Table 1 Summary of study characteristics.

Ahn ¹⁶

South Korea: Inborn infants at a University hospital in Incheon.

Secondary analysis of prospectively collected growth data for a cohort of preterm infants.

Preterm infants born after 32 weeks’ gestation.

n = 238.

Mean GA ³⁶34.8 weeks (1.1).

Range GA: 32–-36 + 6 weeks.

Outborn infants, “socially vulnerable” infants (e.g. those with unmarried mothers, adopted infants), infants with congenital malformations that affect physical growth.

Maternal employment status: Yes/No. Methods for ascertaining maternal employment status not reported.

Weight, length and head circumference measured from birth to 6 months corrected age. Measurements were taken directly either during NICU admission or at outpatient follow-up. Where direct measurement was not possible, growth data were collected from the records.

1. Effect of maternal employment status on variation in weight, length and head circumference growth by 6 months corrected age analysed using a generalized linear model and reported.

2. Graphs of weight gain from birth to 6 months corrected age shown separately for infants of employed and unemployed mothers.

Bocca-Tjeertes ¹⁷

Netherlands: Community-based sample of participants recruited at a routine 4-year well-child visit at Preventive Child Health Care Services (PCHCs).

Secondary analysis of a subgroup of the Lollypop cohort study, with data collected retrospectively from medical records from birth to 4 years, and additional measurements taken at 4 years. The Lollypop study collected data on the growth, development and general health of preterm infants.

Early and moderately preterm infants born before 36 weeks’ gestation.

n = 1123 (infants with growth data available).

Mean GA: 34 weeks.

Range GA: 32 – 35 + 6 weeks.

Infants with congenital malformations and syndromes, infants for who gestational age was not verified “beyond reasonable doubt”.

Maternal education level: low, moderate/high (definitions not reported).

2. Family income: low, moderate/high (definitions not reported).

Maternal education level and family income ascertained through a parental questionnaire administered at the 4-year well-child visit.

Weight and height measurements from birth to 4 years, and head circumference measurements from birth to 12 months. Measurements were obtained from hospital records and records of routine well-child visits at the PCHCs. SD scores for weight at birth, height and head circumference at birth were calculated using the Kloosterman ⁹²and Usher and McLean ⁹³curves respectively. Growth restraint was defined by measurements 2 SD scores below the median growth of the Dutch population, deduced from the fourth nationwide Dutch growth survey with measurements of children born shortly before 1997.

Postnatal growth restraint at 4 years (for weight and height) and 1 year (for head circumference) reported separately according to maternal education level and family income.

Sammy ¹⁸

Kenya: Preterm infants born at the county referral hospital for the former Kitui and Mwingi districts.

Prospective cohort study of preterm infants discharged from the Kitui District Hospital newborn unit, followed up 2 weeks post-discharge.

Preterm infants born at less than 37 weeks’ gestation and discharged from the newborn unit between April and June 2014.

n = 106 (infants with growth data at 2 weeks).

Mean GA and range of GA not calculable as lower limit of gestational age not reported.

More than 50% of infants born between 33 and 36 weeks.

Infants with congenital anomalies likely to interfere with their growth, or with guardians declining to participate in the study.

Maternal education level: no education, primary education, secondary education and tertiary education.

Information on maternal education obtained via a questionnaire-guided interview with guardians at discharge.

Weight, length and head circumference were measured at discharge from the newborn unit, and at 2 weeks post-discharge. Length of hospital stay varied from 2 to 30 days, and therefore time of follow-up was varied.

Number of infants with “optimal growth” (not defined) and “growth deficit” (not defined) reported according to maternal education level.

Teranishi ¹⁹

United Kingdom: children born across England, Scotland and Wales.

Birth cohort study, with a wide range of data collected from parental interviews, medical examinations, school questionnaires, personal questionnaires and attainment tests.

White singleton infants born 3–9 Mar 1958 who had perinatal data and data at 23 years available. Preterm birth was considered a dichotomous variable (GA < 37 weeks vs ≥ 37 weeks).

Number of preterm infants unknown.

Mean GA and range of GA not calculable.

Unknown.

Social class coded by father’s occupation according to the Registrar General’s Classification (I: professionals, II: intermediate, IIINM: skilled non-manual, IIIM: skilled manual, IV: partly manual, V: unskilled manual).

Questionnaire administered to parents at time of birth used to ascertain father’s occupation.

Height was measured at 7, 11 and 16 years. Height was self-reported at 23 years.

Height deficit was defined as the difference in height between low birthweight (LBW, birthweight < 2500 g) and appropriate birthweight (ABW, birthweight ≥ 2500 g) children calculated (mean height of ABW children – mean height of LBW children.)

Graphs of height deficit at ages 7, 11, 16 and 23 are shown separately for preterm infants in social classes I and II and those in classes IV and V.

Holmqvist ²⁰

Sweden: preterm infants born at a University hospital in Lund.

Prospective cohort study of preterm infants followed up until 48 months of age.

Low risk, inborn preterm infants (GA < 37) born Jan-Dec 1979: vaginal delivery, singleton pregnancy, cephalic presentation, preterm labour of unknown aetiology, no additional maternal complications (e.g. maternal diseases, signs of preeclampsia, anaemia or vaginal bleeding), no additional foetal complications (biparietal diameter within 2 SD, “normal” symphysis-fundus height). n = 35.

Mean GA = 34 weeks.

Maternal complications (e.g. maternal diseases, signs of preeclampsia, anaemia or vaginal bleeding), foetal complications (biparietal diameter outside 2 SD, symphysis-fundus height not considered “normal”).

Maternal education: well educated mothers (> 11 school years), less well educated mothers (< 11 school years).

Questionnaire administered at one of the follow-up visits (timing not specified).

Weight, length and head circumference measured at birth, and at 3, 7, 12, 24 and 48 months. Weekly individual growth increments for weight, length and head circumference calculated by dividing difference between 2 consecutive measurements by the number of weeks in between for 5 time periods: birth to 3 months, 3-7 months, 7-12 months, 1-2 years, 2-4 years.

Weekly growth increments in weight, height and head circumference for each timer period compared between infants with well educated mothers and infants with less well educated mothers.

Ghods ²¹

Austria: preterm infants admitted to a NICU at a University hospital in Vienna.

Retrospective cohort study of preterm infants followed up until 66 months of age.

All preterm (GA < 37) or VLBW (birthweight < 1500 g) infants admitted to the NICU born between 2000 and 2002. n = 173.

Mean GA of infants with head circumference (HC) catch-up ⁹² = 28.04 weeks (2.6 weeks).

Mean GA of infants without HC catch-up ⁹² = 29.53 weeks (2.3 weeks).

Infants lost to follow-up, infants whose medical records were incomplete, infants whose head circumference was not measured and could not be interpolated at 1 year corrected age.

Maternal education ≥ 12 years: Yes/No.

2. Home Facilities: Good, Adequate, Inadequate. Assessed based on indirect questioning and assessment of home environment by the interviewer (e.g. light, ventilation, accommodation).

3 Financial Situation: Good, Enough, Inadequate. Estimated based on education level (parent not specified), occupation and social facilities.

Weight, height/length and head circumference (HC) were measured at birth, during the NICU stay and at 3, 6, 9, 12, 24, 40, 54 and 66 months corrected age. Z-scores for measurements taken during the NICU stay were calculated using the data from Fenton et al⁹⁴. Z-scores for measurements taken after discharge were calculated using the WHO 2006 growth standards⁹⁵. HC catch-up defined as an increase of 0.67 Z-score units between birth and 1 year corrected age.

Parental socioeconomic characteristics compared between infants with and without HC catch-up.

Ni ²²

United Kingdom: preterm infants born across the UK and Ireland.

Prospective cohort study of a birth cohort of preterm infants followed up until 19 years of age.

Preterm infants born at < 26 weeks’ gestation between Mar and Dec 1995 and enrolled in the EPICure study who were alive at 16 years and consented to assessment at 19 years. n = 129.

Mean GA of infants with metabolic syndrome at 19 years ⁹² = 24.95 weeks (0.78 weeks).

Mean GA of infants without metabolic syndrome at 19 years ⁹² = 24.76 weeks (1.21 weeks).

Participants with data missing for BMI and central systolic BP at 19 years, participants who died or were lost to follow-up before assessment at 19 years.

1. Maternal education: A level or above, GCSE or below.

2. Parental occupation: Higher (non-manual employment), Lower (manual employment or unemployed).

Parent questionnaire administered when children were 2.5 years old.

Weight was measured at birth, 40 weeks postmenstrual age, and at 2.5, 6, 11 and 19 years. Height was measured at 19 years. Waist and hip circumferences were measured at 19 years.

Z- scores for weight were calculated at each age based on UK population norms. Change in weight z-scores were calculated for birth to term, term to 2.5 years, 2.5 to 6 years and 6 to 11 years. BMI at 19 years was calculated. The waist:hip ratio was calculated at 19 years.

BMI distribution at 19 years is plotted separately for participants with lower (occupation-based) parental socioeconomic status and higher parental socioeconomic status.

Liang ²⁵

China: Preterm infants born at a teaching hospital in Nantong (Jiansu province).

Retrospective cohort study of preterm infants followed up until 18 months, some of whom received the intervention of family integrated care (FIcare group) and some of whom received NICU traditional care (control group) during the NICU admission.

Preterm infants born at > 28 weeks’ and < 34 weeks’ gestation between Jan 2018 and Sep 2020, with birthweight 1000 – 2499 g, admitted to the NICU within 24 hours after birth, with parents possessing basic reading and comprehension skills and the ability to care for their infant. n = 215.

Mean GA for infants in FIcare group ⁹² = 30.03 weeks (1.38 weeks).

Mean GA for infants in control group ⁹² = 29.93 weeks (1.30 weeks).

Infants with congenital genetic metabolic diseases, digestive tract malformations, severe congenital heart disease, central nervous system and endocrine diseases, or other severe congenital growth and development abnormalities; infants who received invasive respiratory therapy; infants who required surgery; infants who received comfort treatment; and infants with parents who had previous mental health history.

1. Primary caregiver education: below junior school, high school and technical secondary school, college degree or above.

2. Average monthly household income: < 3000 CNY, 3000 – 5000 CNY, > 5000 CNY.

Questionnaire administered to parents by researchers (timing and setting of administration not specified).

Weight, length and head circumference measured at 1, 3, 6, 12 and 18 months.

Weight, length and head circumference of infants at 12 months are reported according to average monthly household income, and compared between infants of households in different monthly income categories.

Fu ²⁴

China: Preterm infants born across Jianxing (Zheijang province).

Prospective cohort study of a birth cohort of children followed up until school age.

Singleton infants born before 37 weeks’ gestation and enrolled in the Jianxing Birth Cohort between 1999 and 2013. n = 2125.

Median GA = 36 weeks (IQR 35 – 36 weeks)

Mother-child pairs without complete data for the detection of childhood overweight/obesity at 4 – 7 years corrected age, or missing data for any variables considered in the study.

1. Maternal education: < High school, High school, > High school.

2. Maternal occupation: Farm work/housework, Routine job, Temporary work, Others.

Questionnaire administered to mothers in local clinics via interview during enrolment in the cohort and during pregnancy.

BMI was calculated at 4 – 7 years corrected age. BMI z-scores were calculated at corrected ages between 4 and 5 years using the 2006 WHO Child Growth Standards⁹⁶; overweight and obesity were defined as BMI z-scores between 2 and 3, and > 3 respectively. The 2007 WHO Child Growth Standards ⁹⁷were used for children older than 5 years corrected age; overweight and obesity were defined as BMI z-scores between 1 and 2, and > 2 respectively Although not specified by the authors, height and weight were presumably measured at 4 – 7 years corrected age.

Numbers of children with and without obesity at 4 – 7 years corrected age (mean age 6.8 years +/− 0.9 years) reported for each maternal occupation and maternal education category. Distributions of maternal occupation and maternal education were compared between infants with and without overweight/obesity at 4 – 7 years corrected age.

Sices ²⁶

United States: Preterm infants born at a teaching hospital in Cleveland.

Retrospective analysis of growth and developmental data from a cohort of extremely low birthweight infants followed to 20 months corrected age.

Infants with birthweight 501 – 1000 g born between 1997 and 1999, and admitted to the NICU, who survived until 20 months corrected age. n = 154.

Mean GA ⁹² = 25.9 weeks (1.8 weeks).

Infants with congenital malformations known to have a direct impact on growth, infants without a minimum of two consecutive growth measurement points, infants discharged after the age of 49 weeks.

Maternal education less than high school: Yes/No.

Weight, length and head circumference (HC) were measured at clinic visits at 40 weeks corrected age (39.2 + /− 3.0 weeks), and at 4 (4.4 + /− 0.7), 8 (8.6 + /− 1.0) and 20 months (19.0 + /− 1.2) months corrected age. Weight z-scores were calculated at birth and at 40 weeks using the sex-specific standards of Kramer et al⁹⁸. At birth and 40 weeks, length and HC z-scores were calculated using the Usher and McLean curves⁹³. At 4, 8 and 20 months corrected age, and for infants older than 43 weeks at the first post-discharge follow-up, weight, length and HC z-scores were calculated from US CDC sex-specific normative data⁹⁹. Growth failure defined as a decrease in weight z-score of over 0.67 during any of the three study periods: 40 weeks to 4 months, 4 to 8 months, and 8 to 20 months.

Number of infants with growth failure in each period and the number without growth failure in any period are reported for mothers with less than high school education (and are deducible for mothers with high school education or greater). Distribution of maternal education compared between infants with growth failure in a period, and infants without growth failure in any period.

Peterson ²⁷

United States: Preterm infants born at a teaching hospital in Cleveland and two other tertiary centres.

Analysis of data from two cohorts of very low birthweight infants: 1) recruited at birth and followed up prospectively until school age, 2) recruited at early school age with retrospectively collected perinatal data.

Infants with birthweight < 750 g born between 1982 and 1986 and a matched group of infants with birthweight 750–1499 g from the same hospital and with the same sex, race and birth date within 3 months. n = 128.

Mean GA ⁹² = 27.6 weeks (2.8 weeks).

Missing or unreliable head circumference measurement at school age.

Maternal education less than high school: Yes/No.

Height, weight and head circumference (HC) measured at the school age follow-up at a mean age of 6.8 years (+/− 0.9 years).

HC z-score at school age was calculated using Roche et al’s reference data ¹⁰⁰.

Subnormal HC was defined as a measurement more than 2 SD below the mean for age.

Numbers of infants with and without subnormal HC at school age reported separately for mothers with less than high school education (and are deducible for mothers with high school education or greater). Distribution of maternal education compared between infants with and without subnormal HC at school age.

Kelleher ²³

United States: Preterm infants born at 8 medical centres associated with medical schools across the US.

Analysis of data from the Infant Health and Development Program (IHDP), a national multicentre randomised controlled trial to investigate the effect of early intervention on the cognitive, behavioural and health status of preterm low birthweight infants followed from birth until 3 years of age.

All infants enrolled in the IHDP. Infants born between Jan-Oct 1985 at the 8 participating sites at ≤ 37 weeks’ gestation with birthweight ≤ 2500 g were eligible for the IHDP. n = 771.

Mean GA according to failure to thrive (FTT) case status:

FTT infants: (n = 180) = 33 weeks.

Non-FTT infants: (n = 591) = 33.1 weeks.

Infants who lived outside of the catchment area, infants discharged outside the recruitment time period, infants who died within the first 48 hours of life, all triplets, all quadruplets, and twins of ineligible children. Only one infant from each pair of eligible twins was included. Maternal drug or alcohol abuse, maternal inability to communicate adequately in English, maternal report of psychiatric hospitalisation. Hospitalisation longer than 60 days after 40 weeks corrected age, oxygen support for more than 90 days, severe neurologic abnormality, severe sensory deficit or chromosome-multiple anomaly syndrome. Lack of parental consent, or family refusal of group assignment. Infants lost to follow-up before 30 months. Infants who met some but not all of the criteria for failure to thrive (see “Growth outcomes measured” column).

1. Maternal education: < High school, High school graduate, Some college, ≥ College graduate.

2. (Annual) family income: < $10,000, $10-20,000, > $20,000.

Data gathered via a family interview by a nurse clinician during the newborn’s nursery stay. It is not specified whether mothers directly reported their education level.

Weight, height and head circumference measured at the follow-up clinics at 40 weeks corrected age, and 4, 8, 12, 18, 24, 30 and 36 months corrected age.

FTT defined as: a) Infants coded by the developmental clinician during a health assessment as having FTT and b) Weight < 5^th percentile at ≥ 2 points in time and c) Weight growth during the preceding months was less than average for sex and corrected age as determined by velocity growth curves. Infants who met criteria b) and c) whose growth curves were determined to represent FTT upon blind review by 2 developmental paediatricians were also coded as FTT.

Percentages of infants with and without FTT reported according to maternal education level. Distribution of maternal education compared between infants with and without FTT. Logistic regression analysis conducted to explore the association between maternal education and risk of FTT, adjusting for small for gestational age status at birth, abnormal or suspect neurologic exam at birth, birthweight, maternal age, maternal height and whether or not the infant was living with their father.