Abstract

OBJECTIVE--To audit the use of bronchial arteriography and embolisation for controlling haemoptysis. DESIGN--Retrospective review of radiological and clinical data. SETTING--Brompton and National Heart Hospitals. PATIENTS--35 patients with severe pulmonary disease in whom 58 bronchial arteriograms were obtained between 1 January 1984 and 31 December 1989 with the intention of bronchial artery embolisation for controlling haemoptysis. MAIN MEASURES--Rate of technical success and cessation of haemoptysis; detailed evaluation of patients, particularly those with major haemoptysis (> 100 ml expectorated blood); and retrospective assessment of the appropriateness of the procedure in each. RESULTS--58 procedures were performed, nine of which were unsuitable for detailed analysis. Nine procedures were for minor haemoptysis, which subsequently recurred, and 40 for recent major haemoptysis in 26 patients with cystic fibrosis (16) aspergilloma (six), bronchiectasis (three), and an unknown diagnosis (one). The median total volume of haemoptysis in the episode before the procedure was 680 ml (range 270-2200 ml). Embolisation was technically successful in 33/40 procedures, in 17 of which, however, major haemoptysis recurred within 10 days of the procedure, leaving 16 clinically and technically successful procedures in 15 patients. Five patients (three with aspergilloma, two with cystic fibrosis) died of haemoptysis despite attempted embolisation. CONCLUSION--Success rate of bronchial artery embolisation was 40%(16/40). IMPLICATIONS--Bronchial artery embolisation is probably not justified for minor haemoptysis or when performed more than one week after a major haemoptysis. Repeat arteriograms during a single period of haemoptysis are seldom useful. With these criteria 43% fewer procedures would have been performed with no loss of clinical benefit.