BACKGROUND. Mild Cognitive Impairment is a common condition defined as transitional state between normality and dementia of Alzheimer type. Clinically is characterized by subjective and objective memory loss beyond the expected for age and educational level, although a broad range of cognitive inefficiencies may appear, with preservation of daily living activities. Approximately half the patients convert to dementia within 3 years. Since no all patients convert to dementia it is essential to find reliable predictors so as to start the appropriate treatment as soon as possible. METHOD. Extensive Medline-based search for articles dealing with predictors of conversion to dementia in Mild Cognitive Impairment (MCI). RESULTS. There is a substantial body of literature dealing with predictors of dementia in patients with MCI. These predictors range from a simple delayed recall task on Mini-Mental to sophisticated radiological techniques and CSF biomarkers. Comprehensive neuropsychological tests rarely surpass 70% sensitivity and specificity. The presence of the APOE epsilon 4 allele has been associated with increased risk of conversion but the sensitivity is quite low. CSF biochemical markers are being developed with encouraging results. β-amyloid 42 protein is usually lower in converters than in people with stable cognitive status and tau protein is higher. The sensitivity is substantial but specificity is so far low. An epitope of tau protein (P231) looks more specific of Alzheimers disease and therefore a promising biomarker. In the blood, high β-amyloid protein levels indicate risk of conversion but only a few studies have been published. Hippocampal or entorhinal atrophy on MRI is one of the most used radiological markers of conversion but quantification of atrophy is not simple as it is subject to artifacts and anatomic variations. Proton Magnetic Resonance Spectroscopy (MRS) and Positron Emission Tomography (PET) are emerging as the most promising predictive tools. The highest degree of accuracy ( > 90%) has been achieved by means of PET plus either memory performance or APOE4 genotype. However, the samples of the published studies are mostly small, and these instruments are not widely available. CONCLUSIONS. There is no enough evidence to recommend specific techniques for predictions. Until an accurate marker is developed, a combined use of cognitive tests, APOE genotype, and a neuroradiological technique is probably the best option for prediction purposes depending on availability and experience.