Be Wary Brushing in the Right Upper Lobe To the editor: Fiberoptic bronchoscopy is well established in the diagnosis of lung cancer. Mortality related to bronchoscopy has been estimated at between 0.01% and 0.5% in prospective and retrospective studies.1–4 There are no specific data on mortality or morbidity secondary to hemorrhage after simple endobronchial brush sampling, but life-threatening hemorrhage is very unusual in this situation. A 47-year-old female smoker presented with hemoptysis, weight loss, and lassitude. Examination was unremarkable. Investigations confirmed a right upper lobe opacity on chest x-ray, mild microcytic anemia, and slightly abnormal liver function. Computed tomography scan of the chest showed a large right paratracheal mass with some mediastinal lymphadenopathy and a subcarinal lymph node. Liver ultrasound revealed no evidence of metastases. At bronchoscopy, no endobronchial abnormality was seen. Cytology brushings were taken from the apical segment of the right upper lobe. Immediately, brisk bleeding occurred despite endobronchial adrenaline, flooding the right bronchial tree and spilling across the main carina into the left bronchial tree. She was resuscitated and transferred to intensive care. She was ventilated using high inspiratory airway pressures because of a large gelatinous blood clot found to be obstructing the main carina and proximal main bronchi. This was removed endoscopically by suction and lavage, greatly improving ventilation and gas exchange with no further bleeding. In this case, life-threatening bleeding occurred after cytologic brush sampling from the right upper lobe despite no tumor being visible and using only a bronchoscopy brush. The right upper lobe is a potentially dangerous site because hemorrhage can readily spill across the main carina and into the left bronchial tree as a result of its relative anatomical proximity. The British Thoracic Society bronchoscopy guidelines5 do not recommend routine coagulation checks. In retrospect, the frequency of prebronchoscopy hemoptysis, iron deficiency anemia, and mildly deranged liver function could have signaled a higher risk. However, prebronchoscopy hemoptysis is not an established risk factor for bleeding during bronchoscopy, and screening coagulation tests have been shown not to predict bleeding in patients undergoing fiberoptic bronchoscopy with biopsy.6 We would recommend bronchoscopists take extra care when even brush sampling from the right upper lobe even when there is no endobronchial tumor with a history of hemoptysis and suggestive radiology. A. R. L. Medford, BSc, MRCP Respiratory SpR, Lung Research Group, Division of Medicine, University of Bristol, Southmead Hospital, Westbury-on-Trym, Bristol, UK G. D. Phillips, MD, FRCP Department of Respiratory Medicine, Dorset County Hospital, Dorchester, UK