 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* Department of Pediatrics, Boston University School of Medicine and Boston Medical Center, Boston, Massachusetts
Department of Emergency Medicine, Boston University School of Medicine and Boston Medical Center, Boston, Massachusetts
Departments of Pathology and Laboratory Medicine, Harvard Medical School and Children's Hospital, Boston, Massachusetts
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods. The study was a double-blinded, randomized, pragmatic, clinical trial conducted at 3 urban primary care clinics. Subjects included healthy, full-term infants who were enrolled at their 6-month well-child visit. Infants were randomly assigned to receive standard-dose multivitamins with or without iron (10 mg/day). Parents administered multivitamins by mouth daily for 3 months. Laboratory results at 9 months of age were analyzed for the presence of anemia and/or iron deficiency. Anemia was defined as hemoglobin level <11.0 g/dL. Iron deficiency was initially defined as any abnormal laboratory value of the following: mean corpuscular volume combined with red cell distribution width or zinc protoporphyrin (with blood lead level <10 µg/dL) for most subjects and ferritin, transferrin saturation, or reticulocyte hemoglobin content for a subset. Subsequent analyses defined iron deficiency as any 2 abnormalities of the above laboratory outcomes, except hemoglobin.
Results. The control (n = 138) and intervention (n = 146) groups were equivalent with respect to all important sociodemographic and nutritional variables. At 9 months of age, anemia was found in 21% of infants (n = 58). A total of 229 (81%) had iron deficiency on the basis of 1 abnormal laboratory indicator and 139 (49%) on the basis of 2 abnormal laboratory indicators. No difference existed in the occurrence of anemia and iron deficiency between the intervention and control groups. In the intervention group, 22% and 78% of 138, respectively, were anemic or had 1 abnormal laboratory outcome indicative of iron deficiency. In the control group, 19% and 84% of 144 were anemic or iron deficient. When stratified by adherence, no differences in hematologic outcomes between groups were noted. However, in multivariate logistic regression, infants whose mothers were anemic during pregnancy were 2.15 times more likely than others to have any laboratory abnormality (95% confidence interval: 1.14–4.07). Increasing adherence, regardless of group assignment, was associated with a 0.56 times reduced risk of any abnormality (95% confidence interval: 0.41–0.76).
Conclusion. On the basis of intention-to-treat analysis, multivitamins with iron was not effective in preventing iron deficiency or anemia in 9-month-old infants. However, effective prevention and treatment of maternal anemia during pregnancy and giving multivitamins with or without additional iron during infancy may prove to be important approaches to the prevention of iron deficiency among high-risk children. Because of the consequences of iron deficiency and its high prevalence among low-income infants, additional investigation in these areas is warranted.
Key Words: nutrition • anemia • iron deficiency • vitamins • pediatrics
Abbreviations: WIC, Special Supplemental Nutrition Program for Women, Infants, and Children • Hb, hemoglobin • MCV, mean corpuscular volume • RDW, red cell distribution width • Fer, ferritin • TS, transferrin saturation • CHr, reticulocyte hemoglobin content • CDC, Centers for Disease Control and Prevention • OR, odds ratio • CI, confidence interval
Iron deficiency is the most common nutrient deficiency and cause of anemia in childhood.1 Iron deficiency is associated with numerous adverse health effects, particularly cognitive impairment, that may be irreversible2,3 and has been shown to be a precursor of lead poisoning.4 In the United States, an estimated 720 000 (9%) children aged 1 to 2 years are iron deficient, 240 000 (3%) of whom are anemic.5,6 These conditions disproportionately affect low-income children. Despite reports of declining prevalence of iron deficiency anemia among young children,6–8 data from the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC)9 and local studies10 suggest that rates of iron deficiency and anemia of 15% to 19% persist among low-income infants and toddlers.
Few recent US studies have assessed clinic-based approaches to primary prevention using supplementation with medicinal iron. Instead, studies have focused on primary prevention using dietary interventions, with varying results.11–14 The beneficial effects of ferrous sulfate supplementation have been supported by other recent research conducted overseas. Domellof et al15 found a reduction in the prevalence of anemia in healthy, full-term, breastfed Honduran infants who were given supplemental iron from 4 or 6 months to 9 months of age but not among Swedish infants, who had a much lower baseline prevalence of anemia. At 9 months of age, 9% of the supplemented Honduran infants were anemic, compared with 29% of the infants who received placebo. Similarly, Dijkhuizen et al16 found a 38% reduction of the prevalence of anemia in healthy, full-term Indonesian infants who received iron supplementation compared with infants who received placebo. Because vitamin supplements are widely used among preschool children,17 we believed that parents would be receptive to administering a daily vitamin supplement to their infants. In addition, a recent pediatric commentary in Pediatric News18 called for universal iron prophylaxis for infants and toddlers in lieu of current recommendations of the American Academy of Pediatrics to screen and treat anemic children.19
From 1996 to 1998, we conducted a randomized clinical trial of the use of multivitamins with iron as prophylaxis against iron deficiency anemia in infants from 6 to 9 months of age.20 We chose this age range because full-term infants typically exhaust maternally derived iron stores by then.21 In this study, at 9 months of age, anemia was documented in 11% of the infants in the intervention group who received multivitamins with iron and in 19% in the control group who received multivitamins without iron. After stratifying by adherence, infants in the intervention group were 50% less likely to be anemic. This previous study, however, was limited by poor adherence, loss to follow-up, and limited laboratory data with which to assess iron deficiency. Because of these limitations, we conducted this randomized clinical trial.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study was designed to be consistent with current clinical practice at each site so as to simulate how the intervention would be implemented in actual clinical practice. No laboratory studies were done at enrollment because this is not routinely done in clinical practice and our pilot studies indicated that enrollment would be greatly reduced by parental refusal to permit this additional phlebotomy. No additional patient visits were added to the standard protocols for well-child visits. Parents were reminded about visits and called with reminders after missed appointments as is normally done at these clinics.
Sample size calculations were based on rates of anemia documented in our first study, with an expected reduction of the rate of anemia by 50% among the infants who received multivitamins with iron. Using a prevalence of 22% and assuming an 
of .05 and power of 80%, 196 subjects were needed in each group.
At their 6-month well-child visits, all infants who received care at the study sites were approached and screened for enrollment in the study. Infants with premature birth (<37 weeks); low birth weight (<2500 g); major medical, hematologic, or gastrointestinal conditions; and previous supplement use were excluded from the study. At enrollment, parents were interviewed regarding demographics and dietary history, including age in days, gender, race, birth weight in pounds, duration of breastfeeding, use of whole cow milk, introduction of meats, use of iron-fortified cereal, use of iron-fortified formula, receipt of WIC benefits, and education by their pediatrician regarding iron-rich foods. We assessed household food security and hunger with the short form of the US Department of Agriculture Household Food Security Scale.22
The infants were randomized to 2 study groups. Vitamins were repackaged and relabeled to maintain blinding with randomization coordinated by the Boston Medical Center Pharmacy. Research assistants enrolled subjects at each site into the 2 study groups. Randomization was implemented through previous computer-generated randomized assignment of subject numbers to each group by the pharmacy. Blinding of the researchers, research assistants, and parents was maintained throughout the study. The group assignment listing was kept in a sealed envelope until study completion.
The intervention group received standard, nonprescription infant multivitamin drops with iron. Each milliliter of the vitamins contained the following: elemental iron, 10 mg; vitamin A, 1500 IU; vitamin C, 35 mg; vitamin D, 400 IU; vitamin E, 5 IU; thiamine, 0.5 mg; riboflavin, 0.6 mg; niacin, 8 mg; vitamin B6, 0.4 mg; and vitamin B12, 2 µg. The control group received the same multivitamins without iron. Parents were instructed to administer 1 mL of the multivitamins by mouth to their infants each day. Combined with other dietary sources of iron, it was estimated that infants who received the multivitamins with iron would receive 
2 mg/kg/day elemental iron. Parents were given a 3-month supply of vitamins.
At the 9-month well-child visit, hematologic outcomes were measured. Given high rates of missed and delayed appointments, all infants who had specimens drawn between 9 and 12 months of age were included in analyses. These outcomes included hemoglobin (Hb), zinc protoporphyrin, blood lead, mean corpuscular volume (MCV), and red cell distribution width (RDW). For a subset of the infants from whom sufficient blood could be obtained, additional hematologic outcomes included ferritin (Fer), transferring saturation (TS), and reticulocyte hemoglobin content (CHr). Because the infants who had the additional tests conducted were evenly distributed between both study groups, we used the full number of subjects with a Hb level in each group for reporting the prevalences of summary outcomes, such as "any lab abnormality."
Cutoff values for Hb, RDW, MCV, Fer, and TS followed recommendations of the US Centers for Disease Control and Prevention (CDC).1 The CDC guidelines are the standard of care for primary care practice in the United States and suggest the following cutoff values for children younger than 2 years: Hb <11 g/dL, RDW >14.5%, MCV <77 fL, Fer <15 µg/L, and TS <16%. Although some recent research literature5,15 cites lower cutoff values for Fer, TS, and MCV, we chose to use the current CDC guidelines that are intended for clinical practice, particularly primary care. In defining anemia for black infants, no adjustments of Hb values were made; the CDC does not recommend adjustment of Hb cutoff because the basis of variation in Hb distribution by race has not been determined.1 The zinc protoporphyrin cutoff (>35 µg/dL with a blood lead <10 µg/dL) followed widely used guidelines from the Childhood Lead Poisoning Prevention Laboratory, Massachusetts Department of Public Health. CDC guidelines do not address CHr; therefore, the CHr cutoff (<27 pg) was chosen on the basis of published reference data from the laboratory at which the testing was performed.23
The definition of iron deficiency followed commonly accepted criteria that are used in clinical practice at our study sites. The CDC guidelines, although defining cutoff points for abnormal tests of iron sufficiency, do not offer definitions of iron deficiency based on these tests, alone or in combination, with the exception of low serum Fer concentration.1 MCV and RDW were used as a combined measure to minimize the effect of thalassemia trait in interpreting results. In applying a more rigorous definition of iron deficiency anemia, we also determined the prevalence of iron deficiency on the basis of any 2 abnormal laboratory values. We analyzed TS and CHr both as separate indicators of iron deficiency and as a combined measure.
In addition, we analyzed available prenatal maternal Hb values nearest the delivery date using CDC trimester-specific cutoff values to define anemia.1 Maternal Hb levels can fluctuate as a result of hemodilution near term; however, we chose these near-term values to have greater consistency in the timing of the testing during the pregnancy.
During monthly reminder telephone calls, research assistants assessed adherence by parental report. Adherence was measured by verbal report with the following categories: refused to answer, vitamins given every day (6–7 days per week), vitamins given usually (4–5 days per week), vitamins given sometimes (1–3 days per week), or vitamins never given. Full adherence was defined as "vitamins given every day," and partial adherence was defined as "vitamins given usually" or "vitamins given sometimes." Although we attempted to assess adherence via bottle return, nearly all parents failed to return bottles. Primary analyses were intent-to-treat; subsequent analyses included only those with full adherence and then any known adherence. Adherence level was analyzed as "none," "sometimes," "usually," and "every day," with each category representing a unit increase over the next.
Analysis included comparisons of means using t tests and 
2 tests from bivariate 2 x 2 table analyses. Results were analyzed for all children who completed the study (and had complete blood counts tested, 282 of 284 with laboratory test results) and then stratified on the basis of adherence. Using multivariate logistic regression and stratified analyses, we examined the associations between baseline data and adherence with abnormal hematologic outcomes in the infants. Analyses were conducted using STATA 7.0 for Windows (Stata Corp, College Station, TX).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
A total of 179 infants were in the intervention group, and 197 were in the control group. A total of 284 (76%) infants, 138 (76%) in the intervention group and 146 (74%) in the control group, had laboratory results obtained between 9 and 12 months of age and were included in analyses. Those with laboratory test results that were analyzed did not differ from those without results except that those with results were more likely to receive WIC, a finding that is expected, given WIC requirements for anemia testing on recertification at 1 year of age.
Of the families in the intervention group, 16 moved, were lost to follow-up, or refused blood sampling, and 25 did not have laboratory values obtained until after 12 months of age; in the control group, 21 and 30, respectively, were not included in analyses. Two of 284 children with laboratory tests done between 9 and 12 months did not have a complete blood count, leaving 146 infants in the control group and 138 in the intervention group.
The control and intervention groups were similar with respect to all sociodemographic characteristics, including pregnancy history and household food security, and infant characteristics, including age, gender, race, breastfeeding, WIC enrollment, and birth weight. In addition to being evenly distributed between study group, gender did not affect any hematologic outcome. The only difference between groups was that infants in the intervention group were more likely than control subjects to be accompanied by their mothers (95% vs 85%, P = .03; Table 1).
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our finding of an association between maternal prenatal anemia and iron deficiency in the infant has not been previously reported in the United States and deserves additional study. Limited data from developing countries suggest that a causal link exists,24–28 and a biologically plausible mechanism has been documented in one study.29 The high prevalence of prenatal anemia among low-income women has persisted, with the CDC's Pregnancy Nutrition Surveillance System reporting a third-trimester pregnancy rate of 29% among women enrolled in WIC, a level that has not changed during the previous 7 years of monitoring.30 If maternal anemia is a causal factor, then efforts at primary prevention of iron deficiency in infants could be aimed in an entirely new direction.
Similarly, it is biologically plausible that multivitamins with or without additional iron could be associated with improved hematologic outcomes. For example, the role of vitamin C in improved absorption of dietary iron is well established.31,32 Alternatively, adherence to the use of vitamins may be a proxy for overall better infant care practices or other unidentified maternal characteristics. It is possible that parents may have inflated or overestimated the amount of vitamins actually given to their infants; however, this would not explain the association between adherence and improved hematologic outcomes. If reported use of the vitamins was a proxy for other maternal characteristics, then it seems equally plausible that these characteristics would predispose parents toward truly better adherence and not just a false reporting of adherence estimates.
This study has a number of limitations. Measures of iron deficiency and iron deficiency anemia are imprecise; thus, laboratory diagnosis of this common disorder is not simple.1 The diagnosis must rely on a combination of measures. We chose to use a variety of biochemical and hematologic measures either alone or in combination. Also, we could not assess or control for the prevalence of -thalassemia trait in our study population. Also, while laboratory diagnosis may be imprecise, the results were unchanged when the data were reanalyzed on the basis of 2 abnormal laboratory findings. Our finding of 49% of infants with evidence of iron deficiency is consistent with other studies.
It is possible that some of the children with anemia did not have iron deficiency as has been suggested in recent literature33; however, these children should have been randomly distributed between the 2 groups. We did not assess baseline iron status because this is not standard practice and may have discouraged many parents from enrolling their infants in the study. The study was conducted as a pragmatic randomized clinical trial and followed methods similar to those of other peer-reviewed research.34 As such, we did not alter clinic practice except by adding the additional laboratory measures. Children with preexisting anemia or iron deficiency should have been equally distributed between control and intervention groups as a result of the randomization. Furthermore, Irigoyen et al35 demonstrated that iron deficiency anemia is uncommon among 6-month-old infants who are enrolled in the WIC program. The additional loss to follow-up among enrolled subjects also reflects the difficulty in both providing clinical care and conducting research with low-income populations. We did not quantify other sources of dietary iron, such as meat consumption. The range of meat consumption by the infants should have been evenly distributed between study groups, and we found no differences between groups in terms of qualitative reporting of any meat consumption. In addition, analyses indicated that children who declined to participate or had no blood specimens drawn and those who had blood specimens were similar to the entire cohort. We also carefully followed CONSORT requirements for analyzing and reporting randomized clinical trials.36
As demonstrated in this study and others,10 including the CDC's Pediatric Nutrition Surveillance System,9 the high prevalence of iron deficiency and iron deficiency anemia among low-income infants continues. This is of substantial concern, particularly in light of the recent findings that a period of iron deficiency is associated with subsequent elevated blood lead levels4 and that the impact of elevated lead levels <10 µg/dl on neurocognitive function may be greater than the impact of higher levels.37 The elevated risk of iron deficiency among low-income children has persisted despite numerous recommendations for primary intervention strategies. To date, no evidence satisfactorily explains this increased risk. It is likely that iron deficiency in low-income infants results from a combination of factors, including insufficient dietary iron content, food insecurity, the early introduction of whole milk, and delayed introduction of meat into the infant diet. However, if, as suggested by our data and that of others,24–28 maternal iron deficiency is a causative factor in infantile iron deficiency, then the persistence of this problem among low-income infants would be expected, as the high prevalence of iron deficiency among pregnant low-income women has not changed in recent years.30 This hypothesis deserves additional exploration in the high-risk populations of the United States.
For the primary prevention of iron deficiency in full-term infants, the CDC currently recommends exclusive breastfeeding for the first 4 to 6 months, encouraging the consumption of iron-rich foods and supplemented formula and cereals, avoiding whole cow milk until after 12 months of age and limiting its consumption to 24 oz thereafter, and considering iron supplementation for breastfed infants whose consumption of iron from supplementary foods is judged to be inadequate by 6 months of age.1 Although overseas studies of the use of iron supplementation as a primary prevention strategy have been encouraging, the results of our study suggest that multivitamins with iron may not prevent iron deficiency anemia among the population of urban, low-income infants in the United States. We are hopeful that the use of multivitamins, with or without iron, may yet prove to be a successful primary prevention approach. However, more studies will be necessary to test this strategy. We believe that because of the consequences of iron deficiency and its high prevalence among low-income infants, additional investigation of the prevention of iron deficiency among infants is needed. Until an effective primary prevention strategy can be identified, universal screening for secondary prevention of iron deficiency among high-risk populations of infants and preschool children, as recommended by the CDC,1 is warranted.
|
ACKNOWLEDGMENTS |
|---|
Presented in part at the Ambulatory Pediatrics Association at the Pediatric Academic Societies' Annual Meeting; Baltimore, MD; May 7, 2002, and May 5, 2003.
|
FOOTNOTES |
|---|
Reprint requests to (P.G.) Whittier Street Health Center, 1125 Tremont St, Roxbury, MA 02120. E-mail: pgeltman{at}bu.edu
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. Bauchner Atoms Arch. Dis. Child., October 1, 2007; 92(10): 831 - 831. [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Brotanek, J. Gosz, M. Weitzman, and G. Flores Iron Deficiency in Early Childhood in the United States: Risk Factors and Racial/Ethnic Disparities Pediatrics, September 1, 2007; 120(3): 568 - 575. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
