What is the most relevant standard of success in assisted reproduction?

Is BESST (birth emphasizing a successful singleton at term) truly the best?

Georg Griesinger1,2, Konstantinos Dafopoulos1, Askan Schultze-Mosgau1, Ricardo Felberbaum1 and Klaus Diedrich1

1 Department of Obstetrics and Gynecology, Medical University of Schleswig-Holstein, Campus Luebeck, Ratzeburger Allee 160, D-23538 Luebeck, Germany

2 To whom correspondence should be addressed. e-mail: georg.griesinger{at}frauenklinik.uni-luebeck.de


    Abstract
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
There is much variability and no consensus on the definition of the most relevant outcome parameter after assisted reproduction technology (ART). Descriptive reports, such as annual statistics from national registries on the success of ART programmes, should present treatment success in terms of live birth per ovarian stimulation started, as this is the most relevant information for patients and doctors alike. Addressing concerns about the high rate of multiple pregnancies, rescaling the outcome of ART in large programmes and national audits to the singleton, live birth, might trigger a global change of attitude towards elective single embryo transfer in addition to any legal restrictions imposed. For clinical studies, the outcome measure will depend on the hypothesis tested, and investigators should remain free to choose the appropriate primary outcome measure.

Key words: assisted reproduction/outcome measures/success rates


    Introduction
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
A recent article published in Human Reproduction by Min and colleagues (Min et al., 2004Go) encourages programmes and registries to report the singleton, term gestation, live birth rate per treatment cycle initiated as the primary end-point parameter of assisted reproductive technology (ART), in terms of redefining ‘success’ in infertility treatment. Similarly, an ESHRE consensus meeting report (Land and Evers, 2003Go) recently stated that the essential aim of IVF/ICSI is the birth of one single healthy child, with a twin pregnancy being regarded as a complication.


    The drive towards elective single embryo transfer (eSET)
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
Elective single embryo transfer (eSET) as the future gold standard in ART is evidently the only way to reduce twin pregnancies further (ESHRE Campus Course Report, 2001Go). However, confidence in the widespread adoption of this standard and reliance on the self-limitation of the IVF community has as yet, for several possible reasons, not proven sufficient: (i) twins have become and are still an accepted side effect of a high pregnancy rate after IVF therapy conducted under short-sighted economical pressure on the part of the patient and the stress of competition among centres; (ii) legal premises, such as the German embryo protection law, that ban embryo selection or embryo freezing might render eSET impossible when weighting the chance of pregnancy against the costs and burden of the procedure; and (iii) studies testing the concept of eSET on a large scale are missing. It appears as if the existing data on eSET are not as yet regarded as being sufficiently convincing by the majority of the IVF community to implement eSET as part of routine practice. Thus, in an attempt to further the progress of reducing twin gestations after ART on a wider basis, rescaling the outcome of ART in large programmes and national audits to the singleton pregnancy or singleton live birth might trigger a global change of attitude towards eSET besides legal restrictions.


    How to measure success after ART
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
An outcome parameter is an arbitrarily defined measure enabling the clinician to study, understand and eventually control variables that influence the success of a therapy. Thus, outcome parameters vary in clinical ART studies. If pregnancy rate or live birth rate is the outcome of choice in a randomized controlled trial, the denominator should be the number of patients randomized—no matter if a subject actually received treatment or not, as elaborated in detail elsewhere (Daya, 2003Go).

For national statistics of ART programmes as a source of information for patients and doctors, we advocate as the denominator the number of patients that actually started ovarian stimulation, with live birth as the numerator, as this provides the most relevant information. National annual reports on success rates from England (http://www.hfea.gov.uk), Germany (http://www.deutsches-ivf-register.de) and the USA (http://www.cdc.gov/nccdphp/drh/art.htm), among others, provide this information.

Choosing the appropriate numerator for clinical studies is more complex and depends on the trial hypothesis and the intervention under study. Thus, investigators choose a variety of numerators, such as oocyte number, oocyte fertilization, embryo morphology and implantation, biochemical pregnancy, clinical pregnancy, clinical ongoing pregnancy or live birth, etc.

We advocate that investigators should remain free to choose a primary outcome parameter to detect immediately a change in outcome by a specific intervention and report studies in line with the CONSORT statement (Moher et al., 2001Go). However, all trialists should report the ongoing clinical pregnancy rate, defined as fetal heart activity on ultrasonography at 12 weeks of gestation, before any invasive prenatal testing, independently of whether this is the primary outcome parameter upon which a sample size estimation is based or not. This outcome parameter provides a high level of assurance since the majority of pregnancy losses occur in the first trimester. Uniformly reporting pregnancy after IVF by this outcome parameter will also facilitate incorporation of data from individual studies into meta-analyses. Beyond that, data from interim outcomes, such as number of oocytes, fertilization, implantation, biochemical pregnancy, abortion, etc., should also be provided. Ideally, a minimum consensus should be established, i.e. what outcome measures should be included in the report from a randomized controlled trial. Arbitrarily setting the primary outcome to live birth rate in studies will not only further increase the sample size necessary to achieve sufficient power for a trial, but will also increase the time until publication of data, as well as increase the effort required by both patients and doctors for follow-up and a standardized antenatal and obstetric care.

In that context, we would like to emphasize that ART is primarily established as a treatment to overcome infertility (in the sense of an inability to conceive), which is reflected by our daily clinical practice: a patient who has become pregnant by means of assisted reproduction is usually no longer attended to by the infertility specialist. Thus, as long as we have no knowledge of to what extent the ART procedure itself is causative of the likelihood of miscarriage and adverse events during pregnancy (reviewed by Lambert, 2003Go), it may be inappropriate to use live birth—a measure that is comparatively remote from the intervention performed and studied in a clinical ART trial—as a primary outcome parameter. However, if we decide to do so under the assumption that the treatment itself, rather than underlying (unknown) pathologies associated with infertility, has an impact on the course of pregnancy, gestational age at delivery, live birth and health of the neonate, we will face a variety of confounding factors in our studies imposed by the policy of prenatal diagnosis/antenatal care/obstetric care, which are far from being standardized, and which may vary considerably from centre to centre (and may even be outside a centre’s control).

Setting the primary outcome furthermore to the singleton pregnancy or singleton live birth is of no apparent benefit in the context of a clinical study as long as the intervention under study does not aim specifically at reducing the multiple pregnancy rate.


    Should ‘term gestation’ be part of the main outcome parameter?
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
The answer to this question is no, but gestational age and weight at delivery should be reported in registries for all births after ART, as is done for example in the French registry (http://perso.wanadoo.fr/fivnat.fr). It is unclear why a baby who has no morbidity born at 35 weeks of gestation is not a success of ART treatment, as implied by Min et al. (2004Go). Arbitrarily excluding babies that might be healthy but that have been born prematurely from a success report as suggested by the ‘BESST’ outcome goes beyond a universally acceptable message. This is in agreement with a trend in the field of antenatal care that aims at redefining ‘preterm’ to an earlier time point than 37 weeks of gestation. Not only has the threshold of gestational age for neonatal survival been pushed earlier into pregnancy, but also the incidence of complications attributable to preterm delivery becomes indistinguishable after 34 weeks of gestation in larger preterm infants as compared with term infants (Robertson et al., 1992Go).

Generally, preterm birth is associated with, besides multiplicity, preterm rupture of membranes, cervical incompetence, polyhydramnion, fetal and uterine anomalies, infections, ethnicity, genetic risk, lifestyle and many other factors. Although we know that the rate of preterm deliveries after ART is higher compared with control populations even in singletons (Dhont et al., 1999Go; Perri et al., 2001Go; Wang et al., 2002Go), the specific contribution of ART treatment (e.g. ovarian stimulation, laboratory techniques) to this phenomenon remains obscure. Thus, with current understanding, gestational age at delivery is not a reasonable outcome measure of the success and the quality of an ART programme.


    Should the ‘health’ of children born from IVF be included in the outcome reports?
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
We note that although Min et al. (2004Go) state that ‘success must reflect the delivery of healthy babies’, we find no specific information about the ‘healthiness’ of the term singletons in their paper. Correspondingly, to our knowledge, no national registry provides information on the health of children born from ART.

In the paper by Min et al. (2004Go), assessment of health of the neonates in their programme was done by ‘neonatal examinations conducted by paediatricians’. A single neonatal examination of babies born from ART is insufficient to detect all health problems of the infant. From a patient perspective, reporting the healthiness of children after ART as is done by Min et al. (2004Go) automatically generates the need for information regarding the children classified as ‘not healthy’ and the relative risk of having a child affected by a specific disease after IVF. However, information as such is best retrieved from controlled studies employing standardized diagnostic criteria (e.g. International Classification of Disease Codes for neonatal morbidity or malformation), karyotyping and longer term follow-up regimens. Not all health problems of the neonate will be evident at birth, and ‘adverse outcomes’ of the neonate are heterogenous and vary considerably in impairment of the quality of life of both infant and parents. Such issues have to be addressed at counselling with regard to the specific background of a couple, including the aetiology for infertility. In the context of registries, adding the classification ‘healthy’ to the outcome report is unwarranted and of no clear benefit. However, we acknowledge the need for uniformly recording pregnancy complications, maternal morbidity, fetal morbidity/mortality and malformations in the setting of national audits.


    Conclusions
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
For the patient, the most informative outcome measure of ART in descriptive reports from ART programmes is the singleton live birth rate per treatment cycle started. For the medical community, it is paramount to document further at a larger scale the outcome of IVF; however, gestational age at delivery does not seem appropriate to measure the professional competence of a centre or programme. The health of children born from ART should not be subsumed in one simple outcome parameter. For clinical trials in subfertility, various outcome parameters may be appropriate; however, trialists should be encouraged to report the ongoing clinical pregnancy rate as a minimum requirement.


    References
 Top
 Abstract
 Introduction
 The drive towards elective...
 How to measure success...
 Should ‘term...
 Should the ‘health’...
 Conclusions
 References
 
Daya S (2003) Pitfalls in the design and analysis of efficacy trials in subfertility. Hum Reprod 18,1005–1009.[Free Full Text]

Dhont M, De Sutter P, Ruyssinck G, Martens G and Bekaert A (1999) Perinatal outcome of pregnancies after assisted reproduction: a case–control study. Am J Obstet Gynecol 181,688–695.[Medline]

ESHRE Campus Course Report. (2001) Prevention of twin pregnancies after IVF/ICSI by single embryo transfer. Hum Reprod 16,790–800.[Abstract/Free Full Text]

Lambert RD (2003) Safety issues in assisted reproductive technology: aetiology of health problems in singleton ART babies. Hum Reprod 18,1987–1991.[Abstract/Free Full Text]

Land JA and Evers JL (2003) Risks and complications in assisted reproduction techniques: report of an ESHRE consensus meeting. Hum Reprod 18,455–457.[Abstract/Free Full Text]

Min JK, Breheny SA, MacLachlan V and Healy DL (2004) What is the most relevant standard of success in assisted reproduction? The singleton, term gestation, live birth rate per cycle initiated: the BESST endpoint for assisted reproduction. Hum Reprod 19,3–7.[Abstract/Free Full Text]

Moher D, Schulz KF and Altman DG (2001) The CONSORT statement: revised recommendations for improving the quality of reports of parallel group randomized trials. BMC Med Res Methodol 1–2.

Perri T, Chen R, Yoeli R, Merlob P, Orvieto R, Shalev Y, Ben-Rafael Z and Bar-Hava I (2001) Are singleton assisted reproductive technology pregnancies at risk of prematurity? J Assist Reprod Genet 18,245–249.[CrossRef][Medline]

Robertson PA, Sniderman SH, Laros RK Jr, Cowan R, Heilbron D, Goldenberg RL, Iams JD and Creasy RK (1992) Neonatal morbidity according to gestational age and birth weight from five tertiary care centers in the United States, 1983 through 1986. Am J Obstet Gynecol 166,1629–1641.[Medline]

VandenEede B (1995) Investigation and treatment of infertile couples: ESHRE guidelines for good clinical and laboratory practice. European Society of Human Reproduction and Embryology. Hum Reprod 10,1246–1271.[Medline]

Wang JX, Norman RJ and Kristiansson P (2002) The effect of various infertility treatments on the risk of preterm birth. Hum Reprod 17,945–949.[Abstract/Free Full Text]