Comparison of nonculture blood-based tests for diagnosing invasive aspergillosis in an animal model

ABSTRACT The European Aspergillus PCR Initiative (EAPCRI) has provided recommendations for the PCR testing of whole blood (WB) and serum/plasma. It is important to test these recommended protocols on nonsimulated “ in vivo ” specimens before full clinical evaluation. The testing of an animal model of invasive aspergillosis (IA) overcomes the low incidence of disease and provides experimental design and control that is not possible in the clinical setting. Inadequate performance of the recommended protocols at this stage would require reassessment of methods before clinical trials are performed and utility assessed. The manuscript describes the performance of EAPCRI protocols in an animal model of invasive aspergillosis. Blood samples taken from a guinea pig model of IA were used for WB and serum PCR. Galactomannan and β-d-glucan detection were evaluated, with particular focus on the timing of positivity and on the interpretation of combination testing. The overall sensitivities for WB PCR, serum PCR, galactomannan, and β-d-glucan were 73%, 65%, 68%, and 46%, respectively. The corresponding specificities were 92%, 79%, 80%, and 100%, respectively. PCR provided the earliest indicator of IA, and increasing galactomannan and β-d-glucan values were indicators of disease progression. The combination of WB PCR with galactomannan and β-d-glucan proved optimal (area under the curve [AUC], 0.95), and IA was confidently diagnosed or excluded. The EAPRCI-recommended PCR protocols provide performance comparable to commercial antigen tests, and clinical trials are warranted. By combining multiple tests, IA can be excluded or confirmed, highlighting the need for a combined diagnostic strategy. However, this approach must be balanced against the practicality and cost of using multiple tests.