Autologous endometrial co-culture in patients with repeated failures of implantation after in vitro fertilization-embryo transfer

Purpose: Our purpose was to evaluate the effect of co-culture on preembryo development and clinical outcome. Methods: Enrolled patients underwent a luteal-phase endometrial biopsy. The tissue was then enzymatically digested (collagenase) and the stromal and glandular cells were separated by differential sedimentation rates. These cells were cultured to confluence, released, and then cryopreserved until the patient's in vitro fertilization (IVF)-embryo transfer (ET) cycle. All normally fertilized oocytes were then placed on the co-cultured cells until transfer on day 3. Preembryo development on co-culture was compared to that in the patient's noncocultured previous cycle. Implantation and clinical ,pregnancy rates were compared to those in a control group of patients undergoing IVF during the study period who were matched for age, stimulation, protocol, number of oocytes retrieved, and preembryos transfered. Results: Twenty-nine women underwent 31 cycles of IVF-ET: On day 3 the overall mean number of blastomeres per-preembryo on co-culture compared to that in the patient's previous cycle was 6.3 +/- 1.8 vs. 5.6 +/- 1.2 (P = 0.04). The average percentage of cytoplasmic fragments on co-culture compared to the previous cycle was 16 +/- 9% vs. 19 +/- 9% (P = 0.32) At transfer; after preembryo selection, the mean number of blastomeres per preembryo on co-culture compared to that in the patient's previous cycle was 6.8 +/- 1.6 vs. 6.6 +/- 1.3 (P = 0.5). The implantation and clinical pregnancy rates between co-culture and the matched control group were 15% (14/93) vs. 13% (16/124) (P = 0.79) and 29% (9/31) vs. 25% (10/40) (P = 0.45). Conclusions: There was a significant improvement in the average number of blastomeres per preembryo on co-culture compared to that in the patient's previous noncoculture cycle. The overall implantation and clinical pregnancy rates between co-culture and a marched control group were not significantly different.