Introduction

Adolescence is a critical period for developing higher-order cognition and the corresponding refinement of the structures and functions of related brain regions.[1–4] The development of higher-order cognition is determined by neurobiological and genetic mechanisms as well as household, community, and environmental factors.[2,5,6] In suboptimal conditions, deficits in higher cognitive functions may occur, which can increase the risk of other health problems later in life. For example, evidence has shown that inferior cognitive abilities during adolescence and young adulthood are associated with an increased risk of dementia and related disorders later in life.[7–10] Therefore, it is crucial to identify the modifiable factors that influence the cognitive function of adolescents and young adults.

Previous studies have demonstrated that the family environment, with its complexity and diverse components, plays a critical role in offspring's cognition and behavior.[11–15] The lifestyle is an important factor in the family environment and is associated with children's cognitive development.[16–18] Maternal behaviors, including a healthy diet, physical activity, non-smoking, infrequent alcohol consumption, and sufficient sleep, have been linked to better offspring's cognitive development in early childhood.[16,17,19–22] Paternal lifestyle factors, such as alcohol consumption, demonstrate similar associations, although research in this area is limited.[18] Furthermore, parental exposure to smoking is a risk factor for cognitive function in offspring.[23,24]

Studies in this area have improved our understanding of the influence of lifestyles on children's cognition. However, it is limited in various ways. First, previous studies only considered individual lifestyle factors or lifestyle profiles of one parent.[17,18,25] Thus, they did not capture how individual lifestyle factors coexisted and interconnected within a complex family system. Second, most previous studies investigating parental lifestyles focused on its impact on offspring in early childhood.[19,20,22] However, limited information exists on the overall association between healthy parental lifestyle and offspring's cognitive performance during adolescence and young adulthood. Therefore, we investigated the prospective associations between overall parental lifestyles and cognitive performance in offspring during adolescence and young adulthood using a nationally representative sample from China in this study; We also separately examined the association of specific maternal or paternal healthy lifestyles with offspring's cognitive performance.

Methods Study population

China Family Panel Studies (CFPS), a nationwide, comprehensive, and longitudinal social survey launched by Peking University,[26] includes 25 provinces and represents 95% of China's population.[27] The first survey was conducted in 2010 and was followed up every two years. CFPS collected information on economic activity, education and work status, family relations and family dynamics, population migration, physical and mental health, and other topics at the community, household, and individual levels via face-to-face interviews. Participants aged 10 years or older received cognitive tests.[26] More details and data were available online (http://www.isss.pku.edu.cn/cfps). The Biomedical Ethics Review Committee of Peking University approved CFPS (No. IRB00001052-14010), and all participants provided written informed consent before enrollment.

Our study used data collected in 2010, 2012, 2014, 2016, and 2018, with the 2010 survey used as the baseline. Of all households included in the CFPS data, there were 3464 offspring aged 10–15 years at baseline. Among them, we excluded offspring with missing data on the lifestyle information of both parents, who were disabled or having brain-related diseases, or without valid cognitive test scores in 2012, 2014, 2016, or 2018. Thus, 2531 offspring were included in the final analysis [Supplementary Figure 1, https://links.lww.com/CM9/B737].

Assessment of lifestyle factors in parents

All parental lifestyle factors were assessed using a baseline questionnaire. Questions about dietary information included whether meat, fish and other aquatic products, fresh vegetables and fruits, dairy products, bean products, eggs, pickled food, and fried food were consumed and the frequency of consumption in the past three months. Based on the Dietary Guidelines for Chinese Residents and the Dietary Diversity Score,[28,29] parents who consumed any one of the following six items weekly received one point: Meat, fish and other aquatic products, fresh vegetables and fruits, dairy products, soy products, and eggs. The modified Dietary Diversity Score was defined as the sum of the scores for each food group, ranging from 0 (least healthy) to 6 (most healthy).

Inquiries into tobacco smoking included current smoking status, the number of cigarettes smoked per day for current smokers, years of quitting for past smokers, and other related information. Information on alcohol consumption was obtained by asking parents whether they drank alcohol more than three times a week in the previous month, which was defined as habitual drinking. Questions regarding physical activity included exercise frequency in the previous week and average duration. Parents also reported whether they had a habit of napping and the mean duration.

Definition of the healthy lifestyle score

Previous studies have used and recognized the definitions of never smoking, non-habitual drinking, regular exercise, and healthy diet as healthy behaviors that have been linked to several chronic health outcomes, including cardiovascular disease, diabetes, dementia, and cancer.[30–33] Several studies have also used a short afternoon nap as a healthy indicator and linked it with certain health outcomes.[34–36] Based on previous evidence and expert knowledge, this study constructed healthy lifestyle scores based on the following five low-risk lifestyle factors: (1) Never smoking; (2) Non-habitual drinking (no more than three times per week in the last month); (3) Weekly exercise; (4) Taking a short afternoon nap (<30 min/day); and (5) Modified dietary diversity score ≥5 points (top 35%). Parents received one point if they met any of the criteria for low-risk lifestyle behaviors. Combining these five lifestyle factors resulted in healthy lifestyle scores of 0–5, with higher scores indicating healthier lifestyles.

If both parents had complete lifestyle information, the parental healthy lifestyle score was calculated as the average of the maternal and paternal healthy lifestyle scores. If only one parent completed the corresponding questionnaire, the parental healthy lifestyle score was defined as the available maternal or paternal healthy lifestyle scores.

Assessment of outcome

CFPS developed two sets of cognitive measures and used them alternatively in different years. In the waves of 2010, 2014, and 2018, vocabulary and mathematics tests were conducted to assess "crystallized intelligence", referred to as the ability to use the knowledge acquired through past learning, experience, and education.[37,38] In the waves of 2012 and 2016, immediate word recall, delayed word recall, and number series tests were performed to measure "fluid intelligence", referred to as the ability to reason and solve problems in unique and novel situations.[37,38] These cognitive tests have been widely adopted and validated in previous studies, and suggested as stable and mature instruments to evaluate the cognitive function levels of individuals.[27,38–40] The score ranged 0–34, 0–24, 0–10, and 0–15 for the vocabulary, mathematics, immediate/delayed word recall, and number series tests, respectively, with higher scores indicating better cognitive performance.

We Z-transformed all test scores to unify the dimensions. Overall fluid intelligence was defined as the sum of the Z-transformed immediate word recall, delayed word recall, and number series test scores. Overall crystallized intelligence was defined as the sum of the Z-transformed vocabulary and mathematics test scores.

Assessment of covariates

We adopted covariates at the offspring, parents, and household income levels. Offspring's characteristics included the follow-up year, age (continuous), sex (male or female), living with mother (yes or no), living with father (yes or no), body mass index (BMI, continuous), age at speaking a complete sentence (binary), and age at counting from one to ten (binary). Parental characteristics included age (binary), ethnic group (Han Chinese or ethnic minorities), education levels (illiterate/semi-literate, primary school, junior high school, or high school and above), employment status (unemployed or employed), BMI (<18.5 kg/m2,18.5–23.9 kg/m2, 24.0–27.9 kg/m2, or ≥28.0 kg/m2), and chronic disease history (yes or no). Household characteristics included residence (urban or rural) and annual per-capita household income on a log scale (binary). Offspring's lifestyle factors were considered to further explore the role of offspring's lifestyle in the association between parental lifestyle and offspring's cognitive performance. The offspring reported lifestyle information, including smoking, drinking, exercise, diet, and sleep. Apart from sleep, the questions were similar to those asked of the parents, and we used the same definitions of low-risk lifestyle factors. For sleep, we considered 9–11 h/night for offspring aged 10–13 years and 8–10 h/night for those aged 14–15 years as low risk.[41] We also computed healthy lifestyle scores for the offspring, ranging 0–5.

Statistical analysis

The baseline characteristics of the study population were presented as medians and interquartile ranges (Q1–Q3) (continuous variables) and n (%)(categorical variables). The distribution of these characteristics by parental healthy lifestyle score subgroups was compared using the Kruskal–Wallis and chi-squared tests. Missing categorical values for covariates were assigned to a separate group, and median values were assigned to continuous variables of covariates with missing values.

We first explored the association between parental adherence to a healthy lifestyle and offspring cognitive performance using the generalized estimating equation models. Parental healthy lifestyle scores were evaluated as both continuous (per 1-unit increment) and categorical (in tertiles of parental healthy lifestyle scores, with the bottom tertile as the reference category) variables. The rationale for categorizing lifestyle scores into tertiles in this study was related to the study's hypothesis and statistical power. We adjusted for the offspring's follow-up year, age, and sex in the crude model. In the fully adjusted model, we additionally adjusted for offspring living status (with mother and/or father), maternal age, maternal ethnic group, maternal education level, and maternal employment status, paternal age, paternal ethnic group, parternal education level, and paternal employment status, residency, and household income per capita on a log scale. We performed tests for linear trends by entering each group's median healthy parental lifestyle score as a continuous variable. Analyses were conducted separately according to two sets of cognitive measures (crystallized intelligence and fluid intelligence). Subsequently, we separately evaluated the relationship between overall or individual maternal and paternal healthy lifestyles and offspring's cognitive performance.

In the subgroup analyses, we investigated whether the association between baseline parental healthy lifestyle scores and offspring's cognitive performance varied by offspring's age, sex, residency, parental education level, household income per capita, and parental employment status. Effect modification was detected by adding the interaction terms of parental healthy lifestyle scores and the abovementioned variables. Additionally, several sets of sensitivity analyses were conducted to test the robustness of the findings, and we further explored the role of offspring lifestyle in the associations between parental lifestyle and offspring's cognitive performance [Supplementary Methods 1, https://links.lww.com/CM9/B737].

All statistical analyses were performed using Stata (version 14.1, StataCorp LLC, TX, USA) and SAS software (version 9.4, SAS Institute Inc., Cary, NC, USA). All P values were two-sided, and statistical significance was defined as P <0.05.

Results Participant characteristics

The baseline characteristics of the offspring and parents are shown in Table 1, according to tertiles of the parental healthy lifestyle score. The median age of offspring was 13 (11–14) years, with 49.8% boys. Their maternal and paternal median baseline age was 38 (36–41) years and 40 (37–43) years, respectively. Parents in this study were predominantly Han Chinese (88.6% mothers and 89.2% fathers). Generally, offspring's age, sex, and BMI were similar across healthy parental lifestyle score subgroups. Parents with higher healthy lifestyle scores had higher education levels and household incomes, were more likely to be of Han Chinese, and lived in urban areas.

Table 1 - Baseline characteristics of offspring and parents, overall and subgroups by lifestyle score. Characteristics Overall Tertiles of the healthy lifestyle score Bottom Medium Top H/χ 2 P value Parental healthy lifestyle score, median 2.0 1.5 2.0 3.0 – – Offspring characteristics Number of offspring 2531 882 846 803 – – Age (years) 13 (11–14) 12 (11–14) 13 (11–14) 12 (11–14) 1.27* 0.542 BMI (kg/m2) 17.3 (15.6–19.2) 17.1 (15.6–19.2) 17.3 (15.6–19.2) 17.4 (15.6–19.2) 0.46* 0.796 Sex (male) 1260 (49.8) 434 (49.2) 419 (49.5) 407 (50.7) 0.40† 0.818 Ethnic group (Han) 2239 (88.5) 726 (82.3) 761 (89.0) 752 (93.7) 55.69† <0.001 Living with mother 2479 (98.0) 847 (96.0) 835 (98.7) 797 (99.3) 25.26† <0.001 Living with father 2476 (97.8) 879 (99.7) 817 (96.6) 780 (97.1) 22.01† <0.001 Offspring earlier cognitive ability Age to speak a complete sentence (years) 1.5 (1.3–2.0) 2.0 (1.3–2.0) 1.7 (1.3–2.0) 1.5 (1.2–2.0) 51.45* <0.001 Age to count from 1 to 10 (years) 3.0 (2.0–4.0) 3.0 (2.2–4.2) 3.0 (2.0–4.0) 2.5 (2.0–3.0) 133.88* <0.001 Maternal characteristics Age (years) 38 (36–41) 38 (36–42) 38 (36–41) 38 (36–41) 2.62* 0.270 BMI (kg/m2) 22.0 (20.2–24.1) 22.0 (20.3–24.2) 21.8 (20.0–23.9) 22.2 (20.3–24.4) 8.91* 0.012 Ethnic group (Han) 2070 (88.6) 629 (81.7) 704 (89.2) 737 (94.7) 65.54† <0.001 Education 344.41† <0.001 Illiterate/semi-literate 813 (34.8) 384 (49.9) 297 (37.6) 132 (17.0) – – Primary school 628 (26.9) 212 (27.5) 233 (29.5) 183 (23.5) – – Junior high school 635 (27.2) 158 (20.5) 201 (25.5) 276 (35.5) – – High school and above 261 (11.2) 16 (2.1) 58 (7.4) 187 (24.0) – – Chronic disease history 319 (13.7) 104 (13.5) 126 (16.0) 89 (11.5) 6.79† 0.034 Employment status (employed) 1365 (60.2) 450 (61.0) 480 (62.1) 435 (57.5) 3.71† 0.156 Lifestyle factors Never smoking 2295 (98.2) 745 (96.8) 777 (98.5) 773 (99.4) 15.38† <0.001 Non-habitual drinking 2276 (97.4) 725 (94.2) 781 (99.0) 770 (99.0) 47.25† <0.001 Weekly exercise 365 (15.6) 9 (1.2) 66 (8.4) 290 (37.3) 430.36† <0.001 Taking a short afternoon nap 40 (1.7) 6 (0.8) 10 (1.3) 24 (3.1) 13.63† 0.001 Diverse diet 794 (34.0) 28 (3.6) 175 (22.2) 591 (76.0) 976.36† <0.001 Paternal characteristics Age (years) 40 (37–43) 40 (37–44) 39 (37–43) 40 (37–43) 5.31* 0.070 BMI (kg/m2) 22.5 (20.8–24.7) 22.1 (20.5–24.1) 22.3 (20.4–24.5) 23.1 (21.3–25.5) 58.78* <0.001 Ethnic group (Han) 1863 (89.2) 723 (83.2) 541 (92.1) 599 (94.8) 57.62† <0.001 Education 302.92† <0.001 Illiterate/semi-literate 445 (21.3) 289 (33.1) 98 (16.7) 58 (9.2) – – Primary school 546 (26.1) 250 (28.6) 181 (30.8) 115 (18.2) – – Junior high school 765 (36.6) 284 (32.5) 228 (38.8) 253 (40.0) – – High school and above 337 (16.1) 50 (5.7) 81 (13.8) 206 (32.6) – – Chronic disease history 212 (10.1) 86 (9.9) 64 (10.9) 62 (9.8) 0.51† 0.776 Employment status (employed) 1480 (72.2) 589 (69.8) 423 (72.8) 468 (74.8) 4.59† 0.101 Lifestyle factors Never smoking 545 (26.0) 55 (6.3) 208 (35.4) 282 (44.6) 316.53† <0.001 Non-habitual drinking 1409 (67.3) 474 (54.3) 421 (71.6) 514 (81.3) 128.58† <0.001 Weekly exercise 377 (18.0) 25 (2.9) 70 (11.9) 282 (44.6) 453.49† <0.001 Taking a short afternoon nap 37 (1.8) 5 (0.6) 7 (1.2) 25 (4.0) 25.73† <0.001 Diverse diet 831 (39.7) 98 (11.2) 239 (40.7) 494 (78.2) 686.47† <0.001 Family characteristics Residency (urban) 976 (38.6) 199 (22.6) 273 (32.3) 504 (62.8) 307.98† <0.001 Household income per capita in 2010, log scale 8.5 (7.9–9.0) 8.2 (7.7–8.7) 8.4 (7.9–8.9) 8.9 (8.4–9.4) 295.46* <0.001

The data are presented as medians and interquartile ranges (Q1–Q3) for continuous variables or n (%) for categorical variables. BMI: Body mass index; –: Not available. *H value. †χ2 value.

Associations of parental healthy lifestyle scores with offspring's cognitive performance

In the multivariable-adjusted analysis, we found that a higher healthy parental lifestyle score was associated with better offspring's performance in both fluid and crystallized intelligence [Table 2]. Particularly, the mean differences comparing the top to the bottom tertile of parental healthy lifestyle scores were 0.53 (95% confidence interval [CI]: 0.29–0.77) and 0.35 (95% CI: 0.16–0.54) for fluid and crystallized intelligence in offspring, respectively. Moreover, each 1-unit increment in the parental healthy lifestyle score was associated with a 0.30 (95% CI: 0.16–0.43) and 0.24 (95% CI: 0.13–0.35) increase in the overall fluid and crystallized intelligence score among offspring, respectively. We also observed similar and significant associations for domain-specific cognitive functions, with the corresponding differences of 0.15–0.20 for domains of immediate word recall, delayed word recall, number series test, vocabulary test, and mathematics test.

Table 2 - Associations of parental healthy lifestyle scores with offspring's cognitive performance. Variables Tertiles of the healthy lifestyle score Healthy lifestyle score (per-unit score increase) P trend Bottom Medium Top Coef. (95% CI) Coef. (95% CI) Coef. (95% CI) Coef. (95% CI) Fluid intelligence Number of offspring 750 727 708 2185 Parental healthy lifestyle score, median 1.5 2.0 3.0 2.0 Overall fluid intelligence Crude model Ref. 0.53 (0.32, 0.74) 1.02 (0.82, 1.23) 0.58 (0.46, 0.70) <0.001 Fully adjusted model Ref. 0.34 (0.12, 0.56) 0.53 (0.29, 0.77) 0.30 (0.16, 0.43) <0.001 Domains of fluid intelligence Immediate word recall Crude model Ref. 0.19 (0.10, 0.28) 0.31 (0.22, 0.39) 0.17 (0.12, 0.22) <0.001 Fully adjusted model Ref. 0.13 (0.03, 0.22) 0.15 (0.05, 0.25) 0.08 (0.02, 0.14) 0.013 Delayed word recall Crude model Ref. 0.16 (0.07, 0.25) 0.30 (0.22, 0.39) 0.16 (0.11, 0.21) <0.001 Fully adjusted model Ref. 0.12 (0.03, 0.22) 0.20 (0.09, 0.30) 0.09 (0.03, 0.15) <0.001 Number series test Crude model Ref. 0.20 (0.10, 0.29) 0.44 (0.35, 0.53) 0.26 (0.21, 0.31) <0.001 Fully adjusted model Ref. 0.10 (0, 0.19) 0.18 (0.07, 0.28) 0.11 (0.05, 0.17) 0.001 Crystallized intelligence Number of offspring 720 685 655 2060 Parental healthy lifestyle score, median 1.5 2.0 3.0 2.0 Overall crystallized intelligence Crude model Ref. 0.62 (0.43, 0.80) 1.07 (0.90, 1.26)