Assessment of hepatic encephalopathy with visual evoked potentials compared with conventional methods.

Thirty-six patients with advanced chronic liver disease of predominantly alcoholic etiology and with a documented history or current physical evidence of hepatic encephalopathy were studied and compared to 30 healthy controls. Assessment was made of their mental state, number connection test, venous blood ammonia, electroencephalography and visual evoked potentials with both pattern reversal and flash stimuli. Because of considerable inter- and intraindividual variation in waveform, visual evoked potentials from flash stimuli were considered unreliable. In pattern reversal visual evoked potentials, the latency of the N1 and P1 waves was significantly longer (p less than 0.05) in patients than in controls; however, the wave latencies did not correlate with the mental state score. The mental state score correlated with the number connection test (r = 0.69, p less than 0.001), asterixis (r = 0.36, p less than 0.05), electroencephalography mean dominant frequency (r = 0.44, p less than 0.01) and blood ammonia (r = 0.60, p less than 0.01). In 14 patients studied sequentially, change in the mental state score correlated with change in the number connection test (r = 0.80, p less than 0.01) and asterixis (r = 0.75, p less than 0.01) but not with change in the electroencephalography, blood ammonia or visual evoked potential wave latencies. Although visual evoked potentials are abnormal in patients with alcoholic cirrhosis and encephalopathy, they are less accurate in assessing the level of consciousness than simple bedside evaluation with a number connection test.