Galactomannan Antigen Detection Immunoassay: an Intra-laboratory Reproducibility Study to Identify the Origin of Analytical False Positive Results

N. Guigue 1, S. Lardeux2, A. Alanio1, S. Hamane1, M. Tabouret2, S. Bretagne1; 1Université Paris Diderot, Sorbonne Paris Cité, Paris, France, 2Bio-Rad Lab., Steenvoorde, France


Background: Galactomannan (GM) antigen is widely used for diagnosing and monitoring invasive aspergillosis but has always been hampered by false positive results from various origins. Based on a reproducibility study, our purpose was to characterize the occurrence of analytical false positive among all positive results and to study their impact on conclusions that could be made on between days variations. Methods: We tested 550 consecutive sera from 235 patients over 22 days using the Platelia Aspergillus #62794 kit (Bio-Rad). Each serum sample was tested in 4 replicates (R1, R2, R3, and R4). R1 was performed by 3 different routine technicians (rT) over the period. R2 to R4 were tested by a unique technician (uT) blind of R1 result. R1 and R2 were from the same GM extraction performed by rT. R3 and R4 were from another extraction performed by uT. The samples were classified as confirmed negative (CN, all replicates with ODI <0.5), confirmed positive (CP, all replicates with ODI ≥0.5). The other situations were classified as unreproducible positive (UP) divided in extraction UP (R1 and R2 ODIs ≥0.5, and R3 and R4 ODIs <0.5, or conversely) and ELISA UP (one replicate with ODI ≥0.5). When replicate values of samples around the 0.5 ODI threshold were within the CV of the kit, these samples were classified as non-conclusive. Each positive serum in either test was retested the day after (R5 by rT and R6 to R8 by uT) after a 4°C storage. Results: Through R1-R4 testing of the 550 sera, 520 (94.5%) were CN, 15 (2.7%) CP, 4 (0.7%) extraction UP, 5 (0.9%) ELISA UP, and 6 (1.1%) non-conclusive. Through R5-R8 testing, the classification of TPs and the non-conclusive remained unchanged. In contrast, the 9 extractions and ELISA UPs (1.6% of the whole sampling and 37.5% of the 24 GM positive samples) were retested negative. Conclusion: This is the first laboratory study distinguishing the different origin of false positive results. We showed the good stability of GM at 4°C over 24 hours and that negativation of retesting is due to technical issues on the first analysis. To avoid useless investigations or treatments, this study underlies the need to eliminate analytical false positive results or to perform as soon as possible a second evaluation of the sample.


abstract No: 


Full conference title: 

54th Interscience Conference of Antimicrobial Agents and Chemotherapy
    • ICAAC 54th (2014)