Association between GLM7 glycolipid metabolism index and female infertility: A cross-sectional analysis of NHANES 2013-2020.

This study confirms that the GLM7 index has an independent, significant, and nonlinear positive correlation with infertility risk among American women, with a particularly prominent effect in the low-value segment (<6.84), and has the ability to serve as an early warning tool for reproductive metabolism. Incorporating the GLM7 index into routine prepregnancy assessments is expected to achieve “primary prevention” of metabolic-reproductive comorbidities, providing new scientific evidence and practical paths for reducing the incidence of infertility and improving population fertility.

Infertility is defined as the failure to achieve pregnancy after 12 months or more of regular, unprotected sexual intercourse in couples where neither partner has known causes of impaired fertility. [ 1 ] According to relevant studies, the global prevalence of female infertility was 110.1 million in 2021, [ 2 ] becoming a significant burden in public health. Female factors account for approximately 37% of all infertility cases, [ 3 ] with common etiologies including advanced age, ovulatory disorders, tubal lesions, and endometriosis. [ 4 ] Lifestyle and metabolic abnormalities are emerging as important new contributors to female infertility. [ 5 ] High-sugar and high-fat diets, lack of exercise, staying up late, and mental stress collectively lead to a high incidence of metabolic syndrome such as obesity, hypertension, and impaired glucose tolerance. [ 6 ] These factors increase the risk of infertility through multiple pathways, including insulin resistance (IR), chronic low-grade inflammation, oxidative stress, and mitochondrial dysfunction, which interfere with follicular development. [ 7 , 8 ] Therefore, traditional single metabolic indicators often fail to fully capture this early, multi-dimensional metabolic disorder, resulting in insufficient sensitivity to early reproductive function damage. [ 9 ] Thus, there is an urgent need to identify new biomarkers and clarify the mechanisms underlying infertility to provide a basis for early screening and personalized intervention. The glycolipid metabolism 7 factors (GLM7) index is a novel composite glycolipid index proposed in recent years. It is calculated by standardized weighted combination of 7 routine indicators: age, body mass index (BMI), fasting blood glucose, insulin, triglycerides, low-density lipoprotein cholesterol (LDL-c), and high-density lipoprotein cholesterol (HDL-c). It comprehensively reflects the overall level of energy metabolism and glycolipid homeostasis in a single dimension. Compared with previous scattered and expensive metabolic indicators, the GLM7 index integrates simplicity of calculation, low cost, and high information density. It has been shown to exhibit excellent predictive power in the early warning of chronic diseases such as metabolic syndrome, type 2 diabetes, and atherosclerosis. [ 10 ] However, this tool has not yet been applied in the field of reproductive health. Existing data indicate that the GLM7 index is highly sensitive to minor changes in obesity and IR, enabling it to issue metabolic alerts in the preclinical stage. IR is a well-recognized core mechanism leading to infertility, [ 11 ] suggesting that the GLM7 index has great potential as a new marker for fertility screening. Therefore, clarifying the association between the GLM7 index and female infertility opens up new intervention pathways for reducing the incidence of infertility from the source, with profound public health and clinical practical significance. Based on the nationally representative sample of the National Health and Nutrition Examination Survey (NHANES) database, exploring the association between the GLM7 index and female infertility is expected to provide a new perspective for the etiological research and population risk monitoring of infertility. The GLM7 index integrates multiple key metabolic factors affecting ovarian function, [ 10 ] which theoretically can serve as a potential objective tool for assessing female fertility. Conducting research on the association between the GLM7 index and infertility, this study utilizes a large, widely representative sample to clarify the association between the GLM7 index and infertility among women in various states of the United States. It not only helps to confirm the applicability of the GLM7 index as an early warning tool for fertility but also provides evidence-based support for constructing an infertility risk monitoring system based on routine physical examination indicators, which is of great public health significance for the early screening and treatment of metabolic-reproductive comorbidities.

Conceptualization: Wenya Zhang, Danyang Xia, Ailing Li, Yanyan Zhang, Yan Kang, Yonghui Jiao. Data curation: Wenya Zhang, Liying Xie. Formal analysis: Wenya Zhang, Abdureyimu Amanguli, Yanyan Zhang. Funding acquisition: Yonghui Jiao. Investigation: Danyang Xia, Ailing Li, Abdureyimu Amanguli, Xiaoling Ma. Methodology: Wenya Zhang, Danyang Xia, Abdureyimu Amanguli, Wenting Liu, Xiaoling Ma. Project administration: Ailing Li, Wenting Liu. Resources: Wenya Zhang, Liying Xie, Yan Kang, Xiaoling Ma. Software: Liying Xie, Yan Kang. Supervision: Yonghui Jiao. Validation: Wenya Zhang, Ailing Li, Wenting Liu, Yonghui Jiao. Visualization: Ailing Li. Writing – original draft: Wenya Zhang, Liying Xie, Danyang Xia. Writing – review & editing: Yonghui Jiao.

This study extracted data from the 2013–2020 NHANES. The database collects residents’ health information through questionnaires, physical examinations, and laboratory tests, and all raw data and codes are publicly available on the official website of the U.S. Centers for Disease Control and Prevention ( www.cdc.gov/nchs/nhanes ). This secondary data analysis study did not require additional independent ethical approval. A total of 22,787 female participants aged 18 years and above were initially included. After excluding 14,523 participants with missing questionnaires or non-definitive responses, 8264 participants were initially included. Further exclusion of 4472 participants lacking key laboratory indicators required for calculating the GLM7 index (such as BMI, fasting blood glucose, insulin, triglycerides, LDL-c, and HDL-c) resulted in a final sample of 3792 women for statistical analysis (Fig. 1 ). A flow diagram of eligible participant selection in NHANES (2013–2020). BMI = body mass index, GLM7 = glycolipid metabolism 7 factors, HDL-c = high-density lipoprotein cholesterol, LDL-c = low-density lipoprotein cholesterol, NHANES = National Health and Nutrition Examination Survey. Referring to relevant studies, infertility was defined as self-reported failure to conceive after attempting for more than 1 year. [ 1 ] Based on the response to the question “Have you been unable to conceive after trying for 12 months?” in the NHANES reproductive health questionnaire, participants who answered “Yes” were classified into the infertile group, and those who answered “No” were classified into the non-infertile group. The GLM7 index is a novel glycolipid metabolism index proposed in recent years for the early diagnosis of metabolic diseases. [ 10 ] According to relevant studies, the GLM7 index is calculated as follows: GLM7 = log 10 (Age [years] × BMI [kg/m 2 ] × Fasting Blood Glucose [mg/dL] × Insulin [pmol/L] × Triglycerides [mmol/L] × LDL-c [mmol/L]/HDL-c [mmol/L]). Considering traditional lipid metabolism detection indicators, 7 routine detection indicators (age, triglycerides, FBG, BMI, HDL-c, LDL-c, and insulin) were finally selected. To more clearly demonstrate the impact of these 7 factors on diseases, the GLM7 index was introduced. It has been initially confirmed to have good predictive value for chronic diseases such as metabolic syndrome, type 2 diabetes, and atherosclerosis. The selection of covariates in this study was based on professional judgment and reference to previous studies. [ 12 , 13 ] Included covariates included sociodemographic factors: age, race/ethnicity, education level, marital status, household income, etc; lifestyle variables: smoking status, drinking status; health-related factors: history of hypertension, history of diabetes, BMI, height, weight, fasting blood glucose level, insulin level, triglycerides, HDL-c, etc. These covariates were derived from the demographic, examination, reproductive health questionnaire, and smoking questionnaire sections of the NHANES database. In this study, participants were divided into the “non-infertile group” and the “infertile group.” Continuous variables in the comparative analysis were expressed as mean ± standard deviation, and categorical variables were expressed as number (n) and percentage (%). Differences between the 2 groups were analyzed by t -test. Multiple imputation was used for data with missing values <10% for variables not involved in the calculation of the GLM7 index. Three logistic regression models were used to assess the correlation between the GLM7 index (as a continuous variable and stratified by tertiles) and infertility: Model 1 was an unadjusted model; Model 2 adjusted for age, marital status, race, household income, and education level; and Model 3 further adjusted for smoking, drinking, and history of hypertension and diabetes on the basis of Model 2. Based on Model 3, restricted cubic spline (RCS) curves were further used to explore the nonlinear relationship between the GLM7 index and infertility. Meanwhile, the receiver operating characteristic curve and its area under the curve (AUC) were used to compare the predictive ability of the GLM7 index for infertility between Model 1 and Model 3. In addition, threshold effect analysis of the GLM7 index and infertility was performed to verify the existence of a positive correlation. Finally, interaction and stratified analyses were conducted according to age, race, household income, marital status, education level, smoking status, drinking status, and chronic disease history. All data in this study were analyzed using R statistical software (version 4.3.1), and a two-tailed P  < .05 was considered statistically significant.

A total of 3792 female participants were included in this study, among whom 413 were in the infertile group. Baseline characteristics are shown in Table 1 . Compared with the non-infertile group, women in the infertile group were significantly older, with higher BMI and weight (all P  < .05). In terms of demographic characteristics, the infertile group had a higher proportion of non-Hispanic Whites, married individuals, and those with higher education, while a lower proportion of Mexican Americans and individuals with low income ( P  < .05). At the same time, the prevalence of hypertension in the infertile group was significantly higher ( P  < .001), while their HDL-c level was significantly lower. Most importantly, the core predictive variable of this study – the GLM7 index – was significantly higher in the infertile group than in the non-infertile group (7.04 ± 0.76, P  < .001). Baseline characteristics of all participants stratified by infertility status. Values are presented as mean ± standard deviation or n (%). BMI = body mass index, GLM7 = glycolipid metabolism 7 factors, HDL-c = high-density lipoprotein cholesterol, LDL-c = low-density lipoprotein cholesterol, RIP = ratio of family income to poverty. When the GLM7 index was stratified by tertiles, a significant risk gradient was observed (Table 2 ). In the fully adjusted model (Model 3), compared with the lowest tertile (T1) group, women in the highest tertile (T3) group had a significantly 49% higher risk of infertility (odds ratio [OR] = 1.49, 95% confidence interval [CI]: 1.10–2.02), with a P -trend of .011. To further explore the exact association pattern between the GLM7 index (as a continuous variable) and infertility risk, RCS analysis was performed. The analysis confirmed a significant nonlinear relationship between the 2 (overall P  = .001, nonlinear P  = .009). As shown in Figure 2 , the RCS curve showed a significant nonlinear relationship between the GLM7 index and infertility risk (overall association P  = .001; nonlinear test P  = .009). As the GLM7 value increased from 5 to 10, the risk of infertility showed a nonlinear upward trend. When the GLM7 index increased from 5 to 10, the risk increased sharply in the interval  8.5 (OR = 0.96; 95% CI 0.88–1.05). The lowest risk point was at GLM7 ≈ 7, which can be used as an inflection point. Multivariate logistic regression analysis of GLM7 index in relation to infertility. The P value of statistical significance was bolded. Model 1: No covariates were adjusted. Model 2: Covariates including age, race, marital status, household income, and education level were adjusted. Model 3: Covariates such as age, race, marital status, household income, education level, smoking, drinking, diabetes, and hypertension were adjusted. CI = confidence interval, OR = odds ratio. Nonlinear association between the GLM7 index and the risk of female infertility. CI = confidence interval, GLM7 = glycolipid metabolism 7 factors. As shown in Figure 3 , threshold effect analysis indicated a significant inflection point between the GLM7 index and infertility ( P  = .006). Overall analysis showed a positive association between the 2 (OR = 1.26, 95% CI 1.07–1.48). Segmented regression further revealed that when GLM7 ≤ 6.84, the risk increased by approximately 1 time (OR = 2.12, 95% CI 1.15–3.89); when GLM7 > 6.84, the association disappeared (OR = 1.05, 95% CI 0.89–1.23), suggesting that 6.84 is a potential effect threshold. Threshold effect analysis of the GLM7 index on infertility risk. GLM7 = glycolipid metabolism 7 factors. As shown in Figure 4 , the AUC of the unadjusted model (Model A) was 0.57 (95% CI: 0.54–0.59), indicating that the discriminative accuracy of the GLM7 index for female infertility was low when no confounding factors were adjusted. After including covariates such as age, race, marital status, education level, smoking, drinking, diabetes, and hypertension (Model B), the AUC increased to 0.68 (95% CI: 0.66–0.71), indicating that the predictive ability of the GLM7 index was significantly improved. Receiver operating characteristic (ROC) curves of the GLM7 index for predicting female infertility. (A) Unadjusted model (Model 1). (B) Fully adjusted model (Model 3). AUC = area under the curve, CI = confidence interval, GLM7 = glycolipid metabolism 7 factors. Figure 5 shows the results of the subgroup analysis. In the total population, the GLM7 index was significantly positively correlated with female infertility (OR = 1.35, 95% CI: 1.16–1.58, P  < .001). Heterogeneity test showed that only age stratification had statistical significance ( P  < .001): the OR values of the < 26 years and 26 to 36 years subgroups were 1.62 (95% CI: 1.27–2.07) and 1.41 (95% CI: 1.18–1.68), respectively, both P  < .001; no significant association was found in the ≥ 37 years subgroup ( P  = .12). In other stratifications, Mexican Americans, non-Hispanic Whites, married and other marital statuses, low income (PIR ≤ 1.3) and middle income (1.3 < PIR ≤ 3.5), education below high school and above high school, moderate drinking, missing smoking/drinking data, and populations without diabetes and hypertension all consistently showed a significant positive correlation between the GLM7 index and infertility risk (all P  < .05). Overall, the direction of the association was stable across all subgroups, and the effect size was larger in younger, socioeconomically disadvantaged, and populations with normal glucose metabolism but high GLM7 index. Associations between GLM7 and female infertility in different subgroups according to baseline characteristics. CI = confidence interval, GLM7 = glycolipid metabolism 7 factors, OR = odds ratio. Based on the nationally representative sample database of NHANES 2013–2020, this study is the first to find an independent positive correlation between the GLM7 index and infertility risk among American women. Each 1-unit increase in the GLM7 index was associated with a 29% higher risk of infertility; the highest tertile (T3) had a 49% higher risk than the lowest tertile (T1), with a significant trend P -value, providing strong evidence for a dose–response relationship between the 2. Threshold effect analysis suggested an inflection point at 6.84, with an OR as high as 2.12 in the low-value segment (<6.84), indicating that the GLM7 index has high sensitivity to early reproductive system damage, highlighting its unique value for early screening. After adjusting for multiple variables such as sociodemographic factors, lifestyle factors, and metabolic diseases, the predictive ability of the GLM7 index improved from AUC = 0.57 to 0.68, reaching a moderate discriminative level. Subgroup analysis showed that the association was more prominent in women < 36 years old, low-income groups, and populations with abnormal glucose metabolism. Therefore, the results of our study suggest that young women, socioeconomically vulnerable women, and women in the metabolic critical state are the primary beneficiaries of GLM7 index screening.

This observational analysis was conducted based on nationally representative multi-ethnic population data from the NHANES spanning 2013 to 2020, and the findings preliminarily indicated an independent positive correlation between the glycolipid metabolic composite index GLM7 and the risk of female infertility. After adequate adjustment for multiple potential confounding variables, each 1-unit increment in GLM7 corresponded to a 29% elevated probability of female infertility. When participants were grouped by GLM7 tertiles, women in the highest tertile exhibited a 49% higher risk of infertility relative to those in the lowest tertile, with a statistically significant trend test, demonstrating an overall dose–response pattern between the 2 variables. Threshold effect analysis identified a critical inflection point of 6.84 for GLM7; the OR for infertility was 2.12 among individuals with GLM7 values below this threshold. This finding suggested that GLM7 may sensitively characterize early subtle abnormalities in the reproductive system, conferring potential value as a preliminary screening biomarker for fertility-related metabolic risks. Following confounder adjustment, the area under the receiver operating characteristic curve (AUC) of GLM7 for predicting female infertility increased from 0.57 to 0.68, indicating a moderate discriminatory capacity of this indicator. Stratified subgroup analyses revealed that the aforementioned statistical correlation was more pronounced among women of reproductive age younger than 36 years, individuals with low socioeconomic status, and populations with impaired glucose homeostasis, implying that these subgroups may benefit from metabolic risk stratification screening guided by GLM7 measurements. Restricted by the cross-sectional study design, the present investigation could only describe the statistical association between the 2 variables, and the underlying biological mechanisms remain to be further elucidated through subsequent basic and clinical cohort studies. IR is widely recognized as a critical intermediate pathway linking glycolipid metabolic disorders to dysregulated reproductive endocrine function in females. [ 14 ] Integrating BMI, fasting blood glucose and serum insulin, GLM7 enables comprehensive quantification of the overall severity of systemic IR. [ 15 ] Accumulated evidence has illustrated that IR often coincides with hyperandrogenemia, which may disrupt the regulatory homeostasis of the hypothalamic–pituitary–ovarian axis, while triggering chronic low-grade inflammation and oxidative stress. These metabolic perturbations frequently coexist with multiple reproductive abnormalities, including diminished oocyte quality, ovarian granulosa cell apoptosis, follicular atresia and impaired endometrial receptivity. [ 16 , 17 ] Reduced peripheral insulin sensitivity triggers compensatory hyperinsulinemia. Ovarian tissues abundantly express insulin receptors, and excessive insulin, in synergy with luteinizing hormone, elevates the circulating concentration of biologically active androgens. Persistent hyperandrogenic status tends to disturb the secretory rhythm of hormones governing follicular development. Additionally, insulin may suppress granulosa cell maturation and induce preantral follicular atresia via modulating the MAPK/ERK signaling pathway, thereby forming a cyclic state complicated by both metabolic derangements and reproductive dysfunction. [ 18 , 19 ] Under insulin-resistant conditions, structural and functional disruption of the intestinal barrier initiates systemic inflammatory responses, which may interfere with follicular maturation and embryonic endometrial implantation through the gut–liver–ovarian axis. [ 20 , 21 ] Furthermore, within the IR microenvironment, free fatty acids and advanced glycation end products facilitate the activation of the TLR4/IκKβ/NF-κB signaling cascade, stimulating the release of abundant pro-inflammatory cytokines and reactive oxygen species (ROS) to elicit local ovarian inflammation, which may further exacerbate systemic IR. [ 22 , 23 ] Accumulated ROS tends to damage the genetic material and cellular structure of oocytes and accelerates the decline of ovarian reserve. Chronic inflammatory microenvironments are also correlated with reduced endometrial receptivity, contributing to adverse reproductive outcomes such as implantation failure and early pregnancy loss. [ 24 , 25 ] In light of existing pathophysiological evidence, it is speculated that the glucose- and insulin-related components incorporated into GLM7 may correlate with female reproductive dysfunction primarily through intermediate processes including IR-mediated hyperandrogenemia and localized ovarian chronic inflammation. GLM7 incorporates 3 lipid-related biomarkers. Previous epidemiological evidence has documented that dyslipidemia is frequently accompanied by aggravated ovarian oxidative stress and exacerbated local inflammatory injury. Decreased HDL-c impairs endogenous antioxidant capacity; elevated LDL-c may disrupt ovarian microcirculatory homeostasis and steroidogenesis; excessive triglyceride accumulation in ovarian tissues activates inflammasomes and induces massive ROS production. Co-occurring multiple lipid disturbances are correlated with compromised developmental potential of oocytes. [ 26 , 27 ] Stratified lipid-lowering interventions formulated on the basis of GLM7 results may provide evidence-based references for optimizing ovarian metabolic status and improving ovulatory function among infertile females with glycolipid metabolic abnormalities. RCS regression analyses demonstrated that the risk of female infertility rose rapidly when GLM7 values were below 6.84, whereas the upward trend of infertility risk gradually plateaued beyond this critical threshold. This inflection point may serve as a reference cutoff for evaluating the body’s compensatory capacity against glycolipid metabolic perturbations. Individuals with GLM7 ≤ 6.84 tend to exhibit attenuated compensatory regulation against mild glycolipid disorders, rendering the reproductive endocrine system more vulnerable to metabolic disturbances. Slight metabolic imbalances may initiate an ovarian inflammatory cascade and consequently increase the likelihood of infertility. [ 28 , 29 ] Notably, this threshold was derived exclusively from statistical analyses of observational population data in the current study and cannot be directly adopted as a diagnostic criterion in clinical practice. Further external validation across multicenter cohorts from diverse populations is required to verify the stability of this threshold, facilitating the refinement of early screening and prevention strategies for metabolism-associated female infertility via GLM7-based stratified management. Compared with traditional composite metabolic indices including the ZJU index, TyG index and METS-IR, the GLM7 model incorporating 7 routine physical and biochemical measurements yielded an AUC of 0.68 for the prediction of female infertility in the present study, suggesting its certain reference value in comprehensive fertility risk assessment. [ 13 , 30 , 31 ] The present study preliminarily identified a threshold association between GLM7 and infertility risk among women without polycystic ovary syndrome, providing epidemiological evidence for the development of metabolic stratified intervention strategies, though this indicator cannot be independently applied for clinical diagnosis of female infertility at present. Prior studies have validated the utility of GLM7 in risk prediction for multiple metabolism-related chronic diseases. This research extends the application scope of this composite biomarker to reproductive medicine, offering a novel evaluative tool for preliminary screening among patients with metabolism-associated infertility. Nevertheless, several limitations should be acknowledged. First, the cross-sectional design only allows the detection of static correlations between variables and cannot establish the temporal sequence between exposure and outcome events. Second, information on infertility was collected via self-reported retrospective questionnaires, which inevitably introduces recall bias and potential outcome misclassification. Restricted by variables available in the original database, ovarian reserve markers, specialized reproductive endocrine parameters, environmental exposures and psychological confounders were not incorporated into adjustment models, and residual confounding could not be completely eliminated. The exploratory threshold of 6.84 requires further validation in external population cohorts. With an AUC of 0.68, GLM7 only confers moderate discriminatory performance and is not suitable for standalone clinical infertility screening. Multiple imputation was adopted to handle missing data to mitigate selection bias, yet this statistical method may introduce inherent analytical uncertainty. Future prospective cohort studies are warranted to verify the correlation patterns between GLM7 and female infertility, optimize relevant predictive models, and accumulate additional evidence supporting the clinical application of this metabolic composite index in fertility risk screening among reproductive-aged women.

We appreciate the National Center for Health Statistics for providing the NHANES database and the Natural Science Foundation of Xinjiang Uygur Autonomous Region for the financial support. All authors have reviewed the final manuscript and agreed to the authorship list.