The impact of neonatal morbidities on child growth and developmental outcomes in very low birth weight infants: a nationwide cohort study

Growth in preterm infants has long-term implications for neurodevelopmental outcomes. We aimed to estimate the nationwide growth outcomes from birth to 5 years in infants born under 1500 g and to analyze the effects of major morbidities in preterm infants on growth. In total, 2961 children born in 2013 with a birth weight under 1500 g who underwent an infant health checkup between 2013 and 2018 according to the National Health Insurance Service database were included. Checkups were conducted at 4–6, 9–12, 18–24, 30–36, 42–48, and 54–60 months of age. Information was obtained from the International Classification of Diseases-10 codes or a questionnaire administered during the check-up. At 60 months of age, the mean percentiles of weight, height, and head circumference fell within only the 30–40th percentile of normal growth values. About 30% of infants had growth parameters below the 10th percentile and showed worse neurodevelopmental outcomes. Using multiple logistic regression, infants with bronchopulmonary dysplasia showed a significantly higher incidence of growth restriction in all three categories of weight (odds ratio [OR] 1.50), height (OR 1.33), and head circumference (OR 1.36) at 60 months. Sepsis was associated with growth restriction in weight (OR 1.43) and head circumference (OR 1.33). Periventricular leukomalacia infants had relatively small head circumferences (OR 1.91) and poor developmental screening results (OR 2.89). Conclusion: Catch-up growth remains a major issue in infants born under 1500 g, especially those with some morbidities from preterm birth. Regular checkups to monitor and early intervention to achieve normal growth are essential. What is Known: • Growth in preterm infants has long-term implications for neurodevelopmental and cardiometabolic outcomes. • Data are lacking on the time-serial effects of many preterm morbidities simultaneously on long-term growth outcomes. What is New: • All growth parameters of VLBW infants, including weight, height, and head circumference, fell within the 30–40th percentile of normal growth for infants at 60 months of age, indicating that catch-up growth for VLBW infants remains an issue. • VLBW infants with major preterm morbidities, including BPD, PVL, and sepsis, showed difficulties in achieving normal catch-up growth and neurodevelopment at 60 months of age. What is Known: • Growth in preterm infants has long-term implications for neurodevelopmental and cardiometabolic outcomes. • Data are lacking on the time-serial effects of many preterm morbidities simultaneously on long-term growth outcomes. What is New: • All growth parameters of VLBW infants, including weight, height, and head circumference, fell within the 30–40th percentile of normal growth for infants at 60 months of age, indicating that catch-up growth for VLBW infants remains an issue. • VLBW infants with major preterm morbidities, including BPD, PVL, and sepsis, showed difficulties in achieving normal catch-up growth and neurodevelopment at 60 months of age.


Introduction
Recent improvements in preterm survival have resulted in a shift in the focus for preterm infants toward improving quality of life through monitoring parameters of growth and developmental status [1]. Growth represents an increase in body size and is a complex process that is influenced by genetic, hormonal, and environmental factors [2]. Infants born with very low birth weight (VLBW) face a higher risk of postnatal growth failure [3,4], and differences in postnatal growth patterns according to gestational age in the first 3 months have been reported in infants developing morbidities [5]. Meanwhile, preterm morbidities, including patent ductus arteriosus (PDA), bronchopulmonary dysplasia (BPD), necrotizing enterocolitis (NEC), and late-onset sepsis, have been found to have, to different degrees, influences on growth depending on the morbidity [6].
Growth is often assessed by comparing weight, length, and head circumference (HC) to normal growth references of a given society. Many studies have reported close relationships between postnatal growth and neurodevelopmental outcomes in preterm infants [4,[7][8][9]. Linear growth may be more relevant to neonatal brain development than weight gain, which is unable to differentiate between development of organs and an increase of adipose tissue [10]. Researchers have also reported that growth-restricted extremely preterm infants with postnatal head sparing show better neurodevelopmental outcomes than those without it, especially in regards to motor function [11].
The catch-up growth patterns of preterm infants have remained a matter of debate. Most preterm infants generally tend to start catch-up growth early in the first months of life, and catch-up growth is often achieved within the first 2 years of life [12]. However, many studies have reported that catchup growth can continue after infancy, even into adolescence [13][14][15], and various complications of preterm birth may make achieving catch-up growth and normal neurodevelopment difficult. To date, data are lacking on the time-serial effects of many preterm morbidities simultaneously on longterm growth outcomes. Accordingly, this study aimed to estimate nationwide long-term growth outcomes for infants born under 1500 g from a population-based surveillance system in comparison to those for term infants. Also, we aimed to analyze the effects of major morbidities from preterm birth on growth outcomes.

Patients and data source
We initially identified 430,541 infants who were born in 2013 and underwent an Infant Health Check-up between 2013 and 2018 from the National Health Insurance Service (NHIS) database. Healthcare claims data, such as diagnostic codes, costs of diagnostic tests, and administered procedures, for almost all Korean residents were linked to a health check-up database. Information on birth weight was obtained by the International Classification of Diseases-10 (ICD-10) codes inputted by the hospital or by a questionnaire administered by the Infant Health Screening Program. Extremely low birth weight (ELBW) infants were defined as infants born under 1000 g, and VLBW infants were defined as infants with birth weights between 1000 and 1499 g. We used birth certificate data from Statistics Korea to estimate examination rates of the National Health Screening Program among preterm infants.
The National Health Screening Program for infants and children in Korea was launched in 2007 to monitor current health issues and has been successfully implemented as a primary clinical service. The program is a type of population surveillance system that includes taking medical history, physical examination, anthropometric measurements, screening for visual acuity, developmental screening by Korean Developmental Screening Test (K-DST), oral examination, and questionnaires with anticipatory guidance. The study population involved had their first visits at 4-6 months of age, their second visits at 9-12 months, their third visits at 18-24 months, their fourth visits at 30-36 months, their fifth visits at 42-48 months, and their sixth visits at 54-60 months. All the check-ups were based on the chronological age, not the corrected age.
The K-DST is a screening test that verifies whether infants have achieved normal neurodevelopmental status in six domains of gross/fine motor, cognition, communication, social interaction, and self-control. Tests are administered according to the child's corrected age at the time of the clinic visit. Results are categorized into four groups based on standard deviation (SD) scores: "further evaluation" for scores below − 2 SD, "follow-up test" for those between −2 and −1 SD, "peer-level" for those between − 1 and 1 SD, and "high-level" for those above 1 SD. For children who are categorized as the follow-up test group, short-term checkups are recommended for re-evaluation. To screen for developmental delay, scores below − 1 SD were set as a critical cutoff, which would reflect the infants in need of further evaluation and follow-up test [16]. Additionally, there are some positive questions as red flags for clinically important neurodevelopmental diseases, such as cerebral palsy, language delay, and autism spectrum disorders, and the involved infants/children were referred to medical specialists [17].
Preterm morbidities, including BPD, sepsis, periventricular leukomalacia (PVL), retinopathy of prematurity, and intraventricular hemorrhage, were identified using ICD-10 codes inputted by the hospital. All morbidities were diagnosed by the clinical judgment of the neonatologists according to generally known definitions. We used The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist to assess and improve the quality of the study.

Statistical analyses
Baseline subject characteristics are expressed as means and standard deviations for continuous variables and as percentages for categorical variables. The cohort was stratified according to birth weight groups and the time of entering the screening program. Growth parameters were analyzed by year or according to birth weight group using one-way ANOVA or chi-square test. Univariate analysis was used to determine the risk factors among preterm morbidities for growth failure and delayed development at 60 months of age. The independent risk factors associated with them were determined to use for multiple logistic regression analysis using odds ratios (OR) and 95% confidence intervals (CI). For the multivariate analysis, variables with p-values under 0.05 in the univariate analysis and other clinically important variables including birth weight and sex were used. The Hosmer and Lemeshow goodness-of-fit test was used to assess the suitability of the model, for it can show how well the nomogram was calibrated. In the test, a close approximation between the observed probability and the predicted probability means good calibration and can confirm the exportability of the model. Testing for multicollinearity with variance inflation factors was done. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, North Carolina). P-Values < 0.05 were considered statistically significant.

Ethics statement
This study used NHIS-NSC data (NHIS-2020-1-543) maintained by the NHIS. The authors declare no conflicts of interest with NHIS. In this study, all identifiable variables, including claim-, individual-, and organizational-level identification numbers, were randomly re-generated by the NHIS database to protect patient privacy. The study protocol was approved by the Institutional Review Board (IRB) of Gangnam Severance Hospital (IRB No. 3-2019-0147). Informed consent was waived due to the nature of the study.

Results
From birth statistics, the number of live infants born under 1500 g in 2013 was 2961. The numbers of infants who received the Infant Health Check-up at ages 6 to 60 months were as follows: first visit, 830 (28%) infants; second visit, 1015 (34%); third visit, 1690 (57%); fourth visit, 2602 (88%); fifth visit, 2031 (69%); and sixth visit, 1836 (62%). Mean percentiles of growth parameters at ages 6 to 60 months are shown in Fig. 1. Mean growth percentiles of ELBW and VLBW infants were significantly lower than those of the control group. The mean weight percentile was the highest at age 36 months (33.1), with mean height percentile the highest at 36 months (38.9) and mean head circumference the highest at 24 months (33.0).
The incidences of infants having poor growth parameters less than 10th percentile of age norms at each visit are shown in Fig. 2. VLBW and ELBW infants showed markedly higher incidences of poor growth than infants with normal birth weight from birth to 60 months of age. For the comparison between ELBW and VLBW infants, significant differences in the incidence of infants below the 10th percentile in growth parameters were noted until 36 months of age; however, they became similar after 48 months.
Among ELBW and VLBW infants, hyaline membrane disease, PDA, intraventricular hemorrhage, sepsis, and BPD significantly increased the risk for weighing less than the 10th percentile of normal at 60 months of age by univariate logistic regression analysis. Hyaline membrane disease, sepsis, BPD, and PVL were associated with a risk for having a height below the 10th percentile at 60 months. Hyaline membrane disease, PDA, intraventricular hemorrhage, sepsis, BPD, and PVL were significant risk factors for having a head circumference less than the 10th percentile for age at 60 months (Table 1).
Multiple regression analysis with preterm morbidities adjusted for sex and birth weight was performed to determine independent risk factors for growth restriction (having growth parameters less than the 10th percentile for age at 60 months of age). BPD (OR 1.50; 95% CI 1.16-1.94) and sepsis (OR 1.43; 95% CI 1.14-1.79) were confirmed as independent risk factors for growth restriction in weight, and BPD (OR 1.33; 95% CI 1.02-1.73) was the only risk factor for growth restriction in height. BPD (OR 1.36; 95% CI 1.04-1.78), sepsis (OR 1.33; 95% CI 1.02-1.73), and PVL (OR 1.91 95% CI 1.27-2.88) were risk factors for growth restriction in head circumference (Table 2). Figure 3 shows a serial comparison of mean growth percentiles at ages 12 to 60 months among ELBW and VLBW infants with and without BPD/PVL. All three growth parameters showed significant differences between the infants with and without BPD/PVL. Among ELBW and VLBW infants, the infants with growth restriction in any growth parameter at 60 months of age showed significantly poor outcomes by developmental screening, compared to those who did not have growth restriction (Fig. 4). From multiple regression analysis with preterm morbidities and growth parameters as risks for poor developmental outcomes at the age of 60 months, we noted that PVL (OR 2.89; 95% CI 1.68-4.99) and growth restrictions in head circumference (OR 2.36; 95% CI 1.46-3.80) and height (OR 2.42; 95% CI 1.57-3.74) were confirmed as independent risk factors for requiring further evaluation for developmental delay. All p values for Hosmer and Lemeshow goodness-of-fit test were above 0.05. Variance inflation factors of all variables were between 1 and 2.

Discussion
Postnatal growth can be an excellent indicator of health and developmental outcomes in preterm infants. Several studies have shown a relationship between impaired extra-uterine growth and poor long-term performances in preterm infants [4,7,8]. Our study confirmed an association between postnatal growth failure and neurodevelopmental outcomes, as well as morbidities affecting growth outcomes until 5 years of age, in infants born under 1500 g.
In Korea, the National Health Screening Program for infants and children has facilitated the obtainment of a vast amount of data on postnatal growth patterns for preterm infants. We found that the screening rates of ELBW and VLBW infants were lower at the first (28.0%) and second (34.3%) check-ups and increased over time to around 70%. It is important for neonatologists to encourage caregivers All three growth parameters among ELBW and VLBW infants tended to gradually approach normal growth rates over time. However, achieving growth in only the 30 to 40th percentile of norm reference ages at 60 months of age was noted. In addition, about 30% of the preterm infants showed growth restriction under the 10th percentile of Fig. 2 Incidences of infants having growth parameters less than the 10th percentile of their age among three birth weight groups: ELBW, VLBW, and infants with a birth weight of 2500-4500 g. A Height. B Weight. C Head circumference. ELBW, extremely low birth weight; VLBW, very low birth weight normal reference in all three parameters at the age of 48 and 60 months and showed worse neurodevelopmental outcomes at 60 months, as has been reported by many other studies [4, 7-9, 18, 19]. Similarly, very-preterm-born and small-for-gestational age-born infants in Denmark showed catch-up growth in weight and height at 6 years of age but reached a significantly lower mean z score than appropriate for gestational age (AGA) children. AGA infants also did not achieve catch-up growth completely at 6 years of age [20]. Clinicians should carefully monitor these infants with extended follow-up after discharge and seek more ways to assist catch-up growth. It is widely known that postnatal head growth appears to be spared relative to weight and height [10,11,19]. In our study, head growth seemed to be similarly spared when compared with the other two parameters at the first check-up, although, with time, the proportion at which growth restriction occurred became similar in all three groups (around 30% at 48 and 60 months). It can be assumed that failure to achieve catch-up growth after an extended period of time will eventually lead to the disappearance of the sparing effects of head growth. Catch-up growth of premature babies after discharge should be monitored and addressed as soon as possible.
Growth is influenced by various factors, including comorbidities and sex [9,21]. In our study, infants with BPD showed significant long-term growth restrictions in all three growth parameters. Furthermore, even with adjustment of other morbidities, it was the only meaningful factor that was associated with growth restrictions in all three categories. BPD happens through complex mechanisms and has longterm effects from infancy to even young adults [22,23]. Repetitive hypoxic insults, increased respiratory efforts, use of postnatal steroids, undernutrition due to fluid restriction and feeding difficulty, and frequent infections are factors associated with growth restriction in BPD infants [18,24,25]. Prevention of BPD is not just beneficial for lung function of premature infants, but is also important for overall health conditions and neurodevelopmental outcomes.
In a previous large-scale cohort study, infection in ELBW infants affected growth failure in weight and head circumference at both 36 weeks and 18 to 22 months of corrected gestational age [26]. We found that both ELBW and VLBW infants with sepsis were more likely to experience delayed growth, especially in weight gain (OR, 1.43) and head growth (OR, 1.33). Infection in the neonatal period affects long-term growth, even into pre-school age. Inflammationinduced white matter injury is already a widely known phenomenon in both experimental and clinical models [27][28][29], and delay in head growth in premature infants with sepsis may be associated with this inflammation. Our data suggests that, once the brain is damaged due to sepsis and associated inflammation, recovery is difficult, and the inflammation may cause continued failure in brain growth and loss of brain function. Premature infants are vulnerable to infection and have a relatively higher incidence of sepsis. Therefore, efforts to prevent and treat sepsis in preterm infants are of paramount importance in preventing issues that could affect neurodevelopmental outcomes.
Our research has some strengths. First, with the data from the National Health Insurance Corporation of Korea, we were able to analyze a nationwide database of preterm infants longitudinally with higher accuracy and reliability. Second, because the data were based on a common set time, comparative analysis was possible for specific times.
Finally, prior to our study, there were few studies about the relationship between morbidities caused by prematurity and poor long-term postnatal growth. Our results can help neonatologists focus on preterm infants with certain morbidities in terms of postnatal growth, thereby minimizing growth failure and neurodevelopmental disabilities. On the other hand, we also had some limitations. First, this was a retrospective and observational study and thus can have an unavoidable limit of controlling some confounders. We could only suggest the explanation of the association, not a causal-effect relationship. Second, there was a variance of the number of the infants who received check-ups at each of the investigated times. The group of infants who did not get the exams at the earlier times of the check-ups showed lower mean birth weight and more ELBW infants than those with the check-ups, inferring some unhealthy VLBW infants that cannot be included in the earlier exams. However, later, at the 5th and 6th check-ups, there were no differences of the gestational age, mean birth weight, and sex between the infants with and without the check-ups. Third, the diagnoses of preterm morbidities were confirmed only using ICD-10 codes inputted by the neonatologists of each hospital. The consistency and suitability of all diagnoses cannot be verified. Fourth, as the claim data were based on postnatal age, not on the corrected age, and some growth parameters could be underestimated, especially in case of young infants. Fig. 3 Effects of BPD and PVL on long-term growth outcomes. A Height percentile of BPD and non-BPD infants. B Height percentile of PVL and non-PVL infants. C Weight percentile of BPD and non-BPD infants. D Weight percentile of PVL and non-PVL infants. E Head circumference percentile of BPD and non-BPD infants. F Head circumference percentile of PVL and non-PVL infants. BPD, bronchopulmonary dysplasia; PVL, periventricular leukomalacia. Note that the time scale is different from Fig. 1 Lastly, there are diversities of treatment protocols between NICU units; this might have affected the postnatal growth of preterm infants.

Conclusions
Despite recent improvements in nutritional support and treatments for preterm infants, growth failure is still an unresolved problem for infants born weighing under 1500 g. We found growth failure to be associated with poor neurodevelopmental outcomes and that infants with major morbidities resulting from preterm birth, such as BPD, sepsis, and PVL, are more likely to have growth restrictions.
Neonatologists and pediatricians should complete regular check-ups of preterm infants for many years and seek to monitor for and treat issues related with catch-up growth and neurodevelopment.