In reply
Article Outline
We thank Drs. Quinn and McDermott for their interest in our work.1 In an external validation study, we found that the San Francisco Syncope Rule demonstrated lower sensitivity (89%; 95% confidence intervals: 81%, 97%) than reported by the San Francisco Syncope Rule investigators in their derivation2 and validation studies.3 Our findings are consistent with other published4 and unpublished reports.5, 6, 7 Drs. Quinn and McDermott point to several differences between the original San Francisco Syncope Rule investigations and our study; we believe that these differences are minor and unlikely to qualitatively change our findings.
First, Drs. Quinn and McDermott are concerned that our definition of syncope may have resulted in the enrollment of patients with persistent altered mental status. In our study, we excluded all patients with an abnormal mental status, including patients with baseline cognitive deficits. Thus, we used a definition that was more conservative than the original San Francisco Syncope Rule studies to exclude patients with potential neurologic conditions. Our frequency of patients with stroke/transient ischemic attack (0.4%) is comparable to the San Francisco Syncope Rule derivation (0.4%) and validation studies (0.4%). Both patients with stroke/transient ischemic attack in our study complained of vertigo/unsteady gait in tandem with syncope, and both were documented to have a normal mental status by both the emergency and admitting physicians.
Second, Drs. Quinn and McDermott are concerned that the 3 San Francisco Syncope Rule negative patients in our study who experienced an arrhythmia may not have had a clinically important event. On a 3 physician panel review, all 3 patients had explicit documentation of an arrhythmia on inpatient cardiac monitoring (2 patients experienced ventricular arrhythmia; the third experienced symptomatic paroxysmal supraventricular tachycardia). One patient underwent electrophysiology testing, which did not reveal inducible ventricular tachycardia. A second patient was felt to be a poor automatic implantable cardioverter defibrillator candidate given advanced age and multiple co-morbidities. The final patient required adenosine administration to terminate a symptomatic paroxysmal supraventricular tachycardia.
Finally, Drs. Quinn and McDermott point out that the definition of an “abnormal” ECG was different from the San Francisco Sycope investigations. We used explicit definitions to help clinicians categorize ECGs as normal (including 1st degree block and premature atrial contractions), non-specific ST-T changes, and abnormal (including abnormal conduction intervals). We regarded any abnormalities and non-specific ST-T changes, regardless of whether these changes were old or new, to be positive by the San Francisco Syncope Rule. This conservative definition is likely to upwardly bias estimates of sensitivity compared to the unstructured ECG assessments used by the San Francisco Syncope Rule investigators. On retrospective review of the 4 patients who were classified as San Francisco Syncope Rule negative by the emergency physician but who experienced a cardiac event, the cardiology overread of the ECG was normal in 3 patients and abnormal in 1 patient. If the cardiologist’s ECG interpretation had been used for San Francisco Syncope Rule classification for these 4 patients, the observed San Francisco Syncope Rule sensitivity in our study would have improved to 91% (95%CI: 84%, 99%). This sensitivity is still too low to justify routine application of the San Francisco Syncope Rule and comes at the cost of decreased specificity that will occur as more ECGs are labeled as “abnormal” and the number of false-positive cases increases.
In summary, the issues raised by Drs. Quinn and McDermott are unlikely to have an important effect on our findings. Our results were also robust to multiple sensitivity analyses to assess the effects of missing data, missing follow-up, and experience of the treating physician.
While it is impossible to argue with Drs. Quinn and McDermott’s suggestion that clinicians should not wait “to improve their decisionmaking,” we continue to urge caution regarding widespread application of the San Francisco Syncope Rule, given the lower sensitivities reported by ourselves and others. The San Francisco Syncope Rule was derived on a cohort containing only 79 serious events (including conditions diagnosed during the emergency department visit), and we are concerned about the stability of the San Francisco Syncope Rule in other populations. Although the San Francisco Syncope Rule is a serious contribution to syncope research, further large cohort research analyzing hundreds of delayed, serious clinical events will be required to generate a robust decision instrument.
References
- External validation of the San Francisco Syncope Rule. Ann Emerg Med. 2007;49:420–427
- Derivation of the San Francisco Syncope Rule to predict patients with short-term serious outcomes. Ann Emerg Med. 2004;43:224–232
- Prospective validation of the San Francisco Syncope Rule to predict patients with serious outcomes. Ann Emerg Med. 2006;47:448–454
- . External validation of the San Francisco Syncope Rule in the Australian context. CJEM. 2007;9:157–161
- External validation of the San Francisco Syncope Rule [abstract]. Acad Emerg Med. 2005;12(5 Supplement 1):11
- Can the San Francisco Syncope Rule be safely applied to patients aged 65 years and older who present to the emergency department with syncope or near-syncope?. [abstract] Acad Emerg Med. 2006;13:S69
- Failure to validate the San Francisco Syncope Rule in an independent ED population [abstract]. Acad Emerg Med. 2007;14:S162;-a
PII: S0196-0644(07)01630-7
doi:10.1016/j.annemergmed.2007.10.002
© 2007 American College of Emergency Physicians. Published by Elsevier Inc. All rights reserved.
Refers to article:
- External Validation of the San Francisco Syncope Rule
