Interobserver reliability of video recording in the diagnosis of nocturnal frontal lobe seizures.

ABSTRACT

BACKGROUND: Nocturnal frontal lobe seizures (NFLS) show one or all of the following semeiological patterns (1) paroxysmal arousals (PA brief and sudden recurrent motor paroxysmal behavior); (2) hyperkinetic seizures (HS motor attacks with complex dyskinetic features); (3) asymmetric bilateral tonic seizures (ATS motor attacks with dystonic features); (4) epileptic nocturnal wanderings (ENW stereotyped, prolonged ambulatory behavior). OBJECTIVE: To estimate the interobserver reliability (IR) of video-recording diagnosis in patients with suspected NFLS among sleep medicine experts, epileptologists, and trainees in sleep medicine. METHODS: Sixty-six patients with suspected NFLS were included. All underwent nocturnal video-polysomnographic recording. Six doctors (three experts and three trainees) independently classified each case as "NFLS ascertained" (according to the above specified subtypes PA, HS, ATS, ENW) or "NFLS excluded". IR was calculated by means of Kappa statistics, and interpreted according to the standard classification (0.0-0.20 = slight agreement; 0.21-0.40 = fair; 0.41-0.60 = moderate; 0.61-0.80 = substantial; 0.81-1.00 = almost perfect). RESULTS: The observed raw agreement ranged from 63% to 79% between each pair of raters; the IR ranged from "moderate" (kappa = 0.50) to "substantial" (kappa = 0.72). A major source of variance was the disagreement in distinguishing between PA and nonepileptic arousals, without differences in the level of agreement between experts and trainees. CONCLUSIONS: Among sleep experts and trainees, IR of diagnosis of NFLS, based on videotaped observation of sleep phenomena, is not satisfactory. Explicit video-polysomnographic criteria for the classification of paroxysmal sleep motor phenomena are needed.