Diagnosis of fungal pneumonia (FP) in critically ill patients is challenging. Circulating biomarkers for the diagnosis of FP have limitations and the combination of different assays in serum samples and directly from the target organ may further improve the diagnosis of FP. We prospectively assessed the diagnostic utility of paired galactomannan (GM) in bronchoalveolar lavage fluid (BAL) and serum GM and (1→3)-β-d-glucan (BG) assays in critically ill patients at risk of FP. Patients with FP were classified according to European Organisation for Research and Treatment of Cancer-Mycoses Study Group criteria, with modifications. Out of 847 admissions, 51 patients were eligible. There were nine invasive aspergillosis (IA) cases (four proven, five probable), three proven Pneumocysitis jirovecii pneumonia (PJP) cases and one mixed FP case (probable IA and proven PJP). The diagnostic accuracy as given by the area under the receiver operating characteristic curve in IA cases (proven and probable) for GM in BAL was 0.98 (95% CI, 0.94-1.00), whilst for GM and BG in serum it was 0.85 (95% CI, 0.74-0.96) and 0.815 (95% CI, 0.66-0.96), respectively. For IA cases (proven and probable) AUC for GM in BAL was significantly higher than GM and BG in serum (p0.025 and p0.032, respectively). In one of four proven and one of six probable IA cases, GM in serum remained negative, whereas GM in BAL was positive. In patients with IA, GM (90%) and BG (80%) appeared a mean of 4.3days (range, 1-10days) before Aspergillus was cultured. GM detection in BAL appears to improve the diagnosis of IA in critical patients. © 2010 The Authors. Clinical Microbiology and Infection © 2010 European Society of Clinical Microbiology and Infectious Diseases.