Journal scan

Missed diagnosis of acute cardiac ischaemia in the emergency department J H Pope, T P Aufderheide, R Ruthazer et al N Engl J Med 2000;342:1163–70 Objectives—To describe the incidence of, factors related to, and clinical outcome of a failure to admit patients with acute cardiac ischaemia. Methods—10 689 patients attending 10 US emergency departments with chest pain or other symptoms suggestive of acute cardiac ischaemia were studied in a multi-centre prospective clinical trial. Patients that were sent home attended for repeat examination, ECG and CK-MB within 72 hours of discharge. Results—There was 99% follow up. A total of 1866 (17%) patients had acute cardiac ischaemia (8% MI, 9% unstable angina). Twenty seven per cent had stable angina or other cardiac problems. Fifty five per cent had non-cardiac pain. Nineteen (2.1%) of the 889 patients with acute MI and 22 (2.3%) of the 966 patients with unstable angina were sent home. Factors associated with mistaken discharge were female sex and age <55, non-white race, shortness of breath as main symptom and a normal or non-diagnostic ECG. Patients with acute MI who were sent home had the same crude mortality rates as those admitted to hospital (home 10.5%, hospital 9.7%) but when these rates were adjusted for various risk factors the mortality ratio was almost doubled (1.90). Those sent home with unstable angina did have both a higher crude mortality rate (home 9.8%, hospital 5.5%) and adjusted mortality ratio (1.7). Conclusions—Few patients are mistakenly discharged with acute cardiac ischaemia but their mortality is higher. Absence of typical symptoms or ECG changes are associated with mistaken discharge. Critique—This is an important problem. This paper aimed to identify the incidence of wrongful discharge along with the factors and consequences associated. The actual mortality rates of those sent home with acute MI and those admitted were similar although the risk adjusted mortality ratios were increased but the increases did not achieve statistical significance. Further explanation is required of the methodology of risk adjusted mortality ratios. Another weakness is that 929 patients were excluded from the study. No reason is given for these exclusions and this throws some doubt on the 99% follow up rate. Although excluded patients matched study patients for sex and race no explanation is given as to why they were omitted. There is no mention of the proportion from the overall population who were admitted. Obviously the lower the threshold for admission, the less likely there is to be an error. The study was carried out over seven months in 1993–4. No reason for the delay in publication is given. Further methods to identify acute ischaemia are now more widely available.