OBJECTIVE To determine the ability of pattern-electroretinogram (PERG) to discriminate individual evolution of glaucomatous visual field losses in glaucoma suspects. METHODS We studied 151 eyes of 78 GS, classified depending on their risk to develop glaucomatous damage. The initial visual field was normal. Those eyes showing at least 2 reduced amplitudes in the PERG components (P50, N95 and NP) were considered pathologic. The mean follow-up was 30 months (range: 13-50 months); the VF was evaluated at the end of the study. RESULTS At the beginning of the study the PERG was pathologic in 65 eyes (43%); in 45 of these eyes (69%) the visual field was pathologic at the end of the study. The positive predictive value of the PERG of visual field glaucomatous changes was 69% and the negative predictive value was 80%. The N95 amplitude showed the best predictive value, with 100% and 77% in the high and medium risk groups, respectively. CONCLUSIONS The PERG study can discriminate the glaucoma suspects that will develop visual field glaucomatous losses.