Utility of routine surveillance for invasive aspergillosis (IA) using the asp platia galactomannan assay in a bone marrow transplant unit

Iwuchukwu C, Pursell K, Pitrak D, Thompson K, Matushek S, van Besien K

Author address: 

University of Chicago, Chicago, USA


SCT recipients have high risk for and poor outcomes with IA. This is a very appropriate group to study surveillance for early diagnosis and perhaps preemptive therapy. The purpose of this study was to evaluate the sensitivity, specificity, and predictive values of Asp Ag testing done twice weekly as surveillance in a high-risk population in a BMT unit. The study period extended from Jan1, 2004 to June 30, 2004. We reviewed the records of 60 SCT recipients who had 582 (mean 9.7, range 1-21) Asp Ag assays. Patients were classified as having definite, probable, or possible IA based on accepted criteria established by NIAID MSG/EORTC. During the period of study, 16 patients developed clinical symptoms, signs, and/or radiographic evidence of IA. This included 2 cases of definite IA (3.3 %), 8 cases of probable (13.3 %), and 6 cases (10.0 %) of possible IA. One patient with clinical data consistent with probable IA was excluded after the BAL grew Mucor. The results of Asp Ag testing are shown when the assay was positive once and when the test was repeatedly positive on the same sample. The performance of the test depended on whether or not the diagnosis of IA was limited to definite or probable cases and if the assay was positive twice using the same sample. When only cases of definite or probable IA were classified as having disease, the sensitivity was 30 % and the sensitivity was 78 % (PV+ = 21 %, LR+ 1.36, PV- 84.7 %, LR-0.90 with positive tests that were negative on repeat testing. When a repeat positive test done on the same sample was required for positivity, sensitivity was still 30 %, but specificity increased to 90 %. This increased the PV+ (37.5 %) and LR+ (3.0), but PV- (86.5 %) and LR- (0.78) changed little. When patients meeting criteria for possible IA were included, the sensitivity of the test with the single positive test was 25 %, specificity 34.8 %, PV+ 28.6 %, LR+ 0.96, PV- 73.9, and LR- 2.16. With a positive test on repeat, sensitivity was 25 %, specificity 90.9 %, PV+ 50 %, LR+ 2.75, PV- 76.9 %, and LR- 0.82. Of note, the patients with definite or probable IA with a positive galactomannan test had radiographic evidence of disease before or at the time of testing, not earlier in the course of infection. In summary, the routine surveillance for IA using serum Asp Ag is limited by both poor sensitivity and poor specificity if a test is not repeatedly positive on the same sample. Although false positive results could theoretically be evidence of early sub-clinical infection, this is unlikely given the lack of disease progression without any change in anti-fungal prophylaxis or therapy and the poor sensitivity in established disease.

abstract No: 


Full conference title: 

15th Annual Focus on Fungal Infections
    • FFI 15th (2005)