{"paper_id":"02ddaf49-681a-463b-a32e-77cfaef8ab74","body_text":"RESEARCH Open Access\n© The Author(s) 2025. Open Access  This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 \nInternational License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you \ngive appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the \nlicensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or \nother third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the \nmaterial. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or \nexceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit  h t t p  : / /  c r e a  t i  \nv e c  o m m  o n s .  o r  g / l  i c e  n s e s  / b  y - n c - n d / 4 . 0 /.\nHu et al. BMC Public Health         (2025) 25:1811 \nhttps://doi.org/10.1186/s12889-025-22966-z\nBMC Public Health\n†Anquan Hu and Lu Xiong contributed equally to this work.\n*Correspondence:\nFeng Chen\nfenger0802@163.com\nTao Liu\nltao829@163.com\nFull list of author information is available at the end of the article\nAbstract\nBackground With the increasing global aging population, cognitive impairment, particularly Alzheimer’s disease \n(AD), has become an escalating public health and economic concern. Recent research has increasingly focused on the \nrelationship between female reproductive factors and cognitive health. This study explores the association between \nreproductive history factors and cognitive performance in women aged 60 and older in the US, providing insights for \nthe prevention and management of cognitive impairment.\nMethods We analyzed participants in the National Health and Nutrition Examination Survey (NHANES) between 2011 \nand 2014. The cognitive performance was assessed by the Consortium to Establish a Registry for Alzheimer’s Disease \n(CERAD) Word Learning sub-test, Animal Fluency test (AFT), and Digit Symbol Substitution Test (DSST), in relation to \nreproductive history variables like age of menarche, menopause, reproductive span, number of pregnancies, and \nparity. Statistical analyses included weighted linear regression for continuous variables and weighted chi-square \ntests for categorical variables, with adjustments for age, BMI, alcohol intake, smoking, PIR, education, race/ethnicity, \nhypertension, and diabetes.\nResults A total of 698 (weighted sample was 25,558,437) women aged 60 years or older were included in the study. \nParity negatively impacted cognitive performance, women with ≥ 5 parity showing reductions in AFT (β = -2.1, \np = 0.032), DSST (β = -14, p < 0.001), CERAD trial 1 (β = -0.41, p = 0.031), and CERAD Total scores (β = -1.3, p = 0.033) all \nin model 2. Delayed menopause was positively associated with cognitive function, showing improvements in CERAD \ntrial 1 (β = 1.2, p = 0.002) and total recall (β = 2.1, p = 0.031) both in model 3. Longer reproductive span was linked to \nbetter cognitive function, particularly in immediate recall and processing speed (β = 0.12, p < 0.001 for DSST) in model \n3.\nAssociation of menarche, menopause, \nand reproductive history with cognitive \nperformance in older US women: a cross-\nsectional study from NHANES 2011–2014\nAnquan Hu1†, Lu Xiong1†, Huijun Wei1, Liangyan Yuan1, Yumeng Li2, Heyan Qin2, Feng Chen3* and Tao Liu2*\n\nPage 2 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nIntroduction\nThe global population is experiencing an increase in age, \nleading to a significant rise in cognitive impairment, \nwhich has emerged as a critical issue for public health. \nAging is a complex biological process with profound \nimplications for health and society [ 1]. The aging woman \npopulation is increasing, with women generally living \nlonger than men and comprising the majority of older \npersons, especially at the most advanced ages. In the US, \nAD affects approximately 6.7 million adults aged 65 and \nolder, a number projected to nearly double by 2060 with -\nout medical breakthroughs to mitigate its progression \n[2]. Women, in particular, experience a higher prevalence \nof cognitive decline and AD [ 3]. In the past two decades, \nthe greatest increase in female deaths has been from AD \nand other dementias, with deaths nearly tripling between \n2000 and 2021 [ 4]. Factors such as reproductive history \nand hormonal changes are likely contributing to this \ndisparity in cognitive health. The complex relationship \nbetween female reproductive factors cognitive function \nhas received considerable attention in medical research. \nReproductive events are not only biological milestones \nbut also important determinants of women’s long-term \nhealth, particularly cognitive function later in life [5].\nSome findings illustrate the intricate relationship \nbetween female reproductive health and cognitive devel -\nopment. Song et al. found that older age at menarche and \nmenopause and longer reproductive cycles are associ -\nated with lower risks of mild cognitive impairment and \nAD [6]. Jett et al. highlight the link between female repro-\nductive factors and AD risk, focusing on endogenous \nand exogenous estrogen exposure. Key factors include \nreproductive lifespan, menopause status (spontaneous vs. \ninduced), parity, and hormonal therapies such as contra -\nceptives, menopause hormone therapy, and anti-estrogen \ntreatments. Understanding how these factors influence \nbrain aging through sex-specific pathways is essential for \ndeveloping AD prevention and treatment strategies [ 7]. \nDespite these advances, significant research gaps remain. \nExisting studies have primarily focused on isolated repro-\nductive factors or specific cognitive outcomes, leaving a \nlack of systematic investigation into how different repro -\nductive stages collectively influence multiple cognitive \ndomains, such as language memory, executive function, \nand processing speed, in older women. To address these \ngaps, this study leverages the NHANES 2011–2014 data -\nbase, known for its broad representativeness and sys -\ntematic data collection, providing a unique perspective \nand detailed data support that enhances the innovation \nand generalizability of the research. The novelty of this \nstudy lies in its investigation of the relationship between \nreproductive histories and cognitive function in women \naged 60 and the elder. In particular, multiple cognitive \ntests such as CERAD, AFT, and DSST are used to com -\nprehensively assess various aspects of cognitive function, \nincluding language memory, fluency, executive function, \nprocessing speed, and working memory [ 8]. By further \ninvestigating the potential impact of various reproductive \nhistories on cognitive function, this study aims to provide \nnew evidence to support future intervention strategies \nfor preventing dementia in women.\nMethods\nStudy design and population\nThis research is a cross-sectional analysis utilizing the \nNHANES database. The NHANES is an extensive and \nintricate multistage survey that targets the noninstitu -\ntionalized population of the United States. This survey \nis administered by both the Centers for Disease Con -\ntrol and Prevention (CDC) and the National Center for \nHealth Statistics, with the aim of delivering representa -\ntive national estimates regarding health and nutritional \nconditions [ 9]. In each survey cycle, individuals from \ndiverse geographic regions across the United States are \nselected through a stratified sampling process involv -\ning various census blocks or segments of census block \ngroups, thereby ensuring that NHANES constitutes a \nnationally representative sample [ 10]. The detailed sam -\npling method adopted has been published elsewhere [11]. \nThe NHANES 2011–2014 data have been widely used in \nrecent research to explore various aspects of cognitive \nfunction [12, 13]. From the NHANES 2011–2014 public \nrelease dataset (n = 19,931), we excluded individuals aged \nbelow 60 years (n = 16,299), resulting in 3,632 participants \naged 60 and older. We further excluded males (n = 1,760), \nleaving 1,872 female participants. After excluding popu -\nlations with missing data on key variables and cognitive \ntest scores (n = 1,038), 834 participants remained. Finally, \nwe excluded individuals with missing data on reproduc -\ntive history variables, including age at menarche, age at \nmenopause, number of pregnancies, reproductive span, \nand parity ( n = 136), resulting in 698 participants eligi -\nble for analysis. The detailed selection process for study \ninclusion is illustrated in Fig. 1.\nConclusion Higher parity was negatively correlated with processing speed and memory. In contrast, delayed \nmenopause and a longer reproductive span were positively correlated with global cognition and processing speed. \nThese findings suggest that reproductive factors play a potential role in cognitive aging among older women.\nKeywords Cognitive function, Menarche, Menopause, Reproductive history, NHANES\n\nPage 3 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nMeasurement of cognitive function\nCognitive assessments were conducted by highly trained \ninterviewers during initial private interviews. NHANES \nuses the Consortium to Establish a Registry for Alzheim -\ner’s Disease (CERAD) Word Learning test, Animal Flu -\nency test (AFT), and Digit Symbol Substitution test \n(DSST) to assess different cognitive functions [ 14]. Voice \nrecordings were essential for both the CERAD Word \nLearning (WL) and the Auditory Functional Test (AFT). \nIf permission for the voice recording was not provided, \nonly the Digit Symbol Substitution Test (DSST) was \nadministered. The CERAD-WL evaluates memory, focus-\ning on verbal memory and comprises four trials [ 15]. In \nthe first three trials, participants read aloud ten words \nand then immediately recalled them. The total number \nof correctly recalled words from these trials formed the \nimmediate recall score (range: 0 to 30). After complet -\ning the AFT and DSST (approximately 8–10  min after \nthe CERAD-WL), participants were asked to recall as \nmany of the ten words as possible. The total number of \ncorrect recalls was used to calculate their delayed recall \nscore [16]. The AFT assesses semantic fluency, in which \nFig. 1 Flow chart of participants selected\n \n\nPage 4 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nparticipants are challenged to name as many animals as \nthey can within a 60-second timeframe, with each animal \nnamed contributing a point to their score, is used to eval-\nuate executive function [ 17]. The DSST is a test of sus -\ntained attention, working memory and processing speed. \nParticipants were provided with a test sheet featuring 9 \nnumbers at the top, each paired with a corresponding \nsymbol key. Below the key, there was a sequence of 133 \nnumbers. They were asked to match each number with \nthe correct symbol within 2 min. One point was awarded \nfor each accurate match, with a maximum possible score \nof 133 points [18]. Improved scores on these assessments \nreflect superior cognitive functioning, a conclusion \nbacked by considerable research [19].\nCovariates\nCovariates consisted of various demographic character -\nistics, including sex, age in years, age group (60–69 years, \n70–79 years, ≥ 80 years), race/ethnicity, family income ( \npoverty income ratio, PIR), educational attainment (less \nthan a high school education, some high school, high \nschool graduate, some college or associate’s degree, col -\nlege graduate or more), alcohol intake (non-drinker, 1 to \n< 5 drinks/month, 5 to < 10 drinks/month, or ≥ 10 drinks/\nmonth), and smoking status (current, former, or never \nsmoker), body mass index (BMI), hypertension, and dia -\nbetes [20]. Race/ethnicity in NHANES are self-identified, \nas previously reported [ 21, 22]. The categories included: \nNon-Hispanic White, Non-Hispanic Black, Mexican \nAmerican, Other Hispanic, and Other/multiracial [ 23]. \nThe age of menarche, age of menopause, pregnancy, and \nparity were obtained from the NHANES database using \nthe reproductive health questionnaire [ 24, 25]. Specifi -\ncally, the age of menarche was obtained from RHQ010 \n(Age when first menstrual period occurred), and the age \nof menopause was obtained from RHQ060 (Age at last \nmenstrual period). Women experiencing menopause \nwere classified further according to the cause: natural \nmenopause or surgical menopause. Natural menopause \nrefers to the ending of menstrual periods for over 12 \nconsecutive months resulting from physiological fac -\ntors, while excluding influences like surgical procedures \nor medical interventions [ 26]. Surgical menopause, on \nthe other hand, was attributed to the surgical removal \nof both ovaries before the typical age of natural meno -\npause [ 27]. The number of pregnancies was obtained \nfrom RHQ160 (How many times have been pregnant? \n), and RHQ171 (How many deliveries resulted in a live \nbirth? ) to determine parity. In the research analysis, age \nat menarche was categorized into three groups: age at \nmenarche < 12, 12 ≤ age at menarche ≤ 13, and age at men-\narche > 13. The number of pregnancies was categorized \ninto three groups: number of pregnancies ≤ 2, 3 ≤ num-\nber of pregnancies ≤ 4, and number of pregnancies ≥ 5. \nThe parity was categorized into three groups: parity ≤ 2, \n3 ≤ parity ≤ 4, and parity ≥ 5. Reproductive span was \ndefined as the difference between the age at menopause \nand the age at menarche (age menopause– age men -\narche). Age at menopause was categorized into three \ngroups: age at menopause < 45 defined early menopause, \n45 ≤ age at menopause ≤ 55 defined normal menopause, \nand age at menopause > 55 defined delayed menopause \n[28, 29].\nStatistical analysis\nSince NHANES uses a probability sampling strategy, \nwe applied the 2-year weight (wtsa2yr) for NHANES \n2011–2012 and 2013–2014. When combining the two \ncycles, the final weight (1/2 wtsa2yr) was adjusted [ 30]. \nWe divided WTDRD1 by 2 as the new sample weight. All \nstatistical analyses utilized NHANES sample weights to \nmaintain representativeness. Continuous variables were \nrepresented as medians (Q1, Q3), whereas categorical \nvariables were shown as both unweighted counts and \nweighted percentages, the latter indicating the subject \ndistribution following the application of sample weights. \nP values for continuous and categorical variables were \ncalculated using weighted linear regression and chi-\nsquare tests, respectively. Linear regression was con -\nducted on menarche, menopause age, pregnancy, parity, \nreproductive span, and cognitive function. Multivariable \nmodels were adjusted as follows: model 1:no adjusted; \nmodel 2: adjusted for age; model 3: were adjusted for \nage, BMI, alcohol intake, smoking, PIR, education, race/\nethnicity, hypertension, and diabetes [ 31]. The reference \ngroups for the models were defined as follows: age of \nmenarche < 12 years for menarche age, early menopause \nfor menopause status, ≤ 2 pregnancies for the number of \npregnancies, and parity ≤ 2 for parity. Variables with a lin-\near regression result p < 0.05 were subjected to forest plot \nvisual analysis. All statistical tests were performed using \nR (version 4.2.2. https://www.r-project.org/).\nResults\nBaseline characteristics of the participants\nA total of 698 participants were enrolled in this study \nfrom two cycles of the NHANES database. Based on the \nNHANES sampling design, the weighted data analysis \nprovided a population estimate of 25,558,437, represent -\ning US women aged 60 and older. Data included cogni -\ntive function, reproductive history (including menarche, \npregnancy, parity, menopause, and reproductive span), \nas well as other relevant covariates such as sex, age, race/\nethnicity, smoking status, education attainment, alcohol \nintake, BMI, diabetes status, and PIR.\n\nPage 5 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nMenarche and cognitive function\nOur analysis did not find any significant statistical asso -\nciation between the age at menarche and cognitive func -\ntion in later life. Median scores for the CERAD test, AFT, \nand DSST did not show significant differences among age \nat menarche < 12, 12 ≤ age at menarche ≤ 13, and age at \nmenarche > 13 groups (p > 0.05) (Table 1). Linear regres -\nsion analysis: These results suggest that age at menarche \nTable 1 Characteristics and cognitive function of study participants grouped by age at menarche\nCharacteristic Age of menarche < 12, \nN = 138 (20.40%)1\n12 ≤ Age of menarche ≤ 13, \nN = 366 (54.21%)1\nAge of menarche > 13, \nN = 194 (25.39%)1\np-\nVal-\nue2\nAge group > 0.9\n 60–69 years 73 (50.75%) 175 (47.69%) 101 (49.41%)\n 70–79 year 35 (26.70%) 114 (29.55%) 51 (28.68%)\n ≥ 80 years 30 (22.55%) 77 (22.76%) 42 (21.91%)\nRace/Ethnicity 0.024\n Non-Hispanic White 75 (80.38%) 199 (79.09%) 81 (71.57%)\n Non-Hispanic Black 28 (9.19%) 88 (10.45%) 36 (9.26%)\n Mexican American 14 (3.66%) 31 (3.67%) 25 (6.18%)\n Other Hispanic 17 (4.75%) 27 (2.70%) 19 (4.13%)\n Other/multiracial 4 (2.01%) 21 (4.09%) 33 (8.86%)\nEducation attainment 0.3\n Less Than 9th Grade 9 (4.76%) 30 (5.20%) 28 (7.24%)\n 9-11th Grade 18 (9.90%) 66 (13.77%) 29 (13.70%)\n High School Graduate/GED 45 (35.88%) 80 (21.37%) 53 (29.72%)\n Some College or AA degree 43 (28.52%) 111 (34.01%) 56 (30.29%)\n College Graduate or above 23 (20.95%) 79 (25.64%) 28 (19.04%)\nAlcohol intake 0.3\n 1–5 drinks/month 56 (43.41%) 167 (48.52%) 71 (41.90%)\n 5–10 drinks/month 5 (3.97%) 8 (2.04%) 5 (3.31%)\n ≥ 10 drinks/month 9 (11.53%) 45 (15.72%) 14 (9.02%)\n Non-drinker 68 (41.10%) 146 (33.72%) 104 (45.77%)\nSmoke group 0.8\n Current smoker 16 (12.40%) 44 (12.16%) 22 (9.97%)\n Former smoker 45 (38.42%) 119 (35.19%) 54 (32.81%)\n Never smoker 77 (49.18%) 203 (52.64%) 118 (57.22%)\nBMI group 0.057\n Underweight (< 18.5) 0 (0.00%) 7 (1.33%) 4 (0.91%)\n Normal (18.5 to < 25) 19 (17.39%) 82 (26.90%) 59 (25.81%)\n Overweight (25 to < 30) 39 (28.13%) 115 (33.74%) 55 (35.85%)\n Obese (30 or greater) 80 (54.48%) 162 (38.02%) 76 (37.43%)\nHypertension 109 (74.53%) 283 (74.14%) 141 (67.87%) 0.4\nDiabetes 65 (42.46%) 145 (34.58%) 81 (35.94%) 0.4\nPIR 0.15\n ≤ 1.3 48 (23.30%) 113 (20.78%) 72 (23.32%)\n 1.3 < to ≤ 3.5 54 (41.95%) 158 (41.55%) 85 (53.39%)\n > 3.5 36 (34.74%) 95 (37.67%) 37 (23.29%)\nCERAD1 5.00 (4.00, 6.00) 5.00 (4.00, 6.00) 5.00 (4.00, 6.00) 0.2\nCERAD2 8.00 (6.00, 9.00) 8.00 (6.00, 9.00) 7.00 (7.00, 9.00) 0.7\nCERAD3 9.00 (7.00, 9.00) 8.00 (7.00, 10.00) 8.00 (7.00, 9.00) 0.5\nCERAD total 21.0 (18.0, 24.0) 21.0 (18.0, 24.0) 21.0 (17.0, 24.0) 0.5\nCERAD delay recall 7.00 (5.00, 9.00) 7.00 (5.00, 8.00) 7.00 (5.00, 8.00) 0.15\nAFT 18.0 (14.0, 20.0) 17.0 (14.0, 20.0) 17.0 (14.0, 20.0) 0.8\nDSST 52 (42, 62) 53 (40, 64) 50 (42, 64) > 0.9\n1n (weighted %); Median (IQR)\n2chi-squared test with Rao & Scott’s second-order correction; Wilcoxon rank-sum test for complex survey samples\nAbbreviation: CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; AFT, Animal Fluency test; DSST, Digit Symbol Substitution Test; BMI, body mass \nindex; PIR, poverty income ratio; GED, General educational development; AA, Associate of Arts\n\nPage 6 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \ndid not have a significant impact on cognitive function, \nas assessed by these tests (Supplementary table S1, Sup-\nplementary table S2, Supplementary table S3).\nPregnancy and cognitive function\nSignificant demographic differences were found across \npregnancy groups, particularly in age ( p < 0.001), race/\nethnicity ( p < 0.001), and education ( p = 0.013). DSST \nscores also showed a significant difference ( p < 0.001), \nwith median scores of 54 (IQR 45–67) for ≤ 2 pregnan -\ncies, 51 (IQR 41–64) for 3–4 pregnancies, and 49 (IQR \n33–61) for ≥ 5 pregnancies, suggesting a potential decline \nin cognitive function with more pregnancies. (Table  2). \nLinear regression analysis showed that the associa -\ntion between pregnancies and cognitive function varied \nacross models and domains. For CERAD trial 1, ≥ 5 preg-\nnancies were positively associated with cognitive perfor -\nmance in model 3 ( p = 0.040). In CERAD delayed recall, \n3–4 pregnancies were negatively associated in model 1 \n(p = 0.018). For DSST scores, ≥ 5 pregnancies were nega -\ntively associated in models 1 ( p < 0.001) and 2 ( p < 0.001). \nThese associations were not consistent across all models \nor domains. (Supplementary table S1, Supplementary \ntable S2, Supplementary table S3).\nParity and cognitive function\nOur study found significant differences in cognitive \nfunction related to parity. The demographic analysis \nrevealed disparities in age group ( p < 0.001), race/ethnic-\nity ( p < 0.001), and educational levels ( p < 0.001) across \nparity groups. Cognitive function assessments demon -\nstrated statistically significant differences in CERAD \ntrial 1, CERAD trial 2, CERAD trial 3, CERAD trial total, \nCERAD delay recall, AFT, and DSST (Table  3). Linear \nregression analysis showed that women with ≥ 5 par -\nity had a significant negative impact on AFT ( p = 0.032), \nDSST ( p < 0.001), and CERAD total scores ( p = 0.033) \nin model 2. Women with 3 ≤ parity ≤ 4 showed slight \ndecreases in AFT and CERAD total, with a significant \nreduction in DSST. Parity ≥ 5 was associated with a sig -\nnificant decrease in CERAD trial 1 in models 1 and 2. \nHowever, all associations were non-significant in model \n3, indicating that higher parity’s impact on processing \nspeed and memory may be attenuated after adjustment. \n(Supplementary table S1, Supplementary table S2, Sup -\nplementary table S3).\nParity ≥ 5 was negatively correlated with cognitive func-\ntion, manifesting in lower scores for AFT, DSST, CERAD \ntrial 1 and CERAD trial total in model 1 and model 2, but \nit was insignificant in model 3. The results were visually \ndepicted using a forest plot (Fig. 2).\nMenopause and cognitive function\nThe delayed menopause group had higher alcohol intake \n(p < 0.001) and a greater proportion of normal to over -\nweight BMI (p = 0.041). Hypertension was more prevalent \nin the early menopause group (p = 0.022) (Table 4). Linear \nregression analysis showed that normal menopause was \npositively associated with CERAD trial 1 scores, with \nhigher scores compared to the reference group (p = 0.004) \nin model 1 and ( p = 0.005) in model 2. Delayed meno -\npause had a significant positive effect on CERAD trial 1 \nscores across all models, with the strongest association \nin model 1 ( p < 0.001) and maintaining significance in \nmodel 3 ( p = 0.002). It also showed a strong positive cor -\nrelation with CERAD total scores, particularly in model \n1 (p = 0.007), remaining significant in model 3 ( p = 0.031). \nDelayed menopause also positively affected CERAD \ndelayed recall scores in all models, with the strongest \nassociation in model 1 ( p < 0.001), persisting through to \nmodel 3 (p = 0.006).\nThese findings indicate a positive correlation between \nthe age at menopause onset and cognitive function, with \ndelayed menopause consistently linked to higher cogni -\ntive scores in CERAD trial 1, CERAD total, and CERAD \ndelayed recall. The results were visually depicted using a \nforest plot (Fig. 3).\nReproductive span and cognitive function\nIn linear regression analysis, reproductive span remained \nsignificantly associated with CERAD trial 1 across all \nmodels ( p < 0.001) and CERAD total recall ( p < 0.001 in \nmodels 1 and 2, p = 0.019 in model 3). A strong positive \nassociation with DSST was observed in models 1 and 2 \n(p < 0.001), persisting after full adjustment in model 3 \n(p = 0.018). However, associations with CERAD delayed \nrecall and AFT were not significant in model 3 ( p = 0.12 \nand p = 0.2, respectively). These findings suggest that lon-\nger reproductive span is associated with better cognitive \nperformance, particularly in verbal memory, attention, \nand processing speed (Table 5).\nDiscussion\nThe findings of this study highlight the complex relation-\nships between various physiological stages, reproductive \nhistories, and cognitive function in women in later life. \nWe observed a negative association between higher par -\nity (five or more) and cognitive performance, particularly \nin CERAD trial 1, CERAD trial total, AFT, and DSST. \nHowever, this relationship was attenuated after adjust -\ning for factors such as BMI, alcohol intake, smoking, PIR, \neducation, race/ethnicity, hypertension, and diabetes. \nOur findings also suggest that delayed menopause is pos -\nitively associated with better cognitive performance. An \nelevated risk of several health problems, such as cardio -\nvascular conditions, osteoporosis, and cognitive decline, \n\nPage 7 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nis associated with early menopause [ 32]. Using early \nmenopause as the reference group allowed for a clearer \ncomparison of how different reproductive factors influ -\nence cognitive outcomes. The delayed menopause group \ndisplayed higher scores in immediate recall (CERAD trial \n1), CERAD delayed recall, and the total score of CERAD \ntrials when compared to the early menopause group. \nMoreover, reproductive span demonstrated a steady \npositive correlation with cognitive functioning, especially \nin the context of CERAD trial 1, as well as total recall \nand DSST scores, throughout all models. The consistent \npositive association between delayed menopause and \nTable 2 Characteristics and cognitive function of study participants grouped by number of pregnancies\nCharacteristic Number of pregnancies \n≤ 2, N = 217 (36.94%)1\n3 ≤ Number of pregnancies \n≤ 4, N = 274 (38.78%)1\nNumber of pregnancies \n≥ 5, N = 207 (24.28%)1\np-val-\nue2\nAge group < 0.001\n 60–69 years 122 (59.29%) 135 (42.89%) 92 (42.06%)\n 70–79 years 54 (22.35%) 77 (31.46%) 69 (34.14%)\n ≥ 80 years 41 (18.36%) 62 (25.64%) 46 (23.80%)\nRace/Ethnicity < 0.001\n Non-Hispanic White 128 (82.70%) 143 (78.69%) 84 (67.45%)\n Non-Hispanic Black 33 (5.35%) 64 (10.76%) 55 (15.42%)\n Mexican American 13 (2.26%) 21 (3.46%) 36 (8.78%)\n Other Hispanic 17 (2.62%) 23 (2.94%) 23 (5.67%)\n Other/multiracial 26 (7.07%) 23 (4.15%) 9 (2.69%)\nEducation attainment 0.013\n Less Than 9th Grade 14 (4.03%) 11 (3.06%) 42 (12.17%)\n 9-11th Grade 17 (6.54%) 46 (15.45%) 50 (18.75%)\n High School Graduate/GED 52 (24.37%) 77 (27.53%) 49 (27.88%)\n Some College or AA degree 81 (38.32%) 80 (28.10%) 49 (28.39%)\n College Graduate or above 53 (26.74%) 60 (25.85%) 17 (12.80%)\nAlcohol intake 0.3\n 1–5 drinks/month 100 (50.36%) 101 (41.20%) 93 (46.19%)\n 5–10 drinks/month 8 (3.02%) 6 (2.91%) 4 (2.11%)\n ≥ 10 drinks/month 24 (13.86%) 31 (15.44%) 13 (8.46%)\n Non-drinker 85 (32.76%) 136 (40.45%) 97 (43.24%)\nSmoke group 0.5\n Current smoker 24 (12.72%) 30 (9.68%) 28 (13.19%)\n Former smoker 72 (37.56%) 80 (32.85%) 66 (35.56%)\n Never smoker 121 (49.73%) 164 (57.47%) 113 (51.24%)\nBMI group 0.4\n Underweight (< 18.5) 5 (1.23%) 3 (0.47%) 3 (1.31%)\n Normal (18.5 to < 25) 55 (26.98%) 63 (25.31%) 42 (20.19%)\n Overweight (25 to < 30) 59 (28.81%) 77 (33.37%) 73 (39.32%)\n Obese (30 or greater) 98 (42.98%) 131 (40.85%) 89 (39.17%)\nHypertension 167 (75.50%) 204 (69.74%) 162 (72.87%) 0.5\nDiabetes 83 (38.27%) 102 (30.36%) 106 (43.76%) 0.064\nPIR < 0.001\n ≤ 1.3 56 (17.18%) 68 (16.58%) 109 (37.74%)\n 1.3 < to ≤ 3.5 92 (42.64%) 130 (46.29%) 75 (45.04%)\n > 3.5 69 (40.18%) 76 (37.13%) 23 (17.22%)\nCERAD1 5.00 (4.00, 6.00) 5.00 (4.00, 6.00) 5.00 (4.00, 6.00) 0.6\nCERAD2 8.00 (7.00, 9.00) 7.00 (6.00, 9.00) 7.00 (6.00, 9.00) 0.4\nCERAD3 8.00 (7.00, 9.00) 8.00 (7.00, 9.00) 8.00 (7.00, 9.00) 0.6\nCERAD total 21.0 (19.0, 24.0) 21.0 (17.0, 24.0) 21.0 (17.0, 24.0) 0.4\nCERAD delay recall 7.00 (6.00, 9.00) 7.00 (5.00, 8.00) 7.00 (5.00, 8.00) 0.078\nAFT 17.0 (15.0, 21.0) 17.0 (14.0, 20.0) 17.0 (13.0, 19.0) 0.2\nDSST 54 (45, 67) 51 (42, 64) 49 (33, 62) < 0.001\n1n (weighted %); Median (IQR)\n2chi-squared test with Rao & Scott’s second-order correction; Wilcoxon rank-sum test for complex survey samples\n\nPage 8 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \ncognitive function across various measures suggested \nthat an extended reproductive lifespan may contribute to \nbetter cognitive health. Furthermore, our study highlights \nthe impact of socioeconomic factors on both parity and \ncognitive outcomes. The relationship between number of \npregnancies and cognitive function is a complex and con-\ntroversial topic. In some cognitive tasks, such as CERAD \ntrial 1, women with a higher number of pregnancies per -\nformed better, while in others, like DSST and CERAD \ndelayed recall, they exhibited poorer cognitive perfor -\nmance. This discrepancy may be attributed to differential \nsensitivity of specific cognitive domains to parity, as well \nas to the potential confounding effects of factors such as \nTable 3 Characteristics and cognitive function of study participants grouped by parity\nCharacteristic Parity ≤ 2, N = 304 (49.79%%)1 3 ≤ Parity ≤ 4, N = 275 (37.05%)1 Parity ≥ 5, N = 119 (13.15%)1 p-value2\nAge group < 0.001\n 60–69 years 180 (60.27%) 127 (38.37%) 42 (34.38%)\n 70–79 years 73 (23.02%) 77 (31.64%) 50 (42.26%)\n ≥ 80 years 51 (16.71%) 71 (29.99%) 27 (23.36%)\nRace/Ethnicity < 0.001\n Non-Hispanic White 171 (81.80%) 143 (77.39%) 41 (61.13%)\n Non-Hispanic Black 57 (6.91%) 60 (10.95%) 35 (18.21%)\n Mexican American 18 (2.27%) 28 (4.71%) 24 (10.88%)\n Other Hispanic 22 (2.36%) 27 (3.78%) 14 (6.89%)\n Other multiracial 36 (6.67%) 17 (3.17%) 5 (2.89%)\nEducation attainment < 0.001\n Less Than 9th Grade 14 (2.99%) 19 (4.62%) 34 (18.48%)\n 9-11th Grade 31 (7.47%) 51 (17.28%) 31 (21.60%)\n High School Grad/GED 67 (22.17%) 84 (31.33%) 27 (28.93%)\n Some College or AA degree 116 (38.24%) 72 (26.23%) 22 (24.24%)\n College Graduate or above 76 (29.14%) 49 (20.55%) 5 (6.75%)\nAlcohol intake 0.047\n 1–5 drinks/month 137 (49.79%) 107 (41.77%) 50 (42.01%)\n 5–10 drinks/month 10 (3.01%) 7 (2.42%) 1 (2.73%)\n ≥ 10 drinks/month 37 (15.03%) 27 (14.35%) 4 (2.72%)\n Non-drinker 120 (32.17%) 134 (41.46%) 64 (52.53%)\nSmoke group 0.4\n Current smoker 39 (13.77%) 28 (8.66%) 15 (12.09%)\n Former smoker 96 (35.99%) 88 (35.71%) 34 (31.14%)\n Never smoker 169 (50.24%) 159 (55.63%) 70 (56.77%)\nBMI group 0.5\n Underweight (< 18.5) 7 (1.16%) 4 (1.02%) 0 (0.00%)\n Normal (18.5 to < 25) 74 (27.87%) 63 (22.56%) 23 (18.62%)\n Overweight (25 to < 30) 86 (30.14%) 82 (35.25%) 41 (38.51%)\n Obese (30 or greater) 137 (40.84%) 126 (41.17%) 55 (42.88%)\nHypertension 230 (72.13%) 211 (73.12%) 92 (73.11%) > 0.9\nDiabetes 112 (35.19%) 111 (33.22%) 68 (50.99%) 0.045\nPIR < 0.001\n ≤ 1.3 81 (17.37%) 85 (21.73%) 67 (39.83%)\n 1.3 < to ≤ 3.5 126 (41.90%) 128 (46.62%) 43 (49.44%)\n > 3.5 97 (40.74%) 62 (31.65%) 9 (10.73%)\nCERAD1 5.00 (4.00, 7.00) 5.00 (4.00, 6.00) 5.00 (4.00, 6.00) 0.010\nCERAD2 8.00 (7.00, 9.00) 7.00 (6.00, 9.00) 7.00 (6.00, 8.00) 0.038\nCERAD3 9.00 (7.00, 10.00) 8.00 (7.00, 9.00) 8.00 (7.00, 9.00) 0.016\nCERAD total 21.0 (19.0, 25.0) 20.0 (17.0, 23.0) 21.0 (17.0, 23.0) 0.013\nCERAD delay recall 7.00 (6.00, 9.00) 7.00 (5.00, 8.00) 6.00 (5.00, 7.00) 0.008\nAFT 18.0 (15.0, 22.0) 17.0 (14.0, 20.0) 15.0 (12.0, 19.0) 0.018\nDSST 57 (45, 68) 50 (39, 62) 45 (25, 56) < 0.001\n1n (weighted %); Median (IQR)\n2chi-squared test with Rao & Scott’s second-order correction; Wilcoxon rank-sum test for complex survey samples\n\nPage 9 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nFig. 2 Forest plot of the associations between parity and cognitive test scores for CERAD trial 1, CERAD trial total, AFT, and DSST. Parity ≤ 2 was used as the \nreference group. Parity ≥ 5 was negatively correlated with cognitive function, manifesting in lower scores in model 1 and model 2, but it was insignificant \nin model 3. Model 1: no adjusted; Model 2: adjusted for age; Model 3: were adjusted for age, BMI, alcohol intake, smoking, PIR, education, race/ethnicity, \nhypertension, and diabetes\n \n\nPage 10 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \noverall health status, hormonal fluctuations, and physi -\nological changes.\nIn our study, the parity had a nuanced relationship with \ncognitive function. A higher parity, particularly more than \nfive, was negatively associated with cognitive function, \nespecially in terms of delayed recall and word learning. \nIt was consistent with the research results of others. \nYang et al. discovered an inverse relationship between \nthe number of children and cognitive functioning among \nolder adults. In contrast to older adults who have four \nchildren, those with more than five children exhibited a \nnotable decline in their Mini-Mental State Examination \nTable 4 Characteristics and cognitive function of study participants grouped by menopause status\nCharacteristic Early menopause, N = 264 \n(39.05%)1\nNormal menopause, \nN = 388 (54.26%)1\nDelayed menopause, N = 46 \n(6.69%)1\np-val-\nue2\nAge group 0.8\n 60–69 years 125 (47.47%) 200 (49.20%) 24 (52.55%)\n 70–79 years 78 (27.05%) 111 (30.23%) 11 (26.61%)\n ≥ 80 years 61 (25.48%) 77 (20.57%) 11 (20.85%)\nRace/Ethnicity 0.8\n Non-Hispanic White 143 (77.88%) 190 (77.18%) 22 (77.02%)\n Non-Hispanic Black 64 (11.01%) 76 (8.99%) 12 (10.75%)\n Mexican American 22 (3.48%) 45 (5.21%) 3 (1.83%)\n Other Hispanic 20 (3.04%) 38 (3.70%) 5 (4.24%)\n Other/multiracial 15 (4.59%) 39 (4.92%) 4 (6.16%)\nEducation attainment 0.4\n Less Than 9th Grade 31 (7.96%) 32 (4.02%) 4 (5.13%)\n 9-11th Grade 38 (11.79%) 66 (12.79%) 9 (21.20%)\n High School Grad/GED 76 (27.46%) 90 (26.30%) 12 (21.80%)\n Some College or AA degree 83 (34.04%) 115 (30.73%) 12 (29.64%)\n College Graduate or above 36 (18.76%) 85 (26.17%) 9 (22.23%)\nAlcohol intake < 0.001\n 1–5 drinks/month 119 (52.00%) 158 (41.85%) 17 (41.55%)\n 5–10 drinks/month 6 (1.68%) 6 (2.22%) 6 (13.36%)\n ≥ 10 drinks/month 18 (7.14%) 44 (16.58%) 6 (20.53%)\n Non-drinker 121 (39.18%) 180 (39.34%) 17 (24.57%)\nSmoke group 0.078\n Current smoker 38 (16.00%) 39 (8.60%) 5 (11.06%)\n Former smoker 98 (38.56%) 101 (32.72%) 19 (36.44%)\n Never smoker 128 (45.44%) 248 (58.68%) 22 (52.50%)\nBMI group 0.041\n Underweight (< 18.5) 4 (0.65%) 6 (1.20%) 1 (0.78%)\n Normal (18.5 to < 25) 44 (16.83%) 106 (29.30%) 10 (33.06%)\n Overweight (25 to < 30) 89 (35.79%) 108 (31.20%) 12 (33.31%)\n Obese (30 or greater) 127 (46.74%) 168 (38.30%) 23 (32.85%)\nHypertension 222 (79.82%) 276 (67.70%) 35 (70.66%) 0.022\nDiabetes 126 (43.28%) 148 (33.26%) 17 (23.78%) 0.075\nPIR 0.2\n ≤ 1.3 95 (22.30%) 125 (22.51%) 13 (15.14%)\n 1.3 < to ≤ 3.5 111 (46.86%) 167 (44.30%) 19 (34.38%)\n > 3.5 58 (30.83%) 96 (33.18%) 14 (50.48%)\nCERAD1 5.00 (4.00, 6.00) 5.00 (4.00, 6.00) 6.00 (5.00, 7.00) < 0.001\nCERAD2 7.00 (6.00, 8.00) 8.00 (6.00, 9.00) 8.00 (7.00, 9.00) 0.047\nCERAD3 8.00 (7.00, 9.00) 8.00 (7.00, 9.00) 9.00 (7.00, 10.00) 0.14\nCERAD total 21.0 (18.0, 23.0) 21.0 (18.0, 24.0) 23.0 (20.0, 27.0) 0.006\nCERAD delay recall 7.00 (5.00, 8.00) 7.00 (5.00, 8.00) 8.00 (7.00, 9.00) 0.002\nAFT 17.0 (13.0, 20.0) 18.0 (14.0, 21.0) 17.0 (14.0, 22.0) 0.2\nDSST 49 (40, 61) 54 (42, 65) 58 (45, 67) 0.013\n1n (weighted %); Median (IQR)\n2chi-squared test with Rao & Scott’s second-order correction; Wilcoxon rank-sum test for complex survey samples\n\nPage 11 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nscores [33]. The association between increased parity and \ncognitive decline may involve multiple factors, including \nbrain structure, hormonal regulation, metabolic burden, \nand inflammation. Pritschet et al. discovered that as the \nnumber of pregnancies rises, there is a notable decrease \nin gray matter volume as well as in vital areas such as the \nhippocampus, hypothalamus, thalamus, and brainstem. \nNotably, the volume of the hippocampus persisted in \ndeclining with multiple pregnancies, while a consistent \nreduction was observed in the para-hippocampal cortex. \nSuch alterations in structure can negatively affect mem -\nory, emotional management, and cognitive regulation, \nresulting in a progressive decrease in cognitive function \n[34]. Hoekzema et al. found that the volume of gray mat -\nter decreases during pregnancy, and this reduction does \nnot return to baseline levels for a minimum of two years \nafter giving birth [ 35]. Long-term hormonal fluctuations \nalso contribute to cognitive decline, as elevated levels of \nestrogen and progesterone during pregnancy enhance \nshort-term neuroplasticity [ 36]. However, as parity \nincreases, the adaptive effects of these hormones may \ndiminish, especially during the postpartum period when \nestrogen levels drop, potentially exacerbating neurode -\ngenerative processes [ 37]. Furthermore, higher parity is \nassociated with chronic diseases such as hypertension \nand diabetes, which can trigger inflammatory responses. \nLong-term low-grade inflammation may negatively affect \nthe nervous system, contributing to cognitive impair -\nment or dementia [ 38]. Jung’s research indicates that \nhaving many children may enhance the risk of cognitive \ndeterioration or elevate the likelihood of dementia in \nelderly women by worsening atrophy in the hippocam -\npus or cortex, independent of amyloid factors [39]. While \nour study initially demonstrated significant associations \nTable 5 Linear regression β(95%CI) of the association between reproductive span and cognitive function\nReproductive span\nCharacteristic  Model 1  p-value   Model 2  p-value   Model 3 p-value\nCERAD1 0.04 (0.03, 0.06) < 0.001 0.04 (0.03, 0.06) < 0.001 0.04 (0.02, 0.05) < 0.001\nCERAD2 0.01 (0.00, 0.03) 0.081 0.01 (0.00, 0.03) 0.14 0.01 (-0.01, 0.03) 0.5\nCERAD3 0.02 (0.01, 0.03) 0.003 0.01 (0.00, 0.02) 0.005 0.01 (0.00, 0.02) 0.13\nCERAD total 0.07 (0.04, 0.10) < 0.001 0.07 (0.04, 0.10) < 0.001 0.05 (0.01, 0.09) 0.019\nCERAD delay recall 0.02 (0.00, 0.04) 0.025 0.02 (0.00, 0.04) 0.037 0.02 (0.00, 0.04) 0.12\nAFT 0.06 (0.01, 0.12) 0.031 0.06 (0.01, 0.11) 0.032 0.03 (-0.02, 0.08) 0.2\nDSST 0.28 (0.16, 0.40) < 0.001 0.27 (0.15, 0.39) < 0.001 0.12 (0.03, 0.22) 0.018\nModel 1: no adjusted\nModel 2: adjusted for age\nModel 3: were adjusted for age, BMI, alcohol intake, smoking, PIR, education, race/ethnicity, hypertension, and diabetes\nFig. 3 Forest plot of the associations between menopause and cognitive test scores for CERAD trial 1, CERAD trial total, and CERAD delayed recall. Early \nmenopause was used as the reference group. There existed a positive correlation between the age at menopause and cognitive function. A later onset \nof menopause is associated with higher cognitive scores in CERAD trial 1, CERAD trial total, and CERAD delayed recall. Delayed menopause exhibited the \nmost robust positive correlation with cognitive function, with statistical significance consistently observed across model 1 and model 2, and model 3. \nModel 1: no adjusted; Model 2: adjusted for age; Model 3: were adjusted for age, BMI, alcohol intake, smoking, PIR, education, race/ethnicity, hyperten -\nsion, and diabetes\n \n\nPage 12 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nbetween higher parity (≥ 5) and cognitive outcomes, \nincluding AFT, DSST, and CERAD trials, these associa -\ntions were attenuated and became non-significant after \nadjusting for covariates in fully adjusted models. This \nattenuation suggests that the observed effects of parity \non cognitive outcomes may be explained by other fac -\ntors, such as age, education level, socioeconomic status, \nor comorbidities, which are closely related to both par -\nity and cognitive function. It is essential to recognize that \nparity might not serve as a direct risk factor; instead, it \nmay function as a proxy for a complex interaction of bio -\nlogical, social, and environmental factors that together \naffect cognitive health [ 40]. For instance, higher parity is \nfrequently correlated with lower levels of socioeconomic \nstatus and educational attainment [ 41]. Moreover, physi-\nological alterations linked to multiple pregnancies, such \nas changes in hormone levels, shifts in nutritional status, \nand heightened caregiving duties, could lead to varia -\ntions in cognitive function [ 42]. When examining the \nlink between parity and cognitive function, it is essen -\ntial to consider the interplay of various factors, including \nrace/ethnicity, education level, smoking habits, alcohol \nconsumption, diabetes, and hypertension on cognitive \nperformance.\nDelayed menopause was consistently associated with \nbetter cognitive performance across multiple cognitive \nmeasures. Our study was consistent with the conclusions \nof others. For instance, Needham et al. demonstrated \nthat a later age at menopause is linked to better cogni -\ntive performance, particularly in areas like memory, \nvisuospatial skills, and assessments such as the DSST \nand face-name association tasks [ 43]. Similarly, a study \nusing data from the Medical Research Council’s pioneer-\ning National Survey of Health and Development showed \nthat later menopause is associated with improved perfor -\nmance on various cognitive tests, including the Adden -\nbrooke’s Cognitive Examination - Third Edition total \nscore and verbal fluency [ 44]. The link between delayed \nmenopause and cognitive function is likely attributed to \nprolonged exposure to estrogen, which has been shown \nto have protective effects on brain health. Estrogen con -\ntributes to protecting neurons, promoting the forma -\ntion of synapses, and improving memory and learning \nabilities [ 45]. Research supports the idea that increased \nestrogen exposure throughout a woman’s life is associ -\nated with a reduced risk of AD and that estrogen defi -\nciency negatively impacts brain structure and function \n[46]. Research conducted by Fan et al. demonstrates that \nestrogen receptors exhibit high expression levels in areas \nsuch as the hippocampus and prefrontal cortex, both of \nwhich are essential for cognitive functions like memory \nand attention [ 47]. Additionally, findings from Ishunina \net al. indicate that estrogens might have positive effects \non cognitive functions reliant on the hippocampus, \npotentially acting through the mediation of estrogen \nreceptor alpha [ 48]. In our study, the delayed meno -\npause group had higher alcohol intake, while hyperten -\nsion was more common in the early menopause group. \nIt is known that alcohol may increase estrogen levels in \nthe body [ 49], and since estrogen has a protective effect \non ovarian function, this could potentially delay the onset \nof menopause. Furthermore, alcohol may reduce oxida -\ntive stress in ovarian tissue, thereby protecting ovarian \nfunction and contributing to a later onset of menopause \n[50]. On the other hand, the loss of estrogen’s protective \neffects occurs when estrogen levels decline, impairing its \nability to dilate blood vessels, reduce peripheral vascular \nresistance, and regulate lipid metabolism, all of which are \nessential for cardiovascular protection. In women with \nearly menopause, the decline in estrogen levels weakens \nthese protective effects, leading to increased vascular \nconstriction, elevated peripheral vascular resistance, and \na subsequent rise in blood pressure [51].\nOur study indicates that a longer reproductive span \nis positively associated with better cognitive abilities, \nparticularly in memory and processing speed. Other’s \nstudy has shown that the hormonal changes associated \nwith a longer reproductive span, including prolonged \nestrogen exposure, may protect against age-related cog -\nnitive decline [ 52]. In our findings, a longer reproduc -\ntive span was consistently associated with higher scores \nin CERAD Immediate recall and total recall, reflecting \nenhanced verbal memory. This suggests that hormonal \nfactors across a longer reproductive lifespan contribute \nto enhanced neuroprotection. Additionally, the strong \nassociation between reproductive span and DSST scores \nunderscores the role of reproductive health in supporting \nprocessing speed.\nThese findings underscore the importance of repro -\nductive factors in cognitive aging and provide a basis \nfor identifying risk factors for cognitive health in elderly \nwomen and developing interventions for dementia pre -\nvention. These interventions can range from hormone \ntherapy and lifestyle modifications to advanced medical \nprocedures. Hormone replacement therapy is a widely \nused method for alleviating menopausal symptoms and \nmay also help in postponing the onset of menopause; \nhowever, it is crucial to consider the advantages in rela -\ntion to possible risks [ 53]. Lifestyle changes, such as \nmaintaining a balanced diet rich in phytoestrogens, \nhealthy fats, and essential vitamins, engaging in regular \nphysical activity, and practicing mindfulness, can play a \nsignificant role in managing menopausal symptoms and \npossibly delaying its onset [ 54]. Medical interventions, \nincluding the use of rapamycin to prolong ovarian func -\ntion [ 55] and ovarian tissue transplantation to preserve \nfertility and reverse menopause [ 56], show promise in \nextending the fertility window and delaying menopause.\n\nPage 13 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \nA few limitations of the present study needed to be \nnoted. Firstly, the cross-sectional design restricts our abil-\nity to draw causal conclusions, and the sample may not \nrepresent the full diversity of the population. Secondly, in \nthis observational study, residual and unmeasured con -\nfounding factors cannot be completely excluded. Thirdly, \nthe study used a single-sample database, which has cer -\ntain limitations, and can be further validated using multi-\nsample databases.\nFuture research may explore two main potential direc -\ntions. Firstly, future research should consider longitu -\ndinal designs to track women’s cognitive function from \ntheir childbearing years into later life, allowing for a bet -\nter understanding of temporal relationships and potential \ncausal mechanisms between the number of pregnancies, \nage at menopause, and cognitive function. Secondly, to \nfurther investigate the impact of parity and menopause \non cognitive health, future studies could focus on iden -\ntifying and evaluating biomarkers related to these repro -\nductive events.\nConclusions\nThis study explores the association between female \nreproductive factors and cognitive function in later life. \nThe findings indicate that higher parity is associated with \nreduced cognitive performance, while a later onset of \nmenopause and a longer reproductive span are linked to \nbetter cognitive outcomes. Incorporating reproductive \nfactors into the assessment of risk factors associated with \ncognitive impairment in older adults provides valuable \ninsights into preventing and addressing cognitive decline \nand dementia.\nAbbreviations\nNHANES  National Health and Nutrition Examination Survey\nCERAD  Consortium to Establish a Registry for Alzheimer’s Disease\nAFT  Animal Fluency test\nDSST  Digit Symbol Substitution Test\nBMI  Body mass index\nPIR  Poverty income ratio\nAD  Alzheimer’s disease\nCDC  Centers for Disease Control and Prevention\nGED  General educational development\nAA  Associate of Arts\nSupplementary Information\nThe online version contains supplementary material available at  h t t p  s : /  / d o i  . o  r \ng /  1 0 .  1 1 8 6  / s  1 2 8 8 9 - 0 2 5 - 2 2 9 6 6 - z.\nSupplementary Material 1\nSupplementary Material 2\nSupplementary Material 3\nAcknowledgements\nWe thank all participants in the NHANES. We thank the NHANES research team \nfor providing the data.\nAuthor contributions\nT.L. and F.C. were involved in developing and designing the study concept; \nA.H., H.W., L.Y., Y.L., and H.Q. were involved in the data acquisition and analysis; \nA.H., L.X. contributed to the initial manuscript writing. All authors revised and \nagreed to the final version of this article.\nFunding\nThis work was supported by the National Nature Science Foundation of \nHainan Province [Grant No. 821RC675 for Tao Liu]; the Key science and \ntechnology project of Hainan Province [Grant No. ZDYF2023SHFZ096 for \nTao Liu, ZDYF2024SHFZ058 for Feng Chen]; the National Nature Science \nFoundation of China [Grant No. 82160327 for Tao Liu, 82271977 for Feng \nChen]; Hainan Academician Innovation Platform Scientific Research Project \n[Grant No. YSPTZX202135 for Tao Liu]; Joint Program on Health Science & \nTechnology Innovation of Hainan Province [Grant No. WSJK2024QN075 for \nAnquan Hu]. The Innovation Platform for Academicians of Hainan Province; \nHainan Province Clinical Medical Center\nData availability\nThe data that support the findings of this study are openly available on the \nNHANES website and can be accessed at (URL  h t t p  s : /  / w w w  n .  c d c  . g o  v / n c  h s  / n h \na n e s /, accessed on 15 April 2025).\nDeclarations\nEthics approval and consent to participate\nAll methods were carried out in accordance with relevant guidelines and \nregulations. The Research Ethics Review Board (ERB) of the US National Center \nfor Healthcare Statistics (NCHS) authorized the 2011–2014 NHANES (protocol \nnumber: protocol#2011–17 and continuation of protocol #2011–17) ( h t t p  s : /  / \nw w w  . c  d c .  g o v  / n c h  s /  n h a  n e s  / a b o  u t  / e r b . h t m l). Before initiating data collection \nand the NHANES physical examinations, all eligible individuals had given their \ninformed consent.\nConsent for publication\nNot applicable.\nCompeting interests\nThe authors declare no competing interests.\nAuthor details\n1Department of Geriatric Center, Hainan General Hospital (Hainan \nAffiliated Hospital of Hainan Medical University), Haikou 570311, China\n2Department of Neurology, Hainan General Hospital (Hainan Affiliated \nHospital of Hainan Medical University), Haikou 570311, China\n3Department of Radiology, Hainan General Hospital (Hainan Affiliated \nHospital of Hainan Medical University), Haikou 570311, China\nReceived: 5 March 2024 / Accepted: 28 April 2025\nReferences\n1. Chalise HN. Aging: basic concept. Am J Biomedical Sci Res. 2019;1(1):8–10.\n2. 2023 Alzheimer’s disease facts and figures. Alzheimers Dement. \n2023;19(4):1598–1695.\n3. Nooyens ACJ, Wijnhoven HAH, Schaap LS, Sialino LD, Kok AAL, Visser \nM, Verschuren WMM, Picavet HSJ, van Oostrom SH. Sex differences \nin cognitive functioning with aging in the Netherlands. Gerontology. \n2022;68(9):999–1009.\n4. World Health Organization, Global Health Estimates. Life expectancy and \nleading causes of death and disability [Available from:  h t t p  s : /  / w w w  . w  h o .  i n t  / \nd a t  a /  g h o  / d a  t a / t  h e  m e s  / t h  e m e -  d e  t a i  l s /  G H O /  m o  r t a  l i t  y - a n  d -  g l o b a l - h e a l t h - e s t i \nm a t e s Accessed April 15, 2025.].\n5. Tsai CK, Chen YY, Chou CH, Kao TW, Liang CS, Yang FC, Wang CC, Lee JT, \nChen WL. Female reproductive health and cognitive function. Menopause. \n2020;27(12):1357–62.\n6. Song X, Wu J, Zhou Y, Feng L, Yuan JM, Pan A, Koh WP . Reproductive and hor-\nmonal factors and risk of cognitive impairment among Singapore Chinese \nwomen. Am J Obstet Gynecol. 2020;223(3):410. e411-410 e423.\n\nPage 14 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \n7. Jett S, Malviya N, Schelbaum E, Jang G, Jahan E, Clancy K, Hristov H, Pahlajani \nS, Niotis K, Loeb-Zeitlin S et al. Endogenous and exogenous Estrogen expo-\nsures: how women’s reproductive health can drive brain aging and inform \nAlzheimer’s prevention. Front Aging Neurosci. 2022;14.\n8. Lee S, Min JY, Kim B, Ha SW, Han JH, Min KB. Serum sodium in relation to vari-\nous domains of cognitive function in the elderly US population. BMC Geriatr. \n2021;21(1):328.\n9. CDC. National Health and Nutrition Examination Survey. Available online:  h t t \np  s : /  / w w w  . c  d c .  g o v  / n c h  s /  n h a n e s / In.\n10. Ge S, Dong F, Tian C, Yang CH, Liu M, Wei J. Serum soluble alpha-klotho \nKlotho and cognitive functioning in older adults aged 60 and 79: an analysis \nof cross-sectional data of the National health and nutrition examination \nsurvey 2011 to 2014. BMC Geriatr. 2024;24(1):245.\n11. Johnson CL, Paulose-Ram R, Ogden CL, Carroll MD, Kruszon-Moran D, \nDohrmann SM, Curtin LR. National health and nutrition examination survey: \nanalytic guidelines, 1999–2010. Vital Health Stat 2. 2013(161):1–24.\n12. Liang Y, Yang X, Jin J, Huang J, Wang Z, Zuo C, Wang S, Wang Q, Zhang X. \nDietary selenium intake, hypertension and cognitive function among US \nadults, NHANES 2011–2014. Sci Rep. 2024;14(1):25346.\n13. Fan D, Wang T, Xiang J, Bai Y, Zhang L, Wang X. Neutrophil percentage-to-\nalbumin ratio is associated with cognitive function in adults aged over 60 \nyears: an analysis of data from the NHANES 2011–2014. J Alzheimers Dis. \n2024;102(1):150–61.\n14. Cardoso R, Machado B, Steele P . Association between ultra-processed food \nconsumption and cognitive performance in US older adults: a cross-sectional \nanalysis of the NHANES 2011–2014. Eur J Nutr. 2022;61(8):3975–85.\n15. Botelho J, Leira Y, Viana J, Machado V, Lyra P , Aldrey JM, Pías-Peleteiro JM, \nBlanco J, Sobrino T, Mendes JJ. The role of inflammatory diet and vitamin \nD on the link between periodontitis and cognitive function: a mediation \nanalysis in older adults. Nutrients. 2021;13(3).\n16. Song W, Feng Y, Gong Z, Tian C. The association between dietary inflam-\nmatory index and cognitive performance in older adults aged 60 years and \nolder. Front Nutr. 2022;9:748000.\n17. Wang R, Wang W, Hu P , Zhang R, Dong X, Zhang D. Association of dietary \nVitamin D Intake, Serum 25(OH)D(3), 25(OH)D(2) with cognitive performance \nin the elderly. Nutrients. 2021;13(9).\n18. Wang X, Xiao P , Wang R, Luo C, Zhang Z, Yu S, Wu Q, Li Y, Zhang Y, Zhang H, \net al. Relationships between urinary metals concentrations and cognitive \nperformance among U.S. Older people in NHANES 2011–2014. Front Public \nHealth. 2022;10:985127.\n19. Fan Y, Zhang Y, Li J, Liu Y, Chang H, Jiang Y, Tuo X, Zhou L, Yu Y. Association \nbetween healthy eating index-2015 and various cognitive domains in US \nadults aged 60 years or older: the National health and nutrition examination \nsurvey (NHANES) 2011–2014. BMC Public Health. 2021;21(1):1862.\n20. Christensen K, Gleason CE, Mares JA. Dietary carotenoids and cognitive func-\ntion among US adults, NHANES 2011–2014. Nutr Neurosci. 2020;23(7):554–62.\n21. Heymsfield SB, Smith B, Chung EA, Watts KL, Gonzalez MC, Yang S, Heo \nM, Thomas DM, Turner D, Bosy-Westphal A, et al. Phenotypic differences \nbetween people varying in muscularity. J Cachexia Sarcopenia Muscle. \n2022;13(2):1100–12.\n22. Scinicariello F, Buser MC, Balluz L, Gehle K, Murray HE, Abadin HG, Attanasio \nR. Perfluoroalkyl acids, hyperuricemia and gout in adults: analyses of NHANES \n2009–2014. Chemosphere. 2020;259:127446.\n23. Flanagin A, Frey T, Christiansen SL, Bauchner H. The reporting of race \nand ethnicity in medical and science journals: comments invited. JAMA. \n2021;325(11):1049–52.\n24. Centers for Disease Control and Prevention, National Health and Nutrition \nExamination. Survey 2011–2012 Data Documentation, Codebook, and \nFrequencies: Reproductive Health (RHQ_G) [ h t t p  s : /  / w w w  n .  c d c  . g o  v / N c  h s  / D a  t \na /  N h a n  e s  / P u  b l i  c / 2 0  1 1  / D a t a F i l e s / R H Q _ G . h t m]\n25. Centers for Disease Control and Prevention, National Health and Nutrition \nExamination. Survey 2013–2014 Data Documentation, Codebook, and \nFrequencies: Reproductive Health (RHQ_H) [ h t t p s :   /  / w w w  n  . c d   c . g   o v /  N c   h s /  D  a t  \na / N  h a   n e s  / P u   b l i  c /   2 0 1 3 / D a t a  F i l e  s  / R H Q _ H . h t m]\n26. Nelson LM, Covington SN, Rebar RW. An update: spontaneous premature \novarian failure is not an early menopause. Fertil Steril. 2005;83(5):1327–32.\n27. Rodriguez M, Shoupe D. Surgical menopause. Endocrinol Metabolism Clin. \n2015;44(3):531–42.\n28. Gorsi B, Hernandez E, Moore MB, Moriwaki M, Chow CY, Coelho E, Taylor E, \nLu C, Walker A, Touraine P , et al. Causal and candidate gene variants in a large \ncohort of women with primary ovarian insufficiency. J Clin Endocrinol Metab. \n2022;107(3):685–714.\n29. Muka T, Asllanaj E, Avazverdi N, Jaspers L, Stringa N, Milic J, Ligthart S, Ikram \nMA, Laven JSE, Kavousi M, et al. Age at natural menopause and risk of type 2 \ndiabetes: a prospective cohort study. Diabetologia. 2017;60(10):1951–60.\n30. Armstrong AW, Mehta MD, Schupp CW, Gondo GC, Bell SJ, Griffiths \nCEM. Psoriasis prevalence in adults in the united States. JAMA Dermatol. \n2021;157(8):940–6.\n31. Zhao C, Pu M, Wu C, Ding J, Guo J, Zhang G. Association between composite \ndietary antioxidant index and cognitive function impairment among the US \nolder adults: a cross-sectional study based on the NHANES 2011–2014. Front \nNutr. 2024;11:1471981.\n32. Sochocka M, Karska J, Pszczołowska M, Ochnik M, Fułek M, Fułek K, Kurpas D, \nChojdak-Łukasiewicz J, Rosner-Tenerowicz A, Leszek J. Cognitive decline in \nearly and premature menopause. Int J Mol Sci. 2023;24(7).\n33. Yang HL, Zhang SQ, Zhang S, Wu YY, Luo RD. Fertility experiences and \nlater-life cognitive function among older adults in China. Am J Hum Biol. \n2022;34(10):e23786.\n34. Pritschet L, Taylor CM, Cossio D, Faskowitz J, Santander T, Handwerker \nDA, Grotzinger H, Layher E, Chrastil ER, Jacobs EG. Neuroanatomical \nchanges observed over the course of a human pregnancy. Nat Neurosci. \n2024;27(11):2253–60.\n35. Hoekzema E, Barba-Muller E, Pozzobon C, Picado M, Lucco F, Garcia-Garcia D, \nSoliva JC, Tobena A, Desco M, Crone EA, et al. Pregnancy leads to long-lasting \nchanges in human brain structure. Nat Neurosci. 2017;20(2):287–96.\n36. Paternina-Die M, Martínez-García M, Martín de Blas D, Noguero I, Servin-\nBarthet C, Pretus C, Soler A, López-Montoya G, Desco M, Carmona S. Women’s \nneuroplasticity during gestation, childbirth and postpartum. Nat Neurosci. \n2024;27(2):319–27.\n37. Wang X, Feng S, Deng Q, Wu C, Duan R, Yang L. The role of estrogen in \nAlzheimer’s disease pathogenesis and therapeutic potential in women. Mol \nCell Biochem. 2024.\n38. Miller EC, Conley P , Alirezaei M, Wolfova K, Gonzales MM, Tan ZS, Tom SE, \nYee LM, Brickman AM, Bello NA. Associations between adverse pregnancy \noutcomes and cognitive impairment and dementia: a systematic review and \nmeta-analysis. Lancet Healthy Longev. 2024;5(12):100660.\n39. Jung JH, Lee GW, Lee JH, Byun MS, Yi D, Jeon SY, Jung GJ, Joung H, Shin SA, \nKim YK et al. Multiparity, Brain Atrophy, and Cognitive Decline. Frontiers in \nAging Neuroscience. 2020;12.\n40. Gallagher M, Okonkwo OC, Resnick SM, Jagust WJ, Benzinger TLS, Rapp PR. \nWhat are the threats to successful brain and cognitive aging? Neurobiol \nAging. 2019;83:130–4.\n41. Giudicessi AJ, Saelzler UG, Shadyab AH, Posis AIB, Sundermann EE, Banks SJ, \nPanizzon MS. The mediating role of socioeconomic status on the relation-\nship between pregnancy history and later-life cognition. Climacteric. \n2022;25(6):627–33.\n42. Servin-Barthet C, Martínez-García M, Paternina-Die M, Marcos-Vidal L, Martín \nde Blas D, Soler A, Khymenets O, Bergé D, Casals G, Prats P , et al. Pregnancy \nentails a U-shaped trajectory in human brain structure linked to hormones \nand maternal attachment. Nat Commun. 2025;16(1):730.\n43. Needham LP , Lu K, Nicholas JM, Schott JM, Richards M, James SN. A compre-\nhensive assessment of age at menopause with well-characterized cognition \nat 70 years: A population-based British birth cohort. Maturitas. 2023;170:31–8.\n44. Needham L, Lu K, Nicholas JM, Schott JM, Richards M, James SN. Menopause \nand later-life cognition: findings from the longest‐running population‐based \nbirth cohort. Alzheimer’s Dement. 2021;17:S10.\n45. Jett S, Malviya N, Schelbaum E, Jang G, Jahan E, Clancy K, Hristov H, Pahlajani \nS, Niotis K, Loeb-Zeitlin S, et al. Endogenous and exogenous Estrogen expo-\nsures: how women’s reproductive health can drive brain aging and inform \nAlzheimer’s prevention. Front Aging Neurosci. 2022;14:831807.\n46. Epperson CN, Amin Z, Ruparel K, Gur R, Loughead J. Interactive effects of \nEstrogen and serotonin on brain activation during working memory and \naffective processing in menopausal women. Psychoneuroendocrinology. \n2012;37(3):372–82.\n47. Fan X, Zhao X, Zhu B, Qin H. Retrospective evaluation of the correlation \nbetween previous hospitalizations, the type of current living space, and qual-\nity of family function. Front Psychiatry. 2020;11:215.\n48. Ishunina TA, Fischer DF, Swaab DF. Estrogen receptor alpha and its splice vari-\nants in the hippocampus in aging and Alzheimer’s disease. Neurobiol Aging. \n2007;28(11):1670–81.\n49. Wiggs AG, Chandler JK, Aktas A, Sumner SJ, Stewart DA. The effects of diet \nand exercise on endogenous estrogens and subsequent breast Cancer risk in \npostmenopausal women. Front Endocrinol (Lausanne). 2021;12:732255.\n\nPage 15 of 15\nHu et al. BMC Public Health         (2025) 25:1811 \n50. Freeman JR, Whitcomb BW, Purdue-Smithe AC, Manson JE, Langton CR, \nHankinson SE, Rosner BA, Bertone-Johnson ER. Is alcohol consumption asso-\nciated with risk of early menopause?? Am J Epidemiol. 2021;190(12):2612–7.\n51. Li S, Tan I, Atkins E, Schutte AE, Gnanenthiran SR. The pathophysiology, \nprognosis and treatment of hypertension in females from pregnancy to Post-\nmenopause: A review. Curr Heart Fail Rep. 2024;21(4):322–36.\n52. Marta MA, Margot MCB, Anna BS, Gonzalo SB, Armand González E, Ann B, \nHlin K, Mahnaz S, Eider MAU, Sherezade FJ, et al. Reproductive span as a \nmodifier of the association between AD biomarkers and cognitive decline in \ncognitively unimpaired women. Alzheimer’s Dementia. 2023.\n53. Lau ES, Paniagua SM, Guseh JS, Bhambhani V, Zanni MV, Courchesne P , Lyass \nA, Larson MG, Levy D, Ho JE. Sex differences in Circulating biomarkers of \ncardiovascular disease. J Am Coll Cardiol. 2019;74(12):1543–53.\n54. Castelli V, Grassi D, Bocale R, d’Angelo M, Antonosante A, Cimini A, Ferri C, \nDesideri G. Diet and brain health: which role for polyphenols?? Curr Pharm \nDes. 2018;24(2):227–38.\n55. Dou X, Sun Y, Li J, Zhang J, Hao D, Liu W, Wu R, Kong F, Peng X, Li J. Short-term \nRapamycin treatment increases ovarian lifespan in young and middle-aged \nfemale mice. Aging Cell. 2017;16(4):825–36.\n56. Dolmans M-M, Donnez J, Cacciottola L. Fertility preservation: the challenge of \nfreezing and transplanting ovarian tissue. Trends Mol Med. 2021;27(8):777–91.\nPublisher’s note\nSpringer Nature remains neutral with regard to jurisdictional claims in \npublished maps and institutional affiliations.","source_license":"CC0","license_restricted":false}