[Detection of 1,3-beta-D glucan for diagnosis of invasive fungal infections in hematooncological patients: usefulness for screening of invasive mycosis and for confirmation of galactomannan positive results].

BACKGROUND AND AIMS

1,3-beta-D glucan (BG) -- the antigen of fungal cell wall can be detected by a commercially available test for early detection of invasive fungal infections (IFI). The main advantage of this test is its broad coverage of fungal species. The aim of our study was to evaluate usefulness of BG detection for screening of IFI and for confirmation of galactomannan (GM) positive blood samples. Combination of the results of both tests could lead to correct and early diagnosis of invasive aspergillosis (IA).

PATIENTS AND METHODS

Between January 2005 and July 2007 blood samples were collected in patients from intermediate to high risk of IFI. Moreover, between February and October 2007 all patients that had consecutive positive results of GM had their positive symplex tested also for BG.

RESULTS

In BG screening study, 1154 of blood samples from 104 treatment cycles were tested for BG. The incidence of IFI was 17.3 % (n = 18) and probable or proven IFI was detected in 9 cases (8.6%). The highest sensitivity, specificity, PPV and NPV (88.9 %, 40.7 %, 13.6 % and 97.2 %) were obtained when as criteria for positivity cut off 80 pg/ml and one positive result were used. When consecutive positivity of the test was applied as criterium, cut off 60 pg/ml was found more useful (sensitivity 66.7 %, specificity 47.7 %, PPV 11.8 % and NPV 93.2 %). Low PPV, caused by frequent false positive results, was identified as main limitation of this assay. 65 treatment cycles were positive if 1 sample above 80 pg/ml was used as a cut of for positivity. If consecutive positivity with cut off 60 pg/ml was used, 58 treatment cycles were positive. But in 51 (78.4 %) and 45 (77.5 %) cases, respectively, the positivity was not associated with IFI (false positivity). We did not find any correlation between positive BG assay result and frequency of empirical antifungal treatment, mucositis, yeast colonization, administration of selected antibiotics or infusion solutions or bacteriaemia. In our confirmation study, 40 GM positive episodes in 39 patients were identified. In 31 (78 %) GM positivity was false and was not associated with clinical signs and symptoms of IA. Sensitivity of GM detection in IA was 100 % but PPV only 18 %. Confirmation of consecutive GM positive samples (using cut off index positivity 0,5) by consecutive positivity of BG (with cut off 60 pg/ml) was found very useful for diagnosis of IA -- most of GM false positive results were eliminated and PPV increased to 88 %.

CONCLUSIONS

Our analysis focused on routine use of BG test for panfungal screening of IFI in patients with hematological malignancy and confirmed limited usefulness of this test in such setting. Low sensitivity together with low PPV are major limits of this test. On the other hand, BG testing seems to be a promising tool for confirmation of consecutive GM positive result in serum in patients with IA. Positivity of both tests could increase their PPV of tests and eliminate false positive results.