Normalization of the psychometric hepatic encephalopathy score in Polish population. A prospective, quantified electroencephalography study

Background: The psychometric hepatic encephalopathy score (PHES) is recommended as a gold standard in evaluation of minimal hepatic encephalopathy (HE). Normative databases have been collected in few countries, clearly showing differences among studied groups. Thus, the standardization of PHES for selected populations remains necessary. Aims: To standardize PHES in a large cohort of Polish healthy subjects and to evaluate the normograms in patients with cirrhosis with quantified electroencephalography (EEG). Methods: Three hundred and sixteen (142 males/174 females, aged 44.5 +/- 12.1) normal individuals and 50 (31 males/19 females, aged 52.8 +/- 12.4) patients with cirrhosis without overt HE were included. Key correction variables of psychometric tests were performed. The multivariate linear regression was used to calculate PHES normograms. Results: Age and education levels were identified as predictors of all tests, therefore age- and education-adjusted normograms were developed. A weighted time-errors regression model for line tracing test (LTT) scoring was used. The PHES ranged between +5 and -15 points and the cut-off between normal and pathological PHES was set on <= -5 points. By this cut-off level, PHES had a sensitivity of 57% and specificity of 97% to diagnose minimal HE (AUC = 0.866 + 0.028). In patients with cirrhosis, PHES correlated with severity of liver disease (MELD, r = -0.475, P < 0.001 and Child-Pugh classification, r = -0.452, P < 0.002) and EEG (r = 0.547, P < 0.002). In patients with impaired EEG, PHES was lower than in individuals with unaltered EEG (P < 0.02); however, agreement between these two modalities was limited. Conclusions: Valid Polish PHES normograms, which incorporates w-LTT scoring system have been developed. Future multi-centre international studies are needed to validate widely applicable norms.