SLEEP IN OLDER WOMEN

Poor Sleep is Associated with Poorer Physical Performance and Greater Functional Limitations in Older Women
Suzanne E. Goldman, PhD1; Katie L. Stone, PhD2; Sonia Ancoli-Israel, PhD3; Terri Blackwell, MA2; Susan K. Ewing, MS4; Robert Boudreau, PhD1; Jane A. Cauley, DrPH1; Martica Hall, PhD5; Karen A. Matthews, PhD5; Anne B. Newman, MD, MPH1 Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA; 2San Francisco Coordinating Center and California Pacic Medical Center Research Institute, San Francisco, CA; 3Department of Psychiatry, University of California, San Diego, CA; 4Department of Epidemiology and Biostatistics, University of California, San Francisco, CA; 5Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA
1

Study Objectives: This study examined the association between disturbed sleep and poorer daytime function in older women. Design: Observational study. Participants: 2,889 women, mean age 83.5 years, participating in the 2002-2004 examination of the Study of Osteoporotic Fractures. Interventions: N/A Measurements and Results: Participants wore actigraphs for an average SD of 4.1 0.83 24-hour periods. Actigraphy measured sleep variables were total sleep time and hours awake after sleep onset during the night and daytime napping behavior. Neuromuscular performance measurements included gait speed, chair stands, and grip strength. Functional limitations were assessed as self-reported difculty with one or more of 6 instrumental activities of daily living (IADL). In fully adjusted, multivariable models, women who slept <6 hours per night walked 3.5% slower than those who slept 6.0-6.8 hours. Those who slept 7.5 hours took 4.1% lon-

ger to complete 5 chair stands than those who slept 6.8-7.5 hours. With higher wake after sleep onset (1.6 hours compared to <0.7 hours) gait speed was 9.1% slower; it took 7.6% longer to complete 5 chair stands, and odds of functional limitation were 1.8 (95% CI: 1.4, 2.4) higher. Women with 1.0-1.8 hours of daytime sleep had higher odds (1.4 [95% CI: 1.1, 1.8]) of a functional limitation than those with <0.5 hours. Sleep variables did not appear to be associated with grip strength. Conclusions: Objectively measured poorer sleep was associated with worse physical function. Future research is needed to identify the underlying mechanisms for the association between poor sleep and functional decline. Keywords: Actigraphy, sleep, performance, function, elderly Citation: Goldman SE; Stone KL; Ancoli-Israel S; Blackwell T; Ewing SK; Boudreau R; Cauley JA; Hall M; Matthews KA; Newman AB. Poor sleep is associated with poorer physical performance and greater functional limitations in older women. SLEEP 2007;30(10):1317-1324.

SLEEP PROBLEMS ARE COMMON IN OLDER ADULTS, WITH OVER HALF OF COMMUNITY DWELLING ADULTS REPORTING SOME CHRONIC SLEEP COMPLAINT, GENerally as a result of a decrease in the ability to sleep.1,2 Older adults frequently complain of problems with sleep initiation and maintenance and often exhibit decreased total sleep time, poorer sleep quality, and daytime sleepiness. These sleep complaints are often comorbid with medical illness3-5 and are associated with elevated risk of morbidity and mortality.6-10 Daytime function is an important factor in the health and wellbeing of older adults. Maintenance of function is important for
Disclosure Statement This was not an industry supported study. Dr. Ancoli-Israel has received research support from Sepracr and Takeda; has participated in speaking engagements for Cephalon, King Pharmaceuticals, Neurocrine Biosciences, Pzer, Sano-Aventis, Sepracor, and Takeda; and has consulted for Acadia, Cephalon, GlaxoSmithKline, King Pharmaceuticals, Merck, Neurocrine Biosciences, Neurogen, Pzer, Sano-Aventis, Sepracor and Takeda. Dr. Blackwell has received research support from Eli Lilly. Dr. Cauley has received research support from Merck, Eli Lilly, Pzer, and Novartis; has received honorarium from Merck and Eli Lilly and is on the Speakers Bureau fro Merck. Drs.Goldman, Stone, Boudreau, Hall, Matthews, Newman, and Ms.Ewing have reported no nancial conicts of interest. Submitted for publication October, 2006 Accepted for publication June, 2007 Address correspondence to: Katie L. Stone, PhD, California Pacic Medical Center Research Institute, 185 Berry Street, Lobby 4, 5th oor, Suite 5700, San Francisco, CA 94107-1762; Tel: (415) 600-7422; Fax: (415) 514-8150; E-mail: kstone@sfcc-cpmc.net
SLEEP, Vol. 30, No. 10, 2007 1317

quality of life in these individuals.11 Functional evaluation in the clinical setting can provide critical information on the health status of older adults.12 Frequently used measures of daytime function, including lower extremity function13 and hand grip strength,14 have been shown to be related to elevated risk of disability, morbidity, and mortality. Poor sleep has been associated with reduced physical function in older adults,10,15-17 although these studies all used subjective measures of sleep. Nevertheless, abnormal sleep patterns were reported by over 15% of older adults who also reported ambulatory limitations.14 Studies replicating these ndings with objective measures of sleep and physical functioning are needed. The present study employed data from a large cohort of older women from the multicenter Study of Osteoporotic Fractures, to examine the association of objective measures of sleep and nap patterns with objective measures of neuromuscular performance and subjective measures of daytime function. The hypothesis tested was that older women with short or long nighttime sleep duration, more disrupted sleep, and more daytime sleep would have poorer neuromuscular performance and more functional limitation. METHODS Participants Women were enrolled in The Study of Osteoporotic Fractures (SOF), a longitudinal study of risk factors for fracture. Participants in SOF were community-dwelling women, 65 years of age, and ambulatory at the initial visit 1986-88. Women who reported bilateral hip replacements were excluded. Women were recruited from population-based listings at 4 clinical centers in
Sleep and Daytime Performance in Older WomenGoldman et al

the United States: Baltimore, Maryland; Minneapolis, Minnesota; Portland, Oregon; and the Monongahela Valley near Pittsburgh, Pennsylvania. The initial study cohort consisted of 9704 Caucasian women recruited between 1986-198818 and an additional 662 African American women added in 1996-1998. After enrollment, follow-up clinic visits were conducted approximately every 2 years. A total of 4,727 women participated in an eighth study visit (year 16) between January 2002 and April 2004. At this visit 3,127 women had measurements for both actigraphy and daytime function. To focus on women living independently in the community, 165 women were excluded from this analysis because they resided in either a personal care facility or nursing home. An additional 73 participants who did not have actigraphy data collected in the proportional integrative mode (PIM) were also excluded. The nal analysis sample consisted of 2,889 women. The SOF study was approved by the institutional review board at each participating institution, and all participants provided written informed consent. Sleep Assessment Sleep and nap patterns were assessed objectively using wrist actigraphy (Sleep-Watch-O, Ambulatory Monitoring, Inc., Ardsley, NY). Actigraphs were worn on the non-dominant wrist for an average of four 24-hour periods (mean SD = 4.1 0.83 nights, range 1-9 nights). Details of the actigraphy collection and scoring have previously been reported.19,20 Briey, actigraphy data used in this analysis were collected in the proportional integration mode (PIM), also referred to as the digital integration mode. Sleep and wake cycles were differentiated using the University of California, San Diego (UCSD) sleep-scoring algorithm.21 This algorithm calculates a moving average taking into account the activity levels immediately before and after the current minute to determine if the time point should be coded as sleep or wake. Participants also completed a sleep diary for the duration of time they wore the actigraph, which was used in the editing of the actigraphy data to determine bed time and nal up time, reported naps, and periods of watch removal (e.g., for bathing). Actigraphy variables used in analysis of nighttime sleep were: total sleep time at night (TST) calculated as the hours scored as sleep between bedtime and nal up time; and wake after sleep onset (WASO), a measure of fragmented sleep, calculated as the average number of hours awake between sleep onset (where sleep onset was scored as the completion of 20 continuous minutes of sleep after getting into bed) and nal awakening. Total sleep during the day (i.e., number of hours of napping) was calculated as the average minutes of inactivity (minimal movement detected by the actigraphy) between the nal up time that morning and the subsequent bedtime. There was no minimum duration of inactivity set to qualify a period as a nap. All minutes scored by the scoring algorithm as sleeping during the out of bed period were considered as napping. Actigraphy data for each participant were averaged over all 24-hour periods to reduce night-to-night variability. Assessment of Neuromuscular Performance and Functional Limitations Neuromuscular performance was assessed by trained examiners at the eighth SOF clinic visit and included gait speed measured with the 6-meter usual pace test (meters/sec)22; ability to rise up
SLEEP, Vol. 30, No. 10, 2007 1318

from a chair (without using arms) 5 times22; time to complete 5 chair stands (seconds)23; and the average of right and left grip strength (kg).22,23 Functional limitations were assessed by self-report of difculty performing any of 6 instrumental activities of daily living (IADL) that included walking 2-3 blocks, climbing up or walking down 10 steps, preparing meals, heavy housework, and shopping.24, 25 Presence of a functional limitation was dened as having any difculty with at least one of the 6 IADLs. Other Measurements Information about potential confounders or mediators previously identied as associated with sleep in older adults15,17 or in the SOF cohort26 was collected at the eighth clinic visit, along with the self-reported sleep variables and actigraphy data. Information included medical history, anthropomorphic measurements of weight (kg) and height (m), and self-reported walking for exercise. Depressive symptoms were measured with the Geriatric Depression Scale (range 0-15), with higher scores indicating more depression. Anxiety was measured using the Goldberg Anxiety Scale Score, with higher scores indicating more anxiety. Mental status was measured with the Mini-Mental Status Exam (range 0-30), with higher scores representing better cognitive function.20 A comorbidity index (number of comorbid conditions coded as 0, 1, 2, 3+) was created to measure self-reported history of stroke, diabetes, Parkinson disease, Alzheimer disease, heart disease, congestive heart failure, COPD, and hypertension. Statistical Analysis In descriptive univariate analysis, characteristics associated with disturbed sleep, physical performance, and functional limitation measures in participants attending the eighth study visit were compared to women not included in this analysis. Characteristics were summarized as means SD for continuous variables and counts and percentages for categorical variables. Comparisons were performed using t-tests for normally distributed continuous data, Wilcoxon rank sum tests for skewed continuous data, and Pearsons chi-squared test for categorical data. Sleep parameters were categorized into quartiles to examine potential nonlinear, or U-shaped, relationships with performance and function measurements. The U-shaped distribution between TST and the neuromuscular performance measurements was further evaluated by entering a quadratic term for TST in regression models. Individual sleep variables were examined as the independent variable with gait speed, ability to complete 5 chair stands, chair stand time, grip strength, and functional limitations as the dependent variables. Linear regression models were used to examine the relationship between quartiles of the sleep variables and continuous performance outcomes, and adjusted means with 95% CIs were obtained using the least squared means procedure. Logistic regression models were used to examine the relationship between sleep variables and the dichotomous outcomes. All models were adjusted for potential confounders known to affect both physical function and sleep.26 These included age, race, BMI (body mass index (kg/m2)), comorbidity index, walking for exercise, depression, anxiety, and cognitive function. Although data from 2,889 women were included in analyses, due to missing data, sample sizes differed as noted in the Tables. Data were analyzed using STATA V8.1 (Stata Corporation, College
Sleep and Daytime Performance in Older WomenGoldman et al

Station, TX) and SAS V8 (SAS Institute, Cary, NC ) software. Twosided P-values < 0.05 were considered statistically signicant. RESULTS Comparison of SOF Women with Actigraphy vs. Women without Actigraphy The 2,889 women in the analysis sample were younger, had higher BMI, better self-reported health status, fewer comorbid health conditions, less depression, better cognitive function, and more self-reported sleep problems than the 1,838 women who completed some part of the eighth clinic visit (2002-2004) but who are not included in the analysis sample (Table 1). The women

in the analysis subset were more likely to report walking for exercise and had a faster gait speed; were more likely to be able to complete 5 chair stands, and among those who completed the chair stands they were able to do the test more quickly; and they had stronger average grip strength. Women in the analysis group also had fewer functional limitations. Sleep Measurements As shown in Table 2, on average, the women were asleep for 6.7 hours a night and had 1.3 hours of WASO. During the day, they napped for an average of 1.2 hours. As a comparison, the survey data for women from the 2003 National Sleep Foundation Poll of Sleep in the Elderly40 are shown in Figure 1.

Table 1Comparison of Demographic, Health, and Performance Variables in the SOF Cohort at the 8th Clinic Visit (2002 2004) Analysis Sample (n=2,889) 83.5 3.7 314 (10.9) 2,575 (89.1) 27.0 5.0 551 (19.1) 1,630 (56.5) 648 (22.5) 56 (1.9) 790 (27.4) 1,186 (41.1) 601 (20.8) 308 (10.7) 2.4 2.6 1.4 2.2 27.9 2.0 1,066 (37.4) 0.8 0.2 1,375 (47.5) 2,021 (83.7) 13.7 4.8 16.6 4.1 Remaining Sample (n=1,838)** 84.8 4.5 152 (8.3) 1,686 (91.7) 26.6 5.4 293 (16.1) 882 (48.6) 501 (27.6) 140 (7.7) 371 (21.9) 681 (36.5) 399 (23.5) 244 (14.4) 3.5 3.4 1.4 2.3 27.1 2.6 530 (31.8) 0.7 0.3 277 (15.1) 861 (69.7) 14.5 5.1 15.4 4.7 P-value < 0.001 0.003 0.03 <0.001

Age (years) (mean SD) Race: Black (n (%)) White (n (%)) BMI (kg/m2) (mean SD) Self-reported health status (n (%)) Excellent Good Fair Poor Comorbidity1 (n (%)) None One Two Three Depression (GDS15) (mean SD) Anxiety (mean SD) Mini Mental Status Exam (mean SD) Walks for exercise (n (%)) Gait Speed (m/sec) (mean SD) Functional limitation2 (n (%) with no impairment) Able to complete 5 chair stands (n (%)) Chair Stands Time3 (seconds) (mean SD) Grip Strength4 (kg) (mean SD)
1

<0.001

<0.001 0.57 <0.001 0.002 <0.001 <0.001 <0.001 0.003 <0.001

Comorbid conditions at the eighth clinic visit (stroke, diabetes, Parkinson disease, Alzheimer disease, COPD, heart disease, congestive heart failure, and hypertension). 2 Functional limitations: walking 2-3 blocks, climbing or walking down 10 steps, preparing meals, heavy housework, and shopping. 3 Seconds to complete 5 chair stands. 4 Average of right and left grip strength. ** Variables for cognitive function, body mass index, functional limitation, grip strength, chair stands, and gait speed were only measured in those participants who had clinic or home visits (n=744). Remainder of participants completed and returned a questionnaire only.

Table 2Sleep Measures1 in the Study of Osteoporotic Fractures Participants at the Eighth Clinic Visit (2002-2004) TST (hr) Wake After Sleep Onset (hr) Total Daytime Sleep (hr)
1

n 2,890 2,889 2,845

MEAN (SD) 6.7 (1.3) 1.3 (0.8) 1.2 (1.1)

MEDIAN 6.8 1.1 1.0

RANGE 0.5 12.0 0.1 6.5 0 8.0

Measured with wrist actigraphy (Sleep-Watch-O). TST=average hours of sleep in bed at night; Wake after sleep onset=average hours of wake after sleep onset; Daytime sleep = average hours of sleep, out of bed. SLEEP, Vol. 30, No. 10, 2007 1319 Sleep and Daytime Performance in Older WomenGoldman et al

Table 3Physical Performance Measures by Quartile of TST, Wake After Sleep Onset, and Daytime Sleep in the Study of Osteoporotic Fractures Participants at the Eighth Clinic Visit TST1 (Hours) Mean (95% CI) Gait Speed (m/sec) Unadjusted Adjusted2 Chair Stands Time (sec) Unadjusted Adjusted2 Grip Strength (kg) Unadjusted Adjusted2 n 2,687 2,537 2,120 2,014 2,675 2,516 n 2,686 2,537 2,120 2,014 2,674 2,516 n 2,652 1,981 2,093 1,988 2,637 2,481 < 6.0 0.81 (0.79, 0.82)a 0.83 (0.81, 0.84)a 14.05 (13.63, 14.47)a 13.90 (13.48, 14.32) 16.78 (16.47, 17.09) 16.63 (16.32, 16.93) <0.7 0.90 (0.88, 0.91)a 0.88 (0.86, 0.89)a 13.05 (12.66, 13.44)a 13.22 (12.84, 13.60)a 16.93 (16.62, 17.24)a 16.93 (16.63, 17.23)a < 0.5 0.88 (0.86, 0.90)a 0.88 (0.87, 0.90) 13.33 (12.94, 13.72)a 13.62 (13.23, 14.01) 16.87 (16.56, 17.18)a 16.62 (16.31, 16.92) 6.0 - 6.8 0.85 (0.83, 0.87)a, b 0.86 (0.84, 0.87)a 13.56 (13.15, 13.96) 13.49 (13.10, 13.89) 16.71 (16.41, 17.02) 16.71 (16.41, 17.01) 0.7 1.1 0.85 (0.83, 0.87)a, b 0.84 (0.83, 0.86)a, b 13.43 (13.04, 13.83)b 13.50 (13.11, 13.89)b 16.63 (16.32, 16.93) 16.70 (16.40, 16.99) 0.5 -1.0 0.85 (0.84, 0.87)a, b 0.89 (0.87, 0.91) 13.20 (12.80, 13.61)b 13.25 (12.85, 13.65)b 16.68 (16.37, 16.99) 16.75 (16.45, 17.05) 6.8 -7.5 0.86 (0.84, 0.87)a. c 0.85 (0.83, 0.86) 13.23 (12.82, 13.64) a. c 13.29 (12.89, 13.69)c 16.58 (16.27, 16.89) 16.67 (16.37, 16.97) 1.1 1.6 0.84 (0.82, 0.86)a. c 0.85 (0.83, 0.86)a. c 13.77 (13.36, 14.19)a. c 13.67 (13.27, 14.08)c 16.69 (16.38, 17.00) 16.81 (16.51, 17.11) 1.0 -1.8 0.83 (0.81, 0.84)a, ,b, c 0.88 (0.86, 0.89) 13.51 (13.10,13.92)c 13.44 (13.03, 13.85) 16.75 (16.44, 17.06)c 16.80 (16.51, 17.11) 7.5 0.83 (0.81, 0.85)b, c 0.84 (0.82, 0.85) 13.89 (13.48, 14.31)c 13.84 (13.43, 14.25)c 16.45 (16.13, 16.76) 16.72 (16.42, 17.03) 1.6 0.76 (0.74, 0.77)a, ,b, c 0.80 (0.78, 0.81)a, ,b, c 14.70 (14.25, 15.15)a, ,b,c 14.23 (13.83, 14.76)a, ,b,c 16.25 (15.93, 16.57)a 16.25 (15.93, 16.57)a 1.8 0.80 (0.78, 0.81)a, ,b, c 0.89 (0.87, 0.90) 14.68 (14.24, 15.11)a, ,b,c 14.10 (13.66, 14.54)b 16.29 (15.97, 16.61)a. c 16.64 (16.32, 16.95)

Wake After Sleep Onset1 (Hours) Gait Speed (m/sec) Unadjusted Adjusted2 Chair Stands Time (sec) Unadjusted Adjusted2 Grip Strength (kg) Unadjusted Adjusted2

Daytime Sleep1 (Hours) Gait Speed (m/sec) Unadjusted Adjusted2 Chair Stands Time (sec) Unadjusted Adjusted2 Grip Strength (kg) Unadjusted Adjusted2
1

Measured with wrist actigraphy (Sleep-Watch-O). TST=average hours sleep in bed at night; Wake after sleep onset=average hours of wake after sleep onset; Daytime sleep = average hours of sleep, out of bed. 2 Adjusted for age, race, BMI, depression (GDS15), anxiety (Goldberg Anxiety Score 0-9), cognitive function (MMSE), number of comorbidities (Doctor ever told you: stroke, diabetes, Parkinson disease, Alzheimer disease, chronic obstructive pulmonary disease, heart disease (heart attack or coronary event), congestive heart failure, and hypertension), and walking for exercise. a ,b, c Differences between quartile with same letter are statistically signicant P <0.05.

As TST, WASO, and daytime sleep may be related, correlations were run to determine whether these variables could be considered independent. The correlation between TST and WASO was -0.36 (P <0.001), between TST and daytime sleep was -0.08 (P <0.001) and between wake after sleep on WASO and daytime sleep was -0.002 (P = 0.89). Although the correlations between TST and WASO and TST and daytime sleepiness were signicant, they were weak. Therefore, we still considered each variable separately. Sleep and Neuromuscular Function Gait Speed

Wake After Sleep Onset Women with more WASO had slower gait speed (Table 3). In the fully adjusted multivariable model, women with over 1.6 hours of WASO walked almost 10% slower than women who had less than three-quarters of an hour of nighttime wakening.

SOF
35 30 Percent Women 25 20 15 10 5 0 <6 hours 6-6.9 hours 7-7.9 hours 13.5 25 17.5 10.5 31.5 30

NSF
30.5 27

Total Sleep Time In the unadjusted model for nighttime sleep duration and gait speed women who averaged <6 hours and 7.5 hours of TST performed more poorly on the 6-meter usual pace test (meters/sec) (Table 3). In the fully adjusted model for gait speed, women who averaged <6 hours sleep per night still had 3.5% slower gait speed than the women who averaged 6-6.8 hours of sleep. When the quadratic term for nighttime sleep duration was introduced into the model for gait speed, it was signicant in the unadjusted and fully adjusted models.
SLEEP, Vol. 30, No. 10, 2007 1320

7.5 3

8-8.9 hours

>9 hours

Figure 1Comparison of the percent of women with measured total sleep time vs. the percent of women in the 2003 National Sleep Foundation poll reporting their amount of sleep (NSF 2003) Sleep and Daytime Performance in Older WomenGoldman et al

Table 4Odds of Being Unable to Complete 5 Chair Stands by Quartile of TST, Wake After Sleep Onset, and Daytime Sleep in the Study of Osteoporotic Fractures Participants at the Eighth Clinic Visit TST1 (Hours) Odds ratio (95% CI) Unadjusted Adjusted2 N 2,883 2,672 N 2,882 2,672 N 2,839 2,633 < 6.0 1.22 (0.98, 1.53) 1.19 (0.68, 1.15) <0.7 Referent Referent < 0.5 Referent Referent 6.0- 6.8 0.86 (0.68, 1.08) 0.88 (0.68, 1.15) 0.7 1.1 1.12 (0.88, 1.44) 1.07 (0.82, 1.41) 0.5 -1 1.34 (1.05, 1.70) 1.20 (0.92, 1.57) 6.8 -7.5 referent referent 1.1 1.6 1.72 (1.36, 2.19) 1.54 (1.19, 2.01) 1 -1.8 1.37 (1.07, 1.74) 1.03 (0.78, 1.34) 7.5 0.91 (0.73, 1.15) 1.02 (0.79, 1.32) 1.6 2.76 (2.19, 3.48) 2.12 (1.63, 2.76) 1.8 2.20 (1.75, 2.77) 1.39 (1.07, 1.81)

Wake After Sleep Onset1 (Hours) Odds ratio (95% CI) Unadjusted Adjusted2

Daytime Sleep1 (Hours) Odds ratio (95% CI) Unadjusted Adjusted2

1

Measured with wrist actigraphy (Sleep-Watch-O). TST=average hours sleep in bed at night; Wake after sleep onset=average hours of wake after sleep onset; Daytime sleep = average hours of sleep, out of bed. 2 Adjusted for age, race, BMI, depression (GDS15), anxiety (Goldberg Anxiety Score 0-9), cognitive function (MMSE), number of comorbidities (Doctor ever told you: stroke, diabetes, Parkinson disease, Alzheimer disease, chronic obstructive pulmonary disease, heart disease (heart attack or coronary event), congestive heart failure, and hypertension), and walking for exercise.

Daytime Sleep In the unadjusted model for daytime sleep and gait speed, a longer duration of daytime sleep was associated with a slower gait speed. This association was attenuated in the fully adjusted multivariable model. Chair Stands Total Sleep Time Nighttime sleep duration was signicantly associated with time to complete 5 chair stands in both the adjusted and unadjusted models (table 3). In the fully adjusted model for chair stand time, women who averaged 7.5 hours of sleep per night took 4.1% longer to complete 5 chair stands than those who slept 6.8-7.5 hours/night. The quadratic term for TST was not signicant for the chair stand time outcome. In addition, women who averaged <6 hours and 7.5 hours of TST were not at increased odds of being unable to complete the 5 chair stands (Table 4). Wake After Sleep Onset Women with over 1.6 hours of WASO took 7.6% longer to complete 5 chair stands (Table 3). Higher amounts of WASO were also associated with higher odds of not being able to complete 5 chair stands. We also explored the possibility of an association between short TST and higher WASO with poorer physical function and found minimal correlation between the 2 sleep variables. Daytime Sleep Women with 1.8 hours or more of daytime sleep took 8.7% longer to complete 5 chair stands than those with <1.8 hours of daytime sleep in the unadjusted model (Table 3). In the fully adjusted multivariable models, those with 1.8 hours of sleep took 6.4 % longer than those with 0.5-1 hour of daytime sleep. Higher amounts of daytime sleep were also associated with 2.20 and 1.39
SLEEP, Vol. 30, No. 10, 2007 1321

higher odds of not being able to complete 5 chair stands in the unadjusted and adjusted models, respectively. Hand Grip Strength Total Sleep Time There was no association between TST and grip strength in either the unadjusted or adjusted models (Table 3). Wake After Sleep Onset Women with 1.6 hours of WASO had 4.0% weaker grip strength (Table 3). This association remained in the fully adjusted model. Daytime Sleep In the unadjusted model women with more daytime sleep had weaker grip strength (Table 3). This association was attenuated in the fully adjusted model. Sleep and Functional Limitations Women who slept <6 hours or 7.5 hours/night had higher odds of having a functional limitation than women who slept 6.8-7.7 hours/night in the unadjusted model (Table 5). Higher odds for a functional impairment remained for those who slept 7.5 hours even after complete adjustment. Women with more WASO had higher odds of functional impairment. In the fully adjusted model, women with 1.6 hours of WASO still had 1.80 (95% CI: 1.39, 2.35) times the odds of a functional limitation compared to those with less than 0.7 hours of WASO. Finally, women who slept 1 hour during the day had higher odds of a functional limitation then those with <0.5 hours of daytime sleep. Higher odds remained in the fully adjusted model. Although data are not shown, when the functional limitations were separated into ambulatory limitations and other limitations, similar tendencies were noted.
Sleep and Daytime Performance in Older WomenGoldman et al

Table 5Odds of Having a Functional Limitation1 by Quartile of TST, Wake After Sleep Onset, and Daytime Sleep in the Study of Osteoporotic Fractures Participants at the Eighth Clinic Visit TST2 (Hours) Odds ratio (95% CI) Unadjusted Adjusted3 N 2,889 2,674 n 2,889 2,674 n 2,845 2,634 < 6.0 1.48 (1.19, 1.83) 1.16 (0.90, 1.51) <0.7 Referent Referent < 0.5 Referent Referent 6.0- 6.8 1.19 (0.95, 1.48) 1.09 (0.84, 1.42) 0.7 1.1 1.30 (1.04, 1.64) 1.19 (0.91, 1.55) 0.5 -1 1.42 (1.12, 1.78) 1.21 (0.92, 1.58) 6.8 -7.5 Referent Referent 1.1 1.6 1.69 (1.35, 2.12) 1.39 (1.07, 1.80) 1 -1.8 1.91 (1.52, 2.39) 1.37 (1.05, 1.79) 7.5 1.41 (1.13, 1.74) 1.37 (1.06, 1.76) 1.6 2.77 (2.22, 3.45) 1.80 (1.39, 2.35) 1.8 2.55 (2.04, 3.20) 1.42 (1.08, 1.85)

Wake After Sleep Onset2 (Hours) Odds ratio (95% CI) Unadjusted Adjusted3

Daytime Sleep2 (Hours) Odds ratio (95% CI) Unadjusted Adjusted3

1

Difculty with at least 1 of the following 6 functional limitations: walking 2-3 blocks, climbing or walking down 10 steps, preparing meals, heavy housework, and shopping. 2 Measured with wrist actigraphy (Sleep-Watch-O). TST = Average hours sleep in bed at night; Wake after sleep onset=Average hours Wake After Sleep Onset; Daytime sleep=average hours sleep, out of bed. 3 Adjusted for age, race, BMI, depression (Geriatric Depression Score), anxiety (Goldberg Anxiety Score), cognitive function (MMSE), number of comorbidities (Doctor ever told you: stroke, diabetes, Parkinson disease, Alzheimer disease, chronic obstructive pulmonary disease, heart disease (heart attack or coronary event), congestive heart failure, and hypertension), and walking for exercise.

DISCUSSION This study is one of the rst large-scale studies of community dwelling older women to examine the association of sleep behaviors, neuromuscular performance, and daytime function using objective measurements of sleep and physical performance and subjective measurements of functional limitations. The results suggest that those women with more disrupted sleep as characterized by shorter sleep duration and longer wake time during the night, and those with greater daytime sleepiness as characterized by napping behavior, were at greater risk for poorer neuromuscular performance and poorer daytime function. Women with objective measures of poor sleep walked slower, took longer to complete 5 chair stands, had lower grip strength, and had more trouble performing independent activities of daily living. These results held up even after adjustment for multiple confounders and other explanatory variables. The association of short nighttime sleep time with poorer function is consistent with other data on the relationship between sleep time and self-reports of health problems27 or mortality.7,8,28 Individuals sleeping 6 to 7 hours per night have been reported to have the lowest risk of mortality.7,29-31 The women in our study with a TST between 6.0-7.5 hr tended to perform better on neuromuscular performance tasks. Fragmented sleep occurs chronically in older populations and has been correlated with daytime sleepiness.32 Indeed, frequent nighttime awakening is a cardinal symptom of sleep complaints in the older adult. Sleep that is disrupted by brief awakenings is less restorative than consolidated sleep and may result in signicantly altered sleep stage distribution, the appearance of daytime sleepiness, and decreased performance on reaction time and digit symbol substitution tests.33 A major nding of this study was that older women with more fragmented sleep (as measured by minutes of wake after sleep onset) had worse neuromuscular performance and were more likely to have functional impairments. Although frequent awakenSLEEP, Vol. 30, No. 10, 2007 1322

ings are yet to be associated with mortality,8 the signicant reduction in daytime function associated with these multiple awakenings warrants attention to the complaint in a clinical setting. Previous studies have shown an association between napping and self-reports of daytime sleepiness, self-reported poorer health, impaired physical function or mood, being overweight, mortality, cardiovascular disease, myocardial infarction, and congestive heart failure.6,27 In the current study, napping in these older women was associated with higher odds of having a functional limitation. These associations were similar for the ambulatory limitations as a group and for the other functional limitations group. The relationship between sleep behavior and neuromuscular performance is complex, with this study suggesting that poor sleep as measured by actigraphy may be associated with poorer neuromuscular function and greater functional limitations. Walking speed has been shown to be an independent determinant of self-rated health, with slower walk speed associated with poor self-rated health.34 Physical performance measures of lower extremity function, and particularly gait speed, have been shown to predict the onset of progressive ADL impairment, mobility, and upper extremity disability in older women.35 Usual gait speed over a 6-m course <1 m/sec has been identied as a meaningful cutpoint to identify risk of major adverse health-related outcomes in older adults.36 Further, a change of 0.05 m/sec in gait speed is considered a meaningful change in physical performance in older adults.37 The average gait speed in the SOF cohort is already low and has been strongly related to disability in this cohort.38 Handgrip strength has been shown to be a strong predictor of cause-specic and total mortality in older disabled women.14 It has also been associated with reports of more difculties in the physical activities of daily living. Although there was a signicant association between lower average hand grip strength and more wake time during the night, our data in general did not nd an association between average handgrip strength and the other sleep variables. These women had an average grip strength alSleep and Daytime Performance in Older WomenGoldman et al

ready considered to be low.14,39 Further research to clarify the association between hand grip strength and wake after sleep onset is warranted. Past studies have examined sleep and function with sleep diaries or other subjective reports of sleep. While subjective data are of interest, the correlation between reports of sleep and objective measures of sleep are low, with subjective measures of total sleep time often being underestimated and wake time being overestimated. Although there are no normative data on sleep in older women, the 2003 National Sleep Foundation poll did question women about how much sleep they thought they got a night.40 As Figure 1 shows, the reported hours of sleep are quite different from the measured hours of sleep. This suggests that reports of sleep may not represent the physiological impact of objective measures. One strength of this study was the use of actigraphy to objectively measure sleep. Other strengths include data from a large sample of community dwelling women, both black and white, in a well-established and followed cohort, and performance measurements made with stateof-the-art technology by well-trained clinical staff. This study also has some limitations. The analysis subset was healthier than the overall SOF cohort, and may not be representative of the general cohort of older women. Thus, the observed associations likely underestimate the true associations since unhealthier women were excluded from the analysis. Further, these ndings may not be generalizable to other populations including nonambulatory women, younger women, men, or institutionalized individuals. Finally, results are cross-sectional and one cannot determine if sleep disturbance precedes impairments in performance, or vice versa. In summary, disrupted sleep was associated with poorer neuromuscular performance and more functional limitation in older women. As TST deviated either above or below an average of 7 hours, gait speed was slower, and it took longer to complete 5 chair stands. With greater amounts of wake after sleep onset, gait speed was slower, it required more time to complete 5 chair stands, and grip strength weakened. Finally, as the amount of sleep fragmentation or daytime sleep increased, the odds for a functional impairment and inability to complete the chair stands were higher. These ndings were not fully explained by adjustment for demographic or health conditions. Our study has provided strong evidence to suggest that poor nighttime sleep is associated with poorer neuromuscular performance and more functional limitation. Research to identify the causal association between TST, wake after sleep onset, and the decline in neuromuscular performance, as well as to identify potential avenues to minimize neuromuscular decline is warranted. Longitudinal studies are needed to determine if poor sleep patterns are causally associated with functional decline. ACKNOWLEDGEMENTS This work was supported by NIH Grants AG05407, AR35582, AG05394, AR35584, AR35583, AG08415 and the Aging Training Grant Number: 2, T32, AG000181-16. REFERENCES
1. Ancoli-Israel S. Insomnia in the elderly: A review for the primary care practitioner. Sleep 2000;23(Suppl 1):S23-30. 1323

2. 3.

4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

14.

15. 16.

17.

18. 19. 20.

21. 22.

Martin J, Stepnowsky C, Ancoli-Israel S. Sleep apnea in the elderly. In: McNicholas WT, Phillipson EA, Breathing disorders during sleep London: WB Saunders, 2002:278-87. McCrae CS, Rowe MA, Tierney CG, Dautovich ND, DeFinis AL, McNamara JP. Sleep complaints, subjective and objective sleep patterns, health, psychological adjustment, and daytime functioning in community-dwelling older adults. J Gerontol B Psychol Sci Soc Sci 2005;60:182-9. Ohayon MM. Epidemiology of insomnia: what we know and what we still need to learn. Sleep Med Rev 2002;6:97-111. Vitiello MV, Moe KE, Prinz PN. Sleep complaints cosegregate illness in older adults: clinical research informed by and informing epidemiological studies of sleep. J Psychosom Res 2002;53:555-9. Hays JC, Blazer DG, Foley DJ. Risk of napping: Excessive daytime sleepiness and mortality in an older community population. J Am Geriatr Soc 1996;44:693-8. Kripke DF, Garnkel L, Wingard DL, Klauber MR, Marler MR. Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry 2002;59:131-6. Newman AB, Spiekerman CF, Enright P, et al. Daytime sleepiness predicts mortality and cardiovascular disease in older adults. J Am Geriatr Soc 2000;48:115-23. Newman AB Enright PL, Manolio TA, Haponik EF, Wahl PW. Sleep disturbance, psychosocial correlates, and cardiovascular disease in 5201 older adults. J Am Geriatr Soc 1997;45:1-7. Whitney CW, Enright PL, Newman AB, Bonekat W, Foley D, Quan SF. Correlates of daytime sleepiness in 4578 elderly persons: the cardiovascular health study. Sleep 1998;21:27-36. National Sleep Foundation. 2003 Sleep in America Poll; 2003. Applegate WB, Blass JP, Williams TF. Instruments for the functional assessment of older patients. N Engl J Med 1990;322:1207-14. Guralnik JM, Ferrucci L, Pieper CF, et al. Lower extremity function and subsequent disability: consistency across studies, predictive models, and value of gait speed alone compared with the short physical performance battery. J Gerontol A Biol Sci Med Sci 2000;55:M221-31. Rantanen T, Volpato S, Ferrucci L, Heikkinen E, Fried LP, Guralnik JM. Handgrip strength and cause-specic and total mortality in older disabled women: exploring the mechanism. J Am Geriatr Soc 2003;51:636-41. Foley DJ, Monjan AA, Brown SL. Sleep complaints among elderly persons: an epidemiologic study of three communities. Sleep 1995;18:425-32. Kutner NG SK, Ory MG, Baker DI, FICSIT Group. Older adults perceptions of their health and functioning in relation to sleep disturbance, falling, and urinary incontinence. J Am Geriatr Soc 1994;42:757-62. Foley DJ, Ancoli-Israel S, Britz P, Walsh JK. Sleep disturbances and chronic disease in older adults: results of the 2003 National Sleep Foundation Sleep in America Survey. J Psychosom Res 2004;56:497-502. Cummings SR Black DM, Nevitt MC, et al. Appendicular bone density and age predict hip fracture in women: the Study of Osteoporotic Fractures Research Group. JAMA 1990;263:665-8. Blackwell T, Ancoli-Israel SA, Gehrman PR, Schneider JL, Pdeula KL, Stone KL. Actigraphy scoring reliability in the Study of Osteoporotic Fractures. Sleep 2005;28:1599-603. Blackwell T, Yaffe K, Ancoli-Israel S, et al. Poor sleep is associated with impaired cognitive function in older women: The Study of Osteoporotic Fractures. J Gerontol A Biol Sci Med Sci 2006;61:405-10. Girardin JL, Kripke DF, Mason WJ, Elliott JA, Youngstedt SD. Sleep estimation from wrist movement quantied by different modalities. J Neurosci Methods 2001;105:185-91. Cummings SR, Nevitt MC, Browner WS, et al. Risk factors for hip fracture in white women. N Engl J Med 1995;332:767-74. Sleep and Daytime Performance in Older WomenGoldman et al

SLEEP, Vol. 30, No. 10, 2007

23. Seeley DG Cauley JA, Grady D, Browner WS, Nevitt MC, Cummings SR. Is postmenopausal estrogen therapy associated with neuromuscular function or falling in elderly women? Study of Osteoporotic Fractures Research Group. Arch Intern Med 1995;155:293-9. 24. Pincus T, Summey JA, Soraci SA, Wallston KA, Hummon NP. Assessment of patient satisfaction in activities of daily living using a modied Stanford Health Assessment Questionnaire. Arthritis Rheum 1983;26:1346-53. 25. Gregg EW Mangione C, Cauley JA,et al: Study of Osteoporotic Fractures Research Group. Diabetes and incidence of functional disability in older women. Diabetes Care 2002;25:61-7. 26. Schneider J, Blackwell T, Hillier TA, Stone KL. Characteristics of objective and subjective sleep in older women. The Study of Osteoporotic Fractures (SOF). Sleep 2004;27(Abstract Supplement):A123. 27. Ohayon MM, Vecchierini MF. Normative sleep data, cognitive function and daily living activities in older adults in the community. Sleep 2005;28:981-9. 28. Bursztyn M, Ginsberg G, Stessman J. The siesta and mortality in the elderly: effect of rest without sleep and daytime sleep duration. Sleep 2002;25:187-91. 29. Patel SR, Ayas NT, Malhotra MR, et al. A prospective study of sleep duration and mortality risk in women. Sleep 2004;27:440-4. 30. Ayas NT, White DP, Manson JE, et al. A prospective study of sleep duration and coronary heart disease in women. Arch Int Med 2003;163:205-9. 31. Tamakoshi A, Ohno Y. Self-reported sleep duration as a predictor of all-cause mortality: results from the JACC study, Japan. Sleep 2004;27:51-4. 32. Carskadon MA, Brown ED, Dement WC. Sleep fragmentation in the elderly: relationship to daytime sleep tendency. Neurobiol Aging 1982;3:321-7. 33. Stepanski EJ. The effect of sleep fragmentation on daytime function. Sleep 2002;25:268-75. 34. Jylha M, Guralnik JM, Balfour J, Fried LP. Walking difculty, walking speed, and age as predictors of self-rated health: The Womens Health and Aging Study. J Gerontol A Biol Sci Med Sci 2001;56: M609-17. 35. Onder G, Ferrucci L, Fried LP, Guralnik JM, Pahor M. Measures of physical performance and risk for progressive and catastrophic disability: results from the Womens Health and Aging Study. J Gerontol A Biol Sci Med Sci 2005;60:74-9. 36. Cesari M, Visser M, Newman A, et al. Prognostic value of usual gait speed in well-functioning elders. results from the Health ABC Study. J Am Geriatr Soc 2004;52(4 Suppl):s15-6. 37. Perera S Moody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc 2006;54:743-9. 38. Ensrud KE, Nevitt MC, Yunis C, et al. Correlates of impaired function in older women. J Am Geriatr Soc 1994;42:481-9. 39. Rantanen T, Guralnik J, Sakari-Rantala R, et al. Disability, physical activity, and muscle strength in older disabled women. The Womens Health and Aging Study. Arch Phys Med Rehabil 1999;80:130-5. 40. National Center on Sleep Disorders Research. 2003 National Sleep Disorders Research Plan. Bethesda, MD: National Institutes of Health; 2003:60-63. NIH publication 03-5209.

SLEEP, Vol. 30, No. 10, 2007

1324

Sleep and Daytime Performance in Older WomenGoldman et al