Biological and social reproductive factors and late‐life cognitive function in middle‐aged and older Chinese women
Xueqin Li, Wanyu Zhao, Huirong Zheng, Bangshan Liu, Hongbo He, Jin Liu, Ling Jiang Li, Yan Zhang
TL;DR
A longer reproductive span is linked to better late-life cognition in Chinese women, while having more children is associated with poorer cognitive performance.
Contribution
This study is the first to integrate biological and social reproductive factors to explore sex-specific cognitive aging patterns in middle-aged and older Chinese women.
Findings
Longer reproductive span was associated with better cognitive performance in women.
A higher number of children was linked to poorer cognition in both sexes, especially in women.
These associations remained robust after adjusting for age, education, and lifestyle factors.
Abstract
Few studies have concurrently examined the biological and social reproductive factors in relation to women's cognitive aging. We analyzed 8577 women and 7872 men ≥45 years of age from the China Health and Retirement Longitudinal Study. Biological reproductive factors included reproductive span, age at menarche, and age at menopause; social reproductive factors included number of children and age at first live birth. Multivariable regression models were sequentially adjusted for age, childhood cognition proxy, education, and current health and lifestyle factors. Longer reproductive span was associated with better cognitive performance in women, whereas a higher number of children were linked to poorer cognition in both sexes, particularly in women. These associations remained robust after full adjustment, compared with age at menarche, age at menopause, and age at first birth.…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1| Characteristic | Overall | Short span | Long span |
|
|---|---|---|---|---|
| Age, years, mean (SD) | 62.4 (8.5) | 62.7 (8.7) | 62.1 (8.4) | 0.013 |
| Education, | <0.001 | |||
| Preprimary | 1623 (29.1) | 860 (32.4) | 763 (26.0) | |
| Primary | 2368 (42.4) | 1169 (44.1) | 1199 (40.9) | |
| Lower secondary | 989 (17.7) | 416 (15.7) | 573 (19.5) | |
| Upper secondary or above | 605 (10.8) | 208 (7.8) | 397 (13.5) | |
| Residence, | <0.001 | |||
| Urban | 1092 (21.0) | 424 (17.2) | 668 (24.5) | |
| Intergration zone | 362 (7.0) | 158 (6.4) | 204 (7.5) | |
| Rural | 3741 (72.0) | 1885 (76.3) | 1856 (68.0) | |
| Special zone | 3 (0.1) | 2 (0.1) | 1 (< 1) | |
| Mother education, | 0.043 | |||
| Preprimary | 4080 (88.2) | 1978 (89.4) | 2102 (87.1) | |
| Primary | 451 (9.8) | 200 (9.0) | 251 (10.4) | |
| Lower secondary | 61 (1.3) | 24 (1.1) | 37 (1.5) | |
| Upper secondary or above | 32 (0.7) | 10 (0.5) | 22 (0.9) | |
| Father education, | 0.34 | |||
| Preprimary | 2474 (56.4) | 1206 (57.6) | 1268 (55.3) | |
| Primary | 1503 (34.3) | 701 (33.5) | 802 (35.0) | |
| Lower secondary | 236 (5.4) | 103 (4.9) | 133 (5.8) | |
| Upper secondary or above | 173 (3.9) | 83 (4.0) | 90 (3.9) | |
| Childhood economy, | 0.018 | |||
| Better | 506 (10.6) | 213 (9.3) | 293 (11.8) | |
| Same | 2459 (51.4) | 1187 (51.8) | 1272 (51.1) | |
| Worse | 1817 (38.0) | 892 (38.9) | 925 (37.1) | |
| Childhood safe, | 0.58 | |||
| Very safe | 2334 (50.1) | 1129 (50.6) | 1205 (49.7) | |
| Somewhat safe | 1914 (41.1) | 900 (40.4) | 1014 (41.8) | |
| Not safe | 408 (8.8) | 201 (9.0) | 207 (8.5) | |
| Childhood health | 0.19 | |||
| Healthier | 1665 (34.9) | 768 (33.6) | 897 (36.1) | |
| Average | 2455 (51.5) | 1198 (52.5) | 1257 (50.6) | |
| Less healthy | 645 (13.5) | 317 (13.9) | 328 (13.2) | |
| Marital status, | 0.015 | |||
| Married | 4330 (78.9) | 2021 (77.5) | 2309 (80.1) | |
| Others | 1160 (21.1) | 588 (22.5) | 572 (19.9) | |
| Smoke, | 0.96 | |||
| Never smoke | 4410 (92.8) | 2091 (92.9) | 2319 (92.7) | |
| Former smoke | 119 (2.5) | 56 (2.5) | 63 (2.5) | |
| Current smoke | 224 (4.7) | 104 (4.6) | 120 (4.8) | |
| Drink, | 0.016 | |||
| Never drink | 4200 (75.2) | 2009 (75.7) | 2191 (74.7) | |
| Former drink | 421 (7.5) | 219 (8.3) | 202 (6.9) | |
| Current drink | 965 (17.3) | 425 (16.0) | 540 (18.4) | |
| Disease, | 0.37 | |||
| Yes | 3643 (65.5) | 1748 (66.1) | 1895 (65.0) | |
| No | 1916 (34.5) | 895 (33.9) | 1021 (35.0) | |
| Menarche age | 16.1 (2.1) | 16.9 (1.9) | 15.2 (1.8) | <0.001 |
| Menarche age, | <0.001 | |||
| Early menarche | 3352 (60.7) | 2037 (77.7) | 1315 (45.3) | |
| Late menarche | 2174 (39.3) | 584 (22.3) | 1590 (54.7) | |
| Menopause age | 49.5 (3.6) | 47.0 (3.0) | 51.8 (2.4) | <0.001 |
| Menopause age, | <0.001 | |||
| Early menopause | 2399 (43.4) | 1989 (75.8) | 410 (14.1) | |
| Late menopause | 3131 (56.6) | 634 (24.2) | 2497 (85.9) | |
| Reproductive life span | 33.4 (4.1) | 30.0 (2.8) | 36.5 (2.2) | <0.001 |
| Children | 2.8 (1.4) | 2.9 (1.4) | 2.7 (1.4) | <0.001 |
| Children, | <0.001 | |||
| Fewer children | 2381 (50.1) | 1083 (47.4) | 1298 (52.5) | |
| More children | 2374 (49.9) | 1201 (52.6) | 1173 (47.5) | |
| Age at first live birth | 24.8 (4.2) | 24.7 (4.2) | 24.9 (4.1) | 0.18 |
| Age at first live birth, | 0.013 | |||
| Early age at first birth | 2201 (52.7) | 1091 (54.7) | 1110 (50.8) | |
| Late age at first birth | 1979 (47.3) | 905 (45.3) | 1074 (49.2) | |
| Cognitive z‐score | −0.2 (1.0) | −0.3 (1.0) | −0.1 (1.1) | <0.001 |
| Cognitive function, | <0.001 | |||
| Low | 3082 (55.2) | 1579 (59.5) | 1503 (51.2) | |
| High | 2504 (44.8) | 1074 (40.5) | 1430 (48.8) |
| Model 0 | Model 1 | Model 2 | ||||
|---|---|---|---|---|---|---|
| OR (95% CI) |
| OR (95% CI) |
| OR (95% CI) |
| |
| Number | 5586 | 4077 | 4077 | |||
| Short span | 1 (Reference) | NA | 1 (Reference) | NA | 1 (Reference) | NA |
| Long span | 1.40 (1.26,1.56) | <0.001 | 1.31 (1.15,1.50) | <0.001 | 1.20 (1.03,1.38) | 0.017 |
| Number | 6879 | 5157 | 5157 | |||
| Early menarche | 1 (Reference) | NA | 1 (Reference) | NA | 1 (Reference) | NA |
| Late menarche | 1.56 (1.41,1.72) | <0.001 | 1.29 (1.14,1.46) | <0.001 | 1.11 (0.98,1.27) | 0.105 |
| Number | 6396 | 4776 | 4776 | |||
| Early menopause | 1 (Reference) | NA | 1 (Reference) | NA | 1 (Reference) | NA |
| Late menopause | 1.06 (0.96,1.17) | 0.266 | 1.14 (1.00,1.29) | 0.039 | 1.07 (0.94,1.23) | 0.306 |
| Characteristic | Overall | fewer children | more children |
|
|---|---|---|---|---|
| Age, years, mean (SD) | 60.2 (9.2) | 56.4 (7.2) | 64.9 (9.3) | <0.001 |
| Education, | <0.001 | |||
| Preprimary | 2177 (28.1) | 791 (18.3) | 1386 (40.3) | |
| Primary | 3362 (43.4) | 1806 (41.9) | 1556 (45.3) | |
| Lower secondary | 1469 (19.0) | 1103 (25.6) | 366 (10.7) | |
| Upper secondary or above | 738 (9.5) | 611 (14.2) | 127 (3.7) | |
| Residence, | <0.001 | |||
| Urban | 1482 (19.6) | 1036 (24.6) | 446 (13.3) | |
| Intergration zone | 506 (6.7) | 354 (8.4) | 152 (4.5) | |
| Rural | 5570 (73.7) | 2820 (66.9) | 2750 (82.1) | |
| Special zone | 3 (< 1) | 3 (0.1) | 0 (0.0) | |
| Mother education, | <0.001 | |||
| Preprimary | 6320 (85.0) | 3265 (78.8) | 3055 (92.9) | |
| Primary | 922 (12.4) | 715 (17.2) | 207 (6.3) | |
| Lower secondary | 124 (1.7) | 106 (2.6) | 18 (0.5) | |
| Upper secondary or above | 66 (0.9) | 59 (1.4) | 7 (0.2) | |
| Father education, | <0.001 | |||
| Preprimary | 3790 (53.7) | 1869 (47.0) | 1921 (62.2) | |
| Primary | 2499 (35.4) | 1531 (38.5) | 968 (31.3) | |
| Lower secondary | 460 (6.5) | 333 (8.4) | 127 (4.1) | |
| Upper secondary or above | 315 (4.5) | 241 (6.1) | 74 (2.4) | |
| Childhood economy, | <0.001 | |||
| Better | 865 (11.3) | 518 (12.1) | 347 (10.2) | |
| Same | 4010 (52.2) | 2304 (53.8) | 1706 (50.2) | |
| Worse | 2812 (36.6) | 1464 (34.2) | 1348 (39.6) | |
| Childhood safe, | <0.001 | |||
| Very safe | 3720 (49.8) | 2034 (48.7) | 1686 (51.2) | |
| Somewhat safe | 3116 (41.7) | 1819 (43.5) | 1297 (39.4) | |
| Not safe | 637 (8.5) | 326 (7.8) | 311 (9.4) | |
| Childhood health, | 0.018 | |||
| Healthier | 2720 (35.4) | 1574 (36.7) | 1146 (33.8) | |
| Average | 3931 (51.2) | 2137 (49.9) | 1794 (52.9) | |
| Less healthy | 1025 (13.4) | 575 (13.4) | 450 (13.3) | |
| Marital Status, | <0.001 | |||
| Married | 6330 (81.8) | 3825 (88.9) | 2505 (73.0) | |
| Others | 1406 (18.2) | 478 (11.1) | 928 (27.0) | |
| Smoke, | <0.001 | |||
| Never smoke | 6463 (93.2) | 3707 (94.5) | 2756 (91.6) | |
| Former smoke | 156 (2.3) | 67 (1.7) | 89 (3.0) | |
| Current smoke | 312 (4.5) | 148 (3.8) | 164 (5.5) | |
| Drink, | <0.001 | |||
| Never drink | 5825 (75.2) | 3249 (75.4) | 2576 (75.0) | |
| Former drink | 541 (7.0) | 241 (5.6) | 300 (8.7) | |
| Current drink | 1379 (17.8) | 821 (19.0) | 558 (16.2) | |
| Disease, | <0.001 | |||
| Yes | 5275 (68.4) | 2997 (70.0) | 2278 (66.4) | |
| No | 2440 (31.6) | 1287 (30.0) | 1153 (33.6) | |
| Menarche age | 16.0 (2.1) | 15.7 (2.0) | 16.3 (2.1) | <0.001 |
| Menarche age, | <0.001 | |||
| Early menarche | 3606 (59.6) | 1757 (54.3) | 1849 (65.6) | |
| Late menarche | 2448 (40.4) | 1480 (45.7) | 968 (34.4) | |
| Menopause age | 49.3 (3.7) | 49.4 (3.5) | 49.3 (3.9) | 0.13 |
| Menopause age, | 0.20 | |||
| Early menopause | 2484 (44.6) | 1299 (45.4) | 1185 (43.7) | |
| Late menopause | 3087 (55.4) | 1561 (54.6) | 1526 (56.3) | |
| Reproductive life span | 33.4 (4.1) | 33.6 (3.8) | 33.1 (4.3) | <0.001 |
| Reproductive lifespan, | <0.001 | |||
| Short span | 2284 (48.0) | 1083 (45.5) | 1201 (50.6) | |
| Long span | 2471 (52.0) | 1298 (54.5) | 1173 (49.4) | |
| Children | 2.7 (1.4) | 1.7 (0.5) | 3.9 (1.2) | <0.001 |
| Age at first live birth | 24.6 (4.1) | 24.7 (3.6) | 24.5 (4.7) | 0.068 |
| Age at first live birth, | 0.014 | |||
| Early age at first Birth | 3336 (55.4) | 1846 (54.0) | 1490 (57.2) | |
| Late age at first birth | 2684 (44.6) | 1570 (46.0) | 1114 (42.8) | |
| Cognitive z‐score | −0.2 (1.1) | 0.1 (1.0) | −0.5 (1.0) | <0.001 |
| Cognitive function, | <0.001 | |||
| Low | 4203 (54.3) | 1886 (43.7) | 2317 (67.5) | |
| High | 3543 (45.7) | 2425 (56.3) | 1118 (32.5) |
| Model 0 | Model 1 | Model 2 | ||||
|---|---|---|---|---|---|---|
| OR (95% CI) |
| OR (95% CI) |
| OR (95% CI) |
| |
|
| ||||||
| Number | 7872 | 6934 | 6933 | |||
| Fewer children | 1 (Reference) | NA | 1 (Reference) | NA | 1 (Reference) | NA |
| More children | 0.52 (0.47,0.57) | <0.001 | 0.84 (0.75,0.94) | 0.003 | 0.90 (0.80,1.01) | 0.083 |
|
| ||||||
| Number | 7746 | 6568 | 6568 | |||
| Fewer children | 1 (Reference) | NA | 1 (Reference) | NA | 1 (Reference) | NA |
| More children | 0.38 (0.34,0.41) | <0.001 | 0.70 (0.62,0.79) | <0.001 | 0.87 (0.77,0.99) | 0.034 |
| Number | 6478 | 5186 | 5186 | |||
| Early age at first live birth753 | 1 (Reference) | NA | 1 (Reference) | NA | 1 (Reference) | NA |
| Late age at first live birth | 1.14 (1.03,1.25) | 0.011 | 1.12 (0.99,1.26) | 0.069 | 1.01 (0.89,1.15) | 0.873 |
- —STI2030‐Major Projects
- —National Natural Science Foundation of China10.13039/501100001809
- —Hunan Provincial Natural Science Outstanding Youth Foundation
- —Science and Technology Innovation Program of Hunan Province
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsBirth, Development, and Health · Menopause: Health Impacts and Treatments · Maternal Mental Health During Pregnancy and Postpartum
BACKGROUND
1
Dementia is a growing global health crisis, contributing substantially to morbidity, mortality, and health care costs. According to the World Health Statistics 2024, Alzheimer disease and other dementias rank as the seventh leading cause of death from noncommunicable diseases worldwide and are among the top five in the Americas, Europe, and the Western Pacific Region.1 In the United States, 11.3% of adults 65 years of age or older are affected by Alzheimer disease—73% of whom are 75 years of age or older—and this number is projected to double by 2060.2 The economic burden is equally alarming, as per capita Medicare expenditures for individuals with dementia are nearly three times higher than for those without, and Medicaid spending is over 22 times greater.3
In China, where population aging is accelerating rapidly, dementia poses a major public health concern. An estimated 6.0% of individuals 60 years of age or older have dementia, whereas 15.5% have mild cognitive impairment (MCI).4 Among G20 nations, China bears the highest burden of dementia, with a 322% increase in cases and a 273% rise in related health losses over the past three decades, driven primarily by demographic aging.5 By the end of 2023, ≈297 million people in China were 60 years or older, comprising 21.1% of the total population.6 These trends underscore the need for targeted strategies to delay or prevent cognitive decline.
Women bear a disproportionate burden of dementia globally. In the United States, nearly two‐thirds of patients with Alzheimer's disease are women.2 Similarly, in China, dementia affects 7.0% of women versus 5.0% of men; MCI rates are 17.9% versus 13.0%.4 These differences persist after adjusting for age, suggesting that factors beyond longevity may contribute.5 Emerging evidence highlights the role of reproductive factors in shaping women's cognitive aging. These factors, encompassing both biological and social reproductive dimensions, may jointly influence women's long‐term cognitive health.7 Biologically, estrogen is believed to exert neuroprotective effects. Epidemiological studies have linked longer reproductive span,8 later age at menopause,8, 9, 10 and prolonged exposure to endogenous estradiol11, 12 with reduced risk of cognitive decline and Alzheimer's disease. Conversely, accelerated reproductive aging has been linked to adverse cognitive outcomes.13 A recent cohort study in China reported that extreme menopausal age—whether premature, early, or late—was associated with poorer cognitive outcomes in later life.14 The social dimension of reproduction, commonly measured as number of children, has also been linked to cognitive aging. A higher number of children has been associated with poorer cognitive performance across multiple domains.15, 16 Moreover, recent research among African Caribbean women has indicated that older age at first live birth was significantly associated with better late‐life cognitive function, suggesting that reproductive timing may influence long‐term cognitive health.17
Despite an increasing interest in the intersection of reproductive history and cognitive health, few studies have examined both biological and social reproductive factors in tandem. Moreover, most existing evidence comes from high‐income countries. Few population‐based studies have explored these associations in non–high‐income settings such as China, where social norms around fertility and structural disparities in education and health care may shape these relationships in distinct ways. To address this critical gap, we analyzed longitudinal data from the China Health and Retirement Longitudinal Study (CHARLS) to examine how both biological indicators of ovarian aging (i.e., reproductive span, age at menopause) and social reproductive factors (i.e., number of children) are associated with cognitive function in later life. Drawing on a nationally representative sample of midlife and older adults, we tested the following hypotheses: (1) a longer reproductive span is positively associated with late‐life cognitive performance in women; (2) a higher number of children is negatively associated with late‐life cognition in both sexes, with stronger effects in women. By integrating both biological and social reproductive exposures, this study aims to clarify sex‐specific mechanisms underlying cognitive aging and to inform more equitable and tailored dementia prevention strategies.
METHODS
2
Study population
2.1
RESEARCH IN CONTEXT
- Systematic review: We reviewed the literature using traditional search engines (e.g., PubMed and Google Scholar). Few studies have integrated both the biological and social aspects of ovarian function in relation to cognitive aging in women. Most evidence comes from Western populations, mainly from North America and Europe, with limited data from nationally representative Asian cohorts.
- Interpretation: Using data from more than 8000 Chinese women, this study found that that longer reproductive span was associated with better late‐life cognitive function. A higher number of children was linked to poorer cognition in both sexes, with stronger effects in women. These associations remained robust even after full adjustment, relative to age at menarche, age at menopause, and age at first birth.
- Future directions: Future studies should explore modifiable psychosocial and lifestyle mediators of cognitive aging in diverse populations. Our findings highlight the relevance of reproductive factors in dementia risk assessment and suggest potential intervention targets for women with high parity.
We used data from the China Health and Retirement Longitudinal Study (or CHARLS), a nationally representative longitudinal cohort study of community‐dwelling individuals ≥45 years of age in China. CHARLS collects detailed information on demographic characteristics, health status, economic conditions, and social factors.18 For this study, we included data from Wave 1 (2011) through Wave 4 (2018).19 Eligible participants were ≥45 years of age, had completed at least one wave of cognitive assessments, and had available data on number of children or reproductive lifespan. A total of 16,449 individuals met these criteria.
Among the eligible participants, 8577 women and 7872 men were included in the analysis of the association between number of children and cognitive function. For the analysis of reproductive lifespan and postmenopausal cognitive function, 5586 postmenopausal women were included in the study (Figure 1). Supplementary analyses involved 6879 women for age at menarche and 6396 women for age at natural menopause. Although a small number of participants (n = 117) reported a history of ovarian, cervical, or endometrial cancer, they were retained in the sample because the limited case numbers and missing treatment details precluded meaningful assessment of their impact on reproductive factors.
Study flow chart.
Measures
2.2
Outcome: Cognitive function
2.2.1
Cognitive function was assessed using a composite score derived from neuropsychological tasks adapted from the Health and Retirement Study (HRS),19 covering two domains: episodic memory and executive function. Executive function was assessed by tasks measuring orientation (date naming), calculation (serial sevens), and visuospatial ability (figure drawing), with scores ranging from 0 to 11. Episodic memory was evaluated via immediate and delayed word recall (range: 0–10). A global cognition score was computed by summing the two domain scores (range: 0–21), with higher scores indicating better performance. To facilitate interpretation, a global cognition score was standardized into z‐scores using baseline means and standard deviations, with positive values indicating above‐average cognition and negative values indicating below‐average performance. For further analysis, these z‐scores were dichotomized at zero, with scores ≥0 classified as the high cognitive group and scores <0 as the low cognitive group.
Exposure
2.2.2
Biological reproductive factors
2.2.2.1
The primary exposure was reproductive lifespan, defined as the interval between age at menarche and age at natural menopause. Age at menarche and age at natural menopause were obtained via self‐report. When direct age values were unavailable, age was derived from the reported year of menarche or menopause minus the participant's birth year. All implausible values were identified and excluded based on the interquartile range (IQR) method: values below Q1 − 1.5 × IQR or above Q3 + 1.5 × IQR. For further analysis, reproductive lifespan was dichotomized at the sample median, with values ≥ median classified as long and values < median classified as short.
Social reproductive factors
2.2.2.2
Number of biological children was used to represent the social reproductive dimension. Using data from the 2014 CHARLS Life History Survey, participants were categorized into fewer children (1–2 children) and more children (>2 children) groups based on the sample median. Both women and men were included in the analysis to examine sex‐specific associations between number of children and cognitive function. Age at first birth was calculated as the difference between the participant's birth year and the birth year of the first child.
Covariates
2.2.3
Potential confounders were adjusted for by including childhood cognitive proxy variables, as well as sociodemographic, lifestyle, and health‐related characteristics. Due to the lack of direct measures of childhood cognition in the cohort, childhood cognitive proxies included mother's education, father's education, childhood family economic status, childhood family safety, and childhood physical health, all obtained from the 2014 CHARLS Life History Survey. Other covariates were derived primarily from the most recent wave (2018) and supplemented sequentially by 2015, 2013, and 2011, if missing. Educational level was categorized into four groups—preprimary, primary, lower secondary, and upper secondary and above—based on the International Standard Classification of Education (ISCED) 201120 and CHARLS coding. Other covariates included age, sex (male/female), residence type (urban, integration zone, rural, or special zone), marital status (married vs others), smoking status (never, former, or current smoker), and drinking status (never, former, or current drinker). Chronic disease status was also considered, focusing on conditions relevant to women's reproductive health and cognitive function, including hypertension, dyslipidemia, diabetes, heart disease, stroke, psychiatric problems, and memory‐related diseases.21, 22, 23, 24 It is important to note that age, which is closely related to late‐life cognitive performance, was taken from the same survey wave as the cognitive assessment to ensure temporal consistency. All variable details are described in full in the Supplementary Methods.
Statistical analysis
2.2.4
All analyses were based on cross‐sectional data from the 2018 wave of CHARLS, with information from earlier waves (2011, 2013, and 2015) used only to supplement missing variables. Descriptive statistics were used to characterize the study population. Continuous variables were presented as means and SDs, whereas categorical variables were summarized as counts and percentages. All variables were examined by stratifying participants according to reproductive lifespan (short vs long) and number of children (fewer vs more). Group differences in sociodemographic, lifestyle, and health‐related variables were assessed using independent two‐sample t‐tests for continuous variables and chi‐square tests for categorical variables.
Univariate and multivariable logistic regression models were applied to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the associations of reproductive lifespan and number of children with cognitive function. Potential confounders were selected based on biological plausibility and previous literature linking these factors with cognitive outcomes. A series of multivariable logistic regression models were constructed using a stepwise adjustment approach. Model 1 adjusted for age and childhood cognitive proxies, including residential location, mother's education, father's education, childhood family economic status, childhood family safety, and childhood physical health. Model 2 additionally included adult education, whereas Model 3 further accounted for marital status, drinking status, smoking status, and physical diseases. Finally, in Model 4, we considered the potential collinearity and overlapping explanatory value between childhood cognitive proxies and adult education, and therefore excluded childhood proxies while retaining adult education to better evaluate its independent contribution to cognitive function.
All analyses were performed using Stata version 18.0 (StataCorp LLC, College Station, TX), with two‐tailed p values < 0.05 considered statistically significant.
RESULTS
3
Reproductive lifespan and cognition
3.1
Characteristics of the study population
3.1.1
Table 1 presents the characteristics of 5586 postmenopausal women, stratified by reproductive lifespan based on the median split. Of these, 2653 (47.5%) had a short reproductive lifespan and 2933 (52.5%) had a long lifespan. The mean age ± SD was 62.4 ± 8.5 years. Women with longer reproductive lifespans were more likely to exhibit high cognitive function (48.8% vs 40.5%) and to have fewer children (1–2 children: 52.5% vs 47.4%). They also reported later menarche (54.7%) and menopause (85.9%), whereas earlier reproductive milestones were more common in the short‐span group (menarche: 77.7%; menopause: 75.8%). In addition, women in the long reproductive span group had higher educational attainment (lower secondary: 19.5% vs 15.7%; upper secondary or above: 13.5% vs 7.8%), were more likely to report urban residence, current marriage, and current alcohol use.
Association between reproductive lifespan and cognition
3.1.2
Table 2 presents the associations of reproductive lifespan, age at menarche, and age at menopause with cognitive function among postmenopausal women. A longer reproductive lifespan was significantly associated with higher odds of high cognitive function compared with a shorter span (OR = 1.40, 95% CI: 1.26–1.56). This association remained robust after further adjustment for childhood cognitive proxies (OR = 1.31, 95% CI: 1.15–1.50) and adult education (OR = 1.20, 95% CI: 1.03–1.38), and the trend persisted in other adjusted models, including Model 3 (OR = 1.15, 95% CI: 0.99–1.34) and Model 4 (OR = 1.20, 95% CI: 1.05–1.37) (see Table S1). Consistent results were observed in linear regression analyses, further supporting the positive association between reproductive lifespan and late‐life cognition in women (Table S2). In contrast, the associations of later menarche and later menopause with cognitive function appeared less consistent. Later menarche was significantly associated with a higher likelihood of better cognitive performance only in the unadjusted model (OR = 1.56, 95% CI: 1.41–1.72) and Model 1 (OR = 1.29, 95% CI: 1.14–1.46), whereas later menopause was significantly associated with better cognition only in Model 1 (OR = 1.14, 95% CI: 1.00–1.29). Overall, these findings suggest that reproductive lifespan may serve as a more reliable indicator of late‐life cognitive health than the individual timing of menarche or menopause.
Number of children and cognition
3.2
Characteristics of the study population
3.2.1
Table 3 presents the characteristics of all female participants, stratified by number of children. The mean age of the sample was 60.2 ± 9.2 years. Of the 7746 women 45 years of age or older, 4311 (55.7%) reported having one or two biological children, and 3435 (44.3%) reported having three or more. Women with more children exhibited significantly lower cognitive function than those with fewer children: 2317 (67.5%) in the more children group were classified as having low cognition, compared to 1886 (43.7%) in the fewer children group. Furthermore, women with more children had notably lower educational attainment (lower secondary: 10.7% vs 25.6%; upper secondary or above: 3.7% vs 14.2%). Among postmenopausal women, those with more children were also more likely to have a shorter reproductive lifespan (50.6% vs 45.5%) and earlier menarche (65.6% vs 54.3%). In addition, women with more children tended to be born earlier, had parents with lower levels of education, and reported slightly poorer childhood family economic conditions and greater childhood family insecurity. They were also less likely to be currently married, more likely to reside in rural areas, and exhibited somewhat higher rates of smoking and alcohol use compared with women with fewer children.
Association between number of children and cognitive function
3.2.2
Table 4 summarizes the associations between number of children and cognitive function among middle‐aged and older women and men. In both sexes, having more children was significantly associated with a lower odds of high cognitive function compared to having fewer children. This association was more pronounced in women (OR = 0.38, 95% CI: 0.34–0.41) than in men (OR = 0.52, 95% CI: 0.47–0.57). This pattern persisted after adjustment for childhood cognitive proxies, remaining more pronounced in women. After further adjustment for adult education (Model 2), the association remained significant in women (OR = 0.87, 95% CI: 0.77–0.99), whereas in men the association was attenuated and no longer reached statistical significance (OR = 0.90, 95% CI: 0.80–1.01). The negative association between number of children and late‐life cognition was generally robust in women across other two adjusted models (Model 3: OR = 0.86, 95% CI: 0.76–0.99; Model 4: OR = 0.79, 95% CI: 0.70–0.89; see Table S3), and was further supported by multiple adjusted linear regression analyses (Table S2). Stratified analyses by educational level indicated that this negative association persisted across all educational strata, with the effect being more pronounced among women with higher education (OR decreased from 0.64 in preprimary to 0.47 in upper secondary or above; see Table S4). Regarding covariates, higher educational attainment was consistently associated with an increased odds of high cognitive function in both women and men. Among women, several additional factors—including older age, rural residence, unmarried status, poorer childhood family safety, and poorer childhood physical health—were significantly associated with lower cognitive function in Model 3 (see Table S5). In addition, later age at first live birth was associated with a higher odds of high cognitive function in the unadjusted model (OR = 1.14, 95% CI: 1.03–1.25); however, this association became non‐significant after sequential adjustment for covariates.
DISCUSSION
4
This study is the first to concurrently examine the associations between reproductive factors and cognitive performance in middle‐aged and older women from both physiological and sociocultural perspectives. Drawing on nationally representative longitudinal data from China, we found that a longer reproductive lifespan was positively associated with better cognitive outcomes in postmenopausal women. In contrast, a higher number of biological children was associated with an elevated risk of cognitive impairment, with this adverse association significantly more pronounced in women than in men. These findings provide novel evidence on the cognitive implications of women's reproductive history and underscore the importance of integrating biological and social reproductive factors into dementia risk assessment and prevention strategies.
Accumulating evidence from large‐scale cohorts and meta‐analyses supports a robust association between female reproductive factors and later‐life cognitive health. Reproductive lifespan—an indicator of cumulative estrogen exposure—has been linked consistently to better cognitive outcomes. Studies from the United States and Singapore report that shorter reproductive spans are associated with an increased risk of cognitive impairment.25, 26 Meta‐analyses involving millions of women confirm that shorter reproductive span elevates the risk of dementia and cognitive decline.13, 27 Later menopause, reflecting a longer reproductive lifespan, is consistently associated with better cognition, whereas earlier menopause increases dementia risk across North America,25 Europe,28 and Asia.29 These associations are further supported by pooled analyses from multinational prospective studies.13, 27, 30 Evidence regarding age at menarche and later‐life cognition is more mixed. Some studies suggest that later menarche increases dementia risk,13, 25, 28, 31 whereas others—including findings from the UK Biobank—report the opposite.32 A dose–response meta‐analysis indicated a J‐shaped association,27 and several studies found no significant link.33, 34 Confounding by early‐life adversity and lower socioeconomic status—both associated with earlier menarche—may obscure underlying biological effects. Earlier menarche is also linked to younger maternal age at first birth, which has been associated with poorer cognitive outcomes in later life. Estrogen appears neuroprotective across both direct and indirect measures. Higher postmenopausal estradiol levels are inversely associated with Alzheimer's risk,27 and hormonal contraceptive use15, 26, 28, 31, 32 is linked to slower cognitive decline. However, the cognitive effects of hormone therapy (HT) remain debated, with benefits potentially dependent on formulation, dosage, and especially timing of initiation, ideally within 5 years of menopause.15, 26, 28, 31, 32 Neuroimaging studies further demonstrate that premature ovarian insufficiency is linked to structural and functional brain alterations, such as reduced gray matter volume,35 increased white matter hyperintensities,35 impaired synaptic integrity,36 and accelerated prefrontal cortical aging.37 Our findings from a Chinese cohort confirm that a longer reproductive lifespan is associated with better late‐life cognition in women, supporting a neuroprotective role of prolonged estrogen exposure. Although later menopause and later menarche were also linked to better cognition in some covariate‐adjusted models, these associations appear less robust compared with reproductive lifespan.
Much research has focused on the biological aspects of ovarian aging; however, this study also underscores the cognitive relevance of socially acquired reproductive experiences, particularly parity. Across diverse populations, having a higher number of children has been associated consistently with poorer cognitive outcomes among women. Studies from the United States,15 Latin America,28 Singapore,26 and Korea,31 as well as pooled data from 11 international cohorts,38 report elevated risks of cognitive decline or dementia with an increasing number of births. Notably, evidence from the UK Biobank shows that this association is sex specific: while fatherhood is linked to better cognitive performance in men, motherhood is associated with worse outcomes in memory and fluid intelligence.39 However, data from low‐ and middle‐income countries remain scarce. Using nationally representative CHARLS data, our study expands this literature by including both men and women ≥45 years of age. We found that having three or more children was significantly associated with lower cognitive performance, and that this negative effect was more pronounced in women than in men, independent of sociodemographic and health factors. These findings underscore the need to consider gender‐specific social exposures in cognitive aging research. It is important to note that having children versus being childless, and, among parents, the number of children, reflect conceptually distinct research questions. Since our study focuses on the effect of the number of children rather than the presence versus absence of children, participants without children (≈5%) were excluded from analyses of the number of children. In addition, although China's one‐child policy (1980–2015) may have influenced fertility patterns, most women in this cohort had completed childbearing before strict enforcement, and only 18.1% had a single child, suggesting a limited impact on the observed associations between number of children and late‐life cognitive function.
Several pathways may mediate the inverse association between number of children and late‐life cognition, spanning psychosocial, physiological, and sociocultural domains.40 Psychosocially, childbearing can constrain women's educational and occupational trajectories, as fertility typically declines after the mid‐30s, whereas most first births occur earlier,41 forcing trade‐offs between family and career. This contributes to lower fertility intentions among highly educated women globally.42 Motherhood is also associated with depressive symptoms and sleep disruption—the former encompassing postpartum depression,43, 44 disproportionate burdens of unpaid domestic labor,45, 46 and the psychological stress of childrearing,47 and the latter including pregnancy‐related and parenting‐related sleep deprivation.48, 49 These two domains reinforce one another, and both are established risk factors for cognitive decline.50, 51 Physiological mechanisms may also contribute: adverse pregnancy outcomes, such as recurrent stillbirths,52, 53 miscarriages,52, 53 hypertensive disorders of pregnancy,54, 55, 56, 57, 58 and gestational diabetes59 have each been linked to poorer late‐life cognitive outcomes and increased dementia risk, with neuroimaging evidence showing brain changes resembling those in Alzheimer's disease.60 Finally, sociocultural influences are particularly salient in East Asia, where traditional norms around marriage and childbearing may limit women's educational and occupational opportunities, and workplace barriers and postpartum constraints hinder reentry into the labor force. These structural and cultural pressures can reduce cognitive stimulation, foster social isolation, and shape reproductive choices. For example, despite extensive policy adjustments, fertility rates in Korea continue to decline, reflecting the complex interplay of gender norms, employment constraints, and reproductive decisions.61
Strengths and limitations
4.1
A key strength of this study is that it represents the first population‐based investigation in China to simultaneously examine both biological and sociocultural dimensions of ovarian aging and their associations with cognitive function in middle‐aged and older women. By integrating physiological indicators with reproductive experiences, our findings provide a more nuanced understanding of the long‐term cognitive implications of women's reproductive history. Notably, this study is the first to demonstrate that both reproductive lifespan and number of children are consistently associated with late‐life cognition in women, whereas other reproductive factors showed less‐robust associations. These insights underscore the importance of supporting women in navigating reproductive decisions that balance both immediate and later‐life health outcomes. Future research should explore interventions to mitigate the cognitive burden of childbearing, ranging from biological strategies to extend reproductive health to social policies aimed at reducing caregiving demands and emotional strain disproportionately borne by women. Such efforts may promote more equitable reproductive choices and support healthy cognitive aging.
This study has several limitations. First, the absence of detailed data on ovarian or uterine surgeries precluded differentiation between natural and surgical menopause. The recorded number of participants reporting a history of ovarian, cervical, or endometrial cancer (n = 117) was likely underestimated due to stigma‐related underreporting in older adults. Future studies should improve data completeness in this area. Second, we lacked information on exogenous hormone use, which limited our ability to identify women who may have extended reproductive span through hormone therapy; however, its impact is likely minimal given the low use among Chinese women. Third, although variables such as education, depressive symptoms, and sleep disturbances could theoretically mediate the relationship between childbearing and late‐life cognition, limitations in temporal data precluded accurate mediation analysis. Specifically, education could not be examined as a mediator because most women discontinued schooling after childbearing (only 126 of 6229 women continued schooling after their first child), whereas depressive symptoms and sleep disturbances were measured only a few years (up to ≈7 years) prior to the 2018 cognitive evaluation. Fourth, due to the absence of standardized definitions for reproductive age thresholds, extreme values for reproductive lifespan were excluded, and reproductive lifespan was dichotomized at the median to capture overall trends. Future research should examine cognitive outcomes among women with more extreme reproductive histories and explore standardized cutoff values. Finally, an additional wave of CHARLS data from 2020, collected during the coronavirus disease 2019 (COVID‐19) pandemic, was released in late 2024. Given the pronounced rise in depression and anxiety during this period—factors closely linked to cognitive outcomes—we deemed it inappropriate to combine these data with our pre‐pandemic analyses. As we enter a period of sustained coexistence with COVID‐19, future studies will incorporate these data to refine and extend our findings.
CONCLUSION
5
Longer reproductive span was associated with better cognitive function in postmenopausal women, whereas a higher number of children was linked to poorer cognition in both sexes, particularly in women. These associations were robust and stronger than those for other reproductive factors. Together, they highlight the dual biological and social dimensions of reproductive history in cognitive aging and suggest that number of children may serve as a sex‐specific consideration in dementia prevention strategies.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest. Any author disclosures are available in the Supporting Information.
CONSENT STATEMENT
All participants provided written informed consent.
Supporting information
Supporting Information
Supporting Information
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1WHO . World health statistics 2024: monitoring health for the SD Gs, Sustainable Development Goals. World Health Organization; License: CC BY‐NC‐SA 3.0 IGO. 2024;
- 2Rajan KB , Weuve J , Barnes LL , Mc Aninch EA , Wilson RS , Evans DA . Population estimate of people with clinical Alzheimer's disease and mild cognitive impairment in the United States (2020‐2060). Alzheimers Dement. 2021;17(12):1966‐1975. doi:10.1002/alz.12362 34043283 PMC 9013315 · doi ↗ · pubmed ↗
- 3Alzheimer’s Association . 2024 Alzheimer's disease facts and figures. Alzheimers Dement. 2024;20(5):3708‐3821. doi:10.1002/alz.13809 38689398 PMC 11095490 · doi ↗ · pubmed ↗
- 4Jia L , Du Y , Chu L , et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross‐sectional study. Lancet Public Health. 2020;5(12):e 661‐e 671. doi:10.1016/s 2468-2667(20)30185-7 33271079 · doi ↗ · pubmed ↗
- 5Yang K , Yang X , Yin P , Zhou M , Tang Y . Temporal trend and attributable risk factors of Alzheimer's disease and other dementias burden in China: findings from the global burden of disease study 2021. Alzheimers Dement. 2024;20(11):7871‐7884. doi:10.1002/alz.14254 39312279 PMC 11567818 · doi ↗ · pubmed ↗
- 6Ministry of Civil Affairs of the People's Republic of China . Statistical Bulletin on the Development of Undertakings for the Aged in China (2023). 2024. Retrieved from http://www.mca.gov.cn/
- 7Lopez‐Lee C , Torres ERS , Carling G , Gan L . Mechanisms of sex differences in Alzheimer's disease. Neuron. 2024;112(8):1208‐1221. doi:10.1016/j.neuron.2024.01.024 38402606 PMC 11076015 · doi ↗ · pubmed ↗
- 8Wedatilake Y , Myrstad C , Tom SE , Strand BH , Bergh S , Selbæk G . Female reproductive factors and risk of mild cognitive impairment and dementia: the HUNT study. J Prev Alzheimers Dis. 2024;11(4):1063‐1072. doi:10.14283/jpad.2024.46 39044518 PMC 11937202 · doi ↗ · pubmed ↗