Causal Association Between Sleep Traits and Endometriosis: A Mendelian Randomization Study

Introduction

Endometriosis, characterized by the presence of endometrial-like tissue outside the uterus,¹ affects approximately 6–10% of women of reproductive age globally.² Beyond causing symptoms such as pelvic pain, dysmenorrhea and infertility,^3,4 endometriosis is associated with various long-term complications, including cardiovascular diseases, autoimmune diseases, and ovarian cancer.^5–8 Anatomically, endometriosis can be classified based on lesion location, including uterus, ovary, vagina, intestine, peritoneum, and unspecified sites. The symptoms vary slightly by subtype. Despite extensive research into its pathophysiology, the etiology of endometriosis remains unknown.⁹

Insomnia, one of the most prevalent sleep disorders globally, affects an estimated 30–50% of adults at some point in their lives,¹⁰ with women demonstrating a significantly higher susceptibility than men.¹¹ Clinically, insomnia is defined by difficulties in initiating or maintaining sleep, early morning awakening, and, in some cases, nonrestorative or poor-quality sleep.^11–15 Emerging evidence suggests a potential association between insomnia and the occurrence of endometriosis. This association may be mediated through shared neuroendocrine pathways, given that insomnia frequently co-occurs with emotional disorders such as anxiety and depression.¹⁶ Notably, population-based studies indicate that individuals with stress-related disorders are at a higher risk of subsequent endometriosis diagnosis.¹⁷ This susceptibility may stem from the activation of the sympathetic nervous system (SNS) and the hypothalamic-pituitary-adrenal (HPA) axis in response to stress.¹⁸ The SNS subsequently promotes systemic inflammation via catecholamine release,¹⁹ while the HPA axis stimulates excessive glucocorticoid production, potentially driving inflammation and autoimmune diseases.²⁰ These clinical observations corroborated by rodent studies, which demonstrates that chronic stress exposure accelerates endometriosis progression and exacerbates lesion severity.^21,22

Beyond insomnia, other sleep traits including getting up in the morning, nap during day, sleep disorders, chronotype, snoring, short sleep duration, and long sleep duration may affect reproductive hormone regulation through disruptions in the sleep-wake cycle.²³ Irregular sleep-wake rhythms have been shown to impair estrogen synthesis and metabolism,^24,25 thereby increasing the risk of estrogen-related disorders, including endometriosis.²⁶

Mendelian randomization (MR) is an epidemiological method that uses genetic variants as instrumental variables (IVs) to assess causal associations between exposures and outcomes.²⁷ In this study, we employed a two-sample MR analysis to evaluate the causal relationship between sleep traits and endometriosis. By utilizing genetic variants identified in genome-wide association studies (GWASs), we aimed to determine whether sleep traits are a risk factor for endometriosis, with profound implications for prevention and therapeutic strategies.

Methods

Overview

We conducted a Two-Sample Mendelian randomization (TSMR) using summary level data from Genome-wide meta-analysis association studies.²⁸ The MR analysis relies on three core assumptions. Assumption 1, named relevance hypothesis, states that the selected IVs are strongly related to insomnia. Assumption 2, named independence hypothesis, requires that these IVs remain unaffected by potential confounders. Assumption 3, named exclusionary hypothesis, posits that these IVs affect endometriosis solely through insomnia-related pathways.^29,30

The study included a discovery analysis and a validation analysis. In the discovery analysis, we investigated the causal associations between sleep characteristics and endometriosis, including its anatomical subtypes: uterine, ovarian, vaginal, intestinal, peritoneal, and unspecified endometriosis. In the validation analysis, the causal association between insomnia and endometriosis was further verified. The overall study design is shown in Figure 1.

Figure 1 Framework for Evaluating the Causal Impact of Insomnia on Endometriosis Subtypes.

Exposures

The primary exposure (insomnia) data were derived from the recent large-scale meta-analysis of GWASs conducted by the UK Biobank and 23andMe consortium.²⁸ This analysis incorporated 593,724 insomnia cases and 1,771,286 controls, using a fixed-effect model in METAL. We identified 554 single-nucleotide polymorphisms (SNPs) associated with insomnia (Supplementary Table S1). Given that sex-specific meta-analyses revealed no significant differences in results between males and females, combined-sex summary statistics were employed to maximize statistical power. Additionally, to mitigate the impact of linkage disequilibrium (LD) on analysis outcomes, we set an r² threshold of 0.001 and SNP distance of 10,000 kb.

The secondary exposures included seven sleep-related traits: getting up in the morning, nap during day, sleep disorders, chronotype, snoring, short sleep duration, and long sleep duration (Table 1). Instrumental variables for these traits were obtained from UK Biobank and FinnGen [IEU OpenGWAS project (mrcieu.ac.uk)]. For each trait, we selected SNPs meeting genome-wide significance (P<5e-6) and performed LD clumping with a threshold of 10,000 kb and r² < 0.001.

Table 1 The Data Source of Genetic Instruments for Exposures and Outcomes

Outcomes

For outcome variables, we obtained GWAS summary statistics from the FinnGen Consortium [IEU Open GWAS project (mrcieu.ac.uk)] of European population, which comprised genetic and health data from 500,000 Finnish biobank participants. The FinnGen makes use of extensive longitudinal registry data of all Finnish residents, all of which are publicly available. The dataset included GWAS statistics for endometriosis (N=8288) and its subtypes: uterine (N=2372), ovarian (N=3231), vaginal (N=1360), unspecified (N=1435), peritoneal (N=177), and intestinal endometriosis (N=2953). A summary of GWAS data sources is provided in Table 1.

Statistical Analysis

MR analyses were performed using the TwoSample MR package in R version 4.2.3. Five common MR methods were utilized, including inverse variance weighted (IVW), weighted median,³¹ simple mode, weighted mode, and MR-Egger regression. The IVW method, implemented with a random-effects model to calculate a weighted average of the Wald ratio estimates, served as the primary analytical approach. To ensure the robustness, sensitivity analyses were performed to assess heterogeneity and horizontal pleiotropy. Heterogeneity was evaluated using Cochrane’s Q statistic, with a significance threshold of P < 0.05. Notably, the random-effects IVW model inherently accommodates potential heterogeneity. Horizontal pleiotropy was examined through MR-Egger regression, which tests whether IVs influence the outcome via pathways independent of sleep traits—a violation of MR assumptions 2 and 3. Furthermore, MR-pleiotropic Residuals and Outliers (MR-PRESSO) analysis was employed to identify and remove outlier SNPs (P < 0.05) exhibiting horizontal pleiotropy, thereby minimizing their confounding effects.³² Finally, a leave-one-out analysis was performed to evaluate whether the overall results were disproportionately driven by any single SNP. This comprehensive approach ensured the validity and reliability of the MR findings.

Independent Validation Analysis

To verify the causal association between insomnia and endometriosis, we performed a TSMR analysis using an independent, publicly available database. Genetic instruments for insomnia (n=10,894,596) and endometriosis (n=9,983,671) were obtained from the UK Biobank. We applied the same TSMR methodologies as those used in the discovery analysis.

Result

MR Analysis of Causal Links Between Sleep Traits and the Risk of Endometriosis

In the UK Biobank and 23andMe meta-analysis, a total of 544 independent SNPs (P < 5e-8) significantly associated with insomnia were selected as instrumental variables (IVs). As shown in Figure 2, IVW results indicated that insomnia was associated with increased risk of endometriosis (OR = 1.83, 95% CI: 1.34–2.50), uterine endometriosis (OR = 2.22, 95% CI: 1.30–3.80), ovarian endometriosis (OR = 1.61, 95% CI: 1.03–2.54), vaginal endometriosis (OR > 1.00, 95% CI: 1.03–3.79) and unspecified endometriosis (OR = 2.70, 95% CI: 1.45–5.02). However, no significant associations were observed for intestinal endometriosis (OR = 0.96, 95% CI: 0.17–5.41) and peritoneal endometriosis (OR = 1.49, 95% CI: 0.95–2.35). These findings were corroborated by supplementary analyses using weighted median, simple mode, weighted mode, and MR-Egger regression (Supplementary Tables S2–S9).

Figure 2 Forest plot of causality between insomnia and endometriosis.

From the IEU openGWAS database, we screened independent SNPs associated with sleep traits (P < 5e-6), identifying 24 SNPs for sleep disorders, 201 for nap during day, 209 for getting up in the morning, 64 for chronotype, 131 for snoring, 11 for short sleep duration, and 18 for long sleep duration (Supplementary Table S10). Sleep disorders were associated with a reduced risk of unspecified endometriosis (OR = 0.74, 95% CI: 0.58–0.96). Getting up in the morning was associated with a reduced risk of ovarian endometriosis (OR = 0.56, 95% CI: 0.36–0.87). Sleep duration was associated with a reduced risk of intestinal endometriosis (OR = 0.02, 95% CI: < 0.001–0.86) (Figure 3).

Figure 3 Forest plot of causality between sleep traits (except insomnia) and endometriosis.

Sensitivity Analysis

To assess the robustness, we performed horizontal pleiotropy and heterogeneity tests (Table 2). The F statistic for insomnia-IVs was 49.8 (F > 10), indicating strong instrumental variables. No horizontal pleiotropy was observed for endometriosis and its anatomical subtypes (Egger intercept P > 0.05 for all outcomes). The Cochrane Q statistic indicated significant heterogeneity in the analyses of endometriosis, uterine endometriosis, and ovarian endometriosis (P < 0.05 for both MR Egger Q and IVW Q). Subsequent leave-one-out analysis and MR-PRESSO method did not identify any specific IVs driving heterogeneity, suggesting that the causal estimates remained robust. (Supplementary Figure 1A–G).

Table 2 Sensitivity MR Analyses Between Insomnia and Endometriosis Subtypes

Validation Analysis

MR approaches validated the association between insomnia and endometriosis (Table 3). Consistent effects were observed via IVW method (OR = 1.01, 95% CI: 1.00–1.01), weighted median (OR = 1.01, 95% CI: 1.00–1.03), and MR-Egger regression (OR = 1.00, 95% CI: 1.00–1.01), strengthening evidence for causality.

Table 3 Validation MR Analyses Between Insomnia and Endometriosis

Discussion

In this study, we used TSMR Analysis to investigate the causal associations between sleep characteristics and endometriosis. Our results showed that insomnia was associated with an increased risk of overall endometriosis, as well as specific anatomical subtypes including uterine, ovarian, vaginal and unspecified endometriosis. These findings provide valuable evidence supporting preventive strategies and deepened our understanding of the potential role of sleep disorders in the etiology of endometriosis.

Our findings regarding the causal effect of insomnia on the risk of endometriosis are consistent with a recent Mendelian randomization study,³³ which demonstrated that insomnia is a risk factor for overall endometriosis. However, our study extends this knowledge by further stratifying the analysis based on anatomical subtypes. We found that the impact of insomnia varies across different sites, showing a significant causal link specifically with uterine, ovarian, and unspecified endometriosis. This subtype-specific analysis provides novel insights into the heterogeneity of endometriosis etiology. Furthermore, previous epidemiological studies attributing endometriosis risk to sleep issues have largely focused on night-shift work, emphasizing circadian rhythm disruption as the primary mechanism.^34–39 Notably, insomnia characterized by difficulties in both sleep initiation and maintenance involves distinct pathophysiological mechanisms beyond mere circadian disruption. Specifically, these mechanisms include dysregulated HPA axis activity and elevated cytokine levels. Previous studies suggested that abnormal HPA axis biomarkers related to insomnia were significantly associated with endometriosis,⁴⁰ indicating a mechanistic pathway independent of occupational exposures. Consequently, these findings position insomnia as a modifiable risk factor, providing a new direction for the primary prevention of endometriosis.

Current evidence suggests insomnia may contribute to endometriosis through neuroendocrine dysregulation.^40–43 Normal sleep progression is mediated by melatonin, a chronobiotic hormone synthesized predominantly in the pineal gland that governs circadian rhythmicity.^44–46 Critically, melatonin dysregulation has been implicated in the pathogenesis of endometriosis.^41–43 In endometriosis, melatonin enhances antioxidant defenses via the upregulation of superoxide dismutase (SOD) and catalase (CAT).^41,42 Furthermore, melatonin directly inhibits endometriotic cell invasion by modulating the balance of MMP-3/MMP-9 and TIMP-1/TIMP-3.⁴³ Clinical data reveal that sleep disturbances in patients with insomnia correlate with significant disruption in melatonin release.⁴⁷ Moreover, clinical studies demonstrate that women with endometriosis exhibit significantly lower melatonin levels compared to healthy controls.⁴⁸ These findings suggested that insomnia disrupts the integrated melatonin-mediated “circadian-antioxidant-invasive” regulatory network, providing a mechanistic basis for its role in endometriosis pathogenesis.

The insomnia-endometriosis connection may also operate through the “HPA axis activation-inflammation -endometriosis” cascade. Prolonged sleep disruption triggers hyperactivation of the HPA axis, elevating cortisol levels. As a stress hormone, sustained cortisol elevation is known to dysregulate immune function. Mechanistically, this may involve the persistent activation of inflammatory transcription factors like NF-κB, leading to the upregulation of pro-inflammatory genes. This neuroendocrine dysregulation also amplifies oxidative stress.²⁰ This pathological milieu creates a permissive microenvironment for endometriosis development: inflammatory cytokines (e.g., IL-6) promote angiogenesis and cell adhesion in ectopic lesions, while oxidative DNA damage facilitates endometrial cell survival outside the uterine cavity.^49,50 Together, these findings position insomnia as a significant contributor to endometriosis pathophysiology through interconnected neuroendocrine, inflammatory, and oxidative pathways.

Crucially, insomnia is a highly intervenable and modifiable risk factor. Identifying insomnia as a pathogenic driver suggests that sleep-focused interventions could serve as novel, non-hormonal strategies for the prevention of endometriosis or as adjunctive therapies to alleviate disease progression. This shifts the clinical paradigm from solely managing gynecological symptoms to addressing broader systemic dysregulation, underscoring the value of multidisciplinary management in women’s health.

In contrast to insomnia, other sleep characteristics including getting up in the morning, nap during day, sleep disorders, chronotype, snoring, short sleep duration, and long sleep duration showed no causal association with genetically predicted endometriosis risk. This lack of association might be explained by the fact that characteristics such as snoring and sleep duration share overlapping genetic factors with obesity or metabolic genes.^51,52 Although we conducted horizontal pleiotropic tests, it remained possible that unmeasured pleiotropy could interfere with MR assumptions, leading to false-negative findings. Additionally, while insomnia may promote disease progression through inflammatory or hormonal pathways, other characteristics may lack direct interaction evidence with the core mechanism of endometriosis. Future observational studies or multi-omics analyses will be necessary to clarify mechanistic links between these sleep characteristics and the development of endometriosis.

Our study has several notable strengths. The insomnia exposure data were derived from a large-scale GWAS meta-analysis, which increased statistical power and enhanced the precision and robustness of our results. Importantly, as genetic variants are unlikely to be influenced by environmental or lifestyle factors in the same way as traditional exposures, this MR design strengthens the reliability of our finding by minimizing confounding. Notably, our study extends current knowledge by providing the first genetic evidence linking insomnia to specific anatomical subtypes of endometriosis, where previous observational studies had failed to establish this association. Several limitations warrant consideration. Primarily, both the discovery and validation analysis were limited to individuals of European ancestry, potentially restricting the generalizability of the findings to other ethnic groups. Furthermore, despite sensitivity analyses, MR results may still be influenced by heterogeneity and reverse causality, which represent inherent methodological limitations of the MR approach.

Conclusion

Insomnia is a clinically significant and modifiable risk factor for the development of endometriosis. These results highlight the critical importance of screening and intervening for poor sleep quality, positioning insomnia as a viable target for personalized prevention strategies in women’s reproductive health.