Prediction of acute neurovascular syndromes with prehospital clinical features witnessed by bystanders.

Background: The prompt recognition of an acute neurovascular syndrome by the patient or a bystander witnessing the event can directly influence outcome. We aimed to study the predictive value of the medical history and clinical features recognized by the patients' bystanders to preclassify acute stroke syndromes in prehospital settings.

Methods: We analyzed 369 patients: 209 (56.6%) with acute ischemic stroke (AIS), 107 (29.0%) with intracerebral hemorrhage (ICH), and 53 (14.4%) with subarachnoid hemorrhage (SAH). All patients had neuroimaging as diagnostic gold standard. We constructed clinical prediction rules (CPRs) with features recognized by the bystanders witnessing the stroke onset to classify the acute neurovascular syndromes before final arrival to the emergency room (ER).

Results: In all, 83.2% cases were referred from other centers, and only 16.8% (17.2% in AIS, 15% in ICH, and 18.9% in SAH) had direct ER arrival. The time to first assessment in ≤ 3 h occurred in 72.4% (73.7%, 73.8%, and 64.2%, respectively), and final ER arrival in ≤ 3 h occurred in 26.8% (32.1%, 15.9%, and 28.3%, respectively). Clinical features referred by witnesses had low positive predictive values (PPVs) for stroke type prediction. Language or speech disorder + focal motor deficit showed 63.3% PPV, and 77.0% negative predictive value (NPV) for predicting AIS. Focal motor deficit + history of hypertension had 35.9% PPV and 78.8% NPV for ICH. Headache alone had 27.9% PPV and 95.3% NPV for SAH. In multivariate analyses, seizures, focal motor deficit, and hypertension increased the probability of a time to first assessment in ≤ 3 h, while obesity was inversely associated. Final ER arrival was determined by age and a direct ER arrival without previous referrals.

Conclusion: CPRs constructed with the witnesses' narrative had only adequate NPVs in the prehospital classification of acute neurovascular syndromes, before neuroimaging confirmation.