Things We Do For No Reason: Echocardiogram in Unselected Patients with Syncope

Mendu et al.⁵ performed a single-center, retrospective study of the diagnostic yield of testing for syncope in 2106 consecutive patients older than 65 admitted over the course of 5 years. They retrospectively applied the San Francisco Syncope Rule (SFSR), which patients met if they had congestive heart failure, hematocrit <30%, abnormal ECG, shortness of breath, or systolic blood pressure <90 mm Hg. There were 821 patients (39%) who underwent echocardiogram. Among the 488 with no SFSR criteria, 10 patients (2%) had echocardiogram results that affected management, and 4 patients (1%) had results that helped determine the etiology of syncope.

Anderson et al. studied 323 syncope patients in a single ED observation unit over 18 months.⁶ Patients with high-risk features, including unstable vital signs, abnormal cardiac biomarkers, or ischemic ECG changes, were excluded from the unit. The initial ECG was considered abnormal if it contained arrhythmia, premature atrial or ventricular contractions, pacing, second- or third-degree heart block, or left bundle branch block. Of the 235 patients with a normal ECG who underwent echocardiogram, none had an abnormal study.

Chang et al.⁷ performed a retrospective review of 468 patients admitted with syncope at a single hospital. Charts were reviewed for ECG and echocardiogram results. Abnormal ECGs were defined as those containing arrhythmias, Q waves, ischemic changes, second- and third-degree heart block, paced rhythm, corrected QT interval (QTc) >500 ms, left bundle branch or bifasicular block, Brugada pattern, or abnormal axis. Among 321 patients with normal ECGs, echocardiograms were performed in 192. Eleven of those echocardiograms were abnormal: 3 demonstrated aortic stenosis in patients who already carried the diagnosis, and the other 8 abnormal echocardiograms revealed unexpected left ventricular ejection fractions <45% or other nonaortic valvular pathology. None of the findings were felt to be the cause of syncope.

Han et al.⁸ performed a retrospective cohort study of all syncope patients presenting to a single ED over the course of 1 year. Patients were stratified as high risk if they had chest pain, palpitations, a history of cardiac disease (defined as prior arrhythmia, heart failure, coronary artery disease, or structural heart disease), abnormal cardiac biomarkers, or an abnormal ECG (defined as sinus bradycardia, arrhythmia, premature beats, second- or third-degree heart block, ventricular hypertrophy, ischemic Q or ST changes, or abnormal QT interval). Patients with none of those symptoms or findings were considered low risk. Of those categorized as low risk (n = 115), 47 underwent echocardiogram, only 1 of which was abnormal.

Across studies, the percentage of patients with a normal cardiac history, examination, and ECG with new, significant abnormalities on echocardiogram was 0% in 3 studies (n = 340),^4,6,15 2% in 1 study (10/488 patients),⁵ 2.1% in 1 study (1/47 patients),⁸ and 4.2% in 1 study (8/192 patients).⁷ The 11 echocardiograms with significant findings in the studies by Mendu et al.⁵ and Han et al.⁸ were not further described. The 8 patients with abnormal echocardiograms reported by Chang et al.⁷ had depressed left ventricular ejection fraction or nonaortic valvular disease that did not represent a definitive etiology of their syncope. Given the cost of $1,000 to $2,220 per study,¹⁶ routine echocardiograms in patients with a normal history, examination, and ECG would thus require $60,000 to $132,000 in spending to find 1 new significant abnormality, which may be unrelated to the actual cause of syncope.

The diagnostic yield of echocardiogram is higher in patients with a positive cardiac history or abnormal ECG. In the prospective study by Sarasin et al.¹⁵ a total of 27% of patients with a positive cardiac history or abnormal ECG were found to have an ejection fraction less than or equal to 40%. Other studies reporting percentages of abnormal echocardiograms in patients with abnormal history, ECG, or examination found rates of 8% (26/333),⁵ 20% (7/35),⁶ 28% (27/97),⁸ and 29% (27/93).⁷ It should be noted that not all of these abnormalities were felt to be the cause of syncope. For example, Sarasin et al.¹⁵ reported that only half of the patients with newly identified depressed ejection fraction were diagnosed with arrhythmia-related syncope. Chang et al⁷ reported that 6 of the 27 patients (22%) with abnormal ECG and echocardiogram had the cause of syncope established by echocardiogram.