Response to Letter Regarding Article, “Cardiac Magnetic Resonance With T2-Weighted Imaging Improves Detection of Patients With Acute Coronary Syndrome in the Emergency Department”
We read with interest the letter from Dr Bouzas-Mosquera about our published article.1 We appreciate both his comments and the opportunity to respond.
We used a subjective assessment of the pretest probability of acute coronary syndrome (ACS) by the emergency department (ED) caregiver because it incorporates into a single metric all of the data available to decision makers before imaging. Its limitation is that it is too insensitive to use to discharge ED patients directly home without further testing for ACS. It was used because no well established, universally accepted, and validated risk score exists that can be applied to undifferentiated ED populations with chest pain. Available prediction scores that have been described (eg, Goldman score, Sanchis score, TIMI risk score, and Vancouver rules) have usually been derived from non–ED populations or have not been validated and confirmed to have a very high sensitivity. Our subjective estimate reflects the reality of the ED and the challenge that the emergency physician faces to assess and to risk stratify these patients.
The reported P value was based on the 2log likelihood ratio test with 1 degree of freedom to test the difference between the baseline model and the nested logistic regression models containing cardiac magnetic resonance. It describes the statistical significance of this improved model with cardiac magnetic resonance as compared with the baseline model. Our aim was to provide the reader with an estimate of the improvement of the overall-model predictive value and whether this improvement was statistically significant. Ninety-five percent confidence intervals can provide additional information (area under the receiver-operating-characteristic curve for the ED probability estimates: 0.70, 95% confidence interval 0.54 to 0.85; area under the receiver operating characteristic curve for clinical risk factors, 0.77; 95% confidence interval 0.62 to 0.92; area under the receiver operating characteristic curve for adding magnetic resonance imaging to the model containing the clinical risk factors: 0.96, 95% confidence interval 0.91 to 1.00).
The authors raise an interesting question of whether the presentation of rare events with infinite solutions may be trustworthy. We want to emphasize that the event (ACS) is not rare (13/62), and thus the use of the binominal distribution is appropriate. Also, given the somewhat limited number of ACS cases, we minimized the risk for over fitting the model by combining the number of risk factors as a variable, which may have occurred if additional covariates were considered. Mathematically, this almost perfect classification naturally results in extremely wide confidence intervals. Note that the lower limit of this 95% confidence interval is 11 and that likelihood ratios over 10 are usually considered clinically meaningful. Thus, these confidence intervals represent a best estimate of how cardiac magnetic resonance can improve the predictive value of the model.
Finally, our results confirm prior observations from different groups.2–4 Having a noninvasive imaging test that can improve the diagnostic confidence of the emergency physician by giving insights at the myocardium level and making it possible to detect myocardial edema and myocardial necrosis, to differentiate acute versus old infarct, and even to differentiate the subsets of ACS (unstable angina pectoris versus non–ST-segment elevation myocardial infarction) will be important in the accurate triage of patients with acute chest pain in the ED.
Cury RC, Shash K, Nagurney JT, Rosito G, Shapiro MD, Nomura CH, Abbara S, Bamberg F, Ferencik M, Schmidt EJ, Brown DF, Hoffmann U, Brady TJ. Cardiac magnetic resonance with T2-weighted imaging improves detection of patients with acute coronary syndrome in the emergency department. Circulation. 2008; 118: 837–844.
Kwong RY, Schussheim AE, Rekhraj S, Aletras AH, Geller N, Davis J, Christian TF, Balaban RS, Arai AE. Detecting acute coronary syndrome in the emergency department with cardiac magnetic resonance imaging. Circulation. 2003; 107: 531–537.