Résumé : Is it necessary to analyze two blastomeres in preimplantation genetic diagnosis (PGD) by fluorescence in situ hybridization (FISH) or is one blastomere enough, as suggested by some teams? We analyzed the sensitivity (Se), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV), false positives (FP), false negatives (FN), and the efficiency (Eff) of FISH performed on one (Group I) or two (Group II) blastomeres. Ninety embryos were analyzed (day 3), 19 blastocysts were replaced (day 5), 64 embryos were reanalyzed (day 5), (Group I = 23; Group II = 41). No differences were observed between the two groups for all of the parameters considered, but one false negative was observed in Group I. Furthermore, two embryos from Group II, which had a discordant diagnosis at PGD (one blastomere being normal and one abnormal), were read as abnormal after reanalysis. The accidental biopsy of the normal blastomere could have lead to the selection of these 2 embryos for transfer, causing a misdiagnosis rate of 4.8%. We conclude that embryo reanalysis is a useful tool to test the reliability of PGD in each laboratory: that PGD on two blastomeres is safer because the practice of PGD on one blastomere can result in a false-negative misdiagnosis.