Iron fortification of foods is usually considered the most cost-effective approach to prevent iron deficiency. However, iron is the most difficult mineral to add to foods. When added as water-soluble, highly bioavailable compounds such as ferrous sulfate, the soluble iron rapidly catalyzes fat oxidation resulting in rancid products. In addition, water-soluble iron compounds can cause unacceptable color reactions during storage and food preparation. Thus, food manufacturers are often obliged to use water-insoluble iron compounds to fortify foods and fortification compounds such as elemental Fe powder and ferric pyrophosphate are widely used to fortify cereal flours and infant cereals. However, these compounds never dissolve completely in the gastric juice and are usually far less well absorbed than ferrous sulfate (Hurrell 1997). Ferrous fumarate on the other hand, although almost insoluble in water, readily dissolves in the gastric juice and has been shown to have an equivalent absorption to ferrous sulfate in healthy, Western adults (Hurrell et al. 1989, 2000). Because it is non-water soluble, it causes relatively few sensory problems in the fortified foods and is therefore an interesting food fortificant. Iron absorption from ferrous fumarate has been demonstrated to be significantly higher than from ferric pyrophosphate in European infants (Davidsson et al. 2000) and this compound is currently used to fortify blended cereal flours for food aid programs and commercial infant cereals in Europe. However, based on our recent study in Bangladeshi children, there is now concern that due to lower gastric acid output, young children in developing countries may not be able to absorb ferrous fumarate as well as Western adults (Davidsson et al. 2001a, Sarker et al. 2001, 2003). Clearly, there is a need to evaluate the efficacy of water insoluble iron compounds to prevent the development of iron deficiency/iron deficiency anemia in infants and young children living in developing countries. The aim of this study is to evaluate the efficacy of ferrous fumarate and ferric pyrophosphate, as compared to ferrous sulfate, as food fortificants in preventing development of anemia/IDA in Bangladeshi infants and young children (part I).

A potential cause of low gastric acid secretion in Bangladesh and many developing countries is Helicobacter pylori infection. Although H. pylori-infection appeared to have no influence on absorption of ferrous fumarate in children, the impact of chronic H. pylori infection in adults could be expected to be more pronounced due to long time effects on the gastric mucosa, resulting in reduced gastric acid output. The other aim of the study is therefore, to assess of iron absorption and gastric acid output in adult women of child-bearing age with H. pylori infection (part II).

Two hundred and forty non-anemic Children (Hb>105 g/L) will be randomized to three study groups; ferrous fumarate, ferric pyrophosphate or ferrous sulfate (n=80 per group) in wheat flour- and cow milk-based infant formula and will be fed for 9 months. Hemoglobin, serum ferritin, and transferin receptor will be analyzed at baseline and after 4.5 and 9 months of intervention. Prevalence of anemia and iron deficiency during and after the intervention among the three groups will be compared (part I). We furthermore propose a complementary study to determine the relative absorption of ferrous fumarate (relative to ferrous sulfate) in H. pylori infected and non-infected adult Bangladeshi women (15 each) of 20-40 year of age with IDA using stable isotope technique based on the incorporation of iron stable isotopes into erythrocytes 14 days after administration. Assessment of gastric acid output will also be performed. Iron stature and absorption, and assessment of gastric acid output will be compared before and after therapy in H. pylori infected women (part II). The results of this study are expected to have implications in the prevention and treatment of iron deficiency anemia in developing countries

The efficacy study will test the ability of ferrous fumarate and ferric pyrophosphate to prevent anemia/IDA in non-anemic Bangladeshi infants/young children in comparison to ferrous sulfate during a 9 month feeding trial. Hemoglobin and indicators of Fe status (ferritin and circulating transferrin receptor) and an acute phase reactant (C-reactive protein) will be monitored at baseline and after 4.5 and 9 months of intervention.

Subjects Non-anemic infants and young children (6-24 months; n=240) (Hb >105g/L) will be recruited for the study from the population at Nandipara, a peri-urban community near Dhaka city. All children will be randomized to one of three study groups; ferrous fumarate, ferric pyrophosphate or ferrous sulfate (n=80 per group), and stratified based on H. pylori infection (based on H. pylori stool antigen test, positive or negative) at base line. All infants/young children will be retested for H. pylori infection at the end of the study.

All parents or guardians will be fully informed about the aim and procedures of the study and informed consent will be obtained from at least one parent/guardian per child.

Exclusion criteria Exclusion criteria include children with anemia (Hb<105 g/L), systemic infection or apparent inflammatory process or weight for age of < 70% of NCHS median. The children with exclusion criteria will be appropriately treated or if needed, or will be referred to appropriate health center for treatment.

Sample size:

The sample size is based on the expected anemia prevalence after the intervention, i.e., children with anemia after 9 months of intake of fortified infant cereal. We assume that the prevalence of anemia will be 10% in infants consuming cereal fortified with ferrous sulfate, 30% in infants consuming cereal fortified with ferrous fumarate and 50% anemia in infants consuming cereal fortified with ferric pyrophosphate. To detect a difference of 20% points ((P1 - P2); sulfate versus fumarate and sulfate versus pyrophosphate) with 95% confidence and power=0.80, the sample size is calculated to 62 infants/young children per group.

Assuming a drop out rate of about 25 %, we will recruit 80 infants/young children per group i.e. 240 non-anemic infants/young children.

We assume a high prevalence of anemia (70-80%) in the study population and will therefore need to screen approximately 1000-1200 children.

Fortified food Each serving will consist of 25 g infant cereal based on wheat or rice flour and cow milk (produced especially for the study; 9.3 mg Fe as ferrous fumarate, ferric pyrophosphate or ferrous sulfate/25 g dry cereal will be added during manufacture. Ascorbic acid will be added during manufacture at a molar ratio 3:1 relative to added Fe. Fortified infant cereals will be administered by health workers, one serving per day, 6 days per week, under close supervision by the study physician/investigators. Intake will be monitored weekly and, if needed, fed during 7 days per week to compensate for days when the child was sick or absent.

The fortified food will provide the Recommended Nutrient Intake of iron for infants 7-12 months according to WHO/FAO 2002 (in press; please see PAHO/FNP/USAID consultation 2001) per serving, i.e., 9.3 mg Fe. Fortified infant cereal will be fed once per day, 6 days per week, for 9 months.

Dietary information by simplified food frequency questionnaire (SFFQ), validated and currently being used by the Bangladesh Integrated Nutrition Project (BINP), will also be obtained twice a month. Anthropometrical measurement e.g. weight, height, and mid upper arm circumference (MUAC) will be measured at baseline, at 4.5 months and 9 months of intervention.

All parents or guardians will be fully informed about the aim and procedures of the study and informed consent will be obtained from at least one parent/guardian per child. All anemic infants/young children (diagnosed during the screening study) will be treated with medicinal Fe according to standard treatment practice at ICDDR.B; (ferrous sulfate drops; (2 mg Fe/kg BW/d) for 2 months. Infants/young children with anemia at 4.5 and 9 months (efficacy study) will be treated with medicinal Fe (ferrous sulfate drops; (2 mg Fe/kg BW/d) for 2 months and excluded from the study. Medicinal Fe will be administered by health workers.

Hookworm infection is not prevalent in this community but all children participating in the study will be screened for hookworm infection at the time each blood sampling. Infected children will be treated with Albendazole 400 mg single dose for two times (2 weeks apart) but will not be excluded from the intervention study.

Blood analysis Venous blood samples will be drawn (2 ml in EDTA treated tubes) during the screening study to identify non-anemic infants/young children and at baseline and after 4.5 and 9 months of intervention. Hb will be analyzed using the cyanomethemoglobin method (Sigma kit, Sigma, St Louis, MO). Plasma ferritin and plasma circulating transferrin receptor will be measured by ELISA technique by using commercial kits during the initial screening study (Ramco, Houston, TX). Samples drawn at baseline and after 4.5 and 9 months of intervention will be analyzed at KUMC for plasma ferritin and plasma circulating transferrin receptor by ELISA assays based on monoclonal antibodies (Flowers et al. 1986)/dual monoclonal antibodies (Flowers et al. 1989). Plasma samples will also be analyzed for C-reactive protein (CRP) at KUMC. CRP will be monitored as an acute phase reactant. Dr. Cook and Ms. Flowers are currently developing a monoclonal ELISA technique for CRP, which will be used in this study. Commercial quality control materials (DiaMed, Cressier sur Morat, Switzerland, Sigma and Ramco) will be analyzed together with all series of samples analyzed for Hb and ferritin.

Food analysis Infant cereals will be analyzed for Fe and calcium (Ca) by electrothermal/flame atomic absorption spectroscopy (SpectrAA 400, Varian, Mulgrave, Australia) after mineralization by microwave digestion (MLS 1200) using a HNO3/H2O2 mixture, using standard addition technique to minimize matrix effects. The phytic acid content will be determined by HPLC technique (Sandberg & Ahderinne 1986, Sandberg et al. 1989

Fecal samples; screening for H. pylori infection and hookworm infection Stool specimens will be collected and tested for the presence of H. pylori antigen (HpSA) by a newly developed enzyme immunoassay (EIA). A polyclonal commercial kit (H. pylori antigen EIA; Cat# 740096, Novitech, Freiburg, Germany) will be used. The HpSA test has been demonstrated to be sensitive (>80 %) and specific (>95 %) for the detection of H.pylori infection (Braden et al. 2000, Vakil et al. 2000, Sarker et al. personal communication 2003). Stool samples will also be screened for the presence of parasite eggs, using standard microscopic technique.

Statistics ANOVA will be used to compare variables within each study group as well as between the 3 study groups. Results with significant F ratio (p<0.05) will be further analyzed by Student's t-test using multiple comparison procedure. Different outcome variables, i.e., Hb, ferritin, and TfR concentrations before and after intervention in each group will be compared by paired t-test. Data with skewed distribution will be normalized by log transformation before statistical analysis. Chi-squared test will be used to evaluate binary variables at different time points during the study. Odds Ratios and confidence intervals will be calculated. Multivariate analysis (logistic regression and multiple regression) will also be performed to adjust for confounders, e.g. age, nutritional status.

Inclusion Criteria:

• Children of 9-24 months
• Non-anaemic children (haemoglobin more than 10.5g/L)

Exclusion Criteria:

Exclusion criteria include:

• Children with anemia (Hb<105 g/L), systemic infection or apparent inflammatory process or weight for age of < 70% of NCHS median.
• The children with exclusion criteria will be appropriately treated or if needed, or will be referred to appropriate health center for treatment.