Objectives. To evaluate the impact of introducing complementary foods to exclusively breast-fed infants at 4 vs 6 months on growth from 6 to 12 months, and to compare growth patterns of Honduran infants with those of breast-fed infants in the United States.

Design. Randomized intervention trial from 4 to 6 months and longitudinal study of infants from birth to 12 months.

Setting. Low-income communities in San Pedro Sula, Honduras.

Subjects. Primiparous, breast-feeding mothers and their infants (n = 141) recruited from public maternity hospitals.

Intervention. Infants were randomly assigned to exclusive breast-feeding to 6 months, or exclusive breast-feeding with addition of hygienically prepared, nutritionally adequate complementary foods at 4 months, with or without maintenance of baseline breast-feeding frequency. After 6 months, mothers continued to breast-feed and also fed their infants home-prepared foods after receiving instruction in appropriate feeding practices.

Outcome Measures. Infant weight was measured monthly during the first year of life and infant length monthly from 4 to 12 months.

Statistical Analysis. Growth parameters were compared between the Honduran and US cohorts using multiple-regression and repeated-measures analysis of variance. Stepwise multiple regression was used to identify determinants of infant growth.

Results. There were no differences in growth patterns by intervention group. Mean birth weight of the Honduran infants was significantly less than that of a cohort of breast-fed infants in an affluent US population (n = 46) (2889 ± 482 vs 3611 ± 509 g), but the Honduran infants exhibited rapid catch up in weight in the first few months of life, and the cohorts were similar in weight by 3 months. Mean weight and length gain were similar to those of the US cohort from 4 to 9 months but were lower from 9 to 12 months. Mean length for age was significantly less than that of the US cohort from 4 to 12 months; this was attributable to the difference in maternal height (12 cm shorter in Honduras on average). Within the Honduran cohort, growth velocity of low birth weight infants (<2500 g; n = 28) was similar to that of their normal birth weight peers; thus, the former subgroup remained smaller than the latter throughout the first year of life.

Conclusions. In poor populations, when breast-feeding is exclusive for the first 4 to 6 months, continues from 6 to 12 months, and is accompanied by generally adequate complementary foods, faltering in weight does not occur before 9 months among infants born with birth weights of more than 2500 g.