, 2008). Specific gravity was measured at room temperature with a refractometer (National Instrument Company Inc., Baltimore, MD), which was calibrated with deionized water before each measurement. For comparison with other studies, we also provide general statistics and report on the variability of BPA levels in urine using creatinine-corrected concentrations (μg/g). Creatinine (mg/dL) was measured using a commercially available diagnostic enzyme

method (Vitros CREA slides, Ortho Clinical Diagnostics, Raritan, NJ, USA). We first summarized demographic characteristics for women who provided at least one urine sample. We then calculated descriptive statistics for BPA concentrations at each prenatal visit. BPA concentrations were log-normally distributed, therefore, we log10-transformed concentrations prior buy Tofacitinib to further analysis. To evaluate the within- and between-woman variability and reproducibility of BPA concentrations (uncorrected and corrected for specific gravity and creatinine) and specific gravity in urine samples for women who provided both prenatal samples, we calculated the intraclass correlation coefficient

(ICC) using mixed effect models (Rabe-Hesketh and Skrondal, 2012). The ICC is a measure of reproducibility and commonly used to assess the CH5424802 suitability of biomarkers to properly characterize exposure. An ICC > 0.75 indicates excellent reproducibility, an ICC value between 0.4 and 0.75 indicates fair to good reproducibility, and an ICC of < 0.4 indicates poor reproducibility (Rosner, 2006). Thus, low ICC values indicate great within-person variability and that more samples per person are needed to properly characterize exposure. Previous studies have reported that sample collection time, independent of other factors, may influence urinary BPA concentrations (Calafat et al., 2005 and Mahalingaiah et al., 2008). To test this in our study participants, we used generalized estimating from equation (GEE) models (Jewell and Hubbard, 2009) using log10-transformed urinary BPA concentrations (uncorrected and specific gravity-corrected) as

the dependent variable and sample collection time as the independent variable; sample collection time was assessed as a continuous (military time) variable. Because consumption of processed/packaged foods may be a significant source of BPA, we also assessed collection time as a categorical variable in separate GEE models; collection time categories were based on potential meal times and included: 8:00 am to 11:59 am (assumed to be after breakfast, but before lunch), 12:00 pm to 1:59 pm (could be before or after lunch), 2:00 pm to 5:59 pm (assumed to be after lunch, but before dinner), and 6:00 pm to 8:30 pm (assumed before or after dinner). GEE models were conducted since they provide robust standard errors and take into account the non-independence of repeated urine samples collected from the same individual.