Objective To determine the predictive value of euploid embryos in women with recurrent implantation failure undergoing repeated IVF-ET cycles with PGD (PGD). Design Cohort of IVF-PGD cycles in a tertiary care ART facility. Materials and method(s) Fifty-five consecutive patients with repeated implantation failure (more than three failed IVF-ET cycles) underwent two or more PGD cycles for aneuploidy testing. Mean maternal age was 37.6±5.3 years. Biopsies were performed on day 3. One blastomere was removed from each pre-embryo, fixed and analyzed by multicolor and multi-probe FISH for chromosomes X and Y, 13, 15, 16, 17, 18, 21, and 22. Result(s) Forty-three of 55 patients (78%) undergoing PGD had at least one euploid embryo for transfer. Of these 31 patients (72%) also had at least one euploid embryo available for transfer with the second cycle. Of the 12 (28%) patients with no euploid embryos available for transfer with the second IVF/PGD cycle, five had a third cycle of PGD and two of these had euploid embryos available for transfer. Seventeen of the 31 patients (55%) who had euploid embryos on the second PGD cycle conceived. The ongoing pregnancy and implantation rates in patients with at least one euploid embryo were 40% and 18%, respectively. Twelve of the 55 patients (22%) had no euploid embryos available for transfer on the first PGD cycle, but on the second PGD cycle, six (50%) of these had euploid embryos for transfer. Only two pregnancies were achieved among this group of women, yielding a pregnancy rate of 17%, but both conceptions resulted in miscarriage. Of the six patients with no euploid embryos available after the second PGD cycle, four patients had a third IVF/PGD cycle, but none had euploid embryos available for transfer. Also, among women with euploid embryos available only in either the first or second PGD cycle, but not both, no ongoing pregnancy was achieved. No woman who had a PGD cycle productive of no euploid embryos had an ongoing pregnancy. Significant differences were found in terms of ongoing pregnancy (40%, P<0.05) and implantation rates (18%, P<0.05) in women with euploid embryos available for transfer with the first and second IVF/PGD cycles, compared to women with no euploid embryos available for transfer with either the first or second cycle. The positive predictive value of the first euploid cycle predicting a second euploid cycle was 72%, 95% CI 0.66-0.78. The negative predictive value of an aneuploid cycle was 50%, 95% CI 0.27-0.72. The sensitivity and specificity of the first PGD cycle predicting the second was 84%, 95% CI 0.77-0.91 and 33%, 95% CI 0.18-0.48, respectively. Conclusion(s) Even with a history of recurrent implantation failure, the availability of euploid embryos, especially on two, consecutive PGD cycles is associated with high ongoing pregnancy and implantation rates. Conversely, the absence of euploid embryos for transfer predicts poor reproductive outcome, even if subsequent cycles do yield euploid embryos.