Gender Differences are Noted in Factors Associated with Insomnia in Older Adult Outpatients: A Cross-Sectional Study

Background: Insomnia is a common sleep disturbance in older adults and is associated with many poor health outcomes. This study aimed to investigate the relevance of insomnia in older adult outpatients and to analyze differences between genders in factors associated with insomnia. Methods: This cross-sectional study was conducted in the outpatient clinics of a tertiary hospital in Southern Taiwan from July to September 2018. A total of 400 consecutive subjects aged 60 years or older were recruited. Insomnia was dened as a score of > 6 points on the Athens Insomnia Scale (AIS). Socio-demographics, health behaviors and clinical data were collected by face-to-face interview. Multivariable logistic regression was adopted for statistical analysis of the entire sample and stratied by gender. Results: Participants’ mean age was 74.74 ± 8.54 years, and the majority (93%) had more than one chronic disease. The prevalence of insomnia accounted for 30% (120/400) of all subjects, with males 22.9% (46/201) and females 37.2% (74/199). Gender, appetite, exercise, depressive symptoms, and sleep-related conditions such as short sleep duration, sleeping pills usage, medium-high risk of obstructive sleep apnea(OSA) and restless leg syndrome(RLS) were factors associated with insomnia in older adults. Exercise, sleeping pills usage, and RLS had an independent association with insomnia only in men, while appetite was associated with insomnia in women only. Conclusions: Insomnia is highly prevalent among older adults, predominantly females. Signicant differences are found between genders in factors associated with insomnia. Understanding gender differences may help clinicians to modify associated factors when managing older adults with insomnia.

Sleep-related problems, including sleep duration, sleeping pills usage, obstructive sleep apnea (OSA), and restless leg syndrome (RLS), were screened through yes/no questions or structured questionnaires. Self-reported actual duration of sleep (hours per night) within the last one month excluding the time in bed without sleep was recorded. The sleep duration was then grouped into long and normal sleep duration (>6 hours) and short sleep duration (<6hours) [28]. STOP-Bang Sleep Apnea Questionnaire was used to screen for OSA. Scores of 0-2, 3-4, and 5-8 are categorized as low, moderate, and high risk for OSA [29,30]. A four-question set recommended by the International Restless Legs Syndrome Study Group (IRLSSG) was used to de ne RLS [31]. Subjects who met all four criteria were classi ed into the RLS group.

Statistical analysis
Socio-demographic and clinical data are presented as means +/-SD or frequencies and percentages. Comparisons were made between insomnia and non-insomnia groups using independent student t-test or one-way analysis of variance (ANOVA) for continuous variables and chi-square test for categorical variables. Factors that demonstrated a signi cant association (p<0.05) with insomnia in univariable analysis were entered into multivariable logistic regression analysis to identify the independent relationship. Data were further strati ed by gender, and analyzed using univariable and multivariable logistic regression analysis. The odds ratio and 95% con dence interval of each variable was documented. All data were analyzed using the SPSS version 22 (IBM, Armonk, NY, USA). Two-tailed p-values of <0.05 were considered to be statistically signi cant.

Characteristics of study participants with or without insomnia
Among the 400 subjects who participated in the present study, the mean age was 74.74 years (SD, 8.54), and about one-half (49.8%) were women (Table 1). A majority (93%) of participants had more than one chronic disease. The prevalence of insomnia de ned by the AIS scale was 30% (n=120), and approximately 22.9% (46/201) and 37.2% (74/199) for male and female subjects, respectively. Subjects with insomnia had a higher prevalence of poor appetite and not forming regular exercise habits. Most participants also had short sleep duration, tendency to use sleeping pills, depressive symptoms, and medium-high risk of OSA and RLS. Factors associated with insomnia in older adults Abbreviations: OSA -obstructive sleep apnea, GDS-Geriatric Depression Scale, RLS -restless leg syndrome.

The in uence of gender on factors associated with insomnia
Variables that showed signi cance in the entire sample remained signi cant in univariable analysis of the male and female groups evaluated separately (Table 3). After analyzing by multivariable logistic regression, only short sleep duration, medium-high risk of OSA, and depressive symptoms had the same association with insomnia in both genders. The association of appetite disappeared in the male group, while three factors (sleeping pills usage, non-or irregular exercise, RLS) became non-signi cant in the female group. Non-or irregular exercise, sleeping pills usage, and RLS had an independent association with insomnia only in men, while poor appetite within 3 months was associated with insomnia in women only.

Discussion
This cross-sectional study targeted a total of 400 older adult participants in outpatient clinics to comprehensively investigate the factors associated with insomnia, especially considering the in uence of common sleep-related problems such as sleep duration, sleep medication usage, OSA and RLS, and intending to nd the in uence of gender on the associated factors. Results of the study showed that gender, poor appetite, not exercising regularly, depressive symptoms, and sleep-related conditions such as short sleep duration, sleeping pills usage, medium-high risk of OSA and RLS were factors associated with insomnia. Gender was also found to in uence partially associated factors and insomnia. Exercise, sleeping pills usage, and RLS were relevant to insomnia in males alone, whereas poor appetite was relevant to insomnia in females only. Otherwise, signi cant associations were found between short sleep duration, medium-high risk of OSA and depressive symptoms and insomnia across genders.
The prevalence of insomnia depends on the method of identi cation. In the present study, the prevalence of AISde ned insomnia was 30% for all subjects and is similar to the reports in previous studies, which ranged from 30% to 48% [1]. After considering the in uence of covariates, short sleep duration, sleeping pills usage, medium-high risk of duration was collected by questionnaires, the signi cant results were similar to the results derived from the general population using objective measures such as polysomnographic studies [32]. OSA and RLS are two common primary sleep disorders in older adults [33], and are frequently comorbid with insomnia, because all are associated with agingrelated physiologic and sleep architecture changes [12,34]. In the present study, 65.5% of participants demonstrated moderate-high risk of OSA and 4% of participants presented with the essential diagnostic features of RLS by IRLSSG.
Both disorders are signi cantly associated with insomnia. Older adults tend to take sleeping pills to relieve sleep complaints, however, our report disclosed that those who use sleeping pills still tend to complain of insomnia symptoms. Few studies have focused on older adults who use hypnotic agents, except for those with insomnia.
Although hypnotic agents improve sleep duration and latency, the magnitude of effect from sleep medications is small and potentially increases various adverse effects [35,36]. In addition, a variety of sleep medications have different mechanisms of action and different clinical effects on insomnia, which may in uence patients' subjective feelings regarding sleep.
The present study is the rst to identify the signi cant relationship between poor appetite and insomnia. Some clinicians have attributed poor appetite to be a manifestation of depression, but in the present study, the effect still existed in multivariable logistic regression analysis after adjusting for depressive symptoms identi ed by GDS-5. In addition, poor appetite may result in poor nutritional status, and previous studies have shown that nutritional inadequacy and malnutrition status identi ed by Mini Nutritional Assessment (MNA) scores correlate highly with insomnia [37,38]. Furthermore, poor appetite may reduce the consumption of foods rich in melatonin, a hormone secreted by pituitary gland that is part of the sleep-wake cycle and is reported to improve sleep e ciency [39].
Although some epidemiologic studies reported a negative association between habitual exercise and insomnia, after adjusting only for demographic data and mood symptoms [40,41], the present study did not nd similar signi cant associations after evaluating the effects of sleep-related problems. In addition, based on results of studies using quantitative exercise measurement instead of subjective self-reported questions as in the present study, both physical activity of moderate intensity and exercise training programs improved sleep symptoms and quality obviously [42,43]. The effects of depression on insomnia remained strong across genders in the present study. A review of prospective studies has also recognized depression as a risk factor for sleep disturbances [10]. Various mechanisms have been proposed to help explain the ndings relative to the effects of depression on sleep, including in ammatory regulations, genetic and familial factors, and the effects of social and environmental factors [44].
Similar to most previous studies, results of the present study show a clear relationship between female gender and insomnia [9,10,[14][15][16][17]. However, gender not only in uences the prevalence of insomnia, but also signi cantly affects the factors associated with insomnia in older adults. Few studies have investigated the differences in factors associated with insomnia between genders, especially not in older adults. A study of insomnia in Chinese adults aged 18-65 in Hong Kong revealed that socio-demographic factors (education level, occupation and retirement, marital status and environment) associated with insomnia were signi cantly different between genders, and though some common risk factors are found in both males and females, other risk factors were gender-speci c [18]. Another study of the general population in Korea, however, found that the factors of education and occupation status were not signi cantly associated with insomnia [19]. In addition to the insomnia-associated factors mentioned above, and focusing on the older adult population, exercise and sleeping pills usage were shown to be relevant to insomnia in males alone, whereas appetite was associated with insomnia in females only. Although gender differences in insomnia represent a complex interaction between biological, psychological and social factors, the need remains to explain gender differences in factors associated with insomnia in future studies [45].
The present study evaluated a group of older adult outpatients with a high prevalence of insomnia, and identi ed factors associated with insomnia, summarizing some potentially modi able factors for further intervention. Evaluation with validated questionnaires provided conceivable results that could be followed and compared with other populations. Substantial differences found between genders in factors associated with insomnia added new perspectives in clinical practice.
Nevertheless, the study has several limitations. First, the cross-sectional design limits drawing causative conclusions, and sleep patterns were not evaluated by objective measures such as polysomnography or actigraphy. Second, collection of socio-demographic variables based on self-reported data or simple dichotomous yes/no questions may result in recall bias, especially among this older adult population, and there may still be some potential confounding factors left behind. Since the diagnosis of RLS was based on the four screening criteria of the IRLSSG, some mimicking diagnoses may be inevitably included. Third, subjects were recruited from outpatients in a single medical center and thus results may not be extrapolated to different settings or areas.
Conclusions And Future Study Availability of data and materials The datasets analysed during the current study are available from the corresponding author Chih-Kuang Liang on reasonable request. Please contact (ck.vghks@gmail.com) for data access.

Funding
This study was sponsored by the Kaohsiung Veterans General Hospital [Grant number VGHKS18-CT6-15]. The funding organizations had no role in the design, methods, subject recruitment, data collection, analysis, or preparation of the article.
Authors' contribution YTP drafted the manuscript, conducted the formal analysis, and interpreted the data.
YHH collected the data, drafted the manuscript, conducted the formal analysis, and interpreted the data.
MYC collected the data, interpreted the data and critically revised the manuscript for important intellectual content.
CSC collected the data, conducted the formal analysis, and interpreted the data.
CSS de ned the study background and aims, de ned the question guide and collected the data.
CKL conceptualized the paper, de ned the study background and aims, de ned the question guide, collected the data, interpreted the data and editing of the manuscript.
YCW, MD collected the data.
TY de ned the study background and aims, de ned the question guide, collected the data and contributed to interpretation of data.
LKC contributed to interpretation of data and production of the manuscript.
YTL contributed to interpretation of data and production of the manuscript.