Meat as complementary food for older breastfed infants and toddlers: A randomized, controlled trial in rural China

Abstract

Background

Because of its contribution to dietary diversity and to favorable intakes of micronutrients, including iron and zinc, meat is hypothesized to be a valuable complementary food for the infant and young child. However, the evidence base remains limited.

Objective

To compare the difference in anthropometric measurements of rural Chinese infants and toddlers 6 to 18 months of age who received a daily supplement of meat or cereal for 12 months.

Methods

This cluster-randomized, controlled study provided a daily supplement of either meat (n = 514, 20 clusters) or cereal (n = 957, 40 clusters) starting as a first complementary food at 6 months of age. Anthropometric measurements were assessed longitudinally.

Results

After 12 months of intervention, the meat group (Δ13.01 ± 1.9 cm) had greater (p = .01) linear growth than the cereal group (Δ12.75 ± 1.8 cm) and a smaller decrease in length-for-age z-score (LAZ) over time (−0.43 ± 0.72 in the meat group vs. −0.54 ± 0.67 in the cereal group), after adjustment for baseline length, LAZ, maternal education, work status, and maternal height and weight.

Conclusions

Linear growth was modestly greater in the meat group than in the cereal group. LAZ was substantially negative at 6 months, and the intervention did not prevent ongoing decline over the course of the study.

Introduction

Meat is recommended as a good complementary food for infants and toddlers by the World Health Organization [1, 2] because it is a major source of bioavailable zinc and iron [3–6] and contains vitamin B₁₂ and complete protein. This recommendation carries particular relevance for low-resource populations, where there is a high rate of stunting due at least partially to dietary factors. Consuming meat also increases dietary diversity. The importance of dietary diversity sufficient to meet all micronutrient needs has received appropriate emphasis in recent years [7]. In addition, compared with fortified foods, meat may be more readily available in low-resource populations to meet both macro- and micronutrient needs and achieve optimal dietary diversity, and it may be especially available to rural populations. Pilot data from four diverse, low-resource settings found that meat was fed to 60% of toddlers but to only approximately 25% of infants who started complementary feeding [8]. A much smaller percentage of infants or toddlers consumed fortified foods in these settings [8].

Although the theoretical benefits of consuming meat are well documented, evidence from the current literature is still quite limited on how meat, as a complementary food, affects growth and micronutrient status in infants and toddlers. This article briefly reviews a randomized complementary feeding trial that evaluated growth and micronutrient status in subjects consuming meat vs. cereal in rural China, a low-resource environment with a moderately high stunting rate.

Methods

Study design

A total of 1,471 6-month-old infants were randomized to one of three isocaloric complementary feeding regimens for 12 months: meat, multiple-micronutrient-fortified cereal, or locally produced, nonfortified cereal. The study included 60 administrative villages (clustered) in nine domains in Xichou County, Yunnan Province, China, a rural community with a 30% rate of stunting. The participants were examined by the assessment team at the nearest hospitals at baseline and at 3-month intervals. Linear growth and micronutrient status were the two main outcomes of the study. The two cereal groups were essentially indistinguishable in terms of macro- and micronutrient contents and were therefore combined to assess linear growth. The study received ethical approval from the Shanghai Jiao Tong University in China and the Colorado Multiple Institutional Review Board prior to initiation.

Participants

Infants were identified at approximately 3 months of age by their community doctors. The inclusion criteria were absence of acute or chronic illness, term delivery without serious neonatal complications, and being exclusively breastfed. With informed parental consent, these participants were subsequently enrolled by the deputy head of Xichou Maternal Child Health Services. Data obtained at enrollment included years of maternal schooling and maternal anthropometric measurements obtained at 6 months postdelivery.

Interventions

Fresh, certified safe pork was purchased weekly and minced and accurately weighed into daily 60-g (2-oz) aliquots in individual plastic bags and stored frozen until transported weekly to the district hospitals serving the meat clusters. The community doctors serving the villages receiving the meat regimen collected the required number of 60-g aliquots of meat weekly from the district hospital. The meat was stored frozen by the community doctors in their own facilities, and 60-g servings were distributed every other day to the participants’ homes. The meat was cooked (boiled) by the parents and fed to the infant. If the day’s supply of cooked meat was not completely consumed in one serving, the mother kept the remainder for a maximum of 2 hours at room temperature and fed it to the infant again. The cereal group received either a commercially available product or an isocaloric packaged pressed rice cereal product that was locally produced for this study. Supplies of these control foods were provided weekly to the six hospitals serving all participating villages, from which they were also collected and distributed by the participating community doctors weekly for weighing of leftovers.

Measurements, data and sample collection

Compliance with feeding was checked by the community doctors through weekly histories and by observation of feeding of the supplement. Written reports of compliance and morbidity by the community doctors were collected weekly at the district hospitals and were reviewed with the community doctors by the trained hospital directors and staff.

Anthropometry

An assessment team, composed of the Maternal and Child Health Services team and of highly trained graduate students from Xinhua Hospital affiliated with Jiao Tong University, performed all anthropometric measurements, including length, weight, and head circumference. The same team traveled to each of the six collaborating district hospitals monthly for collection of the longitudinal measurements and data. Length was measured following standard techniques using a portable Seca infantometer (0.1 cm precision) and a Seca electronic scale (5 g precision). If the two length measurements differed by more than 0.4 cm, a third measurement was taken. The mean of the two closest length measurements was recorded. The measurements were performed for each participant at 6, 7, 9, 12, 15, and 18 months of age, and the mothers were weighed at 6 months postpartum.

Dietary intake

Quantitative 24-hour dietary recalls provided by the child’s parents were obtained in a subgroup at 18 months by the assessment team and reviewed the same week by a nutrition-trained member of the Xichou staff, and were converted subsequently in Shanghai into estimates of nutrient intake using the Nutrition Data System for Research (Nutrition Coordinating Center, University of Minnesota, Minneapolis).

Data processing and analysis

The data were analyzed with SAS, version 9.3. Values are presented as means ± SD. One-way ANCOVA was used to test the effect of treatment group on the primary outcome, change in length from 6 to 18 months. Student’s t-test and the chi-square test were used to compare the two groups at baseline. Significant covariates, such as maternal weight, education, and height and infant length, were selected from baseline characteristics that were different between groups. Other outcome measures, including weight, head circumference, body mass index (BMI), and World Health Organization (WHO) z-scores were also tested using the same model. Dietary intakes at 18 months were tested for group effects by one-way ANOVA. Tukey-Kramer multiple comparison was used for post hoc analysis if the model was significant. Equal variance was checked by Levene’s test. Nonparametric testing was conducted when samples were not normally distributed. Repeated-measure ANOVA was used to test the effects of time and treatment group.

Results

A total of 1,471 infants completed the study, including 514 in the meat group and 957 in the two cereal groups combined. Baseline maternal characteristics and infant anthropometric measurements are presented in table 1. Observed differences between the meat and cereal groups were included in the subsequent analyses as covariates.

TABLE 1.

Baseline (6 months) characteristics of study population according to intervention group

Characteristic	Meat group (n = 514)	Cereal group (n = 957)	pa
Maternal characteristics
≥ 9 yr education—no. (%)	324 (63)	482 (49)	< .001
Paid work—no. (%)	57 (11)	41 (4)	< .001
Height (cm)—mean ± SD	154 ± 6	153 ± 6	< .001
Weight (kg)—mean ± SD	54 ± 7	52 ± 6	< .001
Infant anthropometry
Length (cm)—mean ± SD	64.7 ± 2.4	64.4 ± 2.3	.02
Weight (kg)—mean ± SD	7.3 ± 0.9	7.3 ± 0.9	.09
HC (cm)—mean ± SD	42.6 ± 1.3	42.6 ± 1.2	.80
LAZ—mean ± SD	−0.89 ± 0.97	−1.02 ± 0.99	.02
WAZ—mean ± SD	−0.38 ± 0.99	−0.47 ± 0.99	.09
WLZ—mean ± SD	0.31 ± 0.99	0.30 ± 0.95	.80
HCZ—mean ± SD	−0.11 ± 0.90	−0.10 ± 0.88	.94

HC, head circumference; HCZ, head circumference z-score; LAZ, length-for-age z-score; WAZ, weight-for-age z-score; WLZ, weight-for-length z-score

^aStudent’s t-test was used to compare the two groups at baseline.

The results of the primary analysis (change of length over time) and other growth parameters are summarized in table 2. Observed differences between the groups at baseline, including infant length and length-for-age z-score (LAZ) and maternal education, work status, height, and weight, were adjusted between groups for the primary analysis. The overall test for a difference between the groups in change in length from 6 to 18 months was significant, with a greater linear growth in the meat group. Similarly, a smaller decrease in LAZ was observed in the meat group. There was also a trend toward a greater weight gain and a smaller decrease in weight-for-age z-score (WAZ) in the meat group. The results of 24-hour recalls according to treatment group are summarized in table 3.

TABLE 2.

Comparison of changes in anthropometric outcomes from 6 to 18 months (mean ± SD)

Measurement	Meat group (n = 462)	Cereal group (n = 856)	pa
Δ Length (cm)	13.01 ± 1.9	12.75 ± 1.8	.01
Δ Weight (kg)	2.45 ± 0.60	2.38 ± 0.60	.08
Δ HC (cm)	3.80 ± 0.73	3.82 ± 0.65	.66
Δ LAZ	−0.43 ± 0.72	−0.54 ± 0.67	< .01
Δ WAZ	−0.35 ± 0.60	−0.43 ± 0.61	.05
Δ WLZ	−0.47 ± 0.79	−0.50 ± 0.81	.62
Δ HCZ	−0.19 ± 0.56	−0.18 ± 0.50	.87

HC, head circumference; HCZ, head circumference z-score; LAZ, length-for-age z-score; WAZ, weight-for-age z-score; WLZ, weight-for-length z-score

^aOne-way ANCOVA testing the effect of group with baseline length, LAZ, maternal education, work status, and maternal height and weight adjusted.

TABLE 3.

Dietary intakes from 24-hour recalls at 18 months of age according to group (mean ± SD)^a

Variable	Meat group (n = 260)	Cereal group (n = 493)	pb
Energy (kJ)	5,088 ± 1,638	4,862 ± 1,664	.07
Per weight energy (kJ/kg)	525 ± 177	508 ± 181	.21
Fat (g)	27 ± 12	22 ± 12	< .0001
Carbohydrate (g)	210 ± 77	212 ± 78	.67
Available carbohydrate (g)	200 ± 74	204 ± 77	.56
Protein (g)	35 ± 12	29 ± 12	< .0001
Per weight protein (g/kg)	3.6 ± 1.2	3.1 ± 1.2	< .0001
Animal protein (g)	16 ± 9	10 ± 8	< .0001
Vegetable protein (g)	19 ± 8	19 ± 9	.31
Vitamin B12 (μg)	1.2 ± 0.9	1.0 ± 0.8	.04
Calcium (mg)	324 ± 203	291 ± 194	.02
Iron (total) (mg)	8.3 ± 3.4	8.4 ± 3.5	.69
Zinc (total) (mg)	5.2 ± 1.7	4.8 ± 1.8	.003

^aIntakes reported by parents.

^bKruskal-Wallis analyses.

Discussion

Following analysis using the new WHO growth standards, the mean LAZ at 6 months of age was slightly more negative than the mean LAZ for the WHO Global Database including data from national surveys from 54 countries [9]. However, it corresponded closely to the Global Regional mean. The observed decline in LAZ during the following year of intervention was similar to the growth faltering reported globally [9]. At 18 months of age, LAZ was comparable to the global WHO mean; however, it compared favorably with the regional WHO data [9]. Of note, and similar to the results of a similar concurrent study [10], more than 50% of the negative LAZ values for each group at 18 months were accounted for by the already negative LAZ at 6 months. This further emphasizes the importance of the first 6 post-natal months and/or prenatal growth [11] as major contributors to the linear growth retardation that characterizes early postnatal growth in populations living in food-insecure environments. It also indicates that any environmental intervention, including nutrition, commencing even at the beginning of the recognized complementary feeding stage is already challenged by the need for treatment as well as prevention. The short length of these children at the toddler stage would have been much less evident and possibly preventable if there had been no length deficit at the 6-month baseline. A small trial conducted in the United States that included nonstunted breastfed infants showed a significant positive effect on linear growth in infants consuming meat vs. cereal as complementary foods for 4 months [12], suggesting that meat intake may have a more beneficial preventive effect for infants without very early growth faltering.

In contrast to the study in China, the trial referred to above [10], which involved four sites in Guatemala, Pakistan, Democratic Republic of Congo, and Zambia and compared beef vs. multiple-micronutrient-fortified cereal as a first complementary food commencing at 6 months of age, found no difference in linear growth between the meat group and the fortified cereal group. Most notable was the limited impact that even the most effective of these nutrition interventions starting at the beginning of the complementary feeding period had on the major decline in LAZ that occurs between 6 and 18 months of age [9]. WAZ was close to zero at 6 months, with a corresponding high weight-for-length z-score (WLZ). These data did not suggest the likelihood of suboptimal energy intake as a cause of decline in LAZ. This was supported by repeat 24-hour diet records from approximately one-third of participants.

Animal-source foods, especially meat, provide a major contribution to food diversity and also are excellent sources of zinc and iron, both of which are quite low in human milk after the first several months post-partum. This is especially important for communities that lack access to affordable micronutrient-fortified products. Pork is the most widely consumed meat in China. Meat is also an energy-dense food, which is important for infants, given their small gastric volume. Meat was affordable in the market even to poorer households at the beginning of this project. This population was more accustomed than most low-resource populations to feeding meat as an early complementary food, and until recently maternal premastication was widely practiced [13]. Trials in diverse settings, including in the United States [10, 14], confirm ready acceptance of meat by young infants [15].

Conclusions

The effects of supplementing the plant-based complementary foods of this rural Chinese population with meat were insufficient to prevent the characteristic downward trend in linear growth. Nevertheless, in this and similar poor populations worldwide, consumption of meat-based complementary foods reduced the decline of LAZ during late infancy. More attention is clearly needed to the causes of linear growth retardation already evident before the complementary feeding period. Earlier and comprehensive interventions and attention to earlier stages of pre- and postnatal growth are likely necessary.