Candida mannan and anti-mannan in the diagnosis of invasive candidal infections in neutropenic patients
Abstract number: P902
Ellis M., al-Ramadi B., Bernsen R., Kristensen J., Alizadeh H., Hedstrom U.
Objective: Antifungal (AF) therapy given to all pts with febrile neutropenia (FN) is costly and toxic. Beta-D-glucan assay is fungal-non-specific; galactomannan is useful only for aspergillosis. This study focused on serial assay of mannan (M) and mannan-antibodies (MA) to diagnose invasive candidiasis (IC) in pts with FN, hence to aid selection of pts for AF. Previous experience with this assay has been limited to non-neutropenic pts, based on retrospective infrequent sampling.
Methods: 100 patients with acute leukaemia undergoing chemotherapy complicated by FN, given liposomal amphotericin B, were studied prospectively with clinical, microbiological (blood culture), and radiological (CT scans chest, liver, spleen, sinuses) evaluations for the development of IC, based on revised EORTC/MSG diagnostic criteria. M + MA were measured daily using Platelia Candida-specific antigen/antibody ELISA kits (Bio-Rad). Diagnostic cut-offs were determined using ROC curves.
Results: 12 of 86 (14%) eligible pts had IC [C. albicans candidaemia (1), C. tropicalis candidaemia (4), hepatosplenic candidiasis (7)], 24 had invasive mould, 50 persistent FN. These last 2 groups served as the comparison group.
Cut-offs were 0.25 ng/ml and 2.5AU/ml for M and MA, lower than manufacturer's recommendations. All pts with IC developed 1 +ve diagnostic M or MA during persistent FN (FIG1). Optimal overall performance occurred when 2 consecutive positive tests for both M and MA were used. Sensitivity, specificity, PPV and NPV [95% CI] were 0.73[0.390.94], 0.80 [0.690.89], 0.36[0.170.59], 0.95[0.860.99] respectively.
A positive correlation (r = 0.28, p = 0.01) was seen with previously determined beta-D-glucan (BDG) concentrations in these pts.
The first +ve M test occurred at a mean ±s.d. of 8.8 ±8.5 (2 to 23) days prior to clinical/mycological diagnosis of IC.
High MA concentrations were delayed until leucopenia resolved.
The candidal colonisation index (CCI) was 0.5 in 60% of the comparison group.
Conclusions: Using institution-based cut off values, early serial determination of combined M+MA is useful to diagnose IC in FN pts. Low PPV may reflect low prevalence of IC, whilst the high NPV confidently excludes IC. Negative M+MA tests in the presence of +ve BDG test indicates IC rather than mould. Specificity may be low as a result of subclinical candida infection in the comparator group (high CCI). This assay could be used as part of a broad fungal diagnostic strategy aimed at tailoring AF.
Figure: Serial course antigen and antibody pt#12.
|Session name:||19th European Congress of Clinical Microbiology and Infectious Diseases|
|Location:||Helsinki, Finland, 16 - 19 May 2009|
|Back to top|