| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
1 Department of Neurology, Sleep Disorders Program, Vanderbilt University Medical Center, Nashville, Tennessee.
2 Department of Psychiatry, University of California, San Diego.
Departments of 3 Epidemiology and 4 Psychiatry, University of Pittsburgh, Pennsylvania.
5 San Francisco Coordinating Center and California Pacific Medical Center Research Institute, San Francisco.
6 Department of Epidemiology and Biostatistics, University of California, San Francisco.
7 Department of Preventive Medicine, University of Tennessee, Memphis.
8 Clinical Research Branch, National Institute on Aging, Baltimore, Maryland.
Address correspondence to Suzanne E. Goldman, PhD, Vanderbilt University Medical Center, Department of Neurology, Sleep Disorders Program, 1301 Medical Center Drive, Room B-727, Nashville, TN 37232. E-mail: suzanne.e.goldman{at}vanderbilt.edu
|
A |
|---|
|
|
|---|
Methods. We studied 2264 men and women, aged 75–84 years (mean 77.5 years; standard deviation [SD] 2.9), participating in the Year 5 (2001–2002) clinic visit of the Health, Aging, and Body Composition (Health ABC) study. Fatigue was determined using a subscale of the Modified Piper Fatigue Scale (0–50; higher score indicating higher fatigue). Hours of sleep per night, trouble falling asleep, waking up during the night, and waking up too early in the morning were assessed using interviewer-administered questionnaires.
Results. The average fatigue score was 17.7 (SD 8.4). In multivariate models, women had a 3.8% higher fatigue score than men did. Individuals who slept
6 hours/night had a 4.3% higher fatigue score than did those who slept 7 hours/night. Individuals with complaints of awakening too early in the morning had a 5.5% higher fatigue score than did those without these complaints. These associations remained significant after multivariate adjustment for multiple medical conditions.
Conclusion. The association between self-reported short sleep duration (
6 hours), and waking up too early and fatigue symptoms suggests that better and more effective management of sleep behaviors may help reduce fatigue in older adults.
Key Words: Sleep Fatigue Aging
Insomnia and other sleep disorders are also common in older adults, with prevalence rates estimated as high as 50% (15,16). These also can result in decreased life quality (17). Although sleep complaints and fatigue are thought to be related, this association has not been evaluated with consideration of the contribution of comorbidity in community-dwelling older adults. This study assessed the prevalence of fatigue in community-dwelling older adults and hypothesized that fewer hours of sleep and/or difficulty with initiating or maintaining sleep would be associated with higher fatigue scores.
| METHODS |
|---|
|
|
|---|
Assessment of Sleep Behaviors and Fatigue
At the year 5 clinic visit, standardized questionnaires that included detailed questions about sleep, fatigue, medical history, and physical activity were administered by trained study personnel. Sleep and napping behavior were assessed by the following questions: "How many hours of sleep do you usually get at night during a usual week?" and "How many times a week do you nap for 5 minutes or more?" Insomnia symptoms were evaluated with a series of questions in which participants were asked how often they experienced: (i) trouble falling asleep, (ii) waking up during the night and having difficulty getting back to sleep, (iii) waking up too early in the morning and being unable to get back to sleep, and (iv) taking sleeping pills or other medication to help sleep. Responses were categorized as: never, rarely (
1 time/month), sometimes (2–4 times/month), often (5–15 times/month), or almost always (16–30 times/month) (18,19).
Fatigue was measured with a subscale from the Revised Piper Fatigue Scale, a scale originally developed for use in breast cancer patients (12). This scale has also been used in a group of older individuals residing in a long-term care facility (20). Participants were asked the following questions in reference to the past month: (i) How weak did you feel? (ii) How sleepy did you feel during the day? (iii) How lively did you feel? (iv) How tired did you feel? and (v) What was your usual energy level? Responses could range from 0 to 10 with 10 indicating the most severe level. A total fatigue score ranging from 0 (no fatigue) to 50 (highest fatigue) was obtained by summing the scores for each item after reversing the scoring for the questions on liveliness and energy level as in the Revised Piper Fatigue Scale (12).
Assessment of Health Conditions
Information on health conditions was based on participant self-report at the time of the year 5 clinic visit. Self-reported health status was assessed with the question "In general, how would you say your health is?" with response options "excellent," "very good," "fair," "poor," or "don't know." Cardiovascular disease was assessed as self-report of congestive heart failure, coronary heart disease, or stroke. Self-reported dyspnea on exertion was classified separately. Incident cancer was based on adjudicated events occurring before the year 5 visit. Depression was assessed with the 10-item Center for Epidemiologic Studies Depression Scale. Physical activity was assessed by self-report and calculated as the total kilocalories per week engaged in walking and climbing stairs (21). Anthropometric measurements, including height measured with a stadiometer and weight measured with a balance beam scale were also obtained. Body mass index (BMI; kg/m2) was calculated from measured height and weight. Sex and race (black or white) were self-designated at baseline.
Statistical Analysis
Descriptive statistics were performed on all variables to evaluate ranges, frequencies, normalities, and inconsistencies in the data. Differences between race and sex were tested using chi-square tests and Student t tests. Age, race, BMI, depression, self-reported health status, cardiopulmonary disease, cancer, and physical activity were all considered as possible confounders of the association between sleep and fatigue and were treated as covariates (13,22–24).
For analysis purposes, based on the distribution of the data, hours of sleep were categorized as
6, 7, 8, and >8 hours. Insomnia symptoms of trouble falling asleep, waking up during the night, waking up too early in the morning, and taking sleeping medicines were collapsed into two categories: "infrequent" consisting of never, rarely, or sometimes; and "frequent" consisting of experiencing at least one symptom often, or almost always (25,26).
Potential associations between the individual fatigue scale questions, hours slept per night, and the individual insomnia symptoms were examined using Spearman rank order correlation coefficients. The sum of the five fatigue questions was then used in subsequent regression models.
The association between the total fatigue score and each sleep variable was examined using a series of linear regression models. Additional factors found significant in univariate analysis were included using progressively complex multivariable models. Model 1 included each sleep variable adjusted for demographic factors, age, race, and sex. Model 2 added health-related variables, BMI, depressive symptom score, self-reported health status, cardiopulmonary factors, and dyspnea to the variables in Model 1. Results were similar to Model 1 and therefore are not presented. Model 3 added kilocalories per kilogram per week expended walking and climbing stairs to the variables in Model 2. To express the strength of the associations, percent differences were calculated from the regression coefficients with the formula (β x unit / mean fatigue score) (27). Factors found significantly associated with fatigue are reported at p
.05. Analyses were performed using SAS (versions 8.2 and 9.0; SAS Institute, Cary, NC) and STATA (version 8; STATA Corporation, College Station, TX).
| RESULTS |
|---|
|
|
|---|
|
6 hours) and long sleep duration (>8 hours) were associated with higher levels of fatigue relative to sleep duration of 7 hours. Higher levels of fatigue were also associated with trouble falling asleep, waking up during the night, waking up too early, and use of sleep medications (Table 3). Results were similar when all sleep variables were considered together in the same model.
|
|
6 hours) and waking up too early in the morning remained associated with higher fatigue scores, but were attenuated after adjustment. Also associated with higher average fatigue scores were BMI, poorer self-reported health status, high depressive symptoms, prevalent cardiovascular disease, dyspnea, and secondary primary cancer. These associations remained significant, but were also partially attenuated in the fully adjusted model. Individuals who engaged in walking or climbing stairs equivalent to >5.3 kcal/kg/wk had lower fatigue scores than individuals with
0.03 kcal/kg/wk of activity. The final multivariable model explained 41% of the overall variance in fatigue score. Finally, factors associated with high levels of fatigue (score >30) were examined to determine if any were specifically related to these higher levels. Results were consistent with the results of the linear model. | DISCUSSION |
|---|
|
|
|---|
Gender and race were both associated with fatigue. The finding that women had higher fatigue scores than men differed from the one other study of fatigue in older adults where no difference was found in fatigue scores between genders (20). However, the Health ABC cohort was younger (77.5 [2.9] vs 87.8 [4.9] years of age), had a higher percentage of men (47.7% vs 18%), and was considerably larger (2264 vs 308 participants). In the general population, women have been found to have higher levels of fatigue than men (28–30). Furthermore, the gender association would be consistent with similar syndromes such as daytime sleepiness and insomnia, where women tend to report higher levels of these symptoms than do men (15,22,31–34). The association of blacks having lower fatigue scores than whites was not expected. Previous research in other populations has reported that blacks had higher fatigue scores than whites (28,35,36). Whereas the differences might be attributable to different population characteristics such as age, or fatigue scale use, this association was found after we adjusted for possible confounders. This finding is interesting and warrants future research.
The definition and pathophysiology of fatigue are not clear. Our study supports other research showing associations between fatigue, disturbed sleep, and medical conditions (2,3,5,9,30,37). It is difficult to fully differentiate fatigue from sleepiness because sleepiness is often used to define fatigue (7). Removal of the question "How sleepy are you during the day?" from the total fatigue score did not change the associations found between disturbed sleep and total fatigue in this study. Fatigue has been associated with a wide range of sleep disorders and behaviors; however, the causes of fatigue in sleep-disordered populations is relatively unknown (3,22). In patients referred to sleep clinics, subjective fatigue has been shown to be independent of sleep disorder severity and daytime sleepiness (3,22). Results of this study suggest an independent association of self-reported sleep duration, and waking up too early and fatigue. Long and short sleep durations have been associated with increased mortality (38–41).
The direct associations of chronic health conditions with fatigue, as well as with sleep, in this population were significant and cannot be disregarded. Fatigue in chronic illness is pervasive and multidimensional, with different perceived causes and implications (4). In this study, self-reported health status, cardiopulmonary disease, dyspnea, depression, and recurrent or secondary cancer were strong associates of fatigue, even after adjustment for other variables. These associations were consistent with known associations of fatigue and medical comorbidities reported in other populations (2,13,20). To what extent chronic disease might cause fatigue, or chronic disease might cause disturbed sleep (which in turn causes fatigue), is not well defined even in cancer or in sleep apnea patients. Results from this study suggest that while chronic health conditions are highly associated with fatigue, poor sleep may have an independent contribution and warrants further investigation.
Fatigue has been considered in the context of cancer (5,6,12), or other chronic diseases (42,43), although it impacts other populations as well. Inclusion of incident cancer in the models in this study slightly attenuated the associations between fatigue and sleep. However, it did not change the overall significance of the associations. Although fatigue is a common complaint in older adults, it has rarely been addressed as a specific outcome in this population. Studies performed in younger cohorts have reported fatigue symptoms in 12%–25% of the population (8,36). In one group of older adults living in an assisted living facility, more than 50% of these individuals exhibited at least some complaint of mild fatigue (20). In addition to the dissimilar population groups assessed, the use of multiple, nonstandardized fatigue scales makes it difficult to compare the fatigue rates found in the Health ABC cohort with those of these other studies. However, more than 37% of the Health ABC cohort had a fatigue score
20 that would indicate a mild-moderate fatigue level.
This study had several limitations. The cross-sectional nature of the data limits evaluation of the temporal relationship between sleep and fatigue. Fatigue is a subjective syndrome, and there is no gold standard to assess it. Various other fatigue scales, or components of scales, currently in use may provide different correlates of fatigue (22). Severity of disease and presence of other clinical conditions that might be associated with fatigue were not evaluated. Usual sleep time and sleep behaviors were obtained by self-report. Although the validity of self-report sleep data has been demonstrated with actigraphy, some variability between actual and self-report exists. Use of self-report data also precluded the measurement of sleep disorders obtained through polysomnography such as apneas, hypopneas, and periodic limb movement disorders that might also be related to complaints of poor sleep. Finally, it is important to keep in mind that the Health ABC cohort members were well-functioning at baseline, so they may be healthier than other older populations.
Conclusion
The results of this study of well-functioning, community-dwelling older adults suggest that fatigue may be associated with several dimensions of sleep including short sleep duration and disrupted sleep. Although part of this association may be explained by poorer self-reported health status, these observed relationships were independent of other health-related conditions commonly associated with fatigue. Additional studies are needed to help understand the direction of these relationships.
|
A |
|---|
|
|
|---|
CONFLICTS OF INTEREST (FINANCIAL DISCLOSURES)
Sonia Ancoli-Israel is a member of the Advisory Board and/or has participated in speaking engagements supported by Sepracor, Takeda Pharmaceuticals, King, Sanofi-Aventis, Cephalon, Merck, and Neurocrine Biosciences.
Jane A. Cauley receives research funds from Merck & Co, Eli Lilly & Co, Pfizer Pharmaceuticals, and Novartis Pharmaceuticals. She receives honoraria from Merck & Co, and Eli Lilly & Co, and has participated in the speaker's bureau of Merck & Co.
|
F |
|---|
|
|
|---|
Received October 23, 2007
Accepted January 14, 2008
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|