9
CLINICAL INVESTIGATIONS Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women Jeanne E. Maglione, MD, PhD,* Sonia Ancoli-Israel, PhD,* Katherine W. Peters, MS,Misti L. Paudel, MPH, Kristine Yaffe, MD, §¶ ** Kristine E. Ensrud, MD, MPH,‡††‡‡ and Katie L. Stone, PhD OBJECTIVES: To examine the relationship between depressive symptoms and subjective and objective sleep in older women. DESIGN: Cross-sectional. SETTING: Four U.S. clinical centers. PARTICIPANTS: Three thousand forty-five community- dwelling women aged 70 and older. MEASUREMENTS: Depressive symptoms were assessed using the Geriatric Depression Scale, categorizing partici- pants as normal (02, reference), some depressive symp- toms (35), or depressed ( 6). Subjective sleep quality and daytime sleepiness were assessed using the Pittsburgh Sleep Quality Index (PSQI) and Epworth Sleepiness Scale (ESS). Objective sleep measures were assessed using wrist actigraphy. RESULTS: In multivariable-adjusted models, there were graded associations between greater level of depressive symptoms and worse subjective sleep quality and more subjective daytime sleepiness (P-trends < .001). Women with some depressive symptoms (odds ratio (OR) = 1.82, 95% confidence interval (CI) = 1.482.24) and depressed (OR = 2.84, 95% CI = 2.083.86) women had greater odds of reporting poor sleep (PSQI>5). Women with some depressive symptoms (OR = 1.97, 95% CI = 1.472.64) and depressed women (OR = 1.70, 95% CI = 1.122.58) had greater odds of reporting excessive daytime sleepiness (ESS>10). There were also graded associations between greater level of depressive symptoms and objectively mea- sured wake after sleep onset (WASO) (P-trend = .03) and wake episodes longer than 5 minutes (P-trend = .006). Depressed women had modestly higher odds of WASO of 1 hour or longer (OR = 1.37, 95% CI = 1.031.83). Women with some depressive symptoms (OR = 1.49, 95% CI = 1.191.86) and depressed women (OR = 2.04, 95% CI = 1.522.74) had greater odds of being in the highest quartile for number of nap episodes longer than 5 minutes. No associations between depressive symptom level and prolonged sleep latency, poor sleep efficiency, or short or long total sleep time were found. CONCLUSION: Greater depressive symptom levels were associated with more subjective sleep disturbance and objective evidence of sleep fragmentation and napping. J Am Geriatr Soc 60:635–643, 2012. Key words: depression; sleep; actigraphy; elderly; age D epression is a common complaint in older adults. The estimated prevalence of significant depressive symp- toms in community-dwelling older adults is 15%, 1 and depression in this age group has been associated with a number of adverse outcomes, including functional impair- ment, 2,3 medical illness, 4 disability, 5,6 and greater health services utilization. 79 Poor sleep is also common in older adults. In one sur- vey of 9,000 adults aged 65 and older, more than 80% of those surveyed reported at least one problem with sleep, and more than half reported that at least one sleep com- plaint occurred nearly all the time. 10 Furthermore, older adults may underreport the severity of sleep disturbances compared with objective assessments by polysomnography (PSG). 11 Insomnia has been associated with poor out- comes in older adults with depression, 12 and poor subjec- tive sleep quality is a risk factor for suicide in older adults. 13 From the * Department of Psychiatry, University of California at San Diego, La Jolla, California; California Pacific Medical Center Research Institute, San Francisco, California; Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, Minnesota; § Department of Psychiatry, University of California at San Francisco, San Francisco, California; Department of Neurology, University of California at San Francisco, San Francisco, California; ** Department of Epidemiology, University of California at San Francisco, San Francisco, California; †† Department of Medicine, University of Minnesota, Minneapolis, Minnesota; and ‡‡ Center for Chronic Disease Outcomes Research, Veterans Affairs Medical Center, Minneapolis, Minnesota. Address correspondence to Jeanne E. Maglione, Gillin Sleep and Chronobiology Research Center, 9500 Gilman Drive #0733, La Jolla, CA 92093. E-mail: [email protected] DOI: 10.1111/j.1532-5415.2012.03908.x JAGS 60:635–643, 2012 © 2012, Copyright the Authors Journal compilation © 2012, The American Geriatrics Society 0002-8614/12/$15.00

Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

Embed Size (px)

Citation preview

Page 1: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

CLINICAL INVESTIGATIONS

Depressive Symptoms and Subjective and Objective Sleep inCommunity-Dwelling Older Women

Jeanne E. Maglione, MD, PhD,* Sonia Ancoli-Israel, PhD,* Katherine W. Peters, MS,†Misti L. Paudel, MPH,‡ Kristine Yaffe, MD,§¶** Kristine E. Ensrud, MD, MPH,‡††‡‡ andKatie L. Stone, PhD†

OBJECTIVES: To examine the relationship betweendepressive symptoms and subjective and objective sleep inolder women.

DESIGN: Cross-sectional.

SETTING: Four U.S. clinical centers.

PARTICIPANTS: Three thousand forty-five community-dwelling women aged 70 and older.

MEASUREMENTS: Depressive symptoms were assessedusing the Geriatric Depression Scale, categorizing partici-pants as normal (0–2, reference), some depressive symp-toms (3–5), or depressed (�6). Subjective sleep qualityand daytime sleepiness were assessed using the PittsburghSleep Quality Index (PSQI) and Epworth Sleepiness Scale(ESS). Objective sleep measures were assessed using wristactigraphy.

RESULTS: In multivariable-adjusted models, there weregraded associations between greater level of depressivesymptoms and worse subjective sleep quality and moresubjective daytime sleepiness (P-trends < .001). Womenwith some depressive symptoms (odds ratio (OR) = 1.82,95% confidence interval (CI) = 1.48–2.24) and depressed(OR = 2.84, 95% CI = 2.08–3.86) women had greaterodds of reporting poor sleep (PSQI>5). Women with somedepressive symptoms (OR = 1.97, 95% CI = 1.47–2.64)and depressed women (OR = 1.70, 95% CI = 1.12–2.58)had greater odds of reporting excessive daytime sleepiness(ESS>10). There were also graded associations between

greater level of depressive symptoms and objectively mea-sured wake after sleep onset (WASO) (P-trend = .03) andwake episodes longer than 5 minutes (P-trend = .006).Depressed women had modestly higher odds of WASO of1 hour or longer (OR = 1.37, 95% CI = 1.03–1.83).Women with some depressive symptoms (OR = 1.49, 95%CI = 1.19–1.86) and depressed women (OR = 2.04, 95%CI = 1.52–2.74) had greater odds of being in the highestquartile for number of nap episodes longer than 5 minutes.No associations between depressive symptom level andprolonged sleep latency, poor sleep efficiency, or short orlong total sleep time were found.

CONCLUSION: Greater depressive symptom levelswere associated with more subjective sleep disturbanceand objective evidence of sleep fragmentation and napping.J Am Geriatr Soc 60:635–643, 2012.

Key words: depression; sleep; actigraphy; elderly; age

Depression is a common complaint in older adults. Theestimated prevalence of significant depressive symp-

toms in community-dwelling older adults is 15%,1 anddepression in this age group has been associated with anumber of adverse outcomes, including functional impair-ment,2,3 medical illness,4 disability,5,6 and greater healthservices utilization.7–9

Poor sleep is also common in older adults. In one sur-vey of 9,000 adults aged 65 and older, more than 80% ofthose surveyed reported at least one problem with sleep,and more than half reported that at least one sleep com-plaint occurred nearly all the time.10 Furthermore, olderadults may underreport the severity of sleep disturbancescompared with objective assessments by polysomnography(PSG).11 Insomnia has been associated with poor out-comes in older adults with depression,12 and poor subjec-tive sleep quality is a risk factor for suicide in olderadults.13

From the *Department of Psychiatry, University of California at SanDiego, La Jolla, California; †California Pacific Medical Center ResearchInstitute, San Francisco, California; ‡Division of Epidemiology andCommunity Health, University of Minnesota, Minneapolis, Minnesota;§Department of Psychiatry, University of California at San Francisco, SanFrancisco, California; ¶Department of Neurology, University of Californiaat San Francisco, San Francisco, California; **Department ofEpidemiology, University of California at San Francisco, San Francisco,California; ††Department of Medicine, University of Minnesota,Minneapolis, Minnesota; and ‡‡Center for Chronic Disease OutcomesResearch, Veterans Affairs Medical Center, Minneapolis, Minnesota.

Address correspondence to Jeanne E. Maglione, Gillin Sleep andChronobiology Research Center, 9500 Gilman Drive #0733, La Jolla, CA92093. E-mail: [email protected]

DOI: 10.1111/j.1532-5415.2012.03908.x

JAGS 60:635–643, 2012

© 2012, Copyright the Authors

Journal compilation © 2012, The American Geriatrics Society 0002-8614/12/$15.00

Page 2: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

Although insomnia in individuals with depressionwas initially conceptualized as a symptom of depression,recent evidence supports a bidirectional relationshipbetween insomnia and depression. Longitudinal studieshave shown that subjective sleep disturbance is a majorrisk factor for future development of first-onset andrecurrent depressive episodes in younger and olderadults,12,14–16 but the existing literature regarding therelationship between depression and sleep disturbances inolder adults has largely focused on subjective assessmentsof sleep disturbances. There is therefore little informationregarding the relationship between depression and objec-tively measured sleep disturbances in this age group. Thisrelationship was recently explored using actigraphy toassess objective sleep disturbances in a group of commu-nity-dwelling older men. Although a strong, graded asso-ciation between greater level of depressive symptoms andsubjective sleep disturbance was found, objective assess-ment of sleep revealed only a modest association betweengreater level of depressive symptoms and longer sleeplatency.17 To the knowledge of the authors of the currentstudy, no studies have examined the relationship betweendepression and objectively measured sleep in olderwomen.

This relationship was examined in a large group ofcommunity-dwelling older women. It was hypothesizedthat older women who endorsed more depressive symp-toms would be more likely to report subjective sleep dis-turbances and more likely to have objective evidence ofdisturbed nighttime sleep and greater daytime napping.

METHODS

Participants

Participants were women enrolled in the Study of Osteo-porotic Fractures (SOF), an ongoing, multicenter, prospec-tive, observational cohort study of primarily Caucasian,community-dwelling women aged 65 and older from fourmetropolitan areas (Portland, Oregon; Minneapolis, Min-nesota; Pittsburgh and the Monongahela Valley, Pennsyl-vania; Baltimore, Maryland). Women gave informedconsent before enrollment in the study. Between September1986 and October 1988, the 9,704 participants making upthe original cohort were recruited from community listingsand mailed announcements. Between February 1997 and1998, 662 African-American women were also recruited.Women were excluded from participation if they requiredassistance with ambulation or had undergone bilateral hipreplacement. Details regarding the study have been pub-lished previously.18,19 The current analyses focus onwomen participating in SOF Visit 8 (approximately15 years after baseline assessment), which occurredbetween January 2002 and February 2004. There were4,727 participants at Visit 8, representing 84% of activesurvivors; 3,052 of these had objective assessment of sleepby actigraphy, and 3,045 also returned completed Geriat-ric Depression Scale (GDS) questionnaires. The analyses ofnighttime sleep were performed on this subset of 3,045women. Fifty of these women did not have usable daytimedata because they removed their actigraph for longer than10% of waketime or because there were not two nights of

data surrounding the day. Hence, the analyses of numberof naps were performed on 2,995 subjects.

Subjective Measures

Demographic information (age, ethnicity, education) wasrecorded at baseline. At Visit 8, participants completedquestionnaires regarding health status, smoking, alcoholconsumption, caffeine intake, exercise, and medical his-tory. Medications were recorded and categorized accordingto a computerized coding dictionary. Cognition wasassessed using the Mini-Mental State Examination(MMSE), with impairment defined as a MMSE score of 24or lower.19

Depression was assessed using the GDS, a 15-item val-idated self-report questionnaire commonly used for assess-ment of depressive symptoms in older adults.20 Womenwere categorized into three groups (0–2 (none to fewdepressive symptoms), 3–5 (some depressive symptoms),�6 (many depressive symptoms)). To remain consistentwith prior publications using this strategy,17 these groupsare labeled normal (0–2), some depressive symptoms (3–5),and depressed (� 6) throughout the remainder of themanuscript. A standard cutoff of 6 has been shown tohave a sensitivity of 91% and specificity of 65% for diag-nosis of a major depressive episode compared with theDiagnostic and Statistical Manual of Mental Disorders,Fourth Edition (DSM-IV).21

Subjective sleep quality was assessed using the Pitts-burgh Sleep Quality Index (PSQI), a validated 19-item self-report questionnaire. Global PSQI scores range from 0 to21, with higher scores reflecting worse sleep. A total scoregreater than 5 has a sensitivity of 89.6% and specificity of86.5% for distinguishing good sleepers from poor sleep-ers.22 Subjective daytime sleepiness was measured usingthe Epworth Sleepiness Scale (ESS), a self-administeredquestionnaire that asks subjects to rate how likely they areto doze off in different situations. Scores on the ESS rangefrom 0 to 24, with a score greater than 10 indicatingexcessive daytime sleepiness.23,24

Actigraphy

Objective sleep parameters were measured using wristactigraphy, a previously validated25–27 noninvasive methodthat provides information about sleep and wake patternsthrough an accelerometer that detects wrist movement.Actigraphs (SleepWatch-O; Ambulatory Monitoring, Inc.,Ardsley, NY) were worn on participants’ nondominantwrist for at least three consecutive 24-hour periods. Move-ments were recorded and summarized in 1-minute epochs.The data were collected using the proportional integrationmode (PIM)26 and analyzed using ActionW-2 software(Ambulatory Monitoring, Inc.). The University of Californiaat San Diego algorithm was used to distinguish sleep fromwake.28,29 Participants also completed sleep diaries for thetime period they wore the actigraph. The diaries includedinformation about times participants got into and out ofbed and times the actigraph was removed. This informa-tion was used in editing the actigraphy data files to setintervals for when the participant was in bed trying tosleep, and to delete time when the actigraph was removed.

636 MAGLIONE ET AL. APRIL 2012–VOL. 60, NO. 4 JAGS

Page 3: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

In the SOF study, interscorer reliability for editing theactigraphy data files has previously been found to be high(total sleep time (TST) intraclass coefficient 0.95), andTST estimated using actigraphy has been shown to havegood concordance with TST assessed using PSG.25,26

The following variables were used in the analyses:TST (minutes sleep between bedtime and wake time), sleepefficiency (SE, percentage of time asleep while in bed),sleep onset latency (SOL, minutes between bedtime andthe first block of inactivity after bedtime), time awakeafter sleep onset (WASO, minutes awake between sleeponset and wake time), number of long wake episodes(number of wake episodes between sleep onset and waketime > 5 minutes), and number of nap episodes (numberof inactive episodes between wake time and sleeponset > 5 minutes). Daytime data were excluded from theanalysis if the participant took the actigraph off for morethan 10% of the wake period. Actigraphy data werecollected for an average of 4.1 ± 0.8 days. Data for eachvariable were averaged over the recorded time.

Statistical Analyses

Differences in participant characteristics according to levelof depressive symptoms were assessed using the chi-squaretest for categorical variables, analysis of variance for nor-mally distributed continuous data, and the Kruskal–Wallistest for skewed continuous data (caffeine intake and num-ber of medical conditions). Pearson correlations weredetermined to examine the relationship between subjectiveand objective sleep variables in the entire sample and ineach subgroup of women according to level of depressivesymptoms.

Objective and subjective sleep measures were analyzedas continuous variables using linear regression models, andthe least-square means strategy was used to estimate themean and 95% confidence interval (CI) for each sleepparameter according to level of depressive symptoms. Testsof linear trend were performed to detect graded associa-tions. Because the distributions of SE, SOL, WASO, andnumber of nap episodes were skewed, the data were trans-formed for normality (natural log (LN) (P) for WASO andSOL, LN (100 � P) for SE, LN (P + 1) for number napepisodes > 5 minutes).

Sleep variables were also expressed as dichotomousoutcomes. Specific categories were PSQI greater than 5 vs5 or less, ESS greater than 10 vs 10 or less, SE less than85% vs 85% or greater, SOL 30 minutes or more vs lessthan 30 minutes, WASO 1 hour or more vs less than1 hour, eight or more long wake episodes vs fewer thaneight. TST was expressed as a three-level outcome(<5 hours (short sleep duration), 5–8 hours (normal sleepduration), and > 8 hours (long sleep duration)). Becausethere were no previously described clinically relevant cut-points for number of nap episodes, quartiles were used,and the women in the highest quartile were compared withthe rest of the sample. Associations between depressivesymptoms and dichotomous sleep variables were assessedusing logistic regression and a test for linear trend. Theassociation between depressive symptoms and TST wasassessed using a multinomial regression model with normalsleepers serving as the referent category.

Models adjusted for age and site and multivariable-adjusted models were assessed. Covariates were includedin multivariable models if they were known correlates ofsleep disturbance or depression or if they were related tothe level of depressive symptoms in this population(P � .10). These included age, race, site, smoking status,body mass index (BMI), self reported health status, educa-tion, exercise, reported medical conditions, impairments inactivities of daily living, cognitive impairment, antidepres-sant use, benzodiazepine use, and use of medications forsleep.

RESULTS

Characteristics of the Study Population

Characteristics of the sample are shown in Table 1. Ageranged from 70 to 99 (mean 83.6 ± 3.8). The majority ofthe women were Caucasian (89.3%), rated their healthstatus as excellent or good (75.4%), lived alone (60.4%),were not smokers (97.2%), and reported drinking two orfewer alcoholic beverages per week (86.9%). The groupwas diverse with respect to education, with 40.1% havinggraduated from college, 41.9% having completed highschool, and 18.0% having completed less education. Theaverage number of medical conditions reported was2.46 ± 1.61. A minority of the sample population reporteduse of medications for sleep (14.0%) or antidepressants(13.7%). The rate of cognitive impairment (MMSEscore � 24) was low in the study population (5.6%), andthe mean MMSE score was 27.8 ± 2.0. Fifty-two percentof the population reported subjectively poor sleep(PSQI > 5), and 11.0% reported excessive daytime sleepi-ness (ESS > 10).

The sample was stratified into three groups accordingto level of depressive symptoms; 1,966 (64.6%) reportedzero to two depressive symptoms (normal), 718 (23.6%)reported three to five (some depressive symptoms), and361 (11.9%) reported six or more (depressed). Womenwith more depressive symptoms were more likely to beolder, to smoke, to be cognitively impaired, to report poorhealth, to be less educated, and to report use of antidepres-sants, benzodiazepines, and medications for sleep. Theywere also more likely to report impairment in instrumentalactivities of daily living (IADLs), less likely to walk forexercise, more likely to be obese, and reported more medi-cal problems.

In general, subjective and objective sleep measureswere weakly correlated or not significantly correlated inthis sample. For example there was a small but significantinverse correlation between PSQI and SE (correlation coef-ficient (r) = –0.12, P < .001) and small but significantdirect correlations between PSQI and WASO (r = 0.13,P < .001) and SOL (r = 0.08, P < .001), whereas PSQIand TST were not significantly correlated. There were alsosmall but significant correlations between ESS and TST(r = 0.17, P < .001), and WASO (r = 0.07, P < .001) inthe sample. These relationships were similar in all groupsof women with differing levels of depressive symptoms.There were no significant correlations between ESS and SEor SOL in the total sample or in any of the groups ofwomen according to depressive symptom level.

JAGS APRIL 2012–VOL. 60, NO. 4 DEPRESSIVE SYMPTOMS AND SLEEP IN OLDER WOMEN 637

Page 4: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

Level of Depressive Symptoms and Subjective SleepMeasures

Regardless of level of depressive symptoms, estimatedmean PSQI scores for all groups of women were greaterthan the clinical cutpoint of five, suggesting poor subjec-tive sleep quality (Table 2). After adjustment for age andsite, there was a strong graded association between moredepressive symptoms and worse subjective sleep quality(P-trend < .001). There was also a significant associationbetween more depressive symptoms and greater subjectivedaytime sleepiness (ESS; P-trend < .001), although theabsolute differences in mean ESS scores were small, andmeans were in the normal range for all three groups. These

associations remained significant after adjustment formultiple potential confounders.

The odds of subjective sleep disturbances were greaterfor women with more depressive symptoms (Figure 1,Table 4). In age- and site-adjusted models, the odds ofendorsing subjective sleep disturbance were significantlygreater for the group with some depressive symptoms(odds ratio (OR) = 2.01, 95% CI = 1.69–2.40) and evengreater for the depressed group (OR = 3.36, 95%CI = 2.61–4.32) than for the normal group. The odds ofsubjective excessive daytime sleepiness were greater forwomen with some depressive symptoms (OR = 2.16, 95%CI = 1.67–2.79) and for those in the depressed group(OR = 1.90, 95% CI = 1.35–2.67) than for those in the

Table 1. Characteristics of Study Participants According to Level of Depressive Symptoms

Population Characteristics All Participants (N = 3,045)

Level of Depressive Symptoms (Geriatric Depression Scale

Score)

Normal (0–2)Some Depressive

Symptoms (3–5) Depressed (�6) P-value

Age, n (%)70–80 527 (17.3) 390 (19.8) 96 (13.4) 41 (11.4) <.00181–82 853 (28.0) 585 (29.8) 192 (26.7) 76 (21.1)83–85 833 (27.4) 535 (27.2) 192 (26.7) 106 (29.4)86–100 832 (27.3) 456 (23.2) 238 (33.1) 138 (38.2)

African American, n (%) 325 (10.7) 231 (11.7) 60 (8.4) 34 (9.4) .03Self-reported health status, n (%)Excellent or good 2,295 (75.4) 1,690 (86.0) 452 (63.0) 153 (42.4) <.001Fair, poor, or very poor 750 (24.6) 276 (14.0) 266 (37.0) 208 (57.6)

Lives alone, n (%) 1,840 (60.4) 1,167 (59.4) 443 (61.7) 230 (63.7) .22Education, n (%)

�High school diploma 547 (18.0) 325 (16.5) 133 (18.6) 89 (24.7) <.001College or graduate school 1,221 (40.1) 841 (42.8) 268 (37.4) 112 (31.1)

Alcoholic drinks per week, n (%)0–2 2,647 (86.9) 1,682 (85.6) 636 (88.6) 329 (91.1) .023–13 341 (11.2) 243 (12.4) 72 (10.0) 26 (7.2)>13 57 (1.9) 41 (2.1) 10 (1.4) 6 (1.7)

Smoking status, n (%)Never 1,935 (63.6) 1,279 (65.1) 447 (62.3) 209 (57.9) .002Former 1,024 (33.6) 644 (32.8) 248 (34.6) 132 (36.6)Current 85 (2.8) 43 (2.2) 22 (3.1) 20 (5.5)

Caffeine intake, mg/d, mean ± SD 151.1 ± 154.1 152.5 ± 156.1 147.7 ± 151.1 150.2 ± 149.4 .83Current antidepressant use, n (%) 417 (13.7) 181 (9.2) 138 (19.3) 98 (27.1) <.001Current benzodiazepine use, n (%) 222 (7.3) 110 (5.6) 59 (8.2) 53 (14.7) <.001Current nonbenzodiazepine anxiolytic orhypnotic use, n (%)

33 (1.1) 16 (0.8) 10 (1.4) 7 (1.9) .11

Reported use of sleep medication, n (%) 427 (14.0) 230 (11.7) 114 (15.9) 83 (23.0) <.001Cognitively impaired (Mini-Mental StateExamination score < 24), n (%)

163 (5.7) 87 (4.6) 45 (6.9) 31 (9.8) .004

Body mass index, kg/m2, n (%)<25.0 (underweight or normal weight) 1,108 (37.2) 691 (35.5) 269 (38.8) 148 (43.8) .00225.0–29.9 (overweight) 1,131 (38.0) 780 (40.0) 253 (36.5) 98 (29.0)� 30.0 (obese) 740 (24.8) 477 (24.5) 171 (24.7) 92 (27.2)

Walks for exercise, n (%) 1,119 (37.2) 831 (42.7) 203 (28.6) 85 (23.9) <.001Instrumental activity of daily livingimpairments, n (%)

1,609 (53.1) 814 (41.6) 500 (70.0) 295 (82.2) <.001

Number of selected medical conditions,mean ± SDa

2.46 ± 1.61 2.20 ± 1.49 2.85 ± 1.70 3.09 ± 1.72 <.001

a Stroke, diabetes mellitus, Parkinson’s disease, chronic obstructive pulmonary disease, congestive heart failure, myocardial infarction, thyroid disease,

hypertension, other neurological condition, other cardiac condition, osteoarthritis, rheumatoid arthritis, osteoporosis, and cancer.

SD = standard deviation.

638 MAGLIONE ET AL. APRIL 2012–VOL. 60, NO. 4 JAGS

Page 5: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

normal group. These associations remained statistically sig-nificant after adjusting for multiple potential confounders.

Level of Depressive Symptoms and Objective SleepMeasures

After adjustment for age and site, there were graded asso-ciations between greater levels of depressive symptoms andpoorer SE (P-trend < .001), longer SOL (P-trend = .003),greater WASO (P-trend < .001), and more long wake epi-sodes (Table 3; P-trend < .001). After adjustment for mul-tiple variables, graded associations between greater level ofdepressive symptoms and greater WASO and number oflong wake episodes continued to be significant. There wereno associations between level of depressive symptoms andshort or long TST.

In models adjusted for age and site, the odds of someobjectively measured sleep disturbances increased withlevel of depressive symptoms (Figure 1, Table 4). The oddsof having SE less than 85% were greater for women withsome depressive symptoms (OR = 1.38, 95% CI = 1.13–1.68) and for the depressed group (OR = 1.53, 95%CI = 1.17–2.01) than for those in the normal group. Simi-larly, the odds of having WASO of 1 hour or longer weregreater for women with some depressive symptoms(OR = 1.45, 95% CI = 1.22–1.73) and for the depressedgroup (OR = 1.73, 95% CI = 1.36–2.19). The odds ofhaving eight or more long wake episodes were also greaterfor women with some depressive symptoms (OR = 1.55,95% CI = 1.29–1.86) and for the depressed group(OR = 1.84, 95% CI = 1.45–2.33). The odds of being along sleeper (>8 hours of sleep per night) as opposed to anormal sleeper (5–8 hours sleep per night) were greaterfor women with some depressive symptoms (OR = 1.46,1.15–1.86) and for the depressed group (OR = 1.58,1.17–2.15). The odds of being a short sleeper (<5 hourssleep per night) as opposed to a normal sleeper weregreater for women with some depressive symptoms(OR = 1.82, 95% CI = 1.36–2.43), whereas there was noassociation for the depressed group (OR = 1.30, 95%CI = 0.86–1.96). The odds of sleep latency of 30 minutesor more was not significantly greater for women withsome depressive symptoms (OR = 1.12, 95% CI = 0.94–1.33) but was slightly greater for those in the depressedgroup (OR = 1.38, 1.09–1.17). When models wereadjusted for multiple possible confounders, there continuedto be significant graded associations between level ofdepressive symptoms and WASO of 1 hour or longer andeight or more long wake episodes only (Figure 1, Table 4).

Level of Depressive Symptoms and Daytime Napping

There was a significant graded association between level ofdepressive symptoms and number of objectively assesseddaytime naps (Table 3). After adjustment for age and site,women in the depressed group had an average of 3.98(95% CI = 3.66–4.33) nap episodes exceeding 5 minutesin length, compared with 3.57 (95% CI = 3.36–3.79) forwomen with some depressive symptoms, and 2.82 (95%CI = 2.71–2.93) for those classified as normal (P-trend < .001). The odds of having 5.5 or more daytime

Table 2. Comparison of Subjective Sleep Parameters According to Level of Depressive Symptoms

Subjective Sleep Parameter

Level of Depressive Symptoms (Geriatric Depression Scale Score)

Normal (0–2)Some Depressive Symptoms (3–5)Mean (95% Confidence Interval) Depressed (� 6) P-trend

Pittsburgh Sleep Quality Index (0–21)Age and site adjusted (n=3,045) 5.6 (5.4–5.8) 7.1 (6.8–7.3) 8.6 (8.2–8.9) <.001Multivariable adjusted (n=2,785) 5.8 (5.7–6.0) 6.9 (6.6–7.1) 8.1 (7.7–8.5) <.001

Epworth Sleepiness Scale (0–24)Age and site adjusted (n=3,045) 5.3 (5.1–5.5) 6.4 (6.1–6.7) 6.1 (5.8–6.5) <.001Multivariable adjusted (n=2,785) 5.3 (5.2–5.5) 6.3 (6.0–6.5) 6.2 (5.7–6.6) <.001

Multivariable models adjusted for age, race, site, education, exercise, body mass index, smoking status, alcohol intake, self-reported health status, instru-

mental activity of daily living impairments, cognitive impairment, number of reported medical conditions, and current use of antidepressants, benzodiaze-

pines, and medications for sleep.

p-trend< 0.001

p-trend< 0.001

p-trend= 0.020

p-trend= 0.021

Association between Level of Depressive Symptoms and Subjectiveand Objective Sleep Disturbances

0

1

2

3

4

5Normal (referent)Some depressive symptomsDepressed

PSQ

I > 5

ESS

> 10

SE >

85%

SOL≥

30 m

in

WA

SO≥

1 hr

≥ 8 L

ong

Wak

e E

piso

des

TST

< 5

hr

TST

> 8

hr

Subjective Measures Objective Measures

Num

ber N

aps≥

5.5

p-trend< 0.001

Odd

s of

sle

ep d

istu

rban

ce

Figure 1. Odds ratios and 95% confidence intervals (95% CIs)for subjective and objective sleep disturbances are shown forgroups of women with differing levels of depressive symptomscompared with the normal (referent) group. Models repre-sented in the graph are adjusted for multiple possible confound-ers, including age, race, site, education, exercise, body massindex, smoking status, alcohol intake, self-reported health sta-tus, instrumental activities of daily living impairments, cogni-tive impairment, and current use of antidepressants,benzodiazepines, and medications for sleep. PSQI = PittsburghSleep Quality Index; ESS = Epworth Sleepiness Scale; SE =Sleep efficiency; SOL = sleep onset latency; WASO = timeawake after sleep onset; TST = total sleep time.

JAGS APRIL 2012–VOL. 60, NO. 4 DEPRESSIVE SYMPTOMS AND SLEEP IN OLDER WOMEN 639

Page 6: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

nap episodes were greater for women with some depressivesymptoms (OR = 1.71, 95% CI = 1.40–2.08) and forthose in the depressed group (OR = 2.57, 95% CI = 2.02–3.27) than for those in the normal group (Figure 1,Table 4). These associations remained statistically signifi-cant after adjusting for multiple variables.

Additional Analyses

To determine whether greater fragmentation of sleepaccounted for the strong associations observed betweenmore depressive symptoms and self-reported sleep out-comes including sleep quality (PSQI) and daytime sleepi-ness (ESS), WASO and number of long wake episodeswere added separately to the multivariable models withPSQI (and then ESS) as dependent variables. After furtheradjusting for WASO or number of long wake episodes, theassociations between more depressive symptoms and PSQIand ESS remained unchanged (data not shown). To furtherassess the effect of current use of antidepressants, ben-zodiazepines, and medications for sleep on these analyses,the analyses were repeated excluding 594 participants whoreported use of these medications, and the results weresimilar (data not shown).

DISCUSSION

These results suggest a strong, graded association betweenmore depressive symptoms and poorer self-reported sleepquality in community-dwelling older women, a finding thatis consistent with previously published reports.10,17,30–33

Significant associations were also found between moredepressive symptoms and objective measures of sleepfragmentation (greater WASO and number of long wake

episodes). A similar analysis of actigraphic sleep recentlyconducted in a cohort of community-dwelling older menfound a modest association between more depressive symp-toms and longer sleep latency but no significant associa-tions between level of depressive symptoms and objectivemeasures of sleep fragmentation.17 It is possible that thisdifference is related to the populations being tested becausethe study population assessed in that study was male andyounger (mean age 76.4 ± 5.5) than the women in theSOF.

After adjusting for multiple potential confounders, theobjective assessment did not show any association betweennumber of depressive symptoms and TST, SE, or SOL inthe group in the current study. This is in contrast to otherstudies that have found objective evidence of shorter TST,poorer SE, and longer SOL in individuals with depressionaccording to subjective assessment 34 and PSG.35,36 Thisdifference may be related to the method of sleep distur-bance assessment. Furthermore, those studies focused onindividuals who had a diagnosis of major depressive disor-der or met DSM-IV criteria for a major depressive episode,whereas the current analysis focused on level of subjec-tively reported depressive symptoms. Alternatively, it maybe that these specific sleep disturbances are less prominentin elderly women with depression than in other groups ofindividuals with depression. A recent study comparingactigraphic sleep measures in elderly men and womenfound elderly men to have shorter and more fragmentedsleep than elderly women.37 To the knowledge of theauthors of the current study, sex differences have not beenexamined in depressed older adults.

Objective sleep disturbances accounted for little of theassociation between more depressive symptoms and greatersubjective sleep disturbance in this population. There are

Table 3. Comparison of Objective Sleep Variables According Level of Depressive Symptoms

Objective Sleep Parameter

Level of Depressive Symptoms (Geriatric Depression Scale Score)

Normal (0–2)Some Depressive Symptoms (3–5)Mean (95% Confidence Interval) Depressed (�6) P-trend

SE (%)Age and site adjusted (n=3,045) 80.59 (80.16–81.01) 78.42 (77.62–79.19) 78.20 (77.04–79.29) <.001Multivariable adjusted (n=2,785) 80.44 (79.99–80.87) 79.62 (78.83–80.38) 79.76 (78.57–80.89) .12

SOL (minutes)Age and site adjusted (n=3,044) 28.16 (27.09–29.27) 30.24 (28.36–32.24) 31.89 (29.13–34.92) .003Multivariable adjusted (n=2,785) 28.39 (27.28–29.54) 28.11 (26.26–30.09) 29.60 (26.72–32.78) .63

WASO (minutes)Age and site adjusted (n=3,044) 61.68 (60.04–63.36) 71.02 (67.92–74.25) 74.67 (70.11–79.52) <.001Multivariable adjusted (n=2,785) 62.42 (60.72–64.18) 65.83 (62.78–69.02) 67.04 (62.44–71.99) .03

Number of long wake episodes (>5 minutes)Age and site adjusted (n=3,045) 6.53 (6.39–6.67) 7.26 (7.03–7.50) 7.58 (7.25–7.91) <.001Multivariable adjusted (n=2,785) 6.58 (6.44–6.72) 6.90 (6.65–7.14) 7.07 (6.70–7.44) .006Total sleep time (hours)Age and site adjusted (n=3,045) 6.75 (6.69–6.81) 6.71 (6.62–6.81) 6.87 (6.74–7.01) .28Multivariable adjusted (n=2,785) 6.75 (6.69–6.80) 6.74 (6.65–6.84) 6.87 (6.72–7.02) .23

Number of nap episodes (> 5 minutes)Age and site adjusted (n=2,995) 2.82 (2.71–2.93) 3.57 (3.36–3.79) 3.98 (3.66–4.33) <.001Multivariable adjusted (n=2,742) 2.84 (2.73–2.96) 3.43 (3.21–3.66) 3.60 (3.26–3.97) <.001

Multivariable models adjusted for age, race, site, education, exercise, body mass index, smoking status, alcohol intake, self-reported health status, instru-

mental activity of daily living impairment, cognitive impairment, number of reported medical conditions, and current use of antidepressants, benzodiaze-

pines, and medications for sleep. Data transformations: LN (P) for wake after sleep onset (WASO) and sleep onset latency (SOL), LN (100 � P) for sleep

efficiency (SE), and LN (P + 1) for number of nap episodes > 5 minutes.

640 MAGLIONE ET AL. APRIL 2012–VOL. 60, NO. 4 JAGS

Page 7: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

several possible explanations for this observation. Addi-tional factors that were not assessed in this analysis mayalso play a role in the greater subjective perception of poorsleep quality. It is also possible that, in older community-dwelling women, women with and without depressivesymptoms experience and report their sleep problems dif-ferently. An analysis in a cohort of community-dwellingolder men found a similar discrepancy between the associ-ation between depressive symptoms and subjective vsobjective sleep disturbances.17 In particular, older menwith depressive symptoms were much more likely to reportpoor sleep quality than their counterparts without depres-sion, whereas depression was more modestly associatedwith objectively measured sleep disturbance. To theknowledge of the authors of the current study, no otherstudies have evaluated the association between depressivesymptoms and both subjectively and objectively assessedsleep disturbances specifically in older adults.

Several studies comparing subjective reports of sleepquality using the PSQI with objective sleep measurementsusing PSG in healthy older adults have found that subjec-tive reports of sleep quality were not as poor as might beexpected considering the objective sleep disturbances thatwere observed using PSG. This suggests that healthy older

adults may underreport their sleep disturbances or adapttheir perception of acceptable sleep quality.11,38 The datapresented here suggest that older women with depressivesymptoms do not share this tendency to underreport theirproblems with sleep, and it has been previously shown thatin individuals with depression, older adults are more likelyto underestimate their total sleep than younger adultswhen subjective assessments are compared with PSG.39

In this population, women with depressive symptomswere almost twice as likely to report excessive subjectivedaytime sleepiness as those in the normal category. Severalstudies in younger populations have shown a similarrelationship between depression and excessive daytimesleepiness.40,41 Objective measurements also revealed anassociation between more depressive symptoms and daytimeinactivity, raising the possibility that women with moredepressive symptoms may nap more. Although actigraphycannot distinguish definitively between quiet restfulness andsleep, previous studies have found similar associationsbetween depressive symptoms and self-reported napping.42

Another study recently showed an association betweenobjectively measured greater fragmentation of sleep andmore subjective daytime napping in older adults.43 Studiesemploying PSG will be required to clarify this relationship.

Table 4. Association Between Depressive Symptoms and Subjective and Objective Sleep Disturbances

Sleep disturbance

Level of Depressive Symptoms (Reference Normal)

P-value

Some Depressive Symptoms (GDS 3–5) Depressed (GDS � 6)

Odds Ratio (95% Confidence Interval)

Subjective sleep disturbancePittsburgh Sleep Quality Index >5

Age and site adjusted (n=3,044) 2.01 (1.69–2.40) 3.36 (2.61–4.32) <.001Multivariable adjusted (n=2,985) 1.82 (1.48–2.24) 2.84 (2.08–3.86) <.001

Epworth Sleepiness Scale >10Age and site adjusted (n=3,044) 2.16 (1.67–2.79) 1.90 (1.35–2.67) <.001Multivariable adjusted (n=2,985) 1.97 (1.47–2.64) 1.70 (1.12–2.58) <.001

Objective sleep disturbanceSleep efficiency<85%

Age and site adjusted (n=3,044) 1.38 (1.13–1.68) 1.53 (1.17–2.01) <.001Multivariable adjusted (n=2,985) 1.18 (0.94–1.47) 1.27 (0.92–1.74) .08

Sleep onset latency� 30minutesAge and site adjusted (n=3,044) 1.12 (0.94–1.33) 1.38 (1.09–1.73) .005Multivariable adjusted (n=2,985) 0.95 (0.78–1.15) 1.21 (0.92–1.59) .39

Wake after sleep onset� 1hourAge and site adjusted (n=3,044) 1.45 (1.22–1.73) 1.73 (1.36–2.19) <.001Multivariable adjusted (n=2,985) 1.16 (0.95–1.42) 1.37 (1.03–1.83) .02

� 8 long wake episodesAge and site adjusted (n=3,044) 1.55 (1.29–1.86) 1.84 (1.45–2.33) <.001Multivariable adjusted (n=2,985) 1.25 (1.01–1.55) 1.34 (1.00–1.79) .02

Total sleep time<5hours (short sleeper)Age and site adjusted (n=3,044) 1.82 (1.36–2.43) 1.30 (0.86–1.96) .006Multivariable adjusted (n=2,985) 1.39 (0.99–1.95) 0.88 (0.50–1.37) .88

Total sleep time >8hours (long sleeper)Age and site adjusted (n=3,044) 1.46 (1.15–1.86) 1.58 (1.17–2.15) <.001Multivariable adjusted (n=2,985) 1.26 (0.96–1.67) 1.15 (0.79–1.68) .22

� 5.5 nap episodesAge and site adjusted (n=2,995) 1.71 (1.40–2.08) 2.57 (2.02–3.27) <.001Multivariable adjusted (n=2,742) 1.49 (1.19–1.86) 2.04 (1.52–2.74) <.001

Multivariable models adjusted for age, race, site, education, exercise, body mass index, smoking status, alcohol intake, self-reported health status, instru-

mental activity of daily living impairments, cognitive impairment, number of reported medical conditions, and current use of antidepressants, benzodiaze-

pines, and medications for sleep.

GDS = Geriatric Depression Scale.

JAGS APRIL 2012–VOL. 60, NO. 4 DEPRESSIVE SYMPTOMS AND SLEEP IN OLDER WOMEN 641

Page 8: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

The estimated point prevalence of major depressivedisorder in older adults has been reported to range from3% to 10%,44–47 whereas “less than major” depressivesyndromes (including dysthymia, minor depression, andsubsyndromal depression) are nearly twice as common,with a reported prevalence of 9–24%.46,48–51 The womenin the SOF did not undergo any formal psychiatric evalua-tion to determine diagnosis, but the prevalence of differinglevels of depressive symptoms based on GDS score in thiscohort suggests a similar pattern. Given that subsyndromaldepressive syndromes make up the majority of depressivesyndromes in older adults, it is interesting that having somedepressive symptoms conferred a similar greater risk forsubjectively assessed poor sleep quality and excessive day-time sleepiness as being in the depressed group. Althoughthere were significant trends suggesting graded associationsbetween more depressive symptoms and objective measuresof sleep fragmentation in multivariable-adjusted models,specific associations between having some depressive symp-toms and objective measures of sleep fragmentationobserved in age- and site-adjusted models were no longersignificant in multivariable-adjusted models. Future studiesdesigned to better explore the relationship between subsyn-dromal depression and objective sleep disturbances wouldbe useful to investigate this relationship further.

These analyses add to the growing body of literaturesuggesting that subjective sleep disturbance and excessivedaytime sleepiness are common in older adults. More thanhalf of the women in this study endorsed subjective poorsleep quality (PSQI > 5), and more than 10% endorsedexcessive daytime sleepiness. This is similar to what hasbeen reported in other populations, including the NationalSleep Foundation’s Sleep in America survey, in which morethan 50% of respondents reported a sleep complaintoccurring nearly every night. Similarly, in the OsteoporoticFractures in Men study, the prevalence of subjective poorsleep (PSQI > 5) in a large cohort of community-dwellingolder men was reported to be 44.2%, and the prevalenceof excessive daytime sleepiness (ESS > 10) was 12.7%.Subtle differences in the age and other covariates (e.g.,burden of medical illness) of these different samples mayexplain small differences in the prevalence rates of thesesubjectively reported problems.

This study has several strengths. The sample size waslarge and was collected from four separate U.S. locations,and the women were not selected based on depressivesymptoms or on their report of sleep disturbances. Unlikemost previous studies examining the relationship betweendepression and sleep disturbances in older adults, subjec-tive and objective measures of sleep disturbance and day-time sleep were assessed.

There are also several limitations to this analysis. Thestudy was cross-sectional, and it was therefore not possibleto draw any conclusions about causality. The sample con-sisted only of community-dwelling women, the vast major-ity of whom were white and aged 80 and older, and thefindings may not be applicable to other populations.Depressive symptoms were assessed by questionnaire onlyrather than a clinical diagnostic interview, so conclusionsabout psychiatric diagnosis cannot be made. Actigraphyrelies on self-reported times in and out of bed, which maybe inaccurate and could introduce errors into measure-

ments such as SOL, but these errors would not affectWASO and number of long wake episodes at night, whichwere found to be associated with depressive symptoms.In addition, actigraphy cannot definitively distinguishbetween daytime sleep and periods of quiet restfulness.Finally, the effect of adding multiple covariates to many ofthe multivariable-adjusted models underscores the com-plexity of the relationship between these covariates, sleep,and depression in this population. The analyses were notdesigned to explore fully the role that these covariates playin the relationship between depressive symptoms and sleepdisturbances. Future studies designed specifically to addressthis would be useful, particularly because some of thesecovariates may be amenable to intervention.

In conclusion, these analyses add to the body of litera-ture supporting an association between depressive symp-toms and subjective sleep disturbance in older adults.These data also suggest an association between moredepressive symptoms and fragmentation of sleep, as wellas more daytime napping in older women. Longitudinalanalyses will be required to determine whether objectivelymeasured sleep disturbances predict changes in depressionover time.

ACKNOWLEDGMENTS

The authors would like to thank the SOF Research Group.We also thank Loki Natarajan, PhD, and Feng He, MS,for thoughtful comments regarding the statistical analysisused in this study. A portion of the data in this manuscriptwas presented at the Associated Professional Sleep Socie-ties annual meeting, SLEEP 2010, San Antonio, Texas,June 6, 2010.

Conflicts of Interest: Dr. Ancoli-Israel has been a con-sultant to Ferring Pharmaceuticals Inc., GlaxoSmithKline,Merck, NeuroVigil, Inc., Pfizer, Philips Respironics, andPurdue Pharma LP. The other authors have indicated nofinancial conflicts of interest.

Dr. Maglione was supported by National Institutes ofHealth (NIH) Grant R25MH7450 and a Mini-Grant inAging provided by the University of California at SanDiego Academic Geriatric Resource Center (09SD-A7–1-24). Dr. Ancoli-Israel is supported by National Institute onAging Grant AG08415. The SOF is also supported byNIH funding under Grants AG05407, AR35582,AG05394, AR35584, AR35583, R01 AG005407, R01AG027576–22, 2 R01 AG005394–22A1, 2 R01AG027574–22A1, AG05407, AR35582, AG05394,AR35584, AR35583, AG026720, and R01 MH086498.

Author Contributions: Maglione JE: Analysis and inter-pretation of data, preparation of manuscript. Ancoli-Israel S,Ensrud KE, Stone KL: Study concept and design, interpreta-tion of data, critical review of manuscript. Peters KW:Analysis and interpretation of data, critical review of manu-script. Paudel ML: Critical review of manuscript. Yaffe K:Study concept and design, critical review of manuscript.

Sponsor’s Role: None.

REFERENCES

1. Blazer DG. Depression in late life: Review and commentary. J Gerontol A

Biol Sci Med Sci 2003;58A:249–265.

642 MAGLIONE ET AL. APRIL 2012–VOL. 60, NO. 4 JAGS

Page 9: Depressive Symptoms and Subjective and Objective Sleep in Community-Dwelling Older Women

2. Hays JC, Krishnan KR, George LK et al. Psychosocial and physical corre-

lates of chronic depression. Psychiatry Res 1997;72:149–159.3. Bruce ML. Depression and disability in late life: Directions for future

research. Am J Geriatr Psychiatry 2001;9:102–112.4. Frasure-Smith N, Lesperance F. Depression—a cardiac risk factor in search

of a treatment. JAMA 2003;289:3171–3173.5. Alexopoulos GS, Vrontou C, Kakuma T et al. Disability in geriatric depres-

sion. Am J Psychiatry 1996;153:877–885.6. Zeiss AM, Lewinsohn PM, Rohde P et al. Relationship of physical disease

and functional impairment to depression in older people. Psychol Aging

1996;11:572–581.7. Huang BY, Cornoni-Huntley J, Hays JC et al. Impact of depressive symp-

toms on hospitalization risk in community-dwelling older persons. J Am

Geriatr Soc 2000;48:1279–1284.8. Unutzer J, Patrick DL, Simon G et al. Depressive symptoms and the cost of

health services in HMO patients aged 65 years and older. A 4-year pro-

spective study. JAMA 1997;277:1618–1623.9. Luber MP, Meyers BS, Williams-Russo PG et al. Depression and service

utilization in elderly primary care patients. Am J Geriatr Psychiatry

2001;9:169–176.10. Foley DJ, Monjan AA, Brown SL et al. Sleep complaints among elderly

persons: An epidemiologic study of three communities. Sleep 1995;18:425–432.

11. Vitiello MV, Larsen LH, Moe KE. Age-related sleep change: Gender and

estrogen effects on the subjective-objective sleep quality relationships of

healthy, noncomplaining older men and women. J Psychosom Res

2004;56:503–510.12. Franzen PL, Buysse DJ. Sleep disturbances and depression: Risk relation-

ships for subsequent depression and therapeutic implications. Dialogues

Clin Neurosci 2008;10:473–481.13. Turvey CL, Conwell Y, Jones MP et al. Risk factors for late-life suicide: A

prospective, community-based study. Am J Geriatr Psychiatry 2002;10:

398–406.14. Livingston G, Blizard B, Mann A. Does sleep disturbance predict depres-

sion in elderly people? A study in inner London Br J Gen Pract

1993;43:445–448.15. Perlis ML, Smith LJ, Lyness JM et al. Insomnia as a risk factor for onset of

depression in the elderly. Behav Sleep Med 2006;4:104–113.16. Roberts RE, Shema SJ, Kaplan GA et al. Sleep complaints and depression

in an aging cohort: A prospective perspective. Am J Psychiatry

2000;157:81–88.17. Paudel ML, Taylor BC, Diem SJ et al. Association between depressive

symptoms and sleep disturbances in community-dwelling older men. J Am

Geriatr Soc 2008;56:1228–1235.18. Cummings SR, Black DM, Nevitt MC et al. Bone density at various sites

for prediction of hip fractures. The Study of Osteoporotic Fractures

Research Group. Lancet 1993;341:72–75.19. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical

method for grading the cognitive state of patients for the clinician. J Psy-

chiatr Res 1975;12:189–198.20. Brink TL. Clinical Gerontology: A Guide to Assessment and Intervention.

New York: Howarth Press, 1986.

21. Almeida OP, Almeida SA. Short versions of the geriatric depression scale:

A study of their validity for the diagnosis of a major depressive episode

according to ICD-10 and DSM-IV. Int J Geriatr Psychiatry 1999;14:858–865.

22. Buysse DJ, Reynolds CF III, Monk TH et al. The Pittsburgh Sleep Quality

Index: A new instrument for psychiatric practice and research. Psychiatry

Res 1989;28:193–213.23. Johns MW. A new method for measuring daytime sleepiness: The Epworth

Sleepiness Scale. Sleep 1991;14:540–545.24. Johns MW. Reliability and factor analysis of the Epworth Sleepiness Scale.

Sleep 1992;15:376–381.25. Blackwell T, Ancoli-Israel S, Gehrman PR et al. Actigraphy scoring reliabil-

ity in the Study of Osteoporotic Fractures. Sleep 2005;28:1599–1605.26. Blackwell T, Redline S, Ancoli-Israel S et al. Comparison of sleep parame-

ters from actigraphy and polysomnography in older women: The SOF

Study. Sleep 2008;31:283–291.27. Ancoli-Israel S, Cole R, Alessi C et al. The role of actigraphy in the study

of sleep and circadian rhythms. Sleep 2003;26:342–392.

28. Cole RJ, Kripke DF, Gruen W et al. Automatic sleep/wake identification

from wrist activity. Sleep 1992;15:461–469.29. Jean-Louis G, Kripke DF, Mason WJ et al. Sleep estimation from wrist

movement quantified by different actigraphic modalities. J Neurosci Meth-

ods 2001;105:185–191.30. Foley DJ, Monjan A, Simonsick EM et al. Incidence and remission of

insomnia among elderly adults: An epidemiologic study of 6,800 persons

over three years. Sleep 1999;22(Suppl 2):S366–S372.31. Foley D, Ancoli-Israel S, Britz P et al. Sleep disturbances and chronic dis-

ease in older adults: Results of the 2003 National Sleep Foundation Sleep

in America Survey. J Psychosom Res 2004;56:497–502.32. Newman AB, Enright PL, Manolio TA et al. Sleep disturbance, psychoso-

cial correlates, and cardiovascular disease in 5201 older adults: The Car-

diovascular Health Study. J Am Geriatr Soc 1997;45:1–7.33. Reid KJ, Martinovich Z, Finkel S et al. Sleep: A marker of physical and

mental health in the elderly. Am J Geriatr Psychiatry 2006;14:860–866.34. van den Berg JF, Luijendijk HJ, Tulen JH et al. Sleep in depression and

anxiety disorders: A population-based study of elderly persons. J Clin Psy-

chiatry 2009;70:1105–1113.35. Benca RM, Obermeyer WH, Thisted RA et al. Sleep and psychiatric disor-

ders. A meta-analysis. Arch Gen Psychiatry 1992;49:651–668; discussion

669–670.36. Benca RM, Peterson MJ. Insomnia and depression. Sleep Med 2008;9(Sup-

pl 1):S3–S9.37. van den Berg JF, Miedema HM, Tulen JH et al. Sex differences in subjec-

tive and actigraphic sleep measures: A population-based study of elderly

persons. Sleep 2009;32:1367–1375.38. Buysse DJ, Reynolds CF III, Monk TH et al. Quantification of subjective

sleep quality in healthy elderly men and women using the Pittsburgh Sleep

Quality Index (PSQI). Sleep 1991;14:331–338.39. Tsuchiyama K, Nagayama H, Kudo K et al. Discrepancy between subjec-

tive and objective sleep in patients with depression. Psychiatry Clin Neuro-

sci 2003;57:259–264.40. Pallesen S, Nordhus IH, Omvik S et al. Prevalence and risk factors of sub-

jective sleepiness in the general adult population. Sleep 2007;30:619–624.41. Chellappa SL, Araujo JF. Excessive daytime sleepiness in patients with

depressive disorder. Rev Bras Psiquiatr 2006;28:126–129.42. Foley DJ, Vitiello MV, Bliwise DL et al. Frequent napping is associated with

excessive daytime sleepiness, depression, pain, and nocturia in older adults:

Findings from the National Sleep Foundation “2003 Sleep in America” Poll.

Am J Geriatr Psychiatry 2007;15:344–350.43. Goldman SE, Hall M, Boudreau R et al. Association between nighttime

sleep and napping in older adults. Sleep 2008;31:733–740.44. Borson S, Barnes RA, Kukull WA et al. Symptomatic depression in elderly

medical outpatients. I. Prevalence, demography, and health service utiliza-

tion. I. Prevalence, demography, and health service utilization. J Am Geri-

atr Soc 1986;34:341–347.45. Schulberg HC, Mulsant B, Schulz R et al. Characteristics and course of

major depression in older primary care patients. Int J Psychiatry Med

1998;28:421–436.46. Lyness JM, King DA, Cox C et al. The importance of subsyndromal

depression in older primary care patients: Prevalence and associated func-

tional disability. J Am Geriatr Soc 1999;47:647–652.47. Steffens DC, Skoog I, Norton MC et al. Prevalence of depression and its

treatment in an elderly population: The Cache County study. Arch Gen

Psychiatry 2000;57:601–607.48. Judd LL, Akiskal HS. The clinical and public health relevance of current

research on subthreshold depressive symptoms to elderly patients. Am J

Geriatr Psychiatry 2002;10:233–238.49. Judd LL, Akiskal HS, Maser JD et al. A prospective 12-year study of sub-

syndromal and syndromal depressive symptoms in unipolar major depres-

sive disorders. Arch Gen Psychiatry 1998;55:694–700.50. Judd LL, Akiskal HS, Maser JD et al. Major depressive disorder: A pro-

spective study of residual subthreshold depressive symptoms as predictor of

rapid relapse. J Affect Disord 1998;50:97–108.51. Olfson M, Broadhead WE, Weissman MM et al. Subthreshold psychiatric

symptoms in a primary care group practice. Arch Gen Psychiatry

1996;53:880–886.

JAGS APRIL 2012–VOL. 60, NO. 4 DEPRESSIVE SYMPTOMS AND SLEEP IN OLDER WOMEN 643