The Bacon Chow study: Genetic analysis of physical growth in assessment of energy–protein malnutrition

Estimates of the prevalence of energy–protein malnutrition almost universally employ physical growth measurements. In this study we focus on this disease and the role of body size of relatives as mediators of responses in individuals to one type of nutrition intervention: supplementation of pregnant and lactating women. In this study, initiated by Dr. Bacon Chow and others in 1967, during gestation of a first infant a mother was untreated, while during the lactation of the first infant and the gestation and lactation of a second infant she was treated with either a calorie supplement or a placebo. Supplement–placebo group differences were sought in sibling and mother–child correlations in growth from birth to 30 months, in order to assess the role of heredity as a mediator of supplement effects. There were 108 pairs of siblings whose mothers had received a high-calorie–high-protein supplement as described above and 105 pairs of siblings whose mothers had received a placebo. Among the latter, sibling correlations for most measurements are statistically significant at birth, and of the same magnitude seen in previous studies (∼0.5), while among supplemented siblings, birth correlations are unusually low and often insignificant. The sibling correlations in Rohrer's index (wt/L3) differed the most between groups (p < 0.01). Group differences in the sibling correlation tended to disappear over the first 2.5 years of life. Correlations between mothers and their second children in subscapular skinfold tended to be higher in the supplemented than in the placebo group, birth to 30 months. In both supplement groups mother–second child correlations for body weight were higher than mother–first child correlations, suggesting the occurrence of secular changes in the environment unconnected with the treatment. The results suggest that: (1) genetic analysis of components of anthropometric variation may be a more sensitive method than the more conventional comparison of group means in detecting supplement effects; and (2) infant relative weight (Rohrer index), particularly the addition of subcutaneous fat, may be more affected by maternal supplementation than growth in weight or length alone.