The relationship between serum retinol concentrations and subclinical infection in rural Brazilian children

Objective To evaluate the relationship between serum retinol concentrations and subclinical infection in children from rural settlements.


I N T R O D U C T I O N
The main manifestations of Vitamin A Deficiency (VAD), an essential micronutrient related to cell growth, immunity and vision, are associated with increased rates of infant morbidity and mortality [1].It is estimated that 190 million children under the age of 5 years worldwide have low serum retinol concentrations (<0.70μmol/L).The prevalence of VAD in Brazil is high among children aged 6 to 59 months [2], with a predominance of the condition in the northeastern, where there have been rates around 20% [3].
Populations living in poor rural areas tend to be more vulnerable to nutritional deficiencies and infection.In Brazil, the latest form of rural organization, called "settlements", established http://dx.doi.org/10.1590/1678-98652018000300004by government policies has grown extensively over the past few years.However, most of these settlements still do not have the necessary infrastructure, predisposing the population to infections [4,5].
Within this context, the objective of the present study was to assess the relationship between serum retinol concentrations and subclinical infection in children aged 6 to 59 months living in the settlements of the northeastern region of Brazil.

M E T H O D S
A cross-sectional, population-based study was conducted in the settlements of the Instituto Nacional de Colonização e Reforma Agrária (INCRA, National Institute for Colonization and Agrarian Reform), Teresina, Piauí, Brazil.Teresina is the capital of the state of Piauí, one of the poorest states in the country, located in the northeastern region of Brazil.In 2010, the mean per capita income was R$757.57(approximately US$253.00) and the Human Development Index (HDI) was 0.751 [6].According to INCRA data [7], Teresina has nine settlements created through the Federal Agrarian Reform between 2005 and 2008, which benefit 475 families within a total area of 6,746.97hectares.
All children within the eligible age range from the nine settlements were invited to participate in the study, for a total of 132 children.There were 14 losses, including 3 (2.27%)due to refusal of the responsible person, 6 (4.54%) due to the absence of the legal guardian of the child at the residence, and 5 (3.78%) because of insufficient blood sample for biochemical analysis.Thus, the final sample consisted of 118 children who met the following inclusion criteria: living in the settlement, age 6 to 59 months, and free and formal consent of the parents who agreed to the child's participation by signing the free informed consent form.

Data collection
The data were collected from May to October 2013.The field team, which was trained by the supervisors of the project, consisted of nutritionists and nutrition students from the Universidade Federal do Piauí (UFPI, Federal University of Piauí).

Socioeconomic and demographic variables
Socioeconomic and demographic data of the families were collected using questionnaires during the home visits.The criteria of the Brazilian Association of Research Companies [8] were used for the classification of the socioeconomic status of the families.At the end of the interview, a date was scheduled for blood collection, anthropometry, and evaluation of food intake and health of the child.

Measurement of serum retinol and C-Reactive Protein (CRP)
Blood samples (3-5 mL) were collected by venipuncture into tubes without anticoagulant (Greiner Bio-One, São Paulo, Brazil) wrapped in aluminum foil for protection from light.Maintaining the cold chain, the samples were transported to the Laboratory of Experimental Nutrition, Department of Nutrition, UFPI, where they were centrifuged at 3,000rpm and divided into two serum aliquots.The two aliquots were stored in amber and transparent microtubes for the measurement of retinol and C-Reactive Protein, respectively, at 80ºC until the time of analysis.[9].VAD was diagnosed when serum retinol concentrations were less than 0.70μmol/L.With respect to the level of epidemiological importance, the following prevalence classification was adopted: 2 to 10%, mild public health problem; 11 to 19%, moderate public health problem, or >20%, serious public health problem [10].

Measurement of CRP
C-Reactive Protein was measured in the samples at the Laboratory of Clinical Analyses, University Hospital, UFPI, by an immunoturbidimetric assay in an automated Cobas Integra Plus 400 biochemical analyzer (Roche ® Diagnostics, Mannheim, Germany) using the CRP kit (Cassete C-Reactive Protein Roche ® ).A level of CRP>5mg/L was defined to indicate the presence of infection or of an inflammatory process.

Anthropometry
The weight of the children was measured with a digital scale (Plena ® , São Paulo, Brazil; capacity of 150kg) to the nearest 100g.Children younger than 2 years were weighed on the arms of the responsible person whose weight was deduced from the total weight.Height was measured with a fabricated infantometer (measuring range of 100cm and subdivisions in millimeter) in children younger than 2 years and with a stadiometer (Seca ® , Hamburg, 206, Germany) containing a 2.2 measuring tape and subdivisions in millimeter in children older than 2 years.Nutritional status was evaluated based on height/age, weight/age, weight/height, and Body Mass Index (BMI)/age indices [11].

Food intake
Food intake was evaluated using a 24-hour diet recall by interview with the mother or responsible person of the child.If the children attended daycare centers or nursery schools, the employees responsible for the preparation and distribution of the child's meal at the institutions were also interviewed.Photographic records of the utensils and portions [12] were used to help the interviewed subjects identify the amounts of foods ingested and conversion tables [13,14] were applied to transform the home measures into gram (g) or milliliter (mL).
For the estimation of intrapersonal variability, a second 24-hour recall was applied to 40% of the sample selected randomly after an interval of 2 months, repeating the same procedure as used in the first interview.Habitual nutrient intake was estimated using the Multiple Source Method (version 1.0.1)[15].
The amounts of nutrients were calculated using the Nutwin 1.5 software [16].The foods not found in the database of the program were inserted based on chemical composition tables of the foods [17][18][19].For processed foods not included in these tables, the nutritional information on the labels of the products was considered.Regional preparations were included in the program according to the ingredients and home quantities described in the form.
The intake of vitamin A (μg), proteins (g), lipids (g), iron (mg), and zinc (mg) was evaluated, the last four because they are involved in the bioavailability of vitamin A [20][21][22].The reference values proposed by the Institute of Medicine were used for the evaluation of nutrient intake [23].The Estimated Average Requirement (EAR) was adopted as the cut-off for calculation of the prevalence of inadequacy.In the case of nutrients for which no EAR was available, the Adequate Intake (AI) was used.In these cases, it was not possible to estimate the prevalence of inadequacy.The volume of milk consumed by breast-fed children was estimated according to the Brazilian Food Guide for children under the age of 2 years [24]. http://dx.doi.org/10.1590/1678-98652018000300004

Data of the child
A questionnaire was applied to the person responsible for the child to obtain data regarding breast-feeding practices, presence of morbidities in the last 15 days, and use of vitamin complexes.The date of birth of the child and information about vitamin A supplementation in the last 6 months were collected from the Child's Health Card.

Statistical analysis
The data were processed and analyzed using the Statistical Package for the Social Sciences (SPSS Inc., Chicago, Illinois, United States) 20.0.The Kolmogorov-Smirnov test was used to evaluate the normality of the dependent variable (serum retinol concentrations).Natural logarithmic transformation was necessary to obtain normality of the data.Levene's test was used to verify the homogeneity of variances.In bivariate analysis, the Student t-test and analysis of variance were used for comparison between serum retinol concentrations and variables with two or more categories, respectively.Numerical variables were compared using Pearson's correlation coefficient.A level of significance of 5% was adopted.The variables showing p<0.20 in bivariate analysis were included in the multiple linear regression model (backward stepwise method).In this case, significance was assumed when p<0.05.

R E S U L T S
Among the 118 children studied, 59.3% were girls, with a mean age of 30.37 (16.68) months (Table 1).The mean serum retinol level was 1.33 (0.57) μmol/L and the prevalence of VAD was 9.3%.
As can be seen in Table 2, vitamin A intake was inadequate in almost 100.00% of children aged 48 to 59 months (99.98%).Another finding that calls attention is the high percentage of inadequate iron intake, especially among children aged 7 to 11 months (99.66%).Mean zinc intake was higher than the EAR in children older than 12 months, with low percentages of inadequacy.With respect to macronutrients, mean protein intake was above the EAR and AI for the corresponding age groups.Mean fat intake was lower than the AI in children under the age of 1 year.
Analysis of the categorical variables showed that serum retinol concentrations were significantly lower in children living in dwellings made of mud and unfinished masonry (p=0.025) and in those diagnosed with subclinical infection (p=0.008).With respect to the numerical variables studied, vitamin A intake and retinol concentrations showed a weak and positive, but significant association (r=0.118;p=0.049) (Table 3).
The final multiple linear regression model (Table 4) showed that the presence of subclinical infection caused a reduction of 0.728μmol/L in retinol concentrations (p=0.008),explaining 51.1% of the variability in serum retinol in the population studied.

D I S C U S S I O N
The mean serum retinol concentrations of the children studied were within the normal range; however, regarding the concentrations identified as low, the percentage found (9.3%) classifies VAD as a mild public health problem.
Low vitamin A intake, as well as infectious processes, can cause a reduction in blood retinol concentrations.In the first case, there is a decrease in liver stores.In the second case, vitamin A concentrations stored in the liver may be normal, but mobilization of the vitamin is suppressed as a result of the low synthesis of the retinol transport protein, also as a consequence of infectious processes [25].In this study, the presence of subclinical infection was inversely associated with serum retinol concentrations.Similar results have been reported in other studies.Rohner et al. [26] showed a significant increase in the risk of VAD in children diagnosed with inflammation, demonstrated by elevated concentrations of CRP and alpha-1-acid glycoprotein.
The prevalence of VAD was 9.3%, but the prevalence of subclinical infection, assessed by CRP concentrations, was 11.9%.It is important to emphasize that retinol-binding protein is a negative acute phase protein, and that the occurrence of a subclinical infection can interfere with its serum concentrations [27].Excluding from the final sample (n=118) the children with subclinical infection (n=14), the prevalence of VAD is reduced to 7.7%.This reduction may reflect an overestimation of 1.6% of the overall prevalence of low serum retinol concentrations, in agreement with the study of Silva et al. [28] who reported an overestimation of 4.1%.
With respect to the variables related to food intake, vitamin A intake was significantly associated with serum retinol concentrations, although this association was no longer observed in the final Multiple Linear Regression (MLR) model.The perception of the influence of vitamin A intake on serum retinol concentrations has been shared by some authors [29][30][31].
One of the limitations of the present study, in addition to those inherent to dietary surveys, was the fact that two persons were interviewed (the mother or responsible person of the child and employees of daycare centers or nursery schools).Additional limitations include the lack of control of seasonality, since some foods such as fruits are generally influenced by this factor [32], and the existence of different food composition tables, which often do not contain adequate information about the content of regional foods.Despite these limitations, care was taken in this study to minimize potential measurement errors resulting from the use of the data collection instrument.

C O N C L U S I O N
The variables associated with serum retinol concentrations in this population were the presence of subclinical infection.Vitamin A deficiency is a mild public health problem in the children studied.These results indicate the need to implement public policies designed to control infectious diseases and nutritional interventions that stimulate the consumption of foods containing vitamin A in order to prevent and/or combat not only VAD, but also other nutritional deficiencies that are common in this age group.

C O N T R I B U T O R S
LFS TELES participated in the development of the study protocol, data collection and analysis, interpretation of the results, and writing of the manuscript.AA PAIVA was responsible for the design of the study and coordination of the research, and participated in the interpretation of the results and writing and revision of the manuscript.LA LUZIA performed the retinol analysis and participated in the revision of the manuscript.FEL LIMA-FERREIRA participated in the analysis of food intake and revision of the manuscript.CMRG CARVALHO participated in the data analysis and the revision of the manuscript.PHC RONDÓ participated in the revision of the manuscript and retinol analysis.
*Interviewee could not answer.† The child had not health booklet.‡ Assessed by C-Reactive Protein concentrations >5mg/L.BMI: Body Mass Index; SD: Standard Deviation.
* IA: Inadequate intake as reference value.† Dietary reference intake not determined.Other nutrients = EAR: Estimated Average Requirement.BMI: Body Mass Index; SD: Standard Deviation.
* Student t-test and analysis of variance for comparison between serum retinol concentrations and variables with two or more categories, respectively.† Pearson's correlation coefficient.‡ Assessed by C-Reactive Protein concentrations >5mg/L.BMI: Body Mass Index; SD: Standard Deviation.