Measurement of beta-(1,3)-glucan in household dust samples using Limulus amebocyte assay and enzyme immunoassays: an inter-laboratory comparison

Environmental levels of beta-(1,3)-glucan, an inflammatory fungal cell wall component, have been suggested to be related to respiratory symptoms. However there is currently little data comparing beta-(1,3)-glucan detection methods and/or results obtained in different laboratories. The aim of this study was to compare levels of beta-(1,3)-glucans detected in household dust samples (n = 40) using different extraction/detection methods (Limulus amebocyte assay (LAL), inhibition enzyme immunoassay (EIA) and sandwich EIA) in five different laboratories. Dust sample aliquots were sent to participating centres, extracted and analysed for beta-(1,3)-glucan according to standard in-house procedures. Significant differences in the levels of beta-(1,3)-glucan were observed between all laboratories (geometric mean levels ranging from 15.4 mu g g(-1) to 4754 mu g g(-1) dust; p <0.0001) with the exception of those using a similar LAL method. The inhibition EIA used in laboratory D produced mean beta-(1,3)-glucan measurements 80-100 times higher than the LAL assays, 4 times higher than the sandwich EIA in the same lab, 17.6 times those obtained with the EIA in lab E and 363 times those obtained in the EIA in laboratory C. Pearson's correlations generally showed significant associations between methods and laboratories, particularly those using similar methodology (R ranging from 0.5 to 0.8; p <0.001), although some poor and even inverse correlations were observed. Bland-Altman analyses showed moderate to good agreement between most assays, although clear absolute differences were observed. In conclusion, although results obtained with different methods were often significantly correlated and therefore comparable in relative terms, direct comparison of results between laboratories and assays may be inappropriate.