Table 4.

Univariate and multivariate predictors of log-transformed 1-OHP and trans, anti-PheT concentrations

Log 1-OHP
Log trans, anti-PheT
Coefficient (95% CI)*PCoefficient (95% CI)*P
Univariate model
    Occupational exposure0.35 (−0.14 - 0.85)0.10.46 (0.04-0.88)0.03
    Urinary cotinine concentration0.04 (−0.02 - 0.10)0.20.04 (−0.02 - 0.09)0.2
Multivariate model
    Occupational exposure0.30 (−0.20 - 0.80)0.20.43 (0.001-0.86)0.05
    Urinary cotinine concentration0.04 (−0.01 - 0.10)0.10.03 (−0.02 - 0.09)0.2
Multivariate model with interaction term
    Exposure, cotinine at 10th percentile0.36 (−0.22 - 0.95)0.20.58 (0.11-1.05)0.02
    Exposure, cotinine at 90th percentile0.11 (−0.93 - 1.16)0.8−0.13 (−1.04 - 0.77)0.8
  • NOTE: Mixed effect models were used to calculate the coefficients (95% CI) of the predictors.

  • * The exposure coefficient is expressed as pre-exposure = 0 and post-exposure = 1. The smoking coefficient is expressed as the change per 100 μg/g creatinine incremental change in cotinine level.

  • Multivariate models with interaction term are adjusted for age.