Relevance of early seizures for in-hospital mortality in acute cerebrovascular disease.

BACKGROUND

We studied the influence of early poststroke seizures (within the first 48 hours of onset of a first stroke or transient ischemic attack) on in-hospital mortality in 1,099 consecutive patients collected in a prospective stroke registry.

METHODS

Differences in the frequency of demographic characteristics, clinical events, and outcome between patients with and those without epileptic seizures were assessed. To determine the independent predictive value of early seizures on in-hospital mortality, variables related to vital status at discharge (alive, dead) in the univariate analysis, plus age, were studied in two multiple linear regression models. The first predictive model was based on demographic, anamnestic, and clinical variables with a total of 13 variables, and the second model was based on clinical and neuroimaging variables with a total of 16 variables.

RESULTS

A total of 27 patients (2.5%) had epileptic seizures during the first 48 hours of stroke. Advanced age, confusional syndrome, hemorrhagic stroke, large lesions, involvement of parietal and temporal lobes, and occurrence of neurologic and medical complications were significantly more frequent in seizure patients than in nonseizure patients. Overall in-hospital mortality rate was 33.3% in the seizure group and 14.2% in the nonseizure group (p = 0.02). The presence of early seizures was a significant predictive variable both in the model based on clinical variables (odds ratio [OR], 5.5; 95% confidence interval [CI], 1.81 to 16.72) and in the model based on clinical and neuroimaging variables (OR, 6.1; 95% CI, 2.13 to 17.93).

CONCLUSIONS

Seizures at the onset of a first-ever stroke is an independent prognostic factor for in-hospital mortality. Patients with the highest risk of developing epileptic seizures-aged persons with a large hemorrhagic infarction of a parietal lobe-may be candidates to be treated prophylactically against seizures for a few days.