Study design
A cross-sectional study was carried out from 2018 to 2019 with 43 preterm infants (Preterm Group, gestational age from 24 to 33 weeks), with age 9 to 24 months, at the follow-up clinic of the Hospital Municipal Universitário de São Bernardo do Campo (HMU-SBC). The comparison group consisted of 47 healthy term infants (Term Group), adequate for gestational age and weighing more than 2,500 grams, of the same age, in follow-up at the Primary Care Health of the same city.
The HMU-SBC adopts the Kangaroo Method. The place is accredited as a Baby-Friendly Hospital and is a reference for high-risk pregnancies in the city. About 14% of births are preterm, and the breastfeeding rate for preterm infants at hospital discharge is 82%. Preterm newborns with gestational age < 34 weeks or birth weight ≤ 1,500 grams are referred to and monitored at the outpatient clinic up to the age of six by a multidisciplinary team consisting of a pediatrician, a psychologist, a social worker, and a physiotherapist. The follow-up is complementary to that performed by the Primary Care Health.
There were excluded from the sample infants with severe malformations (cardiac, abdominal wall, musculoskeletal, and central nervous system defects), genetic syndromes, chronic non-progressive encephalopathy, oxygen-dependent children, those who did not feed exclusively orally, who had intolerance and food allergies, who were unable to provide telephone contact for invitation to be included in the study, and whose family refused to participate. A total of 43/69 (62.3%) and 47/78 (60.2%) of the eligible infants in the preterm and term groups, respectively, were included in the final study sample.
The Research Ethics Committee of Universidade Federal de São Paulo approved the study under Opinion N° 2.937.127 and we confirm that all methods were perfomed in accordance with the Declaration of Helsinki. The children’s legal guardians signed the informed consent form after the interview and clarifications by the researchers regarding the study’s steps and procedures.
Collected data
General informations
We obtained information on socioeconomic status and maternal health during pregnancy, such as maternal age, parental schooling, per capita income, pre-gestational nutritional condition, and diseases reported during pregnancy.
Data were collected on weight, height, head circumference, gestational age, causes of prematurity, and Apgar score regarding birth. The gestational age was calculated, firstly, according to the date of the last menstruation, secondarily, the ultrasonography data of the first trimester was adopted or the subjective evaluation of the newborn [17].
Birth weight was used to evaluate the adequacy for gestational age using the reference proposed by INTERGROWTH-21 [18]. We considered small (SGA), adequate (AGA), and large (LGA) when the birth weight for gestational age was below the 10th percentile, from the 10th to 90th, and above 90th, respectively.
Anthropometry
The principal investigator obtained all anthropometric and body composition measurements. At the time of the evaluation, weight data were measured on a digital, platform-type scale graduated in grams, with the infant undressed and without diapers. Height was measured with a horizontal stadiometer graduated in millimeters; head and arm circumference with an inextensible measuring tape; and the tricipital and subscapular skinfolds with a calibrated Langer® type adipometer [19].
These measures were used to calculate anthropometric indicators as Z-scores of body mass index (BMIZ), height/age (HAZ) and head circumference/age (HCZ), through the WHO Anthro v.3.2.2. The cutoff points used for the classification were those proposed by the World Health Organization [20]. The corrected age of 40 weeks was used to calculate anthropometric indicators for preterm infants.
Food consumption
Information on type and duration of breastfeeding (exclusive and total), use of infant formulas and whole cow’s milk, age at onset and sequence of introduction of complementary, and iron supplementation frequency and dose were collected.
Three 24-hour recalls were applied to assess food intake, with a maximum interval of 15 days between them. The estimated dietary intake was assessed by averaging the intake values of each nutrient in the recalls available for each infant. The calculation was performed using the DietWin® program, which uses the food composition tables proposed by the United States Department of Agriculture [21] and the Brazilian Food Composition Table [22].
Three, two, and one 24-hour recall were available in 24 (26.7%), 16 (17.7%), and 45 (50%) of the infants included in the study. They were used to calculate total daily ditary intake of energy (kcal and kcal/kg), protein (g and g/kg), iron (mg) and zinc (mg). The main meals were defined as food consumed at lunch or dinner per the traditional Brazilian eating habits, in general, rice, beans, meats, and vegetables [23] and complementary feeding such as the combination of main meals and snacks.
The consumption of infant formula and whole cow’s milk was not added to the calculation of complementary feeding to enable comparison between breastfed and non-breastfed infants. We could not quantify the volume of milk consumed in breastfed children, but the number of times (frequency) that the child received breastfeeding during the day was recorded.
Laboratory tests
A total of 8 ml of blood was collected by peripheral venipuncture with a fasting period of 3 hours and in the morning. The blood aliquot was divided into tubes for collecting metals (Vacuette tube®, without additive), dry tube, and EDTA tube.
The material was immediately transported to the Clinical Analysis Laboratory of the CU-FMABC, where the sample preparation, laboratory analysis, and storage were carried out. Samples that were not analyzed immediately were stored in a freezer at -80°C.
The blood count was performed with the multi-parameter automated hematology analyzer (Cell-Dyn Ruby) using the Multi-Angle Polarized Scatter Separation technology and laser flow cytometry. Anemia was defined by the presence of hemoglobin (Hb) below 11 g/dL [24]. The ultra-sensitive C-Reactive Protein (CRP) was measured by turbidimetry.
Serum and erythrocyte zinc levels were determined by the inductively coupled plasma mass spectrometry method (ICP-MS). Red cell lysis was performed with phosphate buffer. The reference values adopted for serum and erythrocyte zinc were 65 μg/dL [4] and 40 μg/g hemoglobin (µg/gHb) [25], respectively.
Statistical analysis
The collected data were tabulated and consolidated in Excel spreadsheets (Office Microsoft ®), and the analyses were performed in the statistical package SPSS 25.0 (IBM®). Categorical variables were presented as absolute numbers and percentages and compared with the Chi-square test. The distribution of continuous variables was assessed using the Shapiro-Wilk test, histograms, and Kurtosis values. The variables with parametric distribution were presented as mean±standard deviation and compared by the Student's t-test for independent variables. Variables with nonparametric distribution were presented as medians and interquartile ranges (p25 - p75) and compared with the Mann-Whitney test.
The Enter linear regression method was adopted for the multivariate analysis, using serum and erythrocyte zinc as dependent variables. The independent variables were included in the model after the analysis of collinearity, which showed a statistically significant difference between the groups and those with clinical relevance. A significance level of 5% was adopted in all analyses.
Employing α-bidirectional = 0.05 and β = 0.20 allowed the included sample (45 infants per group) to detect a difference of 10 µg/dL of serum zinc between the groups (standardized magnitude of effect of 0.6). For this calculation, we used data from the paper published by Cho et al., 2019 [13], which found mean and standard deviation in serum zinc levels of 81.4 ± 18.7 µg/dL in a group of preterm infants.