Did This Patient Have Cardiac Syncope?

Importance: Syncope can result from a reduction in cardiac output from serious cardiac conditions, such as arrhythmias or structural heart disease (cardiac syncope), or other causes, such as vasovagal syncope or orthostatic hypotension. Objective: To perform a systematic review of studies of the accuracy of the clinical examination for identifying patients with cardiac syncope. Study Selection: Studies of adults presenting to primary care, emergency departments, or referred to specialty clinics. Data Extraction and Synthesis: Relevant data were abstracted from articles in databases through April 9, 2019, and methodologic quality was assessed. Included studies had an independent comparison to a reference standard. Main Outcomes and Measures: Sensitivity, specificity, and likelihood ratios (LRs). Results: Eleven studies of cardiac syncope (N = 4317) were included. Age at first syncope of at least 35 years was associated with greater likelihood of cardiac syncope (n = 323; sensitivity, 91% 95% CI, 85%-97%; specificity, 72% 95% CI, 66%-78%; LR, 3.3 95% CI, 2.6-4.1), while age younger than 35 years was associated with a lower likelihood (LR, 0.13 95% CI, 0.06-0.25). A history of atrial fibrillation or flutter (n = 323; sensitivity, 13% 95% CI, 6%-20%; specificity, 98% 95% CI, 96%-100%; LR, 7.3 95% CI, 2.4-22), or known severe structural heart disease (n = 222; range of sensitivity, 35%-51%, range of specificity, 84%-93%; range of LR, 3.3-4.8; 2 studies) were associated with greater likelihood of cardiac syncope. Symptoms prior to syncope that were associated with lower likelihood of cardiac syncope were mood change or prodromal preoccupation with details (n = 323; sensitivity, 2% 95% CI, 0%-5%; specificity, 76% 95% CI, 71%-81%; LR, 0.09 95% CI, 0.02-0.38), feeling cold (n = 412; sensitivity, 2% 95% CI, 0%-5%; specificity, 89% 95% CI, 85%-93%; LR, 0.16 95% CI, 0.06-0.64), or headache (n = 323; sensitivity, 3% 95% CI, 0%-7%; specificity, 80% 95% CI, 75%-85%; LR, 0.17 95% CI, 0.06-0.55). Cyanosis witnessed during the episode was associated with higher likelihood of cardiac syncope (n = 323; sensitivity, 8% 95% CI, 2%-14%; specificity, 99% 95% CI, 98%-100%; LR, 6.2 95% CI, 1.6-24). Mood changes after syncope (n = 323; sensitivity, 3% 95% CI, 0%-7%; specificity, 83% 95% CI, 78%-88%; LR, 0.21 95% CI, 0.06-0.65) and inability to remember behavior prior to syncope (n = 323; sensitivity, 5% 95% CI, 0%-9%; specificity, 82% 95% CI, 77%-87%; LR, 0.25, 95% CI, 0.09-0.69) were associated with lower likelihood of cardiac syncope. Two studies prospectively validated the accuracy of the multivariable Evaluation of Guidelines in Syncope Study (EGSYS) score, which is based on 6 clinical variables. An EGSYS score of less than 3 was associated with lower likelihood of cardiac syncope (n = 456; range of sensitivity, 89%-91%, range of specificity, 69%-73%; range of LR, 0.12-0.17; 2 studies). Cardiac biomarkers show promising diagnostic accuracy for cardiac syncope, but diagnostic thresholds require validation. Conclusions and Relevance: The clinical examination, including the electrocardiogram as part of multivariable scores, can accurately identify patients with and without cardiac syncope.