Hum. Reprod. Advance Access published online on July 9, 2008
Human Reproduction, doi:10.1093/humrep/den250
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Randomized clinical trial in assessing PGS: necessary but not sufficient
College of Medicine, Florida International University, Building HLS II, Room 672, Miami, FL 33315, USA
1 Correspondence address. Tel: +1-305-348-0613; Fax: +1-305-348-0651; E-mail: simpsonj{at}fiu.edu
 |
Abstract |
|---|
 Top Abstract IntroductionRCTs requiring technical skillsConclusionReferences |
|---|
The randomized clinical trial (RCT) is a powerful experimental design that when properly executed produces generalizable results. Conducting a RCT becomes complex when technical skills are required. Without requisite skills, a RCT may yield misleading results, an elegant RCT unwittingly generating spurious results due to technical inexperience. This pitfall is applicable to procedures used to evaluate assisted reproductive technologies. RCTs assessing the value of preimplantation genetic screening, also called preimplantation genetic diagnosis for aneuploidy testing—require three general prerequisites—proper study design, skilled operators (embryo biopsy), and skilled laboratory cytogeneticists (diagnosis). Lacking either of the latter two, even an elegantly designed RCT is not necessarily valid.
|
Introduction |
|---|
|
TopAbstractIntroductionRCTs requiring technical skillsConclusionReferences |
|---|
Prior to accepting any medical intervention, evidence is needed. This applies to pharmacological agents, surgical procedures or technically demanding diagnostic laboratory procedures. Comparing treated and untreated groups has long been accepted as the proper experimental design; the ‘gold standard’ is a randomized clinical trial (RCT). When subject to proper inclusion/exclusion criteria, randomization, unbiased (‘blinded’) analysis and adjustment for potential confounding variables, the RCT is a powerful experimental design whose results are generalizable. But is the result of an RCT or any elegantly designed trial sufficient to dictate practice? Not necessarily.
|
RCTs requiring technical skills |
|---|
|
TopAbstractIntroductionRCTs requiring technical skillsConclusionReferences |
|---|
Sometimes performing an RCT is relatively straight forward. When comparing a new drug to conventional therapy or absence thereof, few pitfalls should arise. No special diagnostic or technical prowess is necessary. Subjects need only be randomized and outcomes assessed by observers unaware of treatment status.
Conducting an RCT becomes more complex when technical skills are required. Without requisite skills, an RCT may yield misleading results. If a study seeks to compare surgical treatment versus no surgical treatment, the operator's technical prowess becomes paramount. Any complex procedure (like embryo biopsy) is mastered at different rates by different operators. Thus, the point in time vis-à-vis an individual operator's mastery of skill will clearly influence outcome of the procedure being evaluated. Unfortunately, many who design RCTs seem to assume technical skill as axiomatic, unnecessary to validate. Yet if technical skills are lacking the result of even the most elegantly designed RCT is not necessarily valid.
Elegantly designed RCTs in which inexperienced operators perform the procedures being evaluated have in fact generated invalid or arguable results. One example familiar to reproductive medicine specialists involves chorionic villus sampling (CVS), now widely applied throughout the world as an efficacious procedure to detect chromosomal abnormalities and other genetic disorders early in pregnancy. In experienced hands, procedure-related risks barely differ from those of second trimester amniocentesis, especially if CVS is performed using a transabdominal approach (ACOG Practice Bulletin, 2007
; Mujezinovic and Alfirevic, 2007). Yet acceptance of CVS was delayed in Europe and worldwide because a single RCT exceptionally concluded that undue risk existed.
The study that caused the untoward delay was a RCT conducted under aegis of the Medical Research Council (Medical Research Council European trial of chorion villus sampling, 1991); the trial was multi-centered, randomized and well designed. Its end-point was completed pregnancies, which proved to be 4.4% lower following CVS than following amniocentesis. The 4.4% difference in completed pregnancies was assumed to be mostly procedure-related, leading to the conclusion that CVS was a more risky procedure than amniocentesis. Was this interpretation valid and, hence, applicable for the wider medical community? Very likely not. The reason is that experience with CVS by the participating obstetricians was limited, certainly compared with their considerably greater experience with amniocentesis. Experience of operators in the MRC trial was also much less than that of operators in a US National Institutes of Health (NIH) trial, in which all obstetricians had considerable prior experience in performing CVS in ongoing pregnancies (Rhoads et al., 1989; Jackson et al., 1992). In the MRC study, the only requirement for participating obstetricians was the performance of 30 ‘practice’ CVS procedures. Even during the course of the MRC trial, certain centers arguably never gained dexterity, given the very few cases they contributed. Studies have since provided ample evidence that CVS skills improve with experience (Silver et al., 1990; Wijnberger et al., 2000).
That inexperienced operators conducting an elegant RCT can unwittingly generate spurious results due to their technical inexperience applies as well to procedures being evaluated in assisted reproductive technologies (ART).
Preimplantation genetic screening (PGS) requires blastomere removal, and has been offered for the years with ostensible success in selected centers (Gianaroli et al., 1999; Munné et al., 2003, 2006; Verlinsky et al., 2005). Large preimplantation genetic diagnosis (PGD) units have attempted to conduct a RCT to evaluate PGS, well aware that without a RCT the ‘gold standard’ has not been met. Unfortunately, RCTs have proved difficult to conduct in the USA, where most of the larger PGD centers are located. The major reason is that the National Institutes of Health proscribe funding for embryo research; thus, governmental funding for a PGS RCT is not possible in the USA. It might seem that a study could readily emanate from the private sector in the USA. However, impediments exist. First, ART patients in the USA usually pay ‘out of pocket’ for the procedure. They are thus understandably more vested in maximizing their own success than that in participating in a RCT. Second, institutional review boards in the USA often universally follow NIH guidelines required for funded research, even if no governmental funding is provided for a given study. Institutional review boards and their parent authorities fear that their portfolio of government funded research could be threatened collaterally should unfunded ‘embryo research’ go awry. Performing a RCT involving embryo biopsy may simply not be possible in many venues.
Into the breach have come RCTs for PGS conducted in locations not subject to US jurisdiction. Two such studies have received considerable attention raising the question of the future of PGS (Harper et al., 2008). The first was that of Staessen et al. (2004) who carried out a well-designed and executed RCT in couples aged 
37 years old. Beginning from a pool of 400 couples, the study compared results of 148 PGD cycles with that of 141 control cycles not subjected to PGS. Differing from protocol in other centers, two blastomeres were removed; chromosomes X,Y,13,16,18,21 and 22 were interrogated. The implantation rate (IR) (gestational sac plus fetal heart beat) per embryo transferred was 17.1% in the PGS group versus 11.5% in control group (P = 0.09). The percentage implantation beyond 12 weeks almost reached significance for PGS (16.5%) versus controls (10.4%) (P = 0.06). On the other hand, the control group showed 29 completed pregnancies, whereas the PGD group only 21. One possible explanation for failure to achieve significance is deleterious effects of removing two blastomeres.
In the well-publicized RCT on PGS of Mastenbroek et al. (2007a), only a single blastomere was removed. The conclusion was that PGS should not be offered to improve pregnancy rates in couples of advanced maternal age; Collins (2007) agreed. This may well be true for advanced maternal age per se, although in fact PGS is performed more often for more specific reasons (e.g. repeated pregnancy losses; advanced maternal age when there are five or more embryos to test). Irrespective, this author and others have contended that the poor results of Mastenbroek et al. could have reflected untoward effects of embryo biopsy (Munné et al., 2007a,b; Cohen and Grifo, 2007). These points have been answered by the authors Mastenbroek et al. (2007b). Nevertheless, it has been reasoned that the reported blastomere biopsy failure rate of 3% was higher than expected, potentially indicating unrecognized damage that could have occurred in ostensibly ‘successful’ biopsies. Further, the high rate (20%) for ‘no results’ could further have reflected embryo damage because a nucleus that has been lost or damaged poses difficulties in achieving an accurate diagnosis. Distinguishing diagnostic inaccuracy involving an intact nucleus from inability to analyze a damaged nucleus requires determining the diagnostic error rate as recommended in professional society guidelines (The Preimplantation Genetic Diagnosis International Society, 2007). That embryo damage in this particular RCT adversely affected pregnancy rates after PGS can in fact be deduced by examining the IRs observed following various scenarios reported by Mastenbroek et al. (2007a). In some cycles, an embryo was biopsied but no diagnosis made; transfer nonetheless took place (De facto this was a surgical sham.) These cases were still classified as PGS under the ‘intent to treat’ rationale. The IR was only 6%. In cycles in which there was no biopsy and no aneuploidy testing (i.e. no PGS) (a genuine control) embryo transfer yielded an IR of 14.7%. The 59% decrease in IR associated with embryo biopsy (alone) presumably reflects embryo damage. Maintaining embryo integrity after biopsy is integral for the success of PGS, which is predicated on the assumption that the deleterious effect of loss of a single blastomere (10% decreased viability) (Cohen et al., 2007) can be overcome by the transfer of euploid embryos. Improvement in the pregnancy rate must be much >10% compared with non-tested embryos. In reality, transfer of euploid embryos in Mastenbroaek et al. showed a 13% improvement in comparison to pregnancies not undergoing PGS –16.8 versus 14.7% (Munné et al., 2007b; Wilton, 2007).
|
Conclusion |
|---|
|
TopAbstractIntroductionRCTs requiring technical skillsConclusionReferences |
|---|
Surgeons and embryologists must utilize RCTs to assess their results. There is no argument. Similarly, epidemiologists must validate surgical or technical expertise of their operators. Otherwise, RCTs in reproductive medicine that involve technical prowess are neither valid nor generalizable. RCTs specifically assessing PGS require three general prerequisites—proper study design, skilled operators (surgical procedure or embryo biopsy) and accurate genetic diagnosis.
|
References |
|---|
|
TopAbstractIntroductionRCTs requiring technical skillsConclusionReferences |
|---|
ACOG Practice Bulletin. Invasive prenatal testing for aneuploidy. In: ACOG Practice Bulletin 88 (2007) 110. Washington, DC: American College of Obstetrics and Gynecologists. 1459–1467. Obstet Gynecol.[Medline]
Cohen J, Grifo JA. Multicentre trial of preimplantation genetic screening report in the New England Journal of Medicine: an in-depth look at the findings. Reprod Biomed Online (2007) 15:365–366.[Web of Science][Medline]
Cohen J, Wells D, Munné S. Removal of 2 cells from cleavage stage embryos is likely to reduce the efficacy of chromosomal tests that are used to enhance implantation rates. Fertil Steril (2007) 87:496–503.[CrossRef][Web of Science][Medline]
Collins JA. Preimplantation genetics screening in older mothers. N Engl J Med (2007) 357:61–63.
Gianaroli L, Magli MC, Ferraretti AP, Munné S. Preimplantation genetic 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 (1999) 72:837–844.[CrossRef][Web of Science][Medline]
Harper J, Sermon K, Geraedts J, Vesela K, Harton G, Thornhill A, Pehlivan T, Fiorentino F, SenGupta S, de Die-Smulders C, et al. What next for preimplantation genetic screening? Hum Reprod (2008) 23:478–480.
Jackson LG, Zachary JM, Desnick RJ, Golbus MS, Ledbetter DH, Mahoney MJ, Pergament E, Schulman JD, Simpson JL, Fowler SE, and the NIH CVS Amniocentesis Study Group. A randomized comparison of transcervical and transabdominal chorionic villus sampling. N Engl J Med (1992) 327:594–598.[Abstract]
Mastenbroek S, Twisk M, van Echten-Arends J, Sikkema-Raddatz BS, Korevaar JC, Verhoeve HR, Vogel N, Arts E, de Vries J, Bossuyt PM, et al. In vitro fertilization with preimplantation genetic screening. N Engl J Med (2007) a 357:9–17.
Mastenbroek S, van der Veen F, Repping S. In vitro fertilization with preimplantation genetic screening. N Engl J Med (2007) b 357:1770–1771.[Web of Science][Medline]
Medical Research Council European trial of chorion villus sampling. MRC working party on the evaluation of chorion villus sampling. Lancet (1991) 337:1491–1499.[CrossRef][Web of Science][Medline]
Mujezinovic F, Alfirevic Z. Procedure-related complications of amniocentesis and chorionic villus sampling: a systematic review. Obstet Gynecol (2007) 110:678–694.
Munné S, Sandalinas M, Escudero T, Velilla E, Walmsley R, Sadowy S, Cohen J, Sable D. Improved implantation after preimplantation genetic diagnosis of aneuploidy. Reprod Biomed Online (2003) 7:91–97.[Medline]
Munné S, Fischer J, Warner A, Chen S, Zouves C, Cohen J. Preimplantation diagnosis significantly reduces pregnancy loss in infertile couples: a multicentre study. Fertil Steril (2006) 85:326–332.[CrossRef][Web of Science][Medline]
Munné S, Gianaroli L, Tur-Kaspa I, Magli C, Sandalinas M, Grifo J, Cram D, Kahraman S, Verlinsky Y, Simpson J. Substandard application of preimplantation genetic screening may interfere with its clinical success. Fertil Steril (2007) a 88:781–784.[CrossRef][Web of Science][Medline]
Munné S, Cohen J, Simpson JL. In vitro fertilization with preimplantation genetic screening. N Engl J Med (2007) b 357:1769–1770.
Rhoads GG, Jackson LG, Schlesselman SE, de la Cruz FF, Desnick RJ, Golbus MS, Ledbetter DH, Lubs HA, Mahoney MJ, Pergament E, et al. The Safety and efficacy of chorionic villus sampling for early prenatal diagnosis of cytogenetic abnormalities. N Engl J Med (1989) 320:609–617.[Abstract]
Silver RK, MacGregor SN, Sholl JS, Hobart ED, Waldee JK. An evaluation of the chorionic villus sampling learning curve. Am J Obstet Gynecol (1990) 163:917–922.[Web of Science][Medline]
Staessen C, Platteau P, Van AE, Michiels A, Tournaye H, Camus M, Devroey P, Liebaers I, Van Steirtegham A. 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 (2004) 19:2849–2858.
The Preimplantation Genetic Diagnosis International Society (PGDIS). Guidelines for good practice in PGD: program requirements and laboratory quality assurance. Repro Biomed Online (2008) 16:134–147.
Verlinsky Y, Tur-Kapsa I, Cieslak J, Bernal A, Morris R, Taranissi M, Kaplan B, Kulieu A. Preimplantation testing for chromosomal disorders improves reproductive outcome of poor-prognosis IVF patients. Repro Biomed Online (2005) 11:219–225.
Wijnberger LD, van der Schouw YT, Christiaens GC. Learning in medicine: chorionic villus sampling. Prenat Diag (2000) 20:241–246.[CrossRef][Web of Science][Medline]
Wilton LJ. In vitro fertilization with preimplantation genetic screening. N Engl J Med (2007) 357:1770.[Web of Science][Medline]
Submitted on December 21, 2007; resubmitted on April 21, 2008; accepted on May 22, 2008.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Vanneste, T. Voet, C. Melotte, S. Debrock, K. Sermon, C. Staessen, I. Liebaers, J.-P. Fryns, T. D'Hooghe, and J. R. Vermeesch What next for preimplantation genetic screening? High mitotic chromosome instability rate provides the biological basis for the low success rate Hum. Reprod., November 1, 2009; 24(11): 2679 - 2682. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. R. Hernandez What next for preimplantation genetic screening? Beyond aneuploidy Hum. Reprod., July 1, 2009; 24(7): 1538 - 1541. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Fritz Perspectives on the efficacy and indications for preimplantation genetic screening: where are we now? Hum. Reprod., December 1, 2008; 23(12): 2617 - 2621. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||