Table 1 Emergency department syncope studies
No | Year | Study | Variables | Scoring system | Endpoints | Results1 | Strengths | Weakness |
|---|---|---|---|---|---|---|---|---|
1 | 1997 | Martin et al. | • Abnormal ECG | 0 to 4 | 1-year arrhythmias or deaths | 4.4% score 0 | One of the earliest studies | Only long-term outcomes |
• History of ventricular arrhythmia | (1 point for each item) | |||||||
• History of CHF | 57.6% score 3 or 4 | Not validated | ||||||
• Age >45 years | ||||||||
2 | 2002 | OESIL | • Abnormal ECG | 0 to 4 | 1-year mortality | 0% score 0 | Externally validated for | Only long-term outcomes |
• History of cardiovascular disease | (1 point for each item) | 0.6% score | ||||||
• Lack of prodrome | 14% score 2 | up to 6 month outcomes | Modest performance for outcomes up to 6 months | |||||
• Age >65 years | ||||||||
29% score 3 | ||||||||
53% score 4 | ||||||||
3 | 2003 | Sarasin et al. | • Age >65 years | 0 to 3 | Arrhythmias in unexplained ED syncope | 2% score 0 | Studied arrhythmia risk in unexplained syncope | Only inpatients |
• History of CHF | (1 point for each item) | 17% score 1 | Internal validation on historical cohort | |||||
Abnormal ECG | 35% score 2 | |||||||
27% score 3 | ||||||||
No external validation | ||||||||
4 | 2004 | San Francisco Syncope Rule | • Abnormal ECG | No item = No risk | 7-day serious events | Sensitivity 98% | First tool for short-term events | Wide variations in performance |
• History of CHF | ||||||||
• Shortness of breath | ||||||||
• Hematocrit < 30% | ≥ 1 item = risk | Specificity 56% | Most widely validated | ECG variable too broad | ||||
• Triage systolic BP <90 mmHg | ||||||||
Included soft outcomes2 | ||||||||
5 | 2007 | Boston Syncope Rule | • Compilation of 25 plausible variables | ≥ 1 item = risk | 30-day serious events | Sensitivity 97% | A thorough list of variables | No statistical methods |
Specificity 62% | Not practical | |||||||
No external validation | ||||||||
6 | 2008 | STePS | • Abnormal ECG | ≥ 1 item = risk | 10-day and 1-year events | Not Reported | Addresses the role of admissions to hospital | Readmission to hospital was an outcome |
• Trauma | ||||||||
• No prodrome | ||||||||
• Male sex | ||||||||
Not validated | ||||||||
7 | 2008 | EGSYS | • Palpitations before syncope (+4) | Addition of all items | Cardiac syncope probability | 2% score <3 | First study to incorporate variables from history | Not generalizable - Syncope expert always available |
• Abnormal ECG and/or heart disease (+3) | ||||||||
• Syncope during effort (+3) | 2-year total mortality | 13% score 3 | ||||||
• Syncope while supine (+2) | 33% score 4 | |||||||
77% score >4 | Internal validation 92% sensitivity | |||||||
2% score <3 | ||||||||
21% score ≥3 | No robust external validation | |||||||
• Autonomic prodrome (−1) | ||||||||
8 | 2009 | Sun et al. | • Age >90 years (+1) | Addition of all items | 30-day events among older (≥ 60 years) syncope patients | 2.5% score −1, 0 | First study to risk stratify older patients | Retrospective |
• Male sex (+1) | ||||||||
• History of arrhythmia (+1) | 6.3% score 1,2 | Can be applied only to older patients | ||||||
• Triage systolic BP >160 (+1) | Large sample size | |||||||
• Abnormal ECG (+1) | ||||||||
• Abnormal troponin I (+1) | 20% score 3 to 6 | Not validated | ||||||
• Near-syncope (−1) | ||||||||
9 | 2010 | ROSE | • BNP level ≥300 pg/ml | Presence of any item | 1-month serious events | Sensitivity 87% | First study to evaluate the role of BNP in risk stratification | Short-term events included stroke |
• Bradycardia ≤50 in ED/pre-hospital | ||||||||
• Positive fecal occult blood on rectal | Specificity 66% | |||||||
• Anemia – Hemoglobin ≤ 90 g/L | Requires BNP testing that is not widely available | |||||||
• Chest pain with syncope | ||||||||
• Q wave on ECG (except in lead III) | ||||||||
• O2 saturation ≤ 94% on room air | Less than ideal sensitivity |