Efficacy of an iron‐fortified infant cereal to reduce the risk of iron deficiency anemia in young children in East Cameroon

Abstract Complementary foods in Africa are often poor sources of bioavailable iron. We assessed the efficacy of iron‐fortified wheat‐based infant cereal (IC) to reduce the risk of iron deficiency anemia in children aged 18–59 months in Cameroon. A 6‐month double‐blind, cluster‐randomized controlled trial was conducted in 2017 among anemic (hemoglobin 7–11 g/dl) but otherwise healthy children. In conjunction with usual diet, children received two 50 g servings/day of a standard, micronutrient‐fortified IC (providing 3.75 mg iron/serving; n = 106) or the same IC without iron fortification (n = 99). Anthropometric measurements, blood sampling, and systematic deworming were performed in all children at baseline (pre‐intervention), 3, and 6 months. Mean hemoglobin, ferritin adjusted for C‐reactive protein (CRP), serum iron, transferrin saturation, prevalence of anemia, iron deficiency, and iron deficiency anemia as well as anthropometrics were compared between the groups at baseline, 3, and 6 months. Compared to the control group, children consuming the iron‐fortified IC had significantly higher baseline‐adjusted mean hemoglobin (10.0 ± 1.8 vs. 9.7 ± 1.4 g/dl, respectively; p = .023), ferritin adjusted for CRP (16.1 ± 8.3 vs. 9.5 ± 7.5 μg/L, p < .001), serum iron (14.5 ± 3.9 vs. 11.2 ± 4.4 μg/dl; p < .001), and transferrin saturation (19.0 ± 17.4 vs. 10.7 ± 12.5%; p ˂ .001) at 6 months. The prevalence of anemia, iron deficiency, and iron deficiency anemia at 6 months decreased by a larger extent in the iron‐fortified group versus controls (all p < .01). In addition, at 6 months, children in the iron‐fortified group demonstrated higher weight‐for‐age z‐scores (p = .016) compared to the control group. Wheat‐based IC fortified with 7.5 mg ferrous fumarate administered daily for 6 months improved iron and nutritional status and decreased the prevalence of iron deficiency anemia in children aged 18–59 months in Salapoumbé, Cameroon.

proving iron status (Bouhouch et al., 2016) and correcting iron deficiency anemia (Andang'o et al., 2007;Sun et al., 2007) in South America, Asia, and Africa, other studies reported mixed or negative results (Assuncao, Santos, Barros, Gigante, & Victora, 2007;Glinz et al., 2015Glinz et al., , 2017Nestel et al., 2004;Rohner et al., 2010). A systematic review of both efficacy and effectiveness trials concluded that consumption of iron-fortified foods results in improved iron status and reduced risk of anemia, yet significant heterogeneity was observed for most outcomes (Gera, Sachdev, & Boy, 2012). Another systematic review of flour fortification programs in 13 countries (none located in Africa) concluded that only limited evidence exists for the effectiveness of iron-fortified flour in reducing anemia prevalence (Pachon, Spohrer, Mei, & Serdula, 2015), perhaps due to variability in program implementation (e.g., compliance, coverage, feasibility in rural settings), iron compound, and iron levels. Together, these results suggest the need for further studies on the efficacy of iron fortification.
Per WHO recommendations, several countries have routinely adopted the strategy of using iron-fortified foods to combat iron deficiency (Nutrition International, 2018). In Africa, 26 of 53 countries including Cameroon have officially adopted iron fortification of cereal flours to reduce the burden of iron deficiency anemia in children (Food Fortification Initiative, 2018). In Cameroon, iron is added to wheat flour as ferrous fumarate (60 mg Fe/kg) (Food Fortification Initiative, 2018). The time and effort to prepare complementary foods from fortified flour rather than from the same foods eaten by older family members, however, may prevent the successful implementation of this policy in all households. For example, in rural Cameroon, complementary foods are often traditional meals (e.g., maize paste and okra sauce) and made with family staples (potatoes, cassava, rice, beans, and peanuts), with infrequent usage of wheat flour (Mananga, Kana-Sop, Nolla, Tetanye, & Gouado, 2014).
Furthermore, the finding that only ~75% of wheat flour samples in Cameroon were actually fortified, plus the lack of improvement in hemoglobin concentration or anemia prevalence in response to wheat flour fortification in Cameroonian children (who commonly receive antihelminthic medications [Institut National de la Statistique (INS) and ICF International, 2012]), suggests that other interventions are needed (Engle-Stone, Nankap, et al., 2017). Nutrient-rich foods that are consistently fortified, convenient to prepare, and readily accepted by young children, such as commercial infant cereal, are important options that can help improve iron consumption during this crucial period of child development. Therefore, we conducted a cluster-randomized controlled trial to evaluate iron fortification of commercial wheat-based infant cereal (IC) to reduce the risk of iron deficiency anemia in young Cameroonian children.

| ME THODS
This double-blind, cluster-randomized, controlled trial was conducted from February to August 2017 in the locality of Salapoumbé (Cameroon). A random sample of 366 children from the entire locality was initially selected within 30 villages based on local vaccination registers. Further inclusion criteria for this study included age 18-59 months, hemoglobin level between 7 and 11 g/dl and apparent good health. Exclusion criteria were also applied, including iron supplementation in progress per caregiver report, clinical presentation of severe malnutrition (e.g., bilateral pitting edema), diagnosis of any chronic infection (e.g., tuberculosis, HIV), severe acute infection (e.g., severe malaria, pneumonia, meningitis), blood transfusion <3 months prior to enrollment, or allergy/intolerance to cows' milk and/or gluten. From the initial sample of 366 children, 205 were eligible after inclusion and exclusion criteria were applied. Children with weight-for-height z-scores <−3 were not excluded; however, only six of the 205 children had scores <−3 at baseline.
The minimum sample size was calculated for 90% statistical power and α-level of 5%, with a predicted difference for mean hemoglobin of ≥0.9 g/dl and a standard deviation 1.48 g/dl (Eichler et al., 2012) in the iron-fortified group between baseline and study completion at 6 months. The minimum sample size was 71 children per group (142 subjects total). To account for dropouts, the entire eligible population of 205 children was enrolled.
The Salapoumbé locality is in the East region of Cameroon that is quite rural and known to be malaria endemic. Salapoumbé is divided by the Lokomo River, which is large and difficult to cross, thus separating the locality into two settlements. The two settlements are generally ethnically comparable in eating habits and customs as well as at the socioeconomic level, notably in terms of literacy and agriculture. Thus, the south and north banks were considered two clusters and the IC (fortified or control) was randomized by cluster.
This design was implemented to minimize treatment crossover due to potential sharing of IC between families located on the same bank of the river.
The batches of IC were manufactured by Nestlé Central and West Africa Ltd. PMB KIA Accra-Ghana according to international cereal fortification standards . Bags of IC were packaged in waterproof plastic and were identical in appearance except for a code (A or B) designating whether the batch was iron-enriched. The code assignment was unknown to researchers and subjects throughout the study period. The random allocation of IC A or B to the south or north bank was performed by the principal investigator by a flip of a coin. All subjects on each bank received the same IC, either A or B. The study team obtained the support of administrative officials and traditional chiefs of the respective villages to ensure monitoring of the correct distribution of the IC lots.
Community health workers also verified through periodic home visits that children enrolled in the study were receiving the IC packets corresponding to the randomized area. The manufacturer sent the batch code assignment to the principal investigator after the statistical analysis was complete.
Parents of enrolled children completed a sociodemographic characteristics questionnaire based on a protocol previously validated by the Comité National d'Ethique et de la Recherche pour la Santé Humaine du Cameroun (CNERSH). Data obtained included the vaccination history of the child, the occurrence of illness during the previous 2 weeks, and the child's date of birth. Child anthropometric assessments included weight measurement using an electronic portable scale (SECA LK 5051) with a sensitivity of 0.5 g. The scale was calibrated after every 10 measurements. Length or height was measured using a SECA portable infantometer or stadiometer with an accuracy of 0.1 cm. Immediately after the anthropometric measurements were taken by trained personnel, each child's hemoglobin level was measured with a portable hemoglobinometer (HemoCue 201+). A 3 ml sample of venous blood was collected and centrifuged, and the serum was immediately frozen for analysis by a Biotek EL800X machine to measure levels of ferritin, serum iron, transferrin, and C-reactive protein (CRP). All assays were performed according to manufacturer's instructions. Appropriate procedures (e.g., portable freezers or generators) were used in the field to maintain Serum samples were stored frozen in Salapoumbé (which is partly equipped with generators for rural electrification) and transported frozen to Yaoundé in batches for analysis due to the distance between the sites (>700 km). Anthropometric measurements and biological tests were repeated at 3 and 6 months. In addition, the children had systematic deworming with a mebendazole 500 mg tablet every 3 months.
In addition to their habitual diet, all enrolled children were fed two 50 g servings/day (morning and evening) of the assigned IC by their mothers, who had been previously educated on hygienic measures of IC preparation and administration. The fortified IC was prepared with previously boiled warm water. The daily ration of 100 g of IC provided 7.5 mg of iron as ferrous fumarate to children in the iron-fortified group; otherwise, the two cereals were identical in terms of appearance, taste, and dietary composition ( Table 1).
The parents and guardians of the children were provided with IC bags every 15 days on a precise and known date. To guarantee the consumption of the planned quantities of IC for each selected child, each household in the study received additional bags of the same IC intended for all of the enrolled child's siblings. At each date of IC supply, parents were questioned for possible adverse events related to the IC administration.

| Ethics
The study protocol was approved by CNERSH and administrative authorization was received from the Cameroon Ministry of Public Health. The study protocol followed local laws and the Declaration of Helsinki. Prior to data collection, parents of enrolled children signed written informed consent after a detailed explanation of the objectives, nature, and risks of the study. All enrolled children who presented with any comorbid condition during the study received free treatment from study medical staff (nurses or physicians). All children who were not enrolled due to severe anemia or severe malnutrition were referred to the local hospital for routine care (e.g., iron supplementation, blood transfusion, inpatient treatment). At the end of the study after code break, all children in the control group received additional ferrous sulfate for 3 months in accordance with standard guidelines (World Health Organization, 2016); anemic children from either group were referred for treatment.

| Statistical analysis
The primary endpoints of the study were hemoglobin, serum ferritin adjusted to CRP, serum iron, transferrin saturation and prevalence of anemia, iron deficiency, and iron deficiency ane-  (Beilby et al., 1992;Yamanishi, Iyama, Yamaguchi, Kanakura, & Iwatani, 2003). The secondary endpoints were changes in weight, height, and other anthropometric z-scores.
Analyses were conducted using Epi Info software (version 3.5.4, CDC), WHO Anthro (version 3.2.2, WHO), and SPSS (version 20, SPSS Inc., IBM). The WHO Anthro software was used to calculate weight-for-height, height-for-age, and weight-for-age z-scores in comparison with the WHO standard reference population (World Health Organization, 2006). Descriptive variables for the two groups receiving IC are presented by group and compared using standard t tests for continuous measures and chi-square tests for categorical measures. Biomarkers of iron status are presented as unadjusted means and standard deviation. Because these data were non-normally distributed, comparisons between the groups for these measures were done using a nonparametric rank-based ANOVA model at baseline and using a nonparametric rank-based ANCOVA model at 3 and 6 months, including adjustment for the baseline value.
Height and weight as well as the z-scores for the anthropometrics were compared between groups at each time point using a standard parametric ANOVA model (baseline) or ANCOVA model (at 3 and 6 months, again with adjustment for baseline values). Differences between the groups for the prevalence of anemia, iron deficiency, and iron deficiency anemia at baseline, 3, and 6 months were assessed using Fisher's exact test. The threshold of significance was 0.05 for all comparisons.

| RE SULTS
Of the 366 children aged 18-59 months identified in the locality of Salapoumbé, 205 children were included the following: 106 children on the south bank who were cluster-randomized to receive the iron-fortified IC and 99 children on the north bank who received the control IC. Of the 205 children who participated in the study, 153 children were evaluated at the end of the intervention with a similar dropout rate across the two groups (23% in the iron-fortified group and 28% in the control group) (Figure 1). The ICs were well tolerated and no side effects were reported in either group. Table 2 presents the demographic characteristics of the enrolled children. Mean age at enrollment was 32.1 months in the iron-fortified group and 36.1 months in the control group (p-value from t test = .009) but no other significant differences were seen between the groups for parental profession or education or family size (Table 2). Overall about 41% of the children had infection or inflammation (CRP > 5 mg/L) at baseline. At baseline, both groups were comparable in weight, height, and weight-for-age z-scores, while the iron-fortified group had higher height-for-age and lower weight-for-height z-scores than the control group (Table 3). The study population as a whole was mildly to moderately undernourished at baseline, with approximately 56% of children stunted (height-for-age z-score <−2) and 25% underweight (weight-for-age z-score <−2); only 3% of the children were wasted (weight-for-height z-score <−2). At the end of the intervention, the iron-fortified group had a very small (~50 g) but borderline significant higher weight gain (p-value from ANCOVA = .052) and a significantly higher weight-for-age z-score (p = .016) compared to the control group (Table 3). Height, heightfor-age, and weight-for-height z-scores were similar in both groups at the end of the intervention. The overall proportion of stunted and underweight children decreased to 52% and 11%, respectively, at the end of the intervention, with no children considered to be wasted; however, anthropometry measures were missing for 27% of the population.

| D ISCUSS I ON
This study found that in addition to their usual diet, daily con- week (Andang'o et al., 2007) and in an Indian study of children aged 6-15 years who received a daily portion of wheat flour fortified with 6 mg of the same iron compound (Muthayya et al., 2012).
In addition, a recent study in children aged 6-12 months in India using a similar but rice-based IC (providing 3.75 mg iron/day as ferrous fumarate) fed for 6 months showed better iron status and reduced risk for anemia and iron deficiency in association with more favorable neurodevelopmental scores (Awasthi et al., 2020).
However, our results showed greater efficacy than those reported in another study in India in which children aged 6-13 years received a daily portion of rice fortified with 20 mg iron as

TA B L E 3 (Continued)
micronized ground ferric pyrophosphate (Moretti et al., 2006) and a study in children aged 6-14 years in Côte d'Ivoire who received biscuits fortified with 20 mg electrolytic iron four times per week (Rohner et al., 2010). This difference could be explained by the fact that the fortified IC in the present study also provided vitamin C, known to promote gut iron absorption (de Almeida et al., 2003), and vitamin A, the deficiency of which decreases the absorption of iron (Hurrell & Egli, 2010). Differences in results may also be due in part to the varied iron fortificant compounds used in the studies, which have different bioavailabilities (Hurrell & Egli, 2007), as well as variability in the population characteristics including age, baseline iron status, and inflammation.
In addition, in the present study, significant increases in mean serum iron, CRP-adjusted ferritin, and transferrin saturation were observed in the iron-fortified group, and at the end of the intervention, there were significant decreases in the prevalence of anemia, iron deficiency, and iron deficiency anemia in the iron-fortified group. These results are comparable to those described in the study done in India with NaFeEDTA-enriched wheat flour (Muthayya et al., 2012). In contrast, the results differ from those in long-term studies conducted in preschoolers (age 9-71 months) and older children (age 6-11 years) in Sri Lanka (Nestel et al., 2004) and in children under age 6 years in Brazil (Assuncao et al., 2007) that demonstrated no effect of iron-fortified wheat flour on hemoglobin or anemia. While low hemoglobin status may be due to other factors besides low iron intake and thus an inadequate biomarker to test the efficacy of iron fortification, this difference could also be explained by the fact that the population of the present study consisted exclusively of anemic children, whereas the proportion of anemic children was 8% and 30%, respectively, in the other two clinical trials. Furthermore, the results may differ because the fortificant used in the previously mentioned trials was electrolytic iron, known to have a lower bioavailability than ferrous fumarate, which was used in the present study Walter, Pizarro, Boy, & Abrams, 2004). The nonsignificant improvement in hemoglobin levels and indicators of iron status (ferritin, serum iron, and transferrin) and the equally nonsignificant decrease in the prevalence of anemia, iron deficiency, and iron deficiency anemia in the control group is unlikely to be due to the cluster randomization, and such a finding has been described in other studies in India (Muthayya et al., 2012) and Côte d'Ivoire (Rohner et al., 2010). These changes in the control group may be related to the vitamin C and vitamin A content of the control IC or to systematic deworming, known for its beneficial effect on re- ducing digestive parasites and improving absorption of iron from the usual diet (Dreyfuss et al., 2000;Stoltzfus, Dreyfuss, Chwaya, & Albonico, 1997). The increase of iron deficiency and iron deficiency anemia prevalence between midpoint and endpoint in the F I G U R E 2 Prevalence of anemia (panel a), iron deficiency (panel b), and iron deficiency anemia (panel c) in both groups at baseline, 3, and 6 months. Comparisons between groups by Fisher's exact test control group could be explained by the lack of iron supplementation in children recovering from mild malnutrition during the first 3 months of intervention (Macdougall, Moodley, Eyberg, & Quirk, 1982;Velasquez Rodriguez et al., 2007).
There was also indication that nutritional status improved after 6 months of intervention with a small but borderline significant higher weight gain (p = .052) and a significantly higher increase in weightfor-age z-score (p = .016) in the iron-fortified group compared to the control group. Mean weight-for-age z-score at 6 months remained below but moved closer to the WHO median in the fortified group compared to controls. These findings are confirmatory for the beneficial effect of iron fortification of wheat-based IC on the nutritional status of children, as demonstrated in other studies (Aukett, Parks, Scott, & Wharton, 1986;Barth-Jaeggi et al., 2015;Briend, Hoque, & Aziz, 1990). This beneficial effect may be due to iron stimulation of appetite in children (Lawless, Latham, Stephenson, Kinoti, & Pertet, 1994;Sachdev, Gera, & Nestel, 2005;Stoltzfus et al., 2004).
In addition, the proportion of underweight children in the overall study population decreased from 25% to 11%, likely due to the energy content of the ICs. The high proportion of stunting observed in our population has been observed previously (Pondy, 2016)  Cameroon has been reported to be 31 g/day, representing an intake of 1.86 mg/day of iron from this food source (Engle-Stone, Nankap, et al., 2017). Therefore, the iron provided by the fortified IC (7.5 mg/day) likely represented a sizable proportion of total daily intake.
An important limitation of this study is that only two clusters were randomized, and thus, the differences observed at the end of the study between the clusters located north and south of the river could theoretically be due to differences in living conditions in the two areas; however, as the two areas are largely comparable in terms of customs, diet, level of literacy, and agro-pastoral activities, this seems to be an unlikely explanation of the results observed. Moreover, the lack of assessment of malaria in both groups during the intervention did not permit us to conclude on the safety of iron fortification in this endemic malaria area as shown in other studies (Prentice, Verhoef, & Cerami, 2013;Zlotkin et al., 2013).
However, the iron concentration of the IC (3.75 mg/serving) is much lower than that of micronutrient powders (12.5 mg/serving) previously associated with increased risk of hospitalization and mortality in malaria-endemic areas (Sazawal et al., 2006). This lower iron dose within a food matrix is unlikely to promote the appearance of plasma nontransferrin bound iron that can enhance bacterial virulence (Brittenham, 2012). The lack of information on compliance or the amount of cereal consumed is an additional study limitation.
Finally, no formal adjustment for multiplicity was performed to account for multiple primary endpoints (seven parameters). However, if a Bonferroni correction is applied (p-value of interest = .05/7 or .00714), differences between groups for six of the seven endpoints remain statistically significant at study end, with only hemoglobin no longer significant. Adjusting for multiplicity does not change the overall interpretation of the study results.
The strengths of the study include the rural setting and subject selection process, which resulted in a study population that is generally representative of this area of Cameroon and other regions of Africa. In addition, the cluster-randomized design helped minimize treatment "contamination" between intervention and control participants, which can reduce the apparent effectiveness of an intervention (Torgerson, 2001).

| CON CLUS ION
In conclusion, these results suggest that daily consumption of micronutrient-fortified IC enriched with 7.5 mg of iron as ferrous fumarate for 6 months improved hemoglobin and iron status and decreased the prevalence of anemia, iron deficiency, and iron deficiency anemia in children under 5 years of age in a malaria-endemic setting, in which it is more difficult to demonstrate a positive effect due to the high infection rate. In addition, the children receiving the iron-fortified IC had improved growth parameters. Given that most households in this region consume wheat flour, a geographically expanded program of iron fortification of affordable cereals in association with other public health interventions, such as prevention of malaria and promotion of immunization, could help to prevent and correct iron deficiency anemia and improve nutritional status in the locality of Salapoumbé, and possibly throughout Cameroon and beyond.

ACK N OWLED G M ENTS
We would like to thank the Ministry of Public Health of Cameroon,

CO N FLI C T O F I NTE R E S T
NPH is employed by Société des Produits Nestlé SA. No other author has any conflict of interest to report.

E TH I C A L A PPROVA L
This study was approved by the Comité National d'Ethique et de la Recherche pour la Santé Humaine du Cameroun, and administrative authorization was received from the Cameroon Ministry of Public Health.

I N FO R M ED CO N S ENT
Written informed consent was obtained from the parents of all study participants.