Urinary specific gravity measures in the US population: Implications for the adjustment of non-persistent chemical urinary biomarker data

Background: Urinary biomarkers are often corrected for sample dilution using creatinine, which is influenced by sociodemographic factors and certain health conditions. It is unknown whether these factors similarly influence specific gravity. Objectives: To identify predictors of specific gravity and creatinine and compare methods for correcting estimated chemical concentrations for sample dilution using these measures. Methods: We assessed predictors of urinary specific gravity and creatinine among NHANES 2007-2008 participants (n = 7257). We corrected concentrations of mono-n-butyl phthalate (MnBP) for dilution using two methods, each applied to both specific gravity and creatinine: correction using a sample mean of the dilution indicator (i.e., specific gravity or creatinine) and covariate-adjusted standardization. We compared distributions and assessed the agreement of uncorrected or corrected concentrations visually using Bland-Altman plots and statistically by Kendall's Tau a. We stratified all analyses by age category (i.e., 6-19 or 20+ years of age). Results: Gender, race/ethnicity, body mass index, and height were associated with urinary specific gravity and creatinine. Distributions of corrected MnBP concentrations were comparable for both methods and dilution indicators, but agreement between methods was greater for specific gravity. Additionally, specific gravity- and creatinine-corrected MnBP concentrations had slightly greater agreement with each other when corrected using a covariate-adjusted standardization method. Discussion: Specific gravity, like creatinine, is associated with sociodemographic and body composition variables. Accounting for these factors as part of the dilution correction method may be important to minimize bias.