The consequences of solid food introduction (usually referred to as the complementary feeding period) on subsequent health outcome are not well documented. This is surprising, given the radical changes in the dietary pattern of infants, which occur over a relatively short period of time. The change in diet is considerable, from a high-fat, low protein, low fiber single food, to a more complex mixture of foods, which is relatively high in protein with a low-to-medium fat and fiber content.
It has been suggested that for the breast-fed infant the inclusion of red meat as a first complementary food would be of particular advantage to ensure an adequate intake of iron and zinc (1). However, meat also contains many other important and essential nutrients including protein and relatively high levels of certain long-chain polyunsaturated fatty acids one of which, arachidonic acid (AA), is found only in foods of animal origin (2). Recently, the role of animal source foods, including meat, for improving infant health in developing countries was highlighted (3), although the complementary feeding period was not specifically addressed.
There is evidence that haem iron, when added to complementary foods for 6-month-old infants, is a highly bioavailable form of iron (4). In another study (5), meat, including fish as well as animal flesh, was shown to prevent a decrease in hemoglobin in 8-month-old infants, possibly by enhancing iron absorption. However, the intervention lasted only 2 months and determination of the effect of meat was confined to hemoglobin, with no data on accompanying growth and cognitive development. We have been unable to find published evidence of a longitudinal nature on the influence of meat consumption on neurocognitive performance and growth in infants up to 24 months of age. We undertook such a prospective trial on 144 healthy full-term infants to investigate the relationship of meat intake with health outcomes. We have tested the hypothesis that the inclusion of red and white meat as a first complementary food would be associated with improved neurocognitive development and growth and that the effects of meat would be greater in the breast-fed infant compared with the formula-fed infant.
PATIENTS AND MATERIALS
One hundred and forty four infants living in Surrey and Hampshire who weighed at least 2500 g at birth and who were not receiving oral iron supplements, long-term treatment or medical care formed the study cohort. Recruitment took place between May 1996 and December 1999. Subjects were recruited before they were 4 months of age and were followed up at 4-monthly intervals until 24 months of age. Parents were unaware of the specific study objectives although they received routine advice on complementary feeding (known as "weaning") from health visitors and were free to choose the type and amount of meat offered to their children (the study team did not influence this choice). At the time of the study, parents in the UK were advised to introduce solids between 4 and 6 months of age; this was later revised to 6 months in 2001 by the World Health Organization (6) and subsequently by the UK Department of Health in 2003 (7). Approval for the study was obtained from the Ethics Committees of the regions in which patients lived and from the University of Surrey Committee on Ethics, and parents gave written informed consent. A questionnaire was used to obtain information relating to gender of infant, age of mother and socioeconomic group of the parents. Further details of subject recruitment have been previously published (8).
Seven-day weighed food intake diaries were recorded when the infants were 4, 8, 12, 16, 20 and 24 months of age. To ensure consistency in data recording among the families, dietary scales (Selectronic 2200 scales; Salter Housewares Ltd, Tonbridge, Kent, UK), capacity 2 g-2 kg, and blank diet diaries were provided for this purpose. Mothers/carers were given careful instruction (both verbally and in writing) on the preparation and completion of weighed food intake records. The study nurses, together with the parent or carer, reviewed the diaries to identify any anomalous or potentially erroneous entries. The amount of total red meat (mainly beef, pork and lamb) and white meat (mainly chicken and turkey) in grams recorded as being consumed was estimated for each day and summated for the periods used in the statistical analysis. Data were analyzed as follows:
1. Total red and white meat intake (g) from 4-12 months as a continuous variable, i.e. total meat intake over 21 days between 4 and 12 months.
2. Total red and white meat intake (g) from 4-16 months as a continuous variable, i.e. total meat intake over 28 days between 4 and 16 months.
3. Total red and white meat intake (g) from 4-24 months as a continuous variable, i.e. total meat intake over 42 days between 4 and 24 months.
The period of breastfeeding and the age at which infant formula milk, cows' milk and solid foods were introduced were recorded. For the purposes of data analysis, breastfeeding patterns were examined by two different approaches:
1. Whether the infant was breast-fed (either exclusively or partially) for ≥4 months, or formula-fed from birth.
2. Whether the infant was breast-fed (either exclusively or partially) for ≥6 weeks, or formula-fed from birth.
The first approach represents the more polarized of the two, but results in a smaller sample size because infants breast-fed for less than 4 months are excluded.
Daily intakes of other macro and micro (8) nutrients were also calculated.
Body weight, supine body length and head circumference were measured at the six time points when the diet diaries were completed. Equipment was standardized and calibrated to maintain consistency throughout the study. Infants were weighed naked using a Seca 834 digital scale accurate to 10 g from 0-10 kg and to 20 g from 10-20 kg. Smaller infants were weighed in a supine position; older infants capable of sitting unsupported were weighed while seated in the scales. Length was measured using a Harlow Healthcare rollameter accurate to 1 mm. The infant was measured lying supine on the mat with the top of the head held in contact with the top plate by the parent, while the research nurse straightened the knees and brought the base plate into contact with the heels. Two or three readings were taken when possible and an average value was determined. The occipitofrontal head circumference was measured using a Harlow Healthcare Lasso circumference tape. The measurement from the tape was taken when positioned midway between the eyebrows and hairline at the front of the infant's head and passing over the occipital prominence at the back. Growth data were expressed as SD scores by reference to UK population reference data (9).
Neurodevelopment at age 22 months was determined from the mental and motor scales of the Bayley Scales of Infant Development II (10), from which were derived the psychomotor development indices (PDI) and the mental developmental indices (MDI). Two research nurses, who were accredited for the purpose, performed this assessment, and interobserver and intraobserver variation was checked every 6 months to ensure reliability of the assessment.
Meat intakes over the periods 4-12 months, 4-16 months and 4-24 months were not normally distributed and were therefore transformed before further analyses. The appropriate power for transformation was 0.44 for meat intake between 4 and 12 months and 0.42 for meat intake between 4 and 16 months. Analysis of covariance was used to examine the independent effects of meat intake and breastfeeding on later outcomes (growth or neurodevelopment), to look for interactions between meat intake and breastfeeding on later outcome and to examine the influence of potential confounding factors. To examine the effect of meat intake or breastfeeding on growth, measurements at later ages were adjusted for birth weight.
One hundred and forty four infants were recruited into the study. The gender and social economic background based on the mothers' occupation of the study population together with that of the UK profile (11) is given in Table 1. The gender mix was similar to that of the national profile. However, the social profile of the study population differed from the national profile in having a higher proportion of mothers in the higher occupation groups and fewer unemployed or unclassified. The majority of families (97%) were of white ethnic origin, the remainder being black Caribbean, Indian and Chinese. Sixty-four percent of the mothers were between 31-40 years, and 99% were married or living with their partner. Of those parents who responded to the question on smoking, 15% and 22% of mothers and fathers, respectively, smoked at the time the questionnaire was completed.
Eighty three percent of the sample received human milk at birth, 37% at 4 months and 2% at 6 months, in line with national rates (11). The use of cows' milk infant formula was widespread so that 89% of the sample were receiving cows' milk formula by 6 months. Whole-pasteurized cows' milk was used as a mixer to cereals or other foods or as a drink by 13.5% at 6 months of age and 88% by 12 months. The early introduction of solid foods was widespread. At 3 months 29% and at 4 months 90% had received solid foods. Of these infants 45.1% and 45% had received red and white meat by 3 or 4 months, respectively.
Meat intakes during the first 12 months were similar in breast-fed and formula-fed infants (median (25th, 75th percentiles): 283 (183, 453) g/21 day period for those formula-fed or breast-fed for less than 6 weeks compared with 265 (174, 370) g/21 day period for those breast-fed for ≥6 weeks). The median intake of meat for all infants (25 and 75 percentiles) was 278 g (range, 178-425 g) over the 4-24 month period of the study as recorded in the six 7-day diet diaries, which equates to approximately 7 g/d of meat. By 8 months of age, 98% of all infants had received some meat. An examination was undertaken of any association between levels of meat intake and social class but no statistically significant relationship was observed.
The contribution of meat to energy and protein intakes, based on the 7-day diet diaries, is shown in Table 2. For clarity, the results for meat intake in Table 2 are grouped by terciles; there were no statistically significant differences in energy and protein intakes between the diet groups at any age. The contribution of meat to trace element intakes (iron, copper and zinc) based on the 7-day diet diaries and examined by terciles of meat intake has been previously reported (8).
A higher meat intake was positively associated with weight gain during the first year (Table 3). The findings were unchanged when weight SD scores were used (data not shown). When the model was adjusted for energy intake, the meat effect remained and both energy intake (P < 0.003) and meat intake (P < 0.04) were independently associated with weight gain. However, when protein intake was added to the model there was an independent association between protein intake and weight gain (P < 0.003), but the meat effect disappeared (P < 0.3).
Dietary iron and zinc intakes were not independently related to weight gain. Meat intake up to 16 months or 24 months was not associated with weight gain during these periods. Length gain and head circumference gain were not related to meat intake over any time period, and there were no significant effects of breastfeeding on subsequent weight, length or head circumference gain.
The mean (SD and range) for the PDI for the whole cohort (n = 131) was 103.9 (10.4, 70-129). A positive and statistically significant association was observed between the PDI and meat intake during 4 to 12 months and 4 to 16 months (Table 4). When the model was adjusted for potential confounding factors (gender, social class, maternal age, parental smoking, infant length at time of developmental assessment) the meat effect remained, and there was also a positive association between breastfeeding for more than 4 months and later PDI (P = 0.054 for 12-month model, P = 0.051 for 16-month model). There was no interaction between breastfeeding and meat intake on PDI. Other factors significantly associated with a higher PDI in this model were female gender, greater infant length, higher maternal age and higher social class. There was no independent association between protein, iron or zinc intake and PDI at 22 months.
The mean (SD and range) for the MDI for the whole cohort (n = 131) was 99.2 (12.60, 59-137). A positive and statistically significant association was observed between the MDI and breastfeeding at both ≥6 weeks and ≥4 months breastfeeding patterns (Table 5). Meat intake was not associated with MDI. When the models were adjusted for confounding factors the breastfeeding effect remained but the P value for the ≥4 month breastfeeding model shifted to P > 0.05 (Table 5). Other factors significantly associated with MDI were gender, maternal smoking and social class with the associations as for PDI.
The design of this study has facilitated examination of associations between meat and subsequent growth, neurologic development and biochemical indices although we emphasize that a randomized controlled trial is required to demonstrate causality. Meat is a nutrient-dense food, in particular providing high quality protein, iron, zinc and the long-chain polyunsaturated fatty acid, AA. We were able to explore the value of meat in increments but not to compare the data with that from non-meat and white meat eaters (the preferred option) because these two groups were small in number and atypical in a number of ways.
The study population can be described as representing the national population in terms of gender and infant feeding practices. This was not true for other confounding factors, the population here being more affluent than the national population; therefore the benefits of meat reported here could be expected to be of a greater magnitude to the population as a whole. We found here, however, no association between levels of meat intake and social class. There was a greater proportion of women working in higher managerial and professional occupations compared with a national sample, and this no doubt contributed to the 100% cooperation rates for 7-day diet diaries and commitment to the study. The achievement of the collection of such detailed longitudinal data on the infants, and in particular the comprehensive dietary data, also reflects the commitment of the study nurses, who maintained close contact with the families throughout the study.
A high meat intake was associated with more rapid weight gain to 12 months but there was no interaction with breastfeeding patterns, suggesting that the effect of meat intake was the same regardless of whether the infant was breast-fed or formula-fed. There was no specific beneficial effect of meat on the growth of the breast-fed infants. A possible explanation for the lack of evidence of a greater beneficial effect of meat in the breast-fed group may be found in an examination of the complementary feeding patterns of these infants. The majority of the 4-month-old infants (whether breast-fed or formula-fed) were already receiving meat in their diets, so any possible period of suboptimal nutrient intake between 4-6 months was not apparent. The practice of early solid food introduction was common in our cohort and reflected national trends at the time (11).
We do not know if the accelerated weight gain was of lean body mass or adipose tissue, as we did not collect information on body composition. In contrast, length and head circumference gains were not influenced by meat intake. These results were confirmed by an examination of the change in SD scores over the same time points. Thus a child eating large amounts of meat had a high weight gain but this was not associated with a concomitant length gain. Furthermore, whether this observation is good or bad in terms of later health is not known, as the data at 24 months showed no such associations. These findings are in accord with the generally held concept that rapid weight gain (growth) in the first year is inevitably influenced by nutritional factors (nurture) but thereafter growth hormone and genetic factors (nature) exert their role. Our regression model suggested that the association between meat intake and weight gain was not explained by intakes of energy, iron or zinc. In contrast, protein intake was positively related to weight gain and displaced the effect of meat intake. This raises the possibility that the observed association between meat intake and weight gain may be explained by protein intake.
The PDI at 22 months "assesses control of gross and fine muscle groups" (12), including walking up and down stairs with help, walking sideways, standing on one foot with help and grasping a pencil using the pads of the fingertips. The PDI score was significantly higher in infants with the highest meat intakes up to 12 and 16 months of age. On average, infants in the cohort were consuming 7 g/d of meat and we can estimate from the statistical model that every 2.3 g/d of extra meat consumed by an infant was associated with a one-point increase in PDI. When the data were adjusted for confounding factors such as social class, gender, maternal age and length the effect was still apparent.
There are a number of possible mechanisms for this effect. A mother's choice to provide meat in her infant's diet may reflect a social/educational advantage (much like that for the provision of breast milk), and in the analysis we may not have identified or measured all relevant social or demographic differences and therefore failed to "adjust out" the advantage. An alternative explanation is that meat contains a factor or factors that promote psychomotor development (for example, iron or AA). There is some evidence that chronic iron deficiency can cause impairment in an infant's performance on both the PDI and MDI of the Bayley Scales of Infant Development (13). However, in our study we found no independent association between iron intake and cognitive outcome. AA is another possible candidate. The role of AA in psychomotor development has been described (14), and it has been suggested that young term infants may be unable to elongate and desaturate adequate amounts of parent linoleic acid and α-linolenic acid to their functional derivatives, AA and decosahexaenoic acid (DHA), respectively (15). Levels of AA are relatively high in meat, beef contains 1.3% AA and pork contains 1.6%, whereas chicken contains only 0.3% and breast milk 0.5% AA. AA is not found in foods of vegetable origin or in cows' milk. Most infant milk formulas were supplemented with long-chain polyunsaturated fatty acids in the late 1990s (i.e., towards the end of the recruitment period for our study), and follow-on infant milk formulas contain no AA supplement. Therefore the major source of exogenous AA for an infant who was no longer predominantly breast-fed was meat.
MDI "assesses memory, problem solving, discrimination, classification, language and social skills" (12). At 22 months the skills assessed include following directions (combing doll's hair, washing face), pointing to three parts of a doll's body, naming a picture from a book and naming an object presented. In contrast to the findings on PDI, MDI was associated with breastfeeding patterns but not meat intake, even when adjusted for confounding factors such as social class, gender and head circumference. This finding is perhaps not surprising given that there is substantial literature on the advantage to cognitive development/MDI scores in breastfeeding compared with formula feeding (16).
Meat intake, via its effect on protein intake, was associated with increased weight gain up to 12 months. Meat intake may positively influence psychomotor development at 22 months, possibly because of an effect of specific nutrients such as iron or long-chain polyunsaturated fatty acids. The longer-term significance of these findings requires further investigation. However, our results highlight the fact that specific components of the complementary diet may affect later outcome. It is important that such effects are understood to advise parents not only when, but what, foods are most appropriate when they are introduced.
Acknowledgments: We acknowledge the important contributions made to this study by Eddie Wallis-Redworth, Eileen Gossage and Dr. BJ Stordy.
1. Wharton BA. Nutrition in infancy.
British Nutrition Foundation briefing paper. London: British Nutrition Foundation, 1997.
2. Gurr M. Fats. In: Garrow J, James WP, Ralph A, eds., Human Nutrition and Dietetics,
10th ed. Edinburgh: Churchill Livingstone, 2000. Pp. 97-120.
3. Gibson RS, Yeudall F, Droast N, Mtitimini BM, Cullinan TR. Experiences of a community based dietary intervention to enhance micronutrient adequacy of diets low in animal source foods and high in phytate: a case study in rural Malawian children. J Nutr
4. Martinez C, Fox T, Eagles J, Fairweather-Tait S. Evaluation of bioavailability in infant weaning
foods fortified with haem concentrate. J Pediatr Gastroenterol Enter Nutr
5. Engelmann MD, Sandstrom B, Michaelsen KF. Meat
intake and iron status in late infancy: an intervention study. J Pediatr Gastroenterol Enter Nutr
6. World Health Organization. Infant and young child nutrition: Geneva: World Health Organization, 2001. P. 54.2.
8. Taylor A, Wallis-Redworth E, Morgan JB. Meat
consumption in children up to 24 months of age: iron, copper and zinc status. Biological Trace Element Research
9. Freeman JV, Cole TJ, Chinn S, Jones PRM, White EM, Preece MA. Cross-sectional stature and weight reference curves for the UK. Arch Dis Child
10. Bayley N. Manual for the Bayley Scales of Infant Development,
2nd ed. San Antonio: Psychological Corporation, 1993.
11. Hamlyn B, Brooker S, Lleinikova K, Wands S. Infant Feeding 2000.
London: TSO, 2002.
12. Birch EE, Garfield S, Hoffman DR, Uauy R, Birch DG. A randomised controlled trial of early dietary supply of long chain polyunsaturated fatty acids and mental development in term infants
. Dev Med Child Neurol
13. Walter T, De Andraca I, Chadud P, Perales CG. Iron deficiency anaemia: effects on infant psychomotor development
14. Willatts P, Forsyth JS, Dimodugno MK, Varma S, Colvin M. The effects of long-chain polyunsaturated fatty acids on infant attention and cognitive behaviour. In: TJ David (Ed.), Major Controversies in Infant Nutrition.
London: Royal Society of Medicine Press; 1996;7-20.
15. Clark KJ, Makrides M, Neumann MA, Gibson RA. Determination of the optimal ratio of linoleic and alpha linolenic acid in infant formulas. Pediatrics
16. Anderson JW, Johnstone BM, Remley DT. Breast feeding and cognitive development: a meta-analysis. Am J Clin Nutr