Abstracts

Rationale: Cyclic seizures (CS) are characterized by recurrence of seizures at regular intervals and have been mostly described in critically ill patients. The clinical correlate and prognostic significance of CS are currently unknown. Methods: We maintain a prospective database of patients undergoing continuous EEG monitoring (CEEG). We reviewed the EEG of all patients with seizures between May 2011 to May 2014 to identify patients with cyclic seizures, defined as recurrence of seizures at a regular interval with a frequency > 1/h. The frequency (defined by number of seizures per hour in the most frequent hour), duration, location of onset and associated EEG findings were recorded. Medical charts were reviewed to collect data on demographics, etiology, treatment and prognosis. A control group consisted of all consecutive patients with electrographic status epilepticus (SE; defined as at least one seizure > 5 minutes) without CS during the same time period. Statistical analysis was performed with Fisher and Mann-Whitney tests. Results: Among the 260 patients with electrographic seizures, we identified 39 (15%) patients with CS and 29 (12%) with non-cyclic electrographic SE. Patients with CS were older and more frequently had acute or progressive brain injury than controls (Table 1). There was no difference in gender, occurrence of seizure before monitoring, the presence of a clinical correlate during seizures, need for mechanical ventilation, total number of antiepileptic drugs or anesthesics used for treatment or functional status before the event (measured by Glasgow Outcome Scale). There was a trend towards lower scores in Glasgow Coma Scale (GCS) and worse prognosis in the CS group, with greater decline in performance and number of deaths. Patients with CS had more focal seizures than SE, with almost half of them (19) originating from the posterior quadrant (52% of SE had generalized correlate, p = 0.004). Median seizure frequency was 15/h (ranging from 4 to 38), median seizure duration was 90 seconds (range 20 to 240s). Also, CS records had more lateralized periodic or rhythmic patterns (including lateralized periodic discharges (LPDs), bilateral independent periodic discharges (BIPDs), lateralized rhythmic delta activity (LRDA) and lateralized sharp and wave (LSW)) than controls (72% vs 30%, p = 0.003). Conclusions: Cyclic seizures are not uncommon in critically ill patients with seizures, with more than half of patients with electrographic SE having this pattern. They occur mostly in the setting of acute or progressive brain injury, originate most often from the posterior region and carry a prognostic significance similar to, if not slightly worse, than electrographic SE. A larger prospective study is required to investigate if they exert an independent effect on outcome.