Optimization of the cutoff value for the Aspergillus double-sandwich enzyme immunoassay.

Abstract

BACKGROUND: Many health care centers worldwide use the Platelia Aspergillus enzyme immunoassay (PA-EIA; Bio-Rad Laboratories) for diagnosis of invasive aspergillosis (IA). A cutoff optical density (OD) index of 1.5 was originally recommended by the manufacturer, but in practice, most institutions use lower cutoff values. Moreover, a cutoff OD index of 0.5 was recently approved in the United States. In the present study, we set out to optimize the cutoff level by performing a retrospective analysis of PA-EIA values for samples that had been obtained prospectively from adult patients at risk for IA at 2 European health care centers. METHODS: In total, 239 treatment episodes were included of which there were 19 episodes of proven IA and 19 episodes of probable IA. Per-episode and per-test analyses and receiver operating characteristic curves were used to determine the optimal cutoff value. RESULTS: In the per-episode analysis, lowering the cutoff OD index for positivity from 1.5 to 0.5 increased the overall sensitivity by 21% (from 76.3% to 97.4%) but decreased the overall specificity by 7% (from 97.5% to 90.5%). Requiring 2 consecutive samples with an OD index > or = 0.5 resulted in the highest test accuracy, with an improved positive predictive value. At a cutoff OD index of 0.5, the antigen test result was positive during the week before conventional diagnosis in 65% of cases and during the week of diagnosis in 79.5% of cases. CONCLUSIONS: A cutoff OD index of 0.5--identical to the approved cutoff in the United States--improves the overall performance of the PA-EIA for adult hematology patients.