SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Objectives

Study effect of maternal gestational weight gain (GWG) on the relationship between maternal menarcheal age (MMA) and child growth and overweight risk and examine socio-demographics characteristics of excessive GWG.

Design and Methods

The relationships between GWG and MMA in 54,184 women and their children's growth trajectories during first 5 years of life (2000-2005) in south China were tested using longitudinal data analysis with mixed models and logistic regression.

Results

Average MMA was 14.8 (1.3) years; 36.3% of the women had excessive GWG. Excessive GWG interacted with adverse effects of early MMA (if ≤ 13 years), leading to the most rapid growth in offspring and highest risk of overweight at age 4-5 (OR = 5.2 [2.0-13.5]) than others. Women with early menarche, high-education, urban residence, and a routine job were more likely to have excessive GWG than the others.

Conclusions

GWG modify the association between early MMA and offspring's growth and overweight. Controlling for GWG may reduce the adverse influence of early MMA and its own adverse influence on childhood health.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Women with early menarcheal age are reported more likely to have excessive gestational weight gain (GWG) during pregnancy [1] and their children have higher risks of rapid infancy growth and childhood obesity [2, 3] Socio-demographic characteristics related to insufficient and excessive GWG were observed [4, 5]. Although early maternal menarcheal age (MMA) is undesirable, controlling GWG during pregnancy may intervene adverse effects of early MMA on offspring child growth. However, no previous studies have tested the complex interaction effects between MMA and GWG on growth and obesity of the offspring children.

This study examined: (1) Whether GWG modified the effect of MMA timing on the offspring's growth in early childhood; and (2) What women groups were more likely to have excessive GWG.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Study design and subjects

Rich data collected in Jiaxing (a middle-income county) in Southeast China, as part of a large population-based health surveillance system established in four provinces of China in 1993 were used [6]. Among 160,459 enrolled women during 2000-2005, 55,785 recorded their weight at the premarital and the last check-ups during weeks 37-43 of their pregnancies (preterm birth was excluded). Data of 54,184 women and their offspring's growth trajectory from birth to age 5 (total number of children's visits: 523,096) were analyzed after excluding twins, gynadromorphism, and extreme values of child growth referring to World Health Organization (WHO) Child Growth Standards [7].

Assessments and key study variables

Obstetricians examined the women's health before, during and after pregnancy, and filled individual booklets with demographics as well as prenatal care, labor, and delivery information. After birth, the children's growth was regularly followed till age 6 years.

Based on self-reported MMA, the mothers were separated into early (≤13 years), intermediate, and late groups (≥16 years) by the 25th and 75th percentiles (only 0.2% had <12 years). The late group was our reference for analyzing the effect of MMA, since there was a dose-response relationship between MMA and infancy-to-childhood growth.

GWG was calculated as the difference in weight between pre-pregnancy and the last check-up during 37-43 weeks of gestation. GWG was categorized as insufficient, normal, and excessive based on recommendations from the Institute of Medicine (IOM, 2009) for underweight, normal, and overweight women defined by WHO [8].

Covariates included maternal socio-demographic characteristics (e.g., age, education, occupation, and urban-rural residence), age at pregnancy, parity, anthropometrics (height, weight, BMI, and overweight at pre-pregnancy defined as BMI ≥ 25), alcohol drinking, and smoking. Alcohol drinking, smoking, and the history of diabetes or pregnancy-induced hypertension were not included in our final models due to their low prevalence (<0.6%).

Macrosomia, growth trajectory (Z scores of BMI, weight, and height) during the first 5 years, being at-risk of overweight (BMI Z score ≥ 1), and overweight (BMI Z score ≥ 2) according to 2006 WHO Child Growth Standards [7], and rapid weight gain during infancy (the difference of weight Z scores between age 2 and birth >0.67) were used as indicators of childhood health status due to their associations in the development of obesity [9].

Statistical analysis

Using logistic regression and ANCOVA, we tested the risk of macrosomia, rapid infancy growth, and overweight at age 4-5 associated with MMA and GWG. Spline-mixed models with random intercept were fitted to study period-specific childhood growth status by MMA–GWG groups, due to the nonlinear nature of childhood growth. Related covariates (see above) were adjusted for. Our modeling strategies were described elsewhere [10]. General characteristics of pregnant women were tested by chi-square test, t-tests, and multinomial logistic regression models. All analysis was conducted using SAS 9.2 (SAS Institute, Cary, NC).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Effects of GWG and MMA on offspring's growth in early childhood

These women's average menarcheal age (MA) was 14.8 (SD = 1.3) years; their mean age at pregnancy was 24.2 (SD = 2.7) years; mean BMI was 20.1 (SD = 2.1) kg/m2; <1% were obese at pre-pregnancy; 36.3% had excessive GWG. The majority had first-time pregnancies (97.1%), lived in rural areas (87.0%), and had less than a high school education (74.0%). Of their offspring, 2.3% were overweight, 0.5% were obese at 4-5 years old.

At birth, children with excessive GWG were significantly larger in BMI, weight, and height than others; they were more likely to experience macrosomia and faster growth in weight (excessive vs. normal group; macrosomia OR = 2.9 [2.6-3.2], rapid weight growth OR = 1.2 [1.1-1.3].

Dose-response relationships between infancy-childhood growth and MMA existed. The early MMA group had the highest levels of birth weight, ponderal index at birth, weight Z-score change during the first 2 years of life, and weight and BMI Z-score in childhood, but the late MMA group had lowest values; the intermediate MMA group had the middle of them. Children with early MMA (≤ 13 years) were more likely to be macrosomial (OR = 1.5 [1.3-1.7]), rapid growers (OR = 1.3 [1.2-1.4]), and overweight (OR = 1.4 [1.0-1.9]) than those with late MMA (≥16 years). The risk of adverse child growth in the intermediate MMA group was less than the early MMA groups (data not shown). After adjusting for GWG, MMA was still associated with children's BMI, weight and height at birth, and their rapid weight growth during the first 2 years of life.

Influence of GWG on the association between MMA and offspring's rapid growth in early childhood

Children with early MMA and excessive GWG were more likely to have rapid weight gain during their first 2 years of life (OR = 1.6 [1.4-1.8]) and overweight at age 4-5 (OR = 5.2 [2.0-13.5]) (Figure 1). They kept the highest average BMI Z-score, weight Z-score, and height Z-score consistently from birth to age 5 (Figure 2). The growth curves of the intermediate MMA group for BMI, height, and weight were located between the early and late MMA groups; they were not shown in the figure.

image

Figure 1. The risk (OR and 95% CI) of rapid growth during first 2 years of life and overweight at age 4-5 in Chinese children was associated with maternal menarcheal age (MMA) and gestational weight gain (GWG) (a) Risk of Rapid Weight Gain during First 2 years by MMA and GWG. Rapid weight gain during infancy was defined by weight Z-score change during the first 2 years > 0.67 (n = 33, 979); logistic regression was used after adjusting for gestational age, birth weight, breast feeding during first 6 months, maternal height, parity, age at time of pregnancy, residence, occupation and education. (b) Risk of overweight at age 4-5 by MMA and GWG. Childhood overweight was defined as weight Z-score at age 4-5 ≥ 2 (n = 4,124); logistic regression was used after adjusting for gestational age, birth weight, breast feeding during first 6 months (yes vs. no), maternal height, parity (first pregnancy vs. others), age of pregnancy, residence (urban vs. rural), occupation (unemployed, housework or temporary work vs. routine jobs), and education (≥high school vs. less). Number of (El) MMA and (I) GWG was not sufficient for logistic regression of overweight risk at age 4-5. Abbreviation: Lt (late); Im (intermediate); El (early); N (normal); I (insufficient); E (excessive).

Download figure to PowerPoint

image

Figure 2. Mean levels of growth of BMI, weight, and height Z-score during first 5 years of life by gestational weight gain (GWG) and maternal menarcheal age (MMA). The primary spline mixed model with random intercept without covariates and their interactions was as follows: Yij = Σβai (ageijc) + βbi (MMAk) + βci (GWGl) + εij. Note: Yij was the continuous measurements of BMI, weight, and height in child i at measurement j; the term (ageij – c) with age interval C ∈ {0, 3, 6, 9, 12, 18, 24, 30, 36}; the term (MMAk) with the case of k equaled 1 if the MMA of the children's mother was in the kth category (k ∈ {≤13 year, 14, 15, 16 ≥ year}), otherwise equaled 0; the term (GWGl) with the case of l = 1 if the GWG of the children's mother was in the lth category (l ∈ {insufficient, normal, excessive}), otherwise equaled 0; and εij was the residual error. The model was adjusted for sex, gestational age, birth weight, exclusive breast feeding during first 6 months (yes vs. no), parity (first pregnancy vs. others), maternal age at pregnancy and height, education (≥high school vs. less) and occupation (unemployed, housework or temporary work vs. routine jobs), and urban residence (urban vs. rural). Growth in (a) BMI, (b) weight, and (c) height Z-scores for age. Abbreviation: N (normal); E (excessive). Bar was used for standard error of each adjusted mean levels. All P-values of group differences at every age < 0.001. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Predictors of excessive GWG

Table 1 shows those women with early MA were highly educated, had a routine job, were pregnant for the first time, and lived in urban areas were 1.3-1.4 times more likely to have excessive GWG, while they were 0.6-0.9 times less likely to have insufficient GWG (P < 0.05). GWG had a linear relationship with MA (β [SE] = −0.40 [0.02] kg per year, P < 0.01 adjusted for covariates. The early MA and excessive-GWG groups have similar demographic characteristics: tended to be highly educated, are older at pregnancy, have a routine job, and have an urban residence (P <0.05).

Table 1. Association between gestational weight gain (GWG, outcome variable) and pregnant women's socio-demographic characteristics in southeast China (n = 54,184)
Independent variablesBased on linearBased on multinomial logistic models (odd ratio (95% CI))
(maternal characteristics)regression modelb (β (SE))Insufficient vs. normal GWGExcessive vs. normal GWG
  1. Abbreviation: MA (menarcheal age). Linear regression and multinomial logistic regression models were fit. GWG was grouped by Institute of Medicine (IOM, 2009) and World Health Organization (WHO) reference; the normal range of GWG were 12.5-18 kg in underweight (pre-pregnancy BMI < 18.5 kg/m2), 11.5-16 kg in normal (pre-pregnancy BMI: 18.5-24.9 kg/m2) and 7-11.5 kg in overweight (pre-pregnancy BMI ≥ 25.0 kg/m2); those lower than the lower limit was defined as insufficient GWG; those higher than the upper limit was defined as excessive GWG.

  2. a

    Variables of maternal characteristics were transformed as binary responses: residence (urban vs. rural), education (≥ high school vs. less), occupation (unemployed, housework, or temporary work vs. routine jobs), and parity (first pregnancy vs. others).

  3. b

    Every P value for the coefficients from the linear regression model was <0.001.

Having MA at ≤13 year0.7 (0.1)0.9 (0.8-0.9)1.3 (1.2-1.3)
Pre-pregnancy overweight (BMI ≥ 25)−3.3 (0.2)1.5 (1.3-1.8)2.3 (2.0-2.6)
First pregnancy1.8 (0.1)0.7 (0.6-0.7)1.2 (1.1-1.4)
Had ≥ high school1.5 (0.1)0.7 (0.6-0.7)1.3 (1.2-1.3)
Living in urban areas1.8 (0.1)0.7 (0.7-0.8)1.4 (1.3-1.5)
Having a routine job1.8 (0.1)0.6 (0.6-0.7)1.4 (1.3-1.4)

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Based on our large sample and longitudinal data analysis, GWG and MMA affect child growth and obesity. GWG modify the association between maternal early menarche and offspring's growth and overweight. Those children with early MMA and excessive GWG had the highest risks of rapid infancy growth and childhood overweight at age 4-5. Controlling for weight gain during pregnancy may reduce the adverse influence of early MMA and its own adverse influence on childhood health.

Ours and other recent research added evidence on intergenerational effects on early puberty and obesity, such as associations between timing of menarche and offspring's growth, effects of maternal BMI on daughter's age at menarche [11] and similarity of early puberty timing [3]. However, little is understood about the mechanisms of these intergenerational relationships. While heritable factors were estimated to explain 70-80% of variance in pubertal timing [12], individual variations of pubertal development (about 1.7 years) were still remained [13].

Using data from 54,184 women in China, our findings shed light into in-utero modification on intergenerational relationships between MMA and the offspring's growth, though less attention had been given to their possible interactions with GWG. Evidence of adipocytokine researches support that the in-utero environment is a critical period in children's growth. Increased maternal leptin levels during pregnancy enhance the mobilization of maternal fat stores and energy transfer to fetal growth [14] and it may cause childhood rapid growth and obesity development. Thus, the risk of children's rapid growth and obesity by early MMA can be changed by controlling GWG. We found significant MMA effects on children's growth trajectory, although the MA was later than that in Western countries, only 0.2% of these Chinese women had menarche at <12 years, while it was 21.6% in US (2) and 19.6% in UK (3). Also 35.9% of these women had excessive GWG. Chen et al. [15] reported concerns about excessive GWG in recent China. Effort is needed to educate pregnant women and their families regarding optimal GWG for mother and their offspring's wellbeing.

Socio-demographic factors seem affecting GWG and MA timing. The older, well educated, those had a routine job, and lived in urban areas were more likely to have excessive GWG and early MA. These were consistent with findings from developed countries [1, 4, 5] although the ethnic and geographic diversity may exert some influence on GWG. Social environmental changes including declined physical activity [16], more consumption of energy from dietary fat [17], and rising obesity prevalence in urban areas of China could lead to further menarche advancement and excessive GWG, based on related research findings from other developing countries [18, 19].

In conclusion, our findings shed light into in-utero modifications of intergenerational relationships between MMA and the offspring's growth, which is not solely genetically determined. Intervening in the socio-demographic factors that contribute to excessive gestational weight gain might reduce the adverse effect of early maternal menarche on offspring's obesity risk. The mechanisms behind these intergenerational associations require further research.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References