The children were lightly dressed and without shoes when measurements of height, with a portable stadiometer ±1 mm, and weight using a digital electronic scale 150 kg ± 0.1 kg (Omron HBF-400) were done, according to standardized procedures [15] . Anthropometric measurements were standardized into age-sex-specific z-scores, which measure the deviation of each individual child’s value from the reference child population. Z-scores of height-for-age (HAZ), weight-for-height (WAZ) and body mass index-for-age (BMIAZ) were calculated using the software AnthroPlus (version 10.4; World Health Organization), relative to WHO reference data 2007 [16] . The z-score cut-offs were defined according to WHO classification as follows:

The dietary intake of the children was assessed during two consecutive days by the Food Photography 24-hour Recall method (FP24-hR) previously validated and described in detail [17] . Briefly the method consists of a photographic food record, subjects take photographs of all their meals and beverages consumed over a period of time, the method is combined with a 24-hourrecall conducted by nutritionists who visit the subjects after each 24-h period so that they may fill in a questionnaire with detailed information of all of the consumed foods. The method was carried out with small modifications for children under the age of 10; their mothers were responsible for taking the photos before the consumption of the meals, and during the 24-h recall the questions were answered by the mother and child together. Food consumption data were extracted from the questionnaires and the nutrient calculation was done according to the guidelines given in the previously reported method [17] . Moreover, data about the mineral content regarding zinc, iron, calcium and phytate content in the plant-based staple food in this tropical area were included from a previous analytical study [18] .

The intake of energy, protein, fat, carbohydrates, fiber, calcium, iron, phosphorus, zinc, cooper, thiamin, riboflavin, niacin, folate, and vitamins A, C, E, B6, and B12 were calculated. These nutrients were selected in order to shed light on possible deficiencies or excesses. The inadequacy of energy intake was evaluated as intakes below two-thirds of the energy expenditure [20] , energy expenditure was calculated for each child according their age and weight, with equations from FAO/WHO [19] . Proportion of inadequate intake of nutrients was evaluated by the estimated average requirements (EAR) cut-point method [20] [21] , according to the sex and age of each child respectively. Molar ratios Phy:Zn (phytate:zinc) and Phy:Fe (phytate:iron) in the diet of the children are presented as an estimation of the relative bioavailability of iron and zinc in their diet by comparing the ratios with the desirable values. The desirable Phy:Zn is below 15, even though ratios below 5 have also shown a negative effect on zinc absorption [22] , the desirable Phy:Fe for an adequate iron absorption is lower than 1 [23] .

The enrolled children were provided with sterile clean leak-proof stool cups for collection. Diagnosis of intestinal parasites was made following standardized procedures of sedimentation and microscopic examination of stool [24] , using lugol and formalin-ethyl acetate concentration; the intestinal parasites were studied in 40X magnification field (Biological microscope Model Olympus CX31, OLYMPUS Latin America INC.).

Blood samples (5 ml) were drawn from fasting children from the antecubital vein, into trace element-free, heparin-container Vacutainer tubes (Becton Dickinson AB, Stockholm, Sweden).The procedures of blood collection and storage follow the guidelines given by IZiNCG [25] .

Serum zinc was quantified by flame atomic absorption spectrometry (Model 2280, Perkin Elmer Corporation, Norwalk, CT, USA). Before analysis the samples were diluted 10 times with deionized water [26] . To validate the analysis in serum, the reference material Seronorm^TM (Trace elements serum L-1-2SERO AS, Norway) was used. Serum iron and total iron binding capacity (TIBC) were analyzed by colorimetric procedures with the commercial kits (Fer-color kit, and IUBC/TIBC AA, Weiner Laboratories, Argentina). Percentage of transferrin saturation (%TS) was calculated as serum iron/TIBC.

Hemoglobin, hematocrit and red blood cells (RBC) were determined by a microhematocrit reader (Microhematocrit system, Model StatSpin CSD2, CritSpin, Iris International Inc., USA). The mean corpuscular volume (MCV) was calculated by dividing the hematocrit by the total number of red blood cells (Hematocrit/RBC) and multiplied by 10. The mean cell hemoglobin concentration (MCHC) was calculated by dividing the hemoglobin by the hematocrit (hemoglobin/hematocrit).

White blood cells were analyzed in a Neubauer hemocytometer by microscopic examination. Neutrophils, basophils, lymphocytes, eosinophils and monocytes were counted through the morphological evaluation and identification. The count was done with 100X magnification under oil immersion (Biological microscope Model Olympus CX31, OLYMPUS Latin America INC.).

The Statistical Package for Social Sciences (SPSS) version 18.0 (SPSS Inc., IBM corporation 2010, www.spss.com) was used to perform the statistical treatments. The significance level was set up at P values < 0.05. Normality of the data was assessed by Shapiro-Wilk test, and measures of skewness and kurtosis. The majority of the variables do not follow a normal distribution (P < 0.05), thus the results are presented as median, minimum and maximum values. Spearman’s correlations were computed to study the association between serum zinc and iron with the corresponding micronutrient intakes and with phytate intake.

Simple and multiple linear regression analyses were computed to evaluate the effect of the number and type of parasites found in the children and the effect of phytate and mineral intake, on the levels of zinc and iron in serum. The total group of children was divided into children with specific parasites on the one hand, and the group without these parasites, on the other hand, which was used as a control group, where Man-Whitney U test was used to compare and evaluate differences in nutrient intake, anthropometric indicators, hematological and trace element status between the groups (with and without the parasites). Additionally, one way ANOVAs were used, after dividing the total group of children in subgroups with zero to n-number of different species of parasites found, to examine the effect of polyparasitism on the dietary intake, anthropometric indicators, hematologic and trace element status.

Forty-six children participated in the study: 20 boys and 26 girls; the age range was from 4 to 13 years old. Anthropometric measurements are shown in Table 1, the z-scores HAZ indicated that 37% of the children were stunted (short for their age), WAZ indicated that 17% of the children were wasted (thin for their age), BMIAZ indicated that 17% of children were underweight, 83% of children were of normal weight and none of them were overweight.

The dietary evaluation showed that the energy distribution was within the recommended values from WHO [21] , 68E% was from carbohydrates, 14E% was from protein and 19E% was from total fat (Table 2).The dietary patterns are mainly based on carbohydrates from cereals (rice, pasta, bread), tubers (potatoes, cassava), and legumes (fava beans, lentils, peanuts), with protein from small portions of meat or eggs and fat from oil and tallow; the vegetables and fruits are present randomly and in small portions. The percentage of children consuming less than two-thirds of their energy expenditure according to their age and sex was 43%, indicating inadequate diets. The analysis of EAR cut-point to evaluate adequate nutrient intakes showed that intakes of protein and carbohydrates

^aHAZ, z-scores height-for age; ^bWAZ, z-scores weight-for age; ^cBMIAZ, z-scores body mass index-for-age.

are adequate for all the children. However, fiber intake is inadequate in 100% of the children, as well as higher levels of inadequacy were found for vitamin E, A and calcium (higher than 90%), intake of folate and vitamin B12 were inadequate in 65 and 39% of the children respectively. Regarding zinc and iron intake, the diet of the children showed to be inadequate to meet the requirements in 30% and 34% of the children respectively.

Phytate intake was from 0.32 to 1.42 g/d, with corresponding high molar ratios of Phy:Zn and Phy:Fe likely to impair mineral absorption, 22% children presented Phy:Zn higher than 15 and 100% of them had Phy:Fe higher than the desirable value 1. The high level of phytate was due to some of the main foods in the diet being cereals and legumes, which are known to have high levels of phytate; tubers and roots are the main staple food and also contain phytate.

Intestinal parasites were found in 96% of the children. Seven types of parasites were identified; Giardia lamblia (n = 8, number of children presenting this parasite), Entamoeba histolytica (n = 8), Entamoeba coli (n = 5), Ancylostoma duodenale (n = 9), Ascaris lumbricoides (n = 28), Trichuris trichiura (n = 19) and Strongyloides stercolaris (n = 2). Forty-two percent of the children had one type of parasite, 37% had two types of parasites, 13% had up to three types of parasites and 4% had four types of parasites (Table 3).

The levels of zinc and iron in serum are shown in Table 4. The results were compared with reference values compiled by IZiNCG [25] from NHANES III (1988-1994) [5] . Serum zinc was lower in 87% of the children compared to the reference values (<65µg/dl) [25] , indicating zinc deficiency. Iron status was evaluated by three indicators; serum iron, TIBC and transferrin saturation. According to the results, 84% of the children present low levels of iron in serum; approximately 66% of the children were iron deficient and 30% of the children were with anemia, which is characterized by low serum iron, elevated TIBC and low transferrin saturation. Furthermore, approximately 70% of the children presented low levels of hemoglobin, hematocrit and red blood cells, which most likely indicates the risk with anemia in these children.

The white blood cell counts were low in 22% of the children; 63% of the children presented low levels of lymphocytes, 100 % of them presented high levels of eosinophils, and neutrophils were lower in 33% of the children; basophils were at the normal level for 98% of the children and monocytes were lower than the normal level for 100% of the children (Table 4).

^aPercentage of children with inadequate nutrient intake, assessed by the estimated average requirements (EAR) cut-point method, using EAR standard values according to sex and age from WHO [21] ; ^bPercentage of children that meet only two-thirds of their energy expenditure calculated according their age and sex with equations from FAO/WHO [19] ; ^cMedian values of protein, fat and carbohydrates intake and their corresponding percentage of energy intake in parenthesis (%E); ^dPercentage of children with molar ratios phytate:mineral, above the suggested molar ratios of Phy:Zn [25] and Phy:Fe [23] .

Strong negative correlations were found between serum zinc and phytate intake (r = −0.439, P = 0.01), as well as with the molar ratio Phy:Zn (r = −0.301, P = 0.05), suggesting that the phytate level in the diet of the children impairs the zinc absorption. The multiple linear regression analysis for evaluating the significance of the effect of phytate and zinc intake on serum zinc, indicated that phytate intake had a significant negative effect (B₁ = −20.0 ± 6.9 µg/dl, r = −0.458, P = 0.006) on the level of serum zinc; the model shows that serum zinc is decreased by 20.0 ± 6.9 µg/dl for every additional unit (g/d) of phytate in the diet. The effect of zinc intake on the serum Zn was not significant (P = 0.824).

The simple linear regression analysis for evaluating the effect of each specific parasite on the levels of zincandironin serum showed that the parasites: A. lumbricordes, E. hystolytica, E. coli, T. trichiura and A. stercolaris have no significant effect on the serum levels of these trace elements; neither were any significant differences found when comparing dietary intake, anthropometric indicators and trace element status between children with or without each specific of these parasites.

However, simple linear regression models for G. lamblia and A. duodenale showed a significant effect on the

^aFrequency of children with none, 1, 2, 3 or 4 parasites; ^bNumber of children with the specific type of parasite.

^aReference values for children, compiled by IZiNCG [25] , data derived from NHANES II and WHO [13] [14] . Values for TIBC and %TS from Cook, 1992 [10] ; ^bResults of iron parameters from 44 samples. Results of 2 samples were discarded because the reported values were extremely low; ^cTIBC, Total iron binding capacity; ^d%TS, Percentage of transferrin saturation.

levels of serum zinc and iron respectively. The model indicates that the presence of G. lamblia had a significant negative effect (B = −7.4 ± 3.4 µg/dl, r = −0.310, P = 0.038) on the serum zinc of the children. Serum zinc in children with G. lamblia is 7.4 µg/dl lower than serum zinc in children without this parasite. As has previously been shown, phytate intake also had a significant effect on serum zinc; therefore, a multiple linear regression model was applied to evaluate the extent to which the phytate intake and the presence of G. lamblia effect on serum zinc levels; this model showed that the serum zinc is significantly affected by the phytate intake (B₁ = −15.8 ± 4.4 µg/dl, r = −0.482, P = 0.001) and also by the presence of G. lamblia (B₂ = −6.6 ± 3.1 µg/dl, r = −0.310, P = 0.035) (Figure 1(a)), indicating that the serum zinc in children is decreased by 15.8 µg/dl for every additional unit of phytate intake. Moreover, the level of zinc in the group of children with G. lamblia is 6.6 µg/dl lower than in the group of children who are free from G. lamblia. There was no significant effect of the presence of G. lamblia on the serum iron of the children (P = 0.328) (Figure 1(b))

Additionally, the test Man-Whitney U was used to compare the parameters of the dietary intake, anthropometric indicators and trace element status in the group of children infected with G. lamblia (n = 8) and the group of children without this parasite (n = 38). The mean serum zinc level was significantly lower (P = 0.026) in the group of children with G. lamblia (49.6 µg/dl) compared to the other group (56.9 µg/dl). There were no significant differences in the comparison of zinc intake (P = 0.931), phytate intake (P = 0.187) or serum iron (P = 0.373), neither in the anthropometric and hematological indicators.

The simple linear model to evaluate the presence of the parasite A. duodenale, showed a negative significant effect on serum iron (B = −18.2 ± 8.4 µg/dl, r = −0.316, P = 0.035) (Figure 1(c)). This model indicates that

the serum iron in children with A. duodenale is 18.2 µg/dl lower than the level in children who are parasite-free. There was no significant effect on the serum zinc (P = 0.802) (Figure 1(d)). Man-Whitney U test showed that the mean serum iron (24.5 µg/dl) in the group of children with A. duodenale (n = 9) was significantly lower (42.7 µg/dl, P = 0.030) than in the group of children without the parasite (n = 37). There were no significant differences in the levels of iron intake (P = 0.923), phytate intake (P = 0.771), serum zinc (P = 0.372) or serum copper (P = 0.825). Furthermore, no significant differences were found between the nutrient intakes and anthropometric indicators between these two groups. However, significant differences were found in the level of neutrophils (P = 0.012), which was higher in the group of children without the parasite (46%) compared to the children with A. duodenale (38%), and the level of eosinophils was significantly lower (P = 0.008) in the group of children without parasites (21%) compared to the group with A. duodenale (29%).

One way ANOVA analysis for evaluating the effect of polyparasitism showed no significant differences in the parameters of the dietary intake, anthropometric indicators, hematologic and trace elements in children presenting 0, 1, 2, 3 or 4 species of different parasites.