Intrapleural haemorrhage is a serious complication of artificial pneumothorax (A.P.), sometimes endangering life and often leading to obliterative pleurisy and loss of the A.P. It is the purpose of this paper to discuss its causes, incidence, pathology, treatment, and effect on the A.P., on the basis of a study of the records of 37 cases in which intrapleural haemorrhage of not less than one pint (600 ml.) occurred after refilling the A.P. or after thoracoscopy at the Brompton Hospital between 1939 and 1949. INCIDENCE Six of these cases followed A.P. refills, of which approximately 180,000 were given at the hospital during the relevant period, giving an incidence of approximately one in 30,000 refills. An incidence of one haemothorax per 2,500 A.P. refills over three years was quoted by Miller and Rinkel (1947). In order to restrict the present series to cases of severe haemorrhage, only those from which at least 20 oz. of blood or " heavily blood-stained fluid " were aspirated are included. Such criteria, if applied to the cases of Miller and Rinkel, would exclude four of their cases, leaving a comparable incidence of one in 5,000 refills.The incidence of haemothorax following adhesion section at the Brompton Hospital was estimated by examining 814 patients having adhesion section between 1939 and 1949, performed by surgeons and trainees. Haemothorax requiring aspiration of at least 20 oz. (600 ml.) of blood or heavily blood-stained fluid occurred 31 times, representing an incidence of 3.9%. There were a further 39 cases in which the volume of blood aspirated was between 10 and 20 oz. (300-600 ml.), and it is probable that many smaller haemothoraces occurred.The incidence of haemothorax following thoracoscopy and adhesion section necessarily varies with the skill of the surgeon and the number of operators included in the series. Summation of the British series of Brock (1938), Edwards, Penman, and Logan (1944), Laird (1945), Watt (1947), and Wollaston (1947 gives a total of neady 2,500 operations for adhesion section: the incidence of haemothorax of variable size was 3.5 %. Similarly, the combined American series of Graham, Singer, and Ballon (1935), Newton (1940), Goorwitch (1944, and Day, Chapman, and O'Brien (1948) cover more than 8,000 operations for adhesion section: the incidence of haemothorax of variable size was 4.3%.
