RESUMO
BACKGROUND: Isolated transient vertigo can be the only symptom of posterior circulation ischemia. Thus, it is important to differentiate isolated vertigo of a cerebrovascular origin from that of more benign origins, as patients with cerebral ischemia have a much higher risk for future stroke than do those with 'peripheral' vertigo. The current study aims to identify risk factors for cerebrovascular origin of isolated transient vertigo, and for future cerebrovascular events. METHODS: From the files of 339 outpatients with isolated transient vertigo we extracted history, clinical and technical findings, diagnosis, and follow-up information on subsequent stroke or transient ischemic attack (TIA). Risk factors were analyzed using multivariate regression models (logistic or Cox) and reconfirmed in univariate analyses. RESULTS: On first presentation, 48 (14.2%) patients received the diagnosis 'probable or definite cerebrovascular vertigo'. During follow-up, 41 patients suffered stroke or TIA (event rate 7.9 per 100 person years, 95% confidence interval (CI) 5.5-10.4), 26 in the posterior circulation (event rate 4.8 per 100 person years, 95% CI 3.0-6.7). The diagnosis was not associated with follow-up cerebrovascular events. In multivariate models testing multiple potential determinants, only the presentation mode was consistently associated with the diagnosis and stroke risk: patients who presented because of vertigo (rather than reporting vertigo when they presented for other reasons) had a significantly higher risk for future stroke or TIA (p = 0.028, event rate 13.4 vs. 5.4 per 100 person years) and for future posterior circulation stroke or TIA (p = 0.044, event rate 7.8 vs. 3.5 per 100 person years). CONCLUSIONS: We here report for the first time follow-up stroke rates in patients with transient isolated vertigo. In such patients, the identification of those with cerebrovascular origin remains difficult, and presentation mode was found to be the only consistent risk factor. Confirmation in an independent prospective sample is needed.