Elevated Neutrophil Count as a Causal Risk Factor for Endometriosis: A Mendelian Randomization Study

Endometriosis is a chronic, estrogen-dependent inflammatory condition characterized by the presence of endometrial-like tissue outside the uterine cavity, primarily on pelvic organs and the peritoneum. 1 It affects approximately 10% of women of reproductive age, causing debilitating symptoms such as chronic pelvic pain, dysmenorrhea, dyspareunia, and infertility, thereby significantly impairing quality of life. 2 Despite its prevalence and impact, the pathophysiology of endometriosis is not fully understood, and current diagnostic and therapeutic strategies are limited. 3 The prevailing theory for its pathogenesis is Sampson’s theory of retrograde menstruation, where endometrial fragments are transported through the fallopian tubes into the peritoneal cavity. 4 However, as retrograde menstruation occurs in most women, other factors must be involved in the establishment and growth of endometriotic lesions. There is a growing body of evidence highlighting that a dysfunctional immune response and a chronic inflammatory state are central to the disease’s development. 5 The peritoneal fluid of women with endometriosis contains an elevated concentration of activated immune cells, including macrophages, neutrophils, and lymphocytes, along with a complex milieu of cytokines, chemokines, and growth factors that create a pro-inflammatory microenvironment conducive to the survival, adhesion, and proliferation of ectopic endometrial cells. 6 From an immunological perspective, endometriosis involves dysregulation of both innate and adaptive immunity. The innate immune system, as the first line of defense, plays a particularly critical role: macrophages exhibit impaired clearance of ectopic endometrial cells, natural killer (NK) cells show reduced cytotoxicity, and neutrophils contribute to the establishment of a pro-inflammatory peritoneal environment through the release of cytokines, reactive oxygen species, and neutrophil extracellular traps (NETs). This multifaceted immune dysfunction creates a permissive microenvironment that facilitates the implantation, survival, and proliferation of endometrial-like tissue outside the uterine cavity. Circulating blood cells are key components of the systemic immune and inflammatory response. Observational studies have reported alterations in peripheral blood cell counts in women with endometriosis. For instance, some studies have noted higher neutrophil-to-lymphocyte ratios (NLR) and platelet-to-lymphocyte ratios (PLR) in patients with endometriosis, suggesting a systemic inflammatory state. 7 Among these, neutrophils, as first responders of the innate immune system, have been particularly implicated. They are found in increased numbers in the peritoneal fluid and ectopic lesions of endometriosis patients and are known to secrete factors that promote angiogenesis and cell invasion, crucial processes for lesion establishment. 8 , 9 However, observational studies are susceptible to confounding factors and reverse causation, making it difficult to establish a true causal link. For example, it is unclear whether the observed alterations in blood cell counts are a primary factor contributing to the risk of endometriosis or merely a secondary consequence of the chronic inflammation associated with the disease. Distinguishing between causality and association is critical for understanding the disease’s etiology and for developing effective targeted interventions. Mendelian randomization (MR) is a powerful genetic epidemiological method that uses genetic variants, typically single nucleotide polymorphisms (SNPs), as instrumental variables (IVs) to infer the causal effect of an exposure on an outcome. 10 Because genetic variants are randomly allocated at conception, MR is less prone to confounding from environmental or lifestyle factors and is not affected by reverse causation, as the disease state cannot alter an individual’s germline DNA. 11 By leveraging summary statistics from large-scale genome-wide association studies (GWAS), two-sample MR allows for high-powered causal inference without requiring individual-level data. 12 The total white blood cell (WBC) count, a routine clinical marker readily available from complete blood counts, serves as a convenient and cost-effective indicator of systemic inflammation. Investigating WBC count alongside specific leukocyte subtypes in an MR framework allows us to determine whether observed associations reflect a generalized inflammatory state or are attributable to specific cell lineages, thereby offering more precise etiological insights. In this study, we aimed to dissect the causal relationship between specific circulating blood cell populations, with a focus on neutrophils, and endometriosis. We employed a bidirectional two-sample MR design to investigate: (1) the causal effects of six major blood cell types (basophils, eosinophils, lymphocytes, monocytes, neutrophils, and total white blood cells) on the risk of endometriosis; and (2) the potential reverse causal effect of a genetic predisposition to endometriosis on these blood cell counts. By clarifying these relationships, we hope to provide novel insights into the role of systemic inflammation in the pathogenesis of endometriosis and identify potential avenues for future therapeutic research.

This study utilized a bidirectional two-sample Mendelian randomization design to investigate the causal relationship between six circulating blood cell counts and endometriosis. The MR framework relies on three core assumptions for the genetic variants used as instrumental variables (IVs): (1) the IVs are robustly associated with the exposure (the relevance assumption); (2) the IVs are not associated with any confounding factors that could affect both the exposure and the outcome (the independence assumption); and (3) the IVs affect the outcome only through the exposure (the exclusion restriction assumption) 13 ( Figure 1 ). This bidirectional approach allows for the examination of causality in both directions, from blood cells to endometriosis and from endometriosis to blood cells, helping to disentangle the direction of the causal pathway. Considering that various confounding factors are closely related to the pathogenesis of endometriosis and blood cell traits, we conservatively removed SNPs that were associated with body mass index (BMI), smoking behavior, and reproductive factors at genome-wide significance levels (P < 5 × 10 −8 ). Data on these confounding factors were obtained from the GWAS Catalog ( https://www.ebi.ac.uk/gwas ) and the IEU OpenGWAS database ( https://gwas.mrcieu.ac.uk/ ). MR analyses were repeated after excluding these potentially confounding SNPs to verify the robustness of our findings. Figure 1 Workflow of the Mendelian randomization (MR) analyses in the present study. The three core instrumental-variable assumptions are illustrated: Assumption 1 (relevance), the genetic instruments are associated with the exposure; Assumption 2 (Independence), the genetic instruments are not associated with measured or unmeasured confounders; and Assumption 3 (exclusion restriction), the genetic instruments influence the outcome only through the exposure. Cross marks (×) indicate associations or pathways that must be absent for the corresponding instrumental-variable assumptions to be satisfied. A diagram showing Mendelian randomization analysis workflow with assumptions and pathways. The diagram illustrates the workflow of Mendelian randomization analysis, focusing on the causal relationship between circulating blood cell counts and endometriosis. On the left, a box labeled ′SNPs of Circulating Blood Cells (p less than 5e minus 8)′ is connected to ′Blood Cell Consortium′ below it. An arrow labeled ′Assumption 1′ points from this box to a central box labeled ′basophil, eosinophil, lymphocyte, monocyte, neutrophil and total white blood cell counts (Exposure and Outcome)′. A dashed line with a cross mark labeled ′Assumption 2′ extends from the left box to a top box labeled ′Confounders′. Another dashed line with a cross mark labeled ′Assumption 3′ extends from the central box to the left box. The central box is connected by an arrow labeled ′MR analysis′ to another central box labeled ′endometriosis (Exposure and Outcome)′, which is connected by a dashed line with a cross mark labeled ′Assumption 3′ back to the left box. On the right, a box labeled ′SNPs of endometriosis (p less than 5e minus 8)′ is connected to ′FinnGen consortium′ below it. An arrow labeled ′Assumption 1′ points from this right box to the central ′endometriosis′ box. A dashed line with a cross mark labeled ′Assumption 2′ extends from the right box to the ′Confounders′ box. Another dashed line with a cross mark labeled ′Assumption 3′ extends from the ′endometriosis′ box to the right box. The diagram visually represents the assumptions and pathways necessary for the Mendelian randomization analysis. Abbreviations : MR, Mendelian randomization; SNPs, single-nucleotide polymorphisms. Workflow of the Mendelian randomization (MR) analyses in the present study. The three core instrumental-variable assumptions are illustrated: Assumption 1 (relevance), the genetic instruments are associated with the exposure; Assumption 2 (Independence), the genetic instruments are not associated with measured or unmeasured confounders; and Assumption 3 (exclusion restriction), the genetic instruments influence the outcome only through the exposure. Cross marks (×) indicate associations or pathways that must be absent for the corresponding instrumental-variable assumptions to be satisfied. Summary-level data for the six blood cell traits were obtained from a large-scale GWAS meta-analysis by the Blood Cell Consortium, which included data from the UK Biobank and the Blood-Cell-Consortium. 14 The dataset comprised up to 563,946 individuals of European ancestry. The traits investigated were basophil count (GWAS ID: ieu-b-29), white blood cell count (ieu-b-30), monocyte count (ieu-b-31), lymphocyte count (ieu-b-32), eosinophil count (ieu-b-33), and neutrophil count (ieu-b-34). All summary statistics are publicly available from the IEU OpenGWAS project database ( https://gwas.mrcieu.ac.uk/ ). For the forward MR analysis (blood cells → endometriosis), GWAS summary statistics for endometriosis were obtained from the FinnGen consortium, release 12 (R12). FinnGen is a large-scale biobank study that combines genomic information with nationwide health registry data in Finland. 15 The endometriosis endpoint (ID: finngen_R12_N14_ENDOMETRIOSIS) included 18,349 cases and 208,924 controls of Finnish ancestry. For the reverse MR analysis (endometriosis → blood cells), the same FinnGen GWAS data were used as the exposure. A summary of the GWAS datasets used in this study is presented in Table 1 . Table 1 Summary of Genome-Wide Association Study (GWAS) Datasets Used in the Mendelian Randomization Analyses GWAS ID Trait Consortium/Study Population Sample Size Year ieu-b-29 Basophil cell count Blood Cell Consortium European 563,946 2020 ieu-b-30 White blood cell count Blood Cell Consortium European 563,946 2020 ieu-b-31 Monocyte cell count Blood Cell Consortium European 563,946 2020 ieu-b-32 Lymphocyte cell count Blood Cell Consortium European 563,946 2020 ieu-b-33 Eosinophil cell count Blood Cell Consortium European 563,946 2020 ieu-b-34 Neutrophil cell count Blood Cell Consortium European 563,946 2020 finngen_R12_N14_ENDOMETRIOSIS Endometriosis FinnGen (Release 12) Finnish 227,273 2024 Summary of Genome-Wide Association Study (GWAS) Datasets Used in the Mendelian Randomization Analyses For each exposure, we implemented a stringent procedure to select suitable genetic IVs. First, we identified SNPs robustly associated with the exposure trait at a genome-wide significance threshold (P < 5 × 10 −8 ). Second, to ensure the independence of the IVs, we performed clumping using the European 1000 Genomes Project reference panel. We applied a strict clumping algorithm with a linkage disequilibrium (LD) threshold of r 2 < 0.001 and a clumping window of 10,000 kb, effectively removing SNPs in high LD with more significant SNPs. Third, to minimize potential violations of the MR assumptions, we removed palindromic SNPs with ambiguous strand orientation and intermediate allele frequencies (minor allele frequency > 0.42). The F-statistic was calculated for each SNP using the formula F = (Beta/SE) 2 , and also for the overall instrument strength, to quantify the strength of the selected IVs and assess potential weak instrument bias. An F-statistic > 10 is conventionally considered indicative of a sufficiently strong instrument. 16 The primary causal estimates were calculated using the random-effects inverse-variance weighted (IVW) method. The IVW method combines the Wald ratio estimates (SNP-outcome effect divided by the SNP-exposure effect) for each IV in a meta-analysis framework, weighting each estimate by the inverse of its variance. 17 The random-effects model was chosen as the primary method to account for potential heterogeneity among the individual SNP effects. Causal estimates were reported as odds ratios (ORs) with 95% confidence intervals (CIs) for binary outcomes (endometriosis) and as beta coefficients for continuous outcomes (blood cell counts). To assess the robustness of our results and test for violations of MR assumptions, several sensitivity analyses were conducted. Weighted Median Method: This method provides a consistent causal estimate even if up to 50% of the information comes from invalid instruments. It calculates the median of the weighted causal estimates from each SNP, where the weight is proportional to the inverse variance of the instrument-outcome association. 18 MR-Egger Regression: This method can detect and adjust for directional horizontal pleiotropy, which occurs when genetic variants influence the outcome through pathways other than the exposure. The intercept of the MR-Egger regression provides an estimate of the average pleiotropic effect. A non-zero intercept (P < 0.05) suggests the presence of directional pleiotropy. The slope of the regression line provides a pleiotropy-corrected causal estimate. 19 Heterogeneity Test: Cochran’s Q statistic was calculated to assess the heterogeneity among the causal estimates from individual SNPs. Significant heterogeneity (P < 0.05) might indicate the presence of pleiotropy or that the variants influence the outcome through different mechanisms. Pleiotropy Assessment: In addition to the MR-Egger intercept test, we visualized the relationship between SNP effects on the exposure and outcome using scatter plots. Funnel plots were used to assess the symmetry of the causal estimates, where asymmetry may suggest directional pleiotropy. To specifically assess the exclusion restriction assumption—that the genetic instruments influence endometriosis only through the blood cell traits and not through alternative pathways—we interpreted the MR-Egger intercept test as an indicator of directional horizontal pleiotropy. A non-significant intercept (P > 0.05) suggests that the causal estimates are unlikely to be biased by widespread pleiotropic effects, supporting the validity of our primary findings. Leave-One-Out Analysis: To evaluate the influence of any single SNP on the overall causal estimate, we performed a leave-one-out sensitivity analysis. This involved systematically removing one SNP at a time and recalculating the IVW estimate with the remaining SNPs. This helps to identify if the result is driven by a single, potentially pleiotropic, genetic variant. The results for the primary MR analyses were corrected for multiple testing using the Benjamini-Hochberg false discovery rate (FDR) method. An FDR-corrected P-value (P-FDR) < 0.05 was considered statistically significant. P-values between 0.05 and the FDR threshold were considered suggestive of a causal association. All statistical analyses were performed using the TwoSampleMR package (version 0.5.6) in R (version 4.2.1).

After a rigorous selection and clumping process, we identified a set of independent, genome-wide significant SNPs as instrumental variables for each of the six blood cell traits and for endometriosis. The number of SNPs used as instruments for the forward MR analysis (blood cells on endometriosis) were: basophil count (189 SNPs), white blood cell count (455 SNPs), monocyte count (476 SNPs), lymphocyte count (470 SNPs), eosinophil count (425 SNPs), and neutrophil count (395 SNPs). For the reverse MR analysis (endometriosis on blood cells), the number of instrumental SNPs ranged from 22 to 25. The F-statistics for all instruments were substantially greater than 10, indicating that weak instrument bias was unlikely to affect our findings. A summary of the GWAS datasets used in this study is presented in Table 1 . The IVW analysis showed that genetically predicted higher neutrophil count was associated with an 11.3% increase in endometriosis risk per standard deviation increase in cell count (OR = 1.113, 95% CI: 1.038–1.194, P = 0.003). Similarly, higher white blood cell count conferred a 9.6% increased risk (OR = 1.096, 95% CI: 1.028–1.169, P = 0.005). Both associations remained significant after FDR correction (P-FDR = 0.015 for both). No statistically significant causal associations were found for the other four blood cell traits with endometriosis risk after FDR correction. Although the weighted median analysis showed a nominally significant association for basophil count (OR = 1.202, P = 0.010), this was not supported by the primary IVW analysis (P = 0.116) and did not survive multiple testing correction. For eosinophil, monocyte, and lymphocyte counts, the causal estimates were close to the null and not statistically significant. The detailed results of the forward MR analysis, including estimates from sensitivity analyses, are presented in Table 2 and Supplementary Table 1 , and visualized in the forest plot in Figure 2A . Table 2 Heterogeneity and Directional Pleiotropy Analyses for the Causal Effects of Blood Cell Counts on Endometriosis Exposure Method N SNPs Q statistic df Q p-value Egger Intercept SE p-value Basophil cell count IVW (random effects) 189 294.5 188 1.09E-06 – – – MR-Egger 189 293.51 187 1.04E-06 −0.002 0.0022 0.427 White blood cell count IVW (random effects) 455 655.72 454 1.59E-09 – – – MR-Egger 455 655.68 453 1.33E-09 0 0.0016 0.881 Monocyte cell count IVW (random effects) 476 690.45 475 3.46E-10 – – – MR-Egger 476 690.2 474 2.97E-10 −0.001 0.0012 0.677 Lymphocyte cell count IVW (random effects) 470 774.2 469 2.33E-17 – – – MR-Egger 470 773.81 468 1.95E-17 0.001 0.0016 0.629 Eosinophil cell count IVW (random effects) 425 593.17 424 9.94E-08 – – – MR-Egger 425 591.96 423 9.99E-08 −0.001 0.0015 0.353 Neutrophil cell count IVW (random effects) 395 598.47 394 1.23E-10 – – – MR-Egger 395 598.45 393 9.97E-11 0 0.0017 0.919 Notes : A Cochran’s Q p-value < 0.05 indicates significant heterogeneity across the instrument variants. An MR-Egger intercept p-value < 0.05 indicates the presence of directional pleiotropy. Figure 2 ( A ) Forest plot illustrating the causal effect of six blood cell traits on the risk of endometriosis. The squares represent the odds ratios (ORs) and the horizontal lines represent the 95% confidence intervals (CIs). The analysis shows a significant increased risk of endometriosis associated with higher neutrophil count and white blood cell count after FDR correction. ( B ) Forest plot of the reverse MR analysis showing the causal effect of genetic liability to endometriosis on six blood cell counts. The results indicate no significant causal effect of endometriosis on any of the investigated blood cell traits. Beta represents the standard deviation change in cell count per log-odds increase in endometriosis. Two forest plots of blood cell traits and endometriosis, highlighting two elevated odds ratios. The image A showing a forest plot with a table at left and point estimates with horizontal confidence intervals at right. Table columns read, Exposure, No. of SNPs, OR (95 percent CI), P value and P (FDR). The x-axis label is, Odds Ratio (Log Scale), with ticks at 0.90, 1.0, 1.1 and 1.2. A vertical reference line is at 1.0. Rows: Neutrophil cell count, 395, 1.113 (1.038 to 1.194), 0.003, 0.015; White blood cell count, 455, 1.096 (1.028 to 1.169), 0.005, 0.015; Eosinophil cell count, 425, 0.955 (0.903 to 1.009), 0.102, 0.174; Basophil cell count, 189, 1.079 (0.981 to 1.185), 0.116, 0.174; Monocyte cell count, 476, 1.013 (0.965 to 1.062), 0.609, 0.641; Lymphocyte cell count, 470, 0.985 (0.924 to 1.050), 0.641, 0.641. The image B showing a forest plot with a table at left and point estimates with horizontal confidence intervals at right. Table columns read, Outcome (Blood Cell), No. of SNPs, Beta (95 percent CI), P value and P (FDR). The x-axis label is, Beta (SD change in cell count per log-odds increase in endometriosis), with ticks at minus 0.04, minus 0.02, 0, 0.02 and 0.04. A vertical reference line is at 0. Rows: Monocyte cell count, 24, minus 0.012 (minus 0.027, 0.002), 0.101, 0.606; Eosinophil cell count, 23, minus 0.005 (minus 0.018, 0.008), 0.457, 0.922; Basophil cell count, 24, 0.005 (minus 0.009, 0.020), 0.486, 0.922; Neutrophil cell count, 25, 0.002 (minus 0.013, 0.017), 0.801, 0.922; Lymphocyte cell count, 22, 0.001 (minus 0.016, 0.018), 0.901, 0.922; White blood cell count, 22, minus 0.001 (minus 0.017, 0.015), 0.922, 0.922. Heterogeneity and Directional Pleiotropy Analyses for the Causal Effects of Blood Cell Counts on Endometriosis Notes : A Cochran’s Q p-value < 0.05 indicates significant heterogeneity across the instrument variants. An MR-Egger intercept p-value < 0.05 indicates the presence of directional pleiotropy. ( A ) Forest plot illustrating the causal effect of six blood cell traits on the risk of endometriosis. The squares represent the odds ratios (ORs) and the horizontal lines represent the 95% confidence intervals (CIs). The analysis shows a significant increased risk of endometriosis associated with higher neutrophil count and white blood cell count after FDR correction. ( B ) Forest plot of the reverse MR analysis showing the causal effect of genetic liability to endometriosis on six blood cell counts. The results indicate no significant causal effect of endometriosis on any of the investigated blood cell traits. Beta represents the standard deviation change in cell count per log-odds increase in endometriosis. The MR-Egger regression intercept did not significantly deviate from zero for any of the blood cell traits (P-intercept > 0.05 for all), suggesting no significant directional pleiotropy. While there was evidence of heterogeneity for all traits as indicated by Cochran’s Q test (P < 0.001), the use of the random-effects IVW model appropriately accounts for this. The leave-one-out analyses did not identify any single SNP that disproportionately influenced the causal estimates for neutrophil or white blood cell count, confirming the robustness of these findings. We next tested whether a genetic predisposition to endometriosis causally affects blood cell counts. The IVW analysis found no evidence for such a relationship: all P-values exceeded 0.10, with the highest for white blood cell count (P = 0.922). After FDR correction, all associations remained non-significant. These results suggest that the inflammatory state associated with endometriosis is unlikely to be the cause of the systemic alterations observed in these circulating blood cell populations. The detailed results for the reverse MR analysis are presented in Table 3 , Figure 2B and Supplementary Table 2 . Sensitivity analyses for the reverse direction also showed no evidence of pleiotropy (MR-Egger P-intercept > 0.05 for all traits). Table 3 Heterogeneity and Directional Pleiotropy Analyses for the Causal Effects of Endometriosis on Blood Cell Counts Outcome Method N SNPs Q Statistic df Q p-value Egger Intercept SE p-value Basophil cell count IVW (random effects) 24 55.4 23 1.71E-04 – – – MR-Egger 24 55.3 22 1.08E-04 −0.0004 0.002 0.838 White blood cell count IVW (random effects) 22 67.38 21 9.19E-07 – – – MR-Egger 22 61.83 20 3.70E-06 −0.0029 0.0021 0.195 Monocyte cell count IVW (random effects) 24 67.12 23 3.36E-06 – – – MR-Egger 24 65.38 22 3.45E-06 0.0015 0.002 0.453 Lymphocyte cell count IVW (random effects) 22 72.91 21 1.19E-07 – – – MR-Egger 22 70.18 20 1.70E-07 −0.0021 0.0023 0.388 Eosinophil cell count IVW (random effects) 23 42.99 22 0.005 – – – MR-Egger 23 40.12 21 0.007 0.0022 0.0018 0.234 Neutrophil cell count IVW (random effects) 25 79.12 24 8.39E-08 – – – MR-Egger 25 78.37 23 5.83E-08 −0.001 0.0022 0.641 Notes : A Cochran’s Q p-value < 0.05 indicates significant heterogeneity across the instrument variants. An MR-Egger intercept p-value < 0.05 indicates the presence of directional pleiotropy. Heterogeneity and Directional Pleiotropy Analyses for the Causal Effects of Endometriosis on Blood Cell Counts Notes : A Cochran’s Q p-value < 0.05 indicates significant heterogeneity across the instrument variants. An MR-Egger intercept p-value < 0.05 indicates the presence of directional pleiotropy. Scatter plots for the significant associations ( Figure 3A for white blood cell count and Figure 3B for neutrophil count) show a consistent positive trend in the effect estimates. The corresponding forest plots, leave-one-out plots, and funnel plots for these significant findings are provided in Figures 4 and 5 , and Supplementary Figure 1 . Detailed SNP-level information is provided in Supplementary Table 3 and Supplementary Table 4 . Figure 3 Scatter plot of SNP effects on ( A ) white blood cell count and ( B ) neutrophil cell count versus their effects on endometriosis. Each dot represents a single SNP. The slopes of the lines represent the causal estimates from different MR methods. Scatter plots showing SNP effects on white blood cell and neutrophil counts versus endometriosis effects. The image features two scatter plots, A and B, showing SNP effects on cell counts and endometriosis. Plot A′s horizontal axis shows SNP effects on white blood cell count, while the vertical axis shows effects on endometriosis. A dense cluster near the origin indicates SNPs with minimal effects. Causal estimates are represented by lines from various Mendelian Randomization methods: inverse variance weighted, MR Egger, weighted median and weighted mode. Plot B depicts SNP effects on neutrophil cell count (horizontal) and endometriosis (vertical), with a similar dense cluster and MR method lines. Both plots suggest a positive trend in effect estimates, with lines indicating different causal estimates. The legend clarifies the color coding for each MR method line. Figure 4 Leave-one-out sensitivity analysis for the effect of ( A ) white blood cell count and ( B ) neutrophil cell count on endometriosis. Each point represents the IVW estimate after removing the specified SNP, demonstrating that the overall result is not driven by any single variant. The corresponding full-size leave-one-out plots are provided in Supplementary Figure 2 . Two forest plots of leave one out Mendelian randomization effect estimates by single nucleotide polymorphism. The images A and B display forest plots of Mendelian randomization leave-one-out effect estimates by single nucleotide polymorphism (SNP). Both plots have an x-axis labeled ′MR leave one out effect estimate′ ranging from 0.00 to 0.15, with ticks at 0.00, 0.05, 0.10 and 0.15. Image A lists SNPs such as rs1886654, rs507778 and others, ending with rs1371794. Image B includes SNPs like rs114050631, rs16850073 and concludes with rs1371794. Each SNP is represented by a dot with a horizontal confidence interval line. In both images, dots cluster between 0.09 and 0.10. In Image A, rs1886654 is near 0.10 and rs1371794 is near 0.08. In Image B, rs2301557 and rs1371794 are both near 0.08. Figure 5 Funnel plots for assessing heterogeneity and pleiotropy in the Mendelian randomization analysis of ( A ) white blood cell count and ( B ) neutrophil cell count on endometriosis. The symmetrical distribution of the SNP-specific estimates around the inverse-variance weighted (IVW) estimate suggests no substantial directional pleiotropy. β represents the estimated genetic effect size (regression coefficient) of the exposure on endometriosis. The corresponding high-resolution funnel plots are provided in Supplementary Figure 3 . Genetic effect vs precision scatter plots for white blood cell and neutrophil counts in endometriosis. The images A and B display scatter plots assessing heterogeneity and pleiotropy in Mendelian randomization analysis. The horizontal axis represents the estimated genetic effect size (beta) of the exposure on endometriosis, while the vertical axis represents precision, shown as 1 divided by standard error. Both plots feature a funnel-shaped distribution of data points, denser near the center and sparser towards the edges, indicating symmetry around zero. This suggests minimal directional pleiotropy. The legend includes ′Inverse variance weighted′ and ′MR Egger′ methods, both lines positioned near zero, indicating similar estimates. The plots aim to illustrate the consistency of effect estimates, with no substantial bias evident. Scatter plot of SNP effects on ( A ) white blood cell count and ( B ) neutrophil cell count versus their effects on endometriosis. Each dot represents a single SNP. The slopes of the lines represent the causal estimates from different MR methods. Leave-one-out sensitivity analysis for the effect of ( A ) white blood cell count and ( B ) neutrophil cell count on endometriosis. Each point represents the IVW estimate after removing the specified SNP, demonstrating that the overall result is not driven by any single variant. The corresponding full-size leave-one-out plots are provided in Supplementary Figure 2 . Funnel plots for assessing heterogeneity and pleiotropy in the Mendelian randomization analysis of ( A ) white blood cell count and ( B ) neutrophil cell count on endometriosis. The symmetrical distribution of the SNP-specific estimates around the inverse-variance weighted (IVW) estimate suggests no substantial directional pleiotropy. β represents the estimated genetic effect size (regression coefficient) of the exposure on endometriosis. The corresponding high-resolution funnel plots are provided in Supplementary Figure 3 .

Our study provides genetic evidence that an elevated neutrophil count is a causal risk factor for endometriosis. This positions systemic innate inflammation—with neutrophils as a key mediator—as a contributor to disease etiology, not just an afterthought. Clinically, systemic inflammatory profiles might one day help with risk stratification, though further validation is needed. More immediately, our results highlight neutrophil-mediated inflammatory pathways as a target for future therapeutic and preventive strategies. Upcoming work should focus on the functional mechanisms linking neutrophils to endometriosis and whether modulating neutrophil activity could be a viable treatment approach.

This bidirectional MR study provides genetic evidence that an elevated neutrophil count causally increases endometriosis risk. A similar effect was seen for total white blood cell count, with no evidence of reverse causality—indicating that these hematological changes contribute to pathogenesis rather than simply reflecting the disease. Our finding that an elevated neutrophil count is a causal risk factor for endometriosis fits with the growing evidence that links neutrophils in the disease’s inflammatory cascade. Neutrophils are a cornerstone of the innate immune system and are typically the first cells to be recruited to sites of inflammation. 20 In the context of endometriosis, neutrophils are found in increased numbers within the peritoneal fluid and the ectopic endometrial lesions themselves. 8 They contribute to the pro-inflammatory microenvironment by releasing a variety of mediators, including cytokines (eg, IL-8, a potent neutrophil chemoattractant), reactive oxygen species (ROS), and proteolytic enzymes. 21 More recently, studies have highlighted that Neutrophil Extracellular Traps (NETs) are elevated in the peritoneal fluid of women with endometriosis, where activated neutrophils release a web-like structure of DNA, histones, and granular proteins that can trap pathogens but also promote inflammation and tissue damage. 22 Our MR study provides robust genetic support for the hypothesis that a systemic predisposition to higher neutrophil levels is not just correlated with, but is causally involved in, the initiation or progression of the disease. This may occur by enhancing the initial inflammatory response to ectopic endometrial cells in the peritoneum, promoting their survival, implantation, and subsequent neovascularization. 23 This specificity for neutrophils, as opposed to a generalized effect across all leukocyte subtypes, suggests that the observed causal association reflects a neutrophil-specific inflammatory mechanism rather than a non-specific, systemic inflammatory state. The causal association we identified for total white blood cell count is likely driven by its neutrophil component, as neutrophils are the most abundant type of leukocyte. This finding reinforces the broader concept that systemic innate immune activation is a key etiological factor. The lack of a causal association for other immune cell types, such as lymphocytes and monocytes, in our study is also informative. While these cells are known to be present in endometriotic lesions and contribute to the local immune milieu, our results suggest that their systemic baseline levels, as determined by common genetic variants, may not be a primary driver of disease risk. It is possible that the local recruitment and activation of these cells within the peritoneal cavity, rather than their circulating numbers, are more critical to pathogenesis. The fact that monocyte and lymphocyte counts did not show significant causal effects further supports the specificity of the neutrophil-mediated pathway in endometriosis pathogenesis, pointing to a distinct innate immune mechanism rather than a global inflammatory predisposition. The bidirectional design is a key strength, as it lets us test for reverse causation. The null finding in the reverse direction—no effect of endometriosis liability on blood cell counts—argues against the idea that systemic inflammation is merely a downstream consequence of the disease. Observational studies cannot easily make this distinction. If endometriosis itself drove neutrophil increases, the reverse MR would have detected it. Instead, our results point to a primary role for innate immune variation in predisposing individuals to endometriosis, suggesting the inflammatory phenotype may precede clinical disease. Our findings agree with prior observational studies reporting elevated inflammatory markers like C-reactive protein (CRP) and altered immune profiles in women with endometriosis. 24–26 What our work adds is specific causal evidence for neutrophils, clarifying the direction of this relationship. By using the latest FinnGen data—one of the largest GWAS datasets available for endometriosis—we strengthen the evidence for this causal pathway. The MR design is the main strength here, as it reduces confounding and reverse-causation biases that plague observational studies. We used the largest available GWAS for both blood cell traits and endometriosis, and multiple sensitivity analyses (MR-Egger, weighted median, pleiotropy and heterogeneity tests) support the robustness of our key findings. However, several limitations should be considered. First, our analyses were conducted using data primarily from individuals of European ancestry, which may limit the generalizability of our findings to other populations. Future studies in diverse ancestral groups are needed. Second, MR analysis is subject to its own assumptions. While we found no evidence of directional horizontal pleiotropy through the MR-Egger intercept test, we cannot completely rule out the possibility of balanced or other complex forms of pleiotropy. Third, GWAS summary statistics typically represent an average effect across a population and do not allow for stratification by disease stage or subtype, which could be influenced by different inflammatory pathways. The endometriosis data from FinnGen, while large, is based on electronic health records, which might have different diagnostic accuracy compared to surgically confirmed cohorts. Finally, our study points to a causal role for the number of circulating neutrophils but does not elucidate the specific functional changes in these cells that drive the disease process. It is also worth noting that neutrophil-driven inflammatory mechanisms have been implicated in other gynecological conditions, including polycystic ovary syndrome (PCOS), where elevated neutrophil-to-lymphocyte ratios have been reported, and in pelvic inflammatory disease. However, the extent to which the neutrophil-mediated causal pathway identified in this study is specific to endometriosis or represents a shared inflammatory mechanism across gynecological diseases remains to be determined. Future MR studies comparing multiple gynecological outcomes using the same analytical framework would help clarify the disease specificity of this neutrophil-driven causal trend.