Meat as a First Complementary Food for Breastfed Infants: Feasibility and Impact on Zinc Intake and Status

Abstract

Objective: 

This study was undertaken to assess the feasibility and effects of consuming either meat or iron-fortified infant cereal as the first complementary food.

Methods: 

Eighty-eight exclusively breastfed infants were enrolled at 4 months of age and randomized to receive either pureed beef or iron-fortified infant cereal as the first complementary food, starting after 5 months and continuing until 7 months. Dietary, anthropometric, and developmental data were obtained longitudinally until 12 months, and biomarkers of zinc and iron status were measured at 9 months.

Results: 

Mean (±SE) daily zinc intake from complementary foods at 7 months for infants in the meat group was 1.9 ± 0.2 mg, whereas that of the cereal group was 0.6 ± 0.1 mg, which is approximately 25% of the estimated average requirement. Tolerance and acceptance were comparable for the two intervention foods. Increase in head circumference from 7 to 12 months was greater for the meat group, and zinc and protein intakes were predictors of head growth. Biochemical status did not differ by feeding group, but approximately 20% of the infants had low (<60 μg/dL) plasma zinc concentrations, and 30% to 40% had low plasma ferritin concentrations (<12 μg/L). Motor and mental subscales did not differ between groups, but there was a trend for a higher behavior index at 12 months in the meat group.

Conclusions: 

Introduction of meat as an early complementary food for exclusively breastfed infants is feasible and was associated with improved zinc intake and potential benefits. The high percentage of infants with biochemical evidence of marginal zinc and iron status suggests that additional investigations of optimal complementary feeding practices for breastfed infants in the United States are warranted.

INTRODUCTION

Zinc and iron are recognized as “problem nutrients” for the older breastfed infant because of the challenge of obtaining adequate intake from exclusive breastfeeding and the resultant dependence on complementary foods to meet dietary requirements (^1,2). Indeed, it has been estimated that for the 9-month-old breastfed infant, over 90% of the projected requirement for these two trace elements would need to be provided by complementary foods (³).

The importance of complementary foods to meet iron requirements of the older infant has long been recognized, and iron-fortified cereal is commonly recommended as a first complementary food to prevent development of iron deficiency. Iron deficiency in infants and young children, however, continues to be cited as the most common micronutrient deficiency (⁴). With the more recent recognition of the widespread prevalence of zinc deficiency on a global basis (^5,6), often co-existing with iron deficiency, a number of strategies have been proposed to meet both iron and zinc needs of the older infant, including supplementation programs and fortification of staple foods (^1,7). Increased use of animal products has been recognized as an option but has often been considered unrealistic (⁷).

Meat is an excellent source of both iron and zinc and has been recommended at 4 months as an alternative to iron-fortified cereal (^8,9). The American Academy of Pediatrics has also indicated that meats are an appropriate early complementary food (¹⁰), but results of a recent survey indicated that less than 10% of 7 to 8 months olds eat baby food or nonbaby food meats, whereas 43% eat mixed dishes as sources of protein, which are much lower in zinc and iron (¹¹). There have been no trials to examine the feasibility or potential benefit of earlier introduction of meat as a complementary food for exclusively breastfed infants, that is, those who consume no infant formula or other fortified beverages.

We have previously published a comparison of zinc absorption from meat and cereal in a small subgroup of 7-month-old breastfed infants who were participants in the study described in this report was observed. Significantly greater zinc absorption from a test meal of meat compared with a test meal of cereal (¹²). In this report, we describe the results of the larger study that was undertaken to compare the feasibility and effects of randomizing exclusively breastfed infants to either meat or iron-fortified infant cereal as the first complementary food. We hypothesized that meat intake would be associated with greater zinc intake, increased growth, higher scores on developmental testing, and improved zinc and iron biochemical status.

METHODS

Study Design

The design was a randomized controlled trial for which the intervention consisted of assignment of exclusively breastfed infants at 4 months to either pureed beef or iron-fortified infant cereal as the first complementary food. Parents were encouraged to wait until approximately 5 months to start the complementary food. The infants remained on the assigned food (plus fruits and vegetables as desired) until 7 months of age when they were allowed to liberalize their diet to all age appropriate foods, including the alternative study food. Thus, at most, this was a 2 month intervention. Longitudinal dietary, anthropometric, and developmental testing data were obtained monthly from 4 to 7 months and at 9 and 12 months. Biochemical tests of zinc and iron status were obtained at 9 months, when many infants have blood drawn as part of the health maintenance visit. The rationale for continuing to follow the infants after 7 months was to examine the possibility that early exposure to meat might result in greater acceptance (and therefore intake) compared with infants who are first offered meats as one of the last types of complementary food.

Subjects

Eighty-eight infants (43 male, 45 female) enrolled in the study at 4 months. The number of infants decreased during the intervention period, with 79 at 5 months, 76 at 6 months, and 72 at 7 months; 72 finished the entire 12 months of study. Reasons for drop-out included inability to continue with exclusive breastfeeding, return to work, and moving; drop-outs were distributed equally across the two feeding groups, and none dropped out because of objections to the assigned complementary food. Recruitment methods included advertisements in physician's offices and parenting newsletters and word of mouth. For inclusion in the study, the infants were required to be healthy, born at term with birth weights appropriate for gestational age, and to be exclusively breastfed and receiving no routine vitamin or mineral supplements. Participating infants were randomized to receive either beef or infant cereal first, and mothers committed to continue breastfeeding through at least 7 months. The study was approved by the Colorado Multiple Institutional Review Board, and written informed consent was obtained from the parents of the infants.

Determination of the number of infants needed for the study was based on previous investigations on growth patterns of breastfed infants within the Denver metro area (¹³). One hundred subjects were predicted to be sufficient for a 0.05 significance level on a two tailed t-test to have a power of 0.80 for detecting differences of 0.85 or larger in weight-for-length z scores between feeding groups for the combined sexes.

Diets

According to the randomization protocol, infants were started on either iron-fortified infant cereal or commercially prepared pureed beef for the first complementary food by at least 6 months of age. Exact timing of the initiation of complementary feeding before 6 months was left to parents' discretion. Both the beef (“Beef and Beef Gravy,” Second Foods) and cereal (iron fortified infant rice cereal, without zinc, First Foods) were provided for the study by Gerber Products Company (Fremont, MI). Generous monthly supplies of the assigned complementary food were provided to the families at no cost. The infants fed cereal as the first complementary food deferred introduction of meats and other high protein foods until after 7 months when the intervention finished, and similarly, the infants randomized to receive beef deferred infant cereal until after 7 months. From 7 through 12 months, all study restrictions related to complementary foods were lifted, and choices were entirely at parents’ discretion. Both groups were instructed to begin additional complementary foods based on standard pediatric practice (e.g., fruits, vegetables, and teething biscuits) after 1 month on the cereal or beef. Written guidelines were given to parents regarding types of foods, how to mix, and instructions for recording, but amounts of any foods were not dictated by the research team. The dried rice cereal was mixed with either expressed human milk or water on approximately a 1:2 ratio of cereal to fluid. Use of formula was allowed after 7 months, and intake was monitored by questionnaire and on the diet records. Any infant who was found to be reluctant to accept the assigned first food had the option of mixing the food with selected pureed fruits. Fruits have negligible amounts of zinc and iron and are low in phytate and thus are unlikely to affect zinc absorption.

Zinc content of both of the complementary foods was analyzed in our laboratory, as documented previously (¹²). Results showed good agreement with the manufacturers' data: 25 μg Zn/g for the beef and 15 μg Zn/g for the dry cereal. Iron content according to manufacturer's data was 15 μg Fe/g for the beef, and 740 μg Fe/g for the dry cereal.

Three day diet records for complementary foods were completed twice monthly after introduction of complementary foods and through the 7 month visit, including one record mid-month and one record immediately before the next visit. At each visit, the diet records were reviewed with the parents by the research nutritionists in the Pediatric General Clinical Research Center (PGCRC). Mean daily intakes of energy, protein, zinc, and iron were calculated by Nutritionist IV (First Data Bank, San Bruno, CA). With each diet record, parents also recorded the infants’ reaction to the assigned weaning food, using a scale of 1 to 5, with 1 indicating infant refused the food and 5 indicating ready acceptance. Parents’ perception of the infants’ tolerance to the assigned food was also reviewed at each monthly visit.

Anthropometry

Anthropometric measurements including weight, length, and head circumference were measured upon entry into the study at 4 months, monthly (±1 wk) until 7 months, and at 9 months and 12 months. An electronic digital balance (Sartorius Corp, Bohemia, NY), which integrates 100 rapid serial measurements to provide a mean weight to the nearest gram, was used to obtain weight. Lengths were measured by an infant stadiometer, which was accurate to 0.1 cm (Holtain, Ltd, Crosswell, Crymych, Pembs, UK). A laminated tape measure was used to measure head circumference. All measurements were performed in duplicate by the PGCRC research nurse who has been trained in and had extensive experience with infant anthropometry.

Developmental Testing

The Bayley Scales of Infant Development (BSID-II) (¹⁴), Mental and Motor Scales and Infant Behavior Record, were administered at 4, 7, and 12 months. The Mental and Motor scales are standardized norm referred scales appropriate for the assessment of developmental status of infants and toddlers aged birth to 30 months of age. Assessments were performed in an outpatient ambulatory clinic in a room specifically outfitted for developmental assessment of infants and toddlers. The BSID testing was performed by a physical therapist who was completing a doctoral degree in early childhood development and who had received specific training in administration of the scales to children showing both typical and atypical development. Throughout the testing, the investigator administering the BSID was blinded to the infants’ treatment group assignment.

Biochemical Testing

At 9 months of age, blood was obtained by venipuncture for hemoglobin, hematocrit, serum ferritin, somatomedin, and plasma zinc analyses. As noted above, the rationale for obtaining biochemical tests at this age included the opportunity to provide the primary care provider with information about iron status and avoid the usual clinical finger stick at this age. In addition, the risk of iron deficiency is progressively greater in breastfed infants after 6 months if they do not receive adequate iron from nonbreast milk sources. Thus, the desirability of obtaining biochemical measurements at 7 months, coincident with the end of the study intervention, was balanced against the low risk of iron (or zinc) deficiencies at 7 months and the probability that the 9 month check by the primary care physician would still be obtained. Somatomedin concentrations were obtained as a potential nonspecific marker of response to protein and zinc status (¹⁵). Plasma zinc concentrations were determined by atomic absorption spectrophotometry (¹⁶). Hemoglobin concentration and hematocrit were determined in the clinical laboratory at The University of Colorado Hospital by Coulter Maxm Coulter Counter (Beckman Coulter, Inc., Brea, CA). Ferritin analyses were performed at The Children's Hospital clinical laboratory by two site enzyme-linked immunoadsorbent assay using Opus ferritin test system, with magnum cartridges (Dade Behring, Inc., Deerfield, IL). Somatomedin C analyses were performed in the PGCRC core laboratory.

Data Management/Data Analysis

Data were stored in a Microsoft Access program and in paper medical charts, which were maintained in locked file cabinets on the PGCRC. All statistical analyses were performed by SAS, V8 (SAS Institute, Inc., Cary, NC). Means and standard deviations were obtained for all dietary data, growth parameters, biochemical indices, and BSID scores between the two groups at all time periods when data were collected. Weight, length, and head circumference measurements were converted to percentiles and z scores for age (based on exact date of visit) by the CDC Anthropometric Softward Package (Division of Nutrition, Center for Health Promotion and Education, Centers for Disease Control and Prevention, Atlanta, GA).

Tests of Treatment

The goal of the analysis was to estimate differences between the meat and cereal groups in rates of growth over time of the three outcomes weight, length, and head circumference. Analyses were carried out using a two-stage linear mixed model (¹⁷). All mixed models included a fixed effect intercept, a fixed effect for time, representing the population mean rate of change in outcome, and random effects for subject-specific intercepts and slopes. Each model also included a fixed effect for food group as well as an interaction between food group and time. The interaction estimates and tests the difference between groups in the rate of change of the outcome. This analysis is similar to first regressing each subject's outcome on time to obtain a slope (rate of change of outcome), then testing for differences in slopes between the groups, but is preferred in the presence of missing data. Results were similar using t-tests to compare the slopes. Analyses were adjusted for sex, birth weight, and gestation age. Outcomes were analyzed separately, and for each outcome, separate analyses were carried out for months 4 to 7 and months 7 to 12.

Subgroup Growth Analyses

As a sensitivity check, estimating and testing growth rates within sexes was accomplished by modelling the sexes separately using the same methods outlined above. This was also checked with the triple interaction of weaning food × visit × sex. To test whether growth rates differed according to birth weights, the triple interaction of weaning food × visit × birth weight was tested, both as a continuous variable and as a discrete variable whose categories were determined by the quartiles of birth weight.

Nutrient Analyses

To assess the effects of zinc, iron, and protein intakes on growth rate, mixed models were again used. Because of the variable timing of introduction of complementary foods between 5 and 7 months, we used the data from visit 7 only, by which time all subjects were taking complementary foods. These data were used as predictors in growth models for weight, length, and head circumference, for visits the visits from 7 through 12 months (the interval after completion of the complementary food intervention). A multivariate model using each of the above nutrients (simultaneously) was constructed. In addition, univariate models with only one of the hypothesis predictors each were used. Each model was adjusted for the possible confounders of gestational age, birth weight, and sex.

RESULTS

Forty-six infants (23 male, 23 female) were randomized to start meat as the first complementary food, and 42 (20 male, 22 female) were assigned cereal. There were no significant differences in any of the anthropometric variables at the start of the study at 4 months between those in the meat or cereal groups. Mean (±SD) reported birth weights were 3.36 ± 0.51 kg and 3.38 ± 0.37 kg for the meat and cereal groups, respectively.

The scores for tolerance of the assigned weaning food were not different between groups and indicated good tolerance. The acceptance scores were 3.9 ± 1.1 and 3.9 ± 1.0 (mean ± SD) for the cereal and beef groups, respectively.

Only 22 infants (7 meat, 15 cereal) had started complementary foods at the 5 month visit. Mean intakes of energy, protein, zinc, and iron from complementary foods at 5, 7, 9, and 12 months for the two groups are shown in Table 1. At the 5 and 7 month visits, protein and zinc intakes were significantly higher in the meat group, whereas iron intake was significantly higher in the cereal group at 7 months. Energy and nutrient intakes relative to body weight at 7 months are presented in Figure 1. After 7 months, at which time the intervention was lifted, the intakes of all three nutrients converged, and there were no differences in energy or micronutrient intakes between groups, except protein intake was still somewhat higher in the meat group at 9 months (P = 0.04). Energy intakes from complementary foods did not differ between groups at any visit.

TABLE 1: Mean (±SEM) dietary intakes from complementary foods* by group at four follow-up visits

Formula use was minimal before 7 months, with one subject consuming less than 4 oz/d at 4 mo, no infants at 5 months, and 1 to 3 oz/d for two subjects at 6 months (1 in each group). By 9 and 12 months, however, nearly half of the subjects (14 infants in each group) were consuming formula, averaging 8 to 9 oz/d at both time points, which contributed an average of an additional approximately 1.8 mg zinc intake per day and 2.9 to 3.2 mg iron intake per day for these subjects based on standard fortification levels of these two nutrients in infant formulas.

Comparisons of growth rates for the two groups are shown in Table 2 for the intervals 4 to 7 months, and 7 to 12 months. Mean weight-for-length percentiles also did not differ between groups for either interval. Adjusting for birth weight, gestational age, and sex, the only significant difference between food groups was a greater rate of growth in head circumference for the meat group over the 7 to 12 month interval (P = 0.02). By subgroup analyses, neither birth weight nor sex was a significant predictor of growth (ponderal or linear) for either group. When all of the nutrients were modeled together to examine association with growth, none were shown to be significant predictors of any of the anthropometric variables. Univariate models indicated that protein and zinc intakes at the 7 month visit were predictors of head circumference growth from 7 to 12 months (P = 0.01 and 0.007 for protein and zinc, respectively).

TABLE 2: Comparison of growth rates for the groups according to assigned complementary food; model included birth weight, gestational age, and sex

Biochemical data obtained at 9 months are shown in Table 3. Iron and zinc status did not differ between the two groups, with the means for all variables within normal limits. Several infants had plasma zinc concentrations less than 60 μg/dL: five (16%) in the meat group and seven (22%) in the cereal group. An additional 11 infants in the meat group, and 8 in the cereal group had zinc concentrations between 60 and 70 μg/dL, the latter cutoff sometimes used as indication of marginal zinc status. Serum ferritin concentrations were low (<12 mg/L) in 14 (40%) of the infants in the meat group and 10 (30%) in the cereal group. A few subjects were mildly anemic, defined as hemoglobin less than 11.5 g/dL adjusted for Denver's altitude (⁹): eight (23%) in the meat group and five (15%) in the cereal group. Nonparametric tests indicated no differences according to group in the rates of low plasma zinc or ferritin concentrations.

TABLE 3: Biochemical variables (mean ± SEM) at 9 months

The BSID scores also did not differ at any of the individual time points or over time between the two groups for either the Mental or Motor subscores, and all means were within the typical range. The Mental percentile subscores at 12 months for the meat and cereal groups, respectively, were 107.2 ± 13.3 and 108.6 ± 11.2. The 12 month Motor percentile subscales were 94.9 ± 13.1 and 91.2 ± 13.7 for the meat and cereal groups, respectively. The mean Behavior index at 12 months was slightly higher in the meat group, 74.8 ± 21.4 versus 65.3 ± 24.6 (P = 0.08).

DISCUSSION

The results of this study indicate the feasibility of introduction of meat as a first complementary food for breastfed infants as a means of providing both iron and zinc, two essential micronutrients that often become marginal in the older breastfed infant's diet. On the basis of parental reporting, there were no adverse effects, and the acceptability of the two foods was comparable. The similarity of the energy intakes from complementary foods for the two groups at the 5 and 7 month visits further supports acceptance of both of the assigned foods.

The differences in nutrient intakes were quite striking, however, and emphasize the very modest intake of zinc from cereals, fruits, and vegetables, which represent the typical choices of complementary foods before 7 months in this country (¹¹). The group assigned to receive meat had average zinc intakes from complementary foods at 7 and 9 months that were 76% and 104%, respectively, of the estimated average requirement (EAR) of the Food and Nutrition Board of 2.5 mg/d for 7- to 12-month-old infants (¹⁸). Assuming an additional intake of 0.5 mg/d of zinc from human milk at 7 months (¹³), the total daily intake would have been greater than 90% of the EAR at 7 months. For reference, a single small jar of the strained beef provided in the study provides approximately 1.8 mg Zn. On a body weight basis (Fig. 1), at both ages, mean zinc intakes from complementary foods of the infants in the meat group were also very close to the EAR of 0.25 mg/kg: 0.23 and 0.24 mg/kg at 7 and 9 months, respectively (¹⁸).

FIG. 1: Comparison of energy and nutrient intakes relative to body weight according to complementary food group at 7 months (meat = solid bars; cereal = hatched bars). *Significant differences in intake between groups.

In contrast, the infants randomized to cereal had mean zinc intakes from complementary foods at 7 months that were approximately 30% (0.08 mg/kg) of the EAR. Even with addition of the expected intake of approximately 0.5 mg Zn/d from human milk at this age (¹³), the cereal group's mean intake would still have been less than 50% of the EAR. The mean zinc intake observed in this group is nearly identical to that of a slightly larger group of breastfed infants in Denver for whom the intakes of complementary foods were determined by parental choice (¹³). Thus, we conclude that the intake in the current study can be considered typical of current intake from nonzinc fortified infant cereals. Since the completion of this study, zinc has been added to some, but not all, commercial infant cereals. There are no published data, however, on the bioavailability of zinc from these zinc fortified products. Although by the 9 month visit, the mean zinc intake had increased for the cereal group and was not different from those assigned to meat, the means from complementary foods for both groups are nearly equal to the EAR, indicating still a relatively modest intake and less than 75% of the Recommended Dietary Allowance for this age (¹⁸).

As noted earlier, the most commonly reported sources of protein from complementary foods for 7 to 11 month olds are mixed dishes, which include a vegetable or rice or pasta plus some meat (¹¹). These mixtures generally provide a small percentage (approximately 15%) of the zinc and a modest percentage (approximately 30%) of the iron compared with single food pureed meats (¹⁹). Thus, it is quite possible that the low intakes of zinc reported for the cereal group may well persist for older infants who do not consume meats or nonzinc-fortified infant cereals.

Biochemical data indicated a surprising percentage of infants with low plasma zinc and serum ferritin concentrations. To our knowledge, these are the first data to show this high percentage of breastfed infants in the United States with zinc status in the mild to moderately deficient range. This finding is perhaps more surprising in view of the formula intake by approximately half of the subjects. Hotz et al. (²⁰) recently reviewed confounding factors for interpretation of plasma zinc concentration, and suggested cutoffs for serum zinc concentrations based on the data from the National Health and Nutrition Examination Survey (NHANES II). In the NHANES II data set, however, data were not available for infants and children less than 3 years old. The authors recommended use of a cut-off of 65 μg/dL for the 0- to 5-year-old group (²⁰). If that cut-off is used, 36% of the infants in our study had low plasma zinc concentrations.

The limitations of plasma zinc as a biomarker of zinc status are well recognized because it is not very sensitive for mild zinc deficiency. However, in a meta-analysis of zinc supplementation trials in infants and young children, Brown et al. (²¹) found serum zinc concentration to be a significant predictor of the magnitude of weight gain after zinc supplementation and thus concluded that it is a useful biomarker of zinc status. Individuals with normal plasma zinc concentrations may have a growth response to a trial of zinc supplementation, which is still considered to be the gold standard for identification of zinc deficiency. Plasma zinc concentrations can be reduced by acute infections (²²) and by recent meals (²³). We did not obtain a concurrent index of infection or inflammation, such as an erythrocyte sedimentation rate or C-reactive protein, but in this generally healthy population, infection or inflammation appears unlikely to have been a major contributing factor to the observed results. Because infants eat frequently, it is difficult to estimate the potential impact of nonfasting status on the results.

Additional evidence of marginal zinc status in our population was provided by the results of the zinc absorption study in a subset of infants (¹²). The size of the exchangeable zinc pool (EZP), which is thought to represent metabolically active zinc, was measured by stable isotope methodology at 7 months. The mean size of EZP did not differ by feeding group but was lower relative to body weight in this subgroup of infants (¹²) than for any other group of infants that we have studied (²⁴). Furthermore, the size of the EZP was significantly positively correlated with zinc intake from complementary foods (¹²). Although the EZP has not been extensively evaluated specifically as an index of zinc status, studies in adults have demonstrated that the size of this rapidly turning over pool decreases under controlled conditions of zinc depletion (^25,26) and increases with zinc supplementation (²⁷). Thus, until additional data are available from appropriately designed studies, we conclude that these data suggest that a relatively high proportion of the infants in this study had marginal zinc status at 9 months.

Slightly more than a third of the infants had low ferritin concentrations at 9 months, consistent with low iron stores. Although this finding was not associated with evidence of frank iron deficiency, these results emphasize the importance of consumption of bioavailable iron from complementary foods by the older breastfed infant so that development of deficiency, particularly anemia, is prevented (⁴). The intake of iron at 7 months in the meat group was significantly lower than that of the cereal group, but absorption from the heme form of iron would be expected to be much higher than that from the cereal. Some evidence suggests iron absorption is inversely associated with dietary iron intake by 9 months (²⁸), but the extent to which this regulation can prevent iron deficiency from developing in the older infant with low intake remains unclear.

There was no difference in weight gain or linear growth between the two groups, and means were well within the normal distribution. For both groups, there were modest declines (<1 standard deviation, data not shown) in weight- and length-for-age z scores from 4 to 12 months, using either the growth charts available at the time this study was conducted or the current growth charts (data not shown), which include somewhat more breastfed infants (^29,30). The meat group had a significantly higher rate of growth of head circumference during the 7 to 12 month period, which was associated with protein and zinc intakes at 7 months. Because the intervention to assigned complementary food was still in place at the beginning of this interval (7 mo), the protein and zinc intakes primarily reflects meat intake. There would also have been other nutrient intake differences between the groups that are not reported here but which could potentially impact development, including especially vitamin B₁₂.

The practical implications of this apparent impact on head growth cannot be determined from these data. Zinc supplementation has been associated with improved psychomotor development in developing countries with use of the BSID (³¹). In addition, a recent observational study in the United Kingdom reported an association between higher (though modest) meat intake and weight gain during the first year of life and psychomotor development at 22 months of age (³²). There was a trend for a difference in the Behavior subscale of the BSID, with the meat group having a higher score at the 12 month testing point (P = 0.08). The Behavior scale generally indicates an infant's state, orientation to the environment, and engagement with people, which may influence the performance on the Motor and Mental subscales. According to a recent review by Black (³³), the motor and behavioral subscales of the current version of the Bayley (BSID II) may be more sensitive to zinc than the mental scale. Furthermore, zinc deficiency may also influence the child's ability to elicit or use “nurturant interactions” from caregivers, by effects on orientation and engagement (exploration, enthusiasm, positive affect) (³³). The findings reported here provide further basis for examining this aspect of infant development in relation to zinc status.

The results of this study may well have been impacted by the study design. Because the intervention (i.e., the assignment of the complementary food group) was stopped at 7 months, and most of the infants did not receive any complementary foods until about 6 months of age, the actual intervention period was very short. In addition, our subjective impression was that once the intervention was lifted, the parents were anxious to use the alternative food, especially those assigned to cereal, many of whom introduced meat shortly after 7 months. This eagerness may have been attributable to the parents’ awareness of our rationale for conducting the study, which was presented during the consent process. Thus, although we had predicted that early exposure to meat might have a carry-over effect and result in better acceptance and intake of the meat after 7 months, this did not appear to be the case. Infants in the cereal group had no apparent reluctance to accept the meat. There was, therefore, convergence of the intakes by the next visit at 9 months. An additional dietary factor that likely influenced outcomes after 7 months was that the “requirement” to continue to exclusively breastfeed was lifted, and approximately half of the women had initiated formula feeding by the 9 month visit. Finally, the study may have had insufficient power to detect differences in the outcomes, including growth, due to the intervention.

The importance of the results of this study with particular reference to zinc are several fold. First, we have demonstrated that intakes of zinc from cereal-based, nonzinc-fortified complementary foods for breastfed infants in the United States are remarkably low at 7 months and do not necessarily adequately increase thereafter. Second, the biochemical data at 9 months suggest that mild to moderate zinc deficiency may not be uncommon in breastfed infants. Third, a modest intake of meat substantially increases zinc intake, in a highly bioavailable form (¹²), and the infants accepted the meat similarly to other new complementary foods. The significantly higher rate of growth in head circumference, coupled with the trend for a difference in the Behavior subscale, also suggest a possible benefit for the meat group.

The obvious question is whether zinc deficiency is actually a problem for the older breastfed infant. In developing countries, growth limiting zinc deficiency has been demonstrated in 6- to 12-month-old breastfed infants who were weaned to plant-based complementary foods (³⁴). Likewise, prevention of zinc deficiency, along with promotion of breastfeeding and optimal choice of complementary foods, has been proposed to be one of the major public health interventions with great potential to reduce infant and childhood mortality on a worldwide basis (⁵). The prevalence of zinc deficiency in the older infant who is breastfed without formula supplementation through at least the first year of life in the United States, as recommended, is much less clear. Current practices for complementary food choices were not developed to address adequacy of zinc intake. The lack of adequately designed studies, the nonspecific effects of zinc deficiency, and the relatively low rates of exclusive breastfeeding past 6 months all contribute to the knowledge gap for this population.

The national health objective in the United States to increase rates of breastfeeding through at least 12 months to 25% is based on recognition of the associated health benefits of extended breastfeeding (^35-37). Successful attainment of these benefits, however, is predicated on the recognition that appropriate complementary foods will be necessary to meet nutrient requirements for the older infant. The results of this study raise the question of whether current complementary feeding practices in the United States ideally meet that standard.

Acknowledgments

The authors wish to note their appreciation for the contribution of the infant foods by Gerber Products Company (Fremont, MI) and of loaned electric breast pumps by Medela, Inc. (McHenry, IL) and Ameda-Egnell, Inc. (Libertyville, IL). They especially thank the mothers and families of the infants who participated in this study; without their dedication, the study would not have been possible. They also acknowledge the contributions to the study of Drs. Gary Grunwald, Donna Beshgetoor, Susan Hall, Deborah Hamilton Voss, and Sanju Jalla.