Extrauterine growth restriction: Universal problem among premature infants

1 Universidade Federal de Viçosa, Programa de Pós-Graduação em Ciência da Nutrição, Departamento de Medicina e Enfermagem. Av. P.H. Rolfs, s/n., Campus Universitário, 36570-900, Viçosa, MG, Brasil. Correspondência para/Correspondence to: BAC FREITAS. E-mail: <brunnella.freitas@ufv.br>. 2 Universidade Federal de Viçosa, Programa de Pós-Graduação em Ciência da Nutrição, Departamento de Nutrição e Saúde. Viçosa (MG), Brasil. Article based on doctoral dissertation of BAC FREITAS, intitled “Restrição do crescimento extrauterino, amamentação e avaliação da adesão e influência da suplementação com micronutrientes nas prevalências de deficiências de ferro, zinco e vitamina A em prematuros”. Universidade Federal de Viçosa; 2015. Extrauterine growth restriction: Universal problem among premature infants


I N T R O D U C T I O N
Extrauterine Growth Restriction (EUGR) in premature infants is an object of studies and results from a complex interaction between genetic and environmental factors, which affect nutritional requirements, endocrine disorders, and treatments 1,2 .
Growth involves genetic potentials that may or may not be reached, depending on the individual's life conditions from conception until adulthood. Prematurity is a determinant for stunting, with repercussions on adulthood 3,4 . Premature infants who are small for gestational age or with EUGR at hospital discharge are at greater risk of stunting and its long-term consequences 1,2 .
Extrauterine growth restriction is a marker of severe nutritional deficit during the first weeks of life, which can be assessed at 36 weeks of corrected age or at discharge based on reference intrauterine or postnatal growth curves 3,7,8 . Fenton & Kim 9 curves, which became available recently, are considered the standard postnatal growth curves [9][10][11] . As gestational age and birth weight decrease, weight regain becomes harder, consequently increasing the incidence of EUGR. To decrease the incidence of EUGR and its consequences, protocols that improve the nutrient supply for premature infants, especially infants at higher risk, must be implemented and updated 12,13 .
The survival of increasingly premature infants has resulted in many studies With Very-Low-Birth-Weight (VLBW) infants. However, there is a scarcity of studies with premature infants with a gestational age of 32 weeks or more, especially because 70% of premature infants have a gestational age of 34 weeks or more 14 .
In face of the above, the following questions arise: what is the growth rate of premature infants during the first weeks of hospital stay? How do gestational age, birth weight, and morbidities affect growth rate? What factors are associated with EUGR? Are there modifiable factors?
In order to answer these questions, this study assessed the growth rate of premature infants during the first weeks of extrauterine life and investigated the factors associated with extrauterine growth restriction at discharge from a Neonatal Intensive Care Unit (NICU).

M E T H O D S
This cross-sectional, retrospective study used secondary data from the medical records of premature infants admitted to the NICU of the Hospital São Sebastião in Viçosa (MG) from January 1, 2008 to December 31, 2010. weight Z-score ≤-1.29 (10%). The infants were also classified as appropriate for gestational age when they were between the tenth and ninetieth percentiles and large for gestational age when they were above the ninetieth percentile 9,20 .
Fenton & Kim 9 curves were used because they were based on a meta-analysis with the highest number of newborns to date and their respective data; the authors used the data to construct new curves for newborn growth assessment according to gestational age. According to the Portuguese Society of Pediatrics 11 , these curves approach a new standard curve. In addition to assessing fetal growth at birth, they allow monitoring the postnatal growth of premature infants until they reach a corrected age of 50 weeks as the data were harmonized with the World Health Organization 21 curve from the 40 th week onward 9,11 . Qualitative explanatory variables: gestational age was categorized according to the number of completed weeks: <28 weeks (extremely premature); between 28 and 31 weeks (very premature); and between 32 and 36 weeks (moderately premature) 16 ; sex was categorized as female or male; and the small for gestational age variables bronchopulmonary dysplasia 22 , lateonset neonatal sepsis (occurring after the first 48 hours of life) 23 , and transfusion of red blood cell concentrates were categorized in a binary fashion (yes/no). Quantitative explanatory variables: gestational age (weeks and days); corrected age at discharge (weeks and days); birth weight and weight at discharge (grams); birth weight and weight at discharge Z-scores; length of hospital stay (days); duration of mechanical ventilation (days); total days of supplemental oxygen (days); age at start of enteral diet (days); age at start of parenteral nutrition (days); age at start of adequate enteral intake (days; considered as 150 mL/kg/day) 24,25 ; and days without enteral diet.
The study population was characterized by frequencies and measures of central tendency and dispersion. The Kolmogorov-Smirnov test showed that only the variables weight and weight Z-scores had normal distribution, so they were expressed as means and standard deviations. The other quantifiable variables were expressed as medians, 25 th percentile, and 75 th percentile. The qualitative variables were compared by the Pearson's Chi-square test or Fisher's exact test, and the quantitative variables by the Mann-Whitney, Kruskal-Wallis, or Student's t tests.
Bivariate analyses were performed between the explanatory and outcome variables, obtaining, as a measure of effect, Prevalence Ratios (PR) using Poisson regression 26 . Variables with p<0.20 were included in multivariate analysis, and variables with p<0.05 were maintained in the final model.
The weight curves were constructed by adjusting a third-degree polynomial function to smooth the curves and thereby suppress small variations in their curvatures.
The programs Excel (Microsoft Office, Redmond, Washington, United States) version 2010, Stata (Stata Corporation, College Station, Texas, United States) version 9.1, and Statistical Package for the Social Sciences (SPSS Inc., Chicago, Illinois, United States) version 20.0, for Windows were used for creating the database, and codifying, entering, and statistically analyzing the data.
The study was approved by the Research Ethics Committee of the Universidade Federal de Viçosa, under Protocol nº 063/2011, and did not require Signed Informed Consent. The study ensured the secrecy and confidentiality of all study participants' data.

R E S U L T S
A total of 293 premature infants were eligible during the study period. However, 13.3% of the premature infants (n=39) died and were excluded from the study, resulting in a population of 254 premature infants. Table 1 characterizes the study population, which had an EUGR prevalence of 24.0%.
Since the first weeks of life are an important time for growth, curves of the median weight of the premature infants were constructed according to their gestacional age at birth for the first four weeks of life ( Figure 1).
The median birth weight tended to increase as gestational age increased. At six days of life, weight had decreased. At thirteen days, premature infants with gestacional age <33 weeks had regained their weight. However, a critical point appeared for premature infants with gestational age >33: they had had no weight regain by the thirteenth or the twentieth day. At 27 days of life, premature babies with gestational age at birth >33 weeks had not yet regained their weight.
Bivariate analyses of the premature infants' characteristics from birth until NICU discharge were conducted, establishing EUGR as outcome ( Table 2). The significant (p<0.05) variables were: female sex, small for gestational age, late-onset sepsis, smaller measurements at birth and   discharge, longer hospital stay, higher age at start of adequate enteral intake, and longer time without enteral diet.
These variables were submitted to bivariate Poisson regression for the outcome EUGR. The variables that persisted with p<0.20 were included in multivariate analysis: small for gestational age, sex, birth weight, neonatal sepsis, red blood cell concentrates, age at start of enteral diet, age at start of adequate enteral intake, and days without enteral diet. The variable gestational age was a confounding factor.
The variables small for gestational age at birth and days without enteral diet were analyzed by gestational age strata in an attempt to explain the behavior of the weight curves in Figure 1. Small for gestational age did not occur in premature infants with gestational age <28 weeks. Small for gestational age frequency was higher in the ≤32 gestational age ≤36 weeks' strata (p=0.001). This occurred because the number of deaths increased as gestational age decreased (p<0.0001), and all small for gestational age <28 weeks died (data not shown). The variable days without enteral diet also differed by gestational age stratum (p<0.0001): time without diet decreased as gestational age increased (Table 3).
Most (85.0%) of the infants born small for gestational age had EUGR at discharge, and 55.3% of the infants with EUGR were small for gestational age at birth. However, 44.3% of the EUGR were not small for gestational age. In a time line, 15.7% of the population was small for gestational age at birth, but 24.0% had EUGR at discharge. Figure 2 shows how the mean weight Z-scores changed from birth to discharge in small for gestational age and non-small for gestational age infants. Small for gestational age infants had a mean weight Z-score of -1.81 (±0.39) at birth and -2.74 (±0.78) at discharge. Non-small for gestational age infants had a mean weight Z-score of -0.06 (±0.75) at birth and -1.07 (±0.98) at discharge.
The differences between the weight Z-scores at birth and discharge measured by the Student's t test were significant for small for gestational age and non-small for gestational age infants (p<0.0001): the small for gestational age infants had a smaller weight Z-score loss (difference of 0.93), while non-small for gestational age had higher weight loss (difference of 1.01).

D I S C U S S I O N
Extrauterine grown restriction between birth and a corrected age of 40 weeks is a proven risk factor for stunting and neurodevelopment deficit in the first two years of life 3,27 . Therefore, it is important to unveil the factors involved in EUGR genesis and intervene accordingly.  Extrauterine grown restriction prevalence in the sample was 24.0%, and being small for gestational age at birth increased the likelihood of EUGR by 6.14 times, while each day without enteral diet increased its likelihood by 8.0%. In a Brazilian study, de Lima et al. 15 found that EUGR affects 26.0% of premature infants, and small for gestational age at birth has more impact than any other variable. Gianini et al. 28 also associated small for gestational age at birth with EUGR. Other studies have also found this association, but the EUGR rates varied from 16.0 to 63.0% [29][30][31][32] . Shan et al. 32  discharge. Data from the Neonatal Surveys Network of the National Institute of Child and Human Development (NICHD) found that 16% of premature infants were small for gestational age at birth, and 89% of these infants continued small at corrected age 36 weeks, corroborating the association between EUGR and small for gestational age at birth 31,33 .
In a previous study with this population, Freitas et al. 34 found a frequency of small for gestational age at birth of 11.3% using as reference Lubchenco's et al. 35 curves, confirming the association between small for gestational age at birth and higher odds of death, which was corroborated by Larroque et al. 36 However, if the data were revisited and the Fenton & Kim 9 curves were used, this prevalence would increase to 17.7%, which is important when considering the associations between small for gestational age at birth and the outcomes EUGR and death.
The variation in EUGR prevalence can be explained by sample differences, use of different reference curves, and use of different cut-offs to characterize the outcome. Although the American Academy of Pediatrics recommends that extrauterine growth rate should aim to equal intrauterine growth rate of a premature infant with same gestational age 2,29 , the postnatal growth curves seem more appropriate to follow their growth 37 , and currently, the Fenton & Kim 9 and Fenton et al. 10 curves approach the standard for classifying and following newborns until the corrected age of 50 weeks 11 . Lima et al. 15 also disagree with the American Academy of Pediatrics recommendations.
Extrauterine grown restriction was significantly associated with longer time without enteral diet. The median ages at start of parenteral and enteral nutrition were higher than the recommended values -early use of parenteral   nutrition in the first hours of life and minimum enteral diet on the first or second day 38 . However, adherence to the clinical protocols remains far from the ideal even in developed countries 38,39 .
Despite the absence of association in the present study, nutritional practices influence the growth rate of premature infants 7,40,41 , and recommendations should be followed because they are widely supported by scientific evidence.
The study morbidities did not affect premature infant growth, disagreeing with Gianini et al 28 . Other researchers have found that morbidities and care practices during hospital stay, including nutritional practices, influenced newborn growth rate 21,22,38,40,41 , and emphasized the importance of providing adequate amounts of energy and proteins 13,42 .
Growth rate and dynamics during hospital stay are under study. It is fact that the first two weeks of life represent a critical occasion in premature infant growth 13,42 and that inadequate growth during the first weeks after birth can be a marker of EUGR 3 .
The weight curves showed that the median weight of infants with gestacional age <32 weeks tended to recover by 13 days of life. Moreover, moderately premature infants (gestational age at birth between 32 and 36 weeks) tended not to recover the median birth weight until the twentieth or twenty-seventh postnatal day. According to the literature, premature infants need 8 to 24 days to recover their birth weight 28,43 . Initial weight loss, which usually occurs between the fourth and ninth postnatal day, especially on the first five days, is followed by an early growth peak, with growth rates that tend to mimic intrauterine rates from the second week onward 33,43,44 .
The results may have been influenced by the absence of small for gestational age in infants with gestational age <28 weeks due to higher mortality as gestational age decreased. Although gestational age did not remain in the final model, it was associated with EUGR in bivariate analysis, that is, as gestational age increased, the frequency of EUGR also increased, which may in part explain the behavior of the study curves.
There is evidence that the incidence of EUGR increases as gestational age and birth weight decrease, given that smaller premature infants lose more weight and take longer to regain the weight, and that as gestational age decreases, nutritional requirements increase and achievement of satisfactory growth rates becomes less likely. In other words, the chance of reaching a normal weight decreases as the time to achieve a normal growth rate increases 2,41,42,[45][46][47] . On the other hand, Shan et al. 32 associated EUGR with an increase in gestational age, which agrees with our results.
The smaller weight Z-score loss among small for gestational age infants allows one to understand the occurrence of EUGR in non-small for gestational age infants and to elucidate the impact of longer periods without enteral diet on EUGR genesis. Lima et al. 15 also found similar results for weight Z-scores. Embleton et al. 48 reinforce the multiplicity of factors involved in the development of EUGR and associate the accumulation of nutritional deficits in the first weeks after birth to nutritional practices, resulting in EUGR.
On the other hand, Ornelas et al. 49 found that the extrauterine growth rates of premature small for gestacional age and appropriate for gestational age infants differ and that small for gestational age infants have less inclined growth curves than appropriate for gestational age until the fortieth week. Ehrenkranz et al. 50 reported that appropriate for gestational age infants and infants who did not develop severe morbidities had higher growth rates.
Cooke et al. 2 and Cooke 51 affirms that extrauterine growth is influenced by a complex interaction of factors, that it is difficult to establish proper nutrition in ill and clinically unstable premature infants, and that when the energy and protein supplies are based on birth weight, the infants receive less food than they should.
In summary, the study showed that: 24.0% of the study premature infants developed EUGR; small for gestational age and more time without enteral diet were associated with higher odds of developing EUGR; EUGR affected 85.0% of small for gestational age infants; 15.7% of the study sample was born small for gestational age and 24.0% developed EUGR; 55.3 and 44.3% of EUGR cases involved infants who were and were not small for gestational age at birth, respectively; moderately premature infants tended not to regain their median birth weight by the third postnatal week, and this stratum contained a higher concentration of small for gestational age infants; and non-small for gestational age infants lost more weight Z-score between birth and discharge than small for gestational age infants, reinforcing the role of nutritional practices.
The present study emphasizes the need of more studies of moderately premature infants, which include most infants born before the thirtyseventh week of gestation. The limitations include the retrospective characteristic of the study, which entails information bias (medical records) and the lack of further nutritional details.

C O N C L U S I O N
Extrauterine growth restriction is a problem in premature infants, which is associated with small for gestational age at birth and longer periods without enteral diet. However, EUGR genesis is multifactorial, and the factors involved remain unknown. Based on the study results, we suggest more prospective studies with premature infants of all gestational ages and point out that the dissemination of information and implementation of good care practices to improve nutrient supply, individualized for small for gestational age infants, can minimize the problem.