Brazilian cohort: crude and adjusted linear regression analysis to investigate the prediction of metabolic risk z-score by birth order and explore the effects of birth weight and catch up growth after 20 months – Exclusion of first-born children with status of only children.

<p>B = regression coefficient for first-borns; ±95CI = 95% Confidence Interval. Significant p values are shown in bold.</p><p>Brazilian Cohort: Analysis was adjusted as follows:</p><p>Model 1: Unadjusted.</p><p>Model 2: Adjusted for family income; maternal education; household wealth score; breastfeeding for at least six months; maternal smoking during pregnancy; maternal weight at the beginning of the pregnancy, maternal height, and subject smoking at 18 years.</p><p>Model 3: Adjusted for model 2+ birth weight z-score.</p><p>Model 4: Adjusted for model 3+ weight gain z-score birth to 20 months.</p><p>The calculation of the metabolic risk z-score is described in the <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0013907#s2" target="_blank">methods</a> section.