PEDIATRICS vol. 84 No. 3 September 1 989 475

Growth of Vegetarian Children: The Farm Study

Joan M. OConnell, MHS, Michael J. Dibley, MB BS, Janet Sierra, RN,Barbara Wallace, PhD, James S. Marks, MD, MPH, andRay Yip, MD, MPH

From the Division of Nutrition, Center for Health Promotion and Education, Centers forDisease Control, Atlanta, Georgia, and The Farm, Summertown, Tennessee

ABSTRACT. To examine the effects of a vegetarian dieton child growth, height and weight data of 404 vegetarianchildren aged 4 months to 10 years who lived in a collec-tive community in Tennessee were studied. Height forage, weight for age, and weight for height were comparedwith the US growth reference. Birth weights, infant feed-ing patterns, and parental heights were also evaluated inrelation to growth. Most of the height for age, weight forage, and weight for height (n = 833) were within the 25thand 75th percentiles of the US growth reference. Themean height for age and weight for age, however, wereslightly less than the median of the reference population.For different age groups, the mean height ranged from0.2 to 2.1 cm and the mean weight ranged from 0.1 to 1.1kg less than the reference median. The largest heightdifference was observed at 1 to 3 years of age and may bepartly the result of intrinsic irregularities in the USgrowth reference at those ages. By 10 years of age, chil-then from The Farm averaged 0.7 cm and 1.1 kg less thanthe reference median, representing only 0.1 and 0.3 SDfrom the reference. Thus, these children have adequateattained growth, even though it was modestly less thanthat of the reference population. Pediatrics 1989;84:475-481; vegetarian diet, growth, height, weight, birth weight.

To examine the effect of a vegetarian diet onchild growth, we studied height and weight of agroup of infant and vegetarian children ages 4months to 10 years. The growth ofchildren consum-ing vegetarian diets was of interest because of thenumber of children consuming such diets and thespecial dietary needs of children during periods ofrapid growth. It has been reported that approxi-mately 4 percent of adults in the United Statesconsume vegetarian diets. Worldwide, many ethnic

Received for publication Jun 13, 1988; accepted Oct 11, 1988.Reprint requests to (R.Y.) Centers for Disease Control, Divisionof Nutrition, Mail stop A-41, Atlanta, GA 30333.PEDIATRICS (ISSN 0031 4005). Copyright 1989 by theAmerican Academy of Pediatrics.

groups have followed vegetarian diets for centuries.The reasons people adhere to these diets includereligious, ethical, political, and health beliefs andeconomic constraints.7

Although several investigators have addressedthe issue of the growth of vegetarian children, mostof these studies were based on relatively smallnumbers (50 or fewer) of preschool-aged child-ren.2#{176} Aside from the small numbers, drawingconclusions from these studies is somewhat difficultbecause of the variety of health-related beliefs andlife-styles of the study populations and the dietarydifferences of the study populations. For example,some vegetarians have behaviors concerning vita-mm and mineral supplements and conventionalmedical care that differ from those of the generalpopulation.82022

The types of vegetarian diets consumed by thestudied populations must be considered when inter-preting results of the growth studies. The extent towhich vegetarian diets meet dietary recommenda-tions varies greatly733 because the term vegetarianrefers to a broad dietary group of people who avoidmeat in their diets. It is estimated that 90% of thevegetarians in the United States are lacto-ovo-vegetarians, who include milk and dairy productsin their diets. Vegans are vegetarians who avoidall animal products, including milk and dairy prod-ucts. There exist additional types of vegetariandiets, some of which involve further restrictions ondietary intake. Although it may be more difficult toprovide adequate diets for children with restrictedprotein intake compared with children who haveno restrictions, all necessary nutrients, includingvitamins B12 and D, calcium, iron, and zinc, can beprovided with careful planning.819233

To examine the effect of a vegetarian diet onchild growth, in this study we describe the growthof vegan children who lived in a collective commu-nity in Tennessee. In this community, the children

at Moldova, Republic of:AAP Sponsored on April 12, 2015pediatrics.aappublications.orgDownloaded from

476 VEGETARIAN CHILDREN

followed a similar diet and many received vitaminand mineral supplements. Because growth datawere collected annually, there was available forstudy a relatively large number of measurements ofchildren through ten years of age. In addition, weevaluated birth weights, parental heights, andbreast-feeding patterns in relation to growth.

SUBJECTS

The children lived in a community called TheFarm, located in rural central Tennessee. The Farmwas established in 1971 as a collective communitythat met many of its own needs concerning healthcare, education, housing, and, to some extent, foodproduction. The socioeconomic level of the popu-lation of The Farm was unusual in that the com-munity members were highly educated, but theygenerated a comparatively small income fromsources outside the community. The average familyincome was below the federal poverty level for manyyears but gradually increased over time. The ma-jority of the children were white and had lived onThe Farm since birth.

The Farm community was generally well in-formed regarding issues related to vegetarianism,including complementing different protein sources,for example, grains and legumes and nonanimalsources of vitamins and minerals. Until 1983, thepopulation followed a vegan diet, with soybeansbeing their primary source of protein. Supplementsof vitamins A, D, and B,2 were added to the soymilk produced on The Farm. Nutritional yeast(containing vitamin B,2) and other vitamin andmineral supplements were also used. In the fall of1983, some members of the community introducedeggs and dairy products into their diets.

METHODS

1984 Special Growth Survey

In the spring of 1984, we collected height andweight data of 144 children residing on The Farm.We obtained written parental consent from morethan 95% ofthe families. In addition, a few childrenwere excluded from the study because of ethnic,dietary, and health backgrounds that differed fromthe other children raised there or because they hada chronic illness known to affect their growth.

Body length to the nearest millimeter was deter-mined with a measuring board for children youngerthan 2 years of age; standing height was measuredfor older children. We measured their weight to thenearest 28 g (1 oz) with a Detecto pediatric balancefor children less than 2 years of age and to thenearest 112 g (quarter pound) with a beam balance

for children older than 2 years of age. At the sametime, a subset of the population was measured onseparate days to validate measurement techniques;no systematic errors were found.

The Farms Annual Growth Surveys

In addition to the data collected in the 1984special growth survey, we included growth datafrom the annual growth surveys of The Farm in ouranalyses. Height and weight data from four cross-sectional surveys performed in the years 1980 to1983 were included. ETHOS, a Farm research or-ganization, and The Farms health clinic conductedthe annual surveys because members of The Farmcommunity were interested in the physical growthof their children.

The methods used in the 1980 to 1983 growthsurveys were not subject to the same rigor as the1984 special growth survey. We evaluated severalcharacteristics of the earlier surveys to determinethe reliability of those data and to determine thesuitability of combining the data from The Farmannual surveys with the data from the 1984 surveyfor cross sectional analyses.

To determine the reliability of the data collectedin each of the separate cross-sectional surveys, wecompared height for age, weight for age, and weightfor height data between the 1984 special survey andthe annual growth surveys of The Farm. No signif-icant differences were found. Because the surveyswere conducted at the same time of year, childrenwith multiple measurements in the data set werenot represented more than once in the age groupbreakdowns by year. We found no significant dif-ferences in growth, dietary, or demographic databetween children who had multiple measurementswith those who had only one measurement andbetween children measured in 1984 and childrenmeasured in the earlier Farm surveys. Based onthese analyses, all survey data from 1980 to 1984were combined into one data set.

Dietary, Health, and Demographic Data

To supplement the growth data, we extractedself-reported information from two health surveysconducted by ETHOS on The Farm in 1980 and1983. The information included birth date, dietaryhistory, birth weight, parental height, breast-feed-ing patterns, vitamins and mineral supplementa-tion, and place of birth. We also obtained diseasehistories for each child to determine whether anychildren had disorders of nutritional significancethat might cause their exclusion from the study.

We evaluated the reliability of the data reportedin the 1980 and 1983 health surveys. If data existed

at Moldova, Republic of:AAP Sponsored on April 12, 2015pediatrics.aappublications.orgDownloaded from

TABLE 1. Social, Demographic, and Dietary Data ofThe Farm Children

Female95th150

I

130

120

* Based on available data.t Mean birth weight 3389 g.:1:Mean number of months breast-fed was 12 months. Mean age when solid foods were introduced was 5months.

NCHS

EC.)0.)

Cs

C,)

5th

1 10

100

90

T Mean SD ofI Sample Children

N 380

2 3

ARTICLES 477

for a child from both surveys, the information wascross-checked for consistency of reporting. No sig-nificant differences were found in data reported in1980 and 1983. We collected recent dietary historiesin 1984 to update past survey data.

The dietary data provided in the health surveyswere modified food frequencies by year and in-cluded the mothers diet during pregnancy. Weclassified the children as omnivore (nonvegetarian),lacto-ovovegetarian, or vegan for each year basedon the amount of eggs, dairy products, and meatsreported in their diets. We analyzed growth inrelation to the following diet classifications: themothers diet during pregnancy, the childs dietfrom birth to age 2 years, and the childs diet during1984. We were also interested in whether or not thechild had followed a vegan diet since birth.

For each child, we averaged the heights of themother and father for a mean parental height.3435We then compared The Farm mean with the aver-age of the mean heights of 25-year-old men andwomen from a US reference population.36

Statistical Procedures

We compared weight and height measurementsof the children to the National Center for HealthStatistics/Centers for Disease Control growth ref-erence population by age and sex. A growth per-centile rank was calculated for each set of measure-ments for three growth indicators: height for age,weight for age, and weight for height.37 We regardeda percentile value greater than the 95th percentileor less than the 5th percentile as abnormal growth.For each set of measurements, a standard deviationvalue, Z score, was calculated for each growth in-dicator based on the growth reference; a Z score of+1.0 represented 1 SD greater than the mean of thereference population.tm39 The Z scores were thencompared by age group with the reference median(presumed mean). We compared the mean Z scoresof the growth data after stratifying them by meanparental height, birth weight, sex, and diet. Weused Students t tests and analysis of covariance todetermine the significance of these variables inrelation to growth.

RESULTS

We evaluated anthropometric data and healthdata for 404 white children, who had a total of 833measurements. The mean age at the time of meas-urement (n = 833) was 71.6 months; 7.7% of thechildren were less than 2 years of age; 35.7% were2 to 5 years of age, and 56.7% were 6 to 10 years ofage. In Table 1, additional social, demographic, anddietary information are provided. The mean birth

Item No.(%)of

Children*

Birth dataBorn on The Farm 304 (82)Lived on The Farm by 2 y of age 339 (91)Birth wt

Female

NCHS

40

at

.c0)a)

50th

T Mean SDof.1_Sample Children

3 4

N 380

5 6 7 8 9 10Age (years)

Fig 3. Weight for age of girls from The Farm relative toNational Center for Health Statistics (NCHS)/Centersfor Disease Control percentiles.

Male

NCHS

MaleNCHS

95th

at

Ota)

EC.)a,

.

(I)

40

30

20

50th

25th

5th

T Mean SD ofJ_Sample Children10

80

N 389

678 10

Age (years)Fig 2. Height for age of boys from The Farm relative toNational Center for Health Statistics (NCHS)/Centersfor Disease Control percentiles.

478 VEGETARIAN CHILDREN

weight of The Farm children was 3389 g; the per-centage of low birth weight infants (

42#{149}

Female

NCHS

Male

a 26

NCHS

, 95th

at

ata

. 75th

, 50th

, 5th

100 110 110

ARTICLES 479

tically significant at ages 9 and 10 years. Of TheFarm children, 3% were less than the 5th percentileof the weight for age reference. The Farm weightfor height data were slightly greater than those ofthe reference population for most ages. The differ-ences in weight for height were statistically signif-icant at ages 5 and younger and at age 9 years. OfThe Farm children, 1% were classified as smallweight for height (weight for height < 5th percent-ile) and 3% as large weight for height (weight forheight > 95th percentile).

A positive association existed between mean pa-rental height and the height for age Z scores (P