Vitamin A defi ciency and associated factors in children in urban areas

OBJECTIVE: To estimate the prevalence of vitamin A defi ciency and its associated factors in children. METHODS: A cross-sectional population-based study, involving 1,211 children of both sexes, aged between six and 59 months old, was carried out in the urban zone of 9 cities in the state of Paraiba, Northeastern Brazil. Vitamin A status was assessed by serum retinol levels (high performance liquid chromatography – HPLC) and subclinical infection was assessed by C-reactive protein concentrations. Socioeconomic, demographic and sanitation conditions, as well as vitamin A supplement intake, were also evaluated. Children with serum retinol concentrations RESULTS: The prevalence of vitamin A deficiency was 21.8% (95%CI 19.6;24.2), showing an association with subclinical infection and lack of indoor plumbing. The prevalence of vitamin A defi ciency was 21.8% (95%CI 19.6;24.2). After adjustment, vitamin A defi ciency was found to be linked with subclinical infection and lack of indoor plumbing. Vitamin A defi ciency was four times higher (CI95% 1.49;10.16) in children with subclinical infection whose homes were without indoor plumbing, compared to children who were not infected and with indoor plumbing in their homes. CONCLUSIONS: Despite activities aimed at the prevention and control of vitamin A defi ciency, hypovitaminosis A, remains a public health concern among children under fi ve. DESCRIPTORS: Infant. Child, Preschool. Vitamin A Defi ciency, epidemiology. Risk Factors. Socioeconomic Factors. Cross-Sectional Studies. 10 Vitamin A defi ciency in children Queiroz D et al Vitamin A is a micronutrient known to be related to visual functions, epithelial integrity and the functioning of the immune system. Studies indicate that vitamin A defi ciency (VAD) is a public health problem in the Northeast and North and in some regions of the Southeast of Brazil.8,9,12,15,17,a,b In the state of Paraiba, this defi ciency has been recognized since the 1980s.3 Results of a clinical-nutritional study carried out in three regions of this state, between 1981 and 1982, show a high prevalence of Bitot’s spots and corneal scars in children in the sertão (desert region), especially in periods between harvests.16 Cases of corneal xerophthalmia (ulcer/queratomalacia) have also been documented in children in João Pessoa.1 In the following years, clinical manifestations of moderate xerophtalmia and corneal scars, were also observed in children aged between two and 28 months, mainly during the drought which lasted from 1981 to 1984.4 Approximately a decade later, a 16% prevalence of VAD was found in pre-school children in a population based survey, indicating a moderate public health problem.a Since1992 there have been no other population based investigations into this problem in the state. The main causes of VAD can be divided into two broad categories: malnourishment and the presence of infectious processes. Malnourishment encompasses defi cits in intake of Vitamin A rich-food, as well as inadequate consumption of foods containing nutrients important for its absorption. Diet is affected by cultural factors, such as eating habits, individual and family preferences and by socioeconomic factors which affect the capacity to select and buy these foods.2,c Infections can result in reduced serum retinol concentrations in the fi rst 24 hours after onset. The presence of sub-clinical infection may lead to VAD due to reductions in the retinol binding protein,20 regardless of insuffi cient intake of food sources and the hepatic reserves of the vitamin. However, episodes of serious or prolonged infection may affect hepatic reserves, due to reduced food intake, lower absorption, increased biological use and abnormal urinary excretion of retinol serum.7 Among strategies to prevent and control VAD, international organizations have advocated the massive supplementing of Vitamin A using mega doses for children aged from six to 59 months and in women immediately post-partum, considered to be an emergency INTRODUCTION intervention, and this has been adopted by the Brazilian Vitamin A Supplement Program. The state of Paraiba was a pioneer in establishing this program in the 1980s. According to Martins et al (2007), this state showed the highest levels of coverage and regularity between 1996 and 1998, compared to other states included in the Program.10 Bearing in mind the high prevalence of Vitamin A defi ciency in Paraiba found in the 1992 survey and the lack of subsequent studies evaluating trends in this defi ciency, the aim of this study was to analyze the prevalence of VAD in children and associated factors.

Vitamin A is a micronutrient known to be related to visual functions, epithelial integrity and the functioning of the immune system.a,b In the state of Paraiba, this defi ciency has been recognized since the 1980s. 3Results of a clinical-nutritional study carried out in three regions of this state, between 1981 and 1982, show a high prevalence of Bitot's spots and corneal scars in children in the sertão (desert region), especially in periods between harvests. 16Cases of corneal xerophthalmia (ulcer/queratomalacia) have also been documented in children in João Pessoa. 1 In the following years, clinical manifestations of moderate xerophtalmia and corneal scars, were also observed in children aged between two and 28 months, mainly during the drought which lasted from 1981 to 1984. 4 Approximately a decade later, a 16% prevalence of VAD was found in pre-school children in a population based survey, indicating a moderate public health problem.a Since1992 there have been no other population based investigations into this problem in the state.
The main causes of VAD can be divided into two broad categories: malnourishment and the presence of infectious processes.Malnourishment encompasses defi cits in intake of Vitamin A rich-food, as well as inadequate consumption of foods containing nutrients important for its absorption.c Infections can result in reduced serum retinol concentrations in the fi rst 24 hours after onset.The presence of sub-clinical infection may lead to VAD due to reductions in the retinol binding protein, 20 regardless of insuffi cient intake of food sources and the hepatic reserves of the vitamin.However, episodes of serious or prolonged infection may affect hepatic reserves, due to reduced food intake, lower absorption, increased biological use and abnormal urinary excretion of retinol serum. 7ong strategies to prevent and control VAD, international organizations have advocated the massive supplementing of Vitamin A using mega doses for children aged from six to 59 months and in women immediately post-partum, considered to be an emergency INTRODUCTION intervention, and this has been adopted by the Brazilian Vitamin A Supplement Program.The state of Paraiba was a pioneer in establishing this program in the 1980s.According to Martins et al (2007), this state showed the highest levels of coverage and regularity between 1996 and 1998, compared to other states included in the Program. 10aring in mind the high prevalence of Vitamin A defi ciency in Paraiba found in the 1992 survey and the lack of subsequent studies evaluating trends in this defi ciency, the aim of this study was to analyze the prevalence of VAD in children and associated factors.

METHODS
A transversal, population-based study was carried out with children aged six to 59 months old, of both sexes, resident in the urban area of the state of Paraiba between January and March 2007.
The probabilistic sample was calculated based on an estimated 16% a (p) prevalence of VAD, the confi dence interval (E = 1.96) for a 5% margin of error, sample correction factor of (c) of 2.1, as the sample is based on cluster technique, and accuracy (A) (or margin of error) of 3.0%.Based on the formula N = [E 2 .p(1-p).c]/A the minimum sample was calculated to be 1,204 children.Adding 10% for possible losses, a fi nal sample size of 1,324 children was arrived at.Increasing the sample size (sample correction factor of 2.1) has been recommended by the World Health Organization (WHO) 6 when making estimates about the population of geographical areas such as those in this study, considering the marked climatic diversity that shapes the meso-regions of the state and especially the great heterogeneity in the dynamics of urban space use.Thus, with the sample size more than doubled, design effect was minimized with the correction of potential intra-cluster variations.
The children were randomly selected, according to the two-stage sampling technique, using the municipality (fi rst cluster) and the census tract (second cluster).Thus, nine municipalities located in the sertão (desert region) (Belém do Brejo da Cruz, Boa Ventura, Conceição, Malta, Patos, Pedra Branca e São José de Espinhares), the agreste (region between the desert and the forest areas) (Campina Grande) and the zona da mata (forested) regions (João Pessoa).The municipalities of João Pessoa and Campina Grande were selected beforehand as they possessed population density signifi cantly higher than the other municipalities in the state, and due to their location (municipalities in two of the three bioclimatic zones.Considering that the distribution of the health problem in question may be affected by geographical location, including these regions in the sample, represented by their most densely populated municipalities, helps to reinforce the sample's representativeness.The other seven municipalities were randomly selected from within the sertão meso-region, bearing in mind the homogeneity in population density in municipalities in this region. Using a random number table in the randomization of census tracts and residences, observing the upper limit of 40 sample units per cluster (census tract) also served to include the assumptions necessary for a probabilistic selection of the sample units.Parents and guardians in the selected residences were invited to take part in the study, with a refusal rate of approximately 4% due to refusal or absence.Of the 1,324 children originally selected, 1,211 were studied.Some losses occurred due to collecting insuffi cient biological material for the biochemical analysis, as well as mothers' refusals to participate.The mean age was 33 months (SD = 15.7 months), and 51.2% of the children were male.
Blood was collected to analyze concentrations of serum retinol and C-reactive protein, and a questionnaire used to collect socioeconomic and demographic data on the study population.The children's health records were used to extract data on previous megadoses of Vitamin A supplement.
The blood was collected between one to three working days after the household interview, in a location close to the subjects' homes, usually the families' primary care center.When it was not possible to use these centers, the blood was collected at parenting groups, churches or residents' associations.
In order to carry out the laboratory tests, the blood samples (5 mL) were collected first thing in the morning, without the need for fasting, using disposable needles and syringes, and were stored in aluminum tubes to protect them from light.
After removal of the coagulant, the blood was centrifuged at a speed of 3,000 rpm and two serum aliquots were stored in amber and transparent eppendorf tubes, to determine the concentrations of serum retinol and C-reactive protein, respectively.The samples from municipalities in the sertão were stored and transported to the laboratories, maintaining the cold chain.
Serum retinol concentrations were measured using high performance liquid chromatography (HPLC), following the technique proposed by Furr et al, d Micronutrient Research Center of the Universidade Federal da Paraíba (UFPB).C-reactive protein concentrations were assayed using the latex agglutination test in the Biochemistry Laboratory of the Hospital Universitário Lauro Wanderley, UFPB.
WHO cutoff points were used to assess retinolemia, according to retinol serum levels, these being: deficient (< 0.35 μmol/L), low (≥ 0.35 μmol/L and < 0.70 μmol/L), acceptable (≥ 0.70 μmol/L and < 1.05 μmol/L) and adequate (1.05 μmol/L).Children with serum retinol concentrations < 0.70 μmol/L were considered to have a defi ciency.CRP concentrations ≥ 6 mg/L were considered to be cases of subclinical infection. 19iInfo v.6.04b(WHO/CDC, Atlanta, GE, USA) software was used to create the database, using double data entry for 100% of the data.In order to check the consistency of the data entry, the 'validate' application was used and the Statistical Package for Social Sciences version 13.0 software (SPSS Inc., Chicago, IL, USA) was used for the statistical analysis.
VAD (low or defi cient levels of retinol) was a dependent variable, and the independent variables were the socioeconomic and demographic data, previous Vitamin A supplements, the presence of subclinical infection and basic sanitary conditions.The Kolmogorov-Smirnov test was used to assess the assumption of normality of the retinol serum variable, described as mean and standard deviation.Percentages and their respective 95% confi dence intervals were described.
To evaluate the difference in the mean concentrations of serum retinol, according to the presence/absence of subclinical infection, the Student's t test for unpaired data was used.The prevalence and odds ratios were adopted, respectively, as measures of occurrence and epidemiological association.Pearson's Chi-squared test was used to assess the statistical signifi cance when comparing percentages.Forward stepwise logistic regression was used for the simultaneous analysis, with the calculation of the adjusted measures, using p < 0.05 and p > 0.10 as criterion for inclusion and exclusion respectively.The statistical signifi cance of the associations analyzed was based on p value < 0.05.
The test results were sent to the primary health care centers closest to the children's residences, with a letter explain the meaning of the biochemical results, to the health care professionals at the centers.Where this was not possible, the results were sent by mail to the participant's residence, with an explanatory letter for d Furr HC, Tanumiharjo S, Olson JA.Training manual for assessing vitamin A status by use the modifi ed relative dose response assays.Ames: Iowa State University; 1992.the child's mother.In cases where VAD was detected, the mother was advised to seek assistance in the nearest primary health care center.
Table 2 shows the prevalence of VAD according to the variables studied.Only the presence of a subclinical infection proved to be signifi cantly linked to VAD.The prevalence of VAD (37.6%, 95%CI 29.6;45.6)among children with subclinical infection was higher (p < 0.001) than those with no infection (19.8%, 95%CI 17.4;22.2).The variables: number of people in the household and indoor plumbing showed results bordering on statistically signifi cant.There was no observed link between VAD and Vitamin A supplements in the previous six months (p = 0.77).
After adjusting the model in the logistic regression analysis, the variables subclinical infection (p < 0.01) and indoor plumbing (p = 0.02) remained in the model, characterized as the best predictors of VAD.The presence of subclinical infection proved to be a risk factor associated with VAD and indoor plumbing as a protection factor against vitamin A defi ciency.VAD was four times more likely to occur (95%CI 1.49;10.16) in children with a subclinical infection and without indoor plumbing in their home, compared with those with no infection living in a house with indoor plumbing (Table 3).

DISCUSSION
The 21.8% prevalence of VAD can be considered as a serious public health problem, according to WHO criteria.e High prevalence of VAD have been found in representative samples of children under 5 in other states in the Northeast of Brazil, such as Pernambuco (22.3%), b Alagoas (44.8%), 21Bahia (44.7%) 12 and Sergipe (32.1%). 9According to the National Demographic and Health Survey 17.4% of children have VAD, with the highest rates of prevalence found in the Northeast (19.0%) and Southeast (21.6%) of the country.f Both insufficient dietary intake of Vitamin A and infections lead to reduced levels of retinol in the blood, although they are situations requiring different solutions, depending on the nutritional status.In the case of the former, there is a reduction in the hepatic reserves.In the latter, vitamin A stored in the liver may well be at normal levels, it being the transport of Vitamin A which is affected as a result of reduced synthesis of the retinol binding protein as a consequence of the infection process.This means it is necessary to verify whether the vitamin A defi ciency is due to insuffi cient intake  or whether it is a secondary response to an infection. 20ood intake was not assessed in this study, which can be considered a limitation.
Based on the VAD prevalence of 21.8% in the total population of the study and 19.8% in children without infections, a 2.0% reduction can be observed, highlighting the influence of infection on the Vitamin A nutritional status.Corroborating these fi ndings, Wieringa et al 23 (2002) 12 In this study, no statistical differences with regards to VAD and the children's gender were found, a result similar to those observed in other studies. 5,12,22According to the WHO, c differences in rates of VAD prevalence related to sex are attributed to cultural differences in eating habits and are not consistent with the existence of a physiological factor in this association.
In this study, statistically significant correlation between serum retinol and C-reactive protein levels were observed, indicating that the presence of subclinical infection increases the chance of a child having VAD by 2.55 times.A meta-analysis of the effect of subclinical infection on retinol serum levels, identifi ed signifi cantly higher retinol values in individuals with normal C-reactive protein levels, showing the damage infl ammatory processes have on the Vitamin A nutritional status. 19Moreover, in mild infections the reduction in retinol serum, due to infl ammatory processes, may have a transient effect on cases of changes in the metabolizing of vitamin A, not accompanied by reduced hepatic stores.However, in the case of serious or prolonged infections, this alteration may affect hepatic Vitamin A stores, aggravating the defi ciency still further. 7e increased susceptibility of children in homes without indoor plumbing to VAD shows the importance of environmental factors in determining defi ciency.It is assumed that, in addition to infectious processes, economic and structural factors may aggravate the occurrence of VAD.
According to the WHO, c a water supply insuffi cient for drinking, personal hygiene and growing food is among the factors associated with malnutrition, including VAD, and explaining the risk of VAD associated with lack of indoor plumbing in the home.
The synergy between nutritional problems and infections, especially in unsanitary conditions, establishes a vicious circle resulting in varying degrees of retarded growth and development in children which needs to be reversed. 18Analyzing the water-infection variable, the importance of contributing factors to the occurrence of VAD can be observed, such as the high prevalence of infections, lack of sanitation and drinking water, as well as adverse socioeconomic conditions and taboos related to food of plant origin which increase the demand or interfere in the organism's intake and metabolizing of Vitamin A. 13 The results of this study show that, in spite of actions to prevent and control VAD on the part of the state of Paraíba, it is still a serious public health problem, demanding specialist attention from the government organizations which deal the population's health and nutrition.In this context, it is necessary to consider the importance of concerted actions covering sanitation, immunization and the control of diarrheal and parasitic diseases as a form of preventing and controlling Vitamin A defi ciency.
The research protocol was approved by the Committee for Ethical Research of the Universidade Estadual da Paraíba (protocol n o 1128.0.133.000-05 in 20/12/2005).

Table 1 .
Concentration of retinol in children according to presence/absence of infection.Paraiba, Northeastern Brazil, 2007.

Table 3 .
observed an overestimation of the prevalence of VAD of around 4.6% comparing prevalence Odds ratio adjusted for risk of vitamin A defi ciency among children according to infection status and indoor plumbing in the home.Paraiba, Northeastern Brazil, 2007.
a Adjusting for age, sex, vitamin A supplement in the last six months, mother's age, mother's schooling, head of household schooling, per capita income, household type, number of rooms and people in the household, garbage and sewage type.