Childhood iron deficiency is associated with behavioral and cognitive delays. Few studies have explored the relationship between prolonged bottle-feeding and iron-deficiency anemia among toddlers.
To examine the association between prolonged bottle-feeding and iron deficiency in a nationally representative sample of children ages 1 to 3 years.
Design and Methods
The National Health and Nutrition Examination Survey III provides data on the feeding practices of children 1 to 3 years old and contains measures of iron status including transferrin saturation, free erythrocyte protoporphyrin, and serum ferritin. The prevalence of iron deficiency and duration of bottle-feeding were determined for black, white, and Mexican American toddlers. Bivariate and multivariate analyses were performed to examine the association between bottle-feeding duration and iron deficiency.
Among 2121 children ages 1 to 3 years, the prevalence of iron deficiency was 6% among whites, 8% among blacks, and 17% among Mexican Americans (P<.001). With increasing duration of bottle-feeding, the prevalence of iron deficiency among all children increased (3.8%, bottle-fed ≤12 months; 11.5%, bottle-fed 13-23 months; and 12.4%, bottle-fed 24-48 months [P<.001]). At 24 to 48 months of age, 36.8% of Mexican American children were still bottle-fed, compared with 16.9% of white and 13.8% of black children. In multivariate analyses, bottle-feeding for 24 to 48 months and Mexican ethnicity were associated with iron deficiency (odds ratio, 2.8; 95% confidence interval, 1.3-6.0; and odds ratio, 2.9; 95% confidence interval, 1.5-5.6, respectively).
Children with prolonged bottle-feeding and Mexican American children are at higher risk for iron deficiency. Screening practices and nutritional counseling should be targeted at these high-risk groups.
Iron-deficiency anemia in infancy and early childhood is associated with behavioral and cognitive delays. Several studies have documented the delayed mental and psychomotor development and behavioral disturbances which result from iron-deficiency anemia that may persist into adolescence. Adverse consequences include impaired learning,1 decreased school achievement,2,3 and lower scores on tests of mental and motor development.4,5
Iron deficiency is the most common nutritional deficiency in childhood.6 The prevalence of iron deficiency without anemia is much greater than the prevalence of iron deficiency with anemia, with the potential for a much greater public health impact.3,7 In iron deficiency without anemia, central nervous system iron stores decrease before restriction of red cell production; these biochemical changes have been shown to impair behavior in infants.8,9 Children 9 to 24 months old are especially susceptible because this is a period of accelerated growth. The prevalence of iron deficiency in the United States was found to be 9% among toddlers (herein, “toddlers” refers to children ages 12-30 months, and “infants” refers to those less than 12 months old10) in a recent analysis of the National Health and Nutrition Examination Survey (NHANES) III.7 Several studies of infants in various urban areas of the United States reveal an iron deficiency prevalence of 17% to 44%, the highest being among infants of lower socioeconomic status.11 These differences in the prevalence of iron deficiency might contribute to and widen existing racial/ethnic and socioeconomic disparities in academic achievement and other measures of well-being. Because of this, prevention of iron deficiency in early childhood is an important pediatric public health issue.
Sufficient dietary intake of iron is essential for toddlers to maintain a positive iron balance and thus to prevent iron-deficiency anemia. Modifiable dietary habits, such as limiting the amount of cow’s milk intake, may be suitable targets for interventions to prevent iron-deficiency in childhood. A 1995 study indicated that 8% of children in the United States ages 2 to 5 years are still bottle-fed, and that Latino ethnicity, poverty status, urban residence, and low parental education are associated with continued bottle use.12
The American Academy of Pediatrics recommends that infants be introduced to the cup around 6 months of age and be fully weaned from the bottle by 15 months to optimize appropriate feeding patterns.13 While several studies have examined the prevalence and health consequences of prolonged bottle-feeding,12 including its role as a causal factor in early childhood caries,14 only one small study has explored the association between bottle use and iron-deficiency anemia.15 The objective of this study was to examine the association between prolonged bottle-feeding and iron deficiency among a nationally representative sample of children 1 to 3 years of age. We hypothesized that prolonged bottle-feeding is a risk factor for iron deficiency. Clinical experience also suggests that Mexican American children may also be at increased risk because they are more likely to be bottle-fed past 15 months of age.
The data source for these analyses was the NHANES III, a large-scale national survey conducted by the National Center for Health Statistics.16 Administered from 1988 through 1994 in 2 phases of equal length and sample size, the survey includes a sample of 33 994 persons. Both phase I and phase II include representative samples of the civilian noninstitutionalized United States population 2 months of age and older living in households. The persons selected were asked to complete an extensive interview and an examination in a mobile health center. The response rate for the interview component of the survey was 86%. Results were weighted to adjust for nonresponse and to provide national estimates.
All survey data collection instruments were available in Spanish as well as English. Questionnaires developed by the National Center for Health Statistics staff were first translated into Spanish. During the pretests, they were checked in both Spanish and English to ascertain their accuracy in both languages. Precise definitions of questionnaire terms were incorporated into detailed instruction manuals and interviewers received intensive training specific to the survey and its questionnaires.
Independent variables included age and sex. Poverty status was dichotomized as below the federal poverty threshold vs at or above the federal poverty threshold, based on family size and the federal poverty threshold at the time of the survey.17 The child’s race/ethnicity was defined by parental self-identification, and included non-Hispanic white, non-Hispanic black, and Mexican American. Because of small sample sizes, other Hispanic, Asian/Pacific Islander, Native American, other, and multiple race/ethnic groups were excluded from the analyses. Other independent variables included caretaker educational attainment (self-reported highest grade completed, categorized as not a high school graduate, high school graduate, or at least some college); weight-for-height status (using age- and sex-specific weight-for-length percentiles, with at risk for being overweight defined as a weight-for-length status of ≥85th and <95th percentile, and overweight defined as a weight-for-length status of ≥95th percentile [BMI was not used because only weight-for-length measurements were available for children 1-3 years of age]); birth weight (<2500 g vs ≥2500 g); and blood lead level. To assess the appropriateness of the duration of breast milk or formula-feeding in relation to the American Academy of Pediatrics guidelines,13 we also examined the age at which breastfeeding or formula-feeding was discontinued. Exclusive breastfeeding beyond 6 months was not included, since its prevalence was low in the study sample. Laboratory values were measured using standard measurement assays, the details of which are provided elsewhere.18,19
We used the definitions of iron deficiency and anemia previously described by Looker et al7 in their evaluation of the prevalence of anemia in the United States using the same NHANES III database. The diagnosis of iron deficiency was based on 3 laboratory tests of iron status: transferrin saturation, free erythrocyte protoporphyrin, and serum ferritin. An individual was considered iron-deficient if any 2 of these 3 values were abnormal for age and sex. For children between the ages of 1 to 2 years, the cutoff values for tests of iron status are less than 10% transferrin saturation, less than 10 μg/L of serum ferritin, and greater than 1.42 μmol/L of RBCs erythrocyte protoporphyrin. For 3-year-old children, these cutoff values are less than 12%, less than 10 μg/L, and greater than 1.24 μmol/L of RBCs, respectively.
The prevalence of iron deficiency and the duration of bottle-feeding were determined for toddlers from 3 major racial/ethnic groups (black, white, and Mexican American). Bivariate analyses were performed to determine the association between iron deficiency and bottle-feeding duration. Multiple logistic regression was used to examine the association between bottle-feeding duration and iron deficiency, adjusting for age, sex, lead level, weight-for-height status, maternal educational attainment, poverty status, and the age at which breastfeeding or formula-feeding was discontinued.
To account for the complex, multistaged sampling design of the NHANES III, SUDAAN software was used to obtain weighted frequencies, odds ratios, and 95% confidence intervals.20 Logistic regression was used for multivariate analysis and χ2 tests were used to test for differences in proportions.
Interactions were examined for the multivariate model. Because interaction terms did not enter as significant independent variables in the vast majority of analyses, they were not included in the final multivariate model; this was also done to avoid overfitting. We tested for colinearity in the multiple linear regression analysis, but there was none found.
Among the 2121 children ages 1 to 3 years included in the sample, 9% had iron deficiency (Table 1). The prevalence of iron deficiency was 5.7% among white children, 8% among black children, and 16.9% among Mexican American children (P<.001). Iron deficiency was most prevalent among 1-year-old children (16.4%), compared with those ages 2 (5.6%), and 3 (5.9%). Iron-deficiency anemia was present in 1.2% of white children, 3.5% of black children, and 5.5% of Mexican American children (Table 1). In addition, with increasing duration of bottle-feeding, the prevalence of iron deficiency among all children increased. The prevalence of iron deficiency was 3.8% among those infants bottle-fed 12 months or less, 11.5% among those bottle-fed 13 to 23 months, and 12.4% among those bottle-fed 24 to 48 months (P<.001). Among toddlers with both risk factors (Mexican American ethnicity and prolonged bottle-feeding [greater than 12 months]), the prevalence of iron deficiency was 18.5% vs 3.3% among those with neither risk factor. There was no significant association between iron deficiency and sex, poverty status, parental education, weight-for-height status, birth weight, lead level, or the age at which breastfeeding or formula feeding was discontinued. Iron-deficiency anemia was found to be significantly associated with nonwhite race/ethnicity, younger age, bottle-feeding duration, poverty, parental educational attainment, and low birth weight, but not with any of the other factors examined.
We then examined the duration of bottle-feeding by race and ethnicity (Table 2). Most children were bottle-fed for less than 23 months. However, at 24 to 48 months, 36.8% of Mexican American children were still being bottle-fed, compared with 16.9% of white and 13.8% of black children (P<.001).
In the multivariate analysis (Table 3), bottle-feeding into the 24- to 48-month age range and Mexican ethnicity were both associated with a significantly greater adjusted odds of iron deficiency. Mexican American children were 3 times more likely to be iron deficient compared with white children (odds ratio, 2.9; 95% confidence interval, 1.5-5.6), and children bottle-fed for greater than 24 months were almost 3 times as likely to be iron deficient compared with children bottle-fed for 12 months or less (odds ratio, 2.8; 95% confidence interval, 1.3-6.0).
These analyses demonstrate that iron deficiency is a persistent problem in children in the United States, with a prevalence as high as 19% in certain groups. Prolonged bottle-feeding is significantly associated with iron deficiency, with increasing rates of iron deficiency as the duration of bottle-feeding increased. After adjustment for potential confounding factors, both prolonged bottle-feeding and Mexican ethnicity are associated with iron deficiency.
Although iron deficiency has become less common in the United States since the introduction of iron-fortified infant formulas, it remains the leading cause of anemia in infants.21 Prior analyses of the NHANES III (1988-1994) yielded a 3% prevalence of iron-deficiency anemia among children in the United States ages 1 to 3 years.7 Eden and Mir22 found that more than one third of urban children had iron insufficiency, and of these, 7% were iron deficient without anemia, and 10% had iron-deficiency anemia. These authors point out that much work remains to be done to prevent iron deficiency during the second year of life.
The prevalence of iron deficiency was extremely high among Mexican American children, identifying yet another disparity in the health of Latino children in the United States. We also found that Mexican American parents are much more likely than white or black parents to bottle-feed their infants for prolonged periods of time. This has been a consistent clinical observation, and was previously reported in a study utilizing the 1991 National Health Interview Survey.12 This suggests that prolonged bottle-feeding among Mexican American infants may at least be partly responsible for the high prevalence of iron deficiency seen in this group. However, Mexican ethnicity remained significantly associated with iron deficiency even with statistical adjustment for the influence of prolonged bottle-feeding. Therefore, there may be other as yet unidentified culturally mediated infant feeding practices among Mexican American families apart from prolonged bottle-feeding that increase the risk for iron deficiency. Latino normative cultural values regarding desirable infant body habitus may shape dietary practices, such as encouraging infants to drink large volumes of milk or juices.
While the connection of prolonged bottle-feeding and anemia has long been recognized by clinicians, this is the first study that we are aware of to report an association between prolonged bottle-feeding and iron deficiency in a nationally representative sample of children ages 1 to 3 years. We are aware of only one other small study that has explored the association between prolonged bottle use and iron-deficiency anemia. This was a cross-sectional survey of a convenience sample of the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) participants at 3 Bronx, NY sites with large Hispanic and African American populations. Caregivers of 95 WIC-enrolled children ages 18 to 56 months completed a questionnaire with detailed nutritional questions. Bottle-feeding was found to be significantly associated with weight-for-height status and with iron-deficiency anemia (25% of those bottle-fed had anemia vs 15% of those not bottle-fed [P<.05]). This study was limited, however, by its relatively small sample size and geographical and urban focus.15
The American Academy of Pediatrics Committee on Nutrition has stated that excessive ingestion of cow’s milk contributes to iron deficiency by preventing intake of adequate iron-rich foods and increasing gastrointestinal blood loss.13 Indeed, clinical experience suggests that excess consumption of cow’s milk is linked to prolonged bottle-feeding,22 which has been documented in several studies. A study of 34 toddlers at their 18-month health supervision visit showed that toddlers still on the bottle at 18 months drank significantly more cow’s milk than those not on the bottle.21 In another study of 165 children observed from infancy, bottle-fed children consumed more cow’s milk than their weaned counterparts.23 The likely mechanism through which prolonged bottle-feeding is associated with iron deficiency is the consumption of large volumes of non–iron-fortified cow’s milk resulting in gastrointestinal blood loss and the displacement of iron-rich foods from the diet. Iron-fortified formulas designed for toddlers have recently been introduced, but their use could not be ascertained in NHANES III.
A number of social and cultural factors may contribute to prolonged bottle-feeding among parents of various racial/ethnic groups. For example, a qualitative study of cultural influences on feeding practices among Hmong toddlers attending WIC sites in St Paul, Minn, revealed that Hmong caregivers’ child-raising philosophy coupled with feeding practices chosen to adjust to American society contribute to the high rate of prolonged bottle-feeding.24 Consistent with the findings of our study, the prevalence of iron deficiency was high among this Hmong population of children. The Latino population in the United States is growing rapidly, yet no studies have examined cultural influences on infant feeding practices among Latino families.
Certain study limitations should be noted. First, the NHANES III is a cross-sectional survey, limiting inferences of causality. Second, other dietary information relevant to iron deficiency, such as the volume of milk and various iron-rich foods consumed, is lacking in NHANES III. Third, analyses of other Latino subgroups (Puerto Ricans, Cubans, and others) could not be performed because of small sample sizes. Lastly, the data from NHANES III are relatively old now; therefore, a similar analysis will be needed when updated data become available.
Further studies are needed to evaluate the mechanisms through which prolonged bottle-feeding is associated with iron deficiency, and to examine reasons why mothers continue to bottle-feed their infants for so long. These studies will be crucial to the design of interventions aimed at decreasing the duration of bottle-feeding among toddlers in the United States.
Screening and counseling practices should be modified to address the increased risk of iron deficiency among children with prolonged bottle-feeding. Our study findings suggest that prolonged bottle-feeding is a marker of clinical significance because of its association with iron deficiency. Asking parents about bottle-feeding in children older than 12 months may help to identify children at risk for iron deficiency and may indicate the need to evaluate the child’s iron status. As suggested in a recent study, prophylactic supplementation of iron for all children until 2 years of age may need to be considered because current screening by testing hemoglobin is insufficient to identify most cases of iron deficiency.25 Clinicians especially should be aware of the high prevalence of both prolonged bottle-feeding and iron deficiency among Mexican American toddlers. Anticipatory guidance around feeding issues during the first year of life may help encourage parents to wean children from bottle-feeding by 15 months.26 If consistently followed, this practice may help to successfully promote healthy nutrient intake and prevent the harmful effects of iron deficiency in early childhood.
Correspondence: Jane M. Brotanek, MD, MPH, Center for the Advancement of Underserved Children, Department of Pediatrics, Medical College of Wisconsin, 8701 Watertown Plank Rd, Milwaukee, WI 53226 (firstname.lastname@example.org).
Previous Presentation: This study was presented in part as a platform presentation at the annual meetings of the Pediatric Academic Societies; May 3, 2004; San Francisco, Calif; and as a poster at the Academy Health Research Meeting; June 5-6, 2004; San Diego, Calif.
Accepted for Publication: April 28, 2005.
Funding/Support: This study was supported by grant 2-T32-HP 12002-16, BHP-HRSA to the Center for Child Health Research, Rochester, NY, from the National Research Service Award Program.
Acknowledgment: We thank Cynthia Howard, MD, MPH, for sharing her insights and expertise in infant feeding/breastfeeding. We also thank Kelly Conn for her assistance with the preparation of the manuscript.
Brotanek JM, Halterman JS, Auinger P, Flores G, Weitzman M. Iron Deficiency, Prolonged Bottle-Feeding, and Racial/Ethnic Disparities in Young Children. Arch Pediatr Adolesc Med. 2005;159(11):1038-1042. doi:10.1001/archpedi.159.11.1038