Skip to main content
. 2012 Jun 21;120(11):1538–1543. doi: 10.1289/ehp.1104614

Table 3.

Relative change (95% CI) in urinary paraben concentrations as a function of demographic and anthropometric predictors from a multivariate regression model.

Parameters n Subjects (n samples)a MP PP
Relative change (95% CI)b Relative change (95% CI)b
Sex
Female 405 (2,104) 4.55 (3.73, 5.56) 7.81 (6.00, 10.2)
Male 243 (589) 1 (Reference) 1 (Reference)
Race
African American 27 (87) 3.84 (2.40, 6.13) 3.80 (2.05, 7.06)
Asian 43 (163) 0.99 (0.69, 1.43) 0.93 (0.57, 1.49)
Other 31 (149) 1.00 (0.66, 1.51) 0.74 (0.43, 1.28)
Caucasian 547 (2,294) 1 (Reference) 1 (Reference)
BMI (kg/m2)
Obese (≥ 30) 116 (423) 0.79 (0.61, 1.01) 0.77 (0.55, 1.07)
Overweight (25–29.9) 210 (778) 0.95 (0.77, 1.16) 0.91 (0.69, 1.19)
Normal (18.5–24.9) 316 (1,457) 1 (Reference) 1 (Reference)
Underweight (< 18.5) 6 (35) 1.42 (0.57, 3.57) 1.46 (0.43, 4.92)
an = 648 subjects and 2,693 urine samples (reduced sample size due to missing BMI and SG values). bExponentiated adjusted parameter estimates from multivariate regression models (including sex, race, and BMI) are presented due to natural log transformation of the outcome and can be interpreted as a relative (times) change from the reference category of the predictor variable; each paraben is modeled separately.