Table 2.

Associations¹ of grandmothers’ (F0) serum o,p’-DDT with adult F2 Obesity (≥30 v. <30 kg/m²), stratified by grandmothers’ BMI², n=258 triads

	Estimated o,p’-DDT association by grand-maternal (F0) BMI2
	Grandmothers’ (F0) BMI <25 (n=182)			Grandmothers’ (F0) BMI ≥25 (n=76)
Model Adjustment Level	o,p’-DDT Odds Ratio3	(95% CI4)	p-value	o,p’-DDT Odds Ratio3	(95% CI4)	p-value
DDT congeners	3.62	(1.37, 9.56)	0.0092	0.37	(0.13, 1.02)	0.0557
+ F1 BMI5 Adjusted	3.43	(1.47, 8.00)	0.0044	0.36	(0.11, 1.13)	0.0809
+ F1 BMI5 + Race6 Adjusted	2.59	(1.00, 6.73)	0.0505	0.31	(0.09, 1.01)	0.0526

¹Associations are estimated from log-linear models adjusted for family clustering and stratified by F0 BMI (categorized as ≤25 kg/m² vs >25 kg/m²). Unadjusted models include linear terms for p,p’-DDT, p,p’-DDE, o,p’-DDT. Models adjusted for F1 BMI added a linear term for F1 BMI reported at age 30 nearest the age when F1 were pregnant with F2 and nearest the ages when F0 and F2 BMI were captured. Models additionally adjusted for race added a dichotomous variable representing African Americans vs. all other races.

²The rationale to stratify by F0 overweight (BMI <25 vs. >25 kg/m²) is based on results from the model estimating F2 obesity that included linear terms for p,p’-DDT, p,p’-DDE, o,p’-DDT and F0 BMI, and the product term between o,p’-DDT (continuous) X F0 BMI (continuous). The p-value for the product term, P_{interaction (o,p’-DDT X F0 BMI)}= 0.0276.

³Reported o,p’-DDT Odds Ratios are estimated for the median of the 3rd vs. the median of the first tertile. The inter-tertile range for F0 o,p’-DDT at the medians was 0.86 ng/mL.

⁴CI=Confidence Limits

⁵BMI=Body Mass Index, continuous kg/m².

⁶Race is dichotomized as African American (1) vs. all else (0).