Search tips
Search criteria 


Logo of bmcphBioMed Centralsearchsubmit a manuscriptregisterthis articleBMC Public Health
BMC Public Health. 2012; 12: 265.
Published online 2012 June 11. doi:  10.1186/1471-2458-12-265
PMCID: PMC3372444

Determinants of linear growth from infancy to school-aged years: a population-based follow-up study in urban Amazonian children



Although linear growth during childhood may be affected by early-life exposures, few studies have examined whether the effects of these exposures linger on during school age, particularly in low- and middle-income countries.


We conducted a population-based longitudinal study of 256 children living in the Brazilian Amazon, aged 0.1 y to 5.5 y in 2003. Data regarding socioeconomic and maternal characteristics, infant feeding practices, morbidities, and birth weight and length were collected at baseline of the study (2003). Child body length/height was measured at baseline and at follow-up visits (in 2007 and 2009). Restricted cubic splines were used to construct average height-for-age Z score (HAZ) growth curves, yielding estimated HAZ differences among exposure categories at ages 0.5 y, 1 y, 2 y, 5 y, 7 y, and 10 y.


At baseline, median age was 2.6 y (interquartile range, 1.4 y–3.8 y), and mean HAZ was −0.53 (standard deviation, 1.15); 10.2% of children were stunted. In multivariable analysis, children in households above the household wealth index median were 0.30 Z taller at age 5 y (P = 0.017), and children whose families owned land were 0.34 Z taller by age 10 y (P = 0.023), when compared with poorer children. Mothers in the highest tertile for height had children whose HAZ were significantly higher compared with those of children from mothers in the lowest height tertile at all ages. Birth weight and length were positively related to linear growth throughout childhood; by age 10 y, children weighing >3500 g at birth were 0.31 Z taller than those weighing 2501 g to 3500 g (P = 0.022) at birth, and children measuring ≥51 cm at birth were 0.51 Z taller than those measuring ≤48 cm (P = 0.005).


Results suggest socioeconomic background is a potentially modifiable predictor of linear growth during the school-aged years. Maternal height and child’s anthropometric characteristics at birth are positively associated with HAZ up until child age 10 y.

Keywords: Children, Linear growth, Height-for-age Z score, School-aged years, Determinants

Articles from BMC Public Health are provided here courtesy of BioMed Central