Sleep Quality and Subjective Cognitive Decline among Older Adults: The Mediating Role of Anxiety/Depression and Worries

Background Subjective cognitive decline (SCD) in older individuals has been implicated as a possible precursor to Alzheimer's disease. Poor sleep quality and anxiety/depressive symptoms have been linked to the progression of SCD, but these associations and older adults' worries have yet to be fully established in the Chinese older adult population, which is one of the largest in the world. The aim of this study was to explore the relationship between sleep quality, anxiety/depression symptoms, and worries, and SCD prevalence among Chinese community-dwelling older individuals. Methods A total of 707 adults aged between 60 and 99 from Shanghai, China, completed self-report questionnaires that covered their cognitive and mental well-being, as well as demographic information. Subjective cognitive decline (SCD) was evaluated using the memory/cognition syndromes of the old adult self-report (OASR). Sleep quality, anxiety/depression, and worries were measured from their respective sections of the OASR. Results The general linear regression models showed that poorer sleep quality was associated with an increased prevalence of anxiety/depression symptoms, worries, and SCD among older adults. As suggested by the mediation analysis, anxiety/depression and worries were significant mediators in the relationship between sleep quality and SCD prevalence, and these two factors also have a serial mediation effect between sleep quality and SCD prevalence. Conclusions Poorer sleep quality is associated with a higher rate of SCD among older adults, and a higher prevalence of anxiety/depression and worries mediate this relationship, suggesting possible mechanism pathways that lead to SCD. These factors may provide the basis for early, targeted interventions for older adults' mental health preservation and improved quality of life.


Introduction
Intact cognitive function is crucial to the standard of living for individuals progressing from middle to late adulthood.Subjective cognitive decline (SCD) is the self-reported decline of memory and increased confusion in individuals, and it has emerged as a growing issue as demographics skew towards older populations around the world [1,2].In 2016, over 8.5% of the world population was over the age of 65, and this percentage is expected to grow to 17% by 2050, representing an increase from 617 million to 1.7 billion individuals [3].Furthermore, China is expected to gain almost 200 million older individuals alone in the next 20 years [4].SCD has gained traction in recent years as a possible precursor to Alzheimer's disease [5].Longitudinal studies report a higher incidence of and faster conversion rates to objective, later stages of cognitive decline, such as mild cognitive impairment and Alzheimer's disease, in adults with SCD compared to their cognitively intact counterparts [6,7].Tis warrants SCD as an important stage to explore in older adults for possible early Alzheimer's disease interventions.
In recent years, poor sleep quality has been implicated as a possible risk factor in the progression of SCD [8].Crosssectional studies in Korean populations have shown that poorer sleep quality is associated with an increased prevalence of and more severe SCD symptoms for both genders and among various age groups [9][10][11].A longitudinal study by Zhao et al. found support for the link between insomnia and SCD in Canadian middle-aged and older adults [12].
Beyond its association with SCD, poor sleep quality is possibly associated with a higher incidence of objective cognitive function issues although the fndings are mixed.Siddarth et al. found objective memory diferences between adults who reported SCD and adults without SCD [13].However, Exalto et al. reported that poorer sleep quality in adults was associated with more severe SCD symptoms but not cognitive testing, MRI, or cerebrospinal fuid biomarker diferences [14].In addition, there is a lack of SCD and sleep quality literature for Chinese populations, which is necessary considering the increasing incidence of SCD in China.While Wang et al. found a correlation between poor sleep and SCD in Western Chinese adults, it may not be fully generalizable to the Chinese population [15].Terefore, additional research is necessary to corroborate these fndings and apply to more older Chinese adults.
Increased anxiety and depressive symptoms have also been associated with the prevalence of SCD.Depressive symptoms and SCD each independently contribute as risk factors to more severe neurocognitive disorders, and their comorbidity greatly increases this risk [16].Adults with SCD experienced greater perceived stress and had increased anxiety/depressive symptoms during COVID-19-related confnement [17].A cross-sectional study across 47 middle-to low-income countries indicated that depressive symptomatology was associated with more severe SCD across age, gender, and education groups [18].In addition, patients who reported SCD tend to be older, have less education, and have higher levels of anxiety/depression although these associations are not necessarily correlated with objective cognitive decline [19].Higher levels of worry are also associated with greater amyloid-β concentration in the frontal lobe, a likely precursor to Alzheimer's disease, in older German adults [20].However, there remains limited research on how the trait of worry is associated with SCD; while this trait is often confated with anxiety/depression, it difers in that anxiety/depression often revolves around negative self-directed thinking, while worries predominantly originate from external factors, such as environmental and interpersonal stressors.Targeting external stress factors is crucial to developing and delivering cognitive therapies that may improve older individuals' quality of life.
With regards to the connection between sleep quality and anxiety/depression in older adults, there is a wellestablished negative correlation between the two variables.Depressed older individuals experience objectively poorer sleep and exhibit increased nocturnal awakening time [21].Te two factors play into one another as well-older individuals who are anxious and/or depressed are more likely to complain of insomnia, and in turn, insomnia increases the risk for psychiatric disorders [22].In a cross-sectional study of southern Chinese nursing home residents, almost 70% of the residents experienced poor-quality sleep.Tese residents showed increased depressive symptomatology, indicating poor-sleep quality as a marker for depression [23].For northern Chinese nursing home residents, factors such as loneliness and self-care abilities signifcantly afected depression in those with poor sleep quality [24].In addition, the combination of poor sleep quality and anxiety/ depression has been implicated in cognitive decline.Several studies have shown a possible mediating role of sleep quality between anxiety/depression and cognitive decline, though the fndings are limited.Liu et al. indicated only a partial mediating efect from poor sleep quality while Wu et al. showed that sleep quality mediation existed only for males [25,26].However, it is vital to examine anxiety/depression and worries as potential mediators between sleep and SCD because poor sleep has been implicated to increase negative mood states [27].
In terms of the mediating efects of anxiety/depression and worries, recent research suggests that anxiety/depression mediates the relationship between sleep quality and SCD.Xiao et al. and Xu et al. both found that poor sleep quality and SCD or cognitive complaints were mediated by depression and anxiety symptomology in older Chinese adults [28,29].However, a relationship between sleep quality and SCD with worries as a possible mediator has yet to be established.Te primary objective of the present study is to fll the existing gaps in the literature by investigating the relationship between sleep quality, anxiety/depression symptoms, worry traits, and SCD in a substantial sample of older Chinese adults.We hypothesized that poor sleep quality is associated with subjective cognitive decline among older adults and is mediated by anxiety/depression symptoms and worries.

Study Population.
A total of 755 elderly individuals, aged 60 and above, without signifcant mental or physical impairments, were included in this community sample.Te recruitment of this participant group was undertaken by the nursing research teams from Shanghai Longhua Hospital and Shanghai Shuguang Hospital, with a focus on 25 local communities in Shanghai, China.Tis was done through a combination of convenience and community sampling methods in November 2017 [30].Out of the initial group, 707 participants successfully completed the survey and they received compensation of 100 Chinese Yuan each.
Te survey was administered by a team of 150 interviewers, composed of 40 nursing supervisors from Shanghai Longhua Hospital and 110 interns from Shanghai Shuguang Hospital.Prior to conducting the interviews, comprehensive training workshops were conducted to ensure the standardization of interview procedures.During the interviews, the participants were presented with a written survey to follow, while the interviewers verbally presented each item and recorded the participants' responses on a separate copy.If participants were capable of independently completing the items, they were given the option to write their own answers.Detailed sampling and recruitment methods were provided in our previous publication [30].
All participants provided written informed consent, and the research was conducted with approval from the Institutional Review Boards of both Shanghai Shuguang Hospital and Shanghai Longhua Hospital (Ethics approval number: 2020-GZR-02-053X).

2
Journal of Aging Research 2.2.Measure.Te assessment of anxiety/depression symptoms, worry levels, and subjective cognitive decline (SCD) in these older adults was conducted using the memory/cognition syndrome scales of the older adult self-report (OASR) and the older adult behavior checklist (OABCL).Te OASR is a standardized instrument developed by Achenbach et al.
for older adults to self-report their adaptive functioning; behavioral, emotional, and social problems, as well as their substance use (i.e., cigarettes, alcohol, and other drugs) [31].Likewise, the OABCL was developed by Achenbach et al. for people close to the adults to report the adults' functioning, allowing for comparisons between their self-perception and others' perceptions [31].Previous research has demonstrated the cross-cultural applicability of the OASR across 20 diferent societies, including China, Brazil, and Germany [32].Specifcally, the Chinese version of the OASR and OABCL has been validated for use among Chinese-speaking older adult populations in clinical and research settings, enabling the assessment of mental health in this demographic [30,32].Anxiety/depression symptoms were measured using the anxious/depressed syndrome scale of the OASR, which encompasses 20 items, such as "feels worthless" and "selfconscious".Worry levels were evaluated using the worries scale of the OASR, which covers eight items, such as "feels burdensome" and "worries about family".Participants provided ratings for all items on a three-point scale (0 � "not true," 1 � "sometimes true," and 2 � "often true"), which refected the participants' experiences over the preceding 2 months.
Sleep quality was measured using the Pittsburgh Sleep Quality Index (PSQI), which consists of 19 items that cover 7 subscales: sleep quality, sleep latency, sleep duration, sleep efciency, sleep disturbances, use of sleep medications, and daytime dysfunction.Te PSQI has been demonstrated to be reliable and valid in various Chinese populations, including older populations [33][34][35].
Subjective cognitive decline (SCD) was evaluated in accordance with the scoring methodology mentioned above, utilizing the memory/cognition problems syndrome scale within the old adult self-report (OASR).Tis scale comprises nine items that vary from "cannot concentrate" to "forgets names."Scores for SCD were derived by summing the responses for each item within the syndrome, which were presented in a normalized T-score format (M � 50, SD � 10), and higher scores indicated a greater degree of subjective cognitive decline [36].
Demographic variables encompassed age, gender (male or female), education level (high education: high school equivalent and beyond and low education: incomplete high school education), income level (high income: total monthly income exceeded RMB 3000 and low income: total monthly income was below RMB 3000), and marital/living status (lives with spouse/partner: married and cohabiting with spouses or unmarried but cohabiting with partners, and other: unmarried, divorced, widowed, or not cohabiting with a spouse/partner over the last 2 months).In terms of age stratifcation for analysis, the sample was divided into two distinct categories: the young-old group, encompassing individuals aged 60-74 and the old-old group, encompassing those aged 75 and above.Tese factors were all controlled for in the subsequent mediation models.

Data Analyses.
Te current study includes 707 data, excluding 8 individuals who had no OASR record, leaving 699 valid participants.Sample characteristics were summarized by descriptive statistics, including the sample count, percentage, mean, and standard deviation, as appropriate.With adjustments for age, gender, education, income, and marital status, generalized linear models (GLMs) were applied to evaluate the independent efects of the following relationships: (1) sleep quality on anxiety/depression, (2) sleep quality on worries, (3) sleep quality on SCD, and (4) sleep quality, anxiety/depression, and worries on SCD.Separated GLMs were created for both self-reported and other-reported SCD.Te mediation analyses were performed in three stages to identify whether anxiety/depression and worries mediated the relationship between sleep quality and SCD among old adults.A single mediation model investigated the mediation efect separately for anxiety/depression and worries.Nonparametric bootstrapping analyses were deployed to test the mediate efects of anxiety/depression and worries as parallel mediators of the relationship between sleep quality and SCD.Serial mediation models were performed to test the serial causal chain of the two mediators using PROCESS 4.3 (Model 6) [37].Sleep quality, anxiety/depression, worries, and SCD were standardized (t score) before running the analyses.Statistical analyses were processed by RStudio.p values less than 0.05 were considered statistically signifcant.

Sample Characteristics.
Te basic demographic characteristics of old adults stratifed by age and gender were summarized in Table 1.Overall, about three fourths (N � 510) of old adults were young-old adults (aged between 60 and 74) and one fourth were old-old adults (aged above 74).Te gender consisted of 45% (N � 307) males and 55% (N � 375) females.Te mean age of all participants was 69.68 (SD � 7.46), 66.14 (SD � 4.37) for the young-old group and 80.17 (SD � 4.37) for the old-old group.A total of 78 (12%) old adults reported having bachelor's and higher degrees, 179 (26%) old adults obtained high-school degrees, and the rest were at middle school and lower education levels (N � 425, 62%).High-income participants (N � 336) accounted for 49% of the whole sample, and there was a similar ratio of low-income participants (N � 346).About three fourths (N � 492, 72.14%) of old adults lived with their spouse/partner in the past 2 months, and 190 (27.86%) old adults were in "other" marital status.

Association between Sleep Quality and SCD.
Signifcant associations between sleep quality and SCD in both forms were observed in GLMs (Table 3).After controlling for confounding variables (age, gender, education, income, and marital status), sleep quality was positively associated with higher scores in SCD, showing that the old adults who had poor sleep quality (β self-reported � 0.24, p < 0.001; β other-reported � 0.24, p < 0.001) tended to have a higher prevalence of SCD problems.

Mediation Roles of Anxiety/Depression and Worries
between Sleep Quality and SCD 3.5.1.Single Mediation Analysis.Te mediation roles of anxiety/depression and worries are shown separately in Figures 1(a) and 1(b).As shown in Figure 1(a), anxiety/ depression partially mediated the association between sleep quality and SCD (mediation effect self-reported � 0.184, p < 0.001).Te signifcant indirect efect explained 76% of the total efect of sleep quality on SCD, and the direct efect of sleep quality was signifcant (β self-reported � 0.057, p < 0.05).Te mediation efect of worries (mediation efect: β self-reported � 0.246, p < 0.001) was signifcant, while the direct efect of sleep quality (β self-reported � 0.023) on SCD was not signifcant.Terefore, worries fully mediated the relationship between sleep quality and SCD.

Parallel Mediation Analysis.
As shown in Figure 2, the mediation efects of anxiety/depression (β � 0.155, p < 0.001 * * * ) explained 63% of the total efect and worries (β � 0.074, p < 0.001 * * * ) explained 31% of the total efect while controlling for confounding variables.Te direct efect of sleep quality on SCD was insignifcant.Both anxiety/ depression and worries fully mediate the relationship between sleep quality and SCD, suggesting that old adults who had poorer sleep quality were more likely to have higher levels of anxiety/depression and worries.In addition, old adults with a higher prevalence of anxiety/depression and worries tended to report higher levels of SCD symptoms.

Serial Mediation Analysis.
Te serial mediation revealed a causal chain linking of mediators (anxiety/depression and worries) in two specifed directions (Figure 3(a) and 3(b)).Te bootstrapped results of each path are detailed in Table 5. Te results of serial mediation models showed that anxiety/depression and worries fully mediated the relationship between sleep quality and SCD (total mediation efect � 0.229, p < 0.05), which explained 95% of the total efect (total efect � 0.241, p < 0.001).In serial model 1, the mediation efect path of sleep quality ⟶ anxiety/depression ⟶ SCD (mediation efect � 0.155) explained 64% of the total efect.Te path of sleep quality ⟶ anxiety/ depression ⟶ worries ⟶ SCD was signifcant, suggesting that poorer sleep quality leads to an increase in the prevalence of both anxiety/depression and worries, anxiety/depression can cause an increased prevalence in worries as well, and the incidence of both anxiety/depression and worries may contribute to an increase in SCD symptoms.In serial model 2, the path of sleep quality ⟶ worries ⟶ anxiety/depression ⟶ SCD yielded     Journal of Aging Research a high ratio of mediation efect to total efect (mediation efect � 0.169, 70% of the total efect), and the path of sleep quality ⟶ anxiety/depression ⟶ SCD was not signifcant.Te results suggested that poor sleep quality may lead to an increase in the prevalence of worries, thereby causing a rise in anxiety/depression and, therefore, SCD symptoms.

Discussion
Tis cross-sectional study from Shanghai communitydwelling older adults found that self-reported anxiety/depression symptoms and worries are mediating factors between older individuals' sleep quality and SCD.Such relationships are still held even when considering other extraneous factors, such as gender, income, education, and marital status.In addition, a serial mediation efect of anxiety/depression and worries was found connecting sleep quality to SCD.Tese fndings highlight important implications in developing more targeted interventions to slow the progression of SCD into objective cognitive deterioration.

Efect of Sleep on Subjective Cognitive Decline.
Previous cohort studies done in other regions of China found that anxiety/depressive symptoms mediate the relationship between sleep quality and SCD, and our study fndings were consistent with them, thus improving the possibility of generalizing such results to more of the Chinese population [28,29,38].Tese fndings also agree with literature on anxiety and depression independently increasing sleep issues in older adults [39][40][41].In addition, the link between sleep quality and SCD is reafrmed, despite mixed results on the subject-non-Chinese cohort studies Figure 1: (a) Mediation role of anxiety/depression between sleep quality and self-reported SCD.Note.Mediation analyses were performed with 1000 bias-corrected bootstrapped samples.Unstandardized regression coefcients reported from self-reported SCD models were out of brackets and other-reported coefcients were in brackets.Te models are controlled for age, gender, education, income, and marital status.
* � p < 0.05, * * * � p < 0.001.(b) Mediation role of worries between sleep quality and self-reported SCD.Note.Mediation analyses were performed with 1000 bias-corrected bootstrapped samples.Unstandardized regression coefcients reported from self-reported SCD models were out of brackets and other-reported coefcients were in brackets.Te models are controlled for age, gender, education, income, and marital status.* * * � p < 0.001, n.s.� not statistically signifcant.Journal of Aging Research have connected SCD with objective memory loss, but not sleep quality with objective memory loss even though sleep quality is associated with SCD [13,14].Nevertheless, research on such connections in Chinese populations is minimal, and our fndings corroborate the connection between poor sleep and SCD in older residents of Western China [15].Depression has also been previously suggested as a partial mediator between SCD symptoms and decreased objective memory, making our fndings important as anxiety/depression may now play a key role even earlier in cognitive decline mechanisms [42].Given the high incidence of signifcant depressive symptoms in Chinese older adults-up to 60.3% in some regions-targeting anxiety/ depression may be particularly benefcial in reducing older adults' progression of SCD [43].

Worries as a
Mediator.Signifcantly, our fndings illustrate that worries are an additional mediator between sleep quality and SCD.To date, no studies have focused on worries and their impacts on regulating the mechanism between sleep and early stages of cognitive decline.Repetitive negative thinking is associated with SCD, and increased amyloid-beta loads are present in the brains of older individuals who report more worries, which possibly indicates Alzheimer's disease progression [20,44].It has also been suggested that poor sleep provides a pathway for betaamyloid to accumulate in the brain [45].Tis provides a possible mechanism for how worries act as a mediator between sleep quality and SCD-poor sleep incites the accumulation of beta-amyloid, which is associated with worries, and beta-amyloid is a marker for SCD and more severe forms of dementia.Tus, it is reasonable to believe that poor sleep not only acts on SCD directly as suggested by previous cross-sectional studies but also by increasing worries, which act on SCD as well [9][10][11][12].Tis novel fnding is benefcial in providing increased possibilities for interventional strategies beyond simply treating anxiety/depression symptoms and may improve the efcacy of current treatment strategies by targeting worries in older individuals.

Anxiety/Depression and Worries as Serial Mediators.
In addition, our fndings are the frst to support a serial mediation model connecting sleep quality, anxiety/depression, worries, and SCD.To our knowledge, most studies have focused on the mediating efect of anxiety/depression between sleep quality and SCD and have not connected worries as a possible factor in the pathway.Te interchangeability of anxiety/depression and worries as serial mediators, as both serial pathways were signifcant, expands the possible therapeutic interventions for delaying the onset of SCD.While anxiety/depression and worries are closely related, they are still inherently diferent in their categorization in the OASR: anxiety/depression items more often focus on negative thinking towards self while worries are directed more externally and are a reaction to environmental and relationship stressors.Due to the possible confation between anxiety/depression and worries, there is minimal research in older populations that examines the connection between the two variables.Our diferentiation between the two factors is signifcant because it can address older individuals' mental health in a more efective fashion: simultaneously addressing more internalizing anxiety/ depression and externalizing worries may provide even better cognitive outcomes for older individuals.

Limitations.
Te interpretation of the fndings should take into account the inherent limitations of the study.First, it is important to note that this cross-sectional study utilized a survey method to measure the prevalence of cognitive decline.As a result, both explanatory variables and outcomes were collected simultaneously, without a prospective or retrospective follow-up.Consequently, establishing a causal relationship based on these fndings is not currently feasible.
A future longitudinal study may better examine the full causality among all variables.Second, while our model considers worries to be a mediator that results in the development of SCD, other literature suggests worries are a symptom of SCD [42].Further research may be necessary to clarify the relationship between the two factors given the general paucity of research on the possible impacts of worries on SCD.Tird, it is worth noting that the OASR, the instrument employed in this study, was originally developed based on Western samples.As a result, it may not fully capture the complete range of behaviours and traits that are specifc to Chinese culture and the Chinese population.Lastly, the study's community sample specifcally consisted of older adults from Shanghai, a highly developed city in China.Terefore, caution should be exercised when generalizing the results to other geographic regions, especially rural areas, and regions with greater ethnic diversity.In light of these limitations, a comprehensive understanding of the study's fndings necessitates careful consideration of these factors to ensure accurate interpretation and appropriate generalization.

Conclusions
By 2050, close to 20% of the global population will be above the age of 65.Tis looming surge in the older population spotlights a need to ensure that older individuals minimize their risk of SCD and more serious forms of dementia for a healthier aging process.Our study found an independent mediating efect and the serial mediating efect of anxiety/ depression and worries on the relationship between sleep quality and SCD in an older Chinese population while controlling for extraneous factors.Tis is vital to addressing not only SCD prevalence but also the high rates of depression/anxiety in older Chinese adults in the status quo.More research is warranted to examine the long-term outcomes of these individuals, establish a causal relationship through a longitudinal study, and test the efcacy of possible interventions.Our current and future fndings have important implications in creating targeted, personalized treatment plans for protecting the mental health of older individuals.Such interventions are vital to decreasing Figure 1: Continued.

a 1 =Figure 2 :Figure 3 :
Figure2: Parallel mediation model.Mediation efects of sleep quality on SCD through anxiety/depression and worries.Note.Mediation analyses were performed with 1000 bias-corrected bootstrapped samples.Unstandardized regression coefcients for self-reported SCD as outcome were out of brackets and coefcients for other-reported SCD as outcome were in brackets.Te models are controlled for age, gender, education, income, and marital status.* * * � p < 0.001, n.s.� not statistically signifcant.

Table 2 ,
controlling for age, gender, education, income, and marital status; sleep quality was positively associated with anxiety/depression and worries, suggesting that old adults who had poorer sleep Journal of Aging Research quality tended to have higher prevalence of anxiety/depression (β � 0.49, p < 0.001) and worries (β � 0.36, p < 0.001).

Table 1 :
Sample characteristics by age and gender.

Table 2 :
Adjusted general linear regression for the efect of sleep quality on anxiety/depression and worries.Note.β � the change in anxiety/depression or worries while the variable increased by one unit, controlled for all other variables.SE � standard error.

Table 3 :
Adjusted linear regression for the efect of sleep quality on SCD.
*Note.β � the change in SCD while the variable increased by one unit, controlled for all other variables.SE � standard error.* � p < 0.05, * * * � p < 0.001.

Table 4 :
Efect of sleep quality, anxiety/depression, and worries on self-reported and other-reported SCD.
Note. β � the change in SCD while the variable increased by one unit, controlled for all other variables.SE � standard error.* * * � p < 0.001.

Table 5 :
Bootstrap result of serial mediation analysis for anxiety/depression and worries between sleep quality and SCD.