Association between mindfulness and sleep quality in perimenopausal women: A cross-sectional mediation analysis of depressive symptoms

Abstract

BACKGROUND

Perimenopausal women are vulnerable to both sleep disturbances and depressive symptoms. Mindfulness has been associated with better psychological well-being and sleep outcomes, but the specific psychological pathway linking mindfulness to sleep quality in this population remains insufficiently defined. In particular, it remains unclear whether depressive symptoms represent an intermediate pathway between mindfulness and sleep quality among perimenopausal women.

AIM

To examine the associations among mindfulness, depressive symptoms, and sleep quality in perimenopausal women, and to test an a priori mediation model in which depressive symptoms mediate the association between mindfulness and sleep quality.

METHODS

This observational study included perimenopausal women recruited using a convenience sampling approach from a tertiary hospital in Zhengzhou, Henan Province, China, between October 2023 and December 2024. Participants completed the Mindful Attention Awareness Scale, the 10-item Center for Epidemiologic Studies Depression Scale, and the Pittsburgh Sleep Quality Index (PSQI). Spearman correlation analysis, multiple linear regression, and mediation analysis using Mplus were performed. Multiple linear regression and mediation analyses were adjusted for age, body mass index, marital status, income level, number of pregnancies, and presence of complications to control for potential confounding effects. Because of the cross-sectional design, mediation results were interpreted as indirect associations rather than evidence of causal direction.

RESULTS

A total of 693 participants were included. Depressive symptoms were negatively correlated with mindfulness (r = -0.444, P < 0.01) and positively correlated with PSQI scores (r = 0.422, P < 0.01), indicating poorer sleep quality. Mindfulness was negatively correlated with PSQI scores in unadjusted correlation analysis (r = -0.268, P < 0.01). In multivariable analysis, both mindfulness and depressive symptoms were independently associated with sleep quality (P < 0.001). Mediation analysis showed that depressive symptoms significantly mediated the association between mindfulness and sleep quality (indirect effect = -0.176, 95% confidence interval: -0.222 to -0.131), accounting for 63.5% of the total effect.

CONCLUSION

Depressive symptoms were found to statistically mediate the association between mindfulness and sleep quality in perimenopausal women. However, given the cross-sectional design, these findings should be interpreted as indirect associations rather than causal relationships. Longitudinal and interventional studies are needed to further clarify these relationships.

Key Words: Perimenopause; Mindfulness; Depressive symptoms; Sleep quality; Mediation analysis

Core Tip: This cross-sectional study of 693 perimenopausal women found that depressive symptoms mediated 63.5% of the association between mindfulness and sleep quality. Higher mindfulness was associated with lower depressive symptoms, which in turn were associated with better sleep quality. These findings suggest that depressive symptoms may partially explain the mindfulness-sleep quality relationship. Due to the cross-sectional design, results indicate indirect associations only. Longitudinal studies are needed to confirm causal pathways.

INTRODUCTION

With population aging, the mental health of perimenopausal women has become an increasingly important public health concern. Perimenopausal syndrome refers to a cluster of symptoms related to declining ovarian function, often characterized by autonomic dysregulation and neuropsychiatric manifestations[1]. Among these, depressive symptoms are particularly common, affecting approximately 15%-50% of perimenopausal women and occurring at a rate substantially higher than in premenopausal populations[2]. In parallel, sleep disturbances are highly prevalent during this period and are closely linked to emotional dysregulation, resulting in a significant reduction in quality of life[2-4].

Mindfulness, defined as a non-judgmental awareness of present-moment experiences, has emerged as an important psychological construct in contemporary psychiatry. Originating from contemplative traditions and later integrated into Western clinical practice, mindfulness is now widely used as a self-regulation strategy to improve emotional and cognitive functioning[5,6]. Higher levels of mindfulness have been associated with reduced depressive symptoms and improved psychological resilience[7]. In addition, mindfulness has been linked to better sleep outcomes, potentially through mechanisms such as reduced cognitive arousal and decreased emotional reactivity[8].

Sleep quality is a key indicator of mental health and plays a central role in emotional regulation. Poor sleep and depressive symptoms are closely related and may reinforce each other over time[9,10]. However, the present study focused on depressive symptoms as the mediator between mindfulness and sleep quality. This mediation model was informed by theoretical and empirical evidence suggesting that mindfulness primarily influences psychological well-being through improvements in emotional regulation. Higher levels of mindfulness may reduce depressive symptoms by decreasing rumination and emotional reactivity, whereas depressive symptoms are known to be strongly associated with sleep disturbance through mechanisms such as cognitive hyperarousal and negative affect. Therefore, depressive symptoms may serve as a plausible intermediary linking mindfulness and sleep quality, particularly among perimenopausal women who are vulnerable to both mood and sleep disturbances. Existing evidence suggests that mindfulness may affect sleep through several psychological mechanisms, including attenuation of pre-sleep cognitive arousal, reduced rumination, enhanced emotional regulation, and alleviation of depressive symptoms[8,9]. Systematic reviews and meta-analyses have shown that mindfulness-based and related psychological interventions may improve sleep quality and reduce depression and anxiety symptoms in adult populations[11-13]. However, the existing literature has several limitations. Although associations among mindfulness, depressive symptoms, and sleep quality have been reported in previous studies, much of the evidence has focused on intervention outcomes, general adult populations, or pairwise relationships between these variables. Relatively few studies have examined these factors simultaneously within a single analytical framework among perimenopausal women, a population characterized by unique hormonal, psychological, and sleep-related changes. Consequently, it remains unclear whether depressive symptoms may represent an important indirect pathway linking mindfulness and sleep quality in this population. Examining these relationships within a single analytical framework may provide a more comprehensive understanding of the psychological factors associated with sleep quality during the perimenopausal transition. Although mindfulness may be conceptually linked to sleep quality through depressive symptoms, alternative explanations should also be considered. Depressive symptoms and sleep disturbances are known to have bidirectional relationships, and both may be influenced by shared underlying factors such as hormonal changes, menopausal symptom severity, anxiety, and other psychosocial stressors. Therefore, the proposed mediation model represents one of several possible conceptual frameworks for understanding these associations in perimenopausal women.

Against this background, the present study examined the associations among mindfulness, depressive symptoms, and sleep quality in perimenopausal women, and tested an a prespecified mediation model in which depressive symptoms were hypothesized to mediate the association between mindfulness and sleep quality. This model was selected a priori based on theoretical frameworks that conceptualize mindfulness as a psychological resource that is closely related to emotional regulation and depressive symptoms, which in turn are strongly associated with sleep disturbances. Nevertheless, alternative directional and reciprocal models remain plausible and could not be distinguished using the present cross-sectional data.

MATERIALS AND METHODS

Study design and participants

This observational study was conducted among perimenopausal women attending a tertiary hospital in Zhengzhou, Henan Province, China, between October 2023 and December 2024. Participants were recruited using a convenience sampling approach. Potential participants were identified during outpatient visits and screened for eligibility by trained researchers. Eligible women who provided written informed consent were invited to complete the study questionnaires.

Sample size adequacy was evaluated based on the recommendation that multiple linear regression analyses require at least 10-15 participants per predictor variable. Given the number of predictors included in the regression model, the final sample of 693 participants was considered sufficient to ensure adequate statistical power. A post hoc power analysis using G*Power 3.1 indicated that the sample size of 693 provided statistical power greater than 0.99 to detect a small-to-moderate effect size (f² = 0.05) at α = 0.05 in the regression analyses.

Inclusion and exclusion criteria

Inclusion criteria: (1) Age 45-55 years; (2) Perimenopausal status defined according to the Stages of Reproductive Aging Workshop criteria, including women in the menopausal transition or within 12 months after the final menstrual period; (3) Intact uterus and at least one preserved ovary; (4) Ability to understand and complete the questionnaire; and (5) Provision of written informed consent.

Exclusion criteria: (1) Current or recent use of hormone therapy or psychotropic medications; (2) Presence of malignant tumors or severe chronic diseases; (3) Medical conditions that may mimic perimenopausal symptoms, such as hyperthyroidism, tuberculosis, diabetes, or chronic infections; (4) History of substance abuse or dependence; and (5) Severe psychiatric disorders, such as schizophrenia or bipolar disorder, or high suicide risk.

Data collection

Data were collected through face-to-face administration of structured questionnaires by trained researchers. Participants completed the questionnaires independently; assistance was provided when necessary. Completed questionnaires were checked for completeness and accuracy at the time of collection.

A total of 735 questionnaires were distributed. Completed questionnaires were reviewed for completeness and consistency at the time of collection. Questionnaires containing substantial missing data or invalid responses were excluded during data screening. Ultimately, 693 valid questionnaires were included in the final analysis, yielding an effective response rate of 94.3%.

Measures

General characteristics: A self-designed questionnaire was used to collect demographic and clinical information, including age, body mass index (BMI), marital status, reproductive history, medical history, and lifestyle factors.

Depressive symptoms: Depressive symptoms were assessed using the validated Chinese version of the 10-item short form of the Center for Epidemiologic Studies Depression Scale (CES-D-10)[14]. The CES-D-10 was selected because it provides a brief and validated measure of depressive symptoms that reduces respondent burden while maintaining good reliability and validity. Previous studies have demonstrated strong agreement between the CES-D-10 and the full Short Form of the Center for Epidemiologic Studies Depression Scale. Each item is scored from 0 to 3, with total scores ranging from 0 to 30. Higher scores indicate more severe depressive symptoms. The Cronbach’s α coefficient in this study was 0.910.

Mindfulness: Mindfulness was measured using the Mindful Attention Awareness Scale[15]. The scale consists of 15 items rated on a 6-point Likert scale, with total scores ranging from 15 to 90. Higher scores indicate greater mindfulness. The validated Chinese version demonstrated good reliability in this study, with a Cronbach’s α coefficient of 0.890.

Sleep quality: Sleep quality was evaluated using the validated Chinese version of the Pittsburgh Sleep Quality Index (PSQI)[16]. The scale assesses sleep quality over the past month across seven components, with total scores ranging from 0 to 21. Higher scores indicate poorer sleep quality. The Cronbach’s α coefficient in this study was 0.842.

Statistical analysis

Data were analyzed using SPSS version 20.0 and Mplus version 8.3. Continuous variables were presented as mean ± SD or median (interquartile range), depending on distribution. Categorical variables were expressed as n (%).

Because several study variables were not normally distributed, Spearman correlation analysis was used to assess associations among mindfulness, depressive symptoms, and sleep quality. Multiple linear regression analysis was performed to identify factors independently associated with sleep quality, with PSQI score entered as the dependent variable. Mindfulness and depressive symptoms were entered as the primary independent variables. Based on previous literature and univariate analyses, age, BMI, marital status, income level, number of pregnancies, and complications were included as covariates in the regression model to control for potential confounding effects. All variables were entered simultaneously into the multivariable model to estimate their independent associations with sleep quality. Prior to model estimation, key regression assumptions, including linearity, normality of residuals, homoscedasticity, independence of errors, and multicollinearity, were evaluated to ensure the appropriateness of the analyses.

Mediation analysis was conducted using Mplus version 8.3 to examine whether depressive symptoms mediated the association between mindfulness and sleep quality. Indirect effects were estimated using a bias-corrected bootstrap procedure with 5000 resamples, and statistical significance was determined based on 95% bootstrap confidence intervals. The primary mediation model was specified a priori as mindfulness to depressive symptoms to sleep quality, based on the theoretical role of mindfulness in emotional regulation and depressive symptoms. Given the cross-sectional nature of the data, mediation analyses were interpreted as tests of indirect associations rather than evidence of causal direction. Missing data were handled using a complete-case approach, and only participants with complete data on study variables were included in the analyses. A two-sided P < 0.05 was considered statistically significant.

RESULTS

Descriptive statistics of depressive symptoms, mindfulness, and sleep quality

A total of 693 perimenopausal women were included in the analysis. The median CES-D-10 score was 6, with an interquartile range of 4-9. The median Mindful Attention Awareness Scale score was 74, with an interquartile range of 64-83. The median PSQI score was 6, with an interquartile range of 4-9 (Table 1).

Table 1 The score of depression, mindfulness level and sleep quality in 693 perimenopausal women, median (interquartile rage).

Item	Score
CES-D	6 (4-9)
MAAS	74 (64-83)
PSQI	6 (4-9)

CES-D: The Short Form of the Center for Epidemiologic Studies Depression Scale; MAAS: Mindful Attention Awareness Scale; PSQI: Pittsburgh Sleep Quality Index.

Differences in depressive symptoms across participant characteristics

Depressive symptom scores differed significantly across age groups, BMI categories, menopausal status, and presence of complications (all P < 0.05). No significant differences were observed with respect to ethnicity, educational level, smoking status, alcohol consumption, age at menarche, number of pregnancies, income level, irregular vaginal bleeding, or surgical history (all P > 0.05; Table 2).

Table 2 Comparison of depression scores in 693 perimenopausal women with different general data, median (interquartile rage).

Variable	Subgroup	Depression score	H value	P value
Age (years)	< 50	5 (3-8)	6.294	0.012
	≥ 50	6 (4-9)
BMI (kg/m2)	< 18.5	9 (6-10)	9.976	0.018
	18.5 ≤ BMI < 24	6 (4-8.5)
	24 ≤ BMI < 28	6 (4-9)
	≥ 28	5 (3-8)
Ethnicity	Han	6 (4-9)	0.120	0.728
	Other	5.5 (4-9.25)
Educational level	Primary school and below	6 (3-9)	0.382	0.944
	Junior high school	6 (4-9)
	Technical secondary school/senior high school	6 (4-8)
	College and above	5 (4-8)
Smoking status	Current smoker	7 (3-11)	1.602	0.449
	Former smoker	6 (6-8.25)
	Never smoker	6 (4-9)
Drinking status	Current drinker	6 (4-10)	0.869	0.648
	Former drinker	6 (4-8)
	Never drinker	6 (4-9)
Menarche age (years)	Before 15	6 (4-9)	0.042	0.837
	After 15	6 (4-9)
Number of pregnancies	< 5	5 (4-8)	20.74	0.267
	≥ 5	7 (5-11)
Personal income (yuan)	< 2000	6 (4-9)	1.618	0.655
	2000-3999	6 (4-9)
	4000-5999	6 (4-9.5)
	≥ 6000	5 (4-8)
Irregular vaginal bleeding	Yes	6 (4-9)	0.423	0.515
	No	6 (4-8.5)
Menopausal status	Menopausal	6 (4-9)	5.447	0.019
	Non-menopausal	6 (3-8)
Complications	Yes	6 (4-9)	3.935	0.047
	No	6 (4-8)
Surgical history	Yes	5 (4-8)	1.059	0.303
	No	6 (4-9)

BMI: Body mass index.

Correlations among mindfulness, depressive symptoms, and sleep quality

Spearman correlation analysis showed that depressive symptoms were negatively correlated with mindfulness (r = -0.444, P < 0.01) and positively correlated with PSQI scores (r = 0.422, P < 0.01), indicating poorer sleep quality. Mindfulness was negatively correlated with PSQI scores in the unadjusted correlation analysis (r = -0.268, P < 0.01; Table 3).

Table 3 Spearman correlation analysis of depression, mindfulness level and sleep quality in 693 perimenopausal women (r value).

Item	CES-D	MAAS	PSQI
CES-D	1.000a	-0.444a	0.422a
MAAS	-0.444a	1.000a	-0.268a
PSQI	0.422a	-0.268a	1.000a

^aP < 0.01.

CES-D: The Short Form of the Center for Epidemiologic Studies Depression Scale; MAAS: Mindful Attention Awareness Scale; PSQI: Pittsburgh Sleep Quality Index.

Factors associated with sleep quality

In the unadjusted regression analysis, mindfulness showed a very weak negative association with sleep quality (B = -0.036), whereas depressive symptoms showed a positive association with sleep quality (B = 0.343). In the multiple linear regression model adjusting for depressive symptoms and other covariates, mindfulness remained negatively associated with sleep quality, and the magnitude of this association increased (β = -0.038), while the positive association between depressive symptoms and sleep quality became weaker (β = 0.320). In addition, age ≥ 50 years, BMI categories, marital status, income level, and number of pregnancies were also significantly associated with PSQI scores (Table 4). Multicollinearity diagnostics indicated no significant collinearity among variables (all variance inflation factor values < 5). The final model explained a significant proportion of variance in sleep quality (R² = 0.226, adjusted R² = 0.294).

Table 4 Multiple linear regression analysis of influencing factors of sleep quality in 693 perimenopausal women.

Variable	Subgroup	B	β	t value	P value
MAAS		-0.038	-0.126	-3.346	< 0.001
CES-D		0.320	0.384	10.187	< 0.001
Age (years)	≥ 50	0.677	0.090	2.651	0.008
BMI (kg/m2)	18.5 ≤ BMI < 24	-2.409	-0.326	-2.871	0.004
	24 ≤ BMI < 28	-2.318	-0.303	-2.745	0.006
	≥ 28	-3.217	-0.267	-3.555	0.001
Marital status	Married	0.656	0.105	3.070	0.002
Personal income (yuan)	2000-3999	0.135	0.016	0.438	0.661
	4000-5999	0.788	0.078	2.127	0.034
	≥ 6000	0.797	0.074	2.003	0.046
Number of pregnancies	3-4	0.507	0.068	1.776	0.076
	≥ 5	1.613	0.165	4.231	< 0.001
Complications	No	-0.472	-0.063	-1.842	0.066

MAAS: Mindful Attention Awareness Scale; CES-D: The Short Form of the Center for Epidemiologic Studies Depression Scale; BMI: Body mass index.

The primary mediation model was a simple observed-variable mediation model. Therefore, global model fit indices were examined when applicable but were not used as the primary basis for interpretation. The mediation results were interpreted according to the estimated path coefficients, indirect effect, confidence interval, and the prespecified theoretical framework.

Mediation analysis demonstrated that mindfulness was significantly associated with depressive symptoms (β = -0.416, P < 0.01), and depressive symptoms were significantly associated with PSQI scores (β = 0.423, P < 0.01). The direct association between mindfulness and sleep quality was smaller but remained statistically significant (β = -0.101, P < 0.01).

Further analysis showed that depressive symptoms significantly mediated the association between mindfulness and sleep quality. The indirect effect was -0.176 (95% confidence interval: -0.222 to -0.131), accounting for 63.5% of the total effect (Table 5). Given the cross-sectional design, this mediation result was interpreted as an indirect association consistent with the prespecified theoretical model rather than evidence of causal direction.

Table 5 Mediation effect analysis of depression between mindfulness level and sleep quality in 693 perimenopausal women.

Effect type	Effect value	Standard error	95%CI	Effect proportion (%)
Direct effect	-0.101	0.041	-0.181 to -0.021	36.46
Indirect effect	-0.176	0.023	-0.222 to -0.131	63.53
Total effect	-0.277	0.042	-0.359 to -0.195

All path coefficients are statistically significant (P < 0.01). Path coefficients: Mindfulness level to depression (-0.416); depression to sleep quality (0.423); mindfulness level to sleep quality (-0.101). CI: Confidence interval.

DISCUSSION

Depressive symptoms in perimenopausal women: Perimenopausal women are particularly vulnerable to depressive symptoms due to neuroendocrine changes and psychosocial stressors. In this study, 19.19% of participants exhibited depressive symptoms, which is higher than previously reported in community-based samples[17]. This finding is consistent with previous literature reporting elevated rates of depressive symptoms during the perimenopausal transition. Although neuroendocrine factors such as fluctuations in estrogen levels have been proposed as potential contributors to mood disturbances[18,19], the present study did not assess biological markers. Therefore, the underlying mechanisms cannot be evaluated directly and should be interpreted cautiously.

In addition to biological hypotheses proposed in previous studies, depressive symptoms in this population may also be associated with psychosocial factors, including life stress, role transitions, and declining physical health[3,4]. These findings support a biopsychosocial framework for understanding mental health in perimenopausal women and highlight the importance of early identification and intervention.

CONCLUSION

Depressive symptoms significantly mediated the association between mindfulness and sleep quality in perimenopausal women. Higher levels of mindfulness were associated with lower depressive symptoms and better sleep quality in the prespecified mediation model, suggesting that depressive symptoms may partly explain the relationship between mindfulness and sleep quality. Given the cross-sectional design, these findings should be interpreted as indirect associations rather than evidence of causal direction. Future longitudinal and intervention studies are needed to confirm the temporal sequence among mindfulness, depressive symptoms, and sleep quality and to evaluate whether further investigation warrant potential strategies for supporting sleep and psychological well-being[13].

ACKNOWLEDGEMENTS

The authors thank all study participants for their contributions and the research team for their assistance with data collection.