Participants
This is a longitudinal observational study performed with a convenience sample of mother-child pairs, selected between September 2011 to December 2016. The study is part of a larger prospective controlled project entitled Impact of Perinatal Environment Variations on the Health of the Newborn in the First Six Months of Life (IVAPSA) study. Further details about the IVAPSA study [20], as well as some baseline results [21] were previously published.
According to the IVAPSA study, the final sample size estimated was 521mother-child pairs, being 87 for each group of intrauterine environments and 174 pairs for the control group. The sample size calculation was described in detail in a previous publication [20]. Newborns were recruited at three hospitals in the city of Porto Alegre: Hospital de Clínicas de Porto Alegre (HCPA), Hospital Fêmina and Hospital Nossa Senhora da Conceição, the latter two belonging to the Grupo Hospitalar Conceição (GHC) (Fig. 1).
The infants were randomly selected based on hospital medical records. The data collection was carried out with interviews at 6 different times after delivery; 24–48 h, 7 days, 1 month, 3 months, and 6 months. These interviews occurred in the infants’ home or the Clinical Research Center of HCPA. Inclusion criteria consisted of pairs attended in the hospitals and belonging to the area covered by Porto Alegre city, Rio Grande do Sul. Mothers with HIV (Human Immunodeficiency Virus) positive test, newborn twins, who had congenital diseases at birth, gestational age less than 37 weeks, those who required hospital admission, those with a birth weight below 500 g, and pairs who did not attend the six-month interview were excluded from the study.
The IVAPSA sample consisted of mother-child pairs who had different backgrounds of adverse intrauterine environments exposure, such as tobacco, determined by affirmative answer for smoking during pregnancy, regardless of the number of cigarettes used per day; DM, defined as gestational diabetes, type 1 DM or type 2 DM; HD, designated by hypertensive disorders during pregnancy, whether it was classified as preeclampsia and eclampsia, preeclampsia superimposed on chronic hypertension, chronic hypertension; control, represented by mothers who did not have hypertension or DM, as well as those who were non-smokers; and IUGR, characterized by mothers with term infants small for gestational age who were below the percentile 5, according to Alexander curve parameters [22], whose causes were not smoking, HD, or DM. Women with more than one background were excluded from the study. The questionnaires used in this study were specially developed for the IVAPSA research, and are available as supplementary material (Supplementary Files 1, 2, 3 and 4).
Covariates
Covariates were collected in the first and sixth interviews, using structured questionnaires on maternal demographic and socioeconomic information, pregnancy, childbirth and the infant variables. Maternal data were age (in years), education (years of study) and total family income (in Reais – Brazilian currency) and the number of past pregnancies (prior live births). Gestational data consisted of pre-gestational body mass index (BMI), expressed in kilogram by stature squared, Kg/m2, and the group of gestational clinical conditions (DM, HD, tobacco, IUGR and control). Infant data included sex, growth z-scores at birth, the pacifier and bottle-feeding use at 6 months, as well as the timing of introduction of cow’s milk or formula milk (in days).
Feeding practices
For infant feeding practices analysis, data were collected from the second to the sixth interview, referring to infant feeding at hospital discharge; time in minutes of initiation of BF at postpartum, questions related to BF (e.g., Did the child breastfeed on the first day of life? If don’t, what did he receive? What was the feeding at hospital discharge? Does your baby breastfeed? If don’t, when did you stop breastfeeding?), weaning and introduction of CF.
The recommendations of infant nutrition analyzed were those proposed by “Feeding Guide for Brazilian Children Under Two Years” [23] and “Dietary Guidelines for the Brazilian Population” [24], both publications of the Ministry of Health of Brazil.
The BF variables were elaborated according to the World Health Organization (WHO) [25]: EBF, when the child received exclusively BF until 6 months, without any other liquid or solid, except for supplements and medicine and non-exclusive (non-EBF), when the child was breastfeeding, independent of other liquids or feeding offered. Then, a compute of the days of EBF-duration and non-EBF-duration was performed.
Foods consumed in the 24-h dietary recall, collected from the third to the sixth interview regarding all foods consumed in the day before, were transposed in Excel® version 2101 (Microsoft Office® Professional Plus pack), and categorized according to the degree of food processing (NOVA classification), proposed by “Dietary Guidelines for the Brazilian Population”, according to Monteiro and collaborators [26]:
- Natural or minimally processed foods: foods obtained directly from plants or animals, and may or may not be subjected to minimum changes;
- Culinary ingredients: products extracted from natural foods or directly from nature and used for seasoning, cooking, and create culinary preparations;
- Processed foods: products made essentially from the addition of salt or sugar to a natural or minimally processed food;
- Ultra-processed foods: products whose manufacturing involves various stages and processing techniques and various ingredients, many of them exclusively for industrial use.
The last 3 categories of processing (culinary ingredients, processed and ultra-processed foods) were grouped as “processed foods” (PF) for analyzes. After the compilation of the three groups of PF, a weighting was carried out by the numbers of items consumed, scoring more severity the earlier the consumption.
The gravity score of processed foods (GSPF) calculation occurred as follows: infants who consumed PF during the first 6 months of life were categorized according to the time of introduction of reported PF. To perform this categorization, a weighting was made by pointing each reported food item considered as PF with 10 points in the third interview, which represented the intake from the first 15 days of life; 7 points in the fourth interview, which represented the intake from 15 to 30 days of life; 5 points in the fifth interview, which represented the intake from 30 to 90 days of life; and 3 points in the sixth interview, which represented the intake from 90 to 180 days of life.
That is, the same food, if it was offered in more than one data collection, would be scored every time it was mentioned. Afterward, a sum of all points was performed and divided into tertiles, generating groups of mild severity (1–12 points), moderate (13–30 points), and severe (more than 30 points). Infants who did not consume any type of PF during the first 6 months of life, obtained null weighting.
Child growth
Measurements of weight (in kilograms) and height (in centimeters) were collected at birth and in the 6-month interview. Child growth was assessed by WHO Anthro® software (Switzerland, 2005). The growth parameters were evaluated according to the anthropometric indices weight for height (WHZ), body mass index for age (BAZ), and weight for age (WAZ), proposed by WHO [27].
Statistical analysis
Data were processed and analyzed with the statistical program SPSS®, version 18.0 (PASW Inc., Chicago, IL, EUA). Qualitative variables were expressed by absolute number and percentage, and Pearson’s chi-square test was used to detect differences between proportions. The Shapiro-Wilk test was used to test the normality of quantitative variables. Quantitative variables were expressed by the median values [interquartile range], and analyzed by the Kruskal-Wallis test with the Dunn Post Hoc test.
To compare the averages of growth variables between the different GSPF, it was performed an analysis of variance test (ANOVA), with the Tukey Post Hoc test. For asymmetric variables, the distribution of BF duration (exclusive and non-exclusive) was compared using the Kruskal-Wallis test; when significant, it was compared by Dunn Post Hoc test.
To perform the adjusted comparison for variables age, maternal pre-gestational BMI, income, and growth z-scores at birth, the BF time variables (exclusive and non-exclusive) were transformed into logarithms, correcting the values asymmetry. The comparison of growth variables and BF time logs were performed using the ANCOVA test. Results was presented by geometric mean and 95% confidence interval. For all analyzes, a significance level of 5% was considered.
Ethical aspects
The mothers signed, in duplicate, the Free and Clarified Consent Term, at the time that preceded the first interview in the first 48 h after delivery. An identification number was defined for each child, maintaining the anonymity of the participants. The study was approved by the Research Ethics Committee of HCPA and GHC, under protocol numbers 11–0097 and 11–027, respectively.