Infant weight gain and adolescent body mass index: comparison across two British cohorts born in 1946 and 2001.

OBJECTIVE: To investigate how the relationship of infant weight gain with adolescent body mass index (BMI) differs for individuals born during and before the obesity epidemic era.

DESIGN: Data from two British birth cohorts, the 1946 National Survey of Health and Development (NSHD, n=4199) and the 2001 Millennium Cohort Study (MCS, n=9417), were used to estimate and compare associations of infant weight gain between ages 0 and 3 years with adolescent outcomes.

MAIN OUTCOME MEASURES: BMI Z-scores and overweight/obesity at ages 11 and 14 years.

RESULTS: Infant weight gain, in Z-scores, was positively associated with adolescent BMI Z-scores in both cohorts. Non-linearity in the MCS meant that associations were only stronger than in the NSHD when infant weight gain was above -1 Z-score. Using decomposition analysis, between-cohort differences in association accounted for 20%-30% of the differences (secular increases) in BMI Z-scores, although the underlying estimates were not precise with 95% CIs crossing 0. Conversely, between-cohort differences in the distribution of infant weight gain accounted for approximately 9% of the differences (secular increases) in BMI Z-scores, and the underlying estimates were precise with 95% CI not crossing 0. Relative to normal weight gain (change of -0.67 to +0.67 Z-scores between ages 0 and 3 years), very rapid infant weight gain (>1.34), but not rapid weight gain (+0.67 to +1.34), was associated with higher BMI Z-scores more strongly in the MCS (β=0.790; 95% CI 0.717 to 0.862 at age 11 years) than in the NSHD (0.573; 0.466 to 0.681) (p<0.001 for between-cohort difference). The relationship of slow infant weight gain (<-0.67) with lower adolescent BMI was also stronger in the MCS. Very rapid or slow infant weight gain was not, however, more strongly associated with increased risk of adolescent overweight/obesity or thinness, respectively, in the more recently born cohort.

CONCLUSIONS: Greater infant weight gain, at the middle/upper end of the distribution, was more strongly associated with higher adolescent BMI among individuals born during (compared with before) the obesity epidemic. Combined with a secular change towards greater infant weight gain, these results suggest that there are likely to be associated negative consequences for population-level health and well-being in the future, unless effective interventions are developed and implemented.