What a Crab to Diagnose and Monitor Invasive Aspergillosis

Abstract and Commentary

Synopsis: Invasive aspergillosis among neutropenic patients could be reliably diagnosed using a commercial test Fungitell ™ to detect the cell wall component (1->3)-ß-D-glucan of certain fungi, including Aspergillus. However, diagnosis was more accurate when both (1->3)-ß-D-glucan and galactomannan were detected.

Source: Pazos C, et al. A Contribution of (1->3)-ß-D-glucan Chromogenic Assay to Diagnosis and Therapeutic Monitoring of Invasive Aspergillosis in Neutropenic Patients: A Comparison With Serial Screening For Circulating Galactomannan. J Clin Microbiol. 2005;43:299-305.

Serum from forty neutropenic patients was used retrospectively to determine what serum (1>3)-ß-D-glucan (BDG) measurement might contribute to the diagnosis of invasive aspergillosis (IA) among neutropenic patients at high risk of developing invasive aspergillosis (IA).

Pazos and colleagues had already defined 40 of 125 consecutive neutropenia patients at high risk according to the EORTC/MSG definitions as having proven IA (5 cases), probable IA (3 cases), possible IA (3 cases), with their being no evidence of IA found in the remaining cases. The detection of galactomannan was used to provide part of the mycological evidence. The serum left over was stored at -70°C, until tested for BGD using a commercial test Fungitell (formerly known as Glucatell). The assay was first developed as the G-test in Japan over a decade ago and uses amoebocytes from the same hermit crab Limulus polyphemus used to detect endotoxin. Removing Factor C allows detection of (1->3)-ß-D-glucan but not endotoxin (see Figure 2). As little as 32 ng/L (1->3)-ß-D-glucan can be detected, but a threshold of at least 60 ng/L has been set to eliminate background levels. The test has recently been cleared for diagnostic use by the FDA and is now available in the United States.

Pazos et al found that BGD was detected in the serum of all proven cases, 2 of the 3 probable cases, 1 of the 3 possible cases, and in 3 of the 29 cases without IA. The course of BDG in serum was similar to that of galactomannan, but BGD was detected earlier and rose sooner. The sensitivity, specificity, and positive and negative predictive values for both BGD and galactomannan tests were also identical, being 87.5%, 89.6%, 70%, and 96.3%, respectively, assuming that only proven and probable cases represented true cases, and that those classified as not having IA did not have the fungal disease. However, performance of both markers could improve diagnostic efficiency when the results were combined, yielding a sensitivity, specificity, and positive and negative predictive values of 87.5%, 100%, 100%, and 96.3%. This was because each test could be used to identify false reactions in the other test. Pazos et al concluded that a proper, prospective evaluation ought now to be done, given the encouraging results of their study.

Comment by J. Peter Donnelly, PhD

It may seem very odd that after more than a decade in the wilderness, we now have a test for BGD that is commercially available and appears to highly efficient in contributing to the diagnosis of IA in neutropenic patients. Even more surprisingly, combining the test results of BGD detection with those of the galactomannan assay improves the efficiency. But before we get completely carried away, a few observations are worth noting. Firstly, a ratio of =1.5 for the galactomannan was considered positive, whereas, in the United States, the threshold is set at =0.5. Next, only a minority of their patients received any antifungal prophylaxis; 9 of the 40 or 22%, to be exact. In addition, they adopted a threshold of 120 pg/mL for BDG, not the manufacturers level of 60 pg/mL, without saying why. Most importantly, the cases were classified on the basis of the EORTC/MSG criteria, which requires mycological evidence by microscopy, culture, or galactomannan assay of appropriate specimens before a clinically defined case can be upgraded from possible to probable IA. Consequently, failure to obtain such specimens weighs equally with negative results. So, translating these results to other centres demands caution.

Interestingly, the dynamics of BDG were similar to those of galactomannan, indicating that either could be used to monitor treatment. Moreover, twice weekly screening for the presence of both markers appears sufficient. The fact that 1 test helped identify false-positive results with the other test is encouraging, but would add to the cost of screening, which may deter clinicians from ordering either. This would be a pity, since screening of the 2 tests together with an early CT scan might result in lower overall costs by helping better differentiate between those patients who need antifungal therapy and those who don’t. Indeed, if there really are so few cases of IA amongst patients at similar high risk (namely 8 of 154 [5%]), as found by Pazos et al, a strategy that efficiently separates proven and probable cases from the rest may prove very cost-effective, if only these patients are treated with antifungal therapy and the remainder were given empirical therapy for not more than a week, whilst the diagnosis is being pursued or better still none at all. Having got this far, it would now be really helpful if this or another group could show the contribution of PCR alongside BGD and GM to the diagnosis of IA.

The test is tailored to detect 1-3 ß-D-glucan by removing Factor C (a). Factor G is activated by BGD, the product and activated factor B help produce the clotting enzyme and this, in turn, cleaves the subtrate Boc-Leu-Gly-Arg-p-nicroanaline to release the yellow coloured. There is thus a direct relationship between the amount of BGD in the sample and the release of of p-nicroanaline.

Dr. Donnelly, Clinical Microbiologist, University Hospital, Nijmegen, The Netherlands, is Associate Editor of Infectious Disease Alert.


1. Ascioglu S, et al. Defining Opportunistic Invasive Fungal Infections in Immunocompromised Patients With Cancer and Hematopoietic Stem Cell Transplants: An International Consensus. Clin Infect Dis. 2002;34:7-14.