The utility of latency and spectral analysis methods in evoked potential recordings from patients with hepatic encephalopathy

Evoked potentials (EPs) are small phasic potentials that are elicited in conjunction with sensory, motor and cognitive events. EP variables have been assessed in patients with cirrhosis but in general, methods were inadequately standardized and study populations incompletely characterized, leading to some studies questioning the validity of EP’s in diagnosing and monitoring hepatic encephalopathy, while other studies indicated that there is only a low positive yield with these investigations. Few studies have attempted tri-modal sensory and cognitive recordings. Recorded waveforms may demonstrate altered morphology while possessing broadly normal latencies. Since EP analysis is usually performed solely in the time domain, latency measurements do not therefore highlight morphological changes to the waveform and so abnormalities may go unreported. The aim of this study was twofold (i) to measure sensory and cognitive EPs in patients with cirrhosis in relation to their neuropsychiatric status and (ii) to address frequency content in relation to neuropsychiatric status by examining EPs with two spectral techniques, the Fourier Transform (FT) and the Power Spectral Density Estimate (PSD). Seventy patients with biopsy–proven cirrhosis were classified using clinical, psychometric and EEG criteria as unimpaired or as having minimal or overt hepatic encephalopathy (HE). Forty-eight healthy individuals served as controls. Visual (VEPs), brainstem auditory (BAEPs) somatosensory (SSEPs) and cognitive auditory (P300) EPs were recorded under standardized conditions. Significant latency differences were observed in sensory EPs between patients and controls with patient subgroups differences being less significant. The cognitive auditory P300 however, distinguished the patient subpopulations from one another. Frequency shifts are observed in all EP modalities with significant differences also occurring between patient groups. The sensitivity and specificity of the frequency-domain is comparable to that of the time-domain. Paired EP investigations analysed by latency indicate BAEP and P300 best discriminate any degree of encephalopathy; in the frequency domain it is the VEP combined with SEP and in the time-frequency domain it is the SEP. These findings suggest that EPs, when performed as a bank of multimodal tests and with spectral analysis, could provide a sensitive and specific method for the diagnosis and monitoring of hepatic encephalopathy.