Author + information
- Nadia Bandstein, MD,
- Rickard Ljung, MD, PhD,
- Magnus Johansson, MD, PhD and
- Martin J. Holzmann, MD, PhD∗ ()
- ↵∗Department of Emergency Medicine, Karolinska University Hospital, Huddinge, 14186 Stockholm, Sweden
We would like to express our gratitude to Dr. Liu and colleagues and Dr. Cullen and colleagues for their letters regarding our recently published paper (1). Firstly, we do agree that our findings need to be validated in other settings than a university hospital, with a different diversity of ethnicity, socioeconomic status, and prevalence of cardiovascular disease.
All patients who were included in our study had chest pain, an electrocardiogram (ECG) recorded, and at least 1 troponin level measured (1). To our knowledge, troponins are not used for any other reasons than to confirm or to exclude myocardial ischemia. In addition, all patients had a clinical assessment made, which we believe is common practice. Occasionally, patients were assessed clinically after the troponin level was available, and an explanation for the chest pain other than a myocardial infarction (MI) would lead to a discharge home. We believe that this is in line with how patients with chest pain are assessed in most emergency departments (ED).
Seventy-seven percent of admitted patients went home the same or the next day. Naturally, diagnoses such as pneumonia, pulmonary embolism, or atrial fibrillation may have necessitated longer hospital stays. Our primary aim was not to investigate risk mitigation in admitted versus discharged patients. We believe that exercise tests, stress echocardiograms, or coronary angiograms by themselves have no impact on prognosis. We do acknowledge that there may have been patients discharged who may have had a second troponin >14 ng/l if measured. However, the risk of all-cause mortality was not higher in patients discharged versus admitted, and there were only 2 cardiovascular deaths within 12 months in 8,907 patients with troponins <5 ng/l, which indicates an excellent long-term prognosis.
In a random sample of 100 patients, the mean time to measurement of troponins was 2.5 h. Thus, most patients had their first troponin level evaluated before 3 h, and we do agree that if in doubt whether to admit or to discharge the patient, it may be appropriate to have a second troponin measured.
Fifteen patients with troponins <5 ng/l, and normal ECG developed MI. Of these patients, 1 came back at day 18, and 3 developed ST-segment elevation within 1 h of arrival. Another 3 patients had sinus tachycardia and would not be sent home after assessment, we believe, by any physician. That leaves us with 8 of 1,697 (0.47%) patients, which yields a negative predictive value of 99.5%, for patients admitted.
We believe that our results, together with a robust clinical assessment, is helpful for any ED physician, to decide, if in doubt, which patient with chest pain to admit or to send home.
- American College of Cardiology Foundation