Reply to Comments on Staessen et al. (2004)
Centres for Reproductive Medicine and Medical Genetics, Vrije Universiteit Brussel, Laarbeeklaan 101, 1090 Brussels, Belgium
1 To whom correspondence should be addressed. Email: catherine.staessen{at}az.vub.ac.be
Sir,
We appreciate the interest of Mastenbroek and colleagues and of Cohen and Munné in our randomized controlled trial (RCT) studying preimplantation genetic screening (PGS). We started this randomized study in order to evaluate the difference in outcome between blastocyst transfer with or without PGS.
Mastenbroek et al. raise the issue of using implantation rate per embryo as the main outcome measure. We certainly agree that ongoing pregnancy should be the primary outcome of assisted reproduction technology (ART) trials; however, despite the unit of analysis problem, we elected implantation rate to be the primary outcome measure since the technique of aneuploidy screening is not performed on patients, but on embryos. Accordingly, power analysis was done on the basis of numbers of embryos. We felt that our report had to adhere to that design. Other studies reporting on PGS (Gianaroli et al., 1999
; Munné et al., 2003) compare the implantation rate between the tested and the control groups, and provide a valid reference to compare with our own findings.
According to Mastenbroek and colleagues, the implantation rate is an inappropriate measure since the number of embryos transferred depends on the strategy and not on the design. The transfer policy is mentioned in Materials and methods and the results presented are valid for the transfer policy followed. What complicates the situation of course is that different numbers of embryos were transferred between the two groups, although this was not intentional. As we discussed in the paper, our trial might have given different results if an equal number of embryos were transferred in both groups. However, this would have been a difficult and even unethical design given the unpredictable numbers in this subgroup of patients. Eventually a design in which only one embryo would have been transferred in both study arms could have been chosen, but is it ethical to transfer only one embryo in patients above 37 years of maternal age if more are available, especially in the control group? Nevertheless, the study, as it was conducted, has led us to conclude that this RCT provides no arguments in favour of PGS in patients with advanced maternal age (AMA) when there are no restrictions in the number of embryos to be transferred. The allocation of the patients to the study was not double-blinded because it was felt unethical that patients would not be aware of what treatment their embryos would be exposed to. At the level of embryo selection, Mastenbroek et al. express their doubts about the blinding. However, we can reassure them that the blinding was effective: in the IVF laboratory, the embryos are evaluated and judged to be transferable on the basis of morphological characteristics by an independent embryologist and, among those embryos, the genetically normal are transferred.
All relevant data are available in the manuscript allowing recalculations. The results as recalculated by Mastenbroek et al. show an ongoing pregnancy rate of 11% in the PGS group versus 15% in the control group [relative risk (RR) 0.72, 95% confidence interval (CI) 0.43–1.21]. The conclusion of the RCT study thus remains the same, i.e. that in terms of clinical pregnancy rate there is no difference between the control group and the PGS group. In view of this recalculation, the conclusion of our RCT that the effectiveness of PGS to improve IVF/ICSI in women with AMA, at least when there are no restrictions on the number of embryos for transfer, remains unproven, is still valid.
The issue raised by Cohen and Munné brings us to a more delicate point that has been a target of discussion for a long time. The biopsy of one blastomere may be less detrimental for the further development of the embryo, but may jeopardize the accuracy of the diagnosis. The references mentioned by Cohen and Munné are not convincing: Bahçe (2003) is comparing 383 embryos from which in 99% (379) one blastomere was removed and in 1% (four) two cells were removed. The implantation rate was 41.17% after one blastomere biopsy versus 18.6% in the two-blastomere biopsy. The reference of Magli et al. (2004), which actually does not deal with the comparison of one- and two-cell biopsies, indeed mentioned in the Introduction a retrospective analysis of data: 29% implantation after one-cell biopsy versus 16% after two-cell biopsy. No other data, such as the number of embryos analysed, are presented. Cohen and Munné infer that the low implantation rate of blastocysts was due to suboptimal culture conditions, an argument that overlooks the mean age of the patients included and their unfavourable infertility history. Also, there were no exclusions because of higher day 3 FSH, failure on previous attempts or insufficient follicles during stimulation. In a typical European setting with IVF reimbursement, implantation rates tend to be lower because of the lack of selection criteria that tend to include ‘optimal’ patients.
References
Bahçe M (2003) Effect of different biopsy procedures on implantation rates. Proceeding of the Fifth International Symposium on PGD. Antalya, Turkey, 5–7 June 2003. 0–23.
Gianaroli L, Magli C, Ferraretti A and Munné S (1999) Preimplantation diagnosis for aneuploidies in patients undergoing in vitro fertilization with poor prognosis: identification of the categories for which it should be proposed. Fertil Steril 72, 837–844.[CrossRef][Web of Science][Medline]
Magli MC, Gianaroli L, Ferraretti AP, Toschi M, Esposito F and Fasolino MC (2004) The combination of polar body and embryo biopsy does not affect embryo viability. Hum Reprod 19, 1163–1169.
Munné S, Sandalinas M, Escudero T, Velilla E, Walmsley R, Sadowy S, Cohen J and Sable D (2003) Improved implantation after preimplantation genetic diagnosis of aneuploidy. Reprod Biomed Online 7, 91–97.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||