We aimed to determine the demographic, clinical and investigatory factors associated with seizures in encephalitis and develop a risk-stratification score.

We performed an analysis of 203 patients from 24 English hospitals (2005-2008) (Cohort 1). Outcome measures were seizures, inpatient seizures, and status epilepticus. A binary logistic regression risk model was converted to a clinical score and independently validated on 233 patients in the Enceph-UK study (2013-2016) (Cohort 2).

In Cohort 1, 121 (60%) patients had a seizure including 103 (51%) with witnessed inpatient seizures. Seizures were associated with fever (OR [95%CI] 2.21[1.13-4.34], p=0.019) and aetiology (p=0.003). Admission Glasgow coma scale (GCS) ≤8/15 was predictive of subsequent inpatient seizures (OR 5.55[2.10-14.64], p<0.001), including in those not presenting with seizures (OR 6.57[1.37-31.5], p=0.025). Status epilepticus occurred in 19 (9%) and was associated with fever (p=0.009).

A clinical score of seizure risk identified aetiology and admission GCS (AUROC = 0.775 [0.701-0.848]). The same model was validated in Cohort 2 (AUROC = 0.744 [0.677-0.811], p<0.001). Of patients with a low-risk score 14/57 (25%) had an inpatient seizure (reference group), of medium-risk 22/46 (48%; OR 2.82[1.22-6.49], p=0.015), of high-risk 23/31 (74%; 8.83[3.23-24.14], p<0.001) and of very high-risk 13/15 (87%;19.96[4.00-99.50], p<0.001). Although patients with seizures had a shorter pre-admission symptom duration (5[1-12] vs. 9[4-24] days, p=0.01), they had a worse outcome (Glasgow Outcome Scale<5 OR 1.81[1.01-3.23], p=0.04).

A small number of variables can effectively stratify acute seizure risk in patients with encephalitis. These findings can support the development of targeted interventions and aid clinical trial design.