Table 2 Summary of study findings.

Weight gain of infants born to employed mothers was lower than that of infants born to unemployed mothers from birth to 6 months corrected age.

• Weight gain velocity diverged between infants of employed and unemployed mothers at 3 months, resulting in a maximum weight difference of approximately 900 g at 5 months.

• Weight gain velocity of infants of employed mothers increased between 5 and 6 months of age, resulting in a narrowed weight difference of approximately 400 g at 6 months.

There was a significant difference in weight gain over time between infants of employed mothers and infants of unemployed mothers (p < 0.001). There was no significant difference in change in length or head circumference over time between the 2 groups.

In univariate analyses, a low level of maternal education, but not low family income, was associated with head circumference growth restraint compared with the median population growth at 1 year.

• The Odds Ratio of head circumference growth restraint for infants of mothers with a low education level compared to a high education level was 5.3 (95% CI 1.4 –20.8) in the multivariable logistic regression adjusting for gestational age, family income, smoking during pregnancy, maternal age, conception by IVF or ICSI, infant sex, multiple pregnancy, and breastfeeding during the first 6 months of life.

Maternal education level did not have a significant effect on the likelihood of growth deficit at 2 weeks following discharge from the neonatal unit.

• Simple linear regression analysis showed no statistically significant difference in the odds of growth deficit in infants born to mothers with primary, secondary and tertiary level education compared to infants born to mothers with no education.

At all time points considered (7, 11, 16 and 23 years), the height deficit between low birthweight (LBW, < 2500 g) and appropriate birthweight (ABW, ≥  2500 g) infants was lower in preterm infants in social classes I & II (fathers in professional and intermediate occupations) than those in social classes IV & V (fathers in partly manual and unskilled manual occupations).

• This difference in height deficit between preterm infants in different social classes was not analysed using statistical tests.

• There appears to be a stepwise increase in height deficit between ABW and LBW infants at ages 7, 11 and 23: preterm infants in social classes I & II have the lowest height deficit, followed by preterm infants in social classes IV & V, term infants in social classes I & II and finally term infants in social classes IV & V.

• The height deficit in preterm infants in social classes I & II decreases during puberty between 7 and 16 years, whereas it increases for preterm infants in social classes IV & V.

Using Student’s t tests to compare the weekly growth of infants of well educated mothers with infants of less well educated mothers did not show significant differences for any of the time periods.

2. Between birth and 3 months, and 3–7 months, infants of well educated mothers had the highest mean weight increments of all the different infant categories: infants with well educated and less well educated mothers; infants with and without intrapartum foetal acidosis (scalp pH < 7.2); infants with and without neurodevelopmental disorders; infants with GA < 34 weeks and ≥ 34 weeks; male and female infants; all preterm infants in the cohort.

• No formal statistical analysis (e.g. ANOVA) was used to compare multiple infant categories.

3. Between birth and 3 months, 3–7 months and 7–12 months, infants of well educated mothers had the highest mean length increments of all infant categories.

• No formal statistical analysis (e.g. ANOVA) was used to compare multiple infant categories.

In a multivariable logistic regression analysis, Maternal education, the Home Facilities rating, and the Financial Situation rating, were significantly associated with head circumference (HC) catch-up (p = 0.012, p = 0.027, p = 0.001 respectively). Compared to infants without HC catch-up, those with HC catch-up were:

• More likely to have mothers with education ≥ 12 years.

• More likely to have a Home Facilities rating of Adequate and less likely to have a rating of Inadequate or Good.

• More likely to have a Financial Situation rating of Good, and less likely to have a rating of Enough or Inadequate.

While assessing the impact of differential loss to follow-up by socioeconomic status in the Discussion, the authors state, “Lower SES (socioeconomic status) was neither related to birth weight nor outcomes in our study but was associated with greater weight gain from birth to term age.” This association is not mentioned elsewhere and the relevant growth data are not reported.

2. Parental occupational SES moderated the association between birthweight SD score and BMI at 19 years: among participants with lower parental socioeconomic status, there was an inverse association between birthweight z-score and BMI at 19 years which remained significant after adjusting for sex, maternal age and maternal smoking during pregnancy. There was no such association in the higher socioeconomic status group.

3. Table S1 in the Supplementary Information shows the distribution of BMI at 19 years of age in infants with low and high parental occupational SES. Although there is no formal statistical comparison, those with parents of lower SES appear to have higher BMIs, and are particularly overrepresented among participants with BMI > 30.

1. Analysis of variance showed a significant difference in the weight, length and head circumference of infants at 12 months according to average monthly household income (all p < 0.001). The reported growth measurements were as follows:

> 5000 CNY

10.3 (1.39)

73.7 (3.7)

0.9 (1.6)

2. Multiple stepwise regression analyses showed that average monthly household income was significantly associated with weight (p = 0.001), body length (p = 0.029), and head circumference (p < 0.001) at 12 months after adjusting for other variables which were significant in univariate analyses, i.e. birthweight, feeding mode (breastfeeding, no breastfeeding or mixed feeding at 12 months) and sex for weight; intervention group (family integrated care vs usual care), gestational age, birth length and sex for body length; and intervention group, gestational age, birth HC and sex for HC.

1. Chi squared analyses showed no significant association between maternal education level or maternal occupation and the presence of overweight/obesity at 4 to 7 years corrected age.

1. Chi squared analyses showed no significant association between maternal education level and the occurrence of growth failure during any of the three study periods (40 weeks to 4 months, 4 to 8 months, 8 to 20 months).

1. Chi squared analysis showed no significant association between maternal education and the prevalence of subnormal head circumference at school age (mean age 6.8 years corrected age +/− 0.9 years).

1. Chi squared analysis showed a significant association (p < 0.05) between maternal education and the incidence of failure to thrive (FTT) up to 36 months corrected age. Mothers of infants with FTT were less likely to have some college education but not a college degree, and more likely to have a college degree or postgraduate education than mothers of infants without FTT.

2. Chi squared analysis showed no significant association between family income and the incidence of FTT in the cohort.

3. A multivariable logistic regression analysis showed that maternal education ≥ college graduation compared some college education without a college degree was significantly associated with an increased risk of the infant developing FTT (p < 0.005) after adjusting for small for gestational age at birth, abnormal or suspect neurologic exam at birth, birthweight, maternal age, maternal height, and whether or not the infant lived with their father.