Date of Award


Document Type

Open Access Thesis


Epidemiology and Biostatistics



First Advisor

Jihong Liu


Background: Much conflicting results exist in the association between breastfeeding and infant growth. One of these confusions is related to the temporal sequence between breastfeeding practice and infant growth.

Objective: This study aimed at examining the association and investigating a possible reverse causality between breastfeeding and infant growth.

Method: Infant Feeding Practices Survey II, a national longitudinal database with repeated measurements, following women prenatally and until one year postpartum (N=2914) was used. Mixed linear model assessed the impact of breastfeeding from the 2nd, 4th, 6th and 9th months on infant growth at the 3rd, 5th, 7th and 12th months, respectively. Log-linear model assessing reverse causation used infant growth data from the 3rd, 5th and 7th months and breastfeeding data from the 4th, 6th and 9th months respectively, restricting to infants' breastfed in the prior months or being exclusively breastfed in the first 5 months.

Results: Non-exclusively breastfed infants had a linear increase in mean weight-for-age z-score (WAZ) from the 3rd month (0.10) to the 7th month (0.34) while exclusively breastfed infants had a stable WAZ (0.27-0.24) (p-value for interaction=0.003). Non-breastfed infants had a higher WAZ throughout the first year (3rd month=0.20, 12th month=0.67) than infants who were ever breastfed in the first year (3rd month=0.04, 12th

month=0.29) (p<.0001). Weight-for-length z-score (WLZ) showed similar results (p interaction=0.006). Log-linear model showed a 7% (95% Confidence Interval 1.00, 1.14) higher risk of continuing with exclusive breastfeeding with every unit increase in WAZ.

Conclusion: In earlier months WAZ was better in exclusively breastfed infants. Only WAZ showed some possibility of reverse causality suggesting weight gain as a predictor of continuation of exclusive breastfeeding.

Included in

Epidemiology Commons