The most award winning
healthcare information source.
TRUSTED FOR FOUR DECADES.
Abstract & Commentary
Synopsis: A good history and physical examination, plus basic laboratory, ECG, carotid massage, and testing for hypotension can identify 3 of every 4 causes of syncope.
Source: Sarasin FP, et al. Am J Med. 2001;111:177-184.
There is no "gold standard" to diagnose syncope, since it is a symptom of several diseases. This study set out to measure the diagnostic yield of a pre-defined, sequential syncope protocol in a community-based group of patients.
Over a 21-month period, Sarasin and colleagues evaluated and prospectively followed all 788 patients who presented to the emergency department of the major primary and tertiary care hospital in their region with syncope. After excluding 138 patients who either did not complete the protocol (115) or who refused to participate (23), they enrolled 650 patients in the study. They defined syncope as "a sudden, transient loss of consciousness with an inability to maintain postural tone and spontaneous recovery." The patients ranged in age from 18 to 93 years old (mean age 60). Males made up 48% of the group. Thirty-six percent of these patients had known comorbid conditions including coronary artery disease, previous myocardial infarction, heart failure, hypertension, diabetes mellitus, peripheral vascular disease, and chronic obstructive pulmonary disease. All patients had an initial evaluation consisting of a history and physical examination using a standardized protocol, plus hematocrit, serum creatine kinase, serum glucose, electrocardiogram (ECG), carotid massage, and orthostatic blood pressures and pulses. Afterwards, they were divided into 3 groups: those in whom a diagnosis was strongly suspected (446); those in whom a diagnosis was suspected, but needed confirmation (67); and those in whom the diagnosis was unclear (137).
Sarasin et al used predefined criteria to assign the causes of syncope, including attempts to reproduce the symptoms when possible. The causes of syncope were divided into 2 groups, cardiac and noncardiac.
In the first group, history and physical examination yielded a diagnosis in 245. The most common abnormality was vasovagal syncope (37%). Orthostatic hypotension represented almost one quarter of the diagnoses. These patients had drug-related hypotension (most often associated with angiotensin-converting enzyme [ACE] inhibitors), hypovolemia, postprandial hypotension, or idiopathic hypotension. Laboratory evaluation picked up gastrointestinal hemorrhage in 2 patients and hypoglycemia in 3.
In the second group, a cause of syncope was confirmed by selected diagnostic testing in 49. The suspected causes were seizure, stroke/TIA, pulmonary embolism, aortic stenosis, arrhythmia, mastocytosis, and subdural hematoma. The diagnostic procedures used were electroencephalogram (EEG), computed tomography (CT), lower limb venous compression ultrasound, plasma D-dimer quantification, lung scan, echocardiogram, Holter monitor, and consultation. The number of patients with a confirmed diagnosis then was 495.
When the 18 patients without a diagnosis in group 2 were added to the third group, there were 155 patients with no cause for syncope after initial clinical evaluation. Of these, 33 pursued no further investigations, either because the patients refused or were in poor health or for logistical reasons. Of the remaining 122 patients, a diagnosis was confirmed in 30 after extensive work-up. All 30 had abnormal baseline ECGs. The work-up included 24-hour Holter monitoring, ambulatory loop ECG recording, echocardiography, tilt-table testing, and electrophysiological (EP) studies. Interestingly, although the echocardiograms revealed abnormalities in some patients, none of the abnormalities were considered pertinent to syncope. Holter monitoring, ambulatory loop recording, tilt-table testing, and EP testing helped establish diagnoses in 9, 3, 11, and 7 patients, respectively. This increased the number of patients with a diagnosis to 525 (see Table). These patients were followed for 18 months, during which 55 died. Patients with a cardiac cause of syncope were more likely to die. Also during this period, 95 patients (15%) had 1 or more recurrences of syncope.
Comment by Allan J. Wilke, MD
Syncope, a short-lived loss of consciousness, accounts for 3% of all emergency room visits (1.1% in this study). It is the result of a bewildering and disparate number of diseases, some of them potentially life threatening, some of them with specific therapies. Its transitory nature, while ultimately a blessing for the patient, makes diagnosis difficult. It is not amenable to our calls to "come back when you have it again." Unless your patient is attached to telemetry or some cardiac event monitor when syncope occurs, the diagnosis is very much an educated guess, but that’s okay. This study’s value is in its reassurance that we can come to a diagnosis, more often than not, with tools that are readily available to anyone practicing adult medicine.
The study does raise some methodological questions. First, this study was done in Geneva, Switzerland. Does Sarasin et al’s study group resemble the patients I see with syncope? Maybe. Although they do not provide ethnic or racial breakdowns, I assume that their patients were predominantly Swiss and probably overwhelmingly Caucasian. Second, is their definition of syncope too broad? For instance, a seizure might look like "a sudden, transient loss of consciousness with an inability to maintain postural tone and spontaneous recovery." However, I would argue that by keeping the definition broad and then assigning a diagnosis (such as seizure) keeps in perspective the reality that, at first, we do not always know what disease has caused the symptom.
The diagnostic yield of an extensive workup (30/155 or 19%) is disappointing. I would like to see a cost analysis of this. It may make sense to limit an extensive work-up to that patient who has an abnormal ECG at baseline.
Dr. Wilke is an Assistant Professor of Family Medicine, Medical College of Ohio, Toledo, Ohio.