BACKGROUND: We present the first powered prospective study to assess whether sperm aneuploidy can predict the outcome of ICSI. METHODS: Our null hypothesis was that aneuploidy rates (AR) are identical in men who achieve successful (Group A) and unsuccessful (Group B) ICSI outcome. A power calculation yielded a sample number of 56 to achieve 80% power to reject our hypothesis at the 5% significance level. Samples for testing were obtained on the day of embryo transfer and tests were performed on raw pre-preparation samples. Sperm AR of chromosomes 13, 18, 21, X/Y were assessed using fluorescence in-situ hybridization (FISH) techniques (mean of 1223 sperm). RESULTS: There was no significant difference in any patient, seminal, cycle or laboratory characteristic between groups that may have affected outcome. Total AR (2.37 versus 1.18%, P 5 0.01), as well as AR of chromosomes 18, X/Y and 18 1 X/Y (1.48 versus 0.67%, P 5 0.005) were significantly higher in Group B compared with Group A. Regression analysis confirmed these differences to be independent of other variables and showed a 2.6-fold change in odds of achieving a pregnancy for every 1% change in total AR. CONCLUSIONS: Our findings confirm a potential role for aneuploidy testing in the work-up of ICSI patients as a predictor of success, as well as in future genetic counselling. If confirmed, there may also be a place for a study of preimplantation genetic screening to improve ICSI success in men found to have high AR and ICSI failure.