M Cruciani1, P L White2, C Mengoli3, J Löffler4, C O Morton5, L Klingspor6, D Buchheidt7, J Maertens8, W J Heinz4, T R Rogers9, B Weinbergerova10, A Warris11, D E A Lockhart12, B Jones13, C Cordonnier14, J P Donnelly15, R A Barnes16. 1. Azienda ULSS 9 Scaligera, Verona, Italy. 2. Public Health Wales, Cardiff, UK. 3. University of Padua, Padua, Italy. 4. University of Wuerzburg, Wuerzburg, Germany. 5. Western Sydney University, Sydney, Australia. 6. Karolinska Institute, Stockholm, Sweden. 7. Heidelberg University, Mannheim, Germany. 8. Department of Microbiology, Immunology, and Transplantation, KULeuven, Leuven, Belgium. 9. Trinity College Dublin, St James's Hospital Campus, Dublin, Ireland. 10. Department of Internal Medicine - Haematology and Oncology, Masaryk University and University Hospital Brno, Brno, Czech Republic. 11. MRC Centre for Medical Mycology, University of Exeter, UK. 12. University of Aberdeen, Aberdeen, UK. 13. Institute of Infection, Immunity and Inflammation, University of Glasgow, UK. 14. Hopital Henri Mondor, Creteil, France. 15. University of Nijmegen, Nijmegen, The Netherlands. 16. Cardiff University School of Medicine, Cardiff, UK.
Abstract
BACKGROUND: The performance of the galactomannan enzyme immunoassay (GM-EIA) is impaired in patients receiving mould-active antifungal therapy. The impact of mould-active antifungal therapy on Aspergillus PCR testing needs to be determined. OBJECTIVES: To determine the influence of anti-mould prophylaxis (AMP) on the performance of PCR blood testing to aid the diagnosis of proven/probable invasive aspergillosis (IA). METHODS: As part of the systematic review and meta-analysis of 22 cohort studies investigating Aspergillus PCR blood testing in 2912 patients at risk of IA, subgroup analysis was performed to determine the impact of AMP on the accuracy of Aspergillus PCR. The incidence of IA was calculated in patients receiving and not receiving AMP. The impact of two different positivity thresholds (requiring either a single PCR positive test result or ≥2 consecutive PCR positive test results) on accuracy was evaluated. Meta-analytical pooling of sensitivity and specificity was performed by logistic mixed-model regression. RESULTS: In total, 1661 (57%) patients received prophylaxis. The incidence of IA was 14.2%, significantly lower in the prophylaxis group (11%-12%) compared with the non-prophylaxis group (18%-19%) (P < 0.001). The use of AMP did not affect sensitivity, but significantly decreased specificity [single PCR positive result threshold: 26% reduction (P = 0.005); ≥2 consecutive PCR positive results threshold: 12% reduction (P = 0.019)]. CONCLUSIONS: Contrary to its influence on GM-EIA, AMP significantly decreases Aspergillus PCR specificity, without affecting sensitivity, possibly as a consequence of AMP limiting the clinical progression of IA and/or leading to false-negative GM-EIA results, preventing the classification of probable IA using the EORTC/MSGERC definitions.
BACKGROUND: The performance of the galactomannan enzyme immunoassay (GM-EIA) is impaired in patients receiving mould-active antifungal therapy. The impact of mould-active antifungal therapy on Aspergillus PCR testing needs to be determined. OBJECTIVES: To determine the influence of anti-mould prophylaxis (AMP) on the performance of PCR blood testing to aid the diagnosis of proven/probable invasive aspergillosis (IA). METHODS: As part of the systematic review and meta-analysis of 22 cohort studies investigating Aspergillus PCR blood testing in 2912 patients at risk of IA, subgroup analysis was performed to determine the impact of AMP on the accuracy of Aspergillus PCR. The incidence of IA was calculated in patients receiving and not receiving AMP. The impact of two different positivity thresholds (requiring either a single PCR positive test result or ≥2 consecutive PCR positive test results) on accuracy was evaluated. Meta-analytical pooling of sensitivity and specificity was performed by logistic mixed-model regression. RESULTS: In total, 1661 (57%) patients received prophylaxis. The incidence of IA was 14.2%, significantly lower in the prophylaxis group (11%-12%) compared with the non-prophylaxis group (18%-19%) (P < 0.001). The use of AMP did not affect sensitivity, but significantly decreased specificity [single PCR positive result threshold: 26% reduction (P = 0.005); ≥2 consecutive PCR positive results threshold: 12% reduction (P = 0.019)]. CONCLUSIONS: Contrary to its influence on GM-EIA, AMP significantly decreases Aspergillus PCR specificity, without affecting sensitivity, possibly as a consequence of AMP limiting the clinical progression of IA and/or leading to false-negative GM-EIA results, preventing the classification of probable IA using the EORTC/MSGERC definitions.