Is a "color effect" demonstrated for hair analysis of carbamazepine?

Psillakis et al. have reported on the concentration of carbamazepine recovered from hair samples collected from patients receiving this anti-seizure medication under medical supervision and determined that their was a high correlation between dose and quantitation of recovered analyte. The analyte was identified by two techniques, FPIA (Abbott TDx) and HPLC and the correlation was high for both procedures. In the literature on hair analysis some have suggested that analyte concentration in hair is critically dependent on hair color. In reporting their data Psillakis et al. reported the hair color of each patient but made no attempt to analyze their results in relation to color. This article performs a secondary analysis of the Psillakis et al. data in order to determine whether there is a hair color effect discernible in the recovery of carbamazepine from hair. Analysis of this data set for both the FPIA and HPLC by one-way analysis of variance fails to identify a color effect at p = .05. Weighting the data for per-patient dosage values fails to discern a color effect. Examination of all possible two-color comparisons also fails to identify a statistically significant effect for any subset of combinations. These data suggest that carbamazepine does not exhibit a color effect when using either FPIA or HPLC assay methods.