Precision and sensitivity of aminobiphenyl hemoglobin adduct assays in a long-term population study

Exploratory statistical analysis of aminobiphenyl hemoglobin adduct data obtained in a large-scale population study was performed to assess precision and sensitivity over the 7 years required to conduct the analyses. A time-dependent trend toward higher values was observed that may be attributable to aging of the internal standard used throughout the study. A several-fold improvement in sensitivity from the beginning to end of the study was also noted. Repeated analysis of duplicate blood specimens provided a worst-case estimate of the coefficient of variation to be 0.31, attributable almost entirely to sample preparation rather than instrumental analysis. Substantial variability in calibration curves for the deuterated internal standard (standard deviation was ±15% of the mean) was observed. The results obtained here will be used in support of further analyses of the data with respect to factors of epidemiological interest.