2.1. Study design
Data from the Ma’anshan birth cohort (MABC) study in China served as the basis for this research. Between May 2013 and September 2014, 3,474 pregnant women at their first prenatal visit at Maternal and Child Health Hospital in Ma’anshan were recruited. The research procedure has received approval from Anhui Medical University's ethics and research committees (No. 20131195). Prior to being enrolled in the study, the participants provided their written informed permission.
2.2. Study Participants
Inclusion criteria for pregnant women included: above the age of 18, within 14 gestational weeks; living in Ma’anshan city; willing to come to the center for antenatal checkups and delivery; without existed mental illness and able to understand and complete the questionnaire. We included a total of 3273 singleton live births after excluded multiple pregnancies, adverse pregnancy outcomes including embryonic arrest, stillbirth, spontaneous abortion, ectopic pregnancy and therapeutic abortion (n = 201). We further excluded women who had no data for single pregnancy anxiety (n = 56) and children who had no available physical data (n = 8). Finally, 3,154 mother-child pairs were included in current study. The detailed flow chart of the mother-child pair selection is shown in Fig. 1.
2.3. Pregnancy-related Anxiety
Using the Pregnancy-Related Anxiety Questionnaire (PRAQ), maternal anxiety in the 1st trimester (interquartile range: 8 to 11 weeks), 2nd trimester (interquartile range: 25 to 26 weeks), and 3rd trimester were evaluated (interquartile range 33 to 34 weeks). Thirteen items make up the screening scale, divided into three subscales: anxiety for the health of the woman herself (six items), anxiety for the health of the fetus (five items), and anxiety for childbirth (two items). For each question, women were asked to rate their level of anxiety on a scale of 1 to 4: "never worried," "occasionally worried," "often worried," and "always worried.". The higher the score, the more severe the woman's pregnancy-related anxiety symptoms. The total score of the scale ranged from 13 to 52. When the overall score was ≥ 24, women were classified as experiencing pregnancy-related anxiety. Cronbach's coefficients were 0.81, 0.64, 0.78, and 0.74 for the total questionnaire and the three subscales, respectively[22]. The retest reliability coefficients were 0.79, 0.67, 0.75, and 0.76, respectively[3]. The scale, which has been used in screening and research on pregnancy-related anxiety in many regions of China.
2.4. Early Childhood Physical Growth
Children's anthropometric measurements were performed by child health professionals at the Ma’anshan Maternal and Child Health Hospital. The physical examination of children was followed from birth to 6 years of age. Children were followed up from birth to 6 years of age. Physical examinations for children were conducted every 6 months. Height and weight were measured from birth to 6 years. The body fat (BF) examination was performed from 4 to 6 years of age.
Using a mechanical height and weight scale, children's height and weight were measured with an accuracy of 0.1 kg for weight and 0.1 cm for height (model: RRZ-50-RP). Children were requested to remove their shoes and hats and wear light clothes for the measuring. Children's height and weight were measured twice and averaged. Children were measured using the Korean-made InBody J20 body composition analyzer. The test requires children’s cooperation, and the BF can be obtained by keeping still for 1 to 2 minutes.
The z-scores for BMI-for-age and of weight-for-age children were identified and categorized using World Health Organization (WHO) Child Growth Standards and Growth Reference Data [23].
2.5. Early Growth Patterns In Children
2.5.1. Rapid weight gain during infancy
The level of weight gain in children was depended on the difference in WAZ between birth and 12/24/36 months (∆z-scores = z-scores 12/24/36 months - z-scores 0 months). We used the criteria for rapid weight gain (RWG) in infancy defined as a change in weight SD score > + 0.67 from birth to 12/24/36 months, which is the most common and broadly accepted definition of RWG [24]. We classified the level of early weight gain into RWG (Δz-scores > 0.67) and no rapid weight gain (NRWG) (Δz-scores ≤ 0.67).
2.5.2. Child Bmi And Bf Trajectory From 48 To 72 Months Of Age
Group-based trajectory modeling (GBTM) was used to calculate BMI and BF trajectories for children from 48 to 72 months. Children with at least three BMI or BF values were permitted to examine quadratic trajectory models. The latent trajectory patterns of BMI rise in longitudinal BMI and BF data were identified using GBTM. Maximum likelihood techniques were implemented for parameter estimation and model fitting. Bayesian Information Criterion (BIC) and similarity trajectory shapes were used to identify the appropriate number of trajectory groups for different trajectories and to represent the appropriate functions for various trajectories in GBTM. Each person was assigned to the group with the most likely trajectory. Each person was assigned to the group with the most likely trajectory.
2.6. Covariates
Combining literature review and directed acyclic graphs (DAGs), we were able to find possible confounders (Figure S1). The covariates included maternal age, race (Han or others), family monthly income per capita (< 2500, 2500–4000, or > 4000 RMB), residence (urban or rural areas), parental education level (junior high school or below, senior middle school, junior college, or bachelor degree or above), maternal pre-pregnancy BMI, parity (0, ≥ 1), maternal metabolic dysfunctions (yes or no), alcohol use (yes or no), and tobacco use during pregnancy (yes or no), father’s BMI, and children’ s sex (boys or girls). The covariates were shown in Table 1.
Table 1
Characteristics of the included participants in the study [Mean ± SD or n (%)]
Characteristics | Included participants (n = 3154) |
Maternal characteristics | |
Age, years [Mean (SD)] | 26.4(3.6) |
Race [n (%)] | |
Han | 3105(98.4) |
Others | 49(1.6) |
Residence)# [n (%)] | |
Rural | 1224(38.9) |
Urban | 1920(61.1) |
Parity# [n(%)] | |
Multipara | 310(9.8) |
Nulliparous | 2839(90.2) |
Education level# [n (%)] | |
Junior high school or below | 623(19.8) |
Senior middle school | 706(22.4) |
Junior college | 980(31.1) |
Bachelor degree or above | 844(26.8) |
Incomes (RMB)# [n (%)] | |
< 2500 | 819(26.2) |
2500–4000 | 1339(42.8) |
> 4000 | 970(31.0) |
Alcohol use)# [n (%)] | |
No | 2891(91.9) |
Yes | 255(8.1) |
Tobacco use)# [n (%)] | |
No | 3146(99.8) |
Yes | 5(0.2) |
Pre-pregnancy BMI, kg/m2 [Mean (SD)] | 20.9(2.9) |
Pregnancy complications [n (%)] | |
No | 2606(82.6) |
Yes | 548(17.4) |
Father’s characteristics | |
BMIk, g/m2 [Mean (SD)] | 23.3(3.6) |
Education level [n (%)] | |
Junior high school or below | 454(14.4) |
Senior middle school | 873(27.7) |
Junior college | 860(27.3) |
Bachelor degree or above | 961(30.5) |
Child characteristics | |
Sex [n (%)] | |
Female | 1600(50.7) |
Male | 1554(49.3) |
Gestational age, wk [Mean (SD)] | 39.0(1.3) |
Birth weight, g [Mean (SD)] | 3363.2(446.6) |
Breastfeeding duration (months)# [n(%)] | |
< 1 | 1298(42.5) |
1–5 | 787(25.7) |
≥ 5 | 972(31.8) |
# There were missing values for the variables. |
Maternal pre-pregnancy BMI was measured at the first antenatal visit. Maternal age, residence, race, parental education levels, family monthly income per capita, parity, and father’s BMI were collected by questionnaire during recruitment. Medical notes were used to extract data on maternal metabolic disorders, parity, children's sex, birth weight and gestational age. Maternal metabolic dysfunctions included hypertensive disorders during pregnancy, gestational diabetes, and other metabolic dysfunctions (thyroid dysfunction, severe anemia, and polycystic ovary syndrome), and having one or more of these disorders in a woman was defined as having maternal metabolic dysfunctions. Three-monthly questionnaires were used to gather information on alcohol and tobacco use during pregnancy.
In addition, we gathered data on the data of breastfeeding duration (months) for additional sensitivity analysis.
2.7. Statistical Analysis
SPSS was used for all analyses (IBM, version 23.0). Statistical tests were conducted on a two-sided basis, with statistical significance defined as a P-value < 0.05. Participants’ demographic characteristics were reported as the mean ± standard deviation (SD) or percentage (n %).
Several analyses were carried out to determine the relationship between maternal anxiety and offspring’s physical development. Considering the impact of prenatal anxiety on offspring’s physical development and the possible critical period or the cumulative effect of long-term effects, we conducted the following categorical exposure analysis. First, we classified four categories according to the presence or absence of pregnancy-related anxiety: 1st trimester (yes vs. no), 2nd trimester (yes vs. no), 3rd trimester (yes vs. no), and all three trimesters (yes vs. no). Second, we explored the association of new-onset anxiety in the 2nd trimester (vs. no anxiety in first two trimesters) and new-onset anxiety in the 3rd trimester (vs. no anxiety in all three trimesters) with offspring’s physical development. Third, taking into account the results of the previous analysis, we performed a post hoc analysis—of the association between simultaneous 2nd- and 3rd-trimester anxiety (vs. no anxiety both in the 2nd- and 3rd trimesters) and offspring’s physical growth.
Binary logistic regression models were used to explore the association between maternal antenatal anxiety and infant RGW from 0–12/24/36 months.
Longitudinal analyses were performed using generalized estimating equations (GEEs) with random intercepts for each subject to account for the correlation between repeated observations within subjects between exposure to pregnancy-related anxiety and offspring’s physical growth (BMI and BF) from 48 months to 72 months of age. Furthermore, considering the characteristics of children’s early physical development, we used multiple linear regression model to compare the anthropometric outcomes (BMI and BF) of offspring exposed and not exposed to pregnancy-related anxiety at each follow-up time point. And the results are reported as unstandardized B coefficients with 95% CIs after controlling for covariates.
For the BMI and BF trajectory, logistic regression models were used to examine the relationship between maternal antenatal anxiety and children’s BMI/BF trajectories. The normal BMI/BF trajectory was used as the reference group.
Finally, to test the reliability of our findings, we performed two sensitivity analyses. First, maternal anxiety during pregnancy and children’s subsequent anthropometric growth may be mediated by gestational age and birth weight. Therefore, we further adjusted the birth weight z-score in the sensitivity analysis. Second, breastfeeding duration is an important protective factor for children with overweight/obesity. To improve the accuracy of the results, it was taken into consideration as precision variable.