Blood can also be a valuable matrix for measuring biomolecular adducts such as DNA, hemoglobin, or albumin adducts (Anwar 1997).
To avoid obtaining non-representative samples, assessment of adrenocortical function can be standardized using faeces as a hormone matrix for measuring time-averaged response to a stressor [ 9- 11].
Ostrea and colleagues have reported that meconium is a sensitive matrix for measuring markers of gestational drug use and prenatal pesticide exposure [ 2- 6].
Meconium is a useful biological matrix for measuring prenatal tobacco smoke exposure and could be used in epidemiological studies that enroll women and infants at birth.
However, the biologic and technical factors noted above strongly support urine as the optimal matrix for measuring nonpersistent, semivolatile, hydrophilic environmental agents.
We are not aware of previous studies that attempted to validate meconium as a matrix for measuring biomarkers of prenatal tobacco smoke exposure using repeated serum cotinine measures for comparison.
Meconium holds promise as a biological matrix for measuring the intensity and duration of environmental toxicant exposure and future studies should validate the utility of meconium using other environmental toxicants.
Saliva has also been explored as a matrix for measuring selected nonpersistent chemicals, such as atrazine and diazinon (Borzelleca and Skalsky 1980; Denovan et al. 2000; Lu et al. 1997a, 1997b, 1998, 2003).
Because breast milk (particularly colostrum, the first breast milk after giving birth) is rich in fat content, it may be a better matrix for measuring highly lipophilic compounds such as PBDEs (Antignac et al. 2009; Gomara et al. 2007), and thus may be more suitable than cord blood or maternal serum for assessing prenatal exposure to PBDEs.
Such diverse thyroid hormone findings might result from using different matrices (Kimbrough and Krouskas 2001) for measuring exposure (i.e., breast milk or venous blood), various specimen collection time points, and/or studied isomers (dioxins, indicator PCBs) for different populations at unique sample sizes.
The simulations compared pairs of samples of individuals differing in the shape and orientation of their covariance matrices for the measured variables.
