Abstract
Study Objectives:
To examine the association between SDB and subjective measures of daytime sleepiness, sleep quality, and sleep related quality of life in a large cohort of primarily community-dwelling older women, specifically considering the relative importance of sleep duration in mediating these associations.
Design:
Cross-sectional. The functional outcome measures of interest were daytime sleepiness (using the Epworth Sleepiness Scale, ESS), sleep-related symptoms (Pittsburgh Sleep Quality Index, PSQI), and sleep related quality of life (Functional Outcomes of Sleep Questionnaire, FOSQ). ANOVA and regression analyses examined the association between SDB severity (measured by indices of breathing disturbances and overnight oxygen saturation) and sleep time (by actigraphy) and these outcome measures. Regression models were adjusted for age, body mass index (BMI), and a medical comorbidity index. We specifically explored whether associations with indices of SDB were mediated by sleep deprivation by adjusting models for actigraphy-determined average total sleep time (TST) during the night.
Setting:
Community-based sample examined in home and outpatient settings.
Participants:
461 surviving older women from the multicenter Study of Osteoporotic Fractures were examined during Visit 8 from 2002–03. All participants underwent in-home overnight polysomnography for one night and wrist actigraphy for a minimum of 3 24-h periods and completed the above functional outcomes questionnaires.
Interventions:
N/A
Measurements and Results:
Participants were aged 82.9 ± 3.5 (mean ± SD) years, had BMI of 27.9 ± 5.1 kg/m2, and had an apnea-hypopnea index (AHI) of 15.7 ± 15.1. AHI and TST demonstrated a weak correlation (r = −0.15). ESS score individually demonstrated a modest association with AHI, oxygen desaturation, and TST. The association of ESS score and AHI—but not oxygen desaturation—was attenuated to some extent by adjustment for TST. PSQI and FOSQ scores were not associated with measures of SDB severity or TST.
Conclusions:
After adjustment for TST, SDB severity in community-dwelling older women was not independently associated with self-reported daytime sleepiness, although there may be a modest association that is mediated through reduced TST. In older women, SDB severity was not associated with indices of sleep related symptoms or sleep related quality of life.
Citation:
Kezirian EJ; Harrison SL; Ancoli-Israel S Redline S; Ensrud K; Claman DM; Stone KL. Behavioral correlates of sleep-disordered breathing in older women. SLEEP 2007;30(9):1181-1188.
Keywords: Sleep disordered breathing, obstructive sleep apnea, older adults, sleepiness
INTRODUCTION
THE PREVALENCE OF SLEEP DISORDERED BREATHING (SDB) IN OLDER ADULTS IS HIGHER THAN IN MIDDLE-AGED ADULTS. IN A COMMUNITY-DWELLING population of older adults aged 65–95 years, Ancoli-Israel et al reported a SDB prevalence (apnea-hypopnea index, or AHI, ≥20) of 51% in men and 39% in women.1 In contrast, estimates of SDB in middle-aged adults are generally less than 10%2; side-by-side comparison studies confirm a higher prevalence of SDB in older adults compared to the middle-aged.3–7
In spite of the higher prevalence of SDB among older adults, the functional consequences have not been clearly established. In young and middle-aged adults, SDB has been associated with daytime somnolence, sleep related symptoms, and decrements in sleep related quality of life.8 Large population-based studies of older adults have shown an association between SDB severity and daytime sleepiness.1,9–11 However, no prior study has been able to explore whether the associations of SDB and sleepiness/sleep related quality of life are explained by short sleep duration or other objective indices of sleep quality. Aging and age-related comorbidities and medication use are associated with changes in sleep architecture.12–14 Because reduced sleep duration or sleep efficiency may themselves contribute to daytime symptoms, it may be important to consider the independent associations of SDB and fragmented or shorter sleep with functional consequences in studies of older adults.
The objectives of this study were (1) to examine the association between SDB severity and daytime sleepiness, sleep related symptoms, and sleep related quality of life in a large cohort of primarily community-dwelling older women and (2) to determine whether any associations were influenced by adjustment for sleep duration (specifically total sleep time, determined by actigraphy). Based on data from young and middle-aged adults, we hypothesized that SDB severity would be associated with daytime sleepiness, sleep disturbances and poorer sleep related quality of life.
METHODS
Population
Participants were from the Study of Osteoporotic Fractures (SOF), a multicenter prospective cohort study of 9,704 primarily white women aged 65 years or older that began in 1986–1988. Details of the SOF have been described previously.15 During the eighth study visit, conducted from 2002–3, a convenience sample (n=461) of surviving older women from the Minnesota (n=112) and Pittsburgh (n=349) study sites underwent polysomnography and wrist actigraphy. All study participants were evaluated with a clinic interview, anthropometry, performance measures, and actigraphy. This study was approved by the institutional review board of each involved institution.
Polysomnography
In-home PSG data were collected using the Compumedics Siesta Unit (Abbotsville, AU), with monitoring of 2 central electroencephalographic leads (EEG; C1, C2), bilateral electrooculogram (EOG), chin electromyogram (EMG), thoracic and abdominal respiratory effort, airflow (by a nasal-oral thermocouple and nasal pressure recording), finger pulse oximetry, electrocardiogram (ECG), body position, and bilateral leg movements (by piezo sensors). Sleep data were scored centrally by certified scorers blinded to other data. Sleep stages and arousals were scored using standard criteria.16 Apneas were defined as a complete or almost complete cessation of airflow (by thermocouple) associated with >3% oxygen desaturation, and hypopneas were identified as a clearly discernible (at least 30%) reduction in respiratory sensor channels associated with a >3% oxygen desaturation. Apneas associated with no evidence of effort on both thoracic and abdominal channels were considered to be “central,” and otherwise as “obstructive.” Arousals were defined as an abrupt shift in EEG frequency of 3 seconds or more and requiring an increase in chin EMG activity if occurring during REM sleep.17
Calculated variables used as indices of SDB severity were AHI (all apneas plus hypopneas/h of sleep), obstructive apnea-hypopnea index (OAHI, number of obstructive apneas plus hypopneas/h of sleep), obstructive apnea index (OAI, number of obstructive apneas/h of sleep), central apnea index (CAI, number of central apneas/h of sleep), arousal index (ArI, number of arousals/h of sleep), and the percentage of sleep time with oxygen saturation below 90% (SaO2 <90%). AHI was considered as a continuous variable and also as a categorical variable based on tertiles. The percentage of sleep time with oxygen saturation below 90% was markedly skewed, and therefore was considered as a binary outcome based (<2% vs. ≥2% time spent at <90% saturation). All other polysomnography variables were evaluated as continuous variables.
Actigraphy
Actigraphy was used to record wrist activity from which sleep/wake was calculated. Actigraphy allows recordings for multiple nights which reduces the night-to-night variability and captures napping behavior.18 Wrist actigraphy (Sleep-Watch-O, Ambulatory Monitoring, Inc, Ardsley, NY) data were collected for ≥3 consecutive 24-h periods (mean recording time 86.1 ± 14.2 h; mean number of 24-h periods 3.6 ± 0.6). Data collected in the proportional integration mode were used for this analysis.19 Total sleep time was calculated as the mean night sleep time (defined as sleep time within the reported period spent in bed) averaged over a 24-h period using data from the entire recorded period of actigraphy. Napping time was calculated as the mean day sleep time (defined as sleep time within the reported period spent out of bed—even if the subject returned to bed during the day) averaged over a 24-h period using data from the entire recorded period of actigraphy. Actigraphy was not performed concurrently with polysomnography in most cases; however, in a subset of 71 patients, polysomnogram was performed during one of the nights evaluated with actigraphy with a resulting correlation of r = 0.78 for total sleep time.
To test for an association between SDB severity and sleep time, the Pearson correlation coefficient (assuming a normal distribution of the variables) and Spearman's rank correlation coefficient (accounting for non-normal distribution) were calculated for TST and each measure of SDB severity.
Functional Outcome Measures
Daytime sleepiness, sleep symptoms, and sleep related quality of life were each evaluated with the use of established, self-administered questionnaires that have been validated in younger populations.
Daytime sleepiness was quantified using the Epworth Sleepiness Scale (ESS). The ESS measures sleep propensity on a 0–3 scale in eight standardized daily situations. Possible scores range from zero to 24, and higher scores reflect greater sleepiness.20 ESS score was considered as a continuous variable. ESS has not been validated in older adults. Since the ESS includes a question related to driving which may not be as applicable to all older adults, the ESS was considered with and without inclusion of this question; the modified ESS score has a possible score range from zero to 21.
Subjective sleep symptoms, disturbances, and patterns were assessed with the Pittsburgh Sleep Quality Index (PSQI), an instrument validated for use in older populations.21,22 The PSQI measures a broad range of symptoms of sleep disturbances over a one-month period. Responses to 19 questions are categorized into 7 component scores (subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction); these are then combined to create one global score.23 The total global score ranges from zero to 21, and greater scores indicate higher levels of sleep symptoms. PSQI score was considered as a continuous variable.
Sleep related quality of life was evaluated with the Functional Outcome of Sleep Questionnaire (FOSQ) a 30-item instrument that measures the effect of excessive daytime sleepiness on activities of daily living.24 Mean-weighted item scores are used to generate 5 subscales (activity level, vigilance, intimacy and sexual relationships, general productivity, and social outcome) that together produce a composite score. The total score ranges from 5 to 20, and lower scores indicate greater dysfunction. The FOSQ score was analyzed as a continuous variable.
COVARIATES
Study participants underwent a questionnaire assessment of medical history and brief physical examination at each study visit. For this analysis, the following variables were included in the analysis: age, race, body mass index (BMI; kg/m2), physician-diagnosed depression, and self-reported health status. Self-reported health status was scaled as excellent, good, fair, poor, or very poor; the responses were treated as an ordinal variable, with the responses converted to a whole number from 1–5 (excellent, 1; very poor, 5) in regression analysis.
Analysis
Statistical analyses were performed using SAS (SAS Institute, Cary, NC) software. Mean values (with standard deviations where applicable) were calculated for age, race, BMI, physician-diagnosed depression, TST, and the functional measures (ESS, PSQI, and FOSQ scores) separately for the subset of patients undergoing polysomnography and the remainder of the SOF cohort at Visit 8. Similar calculations were then performed after dividing the population undergoing polysomnography into AHI tertiles. One-way analysis of variance (continuous variables) and chi-squared testing (categorical variables) were used to compare the values.
One-way analysis of variance was used to consider the variation in outcome measures among groups separately defined either by SDB severity (using AHI tertiles) or by TST (using tertiles) without adjustment for other variables.
Multiple linear regression analysis was performed to determine the differences in ESS, PSQI, and FOSQ scores associated with the measures of SDB severity. Coefficient estimates are reported with 95% confidence intervals (CI). All regression models were adjusted for age, race, BMI, physician-diagnosed depression, and self-reported health status. To determine the influence of TST on the observed associations between SDB severity and the outcome measures, each model was tested separately with and without adjustment for TST. P-values < 0.05 were considered statistically significant in all analyses.
Post hoc power analysis revealed that this sample had 80% power to detect a mean difference of 1.17 in ESS score, 1.23 in PSQI score, and 0.34 in FOSQ score between the lowest and highest AHI tertiles.
RESULTS
Description of the Sample
Of the 4701 women seen in Visit 8, 461 at the Pittsburgh and Minnesota sites were studied with both polysomnography and wrist actigraphy. Five hundred participants were invited—and 494 completed—the evaluation, but data were unusable in 33 of the participants. When compared to those women in all study sites with actigraphy but without polysomnography, those with both polysomnography and actigraphy (and included in this study) were modestly younger and had a higher BMI and better self-reported health status (Table 1). There were no significant differences between those with and without sleep recordings in any other variables, including TST (derived from actigraphy) and ESS, PSQI, or FOSQ scores.
Table 1.
Comparison Between Subset of Study of Osteoporotic Fractures Cohort Undergoing Polysomnography and Remainder of Cohort
| Polysomnography | No Polysomnography | P-value | |
|---|---|---|---|
| N = 461 | N = 4240 | ||
| Age (years) | 82.9 ± 3.5 | 84.2 ± 4.1 | <0.0001 |
| Race (% African American) | 38 (8.3) | 428 (10.0) | 0.22 |
| Body Mass Index (kg/m2) | 27.9 ± 5.1 | 26.8 ± 5.0 | <0.0001 |
| Physician-diagnosed depression (%) | 52 (11.3) | 600 (14.6) | 0.057 |
| Self-reported health status (%) | |||
| Excellent | 100 (22) | 744 (18) | 0.02 |
| Good | 254 (55) | 2258 (53) | |
| Fair | 97 (21) | 1052 (25) | |
| Poor | 9 (2) | 162 (4) | |
| Very poor | 1 (0.2) | 24 (1) | |
| Epworth Sleepiness Scale | 5.8 ± 3.6 | 5.8 ± 4.2 | 0.69 |
| Pittsburgh Sleep Quality Index | 6.6 ± 3.8 | 6.2 ± 3.7 | 0.22 |
| Functional Outcomes of Sleep Questionnaire | 15.2 ± 0.8 | 15.3 ± 0.9 | 0.20 |
| Actigraphic Total Sleep Time (minutes) | 408.8 ± 70.4 | 405.3 ± 78.8 | 0.88 |
For continuous variables, means are shown with standard deviation in parentheses.
P-values reflect the appropriate test of significance (one-way analysis of variance for continuous variables and chi-squared test for categorical variables) to determine whether there is statistically significant variation among the AHI tertiles.
Polysomnography results are shown in Table 2. The distribution of the AHI was: <5 (23.4%), 5-<15 (38.6), ≥15–30 (24.5%), and >30 (13.5%). The range for AHI was 0 to 90, with the 25th percentile of 5.3 and 75th percentile of 21.4. One hundred thirty-nine subjects (30.2%) demonstrated at least 2% of sleep time with SaO2 <90%.
Table 2.
Distribution of Sleep Disordered Breathing Indices
| Mean ± SD | Median | Inter-Quartile Range | |
|---|---|---|---|
| Apnea-hypopnea index | 15.7 ± 15.1 | 11.4 | 5.3–21.4 |
| Obstructive apnea-hypopnea index | 15.3 ± 14.9 | 11.7 | 5.3–21.3 |
| Obstructive apnea index | 2.5 ± 5.6 | 0.4 | 0.2–3.7 |
| Central apnea index | 0.3 ± 2.9 | 0 | 0 |
| Arousal index | 20.7 ± 12.0 | 18.0 | 12.0–26.3 |
| Percent of sleep time with oxygen saturation below 90% | 4.0 ± 9.7 | 0.7 | 0.1–2.8 |
SD, standard deviation
Table 3 presents the covariates and total sleep time for the entire study population and according to AHI tertiles. Individuals with higher AHI levels had significantly lower levels of average TST by actigraphy. There was no association between AHI and napping time (data not shown).
Table 3.
Subject characteristics by level of AHI
| AHI Tertiles |
|||||
|---|---|---|---|---|---|
| Total (N=461) | <7.06 (N=153) | 7.06–17 (N=154) | >17 (N=154) | P-value | |
| Age (years) | 82.9 ± 3.5 | 82.7 ± 3.6 | 82.4 ± 3.0 | 83.6 ± 3.7 | 0.008 |
| Race (% African American) | 38 (8.2) | 17 (11.1) | 11 (7.1) | 10 (6.5) | 0.22 |
| Body Mass Index (kg/m2) | 27.9 ± 5.1 | 26.6 ± 4.7 | 28.9 ± 5.0 | 28.2 ± 5.3 | 0.0002 |
| Physician-diagnosed depression | 52 (11.3) | 14 (9.2) | 18 (11.7) | 20 (13.0) | 0.56 |
| Self-reported health status (%) | |||||
| Excellent | 100 (22) | 32 (20.1) | 36 (23.3) | 32 (20.8) | 0.89 |
| Good | 254 (55) | 83 (54.3) | 85 (55.2) | 86 (55.9) | |
| Fair | 97 (21) | 36 (23.6) | 30 (19.5) | 31 (20.1) | |
| Poor | 9 (2) | 2 (1.3) | 3 (2.0) | 4 (2.6) | |
| Very poor | 1 (0.2) | 0 | 0 | 1 (0.6) | |
| Total Sleep Time (minutes) | 404.8 ± 70.4 | 415.5 ± 60.5 | 407.6 ± 70.4 | 391.4 ± 77.3 | 0.01 |
For continuous variables, means are shown with standard deviation in parentheses.
P-values reflect the appropriate test of significance (one-way analysis of variance for continuous variables and chi-squared test for categorical variables) to determine whether there is statistically significant variation among the AHI tertiles.
Association Between SDB Measures and TST
The Spearman's rank correlation coefficient for AHI and TST was −0.15 (P = 0.0013), respectively, suggesting a weak correlation whereby higher AHI was associated with lower TST. Other measures of SDB severity demonstrated similar results (data not shown).
ESS
Fifty-two of 461 subjects (11.3%) had an ESS >10. Higher ESS score was associated with both higher AHI and lower TST when considered separately using one-way analysis of variance (Tables 4 and 5). The difference in ESS mean scores between the highest and lowest tertiles was 1.4 and 1.8 for AHI and TST, respectively. Similar results were obtained for the modified ESS score that did not include a driving question (data not shown).
Table 4.
Variation in Sleepiness and Functional Outcome Measures by the Apnea-Hypopnea Index
| AHI Tertiles |
||||
|---|---|---|---|---|
| < 7.06 | 7.06–17.0 | > 17.0 | P | |
| ESS | 5.2 (3.3) | 5.6 (3.6) | 6.6 (3.9) | 0.0026 |
| PSQI | 6.7 (3.6) | 6.4 (3.8) | 6.8 (4.0) | 0.70 |
| FOSQ | 19.0 (1.2) | 19.1 (1.0) | 19.0 (0.9) | 0.85 |
AHI, apnea-hypopnea index; ESS, Epworth Sleepiness Scale; PSQI, Pittsburgh Sleep Quality Index; FOSQ, Functional Outcomes of Sleep Questionnaire
Numbers shown represent mean (standard deviation) for each variable within tertiles. P-values reflect one-way analysis of variance testing to determine whether there is statistically significant variation among the AHI tertiles. Numbers in bold are statistically significant at the P< 0.05 level.
Table 5.
Variation in Sleepiness and Functional Outcome Measures by TST Tertiles
| <6.4 h | 6.4–7.3 h | >7.3 h | P | |
|---|---|---|---|---|
| ESS | 6.7 (4.1) | 5.7 (3.3) | 4.9 (3.3) | <0.0001 |
| PSQI | 6.9 (3.9) | 6.3 (3.6) | 6.7 (3.9) | 0.27 |
| FOSQ | 18.9 (1.1) | 19.0 (1.1) | 19.2 (1.0) | 0.045 |
TST, total sleep time; ESS, Epworth Sleepiness Scale; PSQI, Pittsburgh Sleep Quality Index; FOSQ, Functional Outcomes of Sleep Questionnaire
Numbers shown represent mean (standard deviation) for each variable within tertiles. P-values reflect one-way analysis of variance testing to determine whether there is statistically significant variation among the TST tertiles. Numbers in bold are statistically significant at the P< 0.05 level.
Multiple regression analysis (Table 6) showed a modest association between ESS and AHI, adjusting for age, race, BMI, physician-diagnosed depression, and self-reported health status; an increase in AHI of one standard deviation (approximately 15 units) was associated with an ESS score 0.44 points higher, a difference that is statistically significant but small relative to clinical interpretation. After adjustment for TST, this relationship was attenuated; the point estimate decreased by 32% in magnitude and was no longer statistically significant. In contrast, the association between ESS score and TST was modestly stronger and persisted in Model 2 that included adjustment for AHI; a one standard deviation increase in TST (approximately 75 minutes) was associated with an ESS score 1 point lower. Similar results were obtained when OAHI, CAI, and SaO2 <90% were substituted as the measure of SDB severity, although the association between ESS score and oxygen desaturation was not attenuated by adjustment for TST (data not shown). In contrast, the ESS score was not associated with the OAI, ArI or the dichotomous variable describing oxygen desaturation (data not shown). In all of these models, ESS score was associated with TST; the magnitude of the association was similar to that presented in Table 6, Model 2. Similar results were also obtained with the modified ESS score (data not shown). Adjustment for napping time, including the addition of napping time to TST, had no meaningful effect on these findings (data not shown).
Table 6.
Linear Regression Analysis of the ESS Score with AHI and TST
| Unit | Difference in ESS Score* |
||
|---|---|---|---|
| Model 1 | Model 2 | ||
| AHI | 15.1 | 0.44 (0.10, 0.78) | 0.30 (−0.03, 0.64) |
| TST (minutes) | 77.6 | −0.96 (−1.33, −0.59) | |
| Age (years) | 5 | −0.25 (−0.76, 0.26) | −0.22 (−0.72, 0.28) |
| African American race | 0.75 (−0.54, 2.04) | 0.58 (−0.67, 1.84) | |
| Body mass index | 5 | −0.13 (−0.48, 0.22) | −0.30 (−0.65, 0.05) |
| Physician-diagnosed depression | 0.02 (−1.03, 1.08) | 0.14 (−0.89, 1.17) | |
| Self-reported health status | 0.47 (0.002, 0.93) | 0.52 (0.06, 0.97) | |
| Adjusted R-squared | 0.015 | 0.068 | |
N = 461
Model 1 is a multivariate linear regression analysis of the association between ESS score (dependent variable) and AHI (independent variable), adjusting for age, race, body mass index, physician-diagnosed depression, and self-reported health status. Model 2 adds TST as an independent variable to Model 1 to adjust for potential confounding of the relationship between ESS and AHI.
AHI, apnea-hypopnea index; ESS, Epworth Sleepiness Scale; TST, total sleep time
Regression results shown represent coefficient point estimates with 95% confidence intervals. Numbers in bold are statistically significant at the P< 0.05 level.
PSQI
PSQI score was not associated with AHI or TST on one-way analysis of variance (Tables 4 and 5). On multiple regression analysis, PSQI score was not associated with AHI, other measures of SDB severity, or TST (Table 7). Adjustment for napping time, including the addition of napping time to TST, had no meaningful effect on these findings (data not shown).
Table 7.
Linear Regression Analysis of the PSQI Score with AHI and TST
| Unit | Difference in PSQI Score* |
||
|---|---|---|---|
| Model 1 | Model 2 | ||
| AHI | 15.1 | 0.14 (−0.21, 0.48) | 0.09 (−0.26, 0.45) |
| TST (minutes) | 77.6 | −0.22 (−0.61, 0.17) | |
| Age (years) | 5 | −0.02 (−0.54, 0.49) | −0.005 (−0.52, 0.51) |
| African American race | 0.47 (−0.83, 1.77) | 0.40 (−0.91, 1.72) | |
| Body mass index | 5 | −0.32 (−0.67, 0.04) | −0.34 (−0.70, 0.03) |
| Physician-diagnosed depression | 1.63 (0.57, 2.69) | 1.65 (0.58, 2.73) | |
| Self-reported health status | 1.26 (0.79, 1.73) | 1.26 (0.79, 1.74) | |
| Adjusted R-squared | 0.074 | 0.072 | |
N = 461
Model 1 is a multivariate linear regression analysis of the association between PSQI score (dependent variable) and AHI (independent variable), adjusting for age, race, body mass index, physician-diagnosed depression, and self-reported health status. Model 2 adds TST as an independent variable to Model 1 to adjust for potential confounding of the relationship between PSQI and AHI.
AHI, apnea-hypopnea index; PSQI, Pittsburgh Sleep Quality Index; TST, total sleep time
Regression results shown represent coefficient point estimates with 95% confidence intervals.
FOSQ
FOSQ score was not associated with AHI but did have a modest bivariate association with TST (Tables 4 and 5). After adjusting for age, race, BMI, physician-diagnosed depression, and self-reported health status, however, no significant association was demonstrated between the FOSQ score and AHI, other measures of SDB severity, or TST (Table 8). Adjustment for napping time, including the addition of napping time to TST, had no meaningful effect on these findings (data not shown).
Table 8.
Linear Regression Analysis of the FOSQ Score with AHI and TST
| Unit | Difference in FOSQ Score* |
||
|---|---|---|---|
| Model 1 | Model 2 | ||
| AHI | 15.1 | 0.01 (−0.08, 0.11) | 0.02 (−0.07, 0.12) |
| TST (minutes) | 77.6 | 0.10 (−0.01, 0.21) | |
| Age (years) | 5 | −0.01 (−0.16, 0.13) | −0.03 (−0.17, 0.12) |
| African American race | −0.34 (−0.71, 0.02) | −0.34 (−0.70, 0.03) | |
| Body mass index | 5 | −0.02 (−0.12, 0.07) | −0.01 (−0.11, 0.10) |
| Physician-diagnosed depression | −0.37 (−0.67, −0.07) | −0.33 (−0.63, −0.04) | |
| Self-reported health status | −0.33 (−0.46, −0.20) | −0.33 (−0.46, −0.20) | |
| Adjusted R-squared | 0.068 | 0.070 | |
N = 461
Model 1 is a multivariate linear regression analysis of the association between FOSQ score (dependent variable) and AHI (independent variable), adjusting for age, race, body mass index, physician-diagnosed depression, and self-reported health status. Model 2 adds TST as an independent variable to Model 1 to adjust for potential confounding of the relationship between FOSQ and AHI.
AHI, apnea-hypopnea index; FOSQ, Functional Outcomes of Sleep Questionnaire; TST, total sleep time
Regression results shown represent coefficient point estimates with 95% confidence intervals.
DISCUSSION
This study examined the association between SDB severity and selected functional measures—daytime sleepiness, sleep related symptoms, and sleep related quality of life—in older women, and, for the first time, evaluated in a population-based sample of older women, the extent to which any associations were mediated by short sleep duration. By underscoring the high prevalence of SDB in older adults, this study also highlighted the importance of understanding the morbidity of SDB in this population.
In this population of older women, there was a weak association between TST and measures of SDB severity, including the AHI, as well as an association between TST and most outcome measures. Thus, TST may be a confounder in the association between functional measures and SDB severity, underscoring the importance of considering models with and without adjustment for TST. The fact that the correlation is only weak indicates that the 2 variables are measuring distinct attributes.
Daytime sleepiness, as measured by the ESS, was modestly associated with AHI; however, this association was confounded by TST, which was shorter in individuals with more severe SDB. Overall, the relationship between daytime sleepiness and TST was stronger than the relationship between daytime sleepiness and SDB severity. That being said, even the clinical relevance of a one-point difference in ESS score associated with a 75-minute increase in TST is unclear.
Previous large, population-based studies have shown a relationship between SDB severity and daytime sleepiness. In 1,824 middle-aged and older adults from the Sleep Heart Health Study, Gottlieb et al reported a linear relationship between SDB severity and daytime somnolence, measured by the ESS score.25 A second study of 5,777 older adults similarly demonstrated an association between ESS and both snoring frequency and AHI.10 In a cohort of 718 elderly Japanese American men in the Honolulu-Asia Aging Study of Sleep Apnea, a higher fraction of men with AHI >30 reported excessive daytime somnolence (ESS >10) than those with lower AHI.9 Ancoli-Israel also reported an association between SDB severity and symptoms of sleepiness—such as falling asleep reading while not in bed and falling asleep while in conversation—in a study of 427 community-dwelling elderly subjects.1 Finally, in a study of 4,578 non-institutionalized older adults from the Cardiovascular Health Study, signs and symptoms of SDB (such as loud snoring, awakening with dyspnea or snorting, or frequent awakenings) were associated with ESS scores.11 None of these studies included adjustment for sleep duration.
The finding in this study, that SDB severity, by and large, was not independently associated with ESS score in older women, conflicts with these earlier studies. There are a few possible explanations. First, this study only focused on older women, while previous analyses studied both men and women. Because the age-associated changes in sleep-wake patterns appear to be more pronounced in men than women,12 there may be differences in the association between SDB severity, sleep-wake disturbances, and functional effects between older men and women that are not explored in this study. The ESS score, which has not been validated specifically in older populations, also may have poorer sensitivity and specificity in older populations. However, since older individuals often do not drive, we also explored the utility of a modified ESS score derived by omitting the driving question but did not show any improvement in the association.
By selecting our subjects from among surviving members of the SOF cohort enrolled in 1986, this population was somewhat older than populations examined in previous similar studies. This may limit the generalizability of these findings. On the other hand, although this population of community-dwelling women was, on average, older than the largest population-based study (82.9 vs. 72.4 years), the prevalence of SDB was similar.1 In addition, this analysis produced similar results concerning the association between ESS and SDB before the adjustment for TST.
This population demonstrated, on average, mild-moderate SDB, with a mean AHI of 15.7 ± 15.1, and a different sample population with greater SDB severity may experience greater functional consequences. Regression analyses were designed to minimize this potential bias, but replication of these results in a different sample may prove useful.
A final explanation is that any association between daytime sleepiness (measured by ESS score) and SDB severity may, in part, be mediated by TST. We believe that our inclusion of an objective measure of sleep time may explain some of the differences between our findings and previous studies.
Neither SDB severity nor TST were associated with PSQI scores, and FOSQ scores were only associated with TST in unadjusted analyses. No previous studies have considered the association between PSQI or FOSQ scores and SDB severity in older adults. These findings are not entirely surprising because functional consequences of SDB have shown no consistent association with SDB severity in young- and middle-aged adults.26,27 In addition, weaker associations may be observed in community samples and cohorts of older adults due to a survival bias. Advanced elderly may also experience competitive risk factors, contributing to variation in PSQI and FOSQ scores that are not reflected by measures of SDB severity or TST.
Taken together with the ESS results, these results suggest one of two possibilities: that these measures (ESS, PSQI, and FOSQ scores) are not sufficiently sensitive or specific to detect the functional consequences of SDB in older women, or that SDB in older women may not have the same associated functional consequences seen in young and middle-aged adults. The ESS has obvious problems (such as applicability to individuals who may no longer drive) that limit its application to older adults, but the results were consistent for the modified ESS score that eliminated the question related to driving. It is possible that other subjective measures of sleepiness may have yielded different results. Objective measures of daytime sleepiness, such as the maintenance of wakefulness test or multiple sleep latency test, may have increased sensitivity and specificity in detecting daytime sleepiness, and future investigations may choose to include these in spite of the significant logistical challenges of administering the tests to a population of older women. Although the FOSQ does not carry the same concerns as an evaluation tool for older adults, it has not been validated for this population. However, the PSQI has specifically been validated in a population of older adults.21,22
While sleep disorders are common among the elderly, the impact of sleep disturbances—and the interaction of multiple sleep disturbances—remains to be elucidated. Future work from our group will address these same research questions in a population of older men. However, this study supports the notion that the functional consequences of SDB in older women may be very different than in younger populations.28,29
CONCLUSIONS
In community-dwelling older women, SDB severity was only modestly associated with daytime sleepiness as measured by the ESS. TST was associated with the ESS score in adjusted analyses and with the FOSQ in unadjusted analyses. However, SDB also was weakly correlated with TST, and after accounting for shorter sleep duration observed in individuals with more severe SDB, no independent association between ESS and the AHI or other specific indices of breathing disturbances were observed. Indices of SDB also were not associated with sleep related symptoms (PSQI) or sleep related quality of life (FOSQ) in these elderly women. These results underscore the importance of considering the confounding effects of reduced sleep duration on functional outcomes of interest in studies of SDB in older adults.
ACKNOWLEDGMENTS
Supported by Public Health Service Grants AG05407, AR35582, AG05394, AR35584, AR35583 (Study of Osteoporotic Fractures); and NIA AG08415 (Sonia Ancoli-Israel, PhD). This was not an industry-supported study. There were no off-label or investigational uses of medications or technologies.
Portions of this work were performed at each author's institution.
Parts of this paper were presented at the Associated Professional Sleep Societies 2005 Annual Meeting, Denver, CO
APPENDIX 1
Investigators in the Study of Osteoporotic Fractures Research Group: San Francisco Coordinating Center (California Pacific Medical Center Research Institute and University of California San Francisco): SR Cummings (principal investigator), MC Nevitt (co-investigator), DC Bauer (co-investigator), DM Black (co-investigator), KL Stone (co-investigator), W Browner (co-investigator), R Benard, T Blackwell, PM Cawthon, L Concepcion, M Dockrell, S Ewing, C Fox, R Fullman, SL Harrison, M Jaime-Chavez, L Lui, L Palermo, M Rahorst, D Robertson, C Schambach, R Scott, C Yeung, J Ziarno.
University of Maryland: MC Hochberg (principal investigator), L Makell (clinic coordinator), MA Walsh, B Whitkop.
University of Minnesota: KE Ensrud (principal investigator), S Diem (co-investigator), M Homan (co-investigator), D King (Program Coordinator), N Michels (Clinic Director), S Fillhouer (Clinic Coordinator), C Bird, D Blanks, C Burckhardt, F Imker-Witte, K Jacobson, D King, K Knauth, N Nelson, M Slindee.
University of Pittsburgh: JA Cauley (principal investigator), LH Kuller (co-principal investigator), JM Zmuda (co-investigator), L Harper (project director), L Buck (clinic coordinator), C Bashada, W Bush, D Cusick, A Flaugh, A Githens, M Gorecki, D Moore, M Nasim, C Newman, N Watson.
The Kaiser Permanente Center for Health Research, Portland, Oregon: T Hillier (principal investigator), E Harris (co-investigator), E Orwoll (co-investigator), K Vesco (co-investigator), J Van Marter (project director), M Rix (clinic coordinator), A MacFarlane, K Pedula, J Rizzo, K Snider, T Suvalcu-Constantin, J Wallace.
Footnotes
Disclosure Statement
This was not an industry supported study. Dr. Ancoli-Israel has received research support from Sepracor and Takeda; has participated in speaking engagements for Cephalon, King Pharmaceuticals, Neurocrine Biosciences, Pfizer, Sanofi-Aventis, Sepracor, and Takeda; and has consulted to or been on the advisory board of Acadia, Cephalon, GlaxoSmithKline, Merck, Neurocrine Biosciences, Neurogen, Pfizer, Sanofi-Aventis, Sepracor, and Takeda. Dr. Redline is a co-investigator in a NIH SBIR funded grant to Cleveland Medical Devices, Inc, and has served as a scientific advisor for Cypress Bio. And Organon. Dr. Kezirian, Ensrud, Claman, Stone, and Ms. Harrison have indicated no financial conflicts of interest.
REFERENCES
- 1.Ancoli-Israel S, Kripke DF, Klauber MR, Mason WJ, Fell R, Kaplan O. Sleep-disordered breathing in community-dwelling elderly. Sleep. 1991;4:486–95. doi: 10.1093/sleep/14.6.486. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med. 1993;328:1230–5. doi: 10.1056/NEJM199304293281704. [DOI] [PubMed] [Google Scholar]
- 3.Bixler EO, Vgontzas AN, Lin HM, et al. Prevalence of sleep-disordered breathing in women: effects of gender. Am J Respir Crit Care Med. 2001;163:608–13. doi: 10.1164/ajrccm.163.3.9911064. [DOI] [PubMed] [Google Scholar]
- 4.Bixler EO, Vgontzas AN, Ten Have T, Tyson K, Kales A. Effects of age on sleep apnea in men: I. Prevalence and severity. Am J Respir Crit Care Med. 1998;157:144–8. doi: 10.1164/ajrccm.157.1.9706079. [DOI] [PubMed] [Google Scholar]
- 5.Duran J, Esnaola S, Rubio R, Iztueta A. Obstructive sleep apnea-hypopnea and related clinical features in a population-based sample of subjects aged 30 to 70 yr. Am J Respir Crit Care Med. 2001;163:685–9. doi: 10.1164/ajrccm.163.3.2005065. [DOI] [PubMed] [Google Scholar]
- 6.Redline S, Tosteson T, Tishler PV, Carskadon MA, Millman RP. Studies in the genetics of obstructive sleep apnea. Familial aggregation of symptoms associated with sleep-related breathing disturbances. Am Rev Respir Dis. 1992;145:440–4. doi: 10.1164/ajrccm/145.2_Pt_1.440. [DOI] [PubMed] [Google Scholar]
- 7.Young T, Shahar E, Nieto FJ, et al. Predictors of sleep-disordered breathing in community-dwelling adults: the Sleep Heart Health Study. Arch Intern Med. 2002;162:893–900. doi: 10.1001/archinte.162.8.893. [DOI] [PubMed] [Google Scholar]
- 8.Flemons WW, Tsai W. Quality of life consequences of sleep-disordered breathing. J Allergy Clin Immunol. 1997;99:S750–6. doi: 10.1016/s0091-6749(97)70123-4. [DOI] [PubMed] [Google Scholar]
- 9.Foley DJ, Masaki K, White L, Larkin EK, Monjan A, Redline S. Sleep-disordered breathing and cognitive impairment in elderly Japanese-American men. Sleep. 2003;26:596–9. doi: 10.1093/sleep/26.5.596. [DOI] [PubMed] [Google Scholar]
- 10.Gottlieb DJ, Yao Q, Redline S, Ali T, Mahowald MW. Does snoring predict sleepiness independently of apnea and hypopnea frequency? Am J Respir Crit Care Med. 2000;162:1512–7. doi: 10.1164/ajrccm.162.4.9911073. [DOI] [PubMed] [Google Scholar]
- 11.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. doi: 10.1093/sleep/21.1.27. [DOI] [PubMed] [Google Scholar]
- 12.Redline S, Kirchner HL, Quan SF, Gottlieb DJ, Kapur V, Newman A. The effects of age, sex, ethnicity, and sleep-disordered breathing on sleep architecture. Arch Intern Med. 2004;164:406–18. doi: 10.1001/archinte.164.4.406. [DOI] [PubMed] [Google Scholar]
- 13.Ohayon MM, Vecchierini MF. Normative sleep data, cognitive function and daily living activities in older adults in the community. Sleep. 2005;28:981–9. [PubMed] [Google Scholar]
- 14.Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals: developing normative sleep values across the human lifespan. Sleep. 2004;27:1255–73. doi: 10.1093/sleep/27.7.1255. [DOI] [PubMed] [Google Scholar]
- 15.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. [PubMed] [Google Scholar]
- 16.Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Los Angeles: UCLA Brain Information Service/Brain Research Institute; 1968. [Google Scholar]
- 17.EEG arousals: scoring rules and examples: a preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep. 1992;15:173–84. [PubMed] [Google Scholar]
- 18.Ancoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W, Pollak CP. The role of actigraphy in the study of sleep and circadian rhythms. Sleep. 2003;26:342–92. doi: 10.1093/sleep/26.3.342. [DOI] [PubMed] [Google Scholar]
- 19.Jean-Louis G, Kripke DF, Mason WJ, Elliott JA, Youngstedt SD. Sleep estimation from wrist movement quantified by different actigraphic modalities. J Neurosci Methods. 2001;105:185–91. doi: 10.1016/s0165-0270(00)00364-2. [DOI] [PubMed] [Google Scholar]
- 20.Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14:540–5. doi: 10.1093/sleep/14.6.540. [DOI] [PubMed] [Google Scholar]
- 21.Cole JC, Motivala SJ, Buysse DJ, Oxman MN, Levin MJ, Irwin MR. Validation of a 3-factor scoring model for the Pittsburgh sleep quality index in older adults. Sleep. 2006;29:112–6. doi: 10.1093/sleep/29.1.112. [DOI] [PubMed] [Google Scholar]
- 22.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;56:503–10. doi: 10.1016/S0022-3999(04)00023-6. [DOI] [PubMed] [Google Scholar]
- 23.Buysse DJ, Reynolds CF, 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28:193–213. doi: 10.1016/0165-1781(89)90047-4. [DOI] [PubMed] [Google Scholar]
- 24.Weaver TE, Laizner AM, Evans LK, et al. An instrument to measure functional status outcomes for disorders of excessive sleepiness. Sleep. 1997;20:835–43. [PubMed] [Google Scholar]
- 25.Gottlieb DJ, Whitney CW, Bonekat WH, et al. Relation of sleepiness to respiratory disturbance index: the Sleep Heart Health Study. Am J Respir Crit Care Med. 1999;159:502–7. doi: 10.1164/ajrccm.159.2.9804051. [DOI] [PubMed] [Google Scholar]
- 26.Weaver EM, Kapur V, Yueh B. Polysomnography vs self-reported measures in patients with sleep apnea. Arch Otolaryngol Head Neck Surg. 2004;130:453–8. doi: 10.1001/archotol.130.4.453. [DOI] [PubMed] [Google Scholar]
- 27.Weaver EM, Woodson BT, Steward DL. Polysomnography indexes are discordant with quality of life, symptoms, and reaction times in sleep apnea patients. Otolaryngol Head Neck Surg. 2005;132:255–62. doi: 10.1016/j.otohns.2004.11.001. [DOI] [PubMed] [Google Scholar]
- 28.Ancoli-Israel S, Coy T. Are breathing disturbances in elderly equivalent to sleep apnea syndrome? Sleep. 1994;17:77–83. doi: 10.1093/sleep/17.1.77. [DOI] [PubMed] [Google Scholar]
- 29.Young T. Sleep-disordered breathing in older adults: is it a condition distinct from that in middle-aged adults? Sleep. 1996;19:529–30. doi: 10.1093/sleep/19.7.529. [DOI] [PubMed] [Google Scholar]