Analysis of Factors Associated With Body Mass Index at Ages 18 and 36 Months Among Infants Born Extremely Preterm

This cohort study investigates factors associated with body mass index in the first 36 months of life among infants born extremely preterm.


Introduction
The survival rate of infants born preterm has improved dramatically during recent decades as neonatology has developed. 1 The most notable change was in the 1980s with the development of pulmonary surfactant, which was associated with increased survival among infants born preterm with respiratory distress, 2 and its clinical use was associated with a change in their clinical course. 3 Other therapies, like ventilation techniques, 4 antenatal glucocorticoid therapy, aggressive nutrition policies, 1 and appropriate oxygen supplementation, 5 are also still used. These developments were associated with improvements in the survival rate of infants born preterm 6 ; from 1993 to 1980, the survival rate of infants in the US born weighing from 500 g to 550 g improved from 0% to 50% among Black male infants and from 35% to 70% among White female infants. 7 In 2010, 90% of infants born very preterm (ie, birth weight <1500 g) in Germany survived. 8 In Japan, one of the leading countries in neonatology, mortality rates among infants with very low birth weight at discharge improved from 2003 to 2008, with mortality decreasing from 10.8% to 8.7% among infants born with weights from 501 g to 750 g 9 and decreases in mortality seen even among infants weighing less than 400 g. 10 These developments in neonatology were associated with increased numbers of infants with short gestational age surviving their neonatal period, 5 and their long-term health outcomes have become a matter of much concern. Understanding these outcomes is becomingly increasingly urgent. 11 Studies among children born preterm have found different body shapes compared with children born at term, 12,13 and differences in growth trajectories may be associated with various areas of cognitive function. 14,15 However, the sample size of previous studies was small, [16][17][18] with the largest a uniform sample size of 1320 infants. 19 The mean gestational age of all samples in previous studies was at least 28 weeks, 20 and there was little information about infants born at a gestational age of less than 30 weeks. Moreover, infants with intrauterine growth restriction (IUGR) were not separately analyzed in some studies. [21][22][23][24][25][26] This condition results from uteroplacental insufficiency and is associated with chronic hypoxia and changes in all aspects of the growth process 27 and pathological profile. 28 Infants with IUGR are a high-risk population not only during their hospital stay, 29 but also in their later lives, with increased risk of a wide range of health problems. 30,31 It is important to include information about IUGR in studies of subsequent body shape, but to date this has not been done, to our knowledge. Moreover, given that many of the studies to date have been single-center studies, unmeasured center-specific factors may be associated with changes in the results. 32 We conducted a large multicenter study of infants born extremely preterm (ie, at gestational age <28 weeks) to evaluate factors associated with body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) at ages 18 months and 36 months. sick, and a level 2 NICU provides care to infants who weigh from 1200 g to 1800 g or were born from gestational age 30 weeks to 34 weeks. Every infant born alive and whose birth weight was less than 1500 g or who was born before gestational age 32 weeks was registered, and each participating NICU filled in a common questionnaire.

Methods
All infants born from gestational age 23 weeks to 28 weeks were eligible for this study. Infants were excluded from the study if they died during the hospital stay or follow up period, had unknown sex (ie, no information or unable to determine sex), or were born with a congenital anomaly.  [33][34][35][36][37][38][39][40] LCC is a form of hypotension that occurs after the early neonatal period with unclear underlying pathology that is common in Japan but not widely known in North America or Europe. 41 We calculated BMI at ages 18 months and 36 months using weight and height. Then, using the LMS method and functions developed in previous research for Japanese infants, 42 we calculated BMI z score. Until recently, there has been debate about the best tools for measuring body shape among infants. Recently, BMI has been reported to be a useful measurement among infants born preterm, and it is also reported to be a useful predictor associated with subsequent overweight status 43 that is better associated with metabolic health outcomes. 44,45 Furthermore, BMI curves among infants born preterm were published recently, and the reliability of BMI has been established in this field. 46 Therefore, we presented BMI as a good score to inform long-term metabolic health outcomes among infants born very preterm. This score is also useful because it is easy to measure, which is an important consideration in health checkups.

Statistical Analysis
First, to compare the characteristics of samples, we used t tests to compare BMI by the presence of IUGR. Then, we compared infant characteristics by the presence of multiple pregnancy. We used t tests for gestational age, birth weight, body length at birth, maternal age, and BMI. For proportion of sex, presence of IUGR, and parity more than 1, we used χ 2 tests. We used the same procedure to examine differences between infants with data and those without data. Then, to understand features of complications during pregnancy and the neonatal period, we calculated the percentage of infants with complications, including pregnancy-induced hypertension, CLD, LCC, IVH, and NEC, separately by sex and presence of multiple pregnancy. Statistical tests were 1-sided, and results were considered statistically significant at P < .05 or when 95% CIs did not cross 0.

JAMA Network Open | Pediatrics
We analyzed single pregnancy and multiple pregnancy separately because in multiple pregnancy there is uterine overcrowding and growth rates and growth restriction patterns differ from single pregnancy. 47 Additionally, we regressed BMI and BMI z score calculated at ages 18 months and 36 months against gestational age in weeks, with an interaction term for gestational age and IUGR because IUGR often leads to interventions of caesarian delivery or induced delivery. 48 In this regression, non-IUGR status was set as the reference category. Gestational age was rescaled so that a value of 0 corresponded to gestational age 23 weeks so that the intercept term measured the BMI for an infant born without IUGR at gestational age 23 weeks. Parity, complications during pregnancy, and hospital stay, which were reported in previous studies [33][34][35][36][37][38][39][40] to be associated with subsequent outcomes, were included as covariates. A backward stepwise model-building method was used to remove nonsignificant covariates. 49 The analysis was performed using R statistical software version 4.0.3 (R Project for Statistical Computing). Data were analyzed from April 2018 through June 2021.

BMI
Mean BMI was plotted for each gestational age with 95% CIs separately for boys and girls at ages 18 months and 36 months (Figure). At ages 18 months and 36 months, the mean BMI of infants with IUGR was decreased compared with those without IUGR at all gestational ages, with an increasing trend by gestational age in single pregnancies. However, this tendency was not present in multiple

Linear Regression
We regressed BMI and BMI z score against gestational age in weeks with an interaction term for gestational age and IUGR, separately by sex and single vs multiple pregnancy. The coefficients and their 95% CIs are shown in Table 4.

Discussion
In this cohort study, we collected data on infants born very preterm from the NRNJ database. This is the largest database of infants hospitalized in NICUs in Japan that is suitable for research. We found that gestational age and the presence of IUGR at birth were associated with BMI and BMI z score at ages 18 months and 36 months.
To the best of our knowledge, this was the first study of BMI among infants born very preterm to simultaneously consider gestational age, IUGR, and multiple pregnancy in a large, multicenter sample. Our findings suggest that it is no longer reasonable to discuss BMI without considering gestational age and IUGR information among infants born very preterm. Previous studies of infant and childhood BMI among individuals born extremely preterm had small sample sizes, 19,20 nonuniform samples, 21 or more mature samples with analysis by categories of gestational age. 17,18,24 Moreover, while infants born with IUGR should be analyzed separately because this condition is associated with delays in all aspects of the growth process, 20,27 many complications, 28 and increased risk of long-term health problems, 29-31 many previous studies did not account for this condition. 21,24,25 Similarly, in our study, we analyzed singleton and multiple pregnancies given that multiple pregnancies are associated with a different prognosis, 50 impairment, 40 and feeding disorders. 56,57 Nutritional interventions have been studied among infants born preterm with IVH, 58 and our study findings support the importance of these outcomes for long-term development among these children.
Most previous studies of preterm birth 11,59-61 have focused on neurological or developmental outcomes, and there are few guidelines on body weight measurement or tools for understanding growth patterns among infants born extremely preterm. Our results may be a useful guide for caregivers.
Our study found decreased BMI at ages 18 months and 36 months among infants born preterm.
Although extrauterine growth restriction is a well-known complication at discharge among infants born preterm that is associated with cardiometabolic outcomes and neurodevelopmental outcomes, 62 the causes are unknown. 63 However, infants born preterm are already small at discharge. So far, studies have suggested that gastroesophageal reflux, 64 eating disorders, 65 sleeping disorders, 66 and abnormal fat distribution 26 are associated with growth among infants born preterm.
However, infants born preterm grow in a different environment than the uterus, and epigenetic factors are reported to be associated with growth outcomes, 67,68 but further investigation is required from this perspective.

Limitations
This study has several limitations. The first limitation was the large number of infants with no data.
This could be because caretakers of infants who had good prognoses stopped bringing them to the hospital for health checkups. However, we found no difference in the characteristics among infants with data vs those without data. The second limitation was the lack of information on nutritional status, which prevented us from adjusting for the confounding interaction of nutritional interventions with growth. The third limitation was that there was a small number of infants with decreased gestational age, especially in multiple pregnancies, that may be associated with survivor bias. However, despite this small number, ours is the first study, to our knowledge, to assess growth outcomes among these infants born very preterm and may offer important first insights into their subsequent growth curve. The fourth limitation was that we did not determine severity of IUGR, just presence or absence of this condition. However, in our clinical practice, we use the presence of IUGR rather than severity as a clinical marker, so understanding the long-term outcomes associated with IUGR may be more important clinically than assessing the associations between severity of growth restriction and BMI.

Conclusions
Given that more infants born preterm are surviving their perinatal period and their survival rate continues to improve, 9 it is important to know their long-term outcomes. Japan is at the forefront of neonatology, 10 and revealing the subsequent features of infants born preterm in this country may contribute to neonatology in other countries. This study found that among infants born before gestational age 28 weeks, gestational age was associated with increased subsequent BMI and the presence of IUGR was associated with decreased BMI among infants born from single pregnancies.
These findings suggest that when we follow up with these infants, we should be aware of their baseline characteristics and be conscious of the difficulty in gaining weight in this high-risk population. As more infants survive at younger gestational ages, policy developed using the findings of this study may help ensure they are able to thrive to adulthood and enjoy the full benefits of their survival.