A validation study of BM-AspICU algorithm for diagnosing invasive pulmonary aspergillosis in critically ill patients using post-mortem minimally invasive tissue sampling

BM-AspICU algorithm is a newly proposed algorithm for diagnosing invasive pulmonary aspergillosis (IPA) in critically ill patients. However this algorithm has not been validated prospectively. In this study, the primary objective was to determine the diagnostic accuracy of BM-AspICU algorithm using post-mortem minimally invasive tissue sampling (MITS) (multiple transthoracic biopsies obtained by BARD® needle). In this prospective study, 122 critically ill patients with suspected IPA were screened and 80 were included based on suggestive clinical/imaging features or positive <italic>Aspergillus</italic> from lung; and followed up till death/discharge. Deceased patients were subjected to MITS and samples sent for histopathological examination, fungal stain/culture and PCR by AsperGenius®. Mean age of patients were 45.3 + 16.6 years, 36% were male, and the mean ICU stay was 11.2 + 8.4 days. The most common underlying diagnosis was community-acquired pneumonia (61%) and diabetes was most common comorbidity (26%). ~25% of patients were immunocompromised at the time of diagnosis. CT thorax, serum galactomannan(GM) , mini-BAL GM, blood <italic>Aspergillus</italic> PCR, mini-BAL KOH, mini-BAL fungal culture and mini-BAL <italic>Aspergillus</italic> PCR were positive in 62%, 39% , 65.5%, 48.39% , 11.1%, 7.4% and 83.3% patients respectively. MITS sample histopathology and <italic>Aspergillus</italic> PCR were positive in 4% and 92% patients respectively. Overall, diagnosis by BM-AspICU algorithm had a sensivity and specificity of 87.5% and 50% respectively compared with biopsy. BM-AspICU algorithm has good sensitivity but poor specificity for diagnosing IPA in critically ill patients.