The strategies most frequently used in public interventions intended to reduce and control iron-deficiency anemia are these: diversification of the diet of target populations; pharmacological supplementation with iron preparations, frequently combined with other micronutrients; and fortification of food with iron (1).
In Mexico, a national effort to reduce and control the high prevalence of anemia deficiencies in children 12 to 24 months old (nearly 50% of a national probabilistic sample of toddlers 12–24 months old), of whom 65% were iron deficient and 30% were deficient in zinc (2,3), was undertaken by 3 large nutrition interventions: (1) Oportunidades, (2) the distribution of iron-fortified milk at subsidized prices, and (3) the distribution of iron supplements to members of the poor Indian population. Among them, the social program Oportunidades, launched in 1998, targets families living in extreme poverty and includes money transfers intended for food purchasing that are conditional on their regular attendance at prenatal, postnatal, and well-baby clinics. In addition, infants and toddlers 6 to 23 months old and children 24 to 59 months old who show some degree of malnutrition are provided with a fortified complementary baby food based on powdered milk (Nutrisano). Nutrisano also contains the recommended dietary intake of iron, zinc, and other micronutrients listed in Table 1. It is presented as a powder to be reconstituted with water until it reaches a pap consistency (4).
A study in toddlers 12 to 30 months old at the beginning of the intervention, evaluating the effectiveness of Nutrisano, documented a modest decrease in the prevalence of anemia (from 46% to 40%) after 1 year of intervention and no significant effects on their iron status. These deceptive results were attributed to poor absorption of the iron compound that was used then as a fortifier (hydrogen-reduced iron) (5). Another study comparing the bioavailability of hydrogen-reduced iron, ferrous sulfate (FS), and fumarate, added to Nutrisano as fortificants, demonstrated that the bioavailability of the hydrogen-reduced iron added to Nutrisano as a fortificant was 2.5%, compared with 14.3% for FS (6).
On the basis of these results, a group of experts recommended initially switching to FS as a fortificant of Nutrisano because of its high bioavailability and low cost. Nevertheless, some negative effects on the organoleptic characteristics of Nutrisano observed in shelf-life studies moved the experts to test ferrous gluconate (FG) as an alternative iron source. The water solubility of FG is comparable to that of FS, and it has been reported to have a similar absorption to FS in both rats and humans (7). There is little information in the literature regarding the use of FG in large-scale nutritional interventions. A recent efficacy study of a large-scale nutrition intervention using milk fortified with FG demonstrated its effectiveness in reducing anemia and iron deficiency (8). These results reinforced the decision to explore the addition of FG as a fortificant for Nutrisano.
The present study was designed to compare the efficacy of 2 iron compounds with high bioavailability (FS and FG) added as fortificants to Nutrisano to improve iron status and reduce the prevalence of anemia in toddlers who were beneficiaries of the program Oportunidades, which serves 5 million families living in extreme poverty in Mexico.
PATIENTS AND METHODS
Eleven semirural communities served by the social program Oportunidades were selected. The 11 communities were randomly assigned to receive 1 of 3 versions of Nutrisano: 2 versions fortified with iron, either FS or FG, and 1 version with no iron added (control group). Healthy children living in those communities who were 12 to 30 months old at the beginning of the study were identified from a registry of children younger than 5 years of age. Such a registry is maintained and periodically updated by the local health facilities run by the Health Department.
The food supplement for this study was produced by the same manufacturer supplying Nutrisano for the regular operation of Oportunidades, (LICONSA Co., Mexico City, Mexico). The composition of Nutrisano includes powdered milk (54.9%), maltodextrins (37.2%), sugar (5.7%), flavoring substances, iron, zinc, retinol, riboflavin, vitamin B12, α-tocopherol, vitamin C, and folate in the amounts specified in Table 1. The ingredients were mixed in the manufacturing plant with a mechanical-type “V” mixer until they were homogenized in the following sequence: milk powder, maltodextrins (from hydrolyzed corn starch), flavoring substances, vitamins and minerals premixed with an aliquot of hydrolyzed corn starch, and finally sugar. Nutrisano was prepared by end users, who added approximately 25 mL bacteriologically safe bottled water to 44 g of powder (4 tablespoons) to give a puree consistency. Each serving provides 10 mg of iron and 49.9 mg of vitamin C. Nutrisano was supplied in 264-g foil packets, and mothers were instructed to serve 4 spoonfuls of the powder and feed it to the children between the first and the second meal of the day. The foil envelopes containing the 3 versions of Nutrisano were indistinguishable except for a color-coded strip, to which researchers and field operators were blinded.
Design of the Intervention
Nutrisano was distributed to participating homes for 6 months (March to August 2005). Every day a health worker assisted the mother or personally prepared Nutrisano according to the instructions on the package (4 tablespoons or 44 g of powdered formula mixed with 4 tablespoons of water) and measured the intake of Nutrisano before and after a feeding episode by weighing the bowl containing the Nutrisano portion on a precision electronic scale (Tanita, Tanita Co, Tokyo, Japan). Questionnaires that explored housing characteristics, feeding practices during the first year of life, and other sources of supplemental iron were applied to the mothers or caretakers.
Length and weight were measured at baseline and 6 months after initiation of the intervention, by use of an infantometer (local construction, Mexico City) with a precision of 1 mm and on an electronic scale with a precision of 10 g (Tanita).
Hemoglobin concentration was determined in capillary blood obtained by finger prick and measured in a Portable Photometer Hemocue (HemoCue, Angelholm, Sweden) (9,10). Variability of photometers was assessed during field work at the beginning and end of each working day. A 3-level liquid quality control check (4C-ESControl, Beckman-Coulter, Miami, FL) and the readings of a precalibrated reference cuvette included with the equipment were used for that purpose. The mean difference between duplicates was 0.3 ± 9.9 g/L (P = 0.36 for liquid quality control material and −0.24 ± 3.6 g/L, P = 0.27 for the reference cuvette).
Venous blood samples were drawn from an antecubital vein at baseline and 6 months after initiation of the intervention. Samples were centrifuged at 268g in situ, and serum was stored in color-coded cryovials in liquid nitrogen until delivery to a central laboratory. Commercial kits were used to measure the serum concentrations of ferritin (Dade Behring, Inc., Newark, DE, USA), soluble transferrin receptors (sTfR) (Dade Behring, Marburg, Germany) by enzyme-linked immunosorbent assay, and C-reactive protein (CRP) by nephelometry, using monoclonal antibodies (Behring Nephelometer 100 Analyzer, Behring Laboratories, Messer Grisheim Gmbh, Frankfurt, Germany).
Anemia was defined according to the World Health Organization recommendations as hemoglobin concentrations <110 g/L at sea level (11). Hemoglobin concentrations were adjusted for altitude according to the equation proposed by Cohen and Haas (12). Iron status was assessed by 2 indicators: serum ferritin concentration and sTfR. Serum ferritin is considered to be indicative of depleted iron stores when its concentration is <12 ug/L, and sTfR indicates tissue iron deficiency when its concentration is >6 mg/L. Serum ferritin is very sensitive to concurrent acute inflammation, whereas sTfR is not. Total body iron stores were calculated according to the concentrations of serum ferritin and sTfR and expressed as mg/kg body weight according to the following equation proposed by Cook et al (13).
Adherence to treatment was defined as the proportion of the expected intake to observed intake of Nutrisano expressed in grams, consumed in 1 month.
The intake of Nutrisano was defined as the total amount of Nutrisano consumed during the 6 months of intervention divided by 6 (duration in months of the intervention). Then it was stratified into low, medium, and high tertiles. The daily intake of supplemental iron was defined as the amount of iron contained in the total amount of Nutrisano consumed in 1 month divided by 30.
Sample Size Calculations and Statistical Analysis
The sample size was calculated by use of the following assumptions: a Δ of 12.5 percentage points in the prevalence of depleted iron stores as indicated by serum ferritin concentrations, an α value of 0.05, and a β of 0.8; this resulted in a sample size of 160 children per treatment group. However, because we were unable to determine with certainty the baseline prevalence of depleted iron stores because of the lack of basal CRP values to exclude from analysis the cases of active infection or inflammation, we decided instead to calculate the statistical power based on the evaluation of the prevalence of tissue iron deficiency as indicated by sTfR concentrations. Considering the actual sample size attained and the smallest difference between basal (0.26) and final (0.19) prevalence of tissue iron deficiency (FS group), the power to detect differences of 9 percentage points was of 0.78.
Unadjusted comparisons between means were done by analysis of variance and the Bonferroni test as the post-hoc test, and between proportions using χ2 tests. The effects of treatment were assessed by linear regression models in which final concentrations of serum log-ferritin and sTfR and total iron body content were alternately used as dependent variables. Covariables included treatment, age, sex, tertile of intake of Nutrisano, and basal concentrations of sTfR and total body iron content, when appropriate. Also, CRP concentrations at 6 months were introduced into the models analyzing log-ferritin. A logistic regression model was constructed with the final prevalence of tissue iron deficiency as the dependent variable and age, sex, treatment, the intake of other sources of supplemental iron, and baseline serum ferritin of sTfR as independent variables. An orthogonal contrast test was applied to assess differences among adjusted means by treatment group. Serum ferritin and sTfR data were not normally distributed; thus, they were transformed logarithmically for statistical analysis. Both ferritin and sTfR values are expressed throughout the text as antilogarithms. Statistical analyses were performed with the statistical software Stata version 7.2 for Windows (Stata Corp, College Station, TX). Values in the text are mean ± SD unless otherwise indicated.
Parents or legal guardians signed an informed consent letter after a careful explanation of the objectives, nature, and risks of the study. The protocol was reviewed and approved by the Research, Ethics, and Biohazards Committees from the National Public Health Institute, Cuernavaca, Mexico.
A final sample of 486 children was assembled, from which 71 children (14.6%) were lost to follow-up. Losses were mostly due to refusal of families to pursue the study, moving out of town, or refusal to consume the complementary food (see subjects flow chart in Figure 1).
Baseline age, body weight, length or height, sex distribution, and infant feeding practices were not different among treatment groups, nor were housing characteristics different among groups except that piped water was less available in the control group (P < 0.0001) (Table 2). The baseline daily intake of Nutrisano was lower in the FS group (P < 0.05) than in the FG and control groups. At the end of 6 months, the mean intake of Nutrisano was not different among treatment groups. During the 6 months of the intervention period, 92.5% of children consumed ≥32 g/day of Nutrisano, with no differences among treatment groups. The daily intake of dietary iron was significantly lower in the control group than in the FS and FG groups (P < 0.05). During the first month of intervention, the mean intake of iron from Nutrisano was 8.0 ± 2.5 and 9.3 ± 1.5 mg/day in the FS and FG groups, respectively (P < 0.05), and 0.01 ± 0.003 mg/day for the control group, and it remained essentially unchanged for the 6-month intervention.
Hemoglobin and Indicators of Iron Nutritional Status
No significant changes in hemoglobin concentrations or iron-deficiency anemia were observed between baseline and 6 months of intervention within or among treatment groups.
Iron status indicators were as follows. The unadjusted mean concentrations of serum ferritin increased and those of sTfR decreased significantly from basal to 6 months in both the FG group (ferritin Δ = 6.6 μg/L, sTfR Δ = 0.21 mg/L) and the FS group (ferritin Δ = 5.8 μg/L, sTfR Δ = 0.58 mg/L) (P < 0.02), whereas no significant changes occurred in the control group (ferritin Δ = 1.9 μg/L, sTfR Δ = 0.09 mg/L) (P = 0.7) (Table 3).
The prevalence of tissue iron deficiency according to sTfR >6 mg/dL decreased at the end of the 6-month intervention 7.1 percentage points in the FG group, decreased 13.1 percentage points in the FS group; and increased 0.7 percentage points in the control group (P > 0.05). The final prevalence of tissue iron deficiency in the FG and FS groups was significantly lower than in the control group (P < 0.05). The reduction of the prevalence of tissue iron deficiency was significantly lower in the FG group than in the FS group (P < 0.05) (Table 3).
In separate multiple regression models, no association was found between treatment and concentrations of hemoglobin and serum ferritin (data not presented), controlling for the respective basal concentrations, age, sex, and status as beneficiary of other supplementary iron sources; final ferritin values concurrent with abnormal final CRP concentrations (>5 mg/L) were excluded from the analysis. However, an interaction between treatment and the total intake of Nutrisano on the concentration of ferritin was observed. There were no differences in the adjusted means of ferritin concentration among the FG, FS, and control groups of children in the lowest tertile of total intake of Nutrisano, but adjusted means of serum ferritin in the FS group but not in the FG group were higher in the middle tertile, and those in both the FG and FS groups were higher in the highest tertile of total intake of Nutrisano than in the control group (P = 0.05). No differences between the FG and FS groups were found (Table 4).
The FS and FG groups had adjusted means of sTfR significantly lower (Bonferroni P < 0.001) (Fig. 2A) than the respective adjusted means in the control group, controlling for the respective basal concentrations, age, sex, and status as beneficiary of other food assistance programs. The adjusted means of sTfR for the medium and high tertiles of intake of Nutrisano for the FG group (P = 0.001) and for the medium tertile for the FS group (P = 0.05) were significantly lower, and they were significantly higher for the medium tertile of intake in the control group (P = 0.001) compared with the respective low intake values (Table 4).
In the FS and FG groups, the adjusted means of total iron body stores were significantly higher (Bonferroni P range <0.03 and <0.001, respectively) (Fig. 2B) than the respective adjusted means in the control group, controlling for the respective basal concentrations, age, sex, and status as beneficiary of other food assistance programs. Adjusted means of total iron body stores were significantly higher in the FS group than in the FG group (Bonferroni P < 0.001). The adjusted means of total iron body stores for the medium and high tertile of intake of Nutrisano for the FS and FG groups (P = 0.001) were significantly higher, and they were significantly lower for the medium tertile of intake for the control group (P = 0.001) compared with the respective low intake values (Table 4).
Both FG and FS were efficacious in reducing iron deficiency when added as fortifiers to the complementary food Nutrisano. This was substantiated by the changes in the ferritin and sTfR concentrations and in the body iron stores. In a previous study by our laboratory, we demonstrated a superior bioavailability of FS over ferrous fumarate and hydrogen-reduced iron + Na2EDTA as fortifiers for Nutrisano (6), but the bioavailability of FG was not tested in that study. Theoretical and practical information supported the decision to examine the potential use of FG as an adequate iron source for Nutrisano. First, the water solubilities of FG and FS are similar, and their absorption in both rats and humans is comparable (7), and second, FG has been demonstrated in a large-scale trial to be effective in reducing anemia and iron deficiency when added to milk (8). In contrast to FS, FG did not produce negative organoleptic characteristics in Nutrisano. A metallic aftertaste was identified for Nutrisano fortified with FS by a group of trained judges in a sensorial study (14). Although the impact of both FS and FG in improving iron status was unquestionable in this study, their effect on hemoglobin concentrations was puzzling. We were unable to demonstrate any improvement in the final concentrations of hemoglobin among the experimental or control groups. One reason could be the low prevalence of baseline anemia and that the study sample was not designed to detect differences in that variable. In a recent literature review (15) of 5 of 26 randomized clinical trials, in which oral iron supplements were given to children 0 to 59 months old living in developing countries, no effects were found on the final hemoglobin concentrations, although significant positive effects on iron status indicators were reported in 4 of them. This inconsistent effect on hemoglobin concentrations may be indicative of the various causes of anemia in these study populations.
In support of the latter idea, a study describing the association between anemia and deficiencies in folate and vitamin A in a probabilistic sample of Mexican children younger than 5 years of age found that 38% of anemic children were not iron deficient. Furthermore, 25% of the variability in hemoglobin concentrations was explained by variations in erythrocyte folate concentrations and 10% by serum retinol concentrations (16). Nevertheless, Nutrisano was also fortified with folate and vitamin A, making it unlikely that these factors explain the lack of hemoglobin response in this study. Another partial explanation for the failure of Nutrisano to improve the final concentrations of hemoglobin is the weaker efficacy of interventions with iron-fortified food compared with iron supplements to improve hemoglobin concentrations, as reported by others (17). A further explanation for the lack of response in hemoglobin concentrations in our study could be the known detrimental effect exerted by combinations of iron and zinc in supplements. Such a combination blunts the improvement in anemia (18) compared with iron supplementation alone. It is possible that the combination of 10 mg of zinc and 10 mg of iron in Nutrisano lessened both the response of hemoglobin and that of indicators of iron status in our study.
One strength of this study is its randomized, double-blind, clinical trial design, which allows inferences on the causality of the associations analyzed. Another strength is the positive association between the level of intake of Nutrisano and iron status as indicated by serum ferritin, sTfR, and total iron body stores. Thus, we can ascertain that the daily intake of Nutrisano fortified with FG is effective in controlling and reducing iron deficiency. The satisfactory impact of FG on iron status observed in this study, and other sensory and shelf-life data reported elsewhere (14), has led to the political decision to substitute the hydrogen-reduced iron used formerly as a fortificant with FG. We foresee that such an improvement on the efficacy of Nutrisano fortified with FG to reduce iron deficiency will enhance the capability of the Oportunidades program to improve iron deficiency status in its almost 5 million beneficiary children. Considering the data herein presented and that more than 600,000 children 0 to 24 months old are served by Oportunidades, we estimate that 49,600 children will be saved from iron-deficiency anemia annually by the switch to FG as a fortificant. The increment in the production cost of Nutrisano associated with the substitution of hydrogen-reduced iron by ferrous fumarate was US$3.1/child/year. Economic losses in countries like Egypt, with a prevalence of anemia and gross domestic product similar to that of Mexico, are estimated to be $198 million/year. Compared with the new cost of the yearly supply of Nutrisano for the children saved from iron-deficiency anemia ($2.3 million), the return rate results are 1:86 (19). Studies to assess the effectiveness of the reformulated food supplement Nutrisano are in order.
1. Organización Panamericana de la Salud. Compuestos de hierro para la fortificación de alimentos: guías para América Latina y el Caribe
. Washington, DC: Organizacion Panamericana de la Salud; 2002.
2. Villalpando S, Shamah-Levy T, Ramírez Silva CI, et al
. Prevalence of anemia
in children 1 to 12 years of age: results from a nationwide probabilistic survey in Mexico. Salud Publica Mex 2003; 45(Suppl 4):490–498.
3. Villalpando S, García Guerra A, Ramírez Silva CI, et al
. Iron, zinc and iodine status in Mexican children under 12 years and women 12–49 years of age: a probabilistic national survey. Salud Publica Mex 2003; 45(Suppl 4):520–529.
4. Neufeld L, Sotres-Alvarez D, Gertler P, et al
. Impacto de Oportunidades en el crecimiento y estado nutricional de niños en zonas rurales. In: Hernández-Prado B, Hernandez-Avila M, editors. Evaluación Externa de Impacto del Programa Oportunidades 2004: Alimentación. Cuernavaca: Instituto Nacional de Salud Pública; 2005. pp. 19–45.
5. Rivera JA, Sotres D, Habicht JP, et al
. Impact of the Mexican Program for Education, Health, and Nutrition (Progresa) on rates of growth and anemia
in infants and young children: a randomized effectiveness study. JAMA 2004; 291:2563–2570.
6. Pérez-Expósito AB, Villalpando S, Rivera JA, et al
. Ferrous sulfate
is more bioavailable among preschoolers than other forms of iron in a milk-based weaning food distributed by PROGRESA, a national program in Mexico. J Nutr 2005; 135:64–69.
7. Hurrell R. Improvement of trace element status through food fortification. In: Clydesdale FM, Wiemer KL, editors. Iron Fortification of Foods. Orlando, FL: Academic Press; 1985. pp. 39–53.
8. Villalpando S, Shamah T, Rivera JA, et al
. Fortifying milk with ferrous gluconate
and zinc oxide in a public nutrition program reduced the prevalence of anemia
. J Nutr 2006; 136:2633–2637.
9. Johns WL, Lewis SM. Primary health screening by haemoglobinometry in a tropical community. Bull World Health Org 1989; 67:627–633.
10. Hudson-Thomas M, Brigham KC, Simmons WK. An evaluation of the HemoCue for measuring haemoglobin in field studies in Jamaica. Bull World Health Org 1994; 72:423–426.
11. Centers for Disease Control. CDC criteria for anemia
in children and childbearing-age women. MMWR Morb Mortal Wkly Rep
12. Cohen JH, Haas JD. Hemoglobin correction factors for estimating the prevalence of iron deficiency anemia
in pregnant women residing at high altitudes in Bolivia. Pan Am J Public Health 1999; 6:392–399.
13. Cook JD, Skikne BS. Iron deficiency
: definition and diagnosis. J Intern Med 1989; 226:349–355.
14. Morales J, Vargas F, Cassís L, et al
. Sensorial evaluation of nutritional supplements (PROGRESA) enriched with 3 different forms of iron in a rural Mexican community. J Food Sci 2007; 73:S1–S5.
15. Lannotti LL, Tielsch JM, Black MM, et al
. Iron supplementation in early childhood: health benefits and risks. Am J Clin Nutr 2006; 84:1261–1276.
16. Villalpando S, Pérez-Expósito AB, Shamah-Levy T, et al
. Distribution of anemia
associated with micronutrient deficiencies other than iron in a probabilistic sample of Mexican children. Ann Nutr Metab 2006; 50:506–511.
17. Thi Le H, Brouwer ID, Burema J, Nguyen KC, et al
. Efficacy of iron fortification compared to iron supplementation among Vietnamese schoolchildren. Nutr J 2006; 5:32.
18. Weringa FT, Berger J, Dijkhuizen MA, et al
. Combined iron and zinc supplementation in infants improved iron and zinc status, but interactions reduced efficacy in a multicountry trial in southeast Asia. J Nutr 2007; 137:466–471.
19. Ross J, Horton S. Economic Consequences of Iron Deficiency
; Technical Report; Ottawa: Micronutrient Initiative; 1998.