Supportive Data

In clinical studies submitted to the US Food and Drug Administration, the sensitivity of the test for serum was reported to be 81% for proven or probable invasive aspergillosis (n=31 patients), and the specificity was 89% (n=148 patients). The positive and negative predictive values were reported as 68% and 96% respectively, based on an average prevalence of 14% in the study population. In a low prevalence population (5%), the positive predictive value decreases to 31%; the negative predictive value remains at 96%.(Package insert: Platelia Aspergillus EIA. Bio-Rad; 06/2003)

Accuracy:

The performance characteristics of the Platelia Aspergillus enzyme immunoassay (EIA) for the detection of galactomannan in bronchoalveolar lavage (BAL) fluid were validated at Mayo Clinic Laboratories by comparison of results obtained from an outside reference laboratory using the same assay. These studies demonstrated 95.6% (240/251) agreement between sites (Table 1).

Table 1: Comparison of Platelia Aspergillus Antigen results at Mayo Clinic Laboratories and an outside reference laboratory using BAL fluid (n=251).

Percent Agreement: 95.6% (240/251) (95% Cl; 92.2-97.6)

Kappa value: 0.79

For 10 of the 11 discordant results, testing at Mayo Clinic Laboratories correlated with either serum Aspergillus antigen levels or fungal culture.

Precision:

Intra- and interassay precision was tested for negative, midrange and high-positive, and spiked BAL specimens. The mean index values, standard deviation and percent coefficient of variation were all acceptable, indicating excellent precision. (Tables 2 and 3)

Table 2: Intra-assay precision studies

Positive: >0.5

Negative: <0.5

Table 3: Inter-assay precision studies

Positive: >0.5

Negative: <0.5

Analytical Specificity:

Cross-reactivity studies were performed by testing analyte-negative BAL specimens that had been spiked with varying concentrations of positive control material for the following organisms: Histoplasma capsulatum, Blastomyces dermatitidis, or Cryptococcus neoformans. These studies demonstrated that high concentrations of Histoplasma and Blastomyces antigen in BAL may yield positive results by the Platelia Aspergillus antigen assay. This has been demonstrated in prior published studies,(5) and it is a known limitation of this test that there may be cross-reactivity with dimorphic fungal pathogens.

In addition to the studies above, an analyte-negative BAL specimen was spiked with a pleural fluid that was known to be positive for Streptococcus pneumoniae antigen. This spiked- specimen was tested by the Platelia assay and was negative at all dilutions tested.