A longitudinal study of growth and maturation in rural Thailand.

Bailey Sm, Gershoff Sn, McGandy Rb, A Nondasuta, P Tantiwongse, D Suttapreyasri, Jody C Miller, P McCree

PubMed · 1984 · 29 citations · 0 references

Concepts

Abstract

Mixed longitudinal data on growth and maturation are presented for 1,048 infants and children from Thailand's northern Chiang Mai province. Measurements include recumbent length, weight, head, chest and upper arm circumferences, and triceps and subscapular skinfolds. Skeletal maturity is estimated from handwrist radiographs using the Gruelich-Pyle method. Anthropometric data demonstrate marked growth retardation after the first year, with most attained and velocity statistics falling below the 10th NCHS percentile. Subjects also lag between a half and one ct behind urban middle class Thai peers. Skeletal age falls over 2 ct below U.S. standards by the fourth year. This retardation appears to be of caloric origin and is not related to morbidity. These findings are believed to be the first published for infants and children from Southeast Asia to have a true longitudinal component. In Thailand, as elsewhere in Southeast Asia, our understanding of growth retardation is complicated by inadequate data. Of necessity, growth measures usually have been collected cross-sectionally (Chavalittamrong and Vathakanon, 1978a; Khanjanasthiti et al. 1973; Viseshakul et al. 1977; but see Chavalittamrong et al. 1980). Such data, however informative, do not permit accurate assessment of growth velocity except at very large sample sizes. In addition, they include cohort effects and present difficulties for the investigator seeking a direct comparison to longitudinally-derived standards. To our knowledge, the following are the first published growth data from Southeast Asia to contain a true longitudinal component for infants and young children. Collected in northern rural Thailand, these data also are unusual in representing comparatively healthy subjects. These circumstances offer an unusual opportunity to discriminate certain of the multiple environmental factors present in endemic growth retardation. The present paper will concentrate on describing attained growth and growth velocities using anthropometric and radiographic measures. department of Sociology and Anthropology, Tufts University, Medford, MA 02155 2School of Nutrition, Tufts University 3Department of Health, Bangkok, Thailand 4Mahidol University, Bangkok, Thailand Human Biology, September 1984, Vol. 56, No. 3, pp. 530-557. © Wayne State University Press, 1984 This content downloaded from 157.55.39.41 on Tue, 26 Jul 2016 04:40:24 UTC All use subject to http://about.jstor.org/terms 540 Stephen M. Bailey , et al. METHODS AND MATERIALS From 1971 to 1975, over 2,000 children in 29 villages of Thailand's northern Chiang Mai province were the subjects of a study to ascertain the growth effects of rice fortification (with amino acids, thiamin, vitamin A and iron). All village children between the ages of 6 months and 5 years in 1971 were tracked for five years. Children born during the study were tracked for correspondingly shorter periods. Details of the design and outcome have been published (Gershoff et al. 1975; 1977). Basically, the design included villages that were socioeconomically homogeneous, cooperative, and free of marked food shortages. Although common western descriptors of SES, such as income, are inappropriate to this setting, each village typically had a subsistence base of rice agriculture with a mill and one or two small stores for nonlocal goods. At the beginning of the study, no village had electricity and only one was on a major highway. All villages had schools and some type of health care delivery system; day care centers were built in 18 of the villages as part of the design. All villages were within 60 kilometers of Chiang Mai. In addition, over the five year period, essentially the same team collected data in all villages, and one technician took all anthropometric measurements. Compliance was unusually high, over 99 percent, due to systematic followup. To summarize the results, various fortifications did not lead to statistically significant changes in growth. Accordingly, we have consolidated data from the treatment and control groups for the purpose of the present study. Anthropometric measurements included recumbent length (at all ages), weight, circumferences of the head, chest and middle upper arm, and triceps and subscapular skinfolds. These data were collected on all children in January-February and JulyAugust of each year; a subset of children also had monthly measurements over one year. Hand-wrist radiographs were taken during the January examination, as were biochemical and parasitological data. Beginning in July of 1972, morbidity data were collected on each child every 15 days for one to three years. The current study uses anthropometric and radiographic data collected on 1,048 children each January-February period from 1971 to 1975. The mixed longitudinal nature of these data requires some comment. Because all children were measured at the same occasion, rather than on their birthdays, the cohorts were continuous in age distribution for five years. As a result, when data were calculated by age midpoint, the age interval contained children from several different examination rounds. This content downloaded from 157.55.39.41 on Tue, 26 Jul 2016 04:40:24 UTC All use subject to http://about.jstor.org/terms Growth in Thai Children 541 Seven-year olds, for instance, were found in three different rounds, while five-year olds appeared in all rounds. Thus, although each child's data were purely longitudinal over five years, the calculated median values were actually of mixed longitudinal type. This design, unusual in longitudinal studies, has a number of advantages including lack of seasonal effect and error from interpolating growth between fixed age integers (Goldstein, 1979). Finally, data for Figures 1 through 9 were generated using medians calculated in quarter year overlapping increments in order to maximize information from the continuous age distribution. Tabular data were presented for half year increments generated from the same algorithm. All calculations were carried out using the Statistical Program for the Social Sciences (SPSS).