Late preterm birth and growth trajectories during childhood: a linked retrospective cohort study

Background Evidence suggests that accelerated postnatal growth in children is detrimental for adult cardiovascular health. It is unclear whether children born late preterm (34–36 weeks) compared to full term (≥ 39 weeks), have different growth trajectories. Our objective was to evaluate the association between gestational age groups and growth trajectories of children born between 2006–2014 and followed to 2021 in Ontario, Canada. Methods We conducted a retrospective cohort study of children from singleton births in TARGet Kids! primary care network with repeated measures of weight and height/length from birth to 14 years, who were linked to health administrative databases. Piecewise linear mixed models were used to model weight (kg/month) and height (cm/month) trajectories with knots at 3, 12, and 84 months. Analyses were conducted based on chronological age. Results There were 4423 children included with a mean of 11 weight and height measures per child. The mean age at the last visit was 5.9 years (Standard Deviation: 3.1). Generally, the more preterm, the lower the mean value of weight and height until early adolescence. Differences in mean weight and height for very/moderate preterm and late preterm compared to full term were evident until 12 months of age. Weight trajectories were similar between children born late preterm and full term with small differences from 84–168 months (mean difference (MD) -0.04 kg/month, 95% CI -0.06, -0.03). Children born late preterm had faster height gain from 0–3 months (MD 0.70 cm/month, 95% CI 0.42, 0.97) and 3–12 months (MD 0.17 cm/month, 95% CI 0.11, 0.22). Conclusions Compared to full term, children born late preterm had lower average weight and height from birth to 14 years, had a slightly slower rate of weight gain after 84 months and a faster rate of height gain from 0–12 months. Follow-up is needed to determine if growth differences are associated with long-term disease risk. Supplementary Information The online version contains supplementary material available at 10.1186/s12887-023-04257-x.


Introduction
Late preterm birth, defined as birth between 34 and 36 weeks gestation, accounts for 75% of all preterm births and this percentage has been rising since 1990 [1].Increases in obstetrical interventions, such as induced delivery and caesarean section, are important contributing factors to the increase in late preterm births but much of the reason remains unexplained [2][3][4].Late preterm newborns are not as physiologically or metabolically mature as full term newborns and late preterm birth may lead to adverse health outcomes across the life course [5].A recent systematic review of the association between late preterm birth and cardiometabolic health outcomes found that children and adults born late preterm compared to full term were at increased risk of diabetes (pooled adjusted relative risk 1.24 (95% confidence interval (CI) 1.17, 1.32) from 9 studies), and hypertension (pooled adjusted relative risk 1.21 (95% CI 1.13, 1.30) from 11 studies), but had lower BMI z-scores (standardized mean difference -0.38 (95% CI -0.67, -0.09) from 5 studies) [6].Further, a retrospective cohort study assessing the association between preterm gestational age and childhood cardiometabolic risk (CMR) score found that those born late preterm had higher CMR compared with those born full term [7].As the CMR score tracks risk from childhood into adulthood [8][9][10], these findings suggest that late preterm birth may be an important risk factor for cardiometabolic disorders later in life.Although the risk of late preterm birth on later adverse outcomes is small on the individual level, a large number of children born late preterm are reaching young adulthood when the incidence of cardiometabolic diseases may increase [11].
Growth during infancy and childhood are important indicators of child health and development and have an important influence on health later in life [12].Previous studies have reported that accelerated postnatal growth in weight could be detrimental for adult cardiovascular health [13][14][15].However, most growth trajectory research is limited to children born full term [16,17] or overall preterm (defined as < 37 weeks gestation) [18,19] with few studies investigating growth trajectories by gestational age groups, including late preterm birth.Understanding the effects of late preterm birth and growth trajectories can elucidate the link between late preterm birth and cardiometabolic outcomes for this potentially high-risk population.
The primary objective of this study was to evaluate the association between gestational age groups (very/moderate preterm (< 34 weeks), late preterm (34-36 weeks), early term (37-38 weeks) compared to full term (≥ 39 weeks) and trajectories of weight and height from early infancy to mid-childhood.Our secondary objective was to determine whether weight and height growth rates differed during each growth period for gestational age categories by sex.

Study population and design
We conducted a retrospective cohort study of children born between April 1, 2006, andMarch 31, 2014, followed until 2021, participating in The Applied Research Group for Kids (TARGet Kids!) primary care practice-based research network in Toronto, Canada [20].Inclusion criteria for the TARGet Kids! cohort include children < 6 years of age and receiving primary healthcare at a TARGet Kids! participating site.Exclusion criteria include health conditions affecting growth (e.g., failure to thrive, cystic fibrosis), any acute or chronic conditions (other than asthma and high functioning autism), severe developmental delay and families unable to communicate in English [20].Data from children in TARGet Kids! who had weight and height (or length if < 2 years) measured from birth to 14 years were deterministically linked using individual Ontario Health Insurance Plan (OHIP) numbers (98% success) to population-based health administrative data at ICES (Ontario, Canada).ICES is an independent, non-profit research institute whose legal status under Ontario's health information privacy law allows it to collect and analyze health care and demographic data, without requiring individual participant consent, for the purpose of health system evaluation and improvement.We used the MOMBABY dataset at ICES, which is derived from hospital discharge abstracts, to identify maternal-newborn records for all hospital births (98% of births in Ontario) [21].The MOMBABY dataset was then linked with the Better Outcomes Registry and Network (BORN), a province-wide registry of all births in Ontario, Canada, for more detailed clinical perinatal data [22].For this study, perinatal data from BORN were available at ICES for children born between April 1, 2006, and March 31, 2014.Ethics approval was granted by the Hospital for Sick Children, Unity Health Toronto, and the Hamilton Integrated Research Ethics Boards.All methods were carried out in accordance with relevant guidelines and regulations.Informed consent was obtained from all subjects and/or their legal guardian(s).

Exposures
The primary exposure was gestational age at birth which was calculated based on the best clinical estimate of gestation using a combination of ultrasound and last menstrual period-based estimates [23].For children born in hospitals, gestational age was identified from the MOMBABY dataset while BORN data were used to identify gestational age for children born through home births.Gestational age was analyzed categorically with four groups defined a priori: very/moderate preterm (< 34 weeks), late preterm (34-36 weeks), early term (37-38 weeks), and full term (≥ 39 weeks, used as the reference group) [24,25].
In this study, the very/moderate preterm group was defined as < 34 weeks because it included mostly children born between 32-34 weeks with a small number of children born < 32 weeks.Children born less than 32 weeks of age (considered very preterm) are generally excluded from TARGet Kids! at enrollment.However, some parents or caregivers did not report their child's gestational age on the initial self-reported questionnaire, and these children were included.Once linked to administrative data, we found that 58 children < 32 weeks were included in the cohort, and we kept them in the analysis.

Outcomes
The primary outcomes were the rate of weight change (kg/month) and rate of height change (cm/month) between 0-14 years identified from TARGet Kids! data.Children with only one measure of weight or height were excluded.Histograms were created for weight and height and extreme outliers that represented implausible values were removed (i.e., 8 kg (after 30 months) > weight > 75 kg or 30 cm > height > 200 cm).For the weight model, birthweight was included as the initial measurement.Weight and height were measured by trained TARGet Kids! research staff members using standardized instruments [26] and were analyzed continuously.Our primary interest was to characterize and understand the differences in growth in children born late preterm compared to full term over a range of ages; therefore, our primary outcome was the unstandardized growth measures which we describe by chronological age and sex.The World Health Organization growth reference standards are recommended for full-term children, and while they are often used by correcting preterm infants to term age, this is usually only done until up to 2 or 3 years of age but not beyond and in our study children were followed until up to 14 years of age [27].As a secondary analysis in this paper, we included the WHO standardized values but given the range of age for the outcomes and the primary goal of this research study to understand the impact of preterm on growth measures, we did not correct for gestational age.Z-standardized measures of weight, height, and BMI were adjusted for child's age and sex using the World Health Organization recommendations.Biologically implausible values were identified using cut-points proposed by the World Health Organization (values with z-scores < -5 or > 5) and removed.The secondary outcomes of interest were the rate of change in zweight, zheight, and zBMI per month.

Confounding variables
Confounding variables were selected a priori guided by previous literature and included maternal, child, and sociodemographic characteristics hypothesized to be associated with gestational age at birth and child growth, but not on the causal path between them [6].The included confounders were maternal age at delivery, maternal ethnicity, family income, child age at outcome, and child sex.A table describing the data source and operationalization of variables is provided in Supplemental Table 1.

Statistical analysis
Repeated measures of weight and height were used to estimate the rate of weight and height change for each gestational age group through piecewise linear mixed models with random intercepts.Knots were determined by using a locally estimated scatterplot smoothing (LOESS) curve to visually assess the approximate locations (age in months) at which growth rates substantially changed in slope in our cohort.Based on the LOESS curve, we fit knot points at 3, 12, and 84 months of age which also aligned with knot points in other growth studies [12].Sensitivity analyses were conducted to determine: 1) whether changing the value of knots (different months) or 2) increasing the number of knots based on literature and data distribution changed the results [12].The resulting model with the best fit (4 knots) was selected based on model fit statistics (e.g., Akaike information criterion (AIC) and the Bayesian information criterion (BIC)) and the estimated results were compared to results from knots determined through the LOESS curve (3 knots).As the model fit between the 3 knot and 4 knot model were similar in terms of fit statistics, estimates, and standard errors, we chose the 3-knot model for parsimony (Supplemental Tables 2 and 3).We estimated growth rates for each growth period (period between knot points) comparing the gestational age categories to full term by taking the difference between the beta coefficients.Contrasts were used to estimate the difference in slopes between gestational age categories and obtain corresponding confidence intervals.As a secondary analysis, sex-stratified rates of growth were examined to determine whether growth rates during each growth period differed for boys and girls.All analyses were adjusted for maternal age at delivery, maternal ethnicity, child age, child sex, and family income.
Missing data for each covariate were imputed using multiple imputation and a total of 10 imputed data sets with the PROC MI procedure and pooled using Rubin's rules with PROC MIANALYZE.All analyses were conducted using SAS, version 9.4.

Results
There were 10,897 children overall in TARGet Kids! but 4990 of them had a birthdate outside of the eligible range of this study.Of the 5907 potentially eligible children, 108 were not included in the analysis because they were not eligible for Ontario's health care plan, 770 were excluded because of insufficient weight/height measures, 311 were missing gestational age data, and 295 children were part of multiple births.The final study cohort consisted of 4423 (75%) of the eligible children (Supplemental Fig. 1).The mean number of growth measures per child was 11.Of 4423 children, 2320 (52%) were male and the mean age at the last visit (i.e., the average follow-up time) was 5.9 years (Standard Deviation: 3.1).Late preterm children were more likely than full term children to be male and have lower family income; and their mothers were more likely to be older and be of non-European ethnicity.The characteristics of the study population, overall and by gestational age categories, are presented in Table 1.

Weight trajectories
Children born late preterm had lower weight compared to full term through the age range studied (0-14 years).These results are presented in Fig. 1 and Supplemental Table 4.For late preterm compared to full term, the rate of weight growth was similar from 3-12 months (MD 0.08 kg/month 95% CI -0.04, 0.20) and the rate of weight growth was slower from 84-168 months (MD -0.04 kg/month 95% CI -0.06, -0.03).However, there was little to no evidence of differences in the rate of weight growth from 0-3 months and 12-84 months.These results are presented in Table 2.
Children born early term consistently had similar weight and growth rates for weight to full term (Supplemental Table 4; Table 2).

Height trajectories
Children born late preterm consistently had lower height than full term.These results are presented in Fig. 2 and Supplemental Table 5.For late preterm compared to full term, the rate of height growth was faster from 0-3 months (MD 0.70 cm/month 95% CI 0.42, 0.97) and 3-12 months (MD 0.17 cm/month 95% CI 0.11, 0.22).There was no evidence of differences after 12 months.These results are presented in Table 2.

Z-Standardized trajectories
For children born very/moderate preterm, late preterm, and early term compared to full term, differences in zweight (Supplemental Fig. 2), zheight (Supplemental Fig. 3), and zBMI (Supplemental Fig. 4) and rates of growth were evident.A table describing the z-standardized growth rates and mean differences compared to full term is provided in Supplemental Table 6.

Secondary analysis
Regarding the rate of weight growth, for late preterm compared to full term, slight differences by sex were evident after 84 months where boys grew less slowly than girls.These results are presented in Fig. 3 and Table 3. Differences in the rate of weight growth between girls and boys were more evident when comparing very/moderate preterm to full term where boys (MD -0.07 kg/ month, 95% CI -0.09, -0.04) had a slower rate of growth in weight from 84-168 months than girls (MD 0.03 kg/ month, 95% CI 0.00, 0.05).
Regarding the rate of height growth, for late preterm compared to full term, there did not appear to be differences by sex until 84-168 months where late preterm boys (MD 0.05 cm/month, 95% CI 0.03, 0.08) had a faster rate of growth in height than girls (MD -0.06 cm/month, 95% CI -0.09, -0.03).For very/moderate preterm compared to full term, there was some evidence that boys (MD 0.05 cm/month, 95% CI 0.03, 0.06) had a faster rate of growth in height than girls (MD 0.02 cm/month, 95% CI 0.01, 0.04) from 12-84 months.From 84-168 months, there was some evidence that girls (MD 0.05 cm/month, 95% CI 0.02, 0.09) had a faster rate of growth in height than boys (MD 0.00 cm/month, 95% CI -0.03, 0.03).For early term compared to full term, there were no notable differences between girls and boys when comparing weight or height growth rates.The results for height growth are presented in Fig. 4 and Table 3.

Discussion
Using chronological age, our study results suggest that the mean growth trajectories of children born late preterm and very/moderate preterm differed from that of full term children with the highest degree of differences occurring between very/moderate preterm and full term children.Generally, the more preterm, the lower the mean value of weight and height until early adolescent years.Children born late preterm compared to full term had a slightly slower rate of weight gain after 84 months.The association was stronger in girls than boys.In children born late preterm, differences in height trajectories were observed from 0-12 months of age and did not appear to differ by sex.Children born very/moderate preterm had a slight increased rate of weight gain from 12-84 months but slower rate after 84 months with the slowdown more evident in boys.Further, children born very/moderate preterm had an increased rate of height gain across the age range studied, which was more evident in boys between 12-84 months and more evident in girls between 84-168 months.Overall the rates of growth observed in our study for full term children were quite similar to previous data from four international birth cohorts which were mainly general population studies [12].Howe et al. reported that cohorts displayed rapid weight and height gain in early infancy followed by slower growth in later infancy and early childhood [12].For the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort, where data was available for older ages, rates of weight gain began to increase again from 7 years of age which coincided with our data [12].
Few studies have investigated the association between late preterm birth and weight and height growth trajectories.A Dutch cohort study of 2324 children followed from birth to 4 years of age, reported that increases in weight and height were similar for children of different gestational ages, indicating no catch-up growth [28].Although growth patterns were the same for boys and girls, boys demonstrated greater variability in growth by gestational age, especially at lower gestational ages (≤ 30 weeks) suggesting that preterm boys may be more susceptible than girls to complications of preterm birth that influence growth [28].A Brazilian population-based cohort study of 3,285 children followed from birth to 24 months, reported that children born late preterm had similar weight gain to Fig. 2 Growth curves for height in cm per month from 0 to 150 months by gestational age categories.The lines represent the growth curve, and the shaded areas represent the 95% confidence intervals.The box represents a superimposed area of the growth curve between 0 to 12 months full term children whereas children born late preterm grew on average 1.26 cm more than full term [29].The present study found that children born late preterm compared to full term had higher height growth of 0.70 cm per month but only from 0-3 months of age.Several studies have reported that accelerated postnatal growth in weight could  increase the long-term risk of obesity and cardiovascular disease [13][14][15].We found some evidence of faster postnatal growth in weight during early childhood (3-12 months) for late preterm compared to term.Key strengths of this study included the ability to link clinical data from a large cohort of children to health administrative data from a publicly funded healthcare system to obtain valid prospectively measured data on gestational age at birth and maternal factors.Weight and height were measured objectively from birth to 14 years of age by staff trained using standardized instruments.We were able to adjust for confounding variables that are not routinely available in administrative data through linkage with cohort study questionnaire data and physical measures.This study also had several limitations including the possibility of residual confounding through the lack of variables such as feeding practices, environmental and lifestyle factors.We did not have a population-based representative sample because our study was nested within TARGet Kids! and reflects children recruited from selected primary care practices in one urban area in Canada.Selection bias is possible and the findings from this study may not be generalizable to other populations, especially as our sample had relatively high income and education compared to the overall population.Further, approximately 25% of eligible children were not included in the final sample due to missing data or because they were part of multiple births, limiting the representativeness of our study.As children born less than 32 weeks of age were generally excluded from TARGet Kids! at enrollment, this may have disproportionately excluded preterm children with abnormal growth trajectories.Only a small number of 58 children born < 32 weeks was included in our study.Additionally, we investigated growth until 14 years of age; however, we had less follow-up data at older ages (mean follow up time: 5.9 years), which may have reduced the precision of our estimates for older children.Since we were interested in understanding if categories of preterm children grew differently than full term, we did not correct for gestational age which may mask the differences in growth and is consistent with the methods in previous cohort studies.While it may be important to correct for gestational age when evaluating clinical outcomes for children born preterm, for this research study, we were specifically interested in understanding the differences in growth measures between preterm and full Fig. 4 Growth curves for height trajectories in cm per month from 0 to 150 months by sex and gestational age categories.The lines represent the growth curve, and the shaded areas represent the 95% confidence intervals term and did not adjust away the differences by correcting for gestational age [27,30].

Conclusions
We found that late preterm birth was associated with a slower rate of weight gain after 84 months and faster rate of height gain from 0-12 months compared to full term birth.Follow-up is needed to determine if differences influence long-term disease risk.Advances in neonatal and pediatric care in the past few decades mean that a substantial proportion of children born late preterm are now reaching young to mid-adulthood.There is an urgent need to understand how late preterm birth contributes to increased cardiometabolic disease incidence.More data from larger studies in diverse populations may help to understand the association between late preterm birth and growth trajectories and provide clinical guidance for developing appropriate weight and height gain standards for this large subset of preterm infants.This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long-Term Care (MLTC).Parts of this material are based on data and/or information compiled and provided by Canadian Institute for Health Information.However, the analyses, conclusions, opinions and statements expressed in the material are those of the author(s), and not necessarily those of Canadian Institute for Health Information.This study is based in part on data provided by Better Outcomes Registry and Network ("BORN"), part of the Children's Hospital of Eastern Ontario.The interpretation and conclusions contained herein do not necessarily represent those of BORN Ontario.
We also thank all of the participating families for their time and involvement in TARGet Kids! and are grateful to all practitioners who are currently involved in the TARGet Kids! practice-based research network.TARGet Kids! Collaboration conceptualized and designed the study and critically reviewed the manuscript for important intellectual content.LA conceptualized and designed the study, contributed to the analysis and interpretation of data, and critically reviewed the manuscript for important intellectual content.All authors read and approved the final manuscript.

Fig. 1
Fig. 1 Growth curves for weight in kg per month from 0 to 150 months by gestational age categories.The lines represent the growth curve, and the shaded areas represent the 95% confidence intervals.The box represents a superimposed area of the growth curve between 75 to 150 months

Fig. 3
Fig. 3 Growth curves for weight trajectories in kg per month from 0 to 150 months by sex and gestational age categories.The lines represent the growth curve, and the shaded areas represent the 95% confidence intervals

Table 1
Characteristics of study participants (N = 4423) by gestational age at birth for the outcome of weight/height

Table 2
Mean differences in weight growth rates (kg per month) and height growth rates (cm per month) by gestational age categories during each growth period for study participants (N = 4423)

Table 3
Mean differences in weight growth rates (kg per month) and height growth rates (cm per month) by sex for gestational age categories during each growth period