Thirty-four studies were included. The total number of patients was unclear but, where reported, ranged from 12 to 160 across studies in patients with PNES and 11 to 161 with epileptic seizures. Total number of seizures was unclear but, where reported, ranged to 31 to 280 events in patients with PNES and from 25 to 261 in patients with epileptic seizures. Only 22 studies included a control group with epileptic seizures. Incorporation bias was a potential source of bias in all studies as the ictal signs were considered when interpreting the video-EEG. Only four studies reported that the person who interpreted the ictal signs was blind to the results of the EEG.
There was good evidence that the following signs can be used to rule in PNES: long duration (no data reported; seven controlled studies), fluctuating course (sensitivity 47% to 88%, specificity 96% to 100%; two controlled studies), asynchronous movements (sensitivity 9% to 96%, specificity 93% to 100%; three controlled studies), pelvic thrusting (sensitivity 1% to 44%, specificity 92% to 100%; six controlled studies), side-to-side hear or body movements (sensitivity 15% to 63%, specificity 92% to 100%; five controlled studies), closed eyes (sensitivity 34% to 96%, specificity 74% to 100%; five controlled studies), ictal crying (sensitivity 4% to 37%, specificity 100%; four controlled studies) and memory recall (sensitivity 63% to 88%, specificity 90% to 96%; two controlled studies).
There was good evidence that the following signs can be used to rule in epileptic seizures: occurrence from sleep (sensitivity 31% to 59%, specificity 100%; three controlled studies), postictal confusion (sensitivity 61% to 100%, specificity 84% to 88%; two controlled studies) and stertorous breathing (sensitivity 61% to 91%, specificity 100%; three controlled studies).
There was insufficient evidence on gradual onset, non-stereotyped events, flailing or thrashing movements, opisthotonus, arc en cercle, tongue biting and urinary incontinence.