A retrospective cohort study was conducted on 201 women aged 28-44 years, each of whom underwent one cycle of IVF-embryo transfer with fresh, intracytoplasmic sperm injection (ICSI)-derived 7- to 10-cell embryos, transferred 72 h after oocyte retrieval. Samples of media surrounding separately cultured embryos were collected 46 h post-ICSI and stored for subsequent specific enzyme-linked immunosorbent assay. A total of 594 embryos (from own or donor oocytes) were transferred to 201 women. Group A comprised 159 recipients under 39 years and group B compromised 42 recipients aged 39-44 years. Groups A-1 and B-1 recipients had at least one embryo that tested above the geometric mean for soluble human leukocyte antigen-G (sHLA-G) ('positive expression') transferred. In groups A-2 and B-2, all embryos transferred expressed sHLA-G below the geometric mean ('negative expression'). In group A-1, 72/101 women (71%) achieved ultrasound confirmed (clinical), viable (cardiac activity observed) pregnancies. The implantation rate per embryo (IR) was 38%. In group A-2, 13/58 (22%) achieved viable clinical pregnancies. The IR was 9%. In group B-1, the viable clinical pregnancy rate was 52% (15/29) and the IR was 25% compared with a viable clinical pregnancy rate of 15% (2/13) and an IR of 5% in group B-2. The results of this study suggest that by selecting specific embryos for transfer based on their individual sHLA-G expression, pregnancy and implantation rates can be maximized while the number of embryos transferred can be reduced, thereby minimizing the incidence of high-order multiple pregnancies.