The electroencephalogram (EEG) is a very useful tool for the evaluation of patients with altered levels of consciousness and coma. 1,2 It permits one to rule out status epilepticus, confirm a diffuse or focal process and may orient the clinician towards an underlying etiology. Triphasic waves (TWs) on the EEG were described more than five decades ago and are usually attributed to a metabolic encephalopathy. 3,4 The association of TWs with hepatic encephalopathy is well known. In the past, they were thought of as being pathognomonic of hepatic coma. We now know that TWs are associated with a wide variety of disorders (metabolic and non-metabolic) affecting the brain diffusely. 5,6 Triphasic waves are not epileptiform per se and are usually not associated with seizures. The TWs and generalized periodic or continuous epileptiform discharges occurring in generalized nonconvulsive status epilepticus (GNCSE) may share ABSTRACT: Background: Triphasic waves (TWs) and generalized nonconvulsive status epilepticus (GNCSE) share morphological features that may create diagnostic ambiguity. Objective: To describe electroencephalographic differences between TWs and GNCSE. Methods: We retrospectively compared the electroencephalograms (EEGs) of two groups of patients presenting with decreased level of consciousness; those with TWs associated with metabolic encephalopathy and those with GNCSE. We studied the following: demographics, etiology and EEG morphological features. All EEGs were classified blindly (TWs or GNCSE) by two expert EEGers. Agreement between experts and concordance with clinical diagnosis were measured. Results: We analysed 87 EEGs (71 patients) with TWs and 27 EEGs (13 patients) with GNCSE. Agreement between experts and concordance with clinical diagnosis were excellent. When compared to TWs, epileptiform discharges associated with GNCSE had a higher frequency (mean=2.4Hz vs 1.8Hz) (p<0.001), a shorter duration of phase one (p=0.001), extra-spikes components (69% vs 0%) (p<0.001) and less generalized background slowing (15.1% vs 91.1%) (p<0.001). Amplitude predominance of phase two was common with TWs (40.8% vs 0%) (p=0.01). Lag of phase two was absent in all cases of GNCSE but present in 40.8% of patients with TWs. Noxious or auditory stimulation frequently increased the TWs (51%) while it had no effect on the epileptiform pattern (p=0.008). Conclusions: Certain EEG morphological criteria and the response to stimulation are very helpful in distinguishing TWs from GNCSE. Tous les patients avaient une altération de la conscience. Les caractéristiques suivantes furent analysées: âge, étiologie sous-jacente et plusieurs caractéristiques morphologiques de l'EEG. Tous les EEGs ont été classifiés à l'aveugle (OT ou SEGNC) par deux électroencéphalographistes experts et l'exactitude par rapport au diagnostic clinique a été évaluée de même que la concordance entre les deux experts. Résultats: 87 EEGs (71 patients) avec OT et 27 EEGs (13 patients) avec SEGNC ont été analysés. La concordance entre les experts et l'exacti...