1 Center for Reproductive Medicine, Academic Medical Center, PO Box 22700, 1105 AZ Amsterdam, 2 Department of Clinical Epidemiology & Biostatistics, Academic Medical Center, Amsterdam and 3 Department of Obstetrics and Gynaecology, University Hospital Groningen, Groningen, The Netherlands
4 To whom correspondence should be addressed. Email: s.mastenbroek{at}amc.uva.nl
Sir,
We have read with great interest the manuscript by Staessen et al. (2004) reporting a randomized controlled trial (RCT) of preimplantation genetic screening (PGS) in couples with advanced maternal age (AMA). Several investigators, including Staessen et al. (2004)
, commonly use the term preimplantation genetic diagnosis for aneuploidies (PGD-AS) when referring to PGS. However, we will use the term PGS in this letter in agreement with recently adjusted nomenclature from the ESHRE PGD Consortium (Sermon et al., 2005
).
RCTs studying PGS for AMA have been awaited eagerly since PGS for AMA is applied more and more often in recent years (Sermon et al., 2005) despite a lack of evidence of effectiveness. Only one other RCT has been published thus far (Werlin et al., 2003
), but that RCT included a very small number of patients (seven PGS for AMA versus 12 controls). Several non-randomized studies have been published (Gianaroli et al., 1997
, 1999
; Munne et al., 1999
, 2003
; Kahraman et al., 2000
; Obasaju et al., 2001
; Montag et al., 2004
). Despite deficiencies in the design and analysis of some of these comparative studies, the results seem to indicate that PGS for AMA leads to an improved implantation rate without an increase in clinical pregnancy rate. Only a few studies have reported on ongoing pregnancies and none on live birth, the ultimate outcome of interest to patients. The study by Staessen et al. (2004)
is the first RCT to provide data on the effectiveness of PGS.
Its importance encouraged us to comment on some aspects of the design and analysis of this trial that, in our opinion, attenuate the conclusions that can be drawn.
Several investigators have stressed that the main outcome measure in subfertility trials should be live birth or, if not available, ongoing pregnancy rate (Barlow, 2003; Daya, 2003
; Vail and Gardener, 2003
). Live birth is the most important end result of any fertility treatment. By using implantation rate per embryo as the main outcome measure, Staessen and colleagues steer away from this recommendation. They also introduce a unit of analysis error.
Staessen and colleagues used the embryo implantation rate as their primary measure of effectiveness. The embryo implantation rate is the ratio between the number of gestational sacs with a fetal heartbeat and the total number of embryos transferred. This is an inappropriate measure since the denominator (number of embryos transferred) depends on the strategy, not on the design, as is shown in Table II of the paper. It is incorrect to calculate P-values, odds ratios and even a number needed to treat based on this ratio.
As women were the unit of randomization in this trial, they should also be the unit of analysis and, consequently, the basis for the sample size calculations. The CONSORT flow chart in Figure 1 suggests that women were the unit of analysis, but this is not the case. Furthermore, this is also not an analysis on an intention to treat basis: an analysis that compares outcomes of all randomized subjects, regardless of whether they actually received the treatment, in the groups to which they were originally assigned. (Moher et al., 2001; Daya, 2003
; Vail and Gardener, 2003
)
Despite these deficiencies, all relevant data are available in the manuscript and appropriate rates can be recalculated without unit of analysis error and on an intention to treat basis. The results then show an ongoing pregnancy rate of 11% (22 out of 199) in the PGS group versus 15% (29 out of 190) in the control group [relative risk (RR) 0.72, 95% confidence interval (CI) 0.431.21].
It is unfortunate that this trial does not fully adhere to the CONSORT guidelines for reporting randomized clinical trials. Two of the main sources of bias in RCTs are incomplete follow-up and lack of concealment of allocation. It is unclear if there was a proper concealment of allocation in this RCT and whether there was adequate blinding to the mode of embryo selection during treatment.
The increasing use of PGS warrants quality studies. It is to be regretted that this first RCT of significance suffers from a number of deficiencies in design and analysis. The effectiveness of PGS has not yet been demonstrated. Properly designed and conducted RCTs are needed to guide decisions about the use of PGS to increase the effectiveness of IVF/ICSI in women with AMA.
References
Barlow DH (2003) The design, publication and interpretation of research in subfertility medicine: uncomfortable issues and challenges to be faced. Hum Reprod 18, 899901.
Daya S (2003) Pitfalls in the design and analysis of efficacy trials in subfertility. Hum Reprod 18, 10051009.
Gianaroli L, Magli MC, Munne S, Fiorentino A, Montanaro N and Ferraretti AP (1997) Will preimplantation genetic diagnosis assist patients with a poor prognosis to achieve pregnancy? Hum Reprod 12, 17621767.[Abstract]
Gianaroli L, Magli MC, Ferraretti AP and Munne S (1999) Preimplantation diagnosis for aneuploidies in patients undergoing in vitro fertilization with a poor prognosis: identification of the categories for which it should be proposed. Fertil Steril 72, 837844.[CrossRef][ISI][Medline]
Kahraman S, Bahce M, Samli H, Imirzalioglu N, Yakisn K, Cengiz G and Donmez E (2000) Healthy births and ongoing pregnancies obtained by preimplantation genetic diagnosis in patients with advanced maternal age and recurrent implantation failure. Hum Reprod 15, 20032007.
Moher D, Schulz KF and Altman DG (2001) The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised trials. Lancet 357, 11911194.[CrossRef][ISI][Medline]
Montag M, van der Ven K, Dorn, C and van der Ven H (2004) Outcome of laser-assisted polar body biopsy and aneuploidy testing. Reprod Biomed Online 9, 425429.[ISI][Medline]
Munne S, Magli C, Cohen J, Morton P, Sadowy S, Gianaroli L, Tucker M, Marquez C, Sable D, Ferraretti AP et al. (1999) Positive outcome after preimplantation diagnosis of aneuploidy in human embryos. Hum Reprod 14, 21912199.
Munne 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, 9197.[Medline]
Obasaju M, Kadam A, Biancardi T, Sultan K, Fateh M and Munne S (2001) Pregnancies from single normal embryo transfer in women older than 40 years. Reprod Biomed Online 2, 98101.[Medline]
Sermon K, Moutou C, Harper J, Geraedts J, Scriven P, Wilton L, Magli MC, Michiels A, Viville S and De Die C (2005) ESHRE PGD Consortium data collection IV: MayDecember (2001). Hum Reprod 20, 1934.
Staessen C, Platteau P, Van Assche E, Michiels A, Tournaye H, Camus M, Devroey P, Liebaers I and Van Steirteghem A (2004) Comparison of blastocyst transfer with or without preimplantation genetic diagnosis for aneuploidy screening in couples with advanced maternal age: a prospective randomized controlled trial. Hum Reprod 19, 28492858.
Vail A and Gardener E (2003) Common statistical errors in the design and analysis of subfertility trials. Hum Reprod 18, 10001004.
Werlin L, Rodi I, DeCherney A, Marello E, Hill D and Munne S (2003) Preimplantation genetic diagnosis as both a therapeutic and diagnostic tool in assisted reproductive technology. Fertil Steril 80, 467468.[CrossRef][ISI][Medline]