research support from Finnish Concordia Fund, National Doctoral Programme of Musculoskeletal Disorders and Biomaterials, North Savo Regional Fund of Finnish Cultural Foundation and Saastamoinen Foundation, and an EVO grant from Lapland Hospital District. HK-H and RH received an EVO grant from Lapland Hospital District. LW received grant funding from Eli Lilly, Pfizer, The University of Melbourne, Deakin University and the National Health and Medical Research Council (NHMRC) and is currently supported by a NHMRC Career Development Fellowship (GNT 1064272). We would like to thank Ms. Seija Oinonen for technical help.

ACCEPTED MANUSCRIPT

2

ABSTRACT

T

Background: Osteoporosis and depression are major health problems worldwide. The

IP

association between antidepressants, a treatment for depression, and bone health needs more

SC R

detailed exploration.

Objective: The present study investigates antidepressant medication use and postmenopausal bone loss over time.

NU

Methods: A total of 1,988 women (aged 57-67) participating in the Kuopio Osteoporosis

MA

Risk Factor and Prevention Study (OSTPRE) cohort responded to a postal enquiry and had their femoral neck bone mineral density (BMD) measured in 1999 and again in 2004. Data on antidepressant use was obtained from the National Prescription Register. Multiple regression

TE

D

techniques were used to test the associations, before and after adjustment for anthropometric, medical, physical and lifestyle factors.

CE P

Results: Over the five years of follow-up, 319 (16.0%) women purchased antidepressants. Mean baseline femoral neck BMD for the entire study group was 881 mg/cm2 (SD 123) and

AC

mean 5-year bone loss was 6.0 mg/cm2 (SD 4.7). After adjustments, users of tricyclic antidepressants (TCA) had greater annual BMD loss than non-users (-3.6 mg/cm2 vs. -1.1 mg/cm2; P=.031). Accelerated bone loss was also associated with selective serotonin reuptake inhibitor’s (SSRI) use (P=.001) and use of other antidepressants in a dose-response way, with the latter only among women of low-weight and normal-weight women who had lost weight over the study period. Conclusions: In conclusion, the use of SSRIs seems to accelerate postmenopausal bone loss in a dose-response manner. Associations between TCA and other antidepressant use and bone loss may also exist. Thus, the possibility of increased risk of osteoporosis should be considered when prescribing antidepressants for postmenopausal women.

ACCEPTED MANUSCRIPT

Key words: antidepressants; bone loss; bone mineral density; osteoporosis; postmenopausal

IP

T

women; selective serotonin reuptake inhibitors

SC R

Abbreviations: 5-HT, serotonin; BMD, bone mineral density; BMDfem, femoral neck BMD; BMI, body mass index; DDD, defined daily dose; DXA, dual-energy X-ray absorptiometry;

AC

CE P

TE

D

MA

NU

HT, hormone therapy; NA, noradrenaline

3

ACCEPTED MANUSCRIPT

4

1. Introduction

T

Major depression, regardless of age, is one of the leading causes of disease burden

IP

worldwide1. It is also twice as common in women compared to men and its incidence has

SC R

been shown to increase following menopause2. The use of antidepressants has doubled in the last decade3, with 8% of Europeans using them in 20104. The side effects of antidepressants, e.g., anticholinergic effects, are common and more recently have also been noted the negative

MA

NU

effects on bone5.

Antidepressants target monoamines, particularly serotonin (5-hydroxytryptamine, 5-HT) and noradrenalin (NA) in synapses by blocking their transporters6,7. Interestingly, bone cells

TE

D

(osteoblasts, osteoclasts and osteocytes) have been shown to contain both 5-HT transporters (5-HTT) and 5-HT receptors8,9. Osteoblasts, responsible for bone formation, have also been

CE P

studied to express the NA transporter (NET), with NA increasing both bone formation and resorption10,11. Recently, selective serotonin reuptake inhibitors (SSRIs), the most widely

AC

prescribed antidepressant12, have been shown to inhibit human osteoblast and osteoclast formation and function13.

The association between antidepressant use and bone has been investigated in an increasing number of epidemiological studies, mainly been cross-sectional and focused on SSRIs. Among postmenopausal women, use of SSRIs has been shown to be associated with lower bone mineral density (BMD)14-16 and accelerated bone loss over time14, although some have shown no association17,18. In regards to tricyclic antidepressants (TCA), there has generally been null findings across varying populations of women14,18,19, with one study finding the agents to actually be associated with increases in BMD15.

ACCEPTED MANUSCRIPT

5

Thus, we aimed to investigate the association between different types of antidepressants and

T

BMD change over time in a large group of postmenopausal women. To our knowledge this is

IP

the first large longitudinal and population-based study that examines quantitatively not only

SC R

SSRIs or TCAs, but also other antidepressants and poly use. As postmenopausal women are at high risk of bone loss20, this population group was selected to form the study sample.

NU

2. Methods

MA

2.1. Study design and subjects

The initial postal enquiry of the Kuopio Osteoporosis Risk Factor and Prevention Study

TE

D

(OSTPRE) was sent in 1989 to all women born between 1932-41 and residing in Kuopio Province, Finland (n=14,220). Of the 13,100 respondents, 3,222 were participating to bone

CE P

densitometry with Dual X-ray Absorptiometry (DXA). This densitometry group included 2,025 randomly selected women and all the 1,197 women, who had either menopause within

AC

two years, at least three health behavior risk factors or disease or medication that strongly increase risk of osteoporosis. Further details have been published elsewhere21. The enquiry and measurements were repeated every 5 years.

The present study included a total of 1,988 women (aged 57-67 years in 1999), who responded to the enquiry as well as underwent femoral neck DXA scans both in 1999 (baseline) and 2004 (follow-up). The study was approved by the Ethics Committee of Kuopio University Hospital and all participants provided a written informed consent.

2.2. Clinical measurements

ACCEPTED MANUSCRIPT

6

Areal BMD (g/cm2) was measured at the femoral neck site using Lunar DXA (DPX-IQ in

T

1999-2002 and Lunar Prodigy in 2004-2007). The DPX-IQ and Prodigy results were cross-

IP

calibrated and corrected BMD values were used accordingly21. Trained technicians carried out

SC R

all examinations and performed daily calibrations.

At the time of DXA, height, weight, and grip strength were measured. As an indicator of

NU

physical fitness22,23, grip strength was measured using the Martin Vigorimeter (kPa), a

MA

pneumatic squeeze dynamometer. The mean of three successive measurements in 1999 on the dominant hand was determined. Weight changes between 1999 and 2004 and body mass

CE P

2.3. Questionnaire data

TE

D

index (BMI) in 1999 were used.

Information on age of menopause, smoking, alcohol consumption, use of calcium

AC

supplements and dietary calcium intake was taken from the 1999 questionnaire. Dietary calcium intake was computed as the sum of intake from fluid milk products and cheese. Physical activity and use of bisphosphonates, hormone therapy (HT) (i.e., estrogen, progesterone), vitamin D supplements and corticosteroids were taken from the year 2004 questionnaire. Participants reported chronic health disorders including depression diagnosed by a physician in the 1999 and 2004 questionnaires.

2.4. Register data

ACCEPTED MANUSCRIPT

7

Data from OSTPRE was linked by personal identification codes with data from three national registries: (1) the Hospital Discharge Register (to determine hospitalizations due to

T

depression); (2) the National Register for Work Disability Pensions (to determine pensions

IP

due to depression) and (3) the National Prescription Register (to determine use of

SC R

antidepressants between 1999-2004). The prescription register contains data from all reimbursed prescription medications purchased in any pharmacy in Finland24-26.

NU

Antidepressants were classified according to the Anatomical-Therapeutic-Chemical (ATC) -

MA

codes N06A and N06CA27. The subgroups included (1) TCAs only; (2) SSRIs only; (3) other antidepressants (excluding monoamine oxidase inhibitors); (4) poly use (i.e., use of more than one type of antidepressant). In analysis we also examined persons using any antidepressant

TE

D

(Any AD) and persons using only one type of antidepressant, thus excluding persons in the poly subgroup (Any without other ADs). Participants were originally (i.e., in Table 1)

CE P

regarded as users of antidepressants if they had purchased any antidepressant between the two BMD-measurements. Defined Daily Doses (DDD) was used in further analyses with 1 DDD

AC

corresponding to an average of one day use27. To achieve the best overview of the studied association, the use of antidepressants was treated as a dichotomous and continuous variable. For the dichotomous variable, participants were coded as users (actual users), if they had purchased antidepressants for more than 30 DDDs (the size of a ‘starter’ package).

Depression was treated as a 3-category variable: 1) no depression, 2) depression diagnosed by a physician (enquiries in 1999 and 2004), or 3) determined from a registry (work disability or hospitalization in 1999-2004).

2.5. Statistical analyses

ACCEPTED MANUSCRIPT

8

Statistical analyses were performed using the SPSS statistical package 19.0 for Windows

T

(SPSS Inc., Chicago, IL, USA). Differences between antidepressant groups were examined

IP

using the Chi-squared test for categorical variables or ANOVA for continuous variables.

SC R

Linear regression and general linear modelling were used in multivariable analyses. Covariates and possible confounders were chosen based on earlier literature.

NU

Annual % femoral neck bone loss during the 5-year follow-up [i.e., (((BMDfollow-upBMDbaseline)/time as years between measurements)/BMDbaseline)*100%] was the outcome

MA

variable. Adjusted mean annual bone loss values (mg/cm2 and %) were performed with analysis of covariance by using the GLM procedure. In addition, multiple linear regression

TE

D

was used. First model included baseline age, BMI and weight change over the 5-year followup period and the final model included also smoking, use of alcohol, bisphosphonates,

CE P

corticosteroids, HT, calcium & vitamin D supplements, dietary calcium, grip strength, physical activity, age of menopause and depression. Separate multiple linear regression

AC

models were conducted for each antidepressant category. Interactions were tested by backward method and included in the model if significant. Multicollinearity was not found. Locally weighted scatterplot smoothing (LOESS) curves from simple scatterplots were used to determine the associations between antidepressant use (as DDDs) and annual % bone loss.

3. Results 3.1. Study population characteristics

According to the prescription register, 319 of the 1,988 participants (16.0%) had purchased antidepressants during the 5-year follow-up. Of all the participants, 92 (4.6%) had used only

ACCEPTED MANUSCRIPT

9

TCAs, 118 (5.9%) only SSRIs and 56 (2.8%) only other antidepressants, while 52 (2.6%) women were poly users (Table 1). Moreover, one woman had used MAO inhibitors. The

T

group of other antidepressants included users of mirtazapine (n=40), mianserin (n=7),

SC R

IP

venlafaxine (n=5) and trazodone (n=4).

Mean baseline femoral neck BMD (BMDfem) for the entire sample was 881.1 mg/cm2 (SD 123.1), and the mean 5-year bone loss was -6.0 mg/cm2 (SD 4.7). Characteristics of the study

NU

subjects are shown in Table 1, with differences evident across the groups according to

MA

antidepressant use. For example, weight increase during the follow-up period was highest among users of SSRI and poly users. In addition, poly users had higher rates for hospitalization and work disability pension due to depression as well as self-reported

TE

D

depression than other participants. They also purchased the greatest amount of

CE P

antidepressants.

AC

3.2. Effects of antidepressant use on bone

Annual % femoral neck bone loss did not differ statistically between ‘any antidepressant users’ and non-users (Table 2). However, a trend for increasing bone loss in antidepressant users was seen following adjustment for age, BMI and weight change (model 1, B=-0.120, P=.085) but not in the final model (Table 3). A similar trend was evident when antidepressant use was treated as a continuous variable (i.e., amount of purchased antidepressants) (Table 3).

Following the exclusion of ‘other antidepressants’, which included mirtazapine, mianserin, venlafaxine and trazodone, antidepressant use was associated with bone loss (Table 3).

ACCEPTED MANUSCRIPT

10

3.2.1. TCAs

T

TCAs were associated with bone loss before and after adjustments (Tables 2 and 3). The

IP

strongest skeletal effects were seen among those who had been using for less than one year

SC R

(<360 DDDs) or at a lower dose than recommended for depression (Figure 1). When TCA use was investigated as a continuous variable the relationship was not sustained (Table 3).

MA

NU

3.2.2. SSRIs

SSRI use was significantly associated with bone loss in a dose-dependent fashion, but not once dichotomized (Tables 2 and 3). The association was strong and dose-dependent for those

TE

D

who had purchased medication for more than one year use (Figure 1).

CE P

3.2.3. Other antidepressants

AC

No significant association between use of other antidepressants as a dichotomous variable and bone loss was seen (Tables 2 and 3). In contrast to SSRI and TCA use, use of other antidepressants as a continuous variable was associated with an increase in BMDfem in the univariate model (Table 3). Multivariable regression revealed that BMI and weight change were effect modifiers in the association between use of other antidepressants and BMDfem: low-weight and weight decrease were associated with increased bone loss over time (Table 4).

3.2.4. Poly use

ACCEPTED MANUSCRIPT

11

Poly antidepressant use was not significantly associated with bone loss (Tables 2 and 3).

T

Removing use of other antidepressants did not change the results.

SC R

IP

4. Discussion

This longitudinal study investigated the associations between antidepressant use and postmenopausal femoral neck BMD change over a 5-year period. Accelerated bone loss was

NU

associated with the use of TCAs and SSRIs. In addition, use of other antidepressants (mainly

MA

mirtazapine) was associated with accelerated bone loss, however only in low-weight women.

Previously, Diem et al.14 reported a 0.82% annual mean total hip bone loss in postmenopausal

TE

D

SSRI users compared with a 0.47% loss in non-users in a 5-year longitudinal study. In the present study, the 5-year difference between SSRI users and non-users was somewhat smaller

CE P

(adjusted annual femoral neck bone loss of -0.22% vs. -0.08%, respectively). We also reported a dose-response relationship between bone loss and amount of SSRI use. This is

AC

supported by the work of Seifert et al.28 who found a negative correlation between BMD and duration of SSRI use in a cross-sectional study of young adults.

In the present study, TCA users had significantly greater adjusted bone loss than non-users (0.35% vs. -0.08%, respectively). In contrast to our findings, TCA use has previously been associated with increased BMD among postmenopausal women in a cross-sectional study15, whereas longitudinal studies on pre- and postmenopausal women with 5-year14 and 6-year18 follow-ups have found no association between TCA use and BMD.

ACCEPTED MANUSCRIPT

12

The use of antidepressants other than SSRIs and TCAs was associated with accelerated bone loss in our study, but only in low-weight women and women who lost weight over the study

T

period. A similar weight & antidepressant interaction has been seen in a population based

IP

study investigating use of antidepressants and BMD and heel quantitative ultrasound

SC R

(QUS)29,30. It has been hypothesized that greater fat mass, associated with biological factors, may be playing a role in the association between antidepressant use and bone31. In addition, it

NU

has been previously suggested that weight gain may also affect DXA measurements32.

MA

The mixed findings according to antidepressant group may be due to differences in antidepressant mechanisms and binding affinity to transporters and receptors. TCAs inhibit mainly serotonin and noradrenalin reuptake, but also block α-adrenergic, histamine and

TE

D

muscarine receptors6. SSRIs inhibit largely serotonin reuptake only, whereas mirtazapine, from the group of other antidepressants, blocks 5-HT2, 5-HT3, α2-adrenergic and histamine

CE P

receptors and act as an agonist for the 5-HT1 receptor, with weak effects on monoamine reuptake6. Because bone cells also contain monoamine receptors and transporters8,10,11,

bone cells.

AC

different mechanisms of antidepressant action are likely to lead to differing responses also in

Frequent but low-dose use for other indications such as for chronic pain (TCAs) or for sleep (mirtazapine) may have weakened the observed bone effects. However, the contradictory result showing a stronger relationship with low-dose TCA is likely due to study power, or even may indicate residual confounding. The results also showed that women, who had purchased more than one type of antidepressant, also purchased the highest DDD amount of antidepressants but had no increased bone loss. This result, however, may also be due to low study power. Only SSRIs displayed a dose-response effect on bone loss in the present study.

ACCEPTED MANUSCRIPT

13

When investigating the effect of antidepressants on bone, other medications, health disorders

T

as well as physiological, behavioral and psychological mechanisms should be taken into

IP

account33. One of the main confounding factors is depression itself, since it has been

SC R

suspected to have a negative effect on bone34-36. Depression affects bone via serotonin signaling and likely stress factors34-36. However, separating the skeletal effects of depression and antidepressant use is difficult. In the current study, antidepressant use was independently

NU

associated with bone loss with the association remaining after many important confounders

MA

(e.g., depression) were taken into consideration.

The strengths of this study include a large population-based cohort of postmenopausal

TE

D

women, high response rate, long-term follow-up and the possibility to adjust for several confounding factors including depression and behavioral and nutritional factors. The narrow

CE P

age range increased the homogeneity of the sample and provided information about a group that is considered to be at high risk for both osteoporosis and depression. Use of register data

AC

to study the type of antidepressant, its dose and amount purchased during the follow-up with high accuracy26 were further strengths of this study.

The National Prescription Register provides information on purchased antidepressants. However, a limitation is that we cannot verify actual use, which may weaken results. Still, according to our previous study, the agreement between self-reported regular use and prescription register data was reasonable (71%)26. In addition, the size of the antidepressant subgroups is a further limitation of this study. Lastly, the low bone loss rate may be another limiting factor explaining some of the non-associations.

ACCEPTED MANUSCRIPT

14

Although depression was tested as a confounding factor in this study, the challenge is still the separation of the skeletal effects of antidepressant use and depression from each other.

T

Moreover, not only diagnosed depression, but also low subjective well-being i.e., life

IP

dissatisfaction has been shown to independently increase bone loss in postmenopausal

SC R

women37. In future, longitudinal studies across different age and population groups are needed. Antidepressants other than SSRIs and TCAs need to be focused in respect to bone

NU

health with adequate study samples.

MA

5. Conclusions

In conclusion, the use of SSRIs seems to accelerate postmenopausal bone loss in a dose-

TE

D

response manner. A relationship between TCA and other antidepressant use with bone loss may also exist. Thus, the risk of osteoporosis should be considered when antidepressants are

CE P

prescribed for postmenopausal women.

AC

Acknowledgements

PR, HK-H and RH designed the study. PR conducted data analysis, data interpretation, and the writing of the manuscript. HK-H and RH were responsible for the data and contributed to data interpretation and writing of the manuscript. LW contributed to data interpretation and writing of the manuscript. RH, MT and HK were responsible for the original OSTPRE study design. All the authors contributed to the revision of the manuscript and approved the final version.

References

ACCEPTED MANUSCRIPT

15

World Health Organization. The Global Burden of Disease: 2004 Update. World Health Organization, Geneva, Switzerland, 2008.

2.

Eaton WW, Anthony JC, Gallo J, Cai G, Tien A, Romanoski A, Lyketsos C, Chen LS. Natural history of Diagnostic Interview Schedule/DSM-IV major depression. The Baltimore Epidemiologic Catchment Area follow-up. Arch Gen Psychiatry 1997;54(11):993-9.

3.

Organisation for Economic Co-operation and Development. Pharmaceutical consumption, in OECD, Health at a Glance 2011: OECD Indicators, OECD Publishing. doi: 10.1787/health_glance-2011-39-en

4.

Blanchflower DG and Oswald AJ (2011) Antidepressants and age. IZA Discussion paper No. 5785.

5.

Rizzoli R, Cooper C, Reginster JY, Abrahamsen B, Adachi JD, Brandi ML, Bruyére O, Compston J, Ducy P, Ferrari S, Harvey NC, Kanis JA, Karsenty G, Laslop A, Rabenda V, Vestergaard P. Antidepressant medications and osteoporosis. Bone 2012;51(3):606613.

6.

Frazer A. Pharmacology of antidepressants. J Clin Psychopharmacol. 1997;17 Suppl 1:2S-18S.

7.

Maletic V, Robinson M, Oakes T, Iyengar S, Ball SG, Russel J: Neurobiology of depression: an integrated view of key findings. Int J Clin Pract. 2007;61(12):20302040.

8.

Bliziotes M: Update in serotonin and bone. J Clin Endocrinol Metab. 2010;95(9):41244132.

9.

Warden SJ, Haney EM: Skeletal effects of serotonin (5-hydroxytryptamine) transporter inhibition: Evidence from in vitro and animal-based studies. J Musculoskelet Neuronal Interact. 2008;8(2):121-132.

10.

Jones KB, Mollano AV, Morcuende JA, Cooper RR, Saltzman CL: Bone and brain: A review of neural, hormonal, and musculoskeletal connections. Iowa Orthop J. 2004;24:123-132.

11.

Ma Y, Kruenger JJ, Redmon SN, Uppuganti S, Nyman JS, Hahn MK, Elefteriou F: Extracellular norepinephrine clearance by the norepinephrine transporter is required for skeletal homeostasis. J Biol Chem. 2013;288(42):30105-30113.

12.

National Institute for Clinical Excellence 2004 National Institute for Clinical Excellence: Depression: management of depression in primary and secondary care. London: NICE, 2004. (National Clinical Practice Guideline No 23).

13.

Hodge JM, Wang Y, Berk M, Collier FM, Fernandes TJ, Constable MJ, Pasco JA, Dodd S, Nicholson GC, Kennedy RL, Williams LJ. Selective serotonin reuptake inhibitors inhibit human osteoclast and osteoblast formation and function. Biol psychiatry. 2013;74(1):32-39.

AC

CE P

TE

D

MA

NU

SC R

IP

T

1.

ACCEPTED MANUSCRIPT

16

Diem SJ, Blackwell TL, Stone KL, Yaffe K, Haney EM, Bliziotes MM, Ensrud KE. Use of antidepressants and rates of hip bone loss in older women: the study of osteoporotic fractures. Arch Intern Med. 2007;167(12):1240–1245.

15.

Bolton JM, Targownik LE, Leung S, Sareen J, Leslie WD. Risk of low bone mineral density associated with psychotropic medications and mental disorders in postmenopausal women. J Clin Psychopharmacol. 2011;31(1):56-60.

16.

Mezuk B, Eaton WW, Golden SH, Wand G, Lee HG. Depression, antidepressants, and bone mineral density in a population-based cohort. J Gerontol A Biol Sci Med Sci. 2008; 63(12):1410-1415.

17.

Spangler L, Scholes D, Brunner RL, Robins J, Reed SD, Newtown KM, Melville JL, Lacroix AZ. Depressive symptoms, bone loss, and fractures in postmenopausal women. J Gen Intern Med. 2008;23(5):567–574.

18.

Diem SJ, Ruppert K, Cauley JA, Lian YJ, Bromberger JT, Finkelstein JS, Greendale GA, Solomon DH. Rates of bone loss among women initiating antidepressant medication use in mildlife. J Clin Endocrin Metab. 2013;98(11):4355-4363.

19.

Petronijević M, Petronijević N, Ivković M, Stefanović D, Radonjić N, Glišić B, Ristić G, Damjanović A, Paunović V. Low bone mineral density and high bone metabolism turnover in premenopausal women with unipolar depression. Bone. 2008;42(3):582-590.

20.

Cummings SR, Melton III LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet. 2002;359(9319):1761-1767.

21.

Sirola J, Salovaara K, Tuppurainen M, Jurvelin JS, Alhava E, Kröger H. Sister´s fracture history may be associated with perimenopausal bone fragility and modifies the predictability of fracture risk. Osteoporos Int. 2009;20(4):557-565.

22.

Cruz-Jentoft AJ, Baeyens JP, Bauer JM, Boirie Y, Cederholm T, Landi F, Martin FC, Michel JP, Rolland Y, Schneider SM, Topinková E, Vandewoude M, Zamboni M; European Working Group on Sarcopenia in Older People. Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age Ageing. 2010;39(4);412-423.

23.

Jakobsen LH, Rask IK, Kondrup J. Validation of handgrip strength and endurance as a measure of physical function and quality of life in healthy subjects and patients. Nutrition. 2010:26(5);542-550.

24.

Rikala M, Hartikainen S, Sulkava R, Korhonen MJ. Validity of the Finnish Prescription Register for measuring psychotropic drug exposures among elderly Finns. Drugs Aging 2010;27(4):337-349.

25.

Peura S, Rajaniemi S, Kurkijärvi U: Pharmaceutical Pricing and Reimbursement Information Project: Finland 2007. http://whocc.goeg.at/Literaturliste/Dokumente/CountryInformationReports/Finland_PP RI_2007.pdf

AC

CE P

TE

D

MA

NU

SC R

IP

T

14.

ACCEPTED MANUSCRIPT

17

Rauma PH, Koivumaa-Honkanen H, Kröger H, Tuppurainen MT, Kauhanen J, Honkanen R. The relationship between self-reported and registry-based data on use of psychoactive medications in postmenopausal women. BMC Psychiatry. 2013;13:180.

27.

World Health Organization, Collaborating Centre for Drug Statistics Methodology. Guidelines for ATC classification and DDD assignment 2011. Oslo, 2010.

28.

Seifert CF, Wiltrout TR. Calcaneal bone mineral density in young adults prescribed selective serotonin reuptake inhibitors. Clin Ther. 2013;35(9):1412-1417.

29.

Rauma PH, Pasco JA, Berk M, Stuart AL, Koivumaa-Honkanen H, Honkanen RJ, Hodge JM, Williams LJ: The association between use of antidepressants and bone quality using quantitative heel ultrasound. Aust N Z J Psychiatry. 2015;49(5):437-443.

30.

Rauma PH, Pasco JA, Berk M, Stuart AL, Koivumaa-Honkanen H, Honkanen RJ, Hodge JM, Williams LJ: The association between major depressive disorder, use of antidepressants and bone mineral density (BMD) in men. J Musculoskelet Neuronal Interact. 2015;15(2):177-185.

31.

Saarelainen J, Honkanen R, Kröger H, Tuppurainen M, Jurvelin JS, Niskanen L. Body fat distribution is associated with lumbar spine bone density independently of body weight in postmenopausal women. Maturitas. 2011;69(1):86-90.

32.

Saarelainen J, Kiviniemi V, Kröger H, Tuppurainen M, Niskanen L, Jurvelin J, Honkanen R. Body mass index and bone loss among postmenopausal women: The 10year follow-up of the OSTPRE cohort. J Bone Miner Metab. 2012;30(2):208-216.

33.

Mezuk B, Eaton WW, Golden SH. Depression and osteoporosis: epidemiology and potential mediating pathways. Osteoporos Int. 2008;19(1):1-12.

34.

Cizza G, Primma S, Csako G. Depression as a risk factor for osteoporosis. Trends Endocrinol Metab. 2009;20(8):367-373.

35.

Yirmiya R, Bab I. Major depression is a risk factor for low bone mineral density: A meta-analysis. Biol Psychiatry. 2009;66(5):423–432.

36.

Fernandes BS, Hodge JM, Pasco JA, Berk M, Williams LJ. Effects of depression and serotonergic antidepressants on bone: Mechanics and implications for the treatment of depression. Drugs Aging. 2015 Nov 7:1-5. [Epub ahead of print]

37.

Rauma PH, Koivumaa-Honkanen H, Williams LJ, Tuppurainen MT, Kröger HP, Honkanen RJ. Life satisfaction and bone mineral density among postmenopausal women: cross-sectional and longitudinal associations. Psychosom Med. 2014;76(9):709715.

AC

CE P

TE

D

MA

NU

SC R

IP

T

26.

ACCEPTED MANUSCRIPT 18

Any ADe

TCA only

SSRI only

Other AD only

Poly use

(n=1,669)

(n=319)

(n=92)

(n=118)

(n=56)

(n=52)

continuous variables

mean (SD)

mean (SD)

mean (SD)

mean (SD)

mean (SD)

mean (SD)

Age (yr) a

63.7 (2.8)

63.6 (2.9)

63.4 (2.8)

63.4 (2.9)

63.4 (2.9)

64.3 (2.9)

28.1 (4.6)

28.0 (4.5)

28.4 (4.9)

27.7 (4.2)

27.5 (4.5)

28.8 (4.6)

71.9 (12.2)

71.4 (12.1)

72.0 (12.1)

70.3 (11.9)

70.6 (11.8)

73.7 (12.9)

Weight change (kg) b

0.62 (5.02)

-0.67 (4.75)*

1.25 (5.26)**

0.81 (4.19)

1.26 (5.49)*

880.5 (121.8)

884.2 (130.1)

888.7 (130.3)

866.6 (133.6)

893.8 (128.0)

908.7 (121.4)

2

875.1 (116.6)

874.5 (120.9)

872.7 (121.9)

858.2 (114.6)

891.9 (130.4)

898.5 (119.9)

-5.4 (44.9)

-9.7 (55.1)

-16.1 (56.9)**

-8.4 (59.2)

-0.2 (51.1)

-10.2 (45.9)

-1.1 (9.1)

-1.9 (11.1)

-3.1 (11.0)**

-1.8 (12.4)

-0.00 (0.01)

-1.9 (8.9)

-0.08 (1.04)

-0.16 (1.19)

-0.29 (1.18)*

-0.11 (1.29)

-0.02 (1.16)

-0.19 (0.98)

49.6 (4.7)

49.3 (4.4)

49.0 (5.1)

49.2 (3.9)

50.2 (4.1)

48.9 (4.4)

851.5 (372.6)

833.5 (344.2)

829.9 (361.6)

789.9 (311.1)*

906.2 (397.5)

855.5 (318.6)

71.7 (18.2)

70.0 (19.6)

67.7 (19.4)**

70.2 (17.8)

77.7 (19.8) **

65.1 (21.6) **

-

511.1 (692.2)

228.8 (366.4)

531.5 (641.7)

290.0 (328.1)

1210.2 (993.8)

n (%)

n (%)

n (%)

n (%)

n (%)

29 (9.1)

9 (9.8)

13 (11.0)

4 (7.1)

3 (5.8)

2

BMDfem change in 1999-2004 (mg/cm ) 2

Annual BMDfem change (mg/cm ) Annual % bone loss

b

Age of menopause (yr) Dietary calcium (g/day) Grip strength (kPa)

a

a

Amount of purchased AD in DDDs categorical variables Smoking

n (%)

a

Alcohol intake (g/month)

b

125 (7.5)

TE D

BMDfem in 2004 (mg/cm )

CE P

BMDfem in 1999 (mg/cm )

MA N

0.19 (4.53) 2

AC

Weight (kg)

a

CR

BMI (kg/m )

US

2 a

IP

Non-users

T

Table 1: Characteristics of the study population according to purchased medication between 1999 and 2004

a

None

730 (44.5)

145 (46.6)

46 (52.3)

52 (44.8)

23 (41.8)

23 (45.1)

< 360

888 (54.1)

157 (50.5)

41 (46.6)

60 (51.7)

30 (54.5)

26 (51.0)

≥ 360

24 (1.5)

9 (2.9)

1 (1.1)

4 (3.4)

2 (3.6)

2 (3.9)

Sweaty physical activity c

**

***

none

169 (10.4)

49 (15.8)

11 (12.2)

18 (15.5)

5 (9.3)

15 (30.0)

1-2 times per week

565 (34.7)

107 (34.4)

32 (35.6)

39 (33.6)

21 (38.9)

15 (30.0)

ACCEPTED MANUSCRIPT 19 ≥3 times per week

895 (54.9)

155 (49.8)

47 (52.2)

59 (50.9)

28 (51.9)

20 (40.0)

128 (7.7)

29 (9.1)

7 (7.6)

12 (10.2)

5 (8.9)

5 (9.6)

184 (11.0)

26 (8.2)*

9 (9.8)

12 (10.2)

3 (5.4)

1 (1.9) **

234 (14.0)

41 (12.9)

10 (10.9)

12 (10.2)

11 (19.6)

7 (13.5)

440 (26.4)

96 (30.1) *

27 (29.3)

38 (32.2)

15 (26.8)

15 (28.8)

110 (6.6)

33 (10.3) **

7 (7.6)

15 (12.7)**

5 (8.9)

6 (11.5)

-

511.1 (692.2)

228.8 (366.4)

531.5 (641.7)

290.0 (328.1)

1210.2 (993.8)

58 (18.2)

20 (21.7)

21 (17.8)

17 (30.4)

0 (0.0)

40 (43.5)

26 (22.0)

14 (25.0)

4 (7.7)

16 (17.4)

21 (17.8)

8 (14.3)

6 (11.5)

7 (7.6)

14 (11.9)

6 (10.7)

5 (9.6)

3 (3.3)

10 (8.5)

4 (7.1)

6 (11.5)

1 (1.1)

3 (2.5)

2 (3.6)

6 (11.5)

5 (5.4)

23 (19.5)

5 (8.9)

25 (48.1)

HT

b

Corticosteroids

c

Amount of purchased AD in DDDs

b

0-30

IP CR

Vitamin D supplements

c

US

Calcium supplements

a

MA N

Bisphosphonates c

T

Medication use

-

85 (26.6)

181-360

-

51 (16.0)

361-540

-

32 (10.0)

541-720

-

23 (7.2)

721-900

-

12 (3.8)

> 900

-

58 (18.2)

0 (0)

11 (3.4)***

1 (1.1)*

1 (0.8)*

1 (1.8)**

8 (15.4)***

23 (1.4)

48 (15.0)***

4 (4.3)**

16 (13.6)***

8 (14.3)***

20 (38.5)***

5 (1.6) **

0 (0)

3 (2.5) **

0 (0)

2 (3.8) **

Self-reported depression d New work disability due to depression

b

3 (0.2)

CE P

depression b

AC

Hospitalization (no/yes) due to

TE D

31-180

Abbreviations: AD, antidepressant; BMDfem, femoral bone mineral density; BMI, body mass index; DDD, defined daily dose; HT, hormone therapy; SD, standard deviation; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant a

in 1999; b in 1999-2004; c in 2004; d in 1999 or 2004; e users included one participant who used monoamine oxidase inhibitors

* P-value<.1; ** P-value<.05; *** P-value<.001 compared to non-users (from Anova / Pearson chi-square)

ACCEPTED MANUSCRIPT 20

Table 2: Adjusted mean annual bone loss (mg/cm2 or %) of femoral neck according to antidepressant subgroups mg/cm2

(95% CI)

No use a (Referent)

-1.06

(-1.52, -0.60)

Any

-2.07

(-3.33, -0.80)

.15

-0.17

Any without other ADs

-2.76

(-4.13, -1.39)

.022

-0.24

TCA only

-3.60

(-5.85, -1.35)

.031

SSRI only

-2.81

(-4.80, -0.81)

.097

Other ADs only

1.73

(-1.39, 4.84)

.084

Poly use

-1.00

(-3.97, 1.97)

.97

(95% CI)

-0.08

(-0.13, -0.03)

P value *

CR

IP

T

%

(-0.31, -0.03)

.25

(-0.39, -0.08)

.055

-0.35

(-0.61, -0.10)

.039

-0.22

(-0.44, 0.01)

.25

0.22

(-0.14, 0.57)

.11

-0.07

(-0.41, 0.26)

.98

TE D

MA N

US

P value *

Abbreviations: AD, antidepressant; SE, standard error; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant. Adjusted for age, BMI, baseline BMD,

CE P

weight change, smoking, use of alcohol, bisphosphonates, corticosteroids, hormone therapy, calcium and vitamin D supplements, dietary calcium, physical activity, grip

in 1999-2004

AC

strength, age of menopause and depression. Performed by analysis of covariance with GLM procedure. *Compared to non-users; a No use of antidepressants or use <30 DDDs

ACCEPTED MANUSCRIPT 21

Table 3: Linear regression analyses for the associations between use of antidepressants and 5-year annual % femoral neck bone loss SE

P value

B

SE

SE

P value

No/Yesa

-0.102

0.071

.15

-0.120

0.070

.085

-0.089

0.077

.25

Continuousb

-0.008

0.007

.26

-0.012

.091

-0.007

0.008

.37

No/Yesa

-0.167

0.076

.028

-0.178

0.075

.017

-0.162

0.083

.050

Continuousb

-0.013

0.007

.082

-0.016

0.007

.025

-0.010

0.008

.21

No/Yesa

-0.327

0.127

.010

-0.293

0.125

.019

-0.270

0.129

.037

Continuousb

-0.031

0.026

.22

-0.025

0.025

.33

-0.004

0.027

.89

No/Yesa

-0.088

0.111

.43

-0.130

0.109

.23

-0.163

0.116

.16

Continuousb

-0.025

0.012

.038

-0.034

0.012

.004

-0.046

0.014

.001

No/yesa

0.269

0.171

.12

-1.515*

1.009

.13

-2.198*

1.165

.059

Continuousb

0.087

0.032

.007

-0.338*

0.161

.036

-0.416*

0.172

.016

No/yesa

-0.108

0.148

.46

-0.124

0.146

.39

-0.102

0.179

.57

Continuousb

-0.003

0.009

.74

-0.004

0.009

.63

-0.002

0.011

.86

Any

SSRI only

Other ADs only

CE P

TCA only

0.007

MA N

Any without other ADs

P value

CR

B

US

B

Adjusted model 2 (final model)

TE D

IP

T

Adjusted model 1

AC

Unadjusted

Poly use

ACCEPTED MANUSCRIPT 22 Abbreviations: AD, antidepressant; BMI, body mass index; DDD, defined daily dose; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant. Model 1 includes age, BMI, weight change. Final model includes age, BMI, baseline BMD, weight change, smoking, use of alcohol, bisphosphonates, corticosteroids, hormone

T

therapy, calcium and vitamin D supplements, dietary calcium, grip strength, physical activity, age of menopause, depression and interaction terms. *Included ‘weight

IP

change’*’antidepressant use’ and BMI*’antidepressant use’ interaction terms as well as other significant interactions. No = no use or use <30 DDDs, Yes = use ≥30 DDDs in 1999-2004

b

Antidepressant use in 1999-2004 as a continuous variable (DDDs) divided by 100 and including non-use.

AC

CE P

TE D

MA N

US

CR

a

ACCEPTED MANUSCRIPT

AC

CE P

TE D

MA N

US

CR

IP

T

23

Figure 1: Associations between antidepressant use as DDDs and annual % femoral neck bone loss using LOESS curve. Abbreviations: BMDfem, femoral bone mineral density; DDD, defined daily dose; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant.

ACCEPTED MANUSCRIPT 24

IP

T

Table 4: Adjusted estimated mean annual % femoral neck bone loss according to baseline BMI and 5-year weight change between other

CR

antidepressant users and non-users

Estimated mean annual % femoral neck bone loss 20

25

Weight

Non-

change

users

-10 kg

-2.4

-3.5

-1.1

-2.4

-3.0

-0.6

-2.5

-2.6

-0.1

-5 kg

-1.4

-2.2

-0.8

-1.5

-1.7

-0.2

-1.5

-1.3

0 kg

-0.5

-0.8

-0.3

-0.5

-0.4

0.1

-0.5

5 kg

0.5

0.5

0.0

0.4

0.9

0.5

10 kg

1.4

1.8

0.4

1.4

2.2

0.8

Difference* Non-

Users

CE P

TE D

users

Difference* Non-

MA N

Users

30

US

BMI

35 Users

Difference* Non-

Users

Difference*

-2.5

-2.2

0.3

0.2

-1.5

-0.9

0.6

0.0

0.5

-0.6

0.4

1.0

0.4

1.3

0.9

0.4

1.8

1.4

1.4

2.6

1.2

1.3

3.1

1.8

users

users

AC

Abbreviations: BMI, body mass index. Adjusted for age, BMI, baseline BMD, weight change, smoking, use of alcohol, bisphosphonates, corticosteroids, hormone therapy, calcium and vitamin D supplements, dietary calcium, grip strength, physical activity, age of menopause, depression and interaction terms from linear regression and calculated using different weights. * Difference between non-users and users.

ACCEPTED MANUSCRIPT 25

Highlights SSRIs are associated with increased bone loss in a dose-response manner among postmenopausal women



Tricyclic antidepressants showed an association with bone loss



Use of other antidepressants are associated with bone loss, however only among low-weight women



Increased risk of bone loss should be noticed when prescribing antidepressants for postmenopausal women

AC

CE P

TE D

MA N

US

CR

IP

T

