brand logo

Am Fam Physician. 2009;80(6):558-559

Author disclosure: Nothing to disclose.


Each month, three presenters review an interesting journal article in a conversational manner. These articles involve “hot topics” that affect family physicians or “bust” commonly held medical myths. The presenters give their opinions about the clinical value of the individual study discussed. The opinions reflect the views of the presenters, not those of AFP or the AAFP.

This Month's Article

Birnbaum A, Esses D, Bijur P, Wollowitz A, Gallagher EJ. Failure to validate the San Francisco Syncope Rule in an independent emergency department population. Ann Emerg Med. 2008;52(2):151–159.

Can the San Francisco Syncope Rule successfully determine which patients can be safely discharged from the ED after an acute syncopal episode?

Bob: Family physicians are commonly asked to evaluate patients who have had a transient loss of consciousness. Estimates suggest that 1 to 3 percent of emergency department (ED) visits are for syncope.1 Although these episodes are often benign in origin, they can be a signal of an underlying pathologic process associated with significant morbidity and mortality (e.g., malignant cardiac arrhythmias). The family physician often has to decide whether to hospitalize these patients for further evaluation or discharge them.

The San Francisco Syncope Rule has been proposed as a simple five-step clinical decision rule (CDR) that identifies ED patients with syncope or near-syncope who are at low risk of short-term serious outcomes. The authors derived this rule from a large database of patients who presented to San Francisco General Hospital. They noted that patients with a syncopal episode were at low risk of an adverse outcome if none of the following were present: history of congestive heart failure; hematocrit less than 30 percent; abnormal ECG (defined as new changes or nonsinus rhythm); shortness of breath; or systolic blood pressure less than 90 mm Hg at triage.2

Conversely, if any of these conditions were present, there was a 62 percent chance that a serious underlying condition would subsequently be recognized.2 In 2006, the same authors prospectively validated their rule in 791 consecutive patients who presented to the ED at the University of California, San Francisco after a syncopal episode. They reported that the rule detected 98 percent of the patients with serious underlying conditions. Further, they reported that application of the rule would have prevented a substantial number of admissions.3

What we need to know now is: does this rule have external validity? In other words, can other researchers or physicians come up with similar results outside of the original authors' institution?

What does this article say?

Bob: This study is a prospective, observational study that enrolled patients with an episode of syncope who presented to the ED at Montefiore Medical Center. During the two-year study, the five variables of the San Francisco Syncope Rule were applied to each patient, and each patient was followed for seven days. The decision to admit or discharge the patient was made solely by the ED physician independent of the CDR. A total of 738 patients were evaluated, and 713 were available for follow-up and analysis.

Of the 713 patients, 61 (9 percent) had a serious outcome that was predefined. When the San Francisco Syncope Rule was applied, it did not identify 16 of the 61 serious outcomes (26 percent; 95% CI, 16 to 39). These “misses” included one death, eight arrhythmias, three strokes, and one subarachnoid hemorrhage. The authors concluded that the low sensitivity of the rule limits its generalizability.

Andrea: A sensitivity of 74 percent is quite poor. How is it possible that these authors had such a different result from the original authors, who reported a sensitivity of 98 percent for their CDR?

Bob: One can only speculate. Was there a difference in patient populations? It is interesting to note that the initial derivation and validation studies included children, whereas this external validation study exclusively evaluated adults. Or perhaps it is how the rule is applied in other physicians' hands.

Mark: This is an excellent example of external validity. Whether it's a CDR, drug, or some other intervention, the conclusions of the original authors should always be retested in a group of similar patients, but by other researchers at a different institution, before broadly applying it in clinical settings.

Should we believe this study?

Bob: When validating a CDR, the EBM Working Group proposes that the following standards need to be met: (1) the patients are chosen in an unbiased fashion; (2) the patients represent a wide spectrum of disease; (3) there is a blinded assessment of the criterion standard for all patients; (4) there is an explicit and accurate interpretation of the variables in the rule without the knowledge of the outcome; and (5) there is 100 percent follow-up of enrolled patients.4 It appears that the authors of this study did indeed follow these criteria; the 97 percent follow-up rate appears acceptable.

Mark: When deciding whether or not to use a CDR, the EBM Working Group also suggests reference to a hierarchy of evidence specific to CDRs. Progressively more confidence may be placed in a rule as it moves up the evidence ladder, from a derivation data set only, to a single validation study, to multiple validation studies in different settings (or a large prospective study), and finally, to the highest level where there is at least one prospective validation in a different population and a separate impact analysis on physician behavior. A derivation set of criteria is derived retrospectively and is subject to bias (e.g., confounding variables). Researchers look at patients who did well and who did not do well, and try to find the difference.

Next, the CDR needs to be prospectively validated in a similar population. If the CDR is validated, it is still best to have the rule externally validated by another group of researchers before its widespread use. Because this study cannot replicate the original authors' findings, the widespread use of the CDR (i.e., the San Francisco Syncope Rule) cannot be recommended at this time.

Bob: Interestingly, two other studies in Los Angeles and Australia attempted to validate the San Francisco Syncope Rule. The sensitivity of the syncope rule in these two locations was 89 and 90 percent, respectively.5,6 These values are too low to rely on the rule for discharge decisions.

Andrea: What I found fascinating was the ED physicians' clinical “gestalt.” The ED physicians had sole discretion on admitting or discharging these patients. What was the physicians' sensitivity for identifying a potentially serious event? 100 percent! Although probably an overstatement, the physicians' acumen far exceeded the rule.

Do not take away the wrong message here: the study does not suggest that we admit every patient with syncope. But it does make an excellent point about CDRs, which is that they rarely account for each and every potential variable; therefore, physicians must use their clinical discretion and override the results of a CDR when appropriate.

What should the family physician do?

Bob: Because syncope-related admissions account for $2.4 billion in health care costs in the United States,1 I would be thrilled to find a CDR that reliably distinguishes which patients can be safely discharged from the ED. The San Francisco Syncope Rule by itself is not the answer. Family physicians will do well to take important clinical factors into account, such as those in this syncope rule, but they should not rely on them to the exclusion of their overall judgment. This rule can change your prior probability of disease, but cannot rule serious disease in or out.

Mark: I agree. This rule is “not ready for prime time.”

Andrea: And never discount your clinical “gestalt,” because it's what makes us who we are: doctors.

Main Points

  • External validation of the San Francisco Syncope Rule suggests that the sensitivity of the rule is lower than first reported; therefore, this CDR cannot be generally accepted.

  • Physicians need to apply their clinical acumen and override the results of a CDR when necessary.

EBM Points

  • External validity refers to the ability to generalize the results of a study to other settings.

  • A hierarchy of CDRs exists. No CDR should be widely used until it has been clearly shown to be beneficial in external validity studies.

Continue Reading

More in AFP

More in PubMed

Copyright © 2009 by the American Academy of Family Physicians.

This content is owned by the AAFP. A person viewing it online may make one printout of the material and may use that printout only for his or her personal, non-commercial reference. This material may not otherwise be downloaded, copied, printed, stored, transmitted or reproduced in any medium, whether now known or later invented, except as authorized in writing by the AAFP.  See permissions for copyright questions and/or permission requests.