Poor Sleep Quality is Independently Associated with Physical Disability in Older Adults

Abstract

Study Objective:

We aimed to evaluate the association between sleep quality and physical disability in community-dwelling older adults.

Methods:

There were 213 community-dwelling adults (76 men and 137 women) aged 65 years and above participated into this investigation. The Groningen Activity Restriction Scale and the Pittsburgh Sleep Quality Index were utilized to evaluate physical disability and subjective sleep quality, respectively. Global functional capacity was measured by the 6-minute walk test (6MWT). The Mini Mental State Examination and the Chinese Geriatric Depression Screening Scale were used to evaluate cognitive function and depression.

Results:

Univariate analysis revealed a correlation between physical disability and poor sleep quality, older age, 2 or more comorbidities, depression, functional capacity, and poor cognitive function. However, in the multivariate analyses, depression failed to show significant association with physical disability. In contrast, an independent association was observed between poor sleep quality and physical disability (OR = 2.03; 95% CI: 1.02–4.05).

Conclusions:

In community-dwelling older adults, subjective poor sleep was significantly associated with physical disability, even after controlling for the effects of other established risk factors.

Citation:

Chien MY, Chen HC. Poor sleep quality is independently associated with physical disability in older adults. J Clin Sleep Med 2015;11(3):225–232.

Physical disability, defined as the difficulty in performing activities necessary for independent living, and loss of independence, is a common feature in an aging society with an increasing life expectancy.^1,2 Operationally, the concept of disability involves several dimensions, including basic activities of daily life (ADL), instrumental ADL (IADL), role and social activities, leisure activities, and physical function.^3,4 Although labeled as “physical” disability, the mental component (such as cognitive function and mood status) emphasizes the distinction between the concepts of physical disability and impaired physical performance.⁵ People with old age are reported to have a higher risk of developing disability^6,7; therefore, physical disability in older adults has become a major public health concern as the percentage of older persons increases. Owing to the prevalence and complexity of this issue, it has become increasingly important to understand the related factors of physical disability in older adults.

Several crucial determinants of physical disability in older adults have been identified in the literature. From a physical point of view, functional fitness is closely related to functional disability. Along with normal aging, decreased skeletal muscles, declined maximal strength, and compromised functional capacity lead to weakness, resulting in physical disability.^8,9 From a mental perspective, depression, anxiety, and impaired cognitive function are considered correlates of physical disability in older adults.^10–17 In addition to these well-established determinants of physical disability, sleep disturbance is another important issue deserving further exploration.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Sleep disturbance has been a prevalent health problem in older adults. However, the relationship between sleep disturbance and physical disability in older adults is yet to be fully elucidated.

Study Impact: Subjective poor sleep quality was independently associated with physical disability. Older adults with poor sleep quality deserve clinical attention regarding physical disability.

Sleep disturbance has been a prevalent health problem in older adults. An average of up to 50% of community-dwelling older adults suffer from various degrees of sleep problems.¹⁸ Sleep disturbance affects health outcomes in older adults, such as depressive symptoms, morbidity, and mortality.^19–26 Many studies have reported that sleep disturbance is associated with functional disability in patients with chronic diseases.^27,28 However, few studies have evaluated the relationship between sleep disturbances and physical disability in older adults. Furthermore, the existing findings on this issue are inconsistent.

Some studies have demonstrated a relationship between poor subjective^29,30 and objective³¹ sleep quality with increased physical disability in older adults. In contrast, one study failed to link insomnia with physical disability.³² This discrepancy in findings may be a result of different study settings, varying definitions of sleep disturbance, and an inadequate control of important confounders, such as functional capacity, depression, and cognitive function.^29,30,32,33 Therefore, the relationship between sleep disturbance and physical disability in older adults is yet to be fully elucidated. Within the context of the aging society, it is important to gain an understanding of such an association. The present study aims to examine if the relationship between poor sleep quality and physical disability in older adults remains even after accounting for the effect of important physical and mental confounders.

METHODS

Setting and Participants

Poster campaigns were used to recruit volunteers from communities in the Zhongzheng district of Taipei, Taiwan. Volunteers were ineligible to participate if they had any condition that prevented them from completing our questionnaires or if they did not pass the functional capacity test (e.g., exhibited conditions such as amputation, non-ambulation even with the use of a walker or a cane, visual impairment, or hearing impairment). Volunteers who took prescribed antidepressants were also excluded from the study, due to the effects of antidepressants on sleep.³⁴ Considering the risk of co-linearity with depression, and the fact that there were very few subjects in this category, it was not included as a covariate. Our sample consisted of 236 community-dwelling older adults (aged ≥ 65 years), who voluntarily contacted us during the 2-week duration of the survey. Of these, 23 (9.7%) volunteers were ineligible, including 7 individuals who could not walk despite the use of a walker or cane, 4 individuals who were hearing impaired, 5 individuals with dementia who could not complete the questionnaires, and 7 individuals who were taking antidepressant medication. Trained physical therapists collected the data according to standard protocols. Participants provided written informed consent prior to participation. The study was approved by the institutional ethics committee (NCT01689818).

Participants were interviewed using a structured questionnaire designed to elicit basic information and information on major medical conditions, i.e., conditions requiring a formal diagnoses and long-term treatment, such as diabetes, hyper-tension, cardiovascular disease (myocardial infarction, coronary heart diseases, and congestive heart failure), pulmonary diseases, gastrointestinal diseases, orthopedic problems, benign prostate hyperplasia, and cataracts. The sum of all chronic medical conditions present was combined to provide a comorbidity index. The comorbidity index was included in the multivariate analysis instead of individual medical illnesses. A record was kept of volunteers who were prescribed other sleep-altering drugs, such as hypnotics, steroids, or antihistamines. Assessments included anthropometric measures as well as assessment of physical disability, functional capacity, sleep quality, cognitive function, and depression.

Body weight and height were measured with participants dressed in light clothing and barefoot. Body mass index (BMI) was also calculated and categorized according to the Taiwanese Bureau of Health Promotion standards: < 18.5 = under-weight; 18.5–23.9 = normal weight; 24.0–26.9 = overweight; and > 27 = obese. Waist circumference was measured by using a flexible plastic tape measure. According to Taiwanese criteria, obesity was defined as a waist circumference ≥ 80 cm for women and 90 cm for men. Regular exercise behavior was defined by the frequency and duration of a participant's exercise activities. Exercising ≥ 1 h on ≥ 3 days per week was considered regular exercise behavior.

Evaluation of Physical Disability

The Groningen Activity Restriction Scale (GARS) was utilized to assess physical disability.³⁵ The GARS is a one-dimensional hierarchical scale used to grade the difficulties a person might experience when performing ADL without assistance. The GARS questionnaire contains 18 items that assess daily activities on the basis of 1 of 4 response options ranging from 1 = complete independence to 4 = total lack of independence. The total score (ranging from 18 to 72) provides a measure of the respondents' ability to care for themselves and perform household activities. Previous studies have reported GARS to exhibit good reliability and validity.^35,36 It was shown that comparisons between the GARS and the SF-20 subscale for physical functioning has yielded a strong correlation (r = −0.72), supporting concurrent validity.³⁵ In this study, physical disability was defined if one reported having any difficulty in performing ≥ 1 activity (total score ≥ 19).

Evaluation of Sleep Quality

Sleep quality was measured with the Chinese version of the Pittsburgh Sleep Quality Index (CPSQI).³⁷ The sum of the 7 component scores serves as the global PSQI score (ranging from 0 to 21). Scores > 5 are indicative of poor sleep quality.³⁷ The PSQI is a widely used self-rated questionnaire for measuring sleep quality, with good reliability and validity.^37,38 The CPSQI had an overall reliability coefficient of 0.82–0.83, and a CPSQI > 5 resulted in a sensitivity and specificity of 98% and 55%.³⁷

Evaluation of Functional Capacity and Cognitive Function

The 6-minute walk test (6MWT) was conducted to evaluate functional capacity. This self-paced test evaluated each participant's global functional capacity in distance over time, using a standardized protocol in accordance with the guidelines of the American Thoracic Society.³⁹ Participants were instructed to walk as far as possible for 6 minutes, and the distances they walked were recorded to the nearest meter. Pulse rate and oxygen saturation were monitored continuously and simultaneously. Owing to the relatively optimal scores of the participants on the health parameters, the performance on the 6MWT was categorized by quartiles, to explore the potential non-linear relationship with physical disability.

In addition, the Chinese version of the Mini-Mental State Examination (MMSE) was used to evaluate cognitive function.⁴⁰ The MMSE is a 19-item cognitive screening instrument with a potential score of 0 to 30,^40,41 with higher scores indicating better cognitive functioning. The MMSE is reported to exhibit good reliability and validity, and is widely used in the assessment of older adults.⁴¹ As volunteers with dementia were excluded from the sample, the original score on the MMSE was used as a measure of cognitive function.

Evaluation of Depression

Depression was assessed by using a validated Chinese version of the short form of the Geriatric Depression Scale (GDS).⁴² This tool focuses on non-somatic symptoms of depression and emphasizes on cognitive symptoms. The GDS includes 15 yes/ no self-report questions, and takes approximately 10 minutes to complete, where the participant is expected to circle the appropriate responses. A score ≥ 5 on this tool is indicative of depression.⁴² The overall sensitivity and specificity of GDS are 0.97 and 0.95, and intraclass coefficient of test-retest reliability over 2 weeks is 0.83.⁴³

Statistical Analysis

The statistical analyses were performed using the SPSS for Windows, version 13.0 (SPSS Inc., Chicago, IL, USA). In addition to univariate analyses using the χ² test and the Student t-test, logistic regression analysis was conducted to examine the independent relationship between physical disability and poor sleep quality after controlling for confounders. Only the variables found to be significantly associated with physical disability in the univariate analyses were included in the multiple logistic regression models. A p value < 0.05 was considered statistically significant.

RESULTS

Table 1 and Table 2 summarize the basic characteristics, anthropometric variables as well as clinical features of all participants. A total of 213 participants (76 men and 127 women) completed all physical assessments and were included in data analysis. The average age of the participants was 77.9 ± 7.3 years; a majority of them were in the “overweight” category, with mean BMI of 24.9 ± 3.7 kg/m². A total of 61.5% of the participants were aged ≥ 75 years, and 60.1% of them had ≥ 2 medical illnesses. There were 36 (16.9%) participants reported to use hypnotics, but no participant used antihistamines or long-term steroids during the past month. Overall, the participants scored high on the MMSE (24.8 ± 4.3) and low on the GARS (21.7 ± 6.2). Two-thirds of participants scored ≥ 19 on the GARS and were defined as physically disabled. About half of the participants (52.1%) reported poor sleep quality, as defined by the PSQI.

Table 1.

Characteristics of all participants (N = 213).

Table 2.

Univariate analysis of characteristics of participants by physical disability (N = 213).

Table 2 displays the univariate analysis of physical disability. Older adults aged ≥ 75 years with ≥ 2 comorbidities, performed least well on the 6MWT, had PSQI scores > 5 and GDS scores ≥ 5 had significantly higher occurrence of physical disability than their counterparts. Further, participants who exhibited a poor performance on the 6MWT and the MMSE exhibited a higher likelihood of having a physical disability. Similarly, participants with PSQI scores > 5 were found to have a higher occurrence of depression (χ² = 22.2, df = 1, p < 0.001), had ≥ 2 comorbidities (χ² = 5.4, df = 1, p = 0.02), and had a poor functional capacity (χ² for trend test = 6.43, df = 1, p = 0.01; data not shown in the table). Specifically, 81.6% of the depressed older adults scored > 5 on the PSQI, as compared to the non-depressed subjects participants (36%).

According to the findings in Table 2, age, number of comorbidities, functional capacity (as measured by the 6MWT), poor sleep quality (as assessed by the PSQI), depression (as measured by the GDS) and cognitive function (as measured by the MMSE), were considered as confounders with respect to the relationship between poor sleep quality and physical disability. Therefore, these variables were included in the analysis of logistic regression when examining the independent relationship between poor sleep quality and physical disability. The univariate logistic regression revealed that physical disability was significantly associated with older age (OR: 2.81; 95% CI: 1.56–5.06; p = 0.001), higher numbers of comorbidities (OR: 2.54; 95% CI: 1.42–4.57; p = 0.002), third (OR: 2.47; 95% CI: 1.12–5.46; p = 0.03), second (OR: 2.54; 95% CI: 1.15–5.61; p = 0.02), and lowest (OR: 5.83; 95% CI: 2.37–14.34; p < 0.001) quartile of the distance walked on the 6MWT, poor sleep quality (OR: 2.36; 95% CI: 1.32–4.23; p = 0.004), depression (OR: 2.70; 95% CI: 1.23–5.95; p = 0.01), and lower scores on the MMSE (OR: 0.85; 95% CI: 0.78–0.92; p < 0.001). In the multiple regression analyses, the results showed that, except for depression, second and third quartile of 6MWT, all the other variables found to be significant in the univariate logistic analysis remained significantly associated with physical disability, although we observed a decrease in the effect magnitude (Table 3). Specifically, again the older adults with poor sleep quality were found to have a higher likelihood of suffering from a physical disability (OR: 2.03; 95% CI: 1.02–4.05; p = 0.04), even when the effect of age, functional capacity, depression, and cognitive function had been accounted for. Furthermore, interaction terms, including PSQI × GDS and PSQI × 6MWT, were included in the multiple logistic models, to examine the interaction effects. However, we did not find any significant interaction effects with regard to sleep quality and depression or functional capacity. Finally, when poor sleep quality was excluded from the multiple logistic regression model, depression and physical disability showed a relationship with borderline statistical significance (OR: 2.13; 95% CI: 0.90–5.00; p = 0.08).

Table 3.

Logistic regression for factors associated with physical disability (Groningen Activity Restriction Scale ≥ 19).

DISCUSSION

The present study found that about half of our community-dwelling older adults reported poor sleep quality, as defined by the PSQI. The univariate analysis revealed that older age, poor sleep quality, compromised functional capacity, depression, and worse cognitive function were associated with physical disability. The significant relationship between poor sleep quality and physical disability was also observed after controlling for the other covariates. Physical disability comprises both physical and mental constructs that are closely related to sleep disturbance. Therefore, the present study stands out for its simultaneous inclusion of crucial confounders. The findings of the present study add to the literature on determinants of physical disability in older adults by emphasizing on the independent effect of poor sleep quality, a factor which had been overlooked by previous studies.

The prevalence of sleep disturbance varies depending on the stringency of its definition.¹⁸ It was clear that the PSQI-defined sleep disturbance was more prevalent in our participants (52.1%) than that reported in other studies, which utilized a comparative or more stringent definition of sleep disturbance.^26,32,44 Further, it is interesting to note that the prevalence rate reported in the present study is similar to that reported in studies that evaluated sleep complaints only, i.e., without quantifiers.¹⁸ It may be conjectured that the higher age of the participants and the higher proportion of female participants may have contributed to this observation.

Though we reported an association between sleep disturbance and physical disability, a closer inspection on how the definition of sleep disturbance influences this observation is interesting. Some large-scale studies have reported that objectively measured poor sleep in the older adults is associated with worse physical function.^45,46 Self-reported long sleep duration and excessive daytime sleepiness have also been reported to be associated with declined physical function and activity restriction.^47–50 In the present study, the PSQI-defined sleep disturbance was found to be independently associated with physical disability. This finding was in concordance with previous studies that evaluated sleep disturbance with the PSQI.^29,30 Conversely, a study that evaluated sleep disturbance with a more stringent definition did not find an association.³² Differential relationships between various definitions of sleep disturbance with mortality have also been reported. Specifically, the PSQI-defined sleep disturbance, rather than the DSM-IV defined sleep disorder, predicted elevated mortality risk. It has been speculated that the PSQI-defined sleep disturbance reflects a more recent stress response to physiological or psychological events. In contrast, sleep disturbance defined by DSM-IV criteria for insomnia disorder reflects a long-term sleep problem, and a better outcome on physical disability may reflect an effect of selective survival in the study cohort.²⁶ In addition, the PSQI-defined poor sleep quality is not equivalent to “insomnia,” nor is it correlated with any polysomnographic parameters.^38,51 As sleep disturbance is prevalent in older adults and may just reflect the process of normal aging, identifying high-risk subjects who are vulnerable to adverse health outcomes is of clinical importance. Thus, our findings specifically suggest that older adults with the PSQI-defined poor sleep quality deserve clinical attention regarding physical disability.

Several mechanisms link poor sleep quality with physical disability. Poor sleep quality may lead to physical disability through features of sleep disturbance (impaired attention and fatigue), through predisposition to physical and mental illness (diabetes and depression), or through sleep-related maladaptive behaviors (daytime napping). In contrast, the association between poor sleep quality and physical disability may be spuriously confounded by external/underlying factors such as preclinical dementia and medical illness, which result in both sleep disturbance and physical disability. In the present study, although the nature of the cross-sectional design prevented us to disentangle the “causal” and “temporal” relationship, we have adequately addressed several established physical and mental confounders. However, the simultaneous inclusion of depression and poor sleep quality into the statistical model in the present study may be questioned. Owing to their known bidirectional relationship, poor sleep quality could be a result of depression, due to which the current analytical strategy may not have been appropriate. In fact, even when poor sleep quality was removed from the statistical model, the association between depression and physical disability remained weak. This indicated that depression might have played a minor role with regard to physical disability in the present sample.

Depression is another prevalent mental health problem in the older adults. Geriatric depression has been reported to be correlated with physical disability,¹¹ and is even considered a predictor of subsequent physical disability.^10,11,52 In the present study, 23% of the participants reported depression. This prevalence rate was higher than that reported in other community surveys in Taiwan (9.8% to 21.2%).^19,20 The higher proportion of female and older participants in the study sample may explain this observation.¹⁹ Nevertheless, due to the close relationship between depression and sleep disturbance, the influence of sleep disturbance cannot be overlooked while exploring the link between depression and physical disability.

Controlling for cognitive function and physical fitness is another feature of the present study. Numerous studies had demonstrated the association between declined cognitive function and physical disability, including ADL, IADL, leisure and social activities, and lower extremity mobility.^15,53 In the present study, participants with lower MMSE scores were found to have a higher risk for physical disability. This finding was consistent with that of previous studies. On the other hand, according to the American Thoracic Society,³⁹ the 6MWT evaluates the global and integrated responses of all the systems involved during exercise and may indicate overall functional capacity. A low score on the 6MWT indicates poor cardiopulmonary function, weaker muscle strength, and cognitive impairment.⁵⁴ In the present study, as the effect of cognitive function had been partialed out in the multivariate analysis, the association between the 6MWT and physical disability could be regarded as the net effect of “physical fitness” on physical disability. In our sample, compared with the participants whose scores fell in the highest quartile of the 6MWT, only those with scores in the lowest quartile had a higher likelihood for physical disability. This nonlinear relationship implied a lack of heterogeneity in the study sample or a ceiling effect of the performance on the 6MWT. As the participants of the present study were non-institutionalized, they were relatively high-functioning older adults. The observed high MMSE scores, along with the low GARS scores in the study sample seemed to provide evidence to this argument. If this is true, our findings suggest that poor sleep quality is associated with physical disability in healthy older adults, even when the effect of “physical capacity” is accounted for. The utilization of a high-functioning sample could have augmented the observed association and enabled us to get a clearer picture of the association. From the point of view of preventive medicine, our findings warrant future research that examines the underlying mechanism that links poor sleep quality and physical disability in healthy older adults.

The present study suffers from several limitations. This was a cross-sectional study; therefore, a causal relationship between self-reported physical disability and self-reported poor sleep quality could not be established. A longitudinal study will be necessary to disentangle the bidirectional relationship between poor sleep quality and other factors. Further, we evaluated a relatively small number of participants, which may have led to a low statistical power. Small sample size limited our capacity to include each of the comorbidities as an individual covariate in the multivariate analysis; therefore, only the total number of comorbidities was used. The lack of objective insomnia metrics (such as actigraphy or polysomnography), and screening for primary sleep disorders might have biased our findings and limited the interpretations. In addition, comorbidity may be overestimated by self-report; however, a correlation study conducted in Taiwan found that self-report of medical illness was highly reliable.⁵⁵ Failing to include socioeconomic characteristics such as marital status, education level, and employment status as confounding factors in the present study may have biased our findings away from the null hypothesis. In addition, pain is another important factor that may confound the association between poor sleep quality and physical disability. However, according to the scores of PSQI subscale, only 17 participants (8.0%) reported frequent pain-related sleep disturbance (≥ 3 times/week) in the present study. Small case numbers prevented us from including pain as a covariate. Finally, our recruitment strategy might not have included frail community-dwelling older adults. Therefore, the generalizability of our findings is limited.

In conclusion, the present study indicated that physical disability was significantly associated with poor sleep quality, even after adjusting for covariates in community-dwelling older adults. Large-scale, well-controlled, longitudinal studies are needed to further examine mechanisms that underlie these links.

DISCLOSURE STATEMENT

This was not an industry supported study. The authors have indicated no financial conflicts of interest.