Parental age and its influence on functional dependency and cognitive decline in older adults: a cohort analysis Authors: Zongbao Li, Fan Hu, Shaojie Li, Yao Yao, Yanhui Lu Published: 2026 ABSTRACT ObjectiveTo explore the influence of parental age at childbirth on the long-term health of offspring.DesignProspective cohort study.SubjectsThe baseline data of the Chinese Health and Longevity Longitudinal Survey (CLHLS, 2008–2021) in 2008 and the subsequent follow-up data.ExposureParental age at childbirth.Main outcome measuresFunctional independency (assessed by Activities of daily living, ADL), cognitive function (assessed using Mini-Mental State Examination, MMSE), and all-cause mortality.ResultsOffspring with advanced maternal age (≥ 37 years) were associated with higher odds of ADL limitations (OR = 1.39, 95%CI = 1.04–1.84) and cognitive dysfunction (OR = 1.34, 95%CI = 1.02–1.76). Conversely, advanced paternal age (≥ 43 years) demonstrated negative associations with ADL limitation (OR = 0.52, 95%CI = 0.36–0.73) and cognitive dysfunction (OR = 0.60, 95%CI = 0.43–0.83). While no direct association emerged between parental age and offspring mortality, we identified a significant trend toward reduced mortality risk with increasing paternal age, particularly among offspring of non-educated parents (P for trend = 0.02). Notably, the presence of sibling caregiving was associated with attenuated maternal age-related health risks, indicating a potential moderating role of family support structures.ConclusionOur study identifies observational associations between parental age and offspring health outcomes in later life. Specifically, older maternal age associated with poor health outcomes while older paternal age was associated with better health outcomes in offspring. As societies grapple with aging populations and declining fertility, our study suggests the need for integrated health and social policies that support families in this changing demographic landscape.Supplementary InformationThe online version contains supplementary material available at 10.1186/s41043-026-01308-8. Introduction The paradigm of reproductive decision-making in contemporary society has undergone a significant transformation. This shift, driven by evolving fertility concepts, expanded access to higher education, and increasing socioeconomic pressures, has resulted in a substantial rise in delayed childbearing or voluntary childlessness among young couples in both developed and developing nations. Statistical evidence from the United States illustrates this trend, with the mean maternal age at first birth escalating from 21.4 years in 1970 to 26.3 years in 2014 [1], accompanied by a corresponding increase in paternal age from 27.4 to 30.9 years during the same period [2]. Parallel demographic trends are observed in China, where the postponement of marriage and childbearing has contributed to a marked decline in fertility rates [3, 4]. Of particular scientific concern are the genetic and epigenetic modifications associated with advanced parental age, which may accumulate and be transmitted across generations, potentially compromising offspring long-term health outcomes and longevity. In medicine, women aged 35 and above at the time of delivery are typically defined as being of “advanced maternal age” [5]. For males, although there is no clear medical consensus, studies often define advanced paternal age as 40 or 45 years and older [6, 7]. In real-world research, appropriate adjustments can be made based on one’s own research data, so there remains some controversy regarding what constitutes advanced parental age [8]. The association between parental age and offspring health outcomes has garnered substantial attention across multiple disciplines. Subtle changes in the molecular field of the human body can have significant impacts on function [9]. While extensive research has predominantly focused on the adverse effects of advanced parental age on pregnancy outcomes [7, 10–12] and pediatric health conditions [13–15], relatively few studies have focused on offspring long-term health. Emerging evidence suggests that advanced parental age may be associated with increased risks of various chronic conditions in offspring, including physical morbidity [16–18], psychiatric disorders [19, 20], and cognitive dysfunction [21, 22]. Monaghan et al. [23] systematically reviewed the inverse relationship between parental age and offspring longevity, a phenomenon known as the Lansing Effect. However, it is crucial to acknowledge that current evidence primarily derives from controlled laboratory experiments and animal observations, lacking robust population-based real-world studies. To address the gap where animal experiments cannot replicate the interplay between parental age and the complex socioeconomic, environmental, and lifestyle factors that shape health outcomes in real-world human settings, this study leverages data from the Chinese Longitudinal Healthy Longevity Survey (CLHLS) to examine the association between parental age and long-term offspring health. Grounded in the multidimensional framework of “healthy aging” - encompassing physical health, mental well-being, and social adaptability [24] - we employed three primary health indicators: chronic disease comorbidities, activities of daily living (ADL), and cognitive function. Furthermore, we analyzed the relationship between parental reproductive age and offspring all-cause mortality using 10-year longitudinal follow-up data. This investigation aims to advance our understanding of the evolutionary biological implications of parental age effects while providing empirical evidence to inform public health policies addressing demographic challenges, particularly population aging and declining fertility rates. Materials and methods Study population The CLHLS is a nationally representative longitudinal cohort study, employing a rigorous multi-stage, stratified cluster sampling methodology across 23 of China’s 31 provinces. Data collection was conducted through standardized field surveys, utilizing direct interviews with participants or their proxy respondents (family members) when necessary. The comprehensive survey protocol incorporated face-to-face interviews to collect multidimensional data, including: demographic characteristics (age, sex, ethnicity, and education level); lifestyle factors (smoking status, alcohol consumption, and exercise status); and comprehensive health indicators. This methodological approach ensures both high response rates and data reliability. For detailed information regarding the cohort design, sampling strategy, and data collection procedures, readers are referred to previous methodological publications [25, 26]. A total of 16,954 participants were enrolled in 2008 baseline survey. We applied the following exclusion criteria: (1) missing parental age; (2) biologically implausible parental ages (maternal age < 12 or > 60 years; paternal age < 12 or > 70 years); (3) incomplete data on primary health outcomes (ADL, MMSE, chronic comorbidities) or essential covariates. The final analytical cohort comprised 10,064 participants and the complete participant selection process, including detailed inclusion and exclusion criteria, was illustrated in Fig. 1. We compared the baseline characteristics of the included and excluded populations, which are presented in Supplementary Tables 1, and provided a detailed explanation of the results. Fig. 1Screening criteria for study population. Abbreviations: ADL, Activities of daily living. MMSE, Mini-mental state examination Screening criteria for study population. Abbreviations: ADL, Activities of daily living. MMSE, Mini-mental state examination Parental age was defined as the age of parents at the time of the participant’s birth, calculated using parental death ages and participants’ ages at parental death from the 2008 baseline data. Given the non-normal distribution of parental age data, we employed stratification for analytical purposes. Since standardized classification criteria for parental reproductive age are currently lacking, we developed a categorization scheme based on the typical female reproductive lifespan (13–49 years). Parental age was divided into six intervals spanning six years each: ≤18, 19–24, 25–30, 31–36, 37–42, and ≥ 43 years. To maintain statistical robustness, cases with extreme values (participants born to parents < 13 or > 50 years old) were merged with adjacent age groups. Outcome variables Three variables were used to assess geriatric health status in the offspring: chronic comorbidities, daily living activities, and cognitive function. Chronic comorbidities were ascertained from the 2008 baseline data, which encompassed 22 prevalent age-related non-communicable conditions. To account for biological sex differences, the assessment included gender-specific conditions: prostate disorders for male participants and gynecological conditions (uterine tumors and breast diseases) for female participants. Following established epidemiological criteria, participants presenting with two or more concurrent chronic conditions were classified as having chronic comorbidities. Functional capacity was evaluated using the standardized Activities of Daily Living (ADL) scale, a validated instrument assessing six essential self-care domains: bathing, dressing, toileting, transferring indoors, continence, and eating. Each domain was scored on a 3-point scale reflecting the degree of functional dependence: 1 = independent (no assistance required), 2 = partially dependent (requires some assistance), and 3 = completely dependent (bedridden or requiring full assistance). The cumulative score for ADL ranges from 0 to 18, where a score of 0 indicates complete functional independence, a score of ≥ 1 indicates the presence of functional impairment, and higher scores reflect more severe impairment. This scoring system has been widely validated in geriatric research and demonstrates excellent reliability in assessing functional status among older adults [27]. Cognitive function was assessed using a culturally adapted version of the Mini-Mental State Examination (MMSE), specifically modified for the Chinese population. The assessment comprised five cognitive domains: general ability, reaction ability, attention and calculation power, recall power, and language comprehension coordination power, encompassing 24 items with a maximum total score of 30 points. Scoring followed a dichotomous system, with correct responses awarded 1 point and incorrect or missing responses receiving 0 points. Higher total scores indicated superior cognitive performance. To account for educational attainment effects, we implemented education-stratified cutoff values for cognitive impairment. Specifically, participants were classified as cognitively impaired if their MMSE score fell below the following thresholds: ≤18 for individuals with no formal education (illiterate), ≤ 20 for those with primary education (1–6 years of schooling), and ≤ 24 for those with secondary education or higher (≥ 7 years of schooling). This education-adjusted approach enhances the diagnostic accuracy of cognitive assessment in populations with varying educational backgrounds [28]. For deceased participants, survival time (in months) was determined by calculating the interval between the baseline enrollment date (2008) and the recorded date of death during follow-up assessments (2011, 2014, or 2018). For lost individuals, survival time was computed as the duration between baseline enrollment and the last available follow-up contact. Participants who remained alive throughout the study period were right-censored, with their survival time calculated as the period between baseline enrollment and the final follow-up assessment in 2018. This methodological approach ensured accurate survival time estimation while appropriately handling different censoring scenarios in the longitudinal analysis. Other covariates Socio-demographic variables included age, sex (male/female), ethnicity (Han nationality, other ethnic groups), residence (city or town, rural), body mass index (classification criteria for Chinese population: <18.5, 18.5–24.9, 25.0–27.9.0.9, and > 28.0 kg/m2) [29], education level (illiteracy, primary school, junior high school, senior high school, college+), major occupation before age 60 (Manual workers: agriculture, forestry, fisheries, animal husbandry, staff, service worker, and industrial worker; Non-manual workers: professional personnel, technical personnel, doctors, teachers, governmental personnel, institutional personnel, managerial personnel, self-employer, housewife, and unemployed), and family income (divided into four groups according to the interquartile spacing, the values at the nodes are classified into the higher level group) of participants at baseline in 2008. Meanwhile, we included several lifestyle factors, namely tobacco use (yes/no), alcohol consumption (yes/no), regular physical activity (yes/no), and dietary diversity in our analysis. Dietary diversity was evaluated using a validated questionnaire encompassing nine food groups commonly consumed in the Chinese population. Participants received one point for consuming vegetables and fruits more than once daily, and an additional point each for consuming meat, eggs, fish, dairy products, legumes, nuts, and mushrooms more than once weekly. The cumulative scores from these nine food groups were aggregated and subsequently classified into three distinct categories: low (0–3 points), moderate (4–6 points), and high (7–9 points) dietary diversity. Furthermore, we incorporated covariates related to childhood experiences, specifically access to adequate medical care during illness (“Did you have access to adequate medical services when you were sick during your childhood?“) and frequency of hunger before bedtime (“Did you frequently experience hunger before bedtime as a child?“). Additionally, childhood caregiving status provided by elder siblings, was also considered as a covariate in our analysis. Statistical analysis Baseline characteristics were stratified by paternal and maternal age groups separately. Normally distributed continuous variables are expressed as mean ± standard deviation (SD), while categorical variables are presented as frequencies with percentages (%). Intergroup comparisons employed ANOVA for normally distributed continuous variables, Kruskal-Wallis tests for non-parametric distributions and chi-square tests for categorical variables. Binary logistic regression models examined associations between parental age and three primary outcome variables. Cox proportional hazards models were implemented to assess parental age effects on offspring all-cause mortality. A further multi-factor adjustment was made for all of the above analytical methods, including sociodemographic variables (age, gender, race, residence, BMI, education level, past occupation, and family income), lifestyle variables (tobacco intake, alcohol intake, regular exercise, and dietary diversity), and childhood experience variables (timely treatment in childhood and childhood experiences of food insecurity). We conducted a series of sensitivity analyses to evaluate the robustness of our findings: (1) Exclusion of participants with educated parents (n = 1,994) to minimize residual confounding from intergenerational educational attainment; (2) Sex-stratified analyses to identify potential dimorphic patterns in parental age effects on offspring’s long-term health outcomes; (3) Stratified analyses by birth cohort were conducted to address confounding effects of age, period, and cohort dynamics in intergenerational transmission. (4) Birth order-adjusted subgroup analyses examining parity-specific associations with offspring’s geriatric health status; (5) Sibling structure stratification (presence/absence of older siblings) to account for caregiving role substitution within multigenerational households. Statistical significance was determined using a two-tailed P-value threshold of < 0.05. All statistical analyses were conducted using R version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). Results Basic characteristics of study samples The cohort was stratified by maternal and paternal age categories (Table 1; full baseline characteristics in Supplementary Tables 1–2). Participants (mean age 86.16 ± 13.23 years) comprised 55.0% females, 93.4% ethnic Han, and 40.7% urban residents. Key health indicators revealed 26.6% chronic comorbidity prevalence, 19.0% activity of daily living (ADL) limitations, and 23.4% cognitive impairment. Notably, 14.4% of participants were classified as overweight or obese according to Asian-specific BMI thresholds. Table 1Inter-group comparison of baseline characteristics based on different paternal age/maternal age groupsVariablesGroupsMean or proportionInter-group differenceMaternal age (P-value)Paternal age (P-value)Continuous Age, years-86.16 (13.23)< 0.001*< 0.001* Grouped Age < 7070–7980–89≥ 901326 (13.18%)2026 (20.13%)2641 (26.24%)4071 (40.45%)< 0.001*< 0.001* Gender malefemale4525 (45.0%)5539 (55.0%)< 0.001*< 0.001* Race han nationalityOther ethnic groups9401 (93.4%)663 (6.6%)0.2790.375 Residence city or townrural4092 (40.7%)5972 (59.3%)0.5510.095BMI, Kg/m2<18.518.5–23.924.0–27.9.0.9≥ 28.03150 (31.3%)5464 (54.3%)1174 (11.7%)276 (2.7%)< 0.001*0.001* Education level illiteracyprimary schooljunior high schoolsenior high schoolcollege +5977 (59.4%)2357 (23.4%)1220 (12.1%)337 (3.3%)173 (1.7%)< 0.001*< 0.001* Occupation manualnon-manual7996 (79.5%)2068 (20.5%)0.4070.358a Income, yuanQ1 (<4000)Q2 (4000–9999)Q3 (10000–21525)Q4 (> 21525)2557 (25.4%)2714 (27.0%)2277 (22.6%)2516 (25.0%)0.003*0.078b Tobacco intakeyesno3238 (32.2%)6826 (67.8%)< 0.001*< 0.001* Alcohol intake yesno2860 (28.4%)7204 (71.6%)0.0710.016* Exercise yesno3020 (30.0%)7044 (70.0%)0.3530.002* Dietary diversity Q1 (0–2 points)Q2 (3 points)Q3 (4–5 points)Q4 (6-10points)3808 (37.8%)1845 (18.3%)2455 (24.4%)1956 (19.4%)0.137< 0.001* Timely treatment in childhood yesno3546 (35.2%)6518 (64.8%)0.0830.712 childhood experiences of food insecurity yesno7369 (73.2%)2695 (26.8%)0.6650.363 Elder siblings havehave not6355 (63.1%)3709 (36.9%)< 0.001*< 0.001* Chronic Comorbidities yesno2680 (26.6%)7384 (73.4%)0.1830.065 Limitations on ADL yesno1908 (19.0%)8156 (81.0%)< 0.001*< 0.001* Cognitive Dysfunction yesno2352 (23.4%)7712 (76.6%)< 0.001*< 0.001* Survival time -60.32 (46.55)< 0.001*< 0.001** P < 0.05, P-values were two-tailed and considered significant at P < 0.05a Income is calculated as household income in RMB and grouped into quartilesb Tobacco and alcohol consumption, as well as exercise, were recorded based on past experiences rather than current ones Table 1Inter-group comparison of baseline characteristics based on different paternal age/maternal age groupsVariablesGroupsMean or proportionInter-group differenceMaternal age (P-value)Paternal age (P-value)Continuous Age, years-86.16 (13.23)< 0.001*< 0.001* Grouped Age < 7070–7980–89≥ 901326 (13.18%)2026 (20.13%)2641 (26.24%)4071 (40.45%)< 0.001*< 0.001* Gender malefemale4525 (45.0%)5539 (55.0%)< 0.001*< 0.001* Race han nationalityOther ethnic groups9401 (93.4%)663 (6.6%)0.2790.375 Residence city or townrural4092 (40.7%)5972 (59.3%)0.5510.095BMI, Kg/m2<18.518.5–23.924.0–27.9.0.9≥ 28.03150 (31.3%)5464 (54.3%)1174 (11.7%)276 (2.7%)< 0.001*0.001* Education level illiteracyprimary schooljunior high schoolsenior high schoolcollege +5977 (59.4%)2357 (23.4%)1220 (12.1%)337 (3.3%)173 (1.7%)< 0.001*< 0.001* Occupation manualnon-manual7996 (79.5%)2068 (20.5%)0.4070.358a Income, yuanQ1 (<4000)Q2 (4000–9999)Q3 (10000–21525)Q4 (> 21525)2557 (25.4%)2714 (27.0%)2277 (22.6%)2516 (25.0%)0.003*0.078b Tobacco intakeyesno3238 (32.2%)6826 (67.8%)< 0.001*< 0.001* Alcohol intake yesno2860 (28.4%)7204 (71.6%)0.0710.016* Exercise yesno3020 (30.0%)7044 (70.0%)0.3530.002* Dietary diversity Q1 (0–2 points)Q2 (3 points)Q3 (4–5 points)Q4 (6-10points)3808 (37.8%)1845 (18.3%)2455 (24.4%)1956 (19.4%)0.137< 0.001* Timely treatment in childhood yesno3546 (35.2%)6518 (64.8%)0.0830.712 childhood experiences of food insecurity yesno7369 (73.2%)2695 (26.8%)0.6650.363 Elder siblings havehave not6355 (63.1%)3709 (36.9%)< 0.001*< 0.001* Chronic Comorbidities yesno2680 (26.6%)7384 (73.4%)0.1830.065 Limitations on ADL yesno1908 (19.0%)8156 (81.0%)< 0.001*< 0.001* Cognitive Dysfunction yesno2352 (23.4%)7712 (76.6%)< 0.001*< 0.001* Survival time -60.32 (46.55)< 0.001*< 0.001** P < 0.05, P-values were two-tailed and considered significant at P < 0.05a Income is calculated as household income in RMB and grouped into quartilesb Tobacco and alcohol consumption, as well as exercise, were recorded based on past experiences rather than current ones Inter-group comparison of baseline characteristics based on different paternal age/maternal age groups < 70 70–79 80–89 ≥ 90 1326 (13.18%) 2026 (20.13%) 2641 (26.24%) 4071 (40.45%) male female 4525 (45.0%) 5539 (55.0%) han nationality Other ethnic groups 9401 (93.4%) 663 (6.6%) city or town rural 4092 (40.7%) 5972 (59.3%) <18.5 18.5–23.9 24.0–27.9.0.9 ≥ 28.0 3150 (31.3%) 5464 (54.3%) 1174 (11.7%) 276 (2.7%) illiteracy primary school junior high school senior high school college + 5977 (59.4%) 2357 (23.4%) 1220 (12.1%) 337 (3.3%) 173 (1.7%) manual non-manual 7996 (79.5%) 2068 (20.5%) Q1 (<4000) Q2 (4000–9999) Q3 (10000–21525) Q4 (> 21525) 2557 (25.4%) 2714 (27.0%) 2277 (22.6%) 2516 (25.0%) yes no 3238 (32.2%) 6826 (67.8%) yes no 2860 (28.4%) 7204 (71.6%) yes no 3020 (30.0%) 7044 (70.0%) Q1 (0–2 points) Q2 (3 points) Q3 (4–5 points) Q4 (6-10points) 3808 (37.8%) 1845 (18.3%) 2455 (24.4%) 1956 (19.4%) yes no 3546 (35.2%) 6518 (64.8%) yes no 7369 (73.2%) 2695 (26.8%) have have not 6355 (63.1%) 3709 (36.9%) yes no 2680 (26.6%) 7384 (73.4%) yes no 1908 (19.0%) 8156 (81.0%) yes no 2352 (23.4%) 7712 (76.6%) * P < 0.05, P-values were two-tailed and considered significant at P < 0.05 a Income is calculated as household income in RMB and grouped into quartiles b Tobacco and alcohol consumption, as well as exercise, were recorded based on past experiences rather than current ones No significant between-group differences emerged across maternal/paternal age strata for sociodemographic covariates including race, residence, occupation, family income, timely treatment in childhood, and childhood experiences of food insecurity (all P > 0.05). In terms of health outcomes, the differences in chronic comorbidities across various parental age groups were not statistically significant. However, significant intergroup differences were observed in mobility, cognitive levels, and survival time (P < 0.05). Associations between parental age and health outcomes Following multivariable adjustment, no significant association was observed between parental age and offspring chronic comorbidity prevalence (Fig. 2). Notably, maternal age demonstrated a positive association with both activity of daily living (ADL) limitations and cognitive dysfunction risk, whereas paternal age showed an inverse association with these functional outcomes (P for trend < 0.05). Fig. 2Associations between maternal age/paternal age and the three elderly health indicators in offspring. Adjusted for age, gender, race, BMI, education level, occupation, family income, tobacco consumption, alcohol consumption, exercise, dietary diversity, timely treatment in childhood, childhood experiences of food insecurity and possible elder siblings care Associations between maternal age/paternal age and the three elderly health indicators in offspring. Adjusted for age, gender, race, BMI, education level, occupation, family income, tobacco consumption, alcohol consumption, exercise, dietary diversity, timely treatment in childhood, childhood experiences of food insecurity and possible elder siblings care Regarding mortality outcomes, adjusted regression models revealed no significant association between parental age and offspring all-cause mortality risk (Table 2). Advanced paternal age showed a non-significant trend toward reduced offspring mortality risk (P for trend = 0.06). Table 2Associations between maternal age/paternal age and offspring all-cause mortality in the general population and among individuals with uneducated parentsSubgroupsOverall Participants(n = 10,064)Participants with uneducated parents(n = 8070)No.†HR (95% CI)P-valueNo.HR (95% CI)P-value Maternal age (years) ≤ 18541/3251.06 (0.94, 1.19)0.30423/2580.90 (0.78, 1.02)0.1119 ~ 244025/2368control-3179/1931control-25 ~ 302920/16291.03 (0.90, 1.17)0.702298/13410.98 (0.91, 1.06)0.6831 ~ 361368/7391.03 (0.89, 1.20)0.601067/6111.03 (0.92, 1.14)0.6237 ~ 42832/4451.06 (0.90, 1.25)0.50632/3481.02 (0.89, 1.18)0.74≥ 43378/2071.16 (0.94, 1.42)0.20309/1691.07 (0.88, 1.30)0.49P for trend--0.62--0.24 Paternal age (years) ≤ 18325/2050.97 (0.84, 1.13)0.70264/1711.06 (0.90, 1.25)0.4719 ~ 243165/1877control-2507/1540control-25 ~ 303293/18740.97 (0.83, 1.13)0.702571/15350.99 (0.92, 1.07)0.7531 ~ 361629/8820.94 (0.80, 1.10)0.401284/7120.94 (0.85, 1.04)0.2237 ~ 42969/5360.92 (0.77, 1.11)0.40738/4290.94 (0.82, 1.07)0.33≥ 43683/3390.84 (0.69, 1.02)0.08544/2710.84 (0.71, 0.99)0.04*P for trend--0.06--0.02*† Number of cases and mortality count per groupHR: hazard ratio Table 2Associations between maternal age/paternal age and offspring all-cause mortality in the general population and among individuals with uneducated parentsSubgroupsOverall Participants(n = 10,064)Participants with uneducated parents(n = 8070)No.†HR (95% CI)P-valueNo.HR (95% CI)P-value Maternal age (years) ≤ 18541/3251.06 (0.94, 1.19)0.30423/2580.90 (0.78, 1.02)0.1119 ~ 244025/2368control-3179/1931control-25 ~ 302920/16291.03 (0.90, 1.17)0.702298/13410.98 (0.91, 1.06)0.6831 ~ 361368/7391.03 (0.89, 1.20)0.601067/6111.03 (0.92, 1.14)0.6237 ~ 42832/4451.06 (0.90, 1.25)0.50632/3481.02 (0.89, 1.18)0.74≥ 43378/2071.16 (0.94, 1.42)0.20309/1691.07 (0.88, 1.30)0.49P for trend--0.62--0.24 Paternal age (years) ≤ 18325/2050.97 (0.84, 1.13)0.70264/1711.06 (0.90, 1.25)0.4719 ~ 243165/1877control-2507/1540control-25 ~ 303293/18740.97 (0.83, 1.13)0.702571/15350.99 (0.92, 1.07)0.7531 ~ 361629/8820.94 (0.80, 1.10)0.401284/7120.94 (0.85, 1.04)0.2237 ~ 42969/5360.92 (0.77, 1.11)0.40738/4290.94 (0.82, 1.07)0.33≥ 43683/3390.84 (0.69, 1.02)0.08544/2710.84 (0.71, 0.99)0.04*P for trend--0.06--0.02*† Number of cases and mortality count per groupHR: hazard ratio Associations between maternal age/paternal age and offspring all-cause mortality in the general population and among individuals with uneducated parents † Number of cases and mortality count per group HR: hazard ratio Sensitivity analysis Initial analyses excluding participants with educated parents demonstrated consistent patterns: the null association between parental age and offspring chronic comorbidities persisted, and the relationships with activities of daily living (ADL) limitations and cognitive impairment remained robust (Fig. 3). In the analysis stratified by educational level (Table 2), a significant negative correlation was found between paternal age and all-cause mortality among the offspring of uneducated parents. Compared to the reference group (paternal age 25–30 years), offspring of fathers aged ≥ 43 years exhibited a 16% reduction in all-cause mortality risk (HR 0.84, 95% CI 0.71–0.99), with a statistically significant trend (P for trend = 0.02). Fig. 3Associations between maternal age/paternal age and the three elderly health indicators in offspring with uneducated parents Associations between maternal age/paternal age and the three elderly health indicators in offspring with uneducated parents We performed gender-stratified analyses to examine potential sex-specific associations between parental age and offspring functional outcomes, including ADL and cognitive function (Fig. 4). The inverse associations with paternal age remained consistent across both sexes, whereas maternal age showed stable associations only in male offspring (P for trend < 0.05). Notably, the inverse association of advanced paternal age were more pronounced in sons than in daughters. Fig. 4Gender-stratified associations between maternal age/paternal age and offspring health outcomes: daily living activities and cognitive function Gender-stratified associations between maternal age/paternal age and offspring health outcomes: daily living activities and cognitive function In age-stratified subgroup analyses, we categorized participants into two groups based on an 80-year threshold: an older cohort (≥ 80 years) and a younger cohort (< 80 years) [30, 31]. The associations between parental age and both activities of daily living (ADL) limitations and cognitive dysfunction remained robust in the older cohort (Fig. 5). However, in the younger cohort, these associations were not statistically significant (P > 0.05). Fig. 5Age-stratified associations between maternal age/paternal age and offspring health outcomes: daily living activities and cognitive function. The reported number encompass the total size and the number of individuals with positive outcomes Age-stratified associations between maternal age/paternal age and offspring health outcomes: daily living activities and cognitive function. The reported number encompass the total size and the number of individuals with positive outcomes In birth order-stratified analyses, participants were categorized into two groups: first/second-born offspring and third-or-higher birth order offspring (Table 3). While the inverse associations between advanced paternal age and offspring health outcomes remained consistent across both birth order groups, maternal age effects demonstrated differential patterns. Specifically, the association between maternal age and offspring health outcomes was significantly attenuated in the higher birth order group (third-or-higher), losing statistical significance (P > 0.05). Table 3Birth order-stratified associations between maternal age/paternal age and offspring health outcomes: daily living activities and cognitive functionSubgroupsLimitations on ADLCognitive dysfunctionOR (95% CI)P-valueOR (95% CI)P-valueThe first/second born in family (n = 6,177)Maternal age (years)≤ 180.92 (0.70, 1.21)0.551.00 (0.77, 1.28)0.9919 ~ 24control-control-25 ~ 301.11 (0.93, 1.33)0.261.24 (1.05, 1.46)0.01*31 ~ 361.17 (0.85, 1.60)0.330.92 (0.68, 1.23)0.5637 ~ 421.57 (1.01, 2.40)0.04*1.40 (0.94, 2.07)0.10≥ 431.08 (0.55, 2.04)0.822.13 (1.22, 3.68)< 0.01*P for trend-0.05-0.01*Paternal age (years)≤ 181.02 (0.72, 1.43)0.931.39 (1.02, 1.90)0.04*19 ~ 24control-control-25 ~ 300.80 (0.67, 0.94)< 0.01*0.92 (0.78, 1.07)0.2831 ~ 360.69 (0.53, 0.91)< 0.01*0.78 (0.61, 1.00)0.0637 ~ 420.56 (0.37, 0.83)< 0.01*0.74 (0.51, 1.06)0.10≥ 430.70 (0.40, 1.21)0.220.68 (0.40, 1.11)0.13P for trend-< 0.001*-0.001*The third or later born in family (n = 3,887)Maternal age (years)≤ 181.35 (0.69, 2.57)0.370.57 (0.27, 1.11)0.1119 ~ 24control-control-25 ~ 300.82 (0.63, 1.07)0.141.18 (0.92, 1.50)0.2031 ~ 361.22 (0.87, 1.70)0.241.15 (0.85, 1.57)0.3637 ~ 421.10 (0.71, 1.68)0.671.13 (0.77, 1.66)0.54≥ 431.43 (0.84, 2.43)0.191.38 (0.85, 2.22)0.19P for trend-0.28-0.15Paternal age (years)≤ 180.86 (0.42, 1.71)0.681.09 (0.55, 2.13)0.8019 ~ 24control-control-25 ~ 300.90 (0.67, 1.19)0.440.94 (0.72, 1.22)0.6331 ~ 360.62 (0.44, 0.88)< 0.01*0.63 (0.46, 0.86)< 0.01*37 ~ 420.66 (0.44, 0.99)< 0.05*0.56 (0.39, 0.82)< 0.01*≥ 430.41 (0.25, 0.67)< 0.001*0.60 (0.39, 0.92)0.02*P for trend-< 0.001*-< 0.001* Table 3Birth order-stratified associations between maternal age/paternal age and offspring health outcomes: daily living activities and cognitive functionSubgroupsLimitations on ADLCognitive dysfunctionOR (95% CI)P-valueOR (95% CI)P-valueThe first/second born in family (n = 6,177)Maternal age (years)≤ 180.92 (0.70, 1.21)0.551.00 (0.77, 1.28)0.9919 ~ 24control-control-25 ~ 301.11 (0.93, 1.33)0.261.24 (1.05, 1.46)0.01*31 ~ 361.17 (0.85, 1.60)0.330.92 (0.68, 1.23)0.5637 ~ 421.57 (1.01, 2.40)0.04*1.40 (0.94, 2.07)0.10≥ 431.08 (0.55, 2.04)0.822.13 (1.22, 3.68)< 0.01*P for trend-0.05-0.01*Paternal age (years)≤ 181.02 (0.72, 1.43)0.931.39 (1.02, 1.90)0.04*19 ~ 24control-control-25 ~ 300.80 (0.67, 0.94)< 0.01*0.92 (0.78, 1.07)0.2831 ~ 360.69 (0.53, 0.91)< 0.01*0.78 (0.61, 1.00)0.0637 ~ 420.56 (0.37, 0.83)< 0.01*0.74 (0.51, 1.06)0.10≥ 430.70 (0.40, 1.21)0.220.68 (0.40, 1.11)0.13P for trend-< 0.001*-0.001*The third or later born in family (n = 3,887)Maternal age (years)≤ 181.35 (0.69, 2.57)0.370.57 (0.27, 1.11)0.1119 ~ 24control-control-25 ~ 300.82 (0.63, 1.07)0.141.18 (0.92, 1.50)0.2031 ~ 361.22 (0.87, 1.70)0.241.15 (0.85, 1.57)0.3637 ~ 421.10 (0.71, 1.68)0.671.13 (0.77, 1.66)0.54≥ 431.43 (0.84, 2.43)0.191.38 (0.85, 2.22)0.19P for trend-0.28-0.15Paternal age (years)≤ 180.86 (0.42, 1.71)0.681.09 (0.55, 2.13)0.8019 ~ 24control-control-25 ~ 300.90 (0.67, 1.19)0.440.94 (0.72, 1.22)0.6331 ~ 360.62 (0.44, 0.88)< 0.01*0.63 (0.46, 0.86)< 0.01*37 ~ 420.66 (0.44, 0.99)< 0.05*0.56 (0.39, 0.82)< 0.01*≥ 430.41 (0.25, 0.67)< 0.001*0.60 (0.39, 0.92)0.02*P for trend-< 0.001*-< 0.001* Birth order-stratified associations between maternal age/paternal age and offspring health outcomes: daily living activities and cognitive function Above multivariate analysis revealed a significant association between potential elder sibling caregiving and the prevalence of cognitive dysfunction in offspring (P < 0.01). To further investigate this relationship, we performed sensitivity analysis in participants with elder siblings (Fig. 6). Although the association between parental age and both ADL limitations and cognitive dysfunction demonstrated a flatter trend, it remained statistically significant (P for trend < 0.05). Fig. 6Associations between maternal age/paternal age and offspring cognitive function in participants with older siblings Associations between maternal age/paternal age and offspring cognitive function in participants with older siblings Discussion Our study reveals significant observational associations between parental age and long-term offspring health outcomes. Specifically, advanced maternal age was associated with an increased risk of both activities of daily living (ADL) limitations and cognitive dysfunction in offspring. In contrast, advanced paternal age was associated with a lower risk of these two adverse outcomes. Notably, these associations displayed substantial gender-specific patterns, with the inverse associations of paternal age and the positive associations of maternal age being more pronounced in male offspring. The influence of parental age on offspring health has been well-documented in previous research [32–35]. The predominantly manifestation was adverse perinatal outcomes, including preterm birth, reduced gestational age, and low birth weight [36–38], as well as distinct clinical symptoms and frailty in early development stages [39]. Moreover, certain consequences, including recessive genetic variations and epigenetic modifications, may persist throughout the offspring’s lifespan, potentially influencing healthy aging trajectories and mortality risks [23, 40]. These well-documented associations may provide a plausible explanation for the observed positive association between maternal age and offspring’s long-term health outcomes. Despite these established findings, research in this domain remains limited, particularly regarding long-term outcomes. Our study contributes novel insights to this emerging field of research, providing valuable evidence for future investigations. Longitudinal cohort studies must account for the inherent challenges of distinguishing age, period, and cohort effects, particularly when analyzing data from exceptionally elderly populations. Our study sample, comprising individuals born between 1892 and 1948, a generation that experienced significant historical adversities, including wars, economic deprivation, and widespread infectious diseases. These early-life exposures likely had lasting impacts on health trajectories, potentially influencing long-term outcomes. We stratified the analysis by age to minimize potential confounding effects. The association between parental age and offspring health outcomes remained consistent in individuals aged 80 years and older. Furthermore, given the close relationship between birth order and parental age at childbirth, we also stratified the analysis by parity. The results demonstrated that the associations remained stable in offspring first- and second-born. These analyses further support the robustness of our findings. Interestingly, we observed that the effects of paternal and maternal age on offspring health outcomes were diametrically opposed. While advanced maternal age was associated with increased risks of ADL limitations and cognitive dysfunction, advanced paternal age showed an inverse association with these outcomes. However, this finding should be interpreted with caution, as differential survival and cohort selection bias must be considered. Our cohort had a mean age of 86 years (birth years 1892–1948), representing a highly selected “healthy survivor” cohort. Offspring born to older fathers who inherited deleterious genetic mutations or experienced frailty would have been more likely to die or become disabled earlier in life, thereby being systematically excluded from our analysis of late-life health. This selective survival may substantially influence our findings: the observed inverse association between advanced paternal age and functional outcomes is likely a selection artifact rather than a true biological benefit. Indeed, current biological evidence consistently links advanced paternal age with increased risks of de novo mutations and neurodevelopmental disorders in offspring [6, 41, 42], suggesting that any genuine biological effect would likely be adverse rather than protective. Therefore, the observed inverse association is more plausibly explained as an association observed among long-lived survivors rather than evidence of benefit. Conversely, if only the most robust offspring of older mothers survived, the true detrimental effect of advanced maternal age may be even stronger than our estimates suggest. Substantial evidence indicates that advanced maternal age significantly elevates the risk of congenital abnormalities [43, 44] and multiple diseases [21, 45, 46]. Our analysis did not identify statistically significant associations between parental age and late-life chronic comorbidity prevalence in offspring, though a non-significant trend toward elevated risk was observed with advancing maternal age (P for trend = 0.10). Despite the fact that a wide variety of chronic non-infectious diseases/disorders were collected in the CLHLS cohort, existing evidence has established associations between only a restricted subset of these conditions with parental age [45–48]. This may partially explain the null association observed in our study. Moreover, all data pertaining to chronic noninfectious diseases/disorders were self-reported. Although numerous studies proposed that self-reported data can offer accurate evaluations of disease prevalence [49, 50], systematic errors inherent to this method must be considered. The more pronounced effect in sons observed in our study warrants further investigation. While previous research suggests that X-linked genetic mechanisms—such as the compensatory effect of a second X chromosome in females—may contribute to sex differences in certain diseases [51–54], whether these mechanisms underlie our findings remains speculative and requires further validation. In addition, alternative explanations, including differences in survival rates between sexes, social roles, or health behaviors, may also play a role. Due to the lack of genetic and biomarker data in our study, we cannot empirically test specific mechanistic pathways. Future research integrating comprehensive genetic, biomarker, and life-course data is needed to elucidate the potential mechanisms driving these sex-specific associations. Anyway, Parents giving birth at an advanced age should be provided with sufficient support. Especially for older mothers, targeted interventions—including comprehensive genetic counseling, enhanced maternal health monitoring, and personalized risk stratification—are necessary to alleviate potential health risks for their children. Given the pronounced adverse effects of maternal age on male offspring, health initiatives could focus on early interventions that address cognitive and functional decline in old man, particularly in those with advanced maternal age. The escalating global burden of elder care accompanying population aging has become increasingly pronounced. Existing evidence confirms the critical association between caregiving quality and geriatric health outcomes [55, 56]. Research indicates that among offspring with sibling caregivers, the positive association between advanced maternal age and adverse outcomes weakens, while the negative association between advanced paternal age and adverse outcomes becomes more pronounced, suggesting that early-life caregiving may influence health outcomes in later life. Developing targeted policy interventions to address caregiving deficits among vulnerable child populations caused by parental occupational constraints—including long working hours, high-stress employment environments, and labor migration patterns—may contribute to improved health outcomes in old age. At the same time, potential confounding factors related to family structure and socioeconomic status must be considered. Sibling caregiving is closely associated with family size, rural residence, birth order, and family socioeconomic status—factors that may independently influence offspring health outcomes. Although our analysis adjusted for some covariates, residual confounding may still exist due to the complex interrelationships among these variables. For example, larger family size may reflect both greater availability of sibling caregivers and lower socioeconomic resources, making it difficult to distinguish the specific contribution of caregiving from broader family environmental influences. Therefore, this effect should still be interpreted with caution, as it may partially reflect unmeasured family-level characteristics rather than a direct causal effect of caregiving itself. Future research incorporating detailed data on caregiving intensity, quality, and duration, as well as more comprehensive family socioeconomic information, is needed to further elucidate these complex dynamics. No correlation was found between maternal age and paternal age and the likelihood of their offspring dying at the age of 90 in a cross-sectional study conducted in the United States [57]. Although in our study no significant direct association was found between parental age and all-cause mortality of the offspring either, a trend towards lower mortality risk with increasing paternal age emerged, especially among offspring of non-educated parents. Our findings highlight observational associations that may inform further research into the relationship between parental age and offspring health. As societies experience declining fertility rates, understanding these observed patterns could contribute to evidence-based discussions on family health. In particular, the observed associations with maternal age warrant attention in future studies examining health support needs across generations. The study also underscores gaps in research regarding the long-term implications of parental age, including potential genetic, epigenetic, and socio-environmental pathways. Future studies should explore these dynamics further, incorporating more comprehensive data on family structure, socioeconomic factors, and life-course health behaviors. Limitations Several important limitations should be considered. First, a central methodological limitation concerns the measurement of parental age, which was not directly reported but reconstructed using parental age at death and the participant’s age at that time. This approach is inherently susceptible to recall bias, inaccuracies—particularly when parental death occurred early—and potential misreporting in proxy interviews. Random misclassification may attenuate true associations, whereas systematic misclassification could introduce spurious findings. Therefore, the observed protective association with older paternal age should be interpreted with caution, as it may be partly or entirely attributable to such measurement error rather than a genuine biological or social effect. Second, our cohort had a mean age of 86 years (birth years 1892–1948), representing a highly selected population of survivors. This selective survival may substantially influence our findings: the observed inverse association between advanced paternal age and functional outcomes could reflect survival selection rather than a true biological benefit, as frailer offspring of older fathers likely died earlier. Conversely, if only the most robust offspring of older mothers survived, the true detrimental effect of advanced maternal age could be even stronger than our estimates suggest. Third, potential confounding by birth cohort and historical context cannot be ruled out. Participants were born or grow up during a period of profound social and economic transformation in China (war, famine, Cultural Revolution). Although we adjusted for available covariates, residual confounding remains highly possible. Additionally, the assessment of chronic comorbidities relied on self-reported data without clinical validation, and the etiological heterogeneity of the 22 conditions may have masked disease-specific associations, partially explaining the null findings for this outcome. Finally, as an observational study, we cannot establish causal relationships or rule out residual confounding from unmeasured factors such as urban versus rural residence, personal nutritional status [58], family wealth, or maternal health during pregnancy [59], which may correlate with both parental age and offspring outcomes. Finally, the limited follow-up period and advanced baseline age constrained our ability to detect significant mortality associations, with only a non-significant trend observed. Future research should address these limitations by including younger cohorts with longer follow-up, utilizing clinically validated diagnoses, and applying methods such as sibling comparisons or inverse probability weighting to account for survival bias. Conclusion Overall, this study identifies observational associations between parental age and long-term health outcomes in offspring. Advanced paternal age showed an inverse association with risks of ADL limitations and cognitive dysfunction, whereas advanced maternal age showed positive associations with these outcomes. These findings add to the growing body of observational evidence on intergenerational health patterns and may inform future research and policy discussions in the context of global population aging and declining fertility rates. Longitudinal studies with more comprehensive data on family dynamics, caregiving environments, and health behaviors across the life course are needed to further explore the pathways that may link parental age to offspring aging outcomes. Abbreviations: ADL, Activities of daily living. MMSE, Mini-mental state examination. Adjusted for age, gender, race, BMI, education level, occupation, family income, tobacco consumption, alcohol consumption, exercise, dietary diversity, timely treatment in childhood, childhood experiences of food insecurity and possible elder siblings care. The reported number encompass the total size and the number of individuals with positive outcomes. Supplementary Information Supplementary Material 1 Supplementary Material 1