Extrauterine Growth Restriction in Preterm Infants: How to Achieve Optimal Catch-Up Growth During Hospitalization?


 Extrauterine growth restriction (EUGR), a serious risk that potentially impairs the growth of preterm infants after birth and during childhood, triggers a thought of how to achieve optimal catch-up growth during hospitalization.We aimed to access the incidence of optimal catch-up growth in small-for-gestational-age (SGA), appropriate-for-gestational age (AGA) and large-for-gestational-age (LGA) infants, and identify the factors for optimal catch-up growth in preterm infants during hospitalization. Premature infants admitted to Shanghai Children’s Medical Center within 24 hours after birth from January 1,2016 to December 31, 2018 were enrolled. Prenatal, neonatal etiological, and nutrition data were collected and analyzed to identify factors associated with optimal catch-up growth during hospitalization. 105 (9.9%) of 1065 preterm newborns achieved predischarge optimal catch-up growth, including 27 (17.2%) of 157 SGA, 74 (8.5%) of 868 AGA, and 4 (10.0%) of 40 LGA infants. Logistic regression analysis indicated SGA, avoiding BPD and less days to regain birth weight as the contributing factors of optimal catch-up growth during hospitalization.Conclusion: SGA infants showed greater potential in predischarge optimal catch-up growth. Early abundant nutrition and avoiding BPD are essential for achieving optimal catch-up growth during hospitalization.


Introduction
Extrauterine growth restriction (EUGR) refers to the condition of preterm growth failure where an anthropometric measure is below the target based on the postmenstrual age at a certain point of time [1,2]. As a result of perinatal diseases, neonatal morbidities and inadequate nutrition [3], EUGR is not only associated with adverse health outcomes in the neonatal period, but also increases the risk of neurocognitive impairments in later life [4]. It is imperative to reduce the rate of EUGR and control its harms to the maximum extent possible.
In contrast, catch-up growth, de ned as the upward cross of centiles or recovery from growth restriction, is conventionally considered helpful as it releases the constraints originated from the fetal period [5]. V. Pampanini et al.[6] showed that faster weight gain in preterm infants could lead to higher scores on physical and mental developmental testing. In a review article, Ong et al. [4] summarized 16 observational studies relating catchup growth in head circumference (HC) to positive neurodevelopmental outcomes. However, rapid weight gain was found to be associated with insulin resistance and cardiovascular disease in several observational studies [4]. An meta-analysis of 47,661 individuals from 10 cohorts reported the correlation between accelerated infancy weight gain and obesity in later life [7]. So far, there is no unanimity on the time from which catch-up is evaluated (usually at 1-5 years), and the degree of growth required to be identi ed as catch-up growth (an increase in Z-scores or Z-scores >-1).
The optimal growth pattern of preterm infants still remains controversial [8,9].
This clinical dilemma triggers a thought of how to achieve proper catch-up growth in preterm infants. In our study, we suggest a new concept of "optimal catch-up growth during hospitalization", de ned as an increase of Z-scores from birth to discharge for at least two of three auxological parameters including weight, height and HC, because (i) we tried to evaluate catch-up growth at discharge instead of 1-5 years in hope of early indications and interventions for adverse health outcomes; (ii) the decline of weight Z-scores appears in the overwhelming majority of premature newborns as a result of natural weight loss after birth; (iii) premature infants can hardly achieve an large increase in Zscores due to the time limit; (iv) only one quantitative increase in weight, height or HC can not represent the optimal growth of body.
The objectives of this study were to access the incidence of optimal catch-up growth in small-for-gestational-age (SGA), appropriate-forgestational-age (AGA) and large-for-gestational-age (LGA) infants, and identify the prenatal, etiological and nutritional factors for optimal catchup growth in preterm infants during hospitalization.

Participants
This retrospective, single-center study was conducted from January 1, 2016, to December 31, 2018, at a neonatal intensive care unit of Shanghai Children's Medical Center. The criteria for enrollment were (1) gestation age <37 weeks, (2) admission in 24 h after birth, and (3) length of hospital stay ≥7 days. The exclusion criteria were (1) birth with major congenital anomalies (2) major surgery during hospitalization, (3) death or discharge against medical advice, (4) incomplete medical records. Totally, 1065 preterm infants were eligible and divided into the optimal catchup group (n =105, 9.9%) and catch-up failure group (n =960, 90.1%).

Nutrition strategy
The nutritional schedules of all enrolled infants were determined together by neonatologists and nutritionists in line with the physiological and pathological conditions of patients. Enteral nutrition (EN) were initiated mostly at day 2 after birth, unless postponed or stopped by: (i) digestive tract obstruction caused by congenital gastrointestinal malformation; (ii) doubt or diagnosis of necrotizing enterocolitis (NEC); (iii) hemodynamic instability; (iv) multiple organ dysfunction. The starting and increasing volume of EN depended on the birth weight and digestive ability of newborns. Breast milk feeding was applicable as the rst choice and various preterm formula milk was available to supplement the lack or de ciency of breast milk. Inadequate EN intakes and EN calories over 80cal/kg were indications for the beginning and discontinuation of parenteral nutrition (PN), respectively. PN was prescribed by nutritionists and carried out according to the 2013 version of Chinese guidelines [10] for newborn nutrition support in neonates before and changed to the the 2018 version when it came out.

Data collection
Body measurements, clinical characteristics, and nutrition intake were extracted from the paper and electronic medical charts of each infant.
Approval from the Ethics Committee of Shanghai Children's Medical Center was obtained.
Anthropometric data consists of weight, height and HC. Among prenatal factors, gestational age, gender, multiple births, hypertensive disorders of pregnancy (HDP), gestational diabetes mellitus (GDM), prenatal steroid treatment were collected. Neonatal etiological factors, including neonatal asphyxia, pulmonary surfactant use, invasive ventilator, bronchopulmonary dysplasia (BPD), grade III-IV of intracranial hemorrhage (ICH), neonatal sepsis, NEC, feeding intolerance and parenteral nutrition-associated cholestasis (PNAC) were recorded. For PN, information on the day of initiation and termination, duration, the energy intakes during the rst week after birth, on the rst day after PN termination, initial and maximum intakes of proteins, and lipids were included. For EN, we collected data on the type of enteral feeding (formula milk or breast milk feeding), day of initiation, and days to achieve exclusive EN.

De nitions
Based on the Fenton 2013 growth chart [11], SGA, AGA and LGA are de ned as birth weight <10th percentile, between 10th and 90th percentile, and >10th percentile, respectively. We consider EUGR as the decline of weight Z-scores from birth to discharge >1 SD, and optimal catch-up growth as an increase of Z-scores from birth to discharge for at least two of three parameters (weight, height and HC) .
In the study, duration of supplemental oxygen for at least 28 days is de ned as BPD [12]. Neonatal sepsis is diagnosed with positive blood culture. Since there is still no clear de nition criteria for feeding intolerance, the most comprehensive available de nition: gastric residual volume ≥50%, abdominal distension, emesis or feeding disruption due to the inability of digestion was applied [13]. PNAC is de ned as direct bilirubin ≥34.2 µmol/L when PN continues for ≥14 days without other potential diseases [14]. Infants fed with human milk for more than half of days in the hospital are considered as breast milk feeding.

Statistical analysis
The data were presented as the median (interquartile range) and compared using the Mann-Whitney U test or Kruskal-Wallis test for the abnormal distributed quantitative variables. The categorical variables were expressed as numbers (%) and compared using the chi-square or Fisher's exact test. Statistical signi cance was set at P <0.05. Contributing factors for optimal catch-up growth during hospitalization were estimated using a logistic regression model. All statistical data were analyzed using the SPSS statistical software package version 22.0.

Results
The general and prenatal characteristics of preterm newborns during hospitalization are presented in Table 1. The overall percentile of EUGR and optimal catch-up growth were 60.4% and 9.9%. Gestational age, birthweight, weight, height, HC z-scores at birth and at discharge was noticed remarkably lower, while length of hospital stay was signi cantly longer in the EUGR group than in the non-EUGR group, as well as in the catch-up failure group than in the optimal catch-up group. A decrease in SD of weight, height, and head circumference in all groups was observed. There were no discrepancies in multiple births, HDP and GDM between these groups, but more SGA infants in the non-EUGR group and optimal catch-up group. Table 2 shows the mobidities and nutritional characteristics of preterm infants. The etiological data suggested that the percentile of pulmonary surfactant use, invasive ventilation, BPD, hyperbilirubinemia, and feeding intolerance were signi cantly higher in the EUGR group than in the non-EUGR group, as well as in the catch-up failure group than in the optimal catch-up group. The overall morbidities of neonatal asphyxia, BPD, glycometabolism disorder, hyperbilirubinemia, grade III-IV of ICH, neonatal sepsis, NEC, PNAC and feeding intolerance were 17.2%, 15.9%, 10.2%, 83.2%, 2.3%, 2%, 2.2%, 3.1% and 34.8%, respectively. Infants who received breast milk feeding were in the minority (15.2%). It took more days to achieve full feeds for newborns in the EUGR group and catch-up failure group. Energy intakes on the rst day after PN termination were over 80 kcal/kg. The duration of PN, maximum intakes of lipids and protein were found remarkably higher in the EUGR group compared to the non-EUGR group, as well as in the catch-up failure group compared to the optimal catch-up group. Both of the two paired groups showed notable differences in maximum weight loss, days to regain birth weight, and maximum energy intakes during the rst week after birth.
The ow chart in Fig.1 shows that 27 (17.2%) of 157 SGA, 74 (8.5%) of 868 AGA, and 4 (10.0%) of 40 LGA infants achieved optimal catch-up growth during hospitalization. Fig. 2 exhibits the obvious drop in Z-scores between birth and discharge for each infant in all groups, and Z-scores for weight declined more sharply than Z-scores for height and HC. Growth restriction in SGA infants was aggravated with weight Z-scores below -2 SD, height and HC Z-scores approaching -2 SD at discharge. The Z-scores for weight decreased >1 SD in AGA infants, larger than that in SGA infants (<1 SD).
LGA infants lost their advantages at birth and fall down to nearly 0 in Z-scores for all anthrophometric measurements.
Characteristics of SGA, AGA and LGA preterm infants are presented in table 3. SGA infants had signi cantly lower gestational age, weight at birth and at discharge than AGA and LGA infants, but longer days of hospitalization than AGA infants. The rates of an increase in HC z-scores, optimal catch-up growth, EUGR, HDP, glycometabolism disorder and feeding intolerance were remarkably higher in the SGA group, while the rates of multiple births and invasive ventilator were remarkably higher in the LGA group. The average time for starting EN and PN was 3.3 ±2.7 days and 3.2 ±2.0 days in SGA infants, 2.8 ±2.0 days and 3.4 ±2.0 days in AGA infants, and 3.5 ±3.1 days and 4.7 ±5.5 days in LGA infants, though the median numbers appeared to be the same. The SGA group had less weight loss after birth and higher energy intakes at discharge, and took less days to regain birth weight than the other two groups.

Discussion
Growth assessment of preterm infants Growth curves, such as the Fenton2013 curves, WHO growth curve, intergrowth-21st growth curve etc., visually manifest the growth pattern of preterm infants and are implemented as the assessment tools for EUGR and optimal catch-up growth [15,16]. The widely used Fenton2013 curves may not serve as a perfect growth evaluation tool in China since it is based on large preterm sample from different ethnicities in developed countries. And the recently published intergrowth-21st growth curve is prescriptively developed from data of multiethnic and multi country including USA, UK, Italy, China, India, Brazil, Oman, and Kenya [17] . Gonzalez Garcia et al. [17] compared the intergrowth-21st and fenton2013 growth curves in 635 VLBWIs in India and found that the former was more rigorous in the diagnosis of EUGR and better related to neonatal complications than the latter. An Korea study of 1356 very preterm infants by Kim et al[18] also reached similar results. We will do further research to nd out whether this conclusion is applicable to preterm infants with birth weight ≥1500g and gestational age ≥ 32W, and if the diagnosis by intergrowth-21st growth curve could better predict long-term growth outcomes.
In addition, there are no growth charts at present especially targeting multiple births, which occupied approximately 30% of the investigated preterm infants. Whether the increased rate of multiple births and its distinctive growth pattern is causative for the high incidence of EUGR and low rate of optimal catch-up growth cannot be distinguished from our data. Other than that, physiological weight loss has not been taken into consideration in any of the present growth curves [19]. An ideal postnatal growth standard should be updated, indigenized by districts and races, specialized by pregnancy and neonatal conditions, and further research should be conducted [20].
Growth trajectory of SGA, AGA and LGA preterm infants Notably, the changes of Z-scores differed widely in SGA, AGA and LGA preterm infants. They all encountered an obvious drop in weight, height and HC Z-scores, especially in weight Z-scores in view of the natural weight loss at this early stage, but the degree of loss from large to small was LGA, AGA and SGA in sequence, which seemed to be a convergence that narrowed the gap between Z-scores at discharge and at birth. Similar ndings were described in a cohort study conducted by Vandana et al. [21], where catch-up growth (an increase of z-score > 0.67) in SGA and catch-down growth (an decrease of z-score > 0.67 ) in LGA were regarded as goal-seeking paths that brought infants towards their normal growth trajectories.
The postnatal catch-up growth is essential for improving long-term neurodevelopment outcomes, yet excessive growth in both SGA and non-SGA, especially LGA infants, a group at high risk of metabolic diseases like obesity in puberty and adulthood, can bring about the deviation from the original growth channels and increase body fat rather than muscle content, adding to the risk of cadiometabolic diseases in the future [22]. This trade-off presents an enormous challenge and the solution lies on the the close monitoring of neonatal complications and anthropometric measurements, as well as individualized feeding strategies for preterm infants in varied conditions. The genetic potential of catch-up growth in SGA infants In our study, SGA infants had larger gestational age and higher incidences of optimal catch-up growth during hospitalization than the AGA and LGA groups, which was an almost ubiquitous consequence due to their genetic potential [22]. They were predisposed to a large size but were limited by the prenatal factors such as HDP and multiple pregnancy [5]. In our data, SGA infants experienced less physiologic weight loss and days to regain birth weight despite the higher incidences of glycometabolism disorder and feeding intolerance than AGA and LGA newborns. And with the medium gestational age above 34 weeks, their lungs were better developed and had less needs for pulmonary surfactant therapy and invasive ventilation. This genetic power enabled them to reach higher energy intakes at discharge and therefore erase the prenatal de cit and facilitates the catch-up growth.
Etiological factor associated with catch-up failure The catch-up failure group had higher incidences of pulmonary surfactant use, invasive ventilation, BPD, hyperbilirubinemia and feeding intolerance. Ehrenkranz et al. [23] found that invasive ventilation used in preterm infants for the rst 7 days after birth was related to signi cantly less nutrition intakes during the rst 3 weeks. The therapeutic interventions used for lung disease, such as corticosteroids and diuretics, together with insu cient nutrition caused by uid restriction, might be the reasons for catch-up failure in patients with BPD [12]. Typically, BPD is not diagnosed until 28 days after birth, so it is important to implement early parenteral and enteral nutrition on critically ill premature infants at risk for BPD. In addition to respiratory diseases, phototherapy uesd in Infants with hyperbilirubinemia causes more the uid loss, and feeding intolerance re ects the malfunction of the metabolic system and affects the absorption of nutrients, subsequently resulted in catch-up failure [12].
Nutrition factors contributing to optimal catch-up growth In our study, the initiation of both EN and PN were delayed comparing with the Chinese guideline of nutrition support for neonates in our study for fear of complications such as feeding intolerance, NEC, or upper gastrointestinal hemorrhage. According to the Chinese guideline of nutrition support for neonates, EN and PN support with amino acids can be safely administered within 24 hours after birth [10]. Moreover, the energy intake on the rst day after PN termination was far less than 100 kcal/kg/d, meeting the guideline in both groups, which means that the PN support was stopped prematurely. This preterm termination of PN directly led to the deceleration of growth, thereafter elevating the incidence of catch-up failure at discharge We suggest that the pediatricians determine the enteral energy and consult nutritionists to perform a comprehensive nutrient assessment before deciding on PN termination. Nutritional support has to keep up with medical practices in order to reduce morbidities and impairment of physical and neural growth in the both acute and delayed form [24,25].
Several studies have established that human milk should be the rst choice for feeding preterm and low-birth-weight newborns [26]. Breast milk protects against NEC, and the introduction of formula, even as a human milk forti er rather than as the primary milk, increases the risk of NEC and puts very preterm infants at greater risk for catch-up failure [27]. Nonetheless, the percentage of breast milk feeding in the current study was merely 15% and had no obvious effect on growth. The explanation might be that the de nition we applied here is not exclusive breastfeeding, but breastfeeding for more than half of days in the hospital, which interfered with the nutrition intake and offset the bene t of breast milk [28]. Further research is needed to elucidate the relation between exclusive breastfeeding and growth outcomes.

Conclusions
On balance, the extent of optimal catch-up growth should be just enough to avoid the deleterious effect on neurocognitive function, but not too excessive to induce the disorders of metabolic system. Hence, we considered the optimal catch-up growth during hospitalization as an increase of Z-scores from birth to discharge for at least two of three parameters (weight, height and HC). Whether this de nition could indicate the ideal growth outcomes of preterm infants still needs further research.
Small, appropriate, and large for gestational age preterm newborns had different growth trajectories, where SGA infants showed greater potential in predischarge optimal catch-up growth, and LGA infants exhibited a tendency of downward cross of centiles. Early abundant nutrition and avoiding BPD are essential for achieving optimal catch-up growth during hospitalization.        SGA, small-for-gestational-age;; BPD,bronchopulmonary dysplasia. Figure 1 Growth outcomes of SGA, AGA and LGA infants SGA, small-for-gestational-age; AGA, appropriate-for-gestational-age;

Figure 2
Changes in Z-scores for weight, height and HC in preterm infants. HC, head circumference; SGA, small for gestational age; AGA, appropriate for gestational age; LGA, large for gestational age.