This report analyses rate, predisposing factors, causes and outcome of 458 cardiac arrests occurring during anaesthesia and the following 24 hours and collected in France by a national prospective survey between 1978 and 1982. The overall rate of cardiac arrests was 1 per 433 anaesthetics, resulting in death before the 24th postoperative hour in 79% of cases, i.e. 1 per 547 anaesthetics. Cardiac arrests not related to anaesthesia were three times more frequent than those related to anaesthesia. Among these, one half is partially related to anaesthesia and the other totally. The prognosis of cardiac arrests totally related to anaesthesia is six times better than that of not related arrests (64% of recoveries without sequelae vs 10%). The cardiac arrests not related to anaesthesia occur mainly at the end of surgery and the following 24 hours. The main causes are hypovolaemia, sepsis and severity of surgical illness. Among cardiac arrests related to anaesthesia, a quarter occurred during induction, another during maintenance and a half during recovery from anaesthesia. The first were mainly related to histamine release and had the lowest mortality rate. The last were mainly related to unrecognized postoperative respiratory depression and resulted in the highest mortality. The authors underline the necessity of better preoperative detection of high risk patients, smooth induction of anaesthesia, continuous clinical monitoring of the patient during surgery and the recovery period.