Pesquisa do antigénio Galactomannan em lavados broncoalveolares na detecção precoce de Aspergilose Invasiva em doentes imunodeprimidos

Abstract

Oncological patients undergoing immunosuppressive treatments have a higher risk for acquiring opportunistic infections such as invasive aspergillosis (IA). In the diagnosis of primary pulmonary infection, aspergillosis shows nonspecific clinical signs of infection, therefore conventional methods of diagnosis, have poor sensitivity and specificity. New diagnostic strategies such as the detection of galactomannan antigen (GM) by ELISA. The aim of this study was to validate the GM test in the diagnosis of invasive pulmonary aspergillosis (API) and determine the best cut-off of GM to validate this marker in early diagnosis of API in samples of oncological patients. A prospective cohort study was conducted in bronchial samples (LB/LBA) of 104 patients with suspected pulmonary aspergillosis, during a period of nineteen months in the Department of Microbiology, Porto IPOFG. Patients were defined with API (probable and proven) or without API, according to the revised definitions for the EORTC/ MSG. As a control group (n=41) bronchial samples from patients without evidence of respiratory infection were studied. We evaluated the variables of culture, DNA and cytology for Aspergillus, as clinical variables, such as CT scan and the patient´s risk factors, and antibiotic therapy. Of the 104 patients studied with suspected pulmonary aspergillosis, 13 cases were diagnosed with API (12,5%), and of this, 54% of cases had hematologic disease (n=7/ 13). All had GM positive, 46% positive culture for Aspergillus, 62% with positive DNA, and 46% with positive cytology. In 62% of the cases were submitted to antifungal therapies. For a cut-off GM ≥1, sensitivity of the GM test to diagnose API (probable or proven) was 100% and specificity was 45% with IC95% [0,35 to 0,55]. The positive predictive value for the GM test was 24% and negative predictive value was 93%. The predictive value for the ROC curve in this sample was 0,8 to GM ≥1,3; the GM test had a sensitivity of 85% and a specificity of 65%, higher than that found for a cut-off ≥1. The diagnostic GM test in BAL is useful for the diagnosis of IPA. However, this marker must be associated with others parameters, clinical and laboratory, and its real value has a higher negative predictive value for this infection. To enhance the possibility of false-positive by interference with the treatment of certain antibiotics.

Date of Award	3 Jan 2012
Original language	Portuguese
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Catarina Lameiras (Supervisor) & Maria Manuela Pintado (Co-Supervisor)