Abstracts

Rationale: Unexplained acute loss of consciousness (i.e., syncope) is a frequent reason for admission to the emergency room (ER). Causes of syncope include orthostasis, pain, neurocardiogenic syncope, cardiac arrhythmias, seizures, endocrine (hypoglycemias, Addison's disease) and psychological disorders. In up to 40%, the etiology cannot be determined. Since the exploration of all potential causes of brief losses of consciousness is neither cost-efficient nor feasible in the ER setting, we aimed to determine clinical factors allowing to distinguish patients with a high risk for underlying seizures to direct them towards rapid seizure work-up. Methods: From 2011 to 2012, all patients admitted to an academic tertiary care center with the ICD-10 code of unexplained syncope without known diagnosis of epilepsy or cardiac arrhythmias who were discharged after workup with a diagnosis of seizure/epilepsy or cardiac disorder were included. Demographics, clinical features, and results from diagnostic procedures (cardiac workup, EEG, neuroimaging, serum biomarkers) were assessed. Stepwise logistic regression using stepwise forward and backward selection were applied to identify variables associated with seizures independent from possible clinical confounders. Collinearity between covariates was assessed. The Hosmer-Lemeshow goodness-of-fit test was used to check the final models. Areas under the receiver operating curve (ROC) were calculated for the final models to evaluate discrimination. Results: Among 396 patients with unexplained syncope, the underlying etiologies were most likely seizures in 208 and cardiac events in 188 patients. Age (OR 0.97, 95%CI 0.96-0.98), presence of a neurological (OR 2.76, 95%CI 1.81-4.21), cardiovascular (OR 0.36, 95%CI 0.24-0.54) or psychiatric disorder (OR 3.09, 95%CI 1.90-5.03), presence of a brain lesion on neuroimaging (OR 2.97, 95%CI 1.49-5.91), and low serum phosphate levels (<0.8 mmol/l) (OR 1.90, 95%CI 1.12-3.24) differed significantly between patients with seizures and those with cardiac events. In multivariable analyses, age (OR 0.96; 95%CI 0.95-0.98), neurological disorders (OR 3.66; 95%CI 2.27-5.91), psychiatric disorders (OR 2.89; 95%CI 1.68-4.96), and brain lesions (OR 4.52; 95%CI 2.06-9.92) remained independently associated with seizures. The area under the ROC-curve of these combined factors yielded was 0.771, supporting their ability to discriminate between seizures and cardiac events. Conclusions: Readily available clinical data (as age, history of neurological or psychiatric disorder, and brain lesions on neuroimaging) may allow for rapid identification of underlying seizures rather than cardiac events among patients presenting in the ER with a brief loss of consciousness and no prior history of seizures or arrhythmias. Our results may provide guidance for establishing diagnostic algorithms to tailor further diagnostic procedures to individual needs. The retrospective nature of this study calls for prospective investigations to strengthen these results before a scoring system can be introduced to clinical practice.