Table 3. Multivariate linear regression results on outcome variable BMI (kg/m²) with triclosan and bisphenol A (BPA) categorized by quartile.

		Model 1		Model 2
Parameter	Category	Estimate	p- value	Estimate	p- value
Intercept	-	25.29 (23.71, 26.87)	<0.001	25.92 (22.14, 25.69)	<0.001
Triclosan/creatinine	High	0.26 (−0.27, 0.80)	0.33	0.19 (−0.32, 0.71)	0.47
	Medium	1.04 (0.41, 1.68)	0.002	0.95 (0.32, 1.57)	<0.001
	Low	1.53 (0.85, 2.21)	<0.001	1.37 (0.74, 2.01)	<0.001
	Below detection	0	ref.	0	ref.
NHANES	2007–2008	0.76 (0.09, 1.44)	0.03	0.73 (0.10, 1.37)	0.03
	2005–2006	0.61 (−0.17, 1.39)	0.12	0.60 (−0.15, 1.34)	0.12
	2003–2004	0	ref.	0	ref.
Cotinine	High	−0.37 (−1.26, 0.51)	0.40	−0.44 (−1.28, 0.41)	0.32
	Low	1.18 (0.46, 1.91)	0.002	1.18 (0.46, 1.91)	0.003
	Below detection	0	ref.	0	ref.
BPA/creatinine	High	-	-	1.42 (0.59, 2.25)	0.002
	Medium	-	-	1.45 (0.71, 2.20)	<0.001
	Low	-	-	1.79 (0.99, 2.58)	<0.001
	Below detection	-	-	0	ref.
Age (years)	-	0.02 (0.009, 0.04)	0.002	0.02 (0.01, 0.04)	<0.001

Model 1, without BPA; Model 2,with BPA. Models (1/2) also included sex (p = 0.68/0.62), race (p<0.001/<0.001), socioeconomic status (SES) as measured by poverty index ratio (PIR) (p = 0.16/0.14), sex-race interaction (p<0.001/<0.001), and sex-SES interaction (p = 0.006/0.007).