BACKGROUND: Extensive variations in human surfactant protein D (SP-D) levels in circulation as measured by ELISA exist in the published literature. In order to determine the source of these variations, factors influencing the measurement by ELISA were explored. MATERIALS AND METHODS: Peripheral blood from healthy individuals was collected into various vacutainers during the same blood draw. Recombinant SP-D was diluted into different matrices and used for a standard curve. Samples were analyzed by capture ELISA using one of two distinct detection antibodies. RESULTS: The type of matrix had some effects on detection of recombinant SP-D. The type of anticoagulant used and dilution factor had very little effect, except for in plasma collected in EDTA vacutainers. The extent of variation in published values seemed to be due to the ELISA configuration employed, and, in agreement with this, we found that by switching the detection antibody, there was a 50% decrease in the extrapolated SP-D value of serum and plasma samples. Storage of samples resulted in slight changes in measured SP-D levels. CONCLUSIONS: The ELISA configuration employed to measure circulating levels of SP-D has a significant effect on the extrapolated values. In both configurations tested, the use of EDTA as a coagulant resulted in inconsistent values, and we, therefore, suggest the avoidance of this anticoagulant when assaying for SP-D by ELISA. While the demonstrated effects of several factors on measurement of SP-D may not account for all the disparities amongst the previous studies, they stress that variations in methodologies for measuring the same protein can result in very inconsistent results.