Traditional linear regression models of childbearing female's body mass index(BMI) have limitations in capturing cross-distribution variation in effects. The objective of this study was to examine the trends of food consumption in a sample of Chinese childbearing female and its association with the BMI distribution.


We used samples from the China Health and Nutrition Surveys in 1989, 1991, 1993, 1997, 2000, 2004, 2006, 2009, 2011 and 2015 with at least two waves of data collection. A total of 6452 women aged 15-49 years were included in the final analysis with observations. Longitudinal quantile regression analyses were conducted to investigate the changes of BMI distribution over time.


The intake of cereals, tubers, starches, vegetables and oil decreased successively from 1989 to 2015. The intake of fruits and meat increased through the same time. BMI tended to increase every year, from 21.4 in 1989 to 23.4 in 2015. The QR analyses showed that oil had negative association with BMI. Vegetables and fruits had negative association with BMI only in the upper percentile. Cereals and tubers had positive association with BMI in the 75th percentile and above.


This longitudinal quantile regression suggests that effects of different dietary factors worked differently across the BMI distribution. More attentions should be paid to the problem of weight increasing in childbearing women in China.