Out-of-hospital diagnosis of cerebral infarction versus intracranial hemorrhage.

OBJECTIVE

To establish a model based on clinical and anamnestic data easily available in the out-of-hospital setting, which facilitates the differential diagnosis between cerebral infarction and intracranial hemorrhage.

DESIGN

Retrospective study that simulates a prospective approach.

SETTING

Emergency Department of the University Hospital in Vienna, Austria.

PATIENTS AND

PARTICIPANTS

Data of 224 patients with either intracranial hemorrhage or cerebral infarction were prospectively collected. Uni-and multivariate analysis was performed to identify neurological symptoms and anamnestic data, which were associated with either intracranial hemorrhage or cerebral infarction.

MEASUREMENTS AND

RESULTS

Unilateral weakness or sensory loss was observed more frequently in patients with infarction compared to hemorrhage (69.8 % vs 11.9 %, P < 0.001). The frequency of patients with impaired level of consciousness was significantly higher in the hemorrhage group compared to the infarction group (59.3 % vs 3.8 %, P < 0.001). A multivariate logistic regression analysis showed that hypertension (OR = 0.31, 95 % CI = 0.12-0.76, P = 0.01), diabetes (OR = 0.17, 95% CI = 0.04-0.68, P = 0.01), and unilateral weakness or sensory loss (OR = 0.10, 95 % CI = 0.04-0.26, P < 0.001) were significantly associated with cerebral infarction. Impaired level of consciousness was significantly related to hemorrhage (OR = 13.41, 95 % CI = 3.92-45.91, P < 0.001). On the basis of the logistic regression analysis, we generated a scoring system for the out-of-hospital diagnosis between infarction and hemorrhage. The values of the score lay between -3 and +3. The probability of infarction increases when the score becomes negative, and the probability for hemorrhage increases when the score becomes positive.

CONCLUSION

Our model is a useful guideline for the differential diagnosis between cerebral infarction and intracranial hemorrhage in the out-of-hospital setting, as it is based on easily available clinical and anamnestic parameters.