Background Patient safety has been gained much more attention in recent years.The authors reviewed patients who had cardiac arrest in the operating rooms undergoing noncardiac surgery between January 1989 and December 2001 at the University of Pittsburgh Medical Center,USA.The main objectives of the study were to determine the incidence of intraoperative cardiac arrest,to identify possible causes of cardiac arrest and to explore amenable modifications.Methods With approval by the University of Pittsburgh Institutional Review Board,patients experienced cardiac arrest during surgery were retrieved from medical records,surgical operation and anesthesia records and pathological reports by searching the Medical Archival Retrieval System (MARS),a hospital electronic searching system.Cases of cardiac arrest were collected over a period of thirteen years from the Pre byteria University Hospital (PUH),USA.Results We found 23 cases of intraoperative cardiac arrests occurred in 218 274 anesthesia cases (1.1 per 10 000).Fourteen patients (60.8%) died in the operating room,leading to a mortality rate from all causes of 0.64 per 10 000 anesthetics.Immediate overall survival rate after arrest was 39% (9/23).Half of the patients (12/23) were emergency cases with 41% survival rate (5/12).One fourth of the arrests were trauma patients (6/23).Most arrest patients (87%,20/23) were American Society of Anesthesiologists Physical Status (ASA PS) Ⅳ and Ⅴ,while only three patients were ASA PS-Ⅰ,Ⅱ and Ⅲ,respectively.One case was attributable to an anesthesia-related cardiac arrest and recovered after successful resuscitation.Conclusions Most intraoperative cardiac arrests were not due to anesthesia-related causes.Anesthesia-related cardiac arrests might have a higher survival rate when compared to other possible causes of cardiac arrest in the operating room.