- Email: [email protected]
Antidepressant use and functional limitations in U.S. older adults
Journal of Psychosomatic Research 80 (2016) 31–36
Contents lists available at ScienceDirect
Journal of Psychosomatic Research
Antidepressant use and functional limitations in U.S. older adults Ruopeng An a,⁎, Lingyun Lu b a b
Department of Kinesiology and Community Health, College of Applied Health Sciences, University of Illinois at Urbana-Champaign, USA College of Pharmacy, California Northstate University, USA
a r t i c l e
i n f o
Article history: Received 21 September 2015 Received in revised form 24 November 2015 Accepted 25 November 2015 Keywords: Antidepressant Depression Functional limitation Disability Older adult Health and retirement study
a b s t r a c t Objective: The upsurge in prevalence and long-term use of antidepressants among older adults might have profound health implications beyond depressive symptom management. This study examined the relationship between antidepressant use and functional limitation onset in U.S. older adults. Methods: Study sample came from 2006 and 2008 waves of the Health and Retirement Study, in combination with data from 2005 and 2007 Prescription Drug Study. Self-reported antidepressant use was identified based on the therapeutic classification of Cerner Multum's Lexicon. Functional limitations were classified into those pertaining to physical mobility, large muscle function, activities of daily living, gross motor function, fine motor function, and instrumental activities of daily living. Cox proportional hazard models were performed to assess the effects of antidepressant use on future functional limitation onset by limitation category, antidepressant type, and length of use, adjusted by depression status and other individual characteristics. Results: Antidepressant use for one year and longer was associated with an increase in the risk of functional limitation by 8% (95% confidence interval = 4%–12%), whereas the relationship between antidepressant use less than a year and function limitation was statistically nonsignificant. Antidepressant use was associated with an increase in the risk of functional limitation by 8% (3%–13%) among currently nondepressed participants but not currently depressed participants. Conclusion: Long-term antidepressant use in older adults should be prudently evaluated and regularly monitored to reduce the risk of functional limitation. Future research is warranted to examine the health consequences of extended and/or off-label antidepressant use in absence of depressive symptoms. © 2015 Elsevier Inc. All rights reserved.
Introduction Antidepressants are one of the most commonly prescribed medications in the U.S. [1]. The prevalence of antidepressant use among Americans 12 years of age and above nearly quadrupled from 3% in 1988–1994 to 11% in 2005–2008 [2,3]. Older adults are substantially more likely to use antidepressants compared to younger adults. In 2005–2008, 14.5% of adults 60 years of age and above took antidepressants, compared to 6.1% among adults 18–39 years of age [1]. In addition, a large majority of older Americans taking antidepressants have taken the medication for one or more years [1]. The upsurge in prevalence and long-term use of antidepressants among older adults might profoundly impact many aspects of their physical and mental health beyond depressive symptom management. The side effects of antidepressant use have been extensively documented, which include nausea, sexual dysfunction, fatigue, drowsiness, insomnia, dry mouth, blurred vision, constipation, agitation, anxiety, increased appetite, and weight gain [4]. Older adults could be particularly vulnerable to these medication side effects due to compromised ⁎ Corresponding author at: 1206 South 4th Street, Champaign, 61820, IL, USA. E-mail address: [email protected] (R. An).
http://dx.doi.org/10.1016/j.jpsychores.2015.11.007 0022-3999/© 2015 Elsevier Inc. All rights reserved.
cognitive and physical functioning. Long-term use of antidepressants that act on the serotonin system has been linked to reduced bone mineral density and osteoporosis [5,6], which is found to be associated with an elevated risk of fall and hip fractures in older antidepressant users [7,8]. The health implications of long-term antidepressant use among older adults can be substantial. Many of the medication side effects could potentially interfere with older adults' activities of daily living, increase the risk of functional limitations and disability, and reduce their health-related quality of life. However, to our knowledge, no study has prospectively assessed the impact of antidepressant use on functional limitations of various types in older adults. This study fills in the gap by examining the relationship between antidepressant use and functional limitations using data from a nationally representative sample of older adults. Methods Study setting Individual-level data came from the Health and Retirement Study (HRS), an ongoing longitudinal study that surveys a representative
32
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
sample of U.S. community-dwelling adults 50 years of age and above since 1992. Follow-up interviews are conducted every other year, with an overall response rate over 80% across waves. HRS collects rich information including income, employment, assets, pension plans, health insurance, disability, physical health and functioning, cognitive functioning, and health care expenditures. Survey design, questionnaires, and other details can be found on the HRS web portal (http:// hrsonline.isr.umich.edu/). This study used data from HRS 2006 and 2008 waves that were cleaned and compiled by the RAND Corporation (RAND HRS enhanced fat files and longitudinal dataset version N), merged with data from HRS 2005 and 2007 Prescription Drug Study (PDS). This data arrangement allowed us to examine the effect of previous antidepressant use on current functional limitation onset (i.e., antidepressant use in 2005 and 2007 in relation to functional limitation incidents in 2006 and 2008). HRS was approved by the University of Michigan Human Subjects Review Committee. The present study involved secondary-data analysis of de-identified, publicly available data, and was deemed exempt from human subjects review by the Institutional Review Board of the University of Illinois at Urbana-Champaign. Antidepressant use HRS 2005 and 2007 PDS were a two-wave mail survey designed to track changes in prescription medication utilization among U.S. older adults. At the PDS baseline in 2005, a sample of 5654 individuals was drawn from HRS 2004 wave participants, including those born in 1942 or earlier or already covered by Medicare or Medicaid at some time between 2002 and 2004. People lacking prescription medication coverage or having low income and wealth were oversampled. PDS respondents were asked to list all medications prescribed, report length of use for each listed medication, and document the medication name from the label on the prescription bottle. Medications were then matched to the three-level nested therapeutic classification scheme of Cerner Multum's Lexicon [9]. Based on the definition for antidepressant use of the Centers for Disease Control and Prevention, respondents were considered to be taking antidepressant if any of their medications were matched to the second level of drug categorical codes, specifically code 249 [1]. We further classified antidepressants into six major types: tricyclic antidepressants (TCAs), serotoninnorepinephrine reuptake inhibitors (SNRIs), selective serotonin reuptake inhibitors (SSRIs), atypical antidepressants (Mirtazapine [Axit, Mirtaz, Mirtazon, Remeron, Remeron SolTab, Zispin], Bupropion [Aplenzin, Budeprion XL, Buproban, Wellbutrin, Zyban], Trazodone [Desyrel, Oleptro], Nefazodone [Serzone], Vilazodone [Viibryd], and Vortioxetine [Brintellix]), monoamine oxidase inhibitors (MAOIs), and pharmacotherapeutic combination antidepressants (Amitriptyline/Perphenazine [Triavil], Chlordiazepoxide/Amitriptyline [Limbitrol], Perphenazine/Amitriptyline [Etrafon], and Olanzapine/Fluoxetine [Symbyax]). Functional limitations Functional limitations in HRS were classified into six categories based on validated indices. These indices were adopted for their comparability with other studies that measured functional limitations, their validity and reliability, and consistency across survey waves [10,11]. The six categories of functional limitations include: physical mobility limitation, large muscle function limitation, activities of daily living limitation, gross motor function limitation, fine motor function limitation, and instrumental activities of daily living limitation. Each question asked whether a participant had any difficulty (coded as “yes” or “no”) in performing a specific activity. Physical mobility consists of five activities: walking one block, walking several blocks, walking across a room, climbing one flight of stairs without resting, and climbing several flights of stairs without resting. Large muscle function consists of four
activities: sitting for about two hours, getting up from a chair after sitting for long periods, stooping or kneeling or crouching, and pulling or pushing large objects like a living room chair. Activities of daily living limitation consist of five activities: bathing or showering, eating, dressing, walking across a room, and getting in or out of bed. Gross motor function consists of four activities: walking one block, walking across a room, climbing one flight of stairs without resting, and bathing. Fine motor function consists of three activities: eating, dressing, and picking up a dime from a table. Instrumental activities of daily living consist of three activities: using a telephone, taking medication, and handling money. Functional limitation of a specific category is defined as having difficulty (i.e., an answer of “yes”) in performing at least one of the activities included in that category. Any functional limitation is defined as having one or more of the functional limitations in these six categories. Depression Depressive symptoms were measured by the eight-item Center for Epidemiologic Studies Depression Scale (CES-D), a shortened version of the 20-item CES-D [12]. Participants were asked whether (“yes” or “no”) they felt depressed, felt that everything was an effort, slept restlessly, could not get going, felt sad, felt lonely, enjoyed life, and were happy in the past week. The two positive items (i.e., “enjoyed life” and “was happy”) are reverse-coded, so that a higher score indicates a more depressed mood. The eight-item CES-D total score, ranging from zero to eight, sums up the presence of (coded as one) or absence (coded as zero) from each of the eight feelings. Melchior et al. (1993) reported that the eight-item and 20-item CES-D scales were highly correlated (r = 0.93) and had comparable discriminant validity [13]. A cutoff score of three has been suggested by previous validation studies to indicate clinically relevant depressive symptoms [14]. This cut-off score has a sensitivity of 0.71 and a specificity of 0.79 to predict major depressive episodes [14]. A participant was classified as having a depression onset in a survey wave if one scored three or above on the eight-item CES-D in that wave. Other individual characteristics We controlled both wave-invariant and wave-variant individual characteristics in the regression analyses. Wave-invariant covariates include gender, race/ethnicity (non-Hispanic white, non-Hispanic African American, non-Hispanic other race or multi-race, and Hispanic), and education (education less than high school, high school, college, and education higher than college). Wave-variant covariates include age in years, marital status (married or living with partner, and unmarried, divorced, separated, or widowed), household net wealth (divided into four quartiles based on the wealth distribution in each survey wave), current smoking status, heavy drinking status (defined as one or more drink per day on average or four or more drinks on any occasion in the past three months for women, and two or more drinks per day on average or four or more drinks on any occasion in the past three months for men), depression status (CES-D score of three and above), diagnosis of a chronic condition (hypertension, diabetes, heart disease, stroke, lung disease, arthritis, cancer, and memory related disease), residential census region (Midwest, Northeast, South, and West), and body weight status. Body mass index (BMI) was calculated from self-reported height and weight. Body weight status was classified into four categories based on the international classification of adult BMI values – underweight (BMI b 18.5 kg/m2), normal weight (18.5 kg/m2 ≤ BMI b 25 kg/m2), overweight (25 kg/m2 ≤ BMI b 30 kg/m2), and obesity (BMI ≥ 30 kg/m2) [15]. Study sample PDS 2005 wave data were merged to HRS 2006 wave data, and PDS 2007 wave data were merged to HRS 2008 wave data, both using personal identifiers that uniquely identified each study participant.
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
Participants with missing values on functional limitations and/or other individual characteristics were excluded from statistical analyses. We also excluded those who reported any functional limitation in 2004 wave, resulting in a final sample size of 4242. Among these individuals who were free from any functional limitation in 2004 wave, 3530, 2794, 3230, 1100, 1669, 1039, and 984 developed any functional limitation, physical mobility limitation, large muscle function limitation, activities of daily living limitation, gross motor function limitation, fine motor function limitation, and instrumental activities of daily living limitation by 2008 wave. Statistical analyses Cox proportional hazards regressions were performed to examine the relationship between antidepressant use reported in 2005 and 2007 waves and functional limitation reported in 2006 and 2008 waves, adjusted by both wave-invariant and wave-variant individual characteristics. The key independent variable was a dichotomous variable for any antidepressant use. A participant was considered a “survivor” till a functional limitation was reported, if ever. Participants who died or were lost during follow-up without ever reporting a functional limitation were censored at the last wave when they were interviewed. Participants that were alive and remained free from a functional limitation during the study period were censored at the last (2008) wave of the study. We modeled any functional limitation and each of the six specific functional limitation categories in separate regressions. We further divided antidepressant users into those who took the medication for at least a year and those less than a year, and estimated the relationship between antidepressant use and functional limitation among these two groups in separate regressions. We also performed separate regressions on those depressed versus nondepressed at baseline. To assess the differential effects on functional limitation across antidepressant type, we included four dichotomous variables for TCA, SNRI, SSRI, and atypical antidepressant use as independent variables in the regression (MAOIs and pharmacotherapeutic combination antidepressants were excluded because in total only four study participants took them). Both descriptive statistics and regressions were weighted by the HRS baseline sampling weights. In regression analyses, Eicker–Huber–White sandwich estimator was used to calculate standard errors clustered at individual level to account for potential within-individual serial correlations that might downward bias the standard errors of estimates. All statistical analyses were conducted using Stata 14.1 SE version (StataCorp, College Station, TX). Results Table 1 reports baseline sample characteristics by antidepressant use status. Among those who were free from any functional limitation in 2004 wave, 83% developed any functional limitation, 66% developed physical mobility limitation, 76% developed large muscle function limitation, 26% developed activities of daily living limitation, 39% developed gross motor function limitation, 25% developed fine motor function limitation, and 23% developed instrumental activities of daily living limitation by 2008 wave. Less than one in six (16%) older adults used any antidepressant in 2005 wave. A majority of the users (75%) took antidepressant for one year or longer. Among users, the prevalence of TCA, SNRI, SSRI, and atypical antidepressant use was 13%, 11%, 26%, and 60%, respectively. Antidepressant users had substantially higher incidence of function limitation compared to nonusers. The incidence of any functional limitation, physical mobility limitation, large muscle function limitation, activities of daily living limitation, gross motor function limitation, fine motor function limitation, and instrumental activities of daily living limitation among antidepressant users were 91%, 78%, 85%, 34%, 54%, 31%, and 29%, respectively, compared to 82%, 64%, 75%, 24%, 37%, 23%, and 22% among nonusers.
33
Study participants averaged 73 years of age in 2006, and nearly four fifths (79%) were non-Hispanic whites. Only slightly less than a third (32%) fell within the normal weight range. Underweight was rare (2%), whereas the majority (66%) were overweight or obese (BMI ≥ 25). About one in eleven (9%) study participants were heavy drinkers, and less than one in eight (12%) current smokers. Chronic conditions were common: prevalence of depression 25%, hypertension 64%, diabetes 23%, heart disease 31%, stroke 10%, lung disease 12%, arthritis 69%, cancer 18%, and memory related disease 2%. Table 2 reports the adjusted hazard ratios (AHRs) for any functional limitation estimated in Cox proportional hazard regression. Prior-year antidepressant use prospectively predicted functional limitation onset. Compared with nonusers, antidepressant users were 7% (AHR = 1.07, 95% confidence interval [CI] = 1.03, 1.10) more likely to report a functional limitation during the study period. Women were significantly more likely to report functional limitation than men. The risk for functional limitation closely coincided with aging. Non-Hispanic African Americans appeared less likely to report functional limitation compared with their non-Hispanic white counterparts. Smoking but not heavy drinking was found to be associated with an elevated risk of functional limitation. There appeared to be a monotonic relationship between education level or household net wealth and functional limitation — lower education attainment and household net wealth were consistently associated with higher risk of functional limitation. Overweight and obesity were positively associated with functional limitation. Most chronic conditions, except hypertension and cancer, were founded to be linked with functional limitation. Table 3 reports the estimated effects of antidepressant use on alternative types of functional limitation. Antidepressant use was associated with an increase in the risk of physical mobility limitation by 8% (AHR = 1.08, 95% CI = 1.02, 1.14), large muscle function limitation by 6% (AHR = 1.06, 95% CI = 1.01, 1.11), activities of daily living limitation by 15% (AHR = 1.15, 95% CI = 1.01, 1.32), and gross motor function limitation by 19% (AHR = 1.19, 95% CI = 1.09, 1.31). The effect of antidepressant use on fine motor function limitation (AHR = 1.08, 95% CI = 0.94, 1.25) and instrumental activities of daily living limitation (AHR = 1.06, 95% CI = 0.92, 1.23) were both positive but statistically nonsignificant at P b 0.05. The effect of antidepressant use on functional limitation was centered on long-term users and currently nondepressed older adults. Antidepressant use for one year or longer was associated with an increase in the risk of functional limitation by 8% (AHR = 1.08, 95% CI = 1.04, 1.12), whereas the estimated association between antidepressant use less than a year and function limitation (AHR = 1.03, 95% CI = 0.97, 1.09) was statistically nonsignificant. Antidepressant use significantly predicted functional limitation onset among those currently nondepressed (AHR = 1.08, 95% CI = 1.03, 1.13) but not among currently depressed study participants (AHR = 1.01, 95% CI = 0.99, 1.05). The effect of antidepressant use on functional limitation was not found to differ by medication type (i.e., TCA, SNRI, SSRI, and atypical antidepressants). Discussion Antidepressants have been documented to be effective in treating major depressive symptoms in primary and secondary care settings [16,17]. However, findings from recent systematic reviews are mixed regarding the efficacy of antidepressants in the treatment of depression among older adults. Wilson et al (2001) conducted a meta-analysis on 17 placebo-controlled randomized trials that evaluated the effect of antidepressant treatment on geriatric depression [18]. TCAs, SSRIs and MAOIs were found to be effective in both institutionalized and community-dwelling patients, but lower dose TCA treatment was suggested. In contrast, Taylor and Doraiswamy (2004) performed a systematic review and meta-analysis on 18 antidepressant placebocontrolled randomized trials on patients 55 years of age and above,
34
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
Table 1 Descriptive statistics of study sample. Variable
Entire sample
Any antidepressant use in 2005
No antidepressant use in 2005
Sample size
4242
3584
658
83.2 65.9 76.1 25.9 39.3 24.5 23.2
91.0 78.1 85.0 34.3 53.8 31.3 29.5
81.8 63.6 74.5 24.4 36.7 23.2 22.0
16.0 (14.7, 17.3) 25.2 (21.8, 28.6) 74.8 (71.4, 78.2) 13.2 (10.5, 15.9) 11.4 (8.9, 13.9) 60.2 (56.3, 64.0) 25.5 (22.1, 29.0)
/ / / / / / /
/ / / / / / /
42.2 (40.4, 44.0) 57.8 (56.0, 59.6)
43.5 (41.6, 45.5) 56.5 (54.5, 58.4)
35.1 (30.7, 39.5) 64.9 (60.5, 69.3)
72.5 (72.1, 72.8)
72.4 (72.1, 72.8)
72.6 (71.6, 73.5)
79.3 (77.9, 80.6) 10.2 (9.2, 11.1) 2.6 (1.9, 3.2) 8.0 (7.1, 8.9)
78.1 (76.5, 79.6) 10.7 (9.7, 11.8) 2.8 (2.0, 3.5) 8.4 (7.4, 9.4)
85.6 (82.6, 88.6) 7.2 (5.1, 9.4) 1.4 (0.4, 2.5) 5.8 (3.8, 7.7)
25.9 (24.3, 27.4) 53.4 (51.6, 55.2) 13.6 (12.4, 14.8) 7.1 (6.2, 8.1)
26.7 (25.0, 28.4) 52.9 (50.9, 54.8) 13.4 (12.0, 14.7) 7.1 (6.1, 8.1)
21.6 (18.0, 25.2) 56.2 (51.7, 60.7) 14.9 (11.7, 18.2) 7.3 (4.9, 9.7)
54.7 (52.9, 56.5) 45.3 (43.5, 47.1)
55.1 (53.1, 57.0) 44.9 (43.0, 46.9)
52.7 (48.2, 57.2) 47.3 (42.8, 51.8)
27.4 (25.8, 29.0) 24.7 (23.2, 26.2) 22.9 (21.4, 24.3) 25.0 (23.5, 26.6)
27.6 (25.8, 29.4) 24.3 (22.7, 26.0) 22.7 (21.2, 24.3) 25.3 (23.6, 27.0)
26.2 (22.1, 30.3) 26.5 (22.5, 30.4) 23.6 (19.8, 27.4) 23.7 (19.9, 27.6)
9.4 (8.3, 10.5) 11.5 (10.3, 12.7)
9.7 (8.6, 10.9) 11.4 (10.1, 12.7)
7.6 (5.3, 10.0) 12.3 (9.1, 15.4)
1.9 (1.3, 2.4) 31.9 (30.2, 33.5) 37.6 (35.9, 39.3) 28.6 (27.0, 30.3)
1.7 (1.2, 2.2) 32.9 (31.1, 34.7) 37.8 (35.9, 39.7) 27.6 (25.8, 29.3)
2.6 (0.6, 4.6) 26.5 (22.5, 30.4) 36.6 (32.3, 41.0) 34.3 (30.0, 38.6)
24.9 (23.3, 26.5) 63.7 (61.9, 65.4) 23.2 (21.7, 24.7) 31.1 (29.5, 32.8) 9.5 (8.5, 10.6) 12.2 (11.0, 13.4) 69.0 (67.3, 70.6) 17.6 (16.2, 18.9) 1.8 (1.3, 2.4)
23.2 (21.5, 24.8) 61.3 (59.4, 63.2) 21.8 (20.2, 23.4) 28.5 (26.8, 30.3) 9.0 (7.9, 10.1) 11.0 (9.8, 12.2) 67.5 (65.7, 69.4) 17.7 (16.2, 19.2) 1.7 (1.2, 2.3)
34.2 (29.8, 38.5) 76.0 (72.0, 80.0) 30.4 (26.1, 34.7) 44.8 (40.3, 49.3) 12.5 (9.5, 15.5) 18.5 (14.8, 22.3) 76.5 (72.4, 80.5) 16.9 (13.5, 20.2) 2.4 (1.1, 3.8)
17.0 (15.6, 18.4) 24.5 (23.0, 26.0) 38.4 (36.7, 40.1) 20.1 (18.6, 21.5)
16.4 (14.9, 17.8) 24.7 (23.0, 26.3) 38.8 (36.9, 40.7) 20.1 (18.6, 21.7)
20.3 (16.6, 23.9) 23.8 (19.8, 27.8) 36.1 (31.9, 40.3) 19.9 (16.1, 23.6)
Functional limitation incidence during 2004–2008 (%) Any functional limitation Physical mobility limitation Large muscle function limitation Activities of daily living limitation Gross motor function limitation Fine motor function limitation Instrumental activities of daily living limitation Antidepressant use in 2005 (%) Any antidepressant use Any antidepressant use b1 year among users Any antidepressant use ≥1 year among users Tricyclic antidepressant (TCA) use among users Serotonin-norepinephrine reuptake inhibitor (SNRI) use among users Atypical antidepressant use among users Selective serotonin reuptake inhibitor (SSRI) use among users Gender (%) Male Female Age in 2006 (mean) Age in years Race/ethnicity (%) Non-Hispanic white Non-Hispanic African American Non-Hispanic other race/multi-race Hispanic Education in 2006 (%) Less than high school High school College Higher than college Marital status in 2006 (%) Married or living with partner Unmarried/divorced/separated/widowed Household wealth in 2006 (%) Lowest income quartile Mid-low income quartile Mid-high income quartile Highest-high income quartile Risk behavior in 2006 (%) Heavy drinking Currently smoking Body weight status in 2006 (%) Underweight (BMI b 18.5 kg/m2) Normal weight (18.5 kg/m2 ≤ BMI b 25 kg/m2) Overweight (25 kg/m2 ≤ BMI b 30 kg/m2) Obese (BMI ≥ 30 kg/m2) Chronic condition in 2006 (%) Depression Hypertension Diabetes Heart disease Stroke Lung disease Arthritis Cancer Memory related disease Residential census region in 2006 (%) Northeast Midwest South West
Notes: Study sample consisted of Health and Retirement Study (HRS) participants that were free from any functional limitation in 2004 wave. Statistics were weighted by HRS 2006 wave sampling weights (except for sample size and functional limitation incidence during 2004–2008, which were unweighted). 95% confidence intervals are in parentheses.
but no concrete conclusions were drawn due to various study limitations such as small sizes in individual trials, exclusion of common comorbid conditions, lack of controlled head-to-head comparisons, and other methodological design concerns [19]. Nelson and Devanand
(2011) conducted a systematic review and meta-analysis of antidepressant placebo-controlled randomized trials in older adults with depression and dementia [20]. The efficacy of antidepressant treatment was unconfirmed for that patient population. Parikh (2000) summarized the
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36 Table 2 Adjusted hazard ratios for any functional limitation estimated in Cox proportional hazards regression. Independent variable Antidepressant use Any antidepressant use No antidepressant use Gender Male Female Age Age in years Race/ethnicity Non-Hispanic white Non-Hispanic African American Non-Hispanic other race/multi-race Hispanic Education Less than high school High school College Higher than college Current marital status Married or living with partner Unmarried/divorced/separated/widowed Household wealth Lowest income quartile Mid-low income quartile Mid-high income quartile Highest-high income quartile Risk behavior Heavy drinking Currently smoking Body weight status Underweight (BMI b 18.5 kg/m2) Normal weight (18.5 kg/m2 ≤ BMI b 25 kg/m2) Overweight (25 kg/m2 ≤ BMI b 30 kg/m2) Obese (BMI ≥ 30 kg/m2) Chronic condition Depression Hypertension Diabetes Heart disease Stroke Lung disease Arthritis Cancer Memory related disease Residential census region Northeast Midwest South West
Adjusted hazard ratio 1.07⁎⁎⁎ (1.03, 1.10) Reference 0.92⁎⁎⁎ (0.89, 0.96) Reference 1.01⁎⁎⁎ (1.00, 1.01) Reference 0.90⁎⁎⁎ (0.86, 0.95) 1.07 (0.95, 1.21) 0.99 (0.94, 1.04) Reference 0.98 (0.95, 1.02) 0.92⁎ (0.86, 0.98) 0.84⁎⁎ (0.76, 0.93) 1.00 (0.96, 1.03) Reference Reference 0.98 (0.94, 1.01) 0.94⁎⁎ (0.90, 0.98) 0.85⁎⁎⁎ (0.81, 0.90) 0.97 (0.90, 1.05) 1.06⁎ (1.01, 1.12) 1.11⁎ (1.01, 1.22) Reference 1.06⁎⁎ (1.01, 1.10) 1.15⁎⁎⁎ (1.10, 1.20) 1.14⁎⁎⁎ (1.11, 1.17) 1.03 (0.99, 1.07) 1.04⁎ (1.01, 1.08) 1.07⁎⁎⁎ (1.04, 1.11) 1.06⁎⁎ (1.02, 1.10) 1.09⁎⁎⁎ (1.06, 1.12) 1.39⁎⁎⁎ (1.32, 1.47) 0.98 (0.94, 1.02) 1.08⁎⁎ (1.02, 1.14) Reference 1.05 (1.00, 1.10) 1.06⁎ (1.01, 1.11) 1.01 (0.95, 1.07)
Notes: Study sample (N = 4242) came from the Health and Retirement Study (HRS) 2006–2008 waves and consisted of those who were free from any functional limitation in 2004 wave. Cox proportional hazards regressions were performed to estimate the adjusted hazard ratios for any functional limitation, weighted by the HRS 2006 wave sampling weights. 95% confidence intervals are in parentheses. ⁎ P b 0.05. ⁎⁎ P b 0.01. ⁎⁎⁎ P b 0.001.
challenges in study design and result interpretation for evaluating the efficacy of antidepressant treatment in older adults [21]. Most Phase II and III clinical trials on antidepressants had been on younger adults and excluded very old and frail patients. Design limitations of the few trials conducted in older patients precluded accurate interpretation of differences in efficacy or safety between drugs. In sum, findings from previous studies on the effectiveness of antidepressant use in treating depression among older adults remain mixed, and none of the studies focused on the potential side effect of antidepressant use on functional limitations that are highly prevalent in older adults and significantly influence their daily activities and quality of life. This study examined the associations between antidepressant use and functional limitations among U.S. older adults using data from a
35
Table 3 Estimated associations between antidepressant use and functional limitation by limitation type. Type of functional limitation
Adjusted hazard ratio
Any functional limitation Physical mobility limitation Large muscle function limitation Activities of daily living limitation Gross motor function limitation Fine motor function limitation Instrumental activities of daily living limitation
1.07⁎⁎⁎ (1.03, 1.10) 1.08⁎⁎ (1.02, 1.14) 1.06⁎⁎ (1.01, 1.11) 1.15⁎ (1.01, 1.32) 1.19⁎⁎⁎ (1.09, 1.31) 1.08 (0.94, 1.25) 1.06 (0.92, 1.23)
Notes: Study sample came from the Health and Retirement Study (HRS) 2006–2008 waves, consisting of 4,242 participants who did not have any functional limitation in 2004 wave. Cox proportional hazards regressions were performed to estimate the adjusted hazard ratios for functional limitation, weighted by the HRS 2006 wave sampling weights and adjusted by individual characteristics specified in Table 1. 95% confidence intervals are in parentheses. ⁎ P b 0.05. ⁎⁎ P b 0.01. ⁎⁎⁎ P b 0.001.
nationally representative longitudinal survey. Antidepressant use prospectively predicted future functional limitation onset. Antidepressant use for one year or longer was associated with an increase in the risk of functional limitation by 8%, whereas the relationship between antidepressant use less than a year and function limitation was statistically nonsignificant. Antidepressant use was associated with an increase in the risk of functional limitation by 8% among currently nondepressed participants but not currently nondepressed participants. The relationship between antidepressant use and functional limitations differed across limitation category. The impact of antidepressant use on functional limitations was not found to differ by medication type. The estimated association between antidepressant use and functional limitations was only statistically significant in long-term users. The declining physical, mental, and cognitive conditions could make older adults disproportionately susceptible to the side effects of antidepressants. Over time, these side effects, if not properly managed, would create barriers for older adults to accomplish some of their daily routines. Moreover, the impact of antidepressant use on functional limitations was only identified among currently nondepressed older adults. It is not known whether their depressive symptoms were successfully managed due to antidepressant use, they took antidepressants mainly for preventing a relapse of depression, or they took the medication for other reasons, such as treating insomnia or neuropathic pain [22,23]. We also could not rule out the possibility that the null finding on the relationship between antidepressant use and functional limitation among nondepressed older adults and short-term antidepressant users was a consequence of lacking statistical power owing to small subsample size. Nevertheless, the study results indicate the needs for carefully weighing the benefits and costs, particularly in terms of the elevated risk for functional limitations, for long-term antidepressant use in older adults. The finding also warrants future research on the health consequences of extended antidepressant use in absence of depressive symptoms as well as off-label antidepressant use in treating nondepression conditions among older adults. The estimated impact of antidepressant use on fine motor function and instrumental activities of daily living was statistical nonsignificant. Although we were unable to eliminate the possibility of an underpowered estimation, the null finding might partially be explained by the differences between these two types of functional limitations and the other limitation categories. Whereas the measures for physical mobility limitation, large muscle function limitation, activities of daily living limitation, and gross motor function limitation focused on physicallydemanding labor-intensive tasks (e.g., walking, pulling or pushing, climbing), measures for fine motor function and instrumental activities of daily living tested less physically-demanding but more technical skills (e.g., dime picking, money handling). Antidepressant use could have less of an impact on those technical skills (the positive effect
36
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
through depressive symptom management and the negative impact due to medication side effect canceled out) than on those relatively physically-demanding tasks. A few limitations of the study should be noted. This study was conducted based on data from a prospective cohort study. Despites the effort of controlling for an extensive list of individual characteristics, we were unable to completely eliminate the risk of confounding bias resulted from unobserved differences in individual characteristics due to the lack of a randomized study design. Therefore, the results must be interpreted as correlations rather than causations. The study findings might not be generalizable to the entire aging population as the study sample of community-dwelling older adults were generally less fragile relative to their institutionalized counterparts. As commonly seen in other panel studies on aging population [24], this study might suffer from attrition problem. Banks et al. (2011) found household net wealth to be negatively correlated with attrition rate in HRS [25]. As richer older adults were less likely to develop functional limitations and more likely to retain in the study sample, attrition attributable to wealth could result in underestimation of the effect of antidepressant use on functional limitations. Although the HRS sample size was reasonably large, further stratification by antidepressant type substantially reduced the statistical power and estimation precision, which might explain why no significant differences in the effects of TCA, SNRI, SSRI, and atypical antidepressants on functional limitations were identified. It would potentially be more ideal if we could also exclude those who had any functional limitation in 2006 from the analyses, because data on antidepressant use was only available from 2005 onward. However, the main challenge was that we would be left with only a small fraction of the sample due to the very high incidence of any functional limitation (given the average age of study participants of 73 in 2006), and even smaller samples after stratification by antidepressant use status and length of use, which would substantially compromise the statistical power of our estimation. Chronic condition diagnoses were limited to a few diseases, whereas certain illnesses that are likely to be associated with functional limitations in older adults such as Parkinson's disease, Alzheimer, dementia, osteoporosis, and chronic pain, were not administered in HRS. One question regarding memory related disease was administered in HRS 2006 and 2008 waves and was used in regression analyses, but more detailed classification of that disease category (Alzheimer, dementia, etc.) became only available in the 2010 wave and after. The efficacy of antidepressant use to treat minor depressions has been questioned. Barbui et al. (2011) conducted a systematic review and meta-analysis on placebo-controlled randomized trials that evaluated the effect of antidepressant treatment on minor depression [26]. No clinically meaningful advantage for antidepressants over placebo was found in treating individuals with minor depressive symptoms. Wiese [2011] suggested the selection of an antidepressant medication in treating geriatric depression be based on the best side effect profile and the lowest risk of drug-drug interaction [27]. Considering the high prevalence of comorbidity and vulnerability in older adults, antidepressant use might not be recommended for treating minor depressive symptoms, and long-term antidepressant use should be prudently evaluated and regularly monitored in order to prevent adverse health consequences, including the risk of functional limitations. In conclusion, this study examined the impact of antidepressant use on functional limitations among U.S. older adults using data from a nationally representative longitudinal survey. Antidepressant use prospectively predicted future functional limitation onset. The associations between antidepressant use and functional limitations differed across limitation category, and were only identified among long-term antidepressant users and currently nondepressed older adults. Long-term antidepressant use in older adults should be prudently evaluated and regularly monitored in order to prevent adverse health consequences, including the risk of functional limitations. Moreover, future research is warranted to examine the health consequences of extended
antidepressant use in absence of depressive symptoms as well as offlabel antidepressant use in treating non-depression conditions in older adults. Conflict of interest We have no conflict of interests relating to this manuscript to declare. References [1] L.A. Pratt, D.J. Brody, Q. Gu, Antidepressant use in persons aged 12 and over: United States, NCHS Data Brief 2011 (76) (2005-2008) 1–8. [2] R. Paulose-Ram, M.A. Safran, B.S. Jonas, Q. Gu, D. Orwig, Trends in psychotropic medication use among U.S. adults, Pharmacoepidemiol. Drug Saf. 16 (5) (2007) 560–570. [3] M. Olfson, S.C. Marcus, National patterns in antidepressant medication treatment, Arch. Gen. Psychiatry 66 (8) (2009) 848–856. [4] Mayo Clinic, Antidepressants: Get Tips to Cope with side Effects, 2013 (Available from: http://www.mayoclinic.org/diseases-conditions/depression/in-depth/antidepressants/art-20049305). [5] R. Rizzoli, C. Cooper, J.Y. Reginster, B. Abrahamsen, J.D. Adachi, M.L. Brandi, O. Bruyère, J. Compston, P. Ducy, S. Ferrari, N.C. Harvey, J.A. Kanis, G. Karsenty, A. Laslop, V. Rabenda, P. Vestergaard, Antidepressant medications and osteoporosis, Bone 51 (3) (2012) 606–613. [6] Q. Wu, J. Liu, J.F. Gallegos-Orozco, J.G. Hentz, Depression, fracture risk, and bone loss: a meta-analysis of cohort studies, Osteoporos. Int. 21 (10) (2010) 1627–1635. [7] D. Prieto-Alhambra, H. Petri, J.S. Goldenberg, T.P. Khong, O.H. Klungel, N.J. Robinson, F. de Vries, Excess risk of hip fractures attributable to the use of antidepressants in five European countries and the USA, Osteoporos. Int. 25 (3) (2014) 847–855. [8] M.S. Bakken, A. Engeland, L.B. Engesæter, A.H. Ranhoff, S. Hunskaar, S. Ruths, Increased risk of hip fracture among older people using antidepressant drugs: data from the Norwegian Prescription Database and the Norwegian Hip Fracture Registry, Age Ageing 42 (4) (2013) 514–520. [9] Multum Lexicon database, National Health and Nutrition Examination Survey—1988-2008 data documentation, codebook, and frequencies, 2010 (Available from: http://www.cdc.gov/nchs/nhanes/nhanes1999-2000/RXQ_DRUG.htm). [10] S. Chien, N. Campbell, O. Hayden, et al., RAND HRS Data Documentation, Version N, 2014 (Available from: http://hrsonline.isr.umich.edu/modules/meta/rand/ randhrsn/randhrsN.pdf). [11] R.B. Wallace, A.R. Herzog, Overview of the health measures in the Health and Retirement Study, J. Hum. Resour. 30 (1995) S84–S107. [12] F.J. Kohout, L.F. Berkman, D.A. Evans, J. Cornoni-Huntley, Two shorter forms of the CES-D (Center for Epidemiological Studies Depression) depression symptoms index, J. Aging Health 5 (2) (1993) 179–193. [13] L.A. Melchior, G.J. Huba, V.B. Brown, C.J. Reback, A short depression index for women, Educ. Psychol. Meas. 53 (4) (1993) 1117–1125. [14] C.L. Turvey, R.B. Wallace, R. Herzog, A revised CES-D measure of depressive symptoms and a DSM-based measure of major depressive episodes in the elderly, Int. Psychogeriatr. 11 (2) (1999) 139–148. [15] World Health Organization, The international classification of adult underweight, overweight and obesity according to BMI, 2006 (Available from: http://apps.who. int/bmi/index.jsp?introPage=intro_3.html). [16] B. Arroll, C.R. Elley, T. Fishman, F.A. Goodyear-Smith, T. Kenealy, G. Blashki, N. Kerse, S. Macgillivray, Antidepressants versus placebo for depression in primary care, Cochrane Database Syst. Rev. 3 (2009), CD007954. [17] F.L. Rocha, C. Fuzikawa, R. Riera, C. Hara, Combination of antidepressants in the treatment of major depressive disorder: a systematic review and meta-analysis, J. Clin. Psychopharmacol. 32 (2) (2012) 278–281. [18] K. Wilson, P.G. Mottram, A. Sivananthan, A. Nightingale, Antidepressants versus placebo for the depressed elderly, Cochrane Database Syst. Rev. 2 (2001), CD000561. [19] W.D. Taylor, P.M. Doraiswamy, A systematic review of antidepressant placebocontrolled trials for geriatric depression: limitations of current data and directions for the future, Neuropsychopharmacology 29 (12) (2004) 2285–2299. [20] J.C. Nelson, D.P. Devanand, A systematic review and meta-analysis of placebocontrolled antidepressant studies in people with depression and dementia, J. Am. Geriatr. Soc. 59 (4) (2011) 577–585. [21] C. Parikh, Antidepressants in the elderly: challenges for study design and their interpretation, Br. J. Clin. Pharmacol. 49 (6) (2000) 539–547. [22] W.A. Santos Moraes, P.R. Burke, P.L. Coutinho, C. Guilleminault, A.G. Bittencourt, S. Tufik, D. Poyares, Sedative antidepressants and insomnia, Rev. Bras. Psiquiatr. 33 (1) (2011) 91–95. [23] T. Saarto, P.J. Wiffen, Antidepressants for neuropathic pain, Cochrane Database Syst. Rev. 4 (2007), CD005454. [24] R. Groves, M. Couper, Nonresponse in Household Surveys, John Wiley and Sons, New York, 1998. [25] J. Banks, A. Muriel, J. Smith, Attrition and health in aging studies: evidence from ELSA and HRS, Longit. Life Course Stud 2 (2) (2011) 101–126. [26] C. Barbui, A. Cipriani, V. Patel, J.L. Ayuso-Mateos, M. van Ommeren, Efficacy of antidepressants and benzodiazepines in minor depression: systematic review and meta-analysis, Br. J. Psychiatry 198 (1) (2011) 11–16. [27] B.S. Wiese, Geriatric depression: the use of antidepressants in the elderly, BCMJ 53 (7) (2011) 341–347.
Contents lists available at ScienceDirect
Journal of Psychosomatic Research
Antidepressant use and functional limitations in U.S. older adults Ruopeng An a,⁎, Lingyun Lu b a b
Department of Kinesiology and Community Health, College of Applied Health Sciences, University of Illinois at Urbana-Champaign, USA College of Pharmacy, California Northstate University, USA
a r t i c l e
i n f o
Article history: Received 21 September 2015 Received in revised form 24 November 2015 Accepted 25 November 2015 Keywords: Antidepressant Depression Functional limitation Disability Older adult Health and retirement study
a b s t r a c t Objective: The upsurge in prevalence and long-term use of antidepressants among older adults might have profound health implications beyond depressive symptom management. This study examined the relationship between antidepressant use and functional limitation onset in U.S. older adults. Methods: Study sample came from 2006 and 2008 waves of the Health and Retirement Study, in combination with data from 2005 and 2007 Prescription Drug Study. Self-reported antidepressant use was identified based on the therapeutic classification of Cerner Multum's Lexicon. Functional limitations were classified into those pertaining to physical mobility, large muscle function, activities of daily living, gross motor function, fine motor function, and instrumental activities of daily living. Cox proportional hazard models were performed to assess the effects of antidepressant use on future functional limitation onset by limitation category, antidepressant type, and length of use, adjusted by depression status and other individual characteristics. Results: Antidepressant use for one year and longer was associated with an increase in the risk of functional limitation by 8% (95% confidence interval = 4%–12%), whereas the relationship between antidepressant use less than a year and function limitation was statistically nonsignificant. Antidepressant use was associated with an increase in the risk of functional limitation by 8% (3%–13%) among currently nondepressed participants but not currently depressed participants. Conclusion: Long-term antidepressant use in older adults should be prudently evaluated and regularly monitored to reduce the risk of functional limitation. Future research is warranted to examine the health consequences of extended and/or off-label antidepressant use in absence of depressive symptoms. © 2015 Elsevier Inc. All rights reserved.
Introduction Antidepressants are one of the most commonly prescribed medications in the U.S. [1]. The prevalence of antidepressant use among Americans 12 years of age and above nearly quadrupled from 3% in 1988–1994 to 11% in 2005–2008 [2,3]. Older adults are substantially more likely to use antidepressants compared to younger adults. In 2005–2008, 14.5% of adults 60 years of age and above took antidepressants, compared to 6.1% among adults 18–39 years of age [1]. In addition, a large majority of older Americans taking antidepressants have taken the medication for one or more years [1]. The upsurge in prevalence and long-term use of antidepressants among older adults might profoundly impact many aspects of their physical and mental health beyond depressive symptom management. The side effects of antidepressant use have been extensively documented, which include nausea, sexual dysfunction, fatigue, drowsiness, insomnia, dry mouth, blurred vision, constipation, agitation, anxiety, increased appetite, and weight gain [4]. Older adults could be particularly vulnerable to these medication side effects due to compromised ⁎ Corresponding author at: 1206 South 4th Street, Champaign, 61820, IL, USA. E-mail address: [email protected] (R. An).
http://dx.doi.org/10.1016/j.jpsychores.2015.11.007 0022-3999/© 2015 Elsevier Inc. All rights reserved.
cognitive and physical functioning. Long-term use of antidepressants that act on the serotonin system has been linked to reduced bone mineral density and osteoporosis [5,6], which is found to be associated with an elevated risk of fall and hip fractures in older antidepressant users [7,8]. The health implications of long-term antidepressant use among older adults can be substantial. Many of the medication side effects could potentially interfere with older adults' activities of daily living, increase the risk of functional limitations and disability, and reduce their health-related quality of life. However, to our knowledge, no study has prospectively assessed the impact of antidepressant use on functional limitations of various types in older adults. This study fills in the gap by examining the relationship between antidepressant use and functional limitations using data from a nationally representative sample of older adults. Methods Study setting Individual-level data came from the Health and Retirement Study (HRS), an ongoing longitudinal study that surveys a representative
32
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
sample of U.S. community-dwelling adults 50 years of age and above since 1992. Follow-up interviews are conducted every other year, with an overall response rate over 80% across waves. HRS collects rich information including income, employment, assets, pension plans, health insurance, disability, physical health and functioning, cognitive functioning, and health care expenditures. Survey design, questionnaires, and other details can be found on the HRS web portal (http:// hrsonline.isr.umich.edu/). This study used data from HRS 2006 and 2008 waves that were cleaned and compiled by the RAND Corporation (RAND HRS enhanced fat files and longitudinal dataset version N), merged with data from HRS 2005 and 2007 Prescription Drug Study (PDS). This data arrangement allowed us to examine the effect of previous antidepressant use on current functional limitation onset (i.e., antidepressant use in 2005 and 2007 in relation to functional limitation incidents in 2006 and 2008). HRS was approved by the University of Michigan Human Subjects Review Committee. The present study involved secondary-data analysis of de-identified, publicly available data, and was deemed exempt from human subjects review by the Institutional Review Board of the University of Illinois at Urbana-Champaign. Antidepressant use HRS 2005 and 2007 PDS were a two-wave mail survey designed to track changes in prescription medication utilization among U.S. older adults. At the PDS baseline in 2005, a sample of 5654 individuals was drawn from HRS 2004 wave participants, including those born in 1942 or earlier or already covered by Medicare or Medicaid at some time between 2002 and 2004. People lacking prescription medication coverage or having low income and wealth were oversampled. PDS respondents were asked to list all medications prescribed, report length of use for each listed medication, and document the medication name from the label on the prescription bottle. Medications were then matched to the three-level nested therapeutic classification scheme of Cerner Multum's Lexicon [9]. Based on the definition for antidepressant use of the Centers for Disease Control and Prevention, respondents were considered to be taking antidepressant if any of their medications were matched to the second level of drug categorical codes, specifically code 249 [1]. We further classified antidepressants into six major types: tricyclic antidepressants (TCAs), serotoninnorepinephrine reuptake inhibitors (SNRIs), selective serotonin reuptake inhibitors (SSRIs), atypical antidepressants (Mirtazapine [Axit, Mirtaz, Mirtazon, Remeron, Remeron SolTab, Zispin], Bupropion [Aplenzin, Budeprion XL, Buproban, Wellbutrin, Zyban], Trazodone [Desyrel, Oleptro], Nefazodone [Serzone], Vilazodone [Viibryd], and Vortioxetine [Brintellix]), monoamine oxidase inhibitors (MAOIs), and pharmacotherapeutic combination antidepressants (Amitriptyline/Perphenazine [Triavil], Chlordiazepoxide/Amitriptyline [Limbitrol], Perphenazine/Amitriptyline [Etrafon], and Olanzapine/Fluoxetine [Symbyax]). Functional limitations Functional limitations in HRS were classified into six categories based on validated indices. These indices were adopted for their comparability with other studies that measured functional limitations, their validity and reliability, and consistency across survey waves [10,11]. The six categories of functional limitations include: physical mobility limitation, large muscle function limitation, activities of daily living limitation, gross motor function limitation, fine motor function limitation, and instrumental activities of daily living limitation. Each question asked whether a participant had any difficulty (coded as “yes” or “no”) in performing a specific activity. Physical mobility consists of five activities: walking one block, walking several blocks, walking across a room, climbing one flight of stairs without resting, and climbing several flights of stairs without resting. Large muscle function consists of four
activities: sitting for about two hours, getting up from a chair after sitting for long periods, stooping or kneeling or crouching, and pulling or pushing large objects like a living room chair. Activities of daily living limitation consist of five activities: bathing or showering, eating, dressing, walking across a room, and getting in or out of bed. Gross motor function consists of four activities: walking one block, walking across a room, climbing one flight of stairs without resting, and bathing. Fine motor function consists of three activities: eating, dressing, and picking up a dime from a table. Instrumental activities of daily living consist of three activities: using a telephone, taking medication, and handling money. Functional limitation of a specific category is defined as having difficulty (i.e., an answer of “yes”) in performing at least one of the activities included in that category. Any functional limitation is defined as having one or more of the functional limitations in these six categories. Depression Depressive symptoms were measured by the eight-item Center for Epidemiologic Studies Depression Scale (CES-D), a shortened version of the 20-item CES-D [12]. Participants were asked whether (“yes” or “no”) they felt depressed, felt that everything was an effort, slept restlessly, could not get going, felt sad, felt lonely, enjoyed life, and were happy in the past week. The two positive items (i.e., “enjoyed life” and “was happy”) are reverse-coded, so that a higher score indicates a more depressed mood. The eight-item CES-D total score, ranging from zero to eight, sums up the presence of (coded as one) or absence (coded as zero) from each of the eight feelings. Melchior et al. (1993) reported that the eight-item and 20-item CES-D scales were highly correlated (r = 0.93) and had comparable discriminant validity [13]. A cutoff score of three has been suggested by previous validation studies to indicate clinically relevant depressive symptoms [14]. This cut-off score has a sensitivity of 0.71 and a specificity of 0.79 to predict major depressive episodes [14]. A participant was classified as having a depression onset in a survey wave if one scored three or above on the eight-item CES-D in that wave. Other individual characteristics We controlled both wave-invariant and wave-variant individual characteristics in the regression analyses. Wave-invariant covariates include gender, race/ethnicity (non-Hispanic white, non-Hispanic African American, non-Hispanic other race or multi-race, and Hispanic), and education (education less than high school, high school, college, and education higher than college). Wave-variant covariates include age in years, marital status (married or living with partner, and unmarried, divorced, separated, or widowed), household net wealth (divided into four quartiles based on the wealth distribution in each survey wave), current smoking status, heavy drinking status (defined as one or more drink per day on average or four or more drinks on any occasion in the past three months for women, and two or more drinks per day on average or four or more drinks on any occasion in the past three months for men), depression status (CES-D score of three and above), diagnosis of a chronic condition (hypertension, diabetes, heart disease, stroke, lung disease, arthritis, cancer, and memory related disease), residential census region (Midwest, Northeast, South, and West), and body weight status. Body mass index (BMI) was calculated from self-reported height and weight. Body weight status was classified into four categories based on the international classification of adult BMI values – underweight (BMI b 18.5 kg/m2), normal weight (18.5 kg/m2 ≤ BMI b 25 kg/m2), overweight (25 kg/m2 ≤ BMI b 30 kg/m2), and obesity (BMI ≥ 30 kg/m2) [15]. Study sample PDS 2005 wave data were merged to HRS 2006 wave data, and PDS 2007 wave data were merged to HRS 2008 wave data, both using personal identifiers that uniquely identified each study participant.
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
Participants with missing values on functional limitations and/or other individual characteristics were excluded from statistical analyses. We also excluded those who reported any functional limitation in 2004 wave, resulting in a final sample size of 4242. Among these individuals who were free from any functional limitation in 2004 wave, 3530, 2794, 3230, 1100, 1669, 1039, and 984 developed any functional limitation, physical mobility limitation, large muscle function limitation, activities of daily living limitation, gross motor function limitation, fine motor function limitation, and instrumental activities of daily living limitation by 2008 wave. Statistical analyses Cox proportional hazards regressions were performed to examine the relationship between antidepressant use reported in 2005 and 2007 waves and functional limitation reported in 2006 and 2008 waves, adjusted by both wave-invariant and wave-variant individual characteristics. The key independent variable was a dichotomous variable for any antidepressant use. A participant was considered a “survivor” till a functional limitation was reported, if ever. Participants who died or were lost during follow-up without ever reporting a functional limitation were censored at the last wave when they were interviewed. Participants that were alive and remained free from a functional limitation during the study period were censored at the last (2008) wave of the study. We modeled any functional limitation and each of the six specific functional limitation categories in separate regressions. We further divided antidepressant users into those who took the medication for at least a year and those less than a year, and estimated the relationship between antidepressant use and functional limitation among these two groups in separate regressions. We also performed separate regressions on those depressed versus nondepressed at baseline. To assess the differential effects on functional limitation across antidepressant type, we included four dichotomous variables for TCA, SNRI, SSRI, and atypical antidepressant use as independent variables in the regression (MAOIs and pharmacotherapeutic combination antidepressants were excluded because in total only four study participants took them). Both descriptive statistics and regressions were weighted by the HRS baseline sampling weights. In regression analyses, Eicker–Huber–White sandwich estimator was used to calculate standard errors clustered at individual level to account for potential within-individual serial correlations that might downward bias the standard errors of estimates. All statistical analyses were conducted using Stata 14.1 SE version (StataCorp, College Station, TX). Results Table 1 reports baseline sample characteristics by antidepressant use status. Among those who were free from any functional limitation in 2004 wave, 83% developed any functional limitation, 66% developed physical mobility limitation, 76% developed large muscle function limitation, 26% developed activities of daily living limitation, 39% developed gross motor function limitation, 25% developed fine motor function limitation, and 23% developed instrumental activities of daily living limitation by 2008 wave. Less than one in six (16%) older adults used any antidepressant in 2005 wave. A majority of the users (75%) took antidepressant for one year or longer. Among users, the prevalence of TCA, SNRI, SSRI, and atypical antidepressant use was 13%, 11%, 26%, and 60%, respectively. Antidepressant users had substantially higher incidence of function limitation compared to nonusers. The incidence of any functional limitation, physical mobility limitation, large muscle function limitation, activities of daily living limitation, gross motor function limitation, fine motor function limitation, and instrumental activities of daily living limitation among antidepressant users were 91%, 78%, 85%, 34%, 54%, 31%, and 29%, respectively, compared to 82%, 64%, 75%, 24%, 37%, 23%, and 22% among nonusers.
33
Study participants averaged 73 years of age in 2006, and nearly four fifths (79%) were non-Hispanic whites. Only slightly less than a third (32%) fell within the normal weight range. Underweight was rare (2%), whereas the majority (66%) were overweight or obese (BMI ≥ 25). About one in eleven (9%) study participants were heavy drinkers, and less than one in eight (12%) current smokers. Chronic conditions were common: prevalence of depression 25%, hypertension 64%, diabetes 23%, heart disease 31%, stroke 10%, lung disease 12%, arthritis 69%, cancer 18%, and memory related disease 2%. Table 2 reports the adjusted hazard ratios (AHRs) for any functional limitation estimated in Cox proportional hazard regression. Prior-year antidepressant use prospectively predicted functional limitation onset. Compared with nonusers, antidepressant users were 7% (AHR = 1.07, 95% confidence interval [CI] = 1.03, 1.10) more likely to report a functional limitation during the study period. Women were significantly more likely to report functional limitation than men. The risk for functional limitation closely coincided with aging. Non-Hispanic African Americans appeared less likely to report functional limitation compared with their non-Hispanic white counterparts. Smoking but not heavy drinking was found to be associated with an elevated risk of functional limitation. There appeared to be a monotonic relationship between education level or household net wealth and functional limitation — lower education attainment and household net wealth were consistently associated with higher risk of functional limitation. Overweight and obesity were positively associated with functional limitation. Most chronic conditions, except hypertension and cancer, were founded to be linked with functional limitation. Table 3 reports the estimated effects of antidepressant use on alternative types of functional limitation. Antidepressant use was associated with an increase in the risk of physical mobility limitation by 8% (AHR = 1.08, 95% CI = 1.02, 1.14), large muscle function limitation by 6% (AHR = 1.06, 95% CI = 1.01, 1.11), activities of daily living limitation by 15% (AHR = 1.15, 95% CI = 1.01, 1.32), and gross motor function limitation by 19% (AHR = 1.19, 95% CI = 1.09, 1.31). The effect of antidepressant use on fine motor function limitation (AHR = 1.08, 95% CI = 0.94, 1.25) and instrumental activities of daily living limitation (AHR = 1.06, 95% CI = 0.92, 1.23) were both positive but statistically nonsignificant at P b 0.05. The effect of antidepressant use on functional limitation was centered on long-term users and currently nondepressed older adults. Antidepressant use for one year or longer was associated with an increase in the risk of functional limitation by 8% (AHR = 1.08, 95% CI = 1.04, 1.12), whereas the estimated association between antidepressant use less than a year and function limitation (AHR = 1.03, 95% CI = 0.97, 1.09) was statistically nonsignificant. Antidepressant use significantly predicted functional limitation onset among those currently nondepressed (AHR = 1.08, 95% CI = 1.03, 1.13) but not among currently depressed study participants (AHR = 1.01, 95% CI = 0.99, 1.05). The effect of antidepressant use on functional limitation was not found to differ by medication type (i.e., TCA, SNRI, SSRI, and atypical antidepressants). Discussion Antidepressants have been documented to be effective in treating major depressive symptoms in primary and secondary care settings [16,17]. However, findings from recent systematic reviews are mixed regarding the efficacy of antidepressants in the treatment of depression among older adults. Wilson et al (2001) conducted a meta-analysis on 17 placebo-controlled randomized trials that evaluated the effect of antidepressant treatment on geriatric depression [18]. TCAs, SSRIs and MAOIs were found to be effective in both institutionalized and community-dwelling patients, but lower dose TCA treatment was suggested. In contrast, Taylor and Doraiswamy (2004) performed a systematic review and meta-analysis on 18 antidepressant placebocontrolled randomized trials on patients 55 years of age and above,
34
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
Table 1 Descriptive statistics of study sample. Variable
Entire sample
Any antidepressant use in 2005
No antidepressant use in 2005
Sample size
4242
3584
658
83.2 65.9 76.1 25.9 39.3 24.5 23.2
91.0 78.1 85.0 34.3 53.8 31.3 29.5
81.8 63.6 74.5 24.4 36.7 23.2 22.0
16.0 (14.7, 17.3) 25.2 (21.8, 28.6) 74.8 (71.4, 78.2) 13.2 (10.5, 15.9) 11.4 (8.9, 13.9) 60.2 (56.3, 64.0) 25.5 (22.1, 29.0)
/ / / / / / /
/ / / / / / /
42.2 (40.4, 44.0) 57.8 (56.0, 59.6)
43.5 (41.6, 45.5) 56.5 (54.5, 58.4)
35.1 (30.7, 39.5) 64.9 (60.5, 69.3)
72.5 (72.1, 72.8)
72.4 (72.1, 72.8)
72.6 (71.6, 73.5)
79.3 (77.9, 80.6) 10.2 (9.2, 11.1) 2.6 (1.9, 3.2) 8.0 (7.1, 8.9)
78.1 (76.5, 79.6) 10.7 (9.7, 11.8) 2.8 (2.0, 3.5) 8.4 (7.4, 9.4)
85.6 (82.6, 88.6) 7.2 (5.1, 9.4) 1.4 (0.4, 2.5) 5.8 (3.8, 7.7)
25.9 (24.3, 27.4) 53.4 (51.6, 55.2) 13.6 (12.4, 14.8) 7.1 (6.2, 8.1)
26.7 (25.0, 28.4) 52.9 (50.9, 54.8) 13.4 (12.0, 14.7) 7.1 (6.1, 8.1)
21.6 (18.0, 25.2) 56.2 (51.7, 60.7) 14.9 (11.7, 18.2) 7.3 (4.9, 9.7)
54.7 (52.9, 56.5) 45.3 (43.5, 47.1)
55.1 (53.1, 57.0) 44.9 (43.0, 46.9)
52.7 (48.2, 57.2) 47.3 (42.8, 51.8)
27.4 (25.8, 29.0) 24.7 (23.2, 26.2) 22.9 (21.4, 24.3) 25.0 (23.5, 26.6)
27.6 (25.8, 29.4) 24.3 (22.7, 26.0) 22.7 (21.2, 24.3) 25.3 (23.6, 27.0)
26.2 (22.1, 30.3) 26.5 (22.5, 30.4) 23.6 (19.8, 27.4) 23.7 (19.9, 27.6)
9.4 (8.3, 10.5) 11.5 (10.3, 12.7)
9.7 (8.6, 10.9) 11.4 (10.1, 12.7)
7.6 (5.3, 10.0) 12.3 (9.1, 15.4)
1.9 (1.3, 2.4) 31.9 (30.2, 33.5) 37.6 (35.9, 39.3) 28.6 (27.0, 30.3)
1.7 (1.2, 2.2) 32.9 (31.1, 34.7) 37.8 (35.9, 39.7) 27.6 (25.8, 29.3)
2.6 (0.6, 4.6) 26.5 (22.5, 30.4) 36.6 (32.3, 41.0) 34.3 (30.0, 38.6)
24.9 (23.3, 26.5) 63.7 (61.9, 65.4) 23.2 (21.7, 24.7) 31.1 (29.5, 32.8) 9.5 (8.5, 10.6) 12.2 (11.0, 13.4) 69.0 (67.3, 70.6) 17.6 (16.2, 18.9) 1.8 (1.3, 2.4)
23.2 (21.5, 24.8) 61.3 (59.4, 63.2) 21.8 (20.2, 23.4) 28.5 (26.8, 30.3) 9.0 (7.9, 10.1) 11.0 (9.8, 12.2) 67.5 (65.7, 69.4) 17.7 (16.2, 19.2) 1.7 (1.2, 2.3)
34.2 (29.8, 38.5) 76.0 (72.0, 80.0) 30.4 (26.1, 34.7) 44.8 (40.3, 49.3) 12.5 (9.5, 15.5) 18.5 (14.8, 22.3) 76.5 (72.4, 80.5) 16.9 (13.5, 20.2) 2.4 (1.1, 3.8)
17.0 (15.6, 18.4) 24.5 (23.0, 26.0) 38.4 (36.7, 40.1) 20.1 (18.6, 21.5)
16.4 (14.9, 17.8) 24.7 (23.0, 26.3) 38.8 (36.9, 40.7) 20.1 (18.6, 21.7)
20.3 (16.6, 23.9) 23.8 (19.8, 27.8) 36.1 (31.9, 40.3) 19.9 (16.1, 23.6)
Functional limitation incidence during 2004–2008 (%) Any functional limitation Physical mobility limitation Large muscle function limitation Activities of daily living limitation Gross motor function limitation Fine motor function limitation Instrumental activities of daily living limitation Antidepressant use in 2005 (%) Any antidepressant use Any antidepressant use b1 year among users Any antidepressant use ≥1 year among users Tricyclic antidepressant (TCA) use among users Serotonin-norepinephrine reuptake inhibitor (SNRI) use among users Atypical antidepressant use among users Selective serotonin reuptake inhibitor (SSRI) use among users Gender (%) Male Female Age in 2006 (mean) Age in years Race/ethnicity (%) Non-Hispanic white Non-Hispanic African American Non-Hispanic other race/multi-race Hispanic Education in 2006 (%) Less than high school High school College Higher than college Marital status in 2006 (%) Married or living with partner Unmarried/divorced/separated/widowed Household wealth in 2006 (%) Lowest income quartile Mid-low income quartile Mid-high income quartile Highest-high income quartile Risk behavior in 2006 (%) Heavy drinking Currently smoking Body weight status in 2006 (%) Underweight (BMI b 18.5 kg/m2) Normal weight (18.5 kg/m2 ≤ BMI b 25 kg/m2) Overweight (25 kg/m2 ≤ BMI b 30 kg/m2) Obese (BMI ≥ 30 kg/m2) Chronic condition in 2006 (%) Depression Hypertension Diabetes Heart disease Stroke Lung disease Arthritis Cancer Memory related disease Residential census region in 2006 (%) Northeast Midwest South West
Notes: Study sample consisted of Health and Retirement Study (HRS) participants that were free from any functional limitation in 2004 wave. Statistics were weighted by HRS 2006 wave sampling weights (except for sample size and functional limitation incidence during 2004–2008, which were unweighted). 95% confidence intervals are in parentheses.
but no concrete conclusions were drawn due to various study limitations such as small sizes in individual trials, exclusion of common comorbid conditions, lack of controlled head-to-head comparisons, and other methodological design concerns [19]. Nelson and Devanand
(2011) conducted a systematic review and meta-analysis of antidepressant placebo-controlled randomized trials in older adults with depression and dementia [20]. The efficacy of antidepressant treatment was unconfirmed for that patient population. Parikh (2000) summarized the
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36 Table 2 Adjusted hazard ratios for any functional limitation estimated in Cox proportional hazards regression. Independent variable Antidepressant use Any antidepressant use No antidepressant use Gender Male Female Age Age in years Race/ethnicity Non-Hispanic white Non-Hispanic African American Non-Hispanic other race/multi-race Hispanic Education Less than high school High school College Higher than college Current marital status Married or living with partner Unmarried/divorced/separated/widowed Household wealth Lowest income quartile Mid-low income quartile Mid-high income quartile Highest-high income quartile Risk behavior Heavy drinking Currently smoking Body weight status Underweight (BMI b 18.5 kg/m2) Normal weight (18.5 kg/m2 ≤ BMI b 25 kg/m2) Overweight (25 kg/m2 ≤ BMI b 30 kg/m2) Obese (BMI ≥ 30 kg/m2) Chronic condition Depression Hypertension Diabetes Heart disease Stroke Lung disease Arthritis Cancer Memory related disease Residential census region Northeast Midwest South West
Adjusted hazard ratio 1.07⁎⁎⁎ (1.03, 1.10) Reference 0.92⁎⁎⁎ (0.89, 0.96) Reference 1.01⁎⁎⁎ (1.00, 1.01) Reference 0.90⁎⁎⁎ (0.86, 0.95) 1.07 (0.95, 1.21) 0.99 (0.94, 1.04) Reference 0.98 (0.95, 1.02) 0.92⁎ (0.86, 0.98) 0.84⁎⁎ (0.76, 0.93) 1.00 (0.96, 1.03) Reference Reference 0.98 (0.94, 1.01) 0.94⁎⁎ (0.90, 0.98) 0.85⁎⁎⁎ (0.81, 0.90) 0.97 (0.90, 1.05) 1.06⁎ (1.01, 1.12) 1.11⁎ (1.01, 1.22) Reference 1.06⁎⁎ (1.01, 1.10) 1.15⁎⁎⁎ (1.10, 1.20) 1.14⁎⁎⁎ (1.11, 1.17) 1.03 (0.99, 1.07) 1.04⁎ (1.01, 1.08) 1.07⁎⁎⁎ (1.04, 1.11) 1.06⁎⁎ (1.02, 1.10) 1.09⁎⁎⁎ (1.06, 1.12) 1.39⁎⁎⁎ (1.32, 1.47) 0.98 (0.94, 1.02) 1.08⁎⁎ (1.02, 1.14) Reference 1.05 (1.00, 1.10) 1.06⁎ (1.01, 1.11) 1.01 (0.95, 1.07)
Notes: Study sample (N = 4242) came from the Health and Retirement Study (HRS) 2006–2008 waves and consisted of those who were free from any functional limitation in 2004 wave. Cox proportional hazards regressions were performed to estimate the adjusted hazard ratios for any functional limitation, weighted by the HRS 2006 wave sampling weights. 95% confidence intervals are in parentheses. ⁎ P b 0.05. ⁎⁎ P b 0.01. ⁎⁎⁎ P b 0.001.
challenges in study design and result interpretation for evaluating the efficacy of antidepressant treatment in older adults [21]. Most Phase II and III clinical trials on antidepressants had been on younger adults and excluded very old and frail patients. Design limitations of the few trials conducted in older patients precluded accurate interpretation of differences in efficacy or safety between drugs. In sum, findings from previous studies on the effectiveness of antidepressant use in treating depression among older adults remain mixed, and none of the studies focused on the potential side effect of antidepressant use on functional limitations that are highly prevalent in older adults and significantly influence their daily activities and quality of life. This study examined the associations between antidepressant use and functional limitations among U.S. older adults using data from a
35
Table 3 Estimated associations between antidepressant use and functional limitation by limitation type. Type of functional limitation
Adjusted hazard ratio
Any functional limitation Physical mobility limitation Large muscle function limitation Activities of daily living limitation Gross motor function limitation Fine motor function limitation Instrumental activities of daily living limitation
1.07⁎⁎⁎ (1.03, 1.10) 1.08⁎⁎ (1.02, 1.14) 1.06⁎⁎ (1.01, 1.11) 1.15⁎ (1.01, 1.32) 1.19⁎⁎⁎ (1.09, 1.31) 1.08 (0.94, 1.25) 1.06 (0.92, 1.23)
Notes: Study sample came from the Health and Retirement Study (HRS) 2006–2008 waves, consisting of 4,242 participants who did not have any functional limitation in 2004 wave. Cox proportional hazards regressions were performed to estimate the adjusted hazard ratios for functional limitation, weighted by the HRS 2006 wave sampling weights and adjusted by individual characteristics specified in Table 1. 95% confidence intervals are in parentheses. ⁎ P b 0.05. ⁎⁎ P b 0.01. ⁎⁎⁎ P b 0.001.
nationally representative longitudinal survey. Antidepressant use prospectively predicted future functional limitation onset. Antidepressant use for one year or longer was associated with an increase in the risk of functional limitation by 8%, whereas the relationship between antidepressant use less than a year and function limitation was statistically nonsignificant. Antidepressant use was associated with an increase in the risk of functional limitation by 8% among currently nondepressed participants but not currently nondepressed participants. The relationship between antidepressant use and functional limitations differed across limitation category. The impact of antidepressant use on functional limitations was not found to differ by medication type. The estimated association between antidepressant use and functional limitations was only statistically significant in long-term users. The declining physical, mental, and cognitive conditions could make older adults disproportionately susceptible to the side effects of antidepressants. Over time, these side effects, if not properly managed, would create barriers for older adults to accomplish some of their daily routines. Moreover, the impact of antidepressant use on functional limitations was only identified among currently nondepressed older adults. It is not known whether their depressive symptoms were successfully managed due to antidepressant use, they took antidepressants mainly for preventing a relapse of depression, or they took the medication for other reasons, such as treating insomnia or neuropathic pain [22,23]. We also could not rule out the possibility that the null finding on the relationship between antidepressant use and functional limitation among nondepressed older adults and short-term antidepressant users was a consequence of lacking statistical power owing to small subsample size. Nevertheless, the study results indicate the needs for carefully weighing the benefits and costs, particularly in terms of the elevated risk for functional limitations, for long-term antidepressant use in older adults. The finding also warrants future research on the health consequences of extended antidepressant use in absence of depressive symptoms as well as off-label antidepressant use in treating nondepression conditions among older adults. The estimated impact of antidepressant use on fine motor function and instrumental activities of daily living was statistical nonsignificant. Although we were unable to eliminate the possibility of an underpowered estimation, the null finding might partially be explained by the differences between these two types of functional limitations and the other limitation categories. Whereas the measures for physical mobility limitation, large muscle function limitation, activities of daily living limitation, and gross motor function limitation focused on physicallydemanding labor-intensive tasks (e.g., walking, pulling or pushing, climbing), measures for fine motor function and instrumental activities of daily living tested less physically-demanding but more technical skills (e.g., dime picking, money handling). Antidepressant use could have less of an impact on those technical skills (the positive effect
36
R. An, L. Lu / Journal of Psychosomatic Research 80 (2016) 31–36
through depressive symptom management and the negative impact due to medication side effect canceled out) than on those relatively physically-demanding tasks. A few limitations of the study should be noted. This study was conducted based on data from a prospective cohort study. Despites the effort of controlling for an extensive list of individual characteristics, we were unable to completely eliminate the risk of confounding bias resulted from unobserved differences in individual characteristics due to the lack of a randomized study design. Therefore, the results must be interpreted as correlations rather than causations. The study findings might not be generalizable to the entire aging population as the study sample of community-dwelling older adults were generally less fragile relative to their institutionalized counterparts. As commonly seen in other panel studies on aging population [24], this study might suffer from attrition problem. Banks et al. (2011) found household net wealth to be negatively correlated with attrition rate in HRS [25]. As richer older adults were less likely to develop functional limitations and more likely to retain in the study sample, attrition attributable to wealth could result in underestimation of the effect of antidepressant use on functional limitations. Although the HRS sample size was reasonably large, further stratification by antidepressant type substantially reduced the statistical power and estimation precision, which might explain why no significant differences in the effects of TCA, SNRI, SSRI, and atypical antidepressants on functional limitations were identified. It would potentially be more ideal if we could also exclude those who had any functional limitation in 2006 from the analyses, because data on antidepressant use was only available from 2005 onward. However, the main challenge was that we would be left with only a small fraction of the sample due to the very high incidence of any functional limitation (given the average age of study participants of 73 in 2006), and even smaller samples after stratification by antidepressant use status and length of use, which would substantially compromise the statistical power of our estimation. Chronic condition diagnoses were limited to a few diseases, whereas certain illnesses that are likely to be associated with functional limitations in older adults such as Parkinson's disease, Alzheimer, dementia, osteoporosis, and chronic pain, were not administered in HRS. One question regarding memory related disease was administered in HRS 2006 and 2008 waves and was used in regression analyses, but more detailed classification of that disease category (Alzheimer, dementia, etc.) became only available in the 2010 wave and after. The efficacy of antidepressant use to treat minor depressions has been questioned. Barbui et al. (2011) conducted a systematic review and meta-analysis on placebo-controlled randomized trials that evaluated the effect of antidepressant treatment on minor depression [26]. No clinically meaningful advantage for antidepressants over placebo was found in treating individuals with minor depressive symptoms. Wiese [2011] suggested the selection of an antidepressant medication in treating geriatric depression be based on the best side effect profile and the lowest risk of drug-drug interaction [27]. Considering the high prevalence of comorbidity and vulnerability in older adults, antidepressant use might not be recommended for treating minor depressive symptoms, and long-term antidepressant use should be prudently evaluated and regularly monitored in order to prevent adverse health consequences, including the risk of functional limitations. In conclusion, this study examined the impact of antidepressant use on functional limitations among U.S. older adults using data from a nationally representative longitudinal survey. Antidepressant use prospectively predicted future functional limitation onset. The associations between antidepressant use and functional limitations differed across limitation category, and were only identified among long-term antidepressant users and currently nondepressed older adults. Long-term antidepressant use in older adults should be prudently evaluated and regularly monitored in order to prevent adverse health consequences, including the risk of functional limitations. Moreover, future research is warranted to examine the health consequences of extended
antidepressant use in absence of depressive symptoms as well as offlabel antidepressant use in treating non-depression conditions in older adults. Conflict of interest We have no conflict of interests relating to this manuscript to declare. References [1] L.A. Pratt, D.J. Brody, Q. Gu, Antidepressant use in persons aged 12 and over: United States, NCHS Data Brief 2011 (76) (2005-2008) 1–8. [2] R. Paulose-Ram, M.A. Safran, B.S. Jonas, Q. Gu, D. Orwig, Trends in psychotropic medication use among U.S. adults, Pharmacoepidemiol. Drug Saf. 16 (5) (2007) 560–570. [3] M. Olfson, S.C. Marcus, National patterns in antidepressant medication treatment, Arch. Gen. Psychiatry 66 (8) (2009) 848–856. [4] Mayo Clinic, Antidepressants: Get Tips to Cope with side Effects, 2013 (Available from: http://www.mayoclinic.org/diseases-conditions/depression/in-depth/antidepressants/art-20049305). [5] R. Rizzoli, C. Cooper, J.Y. Reginster, B. Abrahamsen, J.D. Adachi, M.L. Brandi, O. Bruyère, J. Compston, P. Ducy, S. Ferrari, N.C. Harvey, J.A. Kanis, G. Karsenty, A. Laslop, V. Rabenda, P. Vestergaard, Antidepressant medications and osteoporosis, Bone 51 (3) (2012) 606–613. [6] Q. Wu, J. Liu, J.F. Gallegos-Orozco, J.G. Hentz, Depression, fracture risk, and bone loss: a meta-analysis of cohort studies, Osteoporos. Int. 21 (10) (2010) 1627–1635. [7] D. Prieto-Alhambra, H. Petri, J.S. Goldenberg, T.P. Khong, O.H. Klungel, N.J. Robinson, F. de Vries, Excess risk of hip fractures attributable to the use of antidepressants in five European countries and the USA, Osteoporos. Int. 25 (3) (2014) 847–855. [8] M.S. Bakken, A. Engeland, L.B. Engesæter, A.H. Ranhoff, S. Hunskaar, S. Ruths, Increased risk of hip fracture among older people using antidepressant drugs: data from the Norwegian Prescription Database and the Norwegian Hip Fracture Registry, Age Ageing 42 (4) (2013) 514–520. [9] Multum Lexicon database, National Health and Nutrition Examination Survey—1988-2008 data documentation, codebook, and frequencies, 2010 (Available from: http://www.cdc.gov/nchs/nhanes/nhanes1999-2000/RXQ_DRUG.htm). [10] S. Chien, N. Campbell, O. Hayden, et al., RAND HRS Data Documentation, Version N, 2014 (Available from: http://hrsonline.isr.umich.edu/modules/meta/rand/ randhrsn/randhrsN.pdf). [11] R.B. Wallace, A.R. Herzog, Overview of the health measures in the Health and Retirement Study, J. Hum. Resour. 30 (1995) S84–S107. [12] F.J. Kohout, L.F. Berkman, D.A. Evans, J. Cornoni-Huntley, Two shorter forms of the CES-D (Center for Epidemiological Studies Depression) depression symptoms index, J. Aging Health 5 (2) (1993) 179–193. [13] L.A. Melchior, G.J. Huba, V.B. Brown, C.J. Reback, A short depression index for women, Educ. Psychol. Meas. 53 (4) (1993) 1117–1125. [14] C.L. Turvey, R.B. Wallace, R. Herzog, A revised CES-D measure of depressive symptoms and a DSM-based measure of major depressive episodes in the elderly, Int. Psychogeriatr. 11 (2) (1999) 139–148. [15] World Health Organization, The international classification of adult underweight, overweight and obesity according to BMI, 2006 (Available from: http://apps.who. int/bmi/index.jsp?introPage=intro_3.html). [16] B. Arroll, C.R. Elley, T. Fishman, F.A. Goodyear-Smith, T. Kenealy, G. Blashki, N. Kerse, S. Macgillivray, Antidepressants versus placebo for depression in primary care, Cochrane Database Syst. Rev. 3 (2009), CD007954. [17] F.L. Rocha, C. Fuzikawa, R. Riera, C. Hara, Combination of antidepressants in the treatment of major depressive disorder: a systematic review and meta-analysis, J. Clin. Psychopharmacol. 32 (2) (2012) 278–281. [18] K. Wilson, P.G. Mottram, A. Sivananthan, A. Nightingale, Antidepressants versus placebo for the depressed elderly, Cochrane Database Syst. Rev. 2 (2001), CD000561. [19] W.D. Taylor, P.M. Doraiswamy, A systematic review of antidepressant placebocontrolled trials for geriatric depression: limitations of current data and directions for the future, Neuropsychopharmacology 29 (12) (2004) 2285–2299. [20] J.C. Nelson, D.P. Devanand, A systematic review and meta-analysis of placebocontrolled antidepressant studies in people with depression and dementia, J. Am. Geriatr. Soc. 59 (4) (2011) 577–585. [21] C. Parikh, Antidepressants in the elderly: challenges for study design and their interpretation, Br. J. Clin. Pharmacol. 49 (6) (2000) 539–547. [22] W.A. Santos Moraes, P.R. Burke, P.L. Coutinho, C. Guilleminault, A.G. Bittencourt, S. Tufik, D. Poyares, Sedative antidepressants and insomnia, Rev. Bras. Psiquiatr. 33 (1) (2011) 91–95. [23] T. Saarto, P.J. Wiffen, Antidepressants for neuropathic pain, Cochrane Database Syst. Rev. 4 (2007), CD005454. [24] R. Groves, M. Couper, Nonresponse in Household Surveys, John Wiley and Sons, New York, 1998. [25] J. Banks, A. Muriel, J. Smith, Attrition and health in aging studies: evidence from ELSA and HRS, Longit. Life Course Stud 2 (2) (2011) 101–126. [26] C. Barbui, A. Cipriani, V. Patel, J.L. Ayuso-Mateos, M. van Ommeren, Efficacy of antidepressants and benzodiazepines in minor depression: systematic review and meta-analysis, Br. J. Psychiatry 198 (1) (2011) 11–16. [27] B.S. Wiese, Geriatric depression: the use of antidepressants in the elderly, BCMJ 53 (7) (2011) 341–347.