[The causes of perioperative mortality. A trial of the German "CEPOD study."]

We performed a prospective multi-center study in order to determine the causes of 30-day perioperative mortality.

In accordance with the CEPOD-Study and with the kind permission of Dr. N. Lunn, we forwarded two different questionnaires to 135 hospitals. One questionnaire was to be answered by the anaesthetist and the other one by the surgeon involved in cases o perioperative death within the first 30 days after the operation. 12 out of 135 addressed hospitals agreed to participate in the study. These included four small hospitals, six medical centres of medium capacity (about 500 beds) and two University hospitals. In order to obtain an exact description of the events leading to perioperative death, the questionnaires consisted of approximately 60 questions for the collection of demographic data and the surgical as well as anaesthesiological perioperative management.

From 1989 to 1993 more than 300 cases of perioperative death were reported. Only 200 cases could be analyzed due to incompletely answered or unreturned questionnaires. The mean risk-classification (ASA) was 3.46, mean age 74.6 years. Approximately 40 percent of deaths occurred in patients older than 80 years. More than 80 percent of patients had at least one pre-existing cardiovascular disease with prevalence of 41% for pulmonary and gastrointestinal diseases. In the majority of cases abdominal operations were performed, followed by hip-surgery and surgery of the aorta. In 86% of the cases, the surgeon was experienced and had performed the respective operation more than 20 times. In 38.2% an anaesthetist in training was responsible for anaesthesia, but only 11.6% were without supervision of a specialist anaesthetist. The majority of patients received general anaesthesia (78%) and 8.5% had a combination of EDA and general anaesthesia. Regional anaesthesia was performed in 12.5%, local anaesthesia in only 1%. The average blood loss was approximately 1.600 ml (with a very wide range) and 42.5% of the patients needed a transfusion of blood components, primarily in the form of packed red blood cells. Seventeen serious incidents occurred intraoperatively, including three "exitus in tabula". Four patients died shortly after the operation in the ICU, the other ten incidents were managed in the operating room. In 11 of 17 incidents the patients suffered a cardiac arrest; nine patients were resuscitated. Two patients were not resuscitated in view of pre-existing diseases and inoperability. All of the hospitals had an ICU for postoperative care, but two of the smaller hospitals had no recovery rooms. In 22 cases of emergency operations, there was a delay due to a lack of personnel or to logistic problems. In five of these cases, the delay was described as a possible cofactor of perioperative mortality. The most frequent causes of perioperative death were myocardial failure (33.7%) and multi-organ-failure (19.2%), followed by respiratory insufficiency (13%) and septic shock in 9.3%. A necropsy was carried out in only 28 of 200 perioperative deaths (14%); 13% of the cases were discussed in a surgical and only 2.5% in an anaesthesiological mortality-conference. In 9 out of 12 hospitals no mortality-conferences were held. All surgeons and anaesthetists were asked for self-assessment on the basis of an analog scale ranging from 0 and 10 points. The average score was 8.52 points (surgical management) and 9.36 points (anaesthesiological management respectively), which is not always in correspondence with the information provided in the questionnaires.

In order to further reduce perioperative mortality in critically ill patients, every hospital should aim to optimize the structure of the surgical and anaesthesiological departments. A delay due to logistical or personnel problems may be a co-factor in perioperative mortality. Recovery rooms with experienced personnel should be the standard in postoperative anaesthesiological care. (ABSTRACT TRUNCATED)