RESUMO
INTRODUCTION: The Dynamiker® Fungus (1-3)-ß-D-glucan assay (DFA) allows the testing of samples in smaller batches compared to the well-established Fungitell® assay (FA) making the assay cost-effective in centers with small numbers of samples. Evaluations of its performance for the diagnosis of invasive aspergillosis (IA) are limited. Therefore, we compared the two assays and evaluated their clinical performance in diagnosing IA. METHODS: A total of 60 adult hematology patients were screened for IA, 13 with probable IA, 19 with possible IA, and 28 with no IA. Serum specimens (n = 166) were collected twice-weekly and tested for (1-3)-ß-D-glucan (BDG) using FA and DFA which were compared quantitatively with Spearman rank correlation analysis and qualitatively with the Chi-square test. Agreement and error rates were determined using FA as the reference method. Sensitivity, specificity, and positive predictive and negative predictive values in diagnosing IA were calculated. RESULTS: The performance of the DFA was highly consistent with that of the FA, both quantitatively (rs = 0.913) and qualitatively (kappa = 0.725). The agreement was 85% with 8% minor, no major, and 7% very major errors (FA+/DFA-). Using a cut-off value of 20 pg/mL for DFA, very major errors were reduced to 1%, although 5% major errors were detected. BDG levels were lower with DFA than FA (slope 0.653 ± 0.031). Sensitivity, specificity, positive predictive value, and negative predictive value (NPV) was 67%, 53%, 44%, and 74% for FA, and 53%, 67%, 49%, and 71% for DFA, respectively. The optimal BDG positivity threshold calculated did not lead to significant test quality improvement for either assay. However, a higher % of patients with probable IA (62%) had ≥ 2 consecutive positive specimens compared to patients with no IA (FA-BDG 26%, p = 0.10, and DFA-BDG 10%, p = 0.01) leading to improved sensitivity and NPV (71% and 85% for DFA, and 95% and 96% for FA, respectively). CONCLUSION: DFA could be a valuable alternative to the FA, particularly in laboratories with small numbers of samples. The results of the BDG testing should be carefully interpreted in the high-risk setting of patients with hematologic malignancies. Higher NPV was found using as criterion ≥ 2 consecutive positive samples for diagnosing IA.
RESUMO
As conventional microbiological documentation of invasive aspergillosis (IA) is difficult to obtain, serum fungal biomarkers are important adjunctive diagnostic tools. Positivity rates and the kinetic profiles of galactomannan (GM), 1,3-ß-D-glucan (BDG) and Aspergillus DNA (PCR) were studied in high-risk patients with hematologic malignancies. GM, BDG and PCR data from serial serum specimens (n = 240) from 93 adult hematology patients with probable (n = 8), possible (n = 25) and no (n = 60) IA were retrospectively analyzed. Positivity rates and sensitivity/specificity/positive/negative predictive values (NPV) of each fungal biomarker alone and in combination were estimated. The three markers were compared head-to-head and correlated with various biochemical, demographic and patient characteristics. The positivity rates for patients with probable/possible/no IA were 88%/8%/0% for GM (X2 = 55, p < 0.001), 62%/46%/35% for BDG (X2 = 2.5, p = 0.29), 62%/33%/27% for PCR (X2 = 3.9, p = 0.15), 50%/4%/0% for GM + BDG and GM + PCR (X2 = 31, p < 0.001), 50%/8%/22% for BDG + PCR (X2 = 6.5, p = 0.038) and 38%/4%/0% for GM + BDG + PCR (X2 = 21, p < 0.001). Higher agreement (76%) and negative correlation (rs = -0.47, p = 0.0017) was found between GM index and PCR Ct values. The sensitivity and NPV was 45-55% and 90-92% when biomarkers assessed alone and increased to 75-90% and 93-97%, respectively when combined. Weak significant correlations were found between GM, PCR and BDG results with renal/liver function markers (r = 0.11-0.57) with most GM+ and PCR+ samples found in the first and second week of clinical assessment, respectively and BDG later on. Different positivity rates, time profiles and performances were found for the three biomarkers advocating the combination of GM with PCR for the early diagnosis of IA, whereas the high NPV of combined biomarkerscould help excluding IA.