Perspectives on the efficacy and indications for preimplantation genetic screening: where are we now?

Given that the majority of failed implantations and early pregnancies reasonably may be attributed to embryo aneuploidy, the logical foundation for the use of preimplantation genetic screening (PGS) in cycles of in vitro fertilization (IVF) seems undeniably sound. If abnormal embryos having no implantation or advanced developmental potential can be identified accurately and excluded, and if only normal, euploid embryos are transferred, improved outcomes certainly should be expected, at least in women at a high risk of aneuploidy, if not in all women. Those at greatest risk of aneuploidy and therefore presumed most likely to benefit from PGS include women of advanced maternal age, women having recurrent unexplained miscarriages or implantation failure after transfer of good-quality embryos, and women with partners having severe male factor infertility. Results of a recent randomized controlled trial evaluating the efficacy of preimplantation genetic screening (PGS) for the indication of advanced maternal age stimulated a vigorous and passionate debate, which continues (Mastenbroek et al., 2007). This issue of Human Reproduction contains four new contributions relevant to that debate. One presents a subgroup analysis of data from the study by Mastenbroek et al., performed to assess whether the effect of PGS on live birth rates in women of advanced maternal age differed with varying risks for aneuploidy (Twisk et al., 2008). A second reports the results of a new randomized controlled trial examining the impact of PGS on in vitro fertilization (IVF) outcomes in women of advanced maternal age (Hardarson et al., 2008). The third offers an opinion about the wisdom and ethics of performing additional trials of PGS in women of advanced maternal age (Mastenbroek et al., 2008). The last reports the results of a randomized controlled trial evaluating the efficacy of PGS for improving the live birth rates achieved with single embryo transfer in young women (Staessen et al., 2008). A brief review of previous PGS trials will help to place these contributions in proper context and clearer perspective. The concept of PGS is so logical and attractive that its efficacy was generally assumed and accepted without serious question, once its technical feasibility had been demonstrated (Munné et al., 1993a,b). Enthusiasm for PGS was fueled by the results of observational studies comparing the outcomes of IVF with and without PGS and finding that PGS was associated with increased implantation rates (Gianaroli et al., 1999; Munné et al., 2003) and decreased miscarriage rates (Munné et al., 2006). The use of PGS has increased steadily, rapidly in recent years, despite the lack of any substantive evidence that PGS can deliver on its promise for achieving higher live birth rates (Twisk et al., 2006). Some predicted that PGS was destined to become a routine element of all IVF cycles (Verlinsky et al., 2004), and in some highly competitive regions in the USA, marketing even implied that programs which could not or chose not to offer PGS were substandard. The first serious doubts about the efficacy of PGS emerged from the results of a randomized controlled trial that appeared in 2004. Staessen et al. (2004) compared implantation rates in women aged 37 years and older after a single cycle of IVF with and without PGS using fluorescence in situ hybridization (FISH) for seven chromosomes (X, Y, 13, 16, 18, 21 and 22). In the PGS group, two blastomeres were removed from all embryos having six cells or more. Up to three (age 37–39 years), or six blastocysts were transferred (age 40 years), and in the PGS group, only chromosomally normal embryos were transferred. Significantly fewer cycles progressed to transfer in the PGS group (54.7%) than in control group (85.8%). Implantation rates (17.1% in the PGS group versus 11.5% in the control group) were not significantly different. When later analyzed by others, on an intention-to-treat basis, the ongoing pregnancy rate in the PGS group was 11.0%, compared with 15.3% in the controls [relative risk 0.72, 95% confidence interval (CI) 0.43–1.21] (Mastenbroek et al., 2005). The overall pregnancy loss rates (including both preclinical and clinical abortions) also were not different (24.1% in the