RESUMEN
OBJECTIVES: To identify signs and symptoms to differentiate cardiac from neurological syncope in patients over 70 using a standardized questionnaire. DESIGN: Prospective cohort study. SETTING: Five short-stay units in a French university hospital. PARTICIPANTS: One hundred thirty-one in-patients with syncope aged 70 and older. MEASUREMENTS: Patients were interviewed about the signs and symptoms that had been present before, during or after syncope. When possible, a witness who had been present during syncope was also interviewed to compare theirs and the patients' answers. The sensitivity and specificity of 35 questions were calculated among 3 groups defined according to the cause of syncope: cardiac (n = 58), neurological (n = 31) and syncope of unknown origin (n = 42). Statistical analyses were performed to determine discriminating signs and symptoms among the causes and crude agreement was calculated for answers from patients and witnesses. RESULTS: Only 8 and 3 of 35 questions had a sensitivity of at least 0.5 for cardiac and neurological causes respectively. A feeling of impending syncope, thoracic oppression, recall of events preceding syncope and a history of arrhythmia were independently and significantly discriminant among groups. Recall of events preceding syncope (Odds Ratio (OR) = 7.5; 95% confidence interval (CI) = 2.2-25.3) and a personal history of arrhythmia (OR = 4.8; 95% CI = 1.6-14.2) were discriminant between cardiac and neurological causes suggesting mostly a cardiac cause. Agreement between patients and witnesses was only found for questions on the patient's medical history or the circumstances surrounding the onset of syncope. CONCLUSIONS: Recall of events preceding syncope and a history of arrhythmia are strongly suggestive of a cardiac rather than a neurological cause of syncope. Interviews of witnesses are not helpful in suggesting a cause for syncope.