Neuropsychological tests used for the assessment of subclinical hepatic encephalopathy (SHE) may overdiagnose SHE because scores are usually not corrected for age. The aim of this study was to estimate the prevalence of SHE using two easy administrable psychometric tests (Number Connection Test part A [NCT-A], Symbol Digit Test [SDT]) with age-related normal values. In addition, spectral electroencephalogram (EEG) was used, which is the in-house electrophysiological method for quantifying encephalopathy. One hundred and thirty-seven consecutive patients (mean age 49 years, range 17-77) with cirrhosis without any clinical signs of encephalopathy, were screened for SHE. In addition, the Child-Pugh score and the arterial blood ammonia were determined. Patients with concurrent use of alcohol, benzodiazepines or anti-epileptics were excluded. Fifty percent of the patients had an abnormal NCT according to the standard recommended procedure, in contrast only 7% of the patients had an abnormal NCT when scores corrected for age were used. Combining the results of the spectral EEG and the psychometric tests corrected for age yielded a higher prevalence of SHE (23%) than when each test method was used alone (17% vs. 10% abnormal, respectively). Severity of liver disease correlated with the presence of SHE, because the prevalence of abnormal tests increased from 14% in Child-Pugh grade A to 45% in Child-Pugh grade B or C. Age above 40 years and an elevated blood ammonia level were significant determinants related to an abnormal EEG. We conclude that the NCT uncorrected for age markedly overdiagnoses SHE and, therefore, should not be used as a test for the screening of SHE. Using a combination of spectral EEG and two psychometric tests with age-corrected normal values a low prevalence of SHE in patients with Child A liver cirrhosis is found. Older patients with an elevated arterial ammonia are more prone to develop SHE than younger patients with an equal arterial ammonia concentration.