Measuring salivary cortisol in studies of child development: watch out--what goes in may not come out of saliva collection devices

Technical advances that enable the noninvasive measurement of biomarkers in saliva have spawned a generation of investigations that integrate biological variables into behavioral and developmental research. This study examines whether the collection of saliva, using common absorbent devices compromises the measurement of cortisol when saliva specimens have low sample volume. Within subjects (n = 20), saliva samples were prepared to experimentally represent a gradient of lower to higher sample volumes. One aliquot was immediately frozen (no treatment control) and the remaining aliquots were absorbed ("collected") using one of three collection techniques employed in studies of child development (e.g., braided cotton dental rope, Salivette cotton pledget, or hydrocellulose microsponge). The sample volume recovered from each device relative to the initial volume available to be absorbed, and cortisol level recovered from each device relative to the untreated-control condition were measured. Results reveal that for certain collection devices (1) the percent volume recovered is related to the initial volume available to be absorbed, (2) a substantial percentage of cortisol in saliva specimens can remain in absorbent material, and (3) the percent of cortisol recovered can be associated with the initial sample volume available to be absorbed. When research participants, such as young children, produce low volume saliva specimens, some absorbent devices may have the potential to introduce error variance in the measurement of salivary cortisol.