Introduction

Sleep is a critical aspect of the biological life of humankind1. It is necessary for replenishing the energy and alertness for everyday activities and maintaining homeostasis, metabolism, and proper function of the brain and other organs of the body2,3,4. Not only sleep quantity, but also its quality has profound effects on our health5. Investigations in different parts of the world have shown that sleep deprivation is prevalent even among healthy individuals6. Reports show that adults have an average of 6.8 h sleep in weeknights and 7.8 h on weekends; 62% of adults do not feel they are getting enough sleep7. Poor or inadequate sleep has been associated with higher risks of cardiovascular diseases, depression, irritability, Alzheimer’s disease, fall and bone fractures, and chronic pain3.

The rate of aging is on the rise worldwide, and Iran is no exception. In this country, the speed of aging is one of the fastest in the world and more than 22% of Iranians are predicted to be over 65 years in 20508. Sleep disorders are common among the elderly9. Aging is associated with difficulties in falling asleep, staying asleep, and having a deep sleep9. Studies in Iran have demonstrated overall poor sleep quality in older adults10. Therefore, it seems beneficial to assess the sleep status alongside other health-related factors in this population.

Obesity, especially abdominal obesity, is related to major non-communicable diseases such as hypertension, type 2 diabetes, and cardiovascular disease and a decline in disease-free years in the elderly11,12. Not only obesity, but also undesirably altered body composition measures such as reduced muscle mass and elevated body fat are of concern; alterations that gradually occur with aging13. Compared to men, women possess higher rates of obesity and fat mass14,15. They also encounter a higher rate of poor sleep quality16,17.

Several investigations have indicated an inverse association between sleep duration/quality and common obesity markers, such as body mass index (BMI) and waist circumference18,19,20,21,22. However, body composition greatly changes during aging23, and is a better estimator of disease risk than obesity markers24. In recent years, the relationship between sleep duration/quality and body composition components has been noted by researchers25,26 but the influence of gender on such a relationship has not been extensively investigated. Given that body composition is remarkably different between men and women, we questioned whether the relationship between sleep quality and body composition components, such as fat and muscle mass, differs between genders. This relationship is particularly critical for muscle mass, as muscle wasting is one of the major problems of people in old age, increasing the risk of falls and fractures in this population. Women generally possess less muscle mass and bone mineral density than men27; so the mentioned relationship may be more crucial for women. Exploring such relationships may help design strategies to attenuate muscle loss and fat accumulation during aging.

Methods

Study design

The current cross-sectional study was conducted on 305 elderly people in Shiraz, Iran, from November 2021 to April 2022. The sample size was calculated based on previous investigations25 using a correlation coefficient of 0.2 for the association of sleep quality and fat mass percentage, a design effect of 1.5, power of 80%, and significance level of 0.05. Based on the results of linear regression analysis for body fat percentage and sleep quality, the power was estimated 99.5% and 93.2%, for men and women respectively, using R-squared values and the number of covariates in the multiple linear regression model.

The study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects. The protocol was approved by the Ethics Committee of Shiraz University of Medical Sciences with the approval code of IR.SUMS.SCHEANUT.REC.1400.002. Written informed consent was obtained from all participants.

Participants

Elderly people without major diseases were collected from two senior centers under the administration of the General Welfare Organization, and the Abolfazl primary health care center in Shiraz. These samples were selected through simple cluster sampling, in which three clusters were randomly selected out of 11 municipal areas in Shiraz; then one center from each cluster was randomly chosen. Due to the COVID-19 pandemic, participants were recruited by convenience sampling in each center. Thus, all clients who attended the three centers from November 2021 to April 2022 were invited to participate in the study. Inclusion criteria were as follows: age ≥ 65 years, community-dwelling residence in Shiraz, and agreement to participate in the project. People were not included if they were living in nursing or care homes, had surgery, stroke, heart attack, infection, falling, car accident, or hospitalization during the last month, or were afflicted by organ failure, thyroid disorders, serious mental illnesses, cognitive disorders, and dementia. Also, participants using antihistamine and antidepressant medications were excluded.

Data collection

Sleep quality

Sleep quality was evaluated by the Pittsburgh Sleep Quality Index (PSQI) questionnaire, which is a self-rated 18-item questionnaire that assesses sleep quality and sleep disturbances during the past month28. The questionnaire has been validated in the elderly Iranian population and has internal consistency and Cronbach's alpha of 0.81 (McDonald’s omega for the original questionnaire = 0.705)29. The PSQI score is generated from 7 components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction. Each component of the PSQI questionnaire is scored from 0 to 3, producing a total score of 0 to 21, with higher scores indicating worse sleep quality. A cut-off point > 5 is suggested as poor sleep quality28.

Anthropometric measures

Body weight was measured with the lightest possible clothes and no shoes by a digital scale (Glamor BS-801, Hitachi, China) to the nearest 0.1 kg. Height was measured by a stadiometer to the nearest 0.1 cm while the participant had shoes off and the head, shoulders, hips and feet were touching the wall. BMI was calculated by dividing weight in kilograms by the square of height in meters. Waist circumference was measured with a non-elastic tape at the mid-point of the distance between the iliac crest and the lowest rib.

Body composition

Body composition was analyzed by a portable bioelectric impedance analysis instrument (InBody S10, InBody Co., Seoul, Korea) with four electrodes attaching the wrists and ankles by the standard procedure declared by the manufacturer30. Four body composition components were recorded: body fat percentage, skeletal muscle mass percentage, fat-free mass percentage, and visceral adipose tissue area.

Physical activity

Physical activity was measured as a confounding factor. As data collection coincided with the COVID-19 pandemic, and the elderly were a high-risk group, we used a single-item questionnaire, which asked: “In the past week/past month, on how many days have you done a total of 30 min or more of physical activity, which was enough to raise your breathing rate. This may include sport, exercise, and brisk walking or cycling for recreation or to get to and from places, but should not include housework or physical activity that may be part of your job”31. The questionnaire has been validated in adults and has a modest concurrent validity with the Global Physical Activity Questionnaire (r = 0.53)31. Examination of the content validity of the translated questionnaire by an expert panel gave the content validity ratio (CVR) of 0.80 and the content validity index (CVI) of 0.97. Because the questionnaire contained one question, Cronbach’s α could not be determined.

Other information

Demographic information, living conditions, medical history, medications, and smoking habits were asked through a researcher-made checklist.

Statistical analysis

Statistical analyses were performed by Statistical Package for Social Sciences (SPSS), version 16 (SPSS Inc., Chicago, IL, USA). The normality of data was tested with the Kolmogorov-Simonov test and where needed data were log-transformed before performing statistical tests. The chi-square test was used to compare categorized demographic characteristics between the genders. Since a remarkable difference in sleep quality was observed between the genders, tertiles of sleep quality were determined separately in each gender. Anthropometric and body composition measurements across gender-specific tertiles of sleep quality were compared using one-way ANOVA.

The relationship between anthropometric variables including body composition data and gender-specific tertiles of sleep quality (as the independent variable) was examined using linear regression analysis in the crude model and after adjusting for possible confounders: age, education, marital status, income level, loneliness, physical activity, smoking, self-reported mood disorders, and chronic pain. These factors have been reported to influence sleep duration or quality32,33,34,35. To calculate the significance of the difference between men and women in the relationships of sleep quality with anthropometric and body composition measurements, linear regression analysis was performed using data from all participants (i.e. men and women), and gender was added among the confounding variables.

Multivariate relationships were examined using one-way multivariate analysis of variance (MANOVA) and Wilks’ Lambda criterion with body fat percentage, skeletal muscle percentage, fat-free mass percentage, and visceral adipose tissue area as dependent variables, gender-specific tertiles of sleep quality as independent variable, and the above-mentioned factors as covariates. For all statistical tests, the level of significance was set at 0.05.

Results

Overall, 627 individuals aged 65 and older were invited to participate in the study; 290 were not willing to participate, and 32 did not meet the inclusion criteria. Therefore, 305 individuals were included in the analysis. The mean age of the participants was 70.2 ± 5.1 years. They included 92 men (30.2%) and 213 women (69.8%). Of them, 210 (68.9%) were married, 46 (15.1%) had academic education, 254 (83.3%) were living with family, 168 (55.1%) reported to have a low income (related to the first two categories of income listed in Table 1), and 14 (14.1% of men and 0.23% of women) were employed. Demographic characteristics of the participants are presented in Table 1.

Table 1 Demographic characteristics of the participants based on gender.
Full size table

The prevalence of poor sleep quality was significantly higher in women (48.9% in men and 77.0% in women; P < 0.001). Table 2 demonstrates the scores of sleep quality components by gender. PSQI score averaged 6.2 ± 3.6 for men and 8.6 ± 3.7 for women (P < 0.001) (Table 2). Except the time spent in bed, women had significantly higher scores in each component of PSQI: they had shorter sleep duration, longer sleep latency, lower sleep efficiency, more frequent sleep disturbance and sleep medication use, worse daytime dysfunction, lower self-rated sleep quality, and a higher overall PSQI score than men. Fourteen men and 65 women used sleeping pills, 21 (3 men and 18 women) used 1 pill, 19 (2 men and 17 women) used 2 pills, and 39 (9 men and 30 women) used 3 pills during the past month.

Table 2 Sleep quality in participants based on gender.
Full size table

Comparison of anthropometric measures and body composition components between genders revealed that BMI, body fat percentage, and visceral adipose tissue area were significantly higher and skeletal muscle and fat-free mass percentages were significantly lower in women than men (Table 3). The prevalence of excess weight was high among the participants; 54.4% of men and 79.3% of women had overweight/obesity (BMI > 25 kg/m2).

Table 3 Anthropometric measures and body composition components in the participants based on gender.
Full size table

BMI (P ≤ 0.01) and waist circumference (P < 0.01) significantly increased across tertiles (i.e. with worsening) of sleep quality in both men and women (Table 4). However, there was a gender difference in body composition components. For men, no significant pattern was observed across gender-specific sleep quality tertiles, but women demonstrated a significant increase in BF% (P = 0.032) and visceral adipose tissue area (P = 0.010) and a significant decrease in skeletal muscle percentage (P = 0.027) and fat-free mass percentage (P = 0.009) across the tertiles.

Table 4 Anthropometric and body composition measures across gender-specific tertiles of PSQI.
Full size table

The association of anthropometric measurements and tertiles of sleep quality was examined with linear regression (Table 5). In men, BMI and waist circumference showed significant positive associations with PSQI in the crude model and after adjusting for confounding factors including age, education, marital status, income level, loneliness, physical activity, smoking, self-reported mood disorders, and chronic pain. In contrast, body composition variables were not associated with gender-specific PSQI tertiles in the crude or adjusted models in men. For women, all associations were significant with the third tertile of PSQI in the crude and adjusted models. Men and women differed in the examined relationships (P < 0.001) except for waist circumference which did not show a significant difference between genders (P = 0.654) (Table 5).

Table 5 Linear regression analysis for the association of gender-specific tertiles of PSQI and anthropometric and body composition variables.
Full size table

Because there was more than one body composition component, MANOVA test was used to see if there was a significant association between sleep quality tertiles and total body composition data. In agreement with the results of linear regression analysis, men did not show an association between PSQI tertiles and body composition data but women indicated an association in the third tertile of PSQI (Table 6).

Table 6 Multivariate analysis of variance examining the association of gender-specific tertiles of PSQI and all body composition data.
Full size table

Discussion

Main findings

Overall, the results of this study showed considerable differences in the rate of overweight/obesity, body composition, and sleep quality between genders. Women had higher BMI and adiposity and lower skeletal and fat-free mass percentages. They also had shorter sleep duration and poorer overall sleep quality than men. Men showed significant associations between sleep quality and BMI and waist circumference, but not body composition components. Women demonstrated significant associations between all the examined anthropometric and body composition variables and the third PSQI tertile (i.e. the worst sleep quality).

Prevalence of poor sleep quality

The prevalence of poor sleep quality was high in the participants of this study (68.5%). Previous studies on the Iranian elderly have similarly reported a high rate of poor sleep quality (75%)36. However, studies in other parts of the world have reported various rates, ranging from 25% in the USA37 to approximately 45% in China38,39 and Brazil40 to 76% in community-dwelling older adults in Slovenia41, but the rates may be higher in nursing home residents (~ 95%)41,42. The difference in the sleep quality between countries may be due to differences in the economic status of the nations43. Household low income is known as a predictor of poor sleep quality44.

Gender difference in body composition

Women had higher rates of adiposity than men as evidenced by higher BMI, body fat percentage, and visceral adipose tissue area. The rate of obesity based on BMI (i.e. BMI ≥ 30 kg/m2) was 10.9% and 38.5% in men and women, respectively. However, based on the ranges of body fat percentage proposed by Gallagher et al. (> 31% and > 43% for obesity in white men and women aged 60–79 years, respectively)45, 20.9% and 24.9% of men and women had obesity. It is not clear whether BMI or body fat percentage is a better predictor of obesity, but body fat percentage may be a more reliable indicator because it considers gender in the estimation of obesity. Women possess higher fat masses and are exposed to a higher rate of obesity14,15,46. Moreover, women deposit fat mostly in the subcutaneous areas and lower limbs while men store fat mostly around the abdomen, an area associated with the risk of metabolic diseases47. This sexual dimorphism is probably mediated through estrogen and androgen receptor-induced gene expression48. Women have more fatty acid uptake by adipose tissue, but they also have a higher rate of fat oxidation during exercise47. The latter indicates the presence of easily metabolizable fat depots in women which may benefit them in the reproductive process.

Contrary to body fat percentage, women had a lower percentage of skeletal muscle compared to men. With age, muscle mass decreases and fat accumulation increases (possibly because of hormonal alteration and gradual decrease in physical activity), leading to an increased rate of obesity49 and sarcopenic obesity50 in older adults. The cross-sectional design of this study did not permit seeing the gender difference in fat and muscle mass alteration during the aging process but previous studies have reported that the skeletal muscle declines at a faster rate in men during aging process51,52. In this regard, Janssen et al. reported that skeletal muscle percentage in women decreases from 34.1% at the age of 18–29 years to 30.2% at the age of 70 and above (with 11.4% decline), while in men these values are 42.3% and 36.0% (14.9% decline), respectively52.

Gender difference in sleep quality

Women showed a higher prevalence of poor sleep quality (77.0% compared to 48.9% in men). The higher prevalence of poor sleep quality in women has also been reported in previous studies16,17,53,54. This gender difference in sleep quality is partly attributed to the higher rate of affective problems such as depression and anxiety in women55. Also, lower educational levels, living alone, and low income are among the factors that may affect sleep quality especially in the elderly56. However, adjusting for self-reported mood disorders and chronic pain as well as socio-economic (age, education, marital status, income, and loneliness) and lifestyle (physical activity and smoking) factors did not affect the significant difference in sleep quality between genders (data not shown). This finding was also observed in young Australian adults16. The reason of higher rate of poor sleep quality in women is not clear based on the results of this and previous investigations and future studies are needed to elucidate the mechanisms involved.

Poor sleep quality, adiposity, and body composition

Previous investigations have pointed to the inverse association between sleep duration/quality and obesity markers18,19,20,21,22. Various explanations have been proposed for this relationship. For instance, during short sleep, changes in eating behaviors such as irregular meal times and frequent snacking occur and contribute to weight gain57. In addition, poor sleep may interfere with appetite regulation and satiety signals in the hypothalamus, leading to elevated ghrelin and decreased leptin levels.

In line with our findings in women, a relationship between short sleep and muscle mass depletion in Chinese elderly was reported by Fu et al.58 although they did not present the data by gender. The decline in muscle mass may result from hormonal dysregulations during poor sleep. In this regard, Auyeung and colleagues reported that in older men, there is an inverted U-shaped relationship between sleep duration and muscle mass and function, associated also with inverted U-shaped testosterone levels59. Even in healthy young males, one-night total sleep deprivation reduces the muscle protein synthesis, increases plasma cortisol, and decreases plasma testosterone levels60. In fact, sleep affects testosterone and cortisol levels and sleep deprivation may disrupt their balance and increase muscle proteolysis61. Apart from hormonal reasons, poor sleep quality may increase inflammatory markers which may negatively affect muscle anabolism and increase muscle protein breakdown.

Gender difference in the association of adiposity with sleep quality

There was also a gender difference in the association of obesity and body composition components and sleep quality. Men demonstrated a significant positive association between the markers of general and abdominal obesity and any level of poor sleep quality, while women displayed an association (positive for fat-related components and negative for muscle and fat-free items) between all anthropometric measures and the third tertile of sleep quality.

In contrast to the associations found in this study in both genders (i.e. sleep quality and obesity in men and sleep quality and obesity as well as body composition components in women), studies have mostly found an association between sleep duration/quality and obesity markers in women, but not men62,63,64,65,66,67,68. For instance, Mamlaki et al. reported that sleep duration or quality was inversely related to BMI and waist circumference in over 65-year-old women but not their male counterparts64. Also, in a study on over 1 million participants aged ≥ 30 years, both low and high sleep durations were associated with higher BMI in women (indicating a U-shaped relationship), but in men a longer sleep duration was associated with lower BMI67. In line with our results, Patel et al. reported increased BMI and waist circumference in the elderly men and women who slept < 5 h, with greater associations in men, as observed in the current study22. The gender-related discrepancies in the results of some studies may be due to differences in the method of measuring sleep (i.e. sleep duration vs. sleep quality), age of participants, degree of adiposity and fat mass in participants of each gender, and control of confounders.

The gender difference which was observed in the association of body composition components and sleep quality in this study may be the result of the difference in the fat and muscle masses between men and women. The magnitude of difference in BMI and waist circumference between genders was not much (~ 13%) but men and women greatly differed in the percentage of fat (~ 69%) and muscle (~ 21%) mass and visceral adipose tissue area (~ 41%); these findings are in the same line with the reports of previous investigators69,70. Only few studies have explored gender differences in the association of body composition components with sleep duration/quality. Fan et al. did not find an association between sleep duration and body fat percentage in adult men or women68, but two studies found that body fat percentage inversely70 and muscle mass positively69,70 were associated with sleep duration/quality in males but not females. Due to the very small number of investigations, it is hard to speculate the cause of discrepancies, but differences in sleep duration/quality and the rate of obesity and body fat percentage between genders in each study may be involved. For instance, in the studies of Buchmann et al.69 and Nam et al.70 higher rates of adiposity were observed in males than females. Other confounding factors such as genetics, psychological conditions, sociodemographic status, diet, and lifestyle might also have differed between studies and caused such discrepancies.

Strengths and limitations

This study was not without limitations. One of the limitations was the cross-sectional design which did not allow establishing a causal relationship between sleep quality and body composition components. The subjective assessment of sleep quality was not the most accurate way of evaluating sleep, although consistency has been found between studies that have objectively measured sleep quality and those that used self-report measures71. Also, the bioelectric impedance analyzer was not the gold standard for the assessment of body composition components. Therefore, the results need to be replicated with more accurate instruments such as dual-energy X-ray absorptiometry. The sample size was not sufficient to generalize the findings to the Iranian elderly population. Also, diabetes was not among the exclusion criteria. However, this work was one of the first studies that tried to explore gender differences in the association between sleep quality and body composition. Considering gender differences in sleep and adiposity, adjusting for potential confounders such as pain and mood disorders, and assessing both sleep duration and quality were among other strengths of this study.

Conclusions

In conclusion, this study indicated a significant and remarkable difference in the rate of overweight/obesity, percentage of body fat and muscle, and sleep quality between genders. While significant associations were observed between BMI/waist circumference and sleep quality in both genders, only women showed a significant association between body fat and muscle percentages and visceral adipose tissue area on one side and the third tertile of sleep quality (i.e. the worst sleep quality) on the other. Based on these data, a causal relationship between poor sleep quality and increased body fat and decreased muscles cannot be established. However, as body fat and muscle masses are important indicators of health and frailty in the elderly population72,73, improvement of sleep quality may lead to greater advantages in the elderly compared to other age groups.