An important focus in Alzheimer's disease (AD) research is the development of methods for early diagnosis. Despite progress with some other biomarkers, sensitive and specific neuropsychological measures for identifying subjects in the prodromal phase of AD remain the most promising early diagnostic tool. We evaluated the value of the composite score for the Consortium to Establish a Registry for Alzheimer's disease Neuropsychological Battery (CERAD-NB) in Europeans with mild cognitive impairment (MCI) and in control populations. Baseline clinical data were analyzed from 223 healthy elderly and 224 subjects with MCI from the prospective AddNeuroMed study carried out in Finland, France, Greece, Italy, Poland, and the United Kingdom. The total score for CERAD-NB was calculated by the subtest addition method. The CERAD total score, adjusted for age, gender, education, and country, clearly differentiated the control and MCI groups (p < 0.001). The optimal between-groups cut-off point for the CERAD total score derived from ROC analysis yielded 81.5% sensitivity and 75.4% specificity (AUC = 0.848, p < 0.001). The CERAD total score was superior to the Mini-Mental Status Examination, or any single CERAD subtest in discriminating between the control and MCI groups. While the overall level of the CERAD total score varied between the different countries, it remained accurate in differentiating controls and MCI subjects within each country. We conclude that the CERAD total score is an accurate measure for detecting mild cognitive impairment, but implementing specific cut-off points needs to be based upon country specific normative data.