This prospective study was planned to assess whether quantitative EEG (qEEG) can give an estimate of the timing of achievement of three endpoints (loss of activities of daily living, incontinence, and death) in 72 consecutive patients (53 females, 19 males; mean age, 70.8) affected with probable Alzheimer's disease, as defined according to the NINCDS-ADRDA criteria. Power-weighted, log-transformed relative values of the four conventional EEG bands were considered in a central-posterior temporal region for each hemisphere. The hypothesis was tested by the lifereg procedure of the Statistical Analysis System package (first significance level accepted, P < or = 0.01). Because patients were in different stages of the disease, the statistical analysis was performed in the entire group as well as in the subgroup of 41 patients (mean age, 69.6) with mild dementia (scoring 3 or 4 on the global deterioration scale). In the whole group, the loss of activities of daily living was predicted by delta power in either side (P = 0.01), incontinence was predicted by alpha power in the right side (P < 0.01), whereas the statistical significance was not reached for death (P < 0.05). In the subgroup of mild demented patients, the loss of activities of daily living was predicted by delta power in the left side (P = 0.01), incontinence by both delta (P < 0.01) and alpha (P < 0.001) power in the right side, and death was not significantly predicted (P = 0.08). Quantitative EEG is a low-cost, discomfort-free technique which may be used to obtain information on the timing of disease evolution. The results showed in mild Alzheimer's disease appear especially interesting to attempt a prediction of the future time course of the disease from its beginning.