sinaiem.org

This month's Annals (May 2006) has an article from Quinn et al. (QuinnAEM-5-06) purporting to validate the San Fransisco Syncope Rule. Also known as "CHESS", the "rule" is positive if any one of the following is true:

• CHF in the past medical history
• Hematocrit < 30
• EKG shows nonsinus rhythm or new changes
• Shortness of Breath
• Systolic < 90 mmHg

In a prospective cohort series, the rule was 98% sensitive (95%CI = 89-100%), 56% specific in predicting "adverse events" within 30 days. Basically, scoring a negative on the San Fransisco rule can make you feel better about sending a patient home. Though, let's be honest, if the test is really 89% sensitive (as the confidence interval allows) than this rule creates an "unnaceptably high" rate of serious outcomes. More below!

Some notes: This rule was generated after following 50 predictor variables, starting in the year 2000, and this study follows up on Quinn's derivation (Annals, February 2004) which showed 96% sensitivity, but the CI was better: 92-100%.

Any patient with trauma, EtOH, drug use or known seizures was excluded. Syncope was defined as transient LOC with return to baseline.  Docs saw the patient, made their decisions, and only then filled out a web-form describing the patient and whether they fit the rule. After that, the investigators followed the patient (and also made sure the rule/form was interpreted correctly — 5% of the time, it wasn't).

Short-term serious outcomes were defined as death, MI, PE, arrhythmia, CVA, SAH, hemorrhage requiring transfusion, any return visit, hospitalization, or procedure related to syncope (if they went home, and came back within 30 days for the same complaint, it was called serious only if they were admitted the second time).

They followed 760 patients (791 visits) and saw serious outcomes in 108 patients — 13.7%. Half (54) were having or already had a serious outcome (ie, they had a new arrythmia or stroke) and half experienced the serious outcome within 30 days of presenting to the ED. Most of the bad outcomes were arrhythmias (23/54) , and most of those were brady or sick sinus syndromes. 59% (469) of the patients presenting with syncope were admitted. 

If docs had applied the San Fransisco rule to all those who presented with syncope, they would have only admitted 52% — saving 7%… And, of interest, if docs had applied the rule only to people they were planning to admit, they would have admitted 24% fewer of them.

In their study, the rule missed only one serious outcome (within 30 days) – a diabetic guy who syncopized and was eventually diagnosed with vertebral artery TIA (the ED doc admitted him anyway, his narrowing was found and he got stented).

Other thoughts:

Martin (Annals, 1997) was the first to try to risk-stratify syncope, looking at death within 1 year. There are others, but Quinn et al say one year is too long to consider when making decisions on admission.

ACEP recommendations on syncope:

• From the history: if the patient is > 60 years old and has CAD, consider it high-risk. If the patient is < 45 and no CAD, consider it low-risk (these are both Class B). Also, if it sounds like Vasovagal, it's low risk (class C recommendation).
• From the physical, consifer the patient high-risk if they have CHF (class B) or cardiac outflow obstruction (class C).
• Admit if the patient has signs or history of CHF, ventricular arrythmias, valvular heart disease, symptoms of chest pain or ACS, or with an EKG showing long QT, BBB, ischemia or arrhythmia (all level B). They go on to say (Level C recommendations) to admit any syncope patient over 60, with a family history of sudden/unexplained death, a history of CAD or congenital heart disease, or exertional syncope in younger patients without obvious cause.

EMCrit has an exhaustive differential in syncope, as well as some cited papers on risk stratification for syncope (Martin 1997 in Annals, Colivicchi in Eur Heart J 2003) and this great nugget: All patients with positive CTs in one study had a witnessed seizure or an alteration of their neurologic exam (Ann Intern Med 1997 Jun 15;126(12):989-996)

Following Quinn's article, the issue of Annals ran an editorial by Chadwick Miller (Wake Forrest), who said we shouldn't over-apply this rule, we already have good vigilance about syncope. This study was just one center, and as it stands, it's just not sensitive enough, the CI needs to come down through a large multicenter trial.

Posted on Friday, May 26th, 2006 at 10:54 pm by Nick. Filed under Risk Stratification, Syncope, Journal Club.
You may post a comment.