Invasive pulmonary aspergillosis (IPA) is now the leading cause of infectious mortality in stem cell transplant recipients. Despite recent advances in molecular diagnostic techniques the diagnosis remains problematic. The mycological yield from bronchoalveolar lavage (BAL) is poor, especially if the patient has been exposed to antifungal drugs. Galactomannan (GM), a component of Aspergillus’ cell wall is detectable in serum and BAL fluid in IPA, though its role in the latter has not been validated. Exhaled breath condensate (EBC) collection provides a non-invasive means of sampling the airway lining fluid. This method has been used in respiratory disease to detect markers of inflammation and oxidative stress. EBC is collected by breathing into a tube surrounded by a cooled insulated metal sleeve (RTube®, Respiratory Research, USA) and results in 1–2 millilitres of fluid. The application of EBC detection of volatile and non-volatile organic compounds has not been explored in this setting. We report on pilot data of the detection of GM in EBC from a subset of patients at high risk of infection forming part of a prospective study into the early diagnosis of IPA in allogeneic transplant recipients and patients with acute leukaemia. Patients were recruited prior to commencing treatment for their underlying haematological condition; the study period finished at neutrophil recovery (> 0.5 x 109/L) or discharge from hospital. EBC and serum were collected once and twice weekly respectively, throughout the study period. High resolution CT (HRCT) imaging of the chest was carried out after 72 hours of neutropenic fever unresponsive to broad-spectrum antibiotics. Where possible, BAL was carried out in the most affected lobe in patients with an abnormal HRCT chest. GM was measured in EBC, serum and, where applicable, BAL fluid using a commercially available kit (Platelia® Aspergillus, Bio-Rad, France). The likelihood of IPA was assigned using criteria published by European Organisation for Research and Treatment of Cancer/Mycoses Study Group (EORTC/MSG) by 2 investigators blinded to clinical details (Ascioglu et al, CID 2002). Of 39 patients, 6 had proven/probable IPA, 9 possible and 24 had no evidence of IPA. Four of the 6 patients with proven/probable disease had GM-positive BAL fluid; 3 of these 4 had persistently negative GM in serum. In 1 of the 6, the EBC GM was positive 3 weeks earlier than the development of HRCT evidence of IPA and a positive BAL GM. In another, the EBC GM was positive early in the infection, whilst BAL GM remained negative. EBC and serum GM followed the same trend in all 6 patients, although the values for both EBC and serum remained below the currently recognised positive cut-off of 0.5 in 2 patients. In patients with possible or no evidence of IPA, GM was negative in all but 2 of 111 EBC samples. These results suggest that GM can be detected in EBC and follows the same trend as in serum. It may even predate GM detection in BAL fluid or abnormal HRCT signs, thereby obviating the need for more invasive investigations. Elevated EBC GM levels early in the course of neutropenia may represent infection and could be applied to the employment of pre-emptive treatment strategies.

Author notes

Disclosure: Consultancy: Dr Agrawal: Gilead Sciences, Pfizer Ltd, Schering-Plough and Merck Ltd. Research Funding: Dr SG Agrawal, Dr SR Doffman: Unrestricted Educational Grant from Pfizer Ltd, Gilead Sciences Ltd, Schering-Plough.