|Home | About | Journals | Submit | Contact Us | Français|
Iron-deficiency anemia (IDA) is associated with alterations in infant behavior and development that may not be corrected with iron therapy.
To determine if a home-based intervention to foster child development improves behavior and development of infants with IDA.
Infants with IDA and nonanemic infants aged 6 and 12 months were treated with oral iron and randomly assigned to a year of surveillance or intervention. Infants in the surveillance group were visited weekly, and information on iron intake, feeding, and health were recorded. Infants in the intervention were visited weekly, and the home visits included an hour-long program to foster child development by providing support to the mother-infant relationship. The number of infants enrolled was 128 (66 who received intervention) and 149 (70 intervention) at 6 and 12 months, respectively. Psychologists who were unaware of iron status and intervention assignment assessed infants' cognitive, motor, and social-emotional development (Bayley Scales) at the beginning, midpoint, and end of the year; 116 6-month-olds and 134 12-month-olds had at least 2 assessments. Hierarchical linear modeling was used to analyze change over time.
Infants with IDA, regardless of enrollment age, were rated as less positive in social-emotional behavior at baseline. There were significant interactions between iron status and intervention associated with change in cognitive performance and positive social-emotional behavior. Infants with IDA who received intervention had developmental trajectories comparable to those of nonanemic infants in the intervention and surveillance groups, but these infants did not catch up in social-emotional behavior. Infants with IDA who received surveillance showed less increase in cognitive scores and had declines in positive social-emotional ratings.
Home-based intervention to foster child development improved cognitive and social-emotional scores in infants with IDA, but social-emotional differences remained between infants with IDA and those without IDA.
Results of numerous studies have demonstrated that infants with iron-deficiency anemia (IDA) show poorer cognitive, motor, social-emotional, and neurophysiologic functioning than those without.1–5 Studies that included reassessments after a full course of iron therapy have had mixed results. In the majority of these studies, alterations were found to persist in infants with IDA despite therapy,6–10 but complete correction of low cognitive and motor scores has also been reported.11,12 Results of follow-up studies at ages from later infancy to young adulthood have shown poorer outcomes in the cognitive, motor, social-emotional, and neurophysiologic domains in individuals treated in infancy for IDA or chronic iron deficiency. These results indicate long-term risk.3,13
These findings are similar to those reported for studies of undernourished infants, especially in developing countries. In these infants, results for developmental recovery have been more promising when developmental support was provided in addition to improved nutrition.14–23 To our knowledge, the use of this approach in infants with iron deficiency has not been reported. In our study we sought to determine if a home-based intervention program designed to foster child development, provided in combination with iron therapy, could improve behavior and development in infants with IDA.
This intervention study was conducted in conjunction with a randomized trial of developmental/behavioral effects of prevention of IDA in infancy. Infants from working-class communities near Santiago, Chile, who received well-child care in community clinics were considered for study participation. Entrance criteria included birth weight ≥3.0 kg, uncomplicated singleton term birth, and no acute or chronic health problems. Exclusion criteria included illiterate or mentally impaired caregiver, infant in day care, and residence outside the neighborhoods. Sample details have been published previously.24–29
The early-intervention component of the study, with enrollment conducted from 1991 to 1995, involved mothers and their infants who were identified as having IDA or who were clearly non-anemic at either 6 or 12 months.24–29 When the hematology laboratory identified infants with IDA, they also identified the next nonanemic infant on the basis of date and time of venipuncture, a random selection process. The laboratory provided the project coordinator with the identification numbers of infants who were identified for possible study inclusion during periods of 1 to 2 weeks and did not indicate iron status. Anemia was defined as a venous hemoglobin concentration of ≤100 g/L at 6 months and <110 g/L at 12 months; clearly nonanemic as a hemoglobin concentration of ≥115 g/L. Iron deficiency was defined as 2 or more abnormal iron measures (a mean corpuscular volume of <70 fL, free erythrocyte protoporphyrin concentration of ≥100 μg/dL red blood cells [1.77 μmol/L], serum ferritin concentration of <12 μg/L), and/or an increase in hemoglobin concentration of ≥10 g/L after 6 months of iron therapy. Infants with IDA and randomly selected infants without IDA at 6 months were invited to participate in early intervention and neuromaturation components of the project (6-month cohort) and were treated for 1 year with 15 mg per day of elemental iron as oral ferrous sulfate.30 The preventive trial ended when the infants were 12 months old, and then infants with IDA and randomly selected infants without IDA were also invited to join the early-intervention study (12-month cohort) and treated with oral iron (30 mg per day) for at least 6 months.
The institutional review boards of the University of Michigan Medical Center and the Instituto de Nutrición y Tec-nología de los Alimentos, University of Chile, approved the protocol. Signed informed consent was obtained from the parent(s) or guardians of all participating infants.
All infants received weekly home visits during the year of intervention study participation, beginning when the infants were 6 or 12 months old for the 6- and 12-month cohorts, respectively. For surveillance-only visits, project personnel recorded data on the infant's iron intake, feeding, and health. Intervention visits included implementation of a program to support child development, and surveillance data were also recorded. The frequency and duration of the intervention program were determined on the basis of research data available at the time.31,32 After considering several models, we focused the intervention on the mother-child relationship,33,34 because we anticipated that this focus would foster cognitive and social-emotional development, both of which were consistently found to be poorer in infants with IDA compared with infants without IDA.1–5 Consideration of the family context was also an integral part of the program.
Intervention visits, which lasted about an hour, were conducted by monitors, who were professional educators devoted to encouraging the development of individual abilities. The monitors were trained and supervised by project psychologists. Monitors worked half time and visited ñ6 families per week and had weekly group and individual supervision. Three manuals that covered the diagnostic phase and intervention in the first and second year of life were developed to guide training and ongoing work with families (available on request, in Spanish only).
The program consisted of 3 phases: introduction (14% of visits), intervention (70% of visits), and termination (16% of visits). The main introduction-phase objectives were to familiarize the mother with the program and have the monitor assess the strengths and weaknesses of the family and mother/child dyad. The monitor interviewed the mother, observed the mother and infant at home, and reached agreement with the mother about areas of the mother-infant relationship to be addressed. The main intervention-phase objective was to support the mother/infant dyad within the family context according to the agreed-on goals. Monitors showed mothers enjoyable verbal and nonverbal activities with the infants, demonstrated positive feedback, discussed child development issues, provided mothers with related written material, and supported them in their interactions with their infants. Monitors also considered with the mother ways to address problems or concerns related to the behavior and development of the infant and relationships in the family. The main termination-phase objectives were to reinforce the mother's independent abilities to support her child's development, resolve problems, and address ending the program and the year-long relationship with the monitor.
The study was a randomized single-blind clinical trial. Infants were randomly assigned to intervention or surveillance within IDA and nonanemic groups on the basis of a predetermined algorithm. All outcome measures were obtained by psychologists who had no role in the intervention program and were unaware if the infant had IDA or received intervention. Of necessity, parents and home visitors knew if a given infant received intervention.
To have >80% power to detect an intervention impact of 0.50 SD or more on cognitive or motor test scores, the desired sample size was 50 infants per group at each age. Infants were enrolled on the basis of openings in the schedule of the monitors, until a shortage of funds necessitated a halt in enrollment in August 1995. Cell sizes ended up being approximately two thirds of the desired n in the 6-month cohort, because IDA was uncommon at this young age. After study completion, we also noted that infants with IDA were over represented in the 12-month cohort. This result was likely attributable to the study design and limitations in the number of families that monitors could accommodate at a given time. The imbalance was not recognized until the study was completed and the code was broken.
Outcomes were assessed at 3 time points for a period of 1 year: at ages 6, 12, and 18 months for the 6-month cohort and 12,18, and 24 months for the 12-month cohort. Testers were trained to ≥85% agreement before starting the study, with random reliability checks thereafter. Overall cognitive and motor development were measured by use of the original Bayley Scales of Infant Development,35 the version available at the time. To have a single metric to assess change over time, we analyzed raw mental and motor scores. Behavioral measures were derived from a prerelease version of the Behavior Rating Scale36 and based on a previously reported factor analysis conducted for more than 1600 infants who completed the preventive trial at 12 months.24 The factors measured were (1) object orientation, (2) motor quality, (3) negative affect, and (4) positive social responsiveness. To permit analysis of behavioral change over time, we used factor score weights from the preventive trial analysis to construct the 4 factors at each time point.
Background/baseline characteristics at enrollment were compared by use of intention to treat (analysis of variance and χ2, SPSS 16.0 for Windows [SPSS Inc, Chicago, IL]). Effects of intervention, which required at least 2 data points to estimate change over time and thus could not be analyzed by intention to treat, were examined by use of 2-stage hierarchical linear modeling (HLM) of measures within individuals (HLM 6: SSI Scientific Software, Chicago, IL). For the within-person stage (level 1), we modeled 2 changes over time (start to midpoint and midpoint to end) with the intercept fixed and the 2 slopes allowed to vary. Individual age at testing, measured in months, was the time parameter. For the between-person stage (level 2), we modeled the impact of intervention group, iron status group, and the interaction between intervention and iron status during each half-year period. We first examined the 3-way interaction between iron group and intervention group over time for each cohort. When a significant interaction was identified, we examined the impact of the intervention on change in outcome separately according to iron status. Because there were no group differences in background characteristics at the level of randomization (see below), we identified potential covariates on the basis of significant correlations with cognitive or motor scores.
Enrollment included 128 infants at 6 months (66 intervention) and 149 infants at 12 months (70 intervention). Background characteristic are shown in Table 1 and Table 2. Intervention and surveillance groups within IDA and nonanemic groups (the levels of randomization) were similar in background in each cohort. As often observed,37 infants with IDA had lower gestational age or birth weight, and their mothers had less education (and lower IQ in the 6-month cohort); a higher proportion of infants with IDA were male. Maternal education and IQ did not correlate with cognitive test scores in these young infants. Gender and birth weight were the only background characteristics significantly related to developmental test scores; these variables were controlled in all analyses. There were no differences in temperament at 6 months38 (data not shown). Table 3 shows differences between infants with IDA and nonanemic infants in iron status measures.
For infants with IDA in both cohorts, ratings in positive social-emotional responsiveness (fourth behavioral rating factor) were lower than for infants without IDA, with differences of −0.22 SD in the 6-month cohort (95% CI: −0.28 to −0.16; P< .001) and −0.09 SD in the 12-month cohort (95% CI: −0.14 to −0.05; P = .002). Groups were similar in baseline mental and motor raw scores and other behavior-rating factors.
Results for exposure measured by attrition or number of visits seemed comparable. Figs 1 and and22 show the number of participants at each time point. In the 6-month cohort, there were no statistically significant differences in the drop-out rates between the first and second assessments or the second and third assessments. In the 12-month cohort, there was no difference in drop-out rate between the second and third assessments, but an indication that more participants dropped out from the intervention group between the first and second assessments (intervention versus surveillance:χ2 = 3.09; P = .08 for IDA and χ2 = 1.57; P = .21 for nonanemic). The increase in drop-out rates between the first and second assessments for intervention versus surveillance was statistically significant for infants who were dropped from the study because of lack of funding χ2 = 5.85; P = .02) but not for those who were removed for any other reason. Within each cohort, there were no significant differences between groups (iron status by intervention) in the mean number of home visits (Tables 1 and and22).
Infants with IDA showed an excellent response to iron (Table 3); only 5 infants did not have a clear response. However, iron status measures remained lower in infants with IDA after 6 months of treatment, compared with nonanemic infants. Hemoglobin was the only measure obtained after 12 months of iron therapy; levels remained lower in infants with IDA in the 12-month cohort.
The longitudinal analyses included 116 infants enrolled at 6 months (60 intervention) and 134 infants enrolled at 12 months (59 intervention). As expected, all groups showed gains in mental and motor raw scores overtime. Intervention did not affect gain in motor scores (data available on request). For cognitive scores, each cohort showed a significant 3-way interaction (intervention group according to iron status by monthly change in score: P = .03 and P = .04 for the 6- and 12-month cohorts, respectively). Change overtime was similar for nonanemic intervention, nonanemic surveillance, and IDA intervention groups and lower for the IDA surveillance group, as shown in Figs 3 and and44 for the 6- and 12-month cohorts. Table 4 provides details of the HLM parameters and statistics. Changes that occurred during the 1-year period are summarized here.
Raw mental scores in the 6-month cohort increased during the 1-year period by 54.0 points for the nonanemic intervention group (95% CI: 53.0–55.0), 52.6 points for the nonanemic surveillance group (95% CI: 51.5–53.7), and 51 points for the IDA intervention group (95% CI: 49.7–52.3). The increase for the IDA surveillance group was 37.7 points (95% CI: 36.1–39.3), significantly lower than the other groups (P = .01 and P = .03 for first and second half-years). The lower rate of change resulted in significantly lower scores after 6 and 12 months: 12.0 points (95% CI: 11.0–13.0) and 16.0 points (95% CI: 15.0–17.0) lower, respectively.
In the 12-month cohort, raw mental scores increased during the year by 41.1 points for the nonanemic intervention group (95% CI: 39.4–42.8), 40.5 points for the nonanemic surveillance group (95% CI: 39.0–42.0), and 38.7 points for the IDA intervention group (95% CI: 37.2–40.2). The increase for the IDA surveillance group was 27.3 points (95% CI: 25.7–28.9), significantly lower than for the other groups (P = .02 and .03 for first and second half-years, respectively). The lower rate of change resulted in significantly lower scores after 6 and 12 months in the study: 8.3 points (95% CI: 7.3–9.3) and 13.2 points (95% CI: 11.6–14.8) lower, respectively.
As with cognitive scores, changes in social responsiveness scores were similar for the nonanemic intervention, nonanemic surveillance, and IDA intervention groups and lower for infants with IDA who were randomly as-signed to the surveillance group. There was no evidence of catch up in positive social-emotional ratings of infants with IDA who were randomly assigned to the intervention group. The 3-way interaction was statistically significant in each cohort (intervention according to iron status according to monthly change in factor score: P = .002 and .01 in the 6- and 12-month cohorts, respectively; rates of change were similar in the first and second half-years).
Change in positive social-emotional responsiveness over time is shown in Figs 5 and and6.6. Briefly, social-emotional responsiveness factor scores in the 6-month cohort increased during the 1-year period by 0.24 SD for the nonanemic intervention group (95% CI: 0.22–0.26), 0.26 SD for the nonanemic surveillance group (95% CI: 0.22–0.30), and 0.22 SD for the IDA intervention group (95% CI: 0.19–0.25). The IDA surveillance group decreased 0.09 SD (95% CI: −0.06 to −0.12), a result that was significantly lower than the other groups (P = .01 and .02 for the first and second half-years). The decrease resulted in significantly lower scores after 6 and 12 months: 0.20 SD (95% CI: 0.18–0.22) and 0.33 SD (95% CI: 0.30– 0.36) lower, respectively.
In the 12-month cohort, social-emotional responsiveness factor scores increased during the year by 0.19 SD for the nonanemic intervention group (95% CI: 0.17–0.21), 0.19 SD for the nonanemic surveillance group (95% CI: 0.16–0.22), and 0.17 SD for the IDA intervention group (95% CI: 0.13– 0.21). The IDA surveillance group decreased 0.13 SD (95% CI: −0.17 to − 0.09), which was significantly lower than the other groups (P= .02 for both half-year intervals). The decrease resulted in significantly lower scores after 6 and 12 months in the study: 0.14 (95% CI: 0.12–0.16) and 0.30 SD (95% CI: 0.27–0.33) lower, respectively.
The results of this study showed that a home-based intervention in infants with IDA resulted in gains in cognitive and positive social-emotional scores that were similar to those observed in nonanemic infants regardless of intervention. Infants with IDA who were randomly assigned to the surveillance group showed smaller gains in cognitive performance and worsening positive social-emotional responsiveness. However, infants with IDA who received intervention did not catch up to non-anemic infants in social-emotional ratings; the similar rates of change indicate that the initial difference was maintained. Because infants were randomly assigned to groups, the benefits to infants with IDA seem to be attributable to the intervention program.
Comparisons of our findings with those from previous research are not possible, because no such intervention study has been published with respect to IDA. However, our findings are similar to those from studies of undernutrition (or zinc deficiency).14-–23,39 Such studies have shown that development improves when infants with generalized undernutrition or a specific micronutrient deficiency are offered early intervention in addition to health monitoring and nutritional rehabilitation.
The benefits that we observed corresponded to the nature of the intervention, which focused on the mother-infant relationship and aspects of maternal behavior thought to foster cognitive and social-emotional development.40 These are the areas in which infants with IDA who received intervention showed developmental improvements compared with those who received surveillance without intervention. Note that all infants were visited weekly to maintain comparability in the frequency of contact with study personnel. Thus, all infants could be considered to have received some kind of intervention. The impact of the early-intervention program may have been even greater compared with routine health care alone.
In contrast to previous investigators,1,3 we did not observe lower initial cognitive and motor scores in infants with IDA. In our study population, however, infants with IDA showed less positive social-emotional responsiveness and were observed to have delayed development, as assessed on the basis of more brain-based measures.26,29 This pattern of results suggests that in the first year of life global cognitive and motor scores are not as sensitive to IDA effects as are social-emotional or more brain-based measures. This interpretation is supported by our recent study of 9-month-old infants in an inner-city US population, in whom social-emotional measures showed more marked differences related to iron status than overall cognitive and motor scores.41–43 Positive social-emotional behavior was also the area most improved by iron supplementation in the preventive trial component of the Chile project.24 Together such results point to the following conclusions: positive social-emotional behavior may be fostered by iron supplementation as a preventive measure but is not fully corrected once IDA occurs, even with additional intervention. Social-emotional behavior may get worse without intervention, even when IDA is treated with iron.
Another possible reason why we did not observe lower initial cognitive and motor scores is that IDA was not present long enough to have an impact on global scores. No infant had IDA postnatally for >6 months, because IDA was identified and treated at 6 and 12 months. In previous studies, infants were older and probably had IDA longer.2–5 Even in the present study, however, infants with IDA who did not receive intervention showed lower cognitive scores as the year went on, despite iron therapy.
Our study could not be double-blind. Although infants were randomly assigned to surveillance and intervention groups and outcomes were measured by masked testers, mothers and families knew which group their infant was in. In regard to generalizability, it is unclear whether such an intervention program would be equally effective if IDA was more chronic or severe or families more or less advantaged than those in our sample. The labor-intensive nature of the intervention and our use of skilled professionals also limit broad application. Concerted efforts are underway to identify developmentally supportive interventions that are feasible and sustainable on a large scale.20,44 For instance, approaches involving less skilled personnel and homemade toys have shown good results in several contexts.15–18,21–23,39 The optimal age for intervention remains unclear, although there was no differential effect associated with infant age in our study. Processes by which early-intervention programs produce benefits are also not well understood and warrant additional research.
Developmental interventions hold promise to benefit infants with IDA, whereas iron supplementation alone seems to be insufficient. However, the lack of catch up in positive social-emotional behavior that we observed points to the need to prevent IDA from occurring in the first place. Because ˜25% of infants in developing countries have IDA,45,46 and infants everywhere whose families are poor or are members of minority groups are at increased risk,47,48 new avenues to improve the developmental outcomes of these infants are sorely needed.
Infants with IDA, despite iron therapy, show poorer cognitive, motor, social-emotional, and neurophysiologic function than infants without IDA. In malnourished infants, who also show poorer long-term outcome, early-intervention programs have improved developmental recovery.
Developmental interventions have not been reported for IDA, which affects ˜25% of infants worldwide. The results of this study showed that home-based intervention to foster child development improved mental scores and social responsiveness in infants with IDA.
This work was supported by grants from the National Institutes of Health (HD14122 and HD33487) and the Gerber Companies Foundation.
We thank Candice Percansky, MA, formerly of the Ounce of Prevention Fund and Erikson Institute in Chicago, Illinois, for consultation on the early-intervention program and Susan McDonough, PhD, for recommendations about parent-focused early intervention. We also thank the monitors, weekly visitors, psychologists, and supervisors for their dedicated efforts and professionalism and the study families for their participation.
Funded by the National Institutes of Health (NIH).
The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Gerber Companies Foundation.
This trial has been registered at www.clinicaltrials.gov (identifier NCT00998998).
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.