We describe a pilot study to determine the effect of delays in blood sample processing under simulated field conditions on measurement of corticotropin-releasing hormone (CRH) levels in pregnant women. CRH, a peptide secreted by the placenta into the maternal blood, is of interest in epidemiological studies of gestational duration. Many investigators suspected that CRH might break down quickly after collection, and believed the optimal treatment of blood samples for CRH must include immediate processing under chilled conditions and quick freezing of plasma. Epidemiological studies often have logistical constraints that make such rapid processing unfeasible. To examine how delays in the processing of blood samples might affect the level of measured CRH, we collected whole blood samples from 33 pregnant women attending a prenatal clinic in Boston. We compared CRH levels measured following three different processing delays with the levels of samples that were processed immediately after blood collection, the 'gold standard'. The delayed strategies involved placing the freshly collected whole blood in a cooler or refrigerator for up to 22 h prior to processing. Correlation coefficients comparing delayed with gold standard processing exceeded 0.96. These results suggest that CRH may be measured in blood samples that were spun and frozen up to 22 h after blood collection.