The association between sedentary behavior and infertility: a population-based cross-sectional observational study and Mendelian randomization analysis.

This cross-sectional analysis utilized publicly available data from the National Health and Nutrition Examination Survey (NHANES) ( http://www.cdc.gov/nchs/nhanes.htm ), a nationally representative survey designed to evaluate the health and nutritional status of the U.S. population. The NHANES data collection involves comprehensive questionnaires, physical examinations, household interviews, and laboratory tests. Ethical approval for the survey was granted by the Institutional Review Board (IRB) of the National Center for Health Statistics (NCHS), with all participants providing written informed consent. For the current study, we utilized data from the 2013–2014, 2015–2016, and 2017–2018 cycles. Female participants aged 20–49 years were selected, and those with incomplete data on infertility, daily sitting time, relevant covariates, as well as women with a history of hysterectomy or oophorectomy were excluded. A final cohort of 2904 participants was included in the analysis, as shown in Fig.  1 . Fig. 1 Flow chart of the study population. Describes how the sample of participants was composed. NHANES, National Health and Nutrition Examination Survey Flow chart of the study population. Describes how the sample of participants was composed. NHANES, National Health and Nutrition Examination Survey In the next phase, we performed Mendelian Randomization (MR) analysis leveraging summary statistics from a Genome-Wide Association Study (GWAS) to evaluate the causal impact of genetically predicted sedentary behavior on infertility. In our study, sedentary behavior was defined based on the self-reported daily sitting time [ 14 , 15 ], measured by the following question: “How much time do you usually spend sitting (or reclining) on a typical day?” This referred to the number of waking hours spent sitting or reclining at school, at home, getting toand from places, or with friends including time spent sitting at adesk, traveling in a car or bus, reading, playing cards, watchingtelevision, or using a computer. The time spent sleeping is not included [ 16 , 17 ]. Responses were quantified as daily hours. In alignment with prior research on sedentary behavior, we classified sitting time using a binary approach: less than 6 h per day versus 6 h per day or more [ 18 , 19 ]. Furthermore, we applied a four-category classification, subdividing daily sitting time into: 0 to < 4 h, 4 to < 6 h, 6 to 8 h, and ≥ 8 h per day [ 20 , 21 ], with the 0 to < 4 h group serving as the reference category. The outcome variable was the presence of infertility, defined based on participants’ responses. Individuals were classified as having a history of infertility if they answered “yes” to the following questions: “Have you ever tried to get pregnant for at least one year without success? (RHQ074)” [ 22 , 23 ]. Participants who refused to answer or were uncertain about their infertility status were excluded from the analysis to ensure the accuracy and completeness of the data. These variables were included as covariates in the data analysis, as they may impact the relationship between daily sitting time and infertility. The covariates considered were age (in years), race/ethnicity, marital status, education level, smoking status, alcohol intake, family income-to-poverty ratio (PIR), body mass index (BMI), hypertension, diabetes, pelvic infection disease (PID) and age of menarche. Race/ethnicity was categorized as Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, and other races. Marital status was classified as married, widowed, divorced, separated, never married, or cohabiting with a partner. Education level was grouped as less than high school, high school, high school graduate, some college, or college graduate and above. Family PIR was grouped as less than 2 times the poverty line or ≥ 2 times the poverty line. BMI was divided into three categories: underweight/normal (< 25.0 kg/m²), overweight (25.0–30.0 kg/m²), and obese (≥ 30 kg/m²). Smoking status was categorized as never, former, or current based on participant responses to “Do you smoke now?“. Never smokers comprised participants reporting either no lifetime cigarette use or consumption of fewer than 100 cigarettes prior to assessment. Former smokers were defined as those having smoked > 100 cigarettes historically but having ceased tobacco use by the time of interview. Current smokers denoted individuals actively consuming cigarettes during the study period with a lifetime intake of ≥ 100 cigarettes. Alcohol intake was categorized as: Never drinkers: Lifetime alcohol intake < 12 drinks; Former drinkers: Prior consumption without intake during the preceding year; Heavy drinkers: Women consuming ≥ 3 daily drinks or ≥ 4 drinks/occasion on ≥ 5 monthly occasions; Moderate drinkers: Women consuming ≥ 2 daily drinks or ≥ 4 drinks/occasion on ≥ 2 monthly occasions; Mild drinkers: Consumption patterns not fulfilling above criteria [ 24 , 25 ]. Diabetes mellitus diagnosis required fulfillment of at least one criterion: physician diagnosis, glycohemoglobin (HbA1c) ≥ 6.5%, fasting plasma glucose ≥ 7.0 mmol/L, random or 2-hour post-OGTT plasma glucose ≥ 11.1 mmol/L, or current use of glucose-lowering agents/insulin. Hypertension history was determined through affirmative responses to the question: “Has a physician or healthcare provider ever diagnosed you with hypertension?”, resulting in a dichotomous (yes/no) classification. PID was identified by the self-reported questions from Reproductive Health Questionnaire (RHQ078): “Have you ever been treated for a pelvic infection/PID [ 26 ]?” The age of menarche was categorized according to clinical practice into two groups: <15 years and ≥ 15 years. Our analysis also accounted for several key reproductive factors, including live birth outcomes, number of pregnancies, prior consultation for infertility (yes/no), history of contraceptive pill use, sexually active status, and previous sexually transmitted infection diagnoses (chlamydia, gonorrhea, genital warts, genital herpes, or HPV). In addition, social and behavioral confounders such as Healthy Eating Index (HEI) and trouble sleeping (yes/no) were considered. Detailed protocols and measurement procedures for all variables are publicly available at the National Health and Nutrition Examination Survey website ( http://www.cdc.gov/nchs/nhanes/ ). Our study drew on data from a meta-analysis of genome-wide association studies (GWAS) across 51 individual studies, with a focus on three sedentary behavior-related physical activity indicators: leisure screen time (LST), sedentary commuting and sedentary behavior at work [ 27 ]. We obtained genetic association summary statistics related to female infertility from the FinnGen GWAS, which included data from a European cohort tested in 2024, comprising 18,189 cases and 30,160 controls. Detailed information is available in Supplementary Table (1) To identify genetic instruments for Mendelian randomization (MR) analysis of sedentary behavior and infertility, we applied a genome-wide significance threshold ( P < 5 × 10⁻⁸) to select robustly associated variants. Linkage disequilibrium (LD) pruning (r² < 0.001 within 10,000 kb) was performed, followed by extraction of remaining SNPs from GWAS summary statistics. We computed F-statistics to assess instrument strength, excluding weak instruments (F < 10). Detailed data are shown in the Supplementary Table (2) For palindromic SNPs, the alignment function was applied when frequency information was available and the allele frequency was not close to 0.5. SNPs with missing frequency data or those identified as ambiguous were excluded. We ensured that all SNPs were aligned in the same direction across both the exposure and outcome datasets. The harmonization process aligned alleles based on strand orientation, ensuring consistency in effect sizes. Additionally, SNPs exhibiting heterogeneity, as identified using the RadialMR R package, were excluded. SNPs with intermediate allele frequencies (>0.42) were also discarded to meet the strong instrument criteria for Mendelian randomization analyses. Ultimately, 104 leisure screen time, 4 sedentary commuting and 21 sedentary behavior at work-associated SNPs were employed in causal inference analyses. This study follows the STROBE-MR reporting guidelines, and the completed checklist is provided in Supplementary Table 3. All statistical analyses were conducted using the complex survey design framework of NHANES (2013–2018), incorporating sample weights derived from the Mobile Examination Center (MEC) exams, pseudo–primary sampling units (PSUs; sdmvpsu), and pseudostrata (sdmvstra) to account for the multistage stratified sampling structure across survey cycles. Participants were categorized according to infertility status. Weighted Student’s t-tests and weighted Chi-square tests were applied to compare continuous and categorical variables, respectively. Continuous variables were expressed as survey-weighted means and standard errors (SEs), while categorical variables were presented as weighted percentages with 95% confidence intervals (CIs). Weighted logistic regression models were used to estimate odds ratios (ORs) and 95% CIs for the association between daily sitting time and infertility. Both unadjusted and multivariable-adjusted models were employed: Model I was unadjusted; Model II included adjustments for age, race, marital status, education level, family PIR, and BMI; Model III further adjusted for smoking status and alcohol intake; and Model IV adjusted for hypertension, diabetes, pelvic infection disease and age of menarche in addition to the variables in Model III. Subgroup analysis was conducted to evaluate the robustness of the association between daily sitting time and infertility. Restricted cubic spline regression was employed to explore potential non-linear relationships between daily sitting time and infertility. We conducted an additional supplementary sensitivity analysis to further assess the robustness of our findings. Specifically, we included additional reproductive, behavioral, and social covariates—including live birth outcomes, number of pregnancies, prior consultation for infertility, history of contraceptive pill use, previous sexually transmitted infection diagnoses (chlamydia, gonorrhea, genital warts, genital herpes, or HPV), Healthy Eating Index (HEI), and trouble sleeping. Additionally, we conducted a two-sample Mendelian Randomization (MR) analysis using GWAS data. In this analysis, Inverse Variance Weighting (IVW) was applied as the primary method to assess genetically predicted causal relationships [ 28 ]. Complementary analyses included the weighted median, MR-Egger, and weighted mode [ 29 , 30 ]. Sensitivity analyses were essential to evaluate potential heterogeneity and horizontal pleiotropy. The Cochran Q test was performed to assess heterogeneity in effect sizes among genetic instrumental variables. The MR-Egger intercept was used to estimate the level of horizontal pleiotropy, with P < 0.05 indicating potential pleiotropic effects. Leave-one-out analysis was also conducted to determine whether the results were significantly influenced by any single SNP. In multiple testing, an adjusted p -value after Bonferroni correction, p -value < 0.05 / 3 ≈ 0.0167 was considered statistically significant. Statistical analyses were performed using R software ( http://www.R-project.org , The R Foundation) and Free Statistics software versions 1.9.2. For the Mendelian randomization (MR) analysis, the TwoSample MR package in R was used to conduct the MR analyses, utilizing summary-level data from genome-wide association studies (GWAS).

The baseline characteristics of the participants included in this study are summarized in Table  1 . The final analytic sample comprised 2,904 female participants from NHANES 2013–2018, representing approximately 45.2 million U.S. adults. Among them, 393 reported a history of infertility and 2,511 did not. Compared with women without infertility, those with infertility were generally older and had higher body mass indexes (BMIs). They were also more likely to be married, to smoke, to engage in sedentary behaviors, and to have a higher prevalence of hypertension and pelvic inflammatory disease (PID) ( P  < 0.05). Table 1 Weighted baseline characteristics of participants Variables Total Fertility Infertility P -value n  = 2904 n  = 2511 n  = 393 Age , mean (SE) , years 33.93 (8.71) 33.48 (8.78) 36.66 (7.72) < 0.001 Race , n (%) 0.15  Mexican American 472 (10.9) 414 (11.1) 58 (9.6)  Other Hispanic 279 ( 6.9) 249 (7.2) 30 (5.1)  Other Hispanic 1008 (59.1) 854 (58.2) 154 (64.6)  Non-Hispanic Black 628 (12.8) 543 (12.9) 85 (11.9)  Other Race 517 (10.3) 451 (10.6) 66 (8.8) Marital Status , n (%) < 0.001  Married 1295 (47.3) 1053 (44.0) 242 (66.6)  Widowed 21 ( 0.6) 17 (0.4) 4 (2.0)  Divorced 216 ( 7.1) 184 (7.2) 32 (6.7)  Separated 105 ( 2.9) 92 (2.9) 13 (2.7)  Never married 870 (28.6) 807 (31.3) 63 (13.0)  Living with partner 397 (13.5) 358 (14.2) 39 (9.1) Education level , n (%) 0.705  Less Than 9th Grade 133 ( 2.8) 121 (3.0) 12 (1.8)  9-11th Grade 272 ( 6.8) 232 (6.6) 40 (7.6)  High School 546 (18.6) 472 (18.4) 74 (19.7)  Some College 1072 (35.1) 923 (35.0) 149 (35.4)  College Graduate or above 881 (36.8) 763 (36.9) 118 (35.6) Family PIR (%) 2.82 (1.66) 2.79 (1.66) 3.00 (1.64) 0.052 BMI , mean (SE) , kg/m 2 29.47 (8.52) 29.09 (8.26) 31.77 (9.63) 0.001 Smoking status , n (%) 0.009  Never 2052 (67.9) 1799 (68.9) 253 (62.1)  Former 333 (13.6) 274 (13.0) 59 (17.6)  Current 519 (18.5) 438 (18.2) 81 (20.3) Alcohol intake , n (%) 0.120  Never 481 (12.3) 426 (12.6) 55 (10.0)  Former 177 ( 5.2) 144 (4.7) 33 (8.1)  Mild 804 (27.5) 695 (27.6) 109 (26.4)  Moderate 761 (29.7) 663 (29.9) 98 (28.2)  Heavy 681 (25.4) 583 (25.1) 98 (27.3) Hypertension , n (%) 0.001  No 2544 (88.9) 2222 (90.1) 322 (82.1)  Yes 360 (11.1) 289 (9.9) 71 (17.9) Diabetes , n (%) 0.057  No 2665 (93.5) 2314 (94.0) 351 (90.6)  Yes 239 ( 6.5) 197 (6.0) 42 (9.4) Pelvic infection disease , n (%) < 0.001  No 2765 (95.9) 2407 (96.7) 358 (90.8)  Yes 139 ( 4.1) 104 (3.3) 35 (9.2) Age of menarche (SE) , years 12.61 ( 1.73) 12.61 ( 1.71) 12.61 ( 1.87) 0.960 Sedentary (SE) , h 6.51 ( 3.40) 6.43 ( 3.36) 6.97 ( 3.61) 0.011 Data are presented as unweighted number (weighted percentage) for categorical variables and weighted mean (SE) for continuous variable PIR poverty income ratio, BMI body mass index, SE standard error; P  < 0.05 presents significant difference Weighted baseline characteristics of participants Data are presented as unweighted number (weighted percentage) for categorical variables and weighted mean (SE) for continuous variable PIR poverty income ratio, BMI body mass index, SE standard error; P  < 0.05 presents significant difference To further explore the potential association between daily sitting time and infertility, we employed weighted multivariable logistic regression analyses, constructing four sequential models with varying levels of covariate adjustment. The corresponding effect estimates—including odds ratios (ORs), 95% confidence intervals (CIs), and P values—are detailed in Table  2 . Weighted multivariable logistic regression analysis demonstrated that daily sitting time was positively associated with infertility in the unadjusted model (Model I: OR = 1.05, 95% CI: 1.01–1.08, P  = 0.010). In Model IV, which adjusted for all covariates, no statistically significant association was found between daily sitting time and infertility (Model IV: OR = 1.03, 95% CI: 1.00–1.07, P  = 0.066). This lack of association remained consistent when daily sitting time was categorized as either a binary variable or based on quartiles. As shown in Supplementary Fig. 1, the results did not indicate a significant non-linear association between daily sitting time and infertility (after adjusting age, race, marital status, education level, family PIR, BMI, smoking status, alcohol intake, hypertensive, diabetes, pelvic infection disease, age of menarche). Table 2 Multivariate regression analysis of the association between daily sitting time and infertility Variable Model I a P Model II b P Model III c P Model IV d P OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) daily sitting time(h) 1.05 (1.01 ~ 1.08) 0.010 1.04 (1.00 ~ 1.08) 0.032 1.04 (1.00 ~ 1.08) 0.034 1.03 (1.00 ~ 1.07) 0.066 daily sitting time(h)  =6 h 1.29 (1.02 ~ 1.63) 0.037 1.29 (1.00 ~ 1.65) 0.047 1.30 (1.01 ~ 1.68) 0.043 1.28 (0.99 ~ 1.66) 0.061 daily sitting time(h)  < 4 h 1(Ref) 1(Ref) 1(Ref) 1(Ref)  4 to < 6 h 0.77 (0.54 ~ 1.10) 0.152 0.79 (0.54 ~ 1.15) 0.202 0.76 (0.52 ~ 1.11) 0.149 0.74 (0.50 ~ 1.09) 0.117  6 to 8 h 0.97 (0.64 ~ 1.49) 0.902 1.05 (0.66 ~ 1.66) 0.825 1.04 (0.65 ~ 1.67) 0.855 1.04 (0.64 ~ 1.68) 0.881  > 8 h 1.19 (0.91 ~ 1.54) 0.194 1.17 (0.87 ~ 1.57) 0.286 1.16 (0.86 ~ 1.57) 0.299 1.12 (0.82 ~ 1.52) 0.460 Trend test 0.042 0.074 0.073 0.121 a Model I: no adjusted; b Model II: adjusted for age + race + marital status + education level + family PIR + BMI; c Model III: Model II + smoking status + alcohol intake; d Model IV: Model III + hypertension + diabetes + pelvic infection disease + age of menarche Multivariate regression analysis of the association between daily sitting time and infertility a Model I: no adjusted; b Model II: adjusted for age + race + marital status + education level + family PIR + BMI; c Model III: Model II + smoking status + alcohol intake; d Model IV: Model III + hypertension + diabetes + pelvic infection disease + age of menarche To further assess the robustness of the association between daily sitting time and infertility, we conducted subgroup analyses, as shown in Fig.  2 . After adjustment for age, race, marital status, education level, family PIR, BMI, smoking status, alcohol intake, hypertension, diabetes, pelvic inflammatory disease, and age at menarche, no statistically significant interactions were observed across any subgroup (all P for interaction > 0.05). Fig. 2 Subgroup analysis for the association between daily sitting time and infertility. Note: Adjusted for age, race, marital status, education level, BMI, family PIR, hypertension, diabetes, pelvic infection disease, age of menarche Subgroup analysis for the association between daily sitting time and infertility. Note: Adjusted for age, race, marital status, education level, BMI, family PIR, hypertension, diabetes, pelvic infection disease, age of menarche In addition, a sensitivity analysis was performed, as shown in Table  3 , incorporating extended reproductive, behavioral, and social covariates into the adjusted model to further assess the robustness of the findings. The results remained consistent with those of the primary model, thereby reinforcing the stability and reliability of our conclusions. Table 3 Sensitivity analysis of the association between daily sitting time and infertility Variable Model I a P Model II b P OR (95%CI) OR (95%CI) daily sitting time(h) 1.10 (1.05 ~ 1.15) < 0.010 0.99 (0.88 ~ 1.12) 0.833 daily sitting time(h)  =6 h 1.55 (1.20 ~ 2.01) 0.002 0.78 (0.43 ~ 1.41) 0.332 daily sitting time(h)  < 4 h 1(Ref) 1(Ref)  4 to  8 h 1.61 (1.12 ~ 2.34) 0.013 0.99 (0.24 ~ 4.10) 0.984 Trend test 0.002 0.691 a Model I: no adjusted; b Model II: adjusted for age + marital status + education level + BMI + smoking status + alcohol intake + hypertension + diabetes + pelvic infection disease + age of menarche + live birth outcomes + number of pregnancies + prior consultation for infertility + history of contraceptive pill use + previous sexually transmitted infection diagnoses + HEI + trouble sleeping Sensitivity analysis of the association between daily sitting time and infertility a Model I: no adjusted; b Model II: adjusted for age + marital status + education level + BMI + smoking status + alcohol intake + hypertension + diabetes + pelvic infection disease + age of menarche + live birth outcomes + number of pregnancies + prior consultation for infertility + history of contraceptive pill use + previous sexually transmitted infection diagnoses + HEI + trouble sleeping Given the absence of a statistically significant association between daily sitting time and infertility risk in the previous weighted multivariate logistic regression analysis, we next performed a Mendelian randomization (MR) analysis to further investigate the potential causal relationship between sedentary behavior and infertility risk. After applying the Bonferroni correction, the results from the inverse variance weighted (IVW) analysis (as shown in Table  4 ) provide no statistically robust evidence to support a causal relationship between genetically predicted leisure screen time (OR = 1.11, 95% CI: 0.10–1.24, P  = 0.052), sedentary commuting (OR = 1.19, 95% CI: 0.88–1.62, P  = 0.257), or sedentary behavior at work (OR = 0.99, 95% CI: 0.83–1.19, P  = 0.930) and infertility. The scatter plot is shown in Fig.  3 , which visually illustrates the impact of these exposure factors on infertility and was evaluated through different MR methods. Sensitivity analysis showed that the MR-Egger intercept method did not detect significant horizontal pleiopotency ( P  > 0.05). Cochran’s Q test using individually selected SNPS showed significant heterogeneity in leisure screen time ( P  < 0.05), as shown in Supplementary Table 4. The “leave-one-out” method was used to evaluate the impact of individual SNPS, and the results showed that there was no significant effect on the size of the effect, as shown in Fig.  4 . Table 4 Mendel randomization analysis results Exposure Method SNPs OR (95%CI) P -value leisure screen time IVW 104 1.11(0.10–1.24) 0.052 MR Egger 104 0.88(0.52–1.48) 0.634 Weighted median 104 1.13(0.97–1.30) 0.109 Weighted median 104 1.30(0.88–1.91) 0.194 sedentary commuting IVW 4 1.19(0.88–1.62) 0.257 MR Egger 4 2.39(0.84–6.78) 0.243 Weighted median 4 1.12(076−1.64) 0.565 Weighted median 4 1.07(0.63–1.81) 0.829 sedentary behavior at work IVW 21 0.99(0.83–1.19) 0.930 MR Egger 21 1.50(0.78–2.88) 0.236 Weighted median 21 1.07(0.84–1.35) 0.586 Weighted median 21 1.09(0.75–1.57) 0.658 SNP single nucleotide polymorphism, OR odds ratio, 95%CI 95% confidence interval Mendel randomization analysis results SNP single nucleotide polymorphism, OR odds ratio, 95%CI 95% confidence interval Fig. 3 Scatter plots of results from Mendelian randomization analysis. Note: A leisure screen time; B sedentary commuting; C sedentary behavior at work Scatter plots of results from Mendelian randomization analysis. Note: A leisure screen time; B sedentary commuting; C sedentary behavior at work Fig. 4 “Leave-one-out”forest map of the causal relationship. Note: A leisure screen time; B sedentary commuting; C sedentary behavior at work “Leave-one-out”forest map of the causal relationship. Note: A leisure screen time; B sedentary commuting; C sedentary behavior at work

According to our research, no statistically significant evidence was found to support the association between sedentary behavior and infertility. Future research involving larger cohorts with harmonized sedentary behavior phenotypes and stronger genetic instruments is warranted to validate and expand upon these results.

In this study, we integrated observational data from the nationally representative NHANES 2013–2018 cross-sectional survey with two-sample Mendelian randomization (MR) analyses to examine the relationship between sedentary behavior and infertility. Overall, no statistically significant association was observed between sedentary behavior and infertility in the general population. Sedentary behavior is widely recognized as one of the most common lifestyle patterns of the 21st century [ 31 ]. Some researchers suggest that women who maintain sedentary lifestyles throughout their reproductive years may experience hormonal dysregulation, and that prolonged inactivity could accelerate the decline in ovarian reserve, thereby impairing fertility [ 32 ]. Despite a growing body of evidence underscoring the general health risks associated with sedentary behavior, research examining its specific role in infertility remains scarce and yields inconsistent conclusions. In the present study, after adjusting for multiple confounders and performing extensive sensitivity analyses, we found no statistically significant association between self-reported sedentary behavior and infertility among women of reproductive age. Our results are in agreement with several population based investigations that similarly failed to identify a clear link between sedentary behavior and female reproductive capacity. For example, Russo et al. conducted a prospective cohort study among women with a history of pregnancy loss and observed no significant relationship between sedentary behavior and fecundability [ 10 ]. Likewise, Esmaeilzadeh et al. reported no notable differences in physical activity patterns or body-mass index between fertile and infertile women in northern Iran [ 11 ]. Conversely, a number of clinical and case–control studies have yielded opposite results. Foucaut et al. demonstrated that prolonged sedentary behavior and low physical activity were associated with idiopathic infertility in both men and women, potentially mediated through alterations in body composition and metabolic dysregulation [ 7 ]. Similarly, Dhair and Abed found that insufficient physical activity and extended sitting time markedly increased the risk of primary infertility among women in the Gaza Strip [ 8 ]. Furthermore, Domar et al., in a study of 12,800 IVF patients, reported that unhealthy lifestyle behaviors including prolonged sitting, smoking, and excessive alcohol consumption were highly prevalent and adversely affected fertility outcomes [ 32 ]. Such discrepancies may reflect variations in study design, participant characteristics, and the operational definitions of sedentary behavior. MR provides a valuable complement to observational analyses by using genetic variants randomly assigned at conception, thereby mitigating confounding and reverse causation in a manner analogous to randomized controlled trials [ 34 ]. In our two-sample Mendelian randomization (MR) analysis, no statistically significant causal association was observed between sedentary behavior and infertility, consistent with the observational findings. This conclusion was further supported by additional GWAS datasets. To our knowledge, this is the first study to provide an integrated evaluation of sedentary behavior and infertility using both observational and genetic epidemiological approaches. It is important to note, however, that NHANES assessed total daily sitting time via self-report, whereas MR instruments captured domain-specific sedentary traits (e.g., leisure screen time, sedentary commuting, sedentary behavior at work). The limited number of SNPs and the mismatches between phenotypes may have weakened the associations and contributed to attenuated MR signals. Future studies will require larger GWAS and harmonized sedentary phenotypes to develop stronger and more comprehensive genetic instruments, thereby improving the reliability and interpretability of MR in this field. This study has several notable strengths. First, it leveraged NHANES, a large nationally representative dataset, thereby enhancing the generalizability of our findings. Second, we employed extensive covariate adjustments and stratified subgroup analyses, which reduced potential confounding and enabled evaluation of differential effects across key populations. Third, by integrating observational and MR approaches, and validating consistency across multiple sensitivity analyses, we improved the robustness and credibility of our conclusions. Nonetheless, several limitations must be acknowledged. First, both infertility and daily sitting time in NHANES were self-reported, which may be subject to recall bias and lacks domain-specific or temporal detail. Second, although we adjusted for several established confounders, residual and unmeasured confounding may remain. For example, infertility-related medical histories such as endometriosis, polycystic ovary syndrome (PCOS) and duration of infertility were not available in NHANES but could influence the observed associations. Third, the cross-sectional design of NHANES precludes causal inference. Future prospective studies with larger cohorts, detailed reproductive histories, and objective sedentary behavior measurements are needed to validate and expand these observations.

Infertility is defined as the inability to achieve a clinical pregnancy after at least 12 months of regular, unprotected intercourse. According to recent estimates from the World Health Organization, approximately one in six individuals worldwide will experience infertility at some point in their lives, based on data spanning 1990 to 2021 [ 1 ]. In the United States, about 12.7% of women of reproductive age seek medical treatment for infertility each year [ 2 ]. Beyond being a deeply personal issue, infertility has become a major public health concern with profound implications for human development. Its rising prevalence has been linked to a range of factors, including unhealthy lifestyle behaviors, environmental exposures and psychological stress [ 3 ]. In recent years, sedentary behavior has gained increasing attention as a potential determinant of reproductive health [ 4 , 5 ]. Sedentary behavior is defined as any waking activity undertaken in a sitting, reclining, or lying posture with an energy expenditure of ≤ 1.5 metabolic equivalents (METs), and it is highly prevalent across populations [ 6 ]. Numerous studies have examined the association between sedentary behavior and female fertility, yet the findings remain inconsistent. Several multicenter case–control studies have reported that prolonged sedentary behavior is associated with higher odds of infertility, suggesting that a physically inactive lifestyle may increase the risk of primary infertility among women. For example, Foucaut et al. found an association between sedentary behavior and infertility in a clinic-based sample of men and women with idiopathic infertility, while Dhair and Abed observed similar findings in a case-control study among women in Gaza Strip [ 7 , 8 ]. In contrast, evidence from a large population-based cohort study found no significant association between sitting time and infertility [ 9 ]. Additionally, studies by Russo et al. and Esmaeilzadeh et al. failed to identify a meaningful relationship between physical activity or sedentary behavior and female fertility in their clinic-based cohorts [ 10 , 11 ]. Taken together, the existing literature offers mixed and sometimes contradictory evidence regarding the impact of sedentary behavior on infertility. Moreover, most studies to date have relied on clinical samples, with only a limited number employing large-scale, population-based cohorts, thereby restricting the generalizability of their conclusions. Mendelian randomization (MR) provides an alternative approach to strengthen causal inference. This method uses genetic variants—typically single nucleotide polymorphisms (SNPs)—as instrumental variables to evaluate causal relationships through statistical analysis. Because these variants are randomly allocated at conception, MR is inherently less susceptible to confounding and reverse causation than conventional observational designs [ 12 , 13 ]. As such, MR offers a robust and biologically grounded framework for assessing the potential causal effect of sedentary behavior on infertility. In the present study, we combined data from the nationally representative NHANES surveys (2013–2018) with two-sample MR analyses to provide a comprehensive evaluation of the association between sedentary behavior and infertility.

Supplementary Material 1. Supplementary Material 1.