IMPORTANCE: Prenatal psychosocial stress and nutrition may each program offspring adiposity, an important predictor of lifelong cardiometabolic health. Although increased stress and poor nutrition have been found to co-occur in pregnancy, little is known about their combined longitudinal associations in the offspring.
OBJECTIVE: To investigate whether the associations of the Dietary Inflammatory Index (DII) with offspring adiposity differ by prenatal stress levels and whether these associations change with age.
DESIGN, SETTING, AND PARTICIPANTS: Project Viva, a prospective prebirth cohort study of mother-child dyads in Massachusetts, included singleton children of mothers enrolled between April 1999 and July 2002, with follow-up visits at early childhood, midchildhood, and early adolescence. Data analysis was performed from October 31, 2020, to October 31, 2022.
EXPOSURES: Food frequency-derived DII score in pregnancy was the exposure. Effect modifiers included stress-related measures in pregnancy; depressive symptoms assessed using the Edinburgh Postnatal Depression Scale (EPDS), dichotomized at scores greater than or equal to 13 vs less than 13; and census tract-level social vulnerability (overall Social Vulnerability Index and its 4 main subindices), dichotomized at the 75th percentile.
MAIN OUTCOMES AND MEASURES: Overall adiposity, comprising sex- and age-standardized body mass index (BMI z), sum of subscapular and triceps skinfolds, fat mass index (FMI), and body fat percentage estimated using bioelectrical impedance analysis (BIA) and dual x-ray absorptiometry (DXA); and central adiposity, comprising waist circumference, ratio of subscapular to triceps skinfolds, and DXA-derived trunk FMI.
RESULTS: Among 1060 mother-child dyads, mean (SD) maternal age was 32.6 (4.6) years, and 811 (77%) mothers were non-Hispanic White. Mean (SD) DII score was -2.7 (1.3) units, Social Vulnerability Index level was 38th (27th) percentile, and 8% of mothers had depressive symptoms. Mean (SD) age of the children was 3.3 (0.3) years at the early childhood visit, 7.9 (0.8) years at the midchildhood visit, and 13.2 (0.9) years at the early adolescence visit. In adjusted analyses, children born to mothers in the highest (vs lowest) quartile of DII had slower decrease in BMI z scores (β, 0.03 SD units/y; 95% CI, 0.01-0.05 SD units/y), and faster adiposity gain (eg, BIA total FMI β, 0.11 kg/m2/y; 95% CI, 0.03-0.19 kg/m2/y) over time. Associations of prenatal DII quartiles with childhood adiposity were stronger (eg, BIA total FMI quartile 4 vs quartile 1 change in β, 1.40 kg/m2; 95% CI, 0.21-2.59 kg/m2) among children of mothers with high vs low EPDS scores in pregnancy, although EPDS scores did not modify the change over time. Associations of prenatal DII with adiposity change over time, however, were greater among children whose mothers lived in neighborhoods with a high (BIA percentage body fat: β, 0.55% per year; 95% CI, 0.04%-1.07% per year) vs low (β, 0.13% per year; 95% CI, -0.20 to 0.46% per year), percentage of racial and ethnic minorities, and residents with limited English-language proficiency.
CONCLUSIONS AND RELEVANCE: The findings of this cohort study suggest that it may be useful to simultaneously evaluate prenatal diet and psychosocial stress in women as targets for interventions intended to prevent excess childhood adiposity.
