Human Reproduction, Vol.30, No.12 pp. 2703– 2707, 2015
Advanced Access publication on October 13, 2015 doi:10.1093/humrep/dev263
OPINION
Cumulative live birth rate: time
for a consensus?
Abha Maheshwari 1,*, David McLernon 2, and Siladitya Bhattacharya2
1
Aberdeen Fertility Centre, NHS Grampian, Foresterhill, Aberdeen, UK 2Division of Applied Health Sciences, University of Aberdeen, Polwarth
Building, Foresterhill, Aberdeen AB24 2ZD, UK
*Correspondence address. Aberdeen Fertility Centre, NHS Grampian, Foresterhill, Aberdeen AB25 2ZL, UK. Tel: +44-1224-438426;
E-mail: [email protected]
Submitted on August 2, 2015; resubmitted on September 13, 2015; accepted on September 24, 2015
abstract: Traditionally, IVF success rates have been reported in terms of live birth per fresh cycle or embryo transfer. With the increasing use
of embryo freezing and thawing it is essential that we report not only outcomes following fresh but also those after frozen embryo transfer as a
complete measure of success of an IVF treatment. Most people agree that an individual’s chance of having a baby following fresh and frozen embryo
transfer should be described as cumulative live birth rate. However, views on the most appropriate parameters required to calculate such an
outcome have been inconsistent. There is an additional dimension—time for all frozen embryos to be used up by a couple, which can influence
the outcome. Given that cumulative live birth rate is generally perceived to be the preferred reporting system in IVF, it is time to have an
international consensus on how this statistic is calculated, reported and interpreted by stakeholders across the world.
Key words: cumulative live birth rate / fresh embryo transfer / frozen embryo transfer / IVF, ICSI / pregnancy / life tables / embryo transfer
Background
The purpose of publishing data on the outcome of in vitro fertilization
(IVF) is to provide stakeholders with information that addresses their
specific needs. Clinicians need to be able to audit and evaluate care;
consumers value a reliable way of understanding rates of success and
complications while purchasers, regulators and politicians are interested
in the demand for treatment, developments in service delivery and
trends in clinical outcomes. For all, an ideal outcome is one, which
provides a meaningful summary of the effectiveness of the treatment
and its safety (Tiitinen et al., 2004).
Traditionally, the success of IVF has been reported on the basis of the
outcome of a treatment initiated with the intention of replacing one or
more fresh embryos by selecting the best embryo(s) for transfer within
the uterus. Since the first live birth associated with a thawed frozen
embryo was reported in 1984, freezing surplus embryos has become a
feasible and increasingly popular option. With refinement of technology
in recent years, the numbers of thawed frozen embryo transfers have
increased (De Mouzon et al., 2010) as have pregnancy rates associated
with them (Roque et al., 2013). This practice has been further encouraged by the strategy of single embryo transfer and an emphasis on
preventing ovarian hyperstimulation in at-risk women. Reporting IVF
success rates based on a fresh embryo transfer alone is meaningless in
these situations and therefore outdated. The reporting of an IVF treatment should not only incorporate outcomes associated with fresh
embryo transfers but also those resulting from the replacement of
surplus frozen thawed embryos to provide an all-inclusive success rate
which is comprehensive, relevant and meaningful. In this article we
discuss how this can be reported, what the challenges are, and how
these can be overcome.
Cumulative live birth rate: an
all-inclusive success rate
Cumulative live birth rate (CLBR) has been suggested as a suitable way of
reporting success of an IVF programme which incorporates fresh as well
as thawed frozen embryo transfer (Germond et al., 2004). The term cumulative means increasing in quantity by successive additions. In the
context of IVF it implies capturing the totality of live birth episodes following successive treatments. Conventionally, most IVF studies examine and
report on the ‘per cycle’ or ‘per embryo transfer episode’ chance of pregnancy and delivery (Olivius et al., 2002). However, after an unsuccessful
fresh cycle, couples often want to know what their chance is of having a
live birth if they continue with further ART treatment—either utilizing
any remaining frozen embryos accruing from their initial attempt or
after further fresh treatments. Thus, from the patients’ perspective,
the cumulative rate of live birth is more important (Malizia et al., 2013)
since it better summarizes the chance of a live birth over an entire treatment period (Malizia et al., 2009) and is most pertinent to a couple’s
decision on whether to undertake further IVF or not (Hull, 1994).
For clinicians, CLBR per oocyte retrieval is more meaningful as it is a
much better indicator of quality and success in IVF in its totality as cryopreservation has become an integral part of IVF (Garrido et al., 2011;
& The Author 2015. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: [email protected]
2704
Li et al., 2013). It has been suggested that the outcome of a complete IVF
cycle (including cryopreservation) allows better comparisons between
different centres (Lintsen et al., 2010) which may possess different strategies for freezing and extended culture of embryos. Reporting CLBR
rather than success rates based on fresh embryo transfer will be more
appropriate for making economic and political decisions with regards
to treatment efficacy and costs of reimbursement.
Despite these arguments, most national registries do not routinely
report cumulative rates. This is due to a number of reasons as described
below:
Definition of cumulative
live birth rate
Defining cumulative birth rate itself is a major challenge. Computing rate
of an event needs agreement about the most appropriate numerator and
the denominator.
Numerator for cumulative live birth rate
Some studies have used first live birth (Thurin-Kjellberg et al., 2009; Luke
et al., 2012; Stern et al., 2013; Bodri et al., 2014) as the preferred numerator while others have included all live birth episodes from an index
stimulation cycle (Li et al., 2014). There are advantages and disadvantages of both approaches. Using the number of women who have at
least one live birth as the numerator highlights the outcome for which
couples initially sought treatment. However, as it is unable to capture
the total benefit associated with a single egg collection, it cannot be
used as a measure of ART efficiency.
There are no existing data to indicate what stakeholders prioritize as
their final outcome—a single live birth or all live births from one egg collection. Preferences may also depend on their perspective—thus couples
undergoing treatment, commissioners or clinics may differ in their choice.
Using all live births as a numerator may give a rate of more than unity (e.g.
200%), which may not be intuitive or mathematically acceptable.
Choosing the right numerator is critical—not just in terms of communicating IVF results to patients and the public, but also in evaluating the
efficacy of newly introduced treatments.
Denominator for cumulative live birth rate
The denominator could be all women who attempt ovarian stimulation
as part of IVF or all those who have undergone egg collection. In addition,
there is no consensus on how results should be interpreted for those
who return for a repeat cycle of IVF after an initial live birth. In some
studies these women were not re-enrolled in the data collection
system after their first delivery (Bodri et al., 2014), while other researchers included them in any analysis as a ‘new patient under observation’ if
they underwent further ART (Gnoth et al., 2011). Women who failed to
return (either because they moved to another IVF centre or stopped
treatment for any other reason) were censored after the last treatment
(Gnoth et al., 2011). However, this makes the egg collection rather than
the woman the unit of analysis.
There are two further components of the denominator:
(a) Unit of IVF treatment.
Assuming that the woman is the ideal denominator, it is important
to determine the number of episodes of IVF treatment over which
Maheshwari et al.
the cumulative live birth rate will be estimated. Some authors have
suggested this should be all successive embryo transfers until live
birth is achieved (Garrido et al., 2011) or only cycles which have
cryopreserved embryos leading to fresh and some frozen embryo
transfers (Bodri et al., 2014). Other studies assessed cumulative
live birth rates over multiple fresh or frozen treatments which can
be difficult to interpret. Much more intuitive is a policy of reporting
cumulative rates over multiple complete cycles of treatment (where
a complete cycle is defined as all fresh and frozen embryo transfers
resulting from one episode of ovarian stimulation). However, this
approach assumes that couples use up all their frozen embryos
from previous egg collections before proceeding to the next egg collection, which may not be the case for all.
Cumulative live birth rate can be reported over a course of a complete therapeutic episode but there is no universal agreement as to
how many cycles of treatment constitute ‘complete therapy’ (Daya,
2005) and when ‘treatment’ is deemed to start and finish. It can
vary from a single egg collection to any number of egg collections
(Ke et al., 2013). The National Institute of Health & Care Excellence
(NICE) defines ‘3 full cycles’ as 3 egg collections including all the fresh
and frozen embryo transfers related to these. For a woman undergoing a number of egg collections, the cumulative live birth rate could be
reported over multiple complete cycles of treatment (see Table I).
The denominator here is per woman and it does not take account
of drop off after first cycle. Moreover, this approach makes contemporaneous reporting of current or recent results impossible as one
would have to wait till couples have completed a number of treatments. This will lead to presentation of results, which are outdated
and unrepresentative of current practice, and therefore less relevant
to most stakeholders, when presented at the end of three cycles.
Reporting the number of cycles is of significance in situations where
some centres perform multiple egg collections for a woman, create
embryos and freeze them. Once a desired number of frozen
embryos are available, they are thawed and the best quality ones
are transferred. There are sometimes specific cases—such as
women with suboptimal ovarian reserve in which patients would
choose to undergo several potentially risky (ovarian stimulation and
Table I How the conservative cumulative live birth rate
is calculated over three complete cycles of IVF.
Cycle 1
Cycle 2
Cycle 3
........................................................................................
Number of women
entering treatment
N1
N2
N3
No. of women with at
least one live birth
n1
n2
n3
Live birth rate per cycle
n1/N1
n2/N2
n3/N3
Cumulative live birth rate
n1/N1
(n1 + n2)/N1 (n1 + n2 + n3)/N1
N2 ¼ N1 2 n1 2 w1.
N3 ¼ N1 2 n1 2 n2 2 w1 2 w2.
wi ¼ the number of women who discontinued treatment without a live birth after
cycle i.
Here, denominator ¼ N1; numerator ¼ number of women with at least one live birth;
Cycle i includes all fresh and frozen embryo transfers resulting from the ith episode of
ovarian stimulation.
2705
Cumulative live birth
egg collection) procedures in order to bank embryos prior to full utilization. In these cases, using live birth as a numerator but specifying the
number of cycles it took to achieve this outcome will be informative.
Another reason for specifying the number of cycles is that the precise
protocol used may influence the number of stimulation episodes over
a given period of time. Thus, a couple may receive more spontaneous
(modified natural) cycles or antagonist cycles in comparison with long
down-regulation cycles using GnRH agonists, which will impact on the
overall outcomes.
(b) Time horizon.
Close linked with the number of treatments/cycles is the issue of
the time horizon over which treatment is provided. There needs to
be balance between contemporaneous data reporting and providing
adequate time for couples to use their frozen embryos. Various
methods have been advocated to take account of this. In some previous reports, women with frozen embryo(s) which have not been
thawed by the end of an agreed follow-up period have been
excluded from cumulative live birth analyses (Li et al., 2014). In
others, conservative cumulative live birth rates have been calculated
by dividing the number of women achieving live birth within a predetermined number of cycles, by the total number of women starting
treatment. A third method estimates the cumulative live birth delivery rate after a specified number of cycles using a life-table analysis,
by taking into account the effects of censoring (drop-out). The
outcome measure associated with this method is referred to as
‘Expected cumulative delivery rate’ (De Brucker et al., 2009).
These three methods cannot be compared directly as they are
based on different assumptions.
Once we agree on the numerator and denominator, we need to
specify the intended method of analysis.
Method of analysis
Life-table analysis has been used to estimate the success rates of various
practices in the field of reproductive medicine including assisted reproduction (Hull et al., 1992). To use this method it is assumed that women who
discontinue the treatment would have had similar chances of live birth as
those who continue with further treatments—the independent censoring
assumption (Stolwijk et al., 1996). However, some studies have found that
women discontinue the treatment because of poor treatment prognosis,
i.e. that the reason for censoring was not independent of the outcome
(Doody, 1993; Walters, 1994). However, others (Haan et al., 1991)
have not found any over-representation of patients with a poor prognosis
in the group of dropouts.
to the fact that length of time it takes to complete one full cycle of treatment will depend on the number of frozen embryos and the number of
menstrual cycles required to transfer all these embryos. Moreover,
women may have a gap between various episodes of embryo transfers
for various reasons. It has been suggested (Engmann et al., 1999) that
when calculating cumulative delivery rates by life-table analysis, the age
of each woman should be recorded at the start of her first treatment
cycle, even though this can cause an underestimation of age-related cumulative delivery rates. Another disadvantage of life-table analysis is that
it also tends to overestimate the cumulative delivery rates when the
group is too small (De Brucker et al., 2009).
Other model based approaches can be used to estimate the probability of cumulative live birth in individual couples based on characteristics of
the couple and their treatment (Missmer et al., 2011). These are beyond
the scope of this discussion.
Discussion
Until agreement is reached on a consistent approach to reporting outcomes, it will be difficult to compare CLBR in ART across different
studies or national registry reports. Some authors tend to publish
three different outcomes: the conditional live birth rate at a specific
cycle; a conservative estimate of the cumulative live birth rate, which
was based on the assumption that none of the women who did not
return for a subsequent cycle would have had a live birth; and an
optimal estimate of the cumulative live birth rate, which was based on
the assumption that women who did not return for a subsequent cycle
would have had the same live birth rates as those who did return
(Luke et al., 2012). This latter method represents an ideal world scenario
where there are no barriers to treatment continuation. However, while
helpful from a research perspective, this is unhelpful for comparing clinics
and treatments by some stakeholders. Therefore, it is time to strive
towards a consensus on a standard numerator, denominator (including
unit of treatment and time horizon) and a method of analysis from the
range that is available (Table II).
Preferences of stakeholders
Preferences of stakeholders on the best way of expressing cumulative IVF
outcomes are not available in the literature. It may be that different
groups have different perspectives and one definition may not be suitable
for all. The suggested time horizon must be sensitive to the needs of
couples undergoing treatment—an issue, which needs to be explored
further.
Recommendations
Use of life table approach for multiple egg collections
The use of life table approach for multiple egg collections is associated
with some limitations. In life table approach, used in this context, it is
assumed that events are happening in consecutive menstrual cycles.
Therefore, two women, each of whom have had three cycles of treatment, are considered similar for the purposes of comparison and data.
However, in reality, there will be a difference in the time period during
which each of them have these three full IVF cycles (i.e. fresh as well as
linked frozen embryo transfer). There will also be a difference in the
age at which second and third egg collections are performed in different
women, even if first egg collection was done at the same age. This is due
To take account of the above factors and to avoid undue confusion we
would recommend reporting of cumulative live birth rate over short-,
medium- and long-term time horizons. Although, in the current electronic era, stakeholders could theoretically choose to view their preferred
outcomes by means of drop down menus, such a laissez-faire approach
towards data dissemination has the potential to complicate rather than
simplify the process and lead to serious inconsistency in interpretation
of results.
The most important requirement is to have a clearly defined numerator and denominator that clinicians/researchers must report alongside
the outcome rate, as well as clarity around the statistical method used to
2706
Maheshwari et al.
Table II Components for cumulative live birth rate.
List of numerators
List of denominators
Units of treatment
Method of analysis
.............................................................................................................................................................................................
All live birth episodes
Number of women who attempted stimulation
Per one/multiple egg collection
Crude rates
Conditional rates
Life table
At least one live birth episode
Number of women who had egg collection
Per n number of embryo transfers
Per fixed time period
† Conservative estimate
† Optimistic estimate
Table III Suggestions for reporting cumulative live birth
rate.
Suggested parameters
Account for
........................................................................................
Short
term
Live birth episodes/woman/
one egg collection in a 2-year
period
First live birth
Medium
term
Live birth episodes/woman/
one egg collection in a 5-year
period
All live birth episodes
(including second and further
live birth) from one egg
collection
Long term
Live birth episodes per
woman/3 egg collections
over 10 years
All live births from 3 egg
collections
compute it. We may have to accept a three-step approach towards
reporting cumulative live birth rate in the short, medium and long term.
Short term. To measure short-term outcomes, CLBR can be presented as
live birth episodes per woman per egg collection over a 2-year period to
account for the first live birth. A 2-year period is chosen as women are
likely to use all their frozen embryos but unlikely to have two live birth
episodes during this period.
Medium term. It is assumed that most, if not all, frozen embryos from a
single egg collection will have been thawed and transferred within 5
years. To measure medium-term outcomes CLBR should be reported
as live birth episodes per woman per one egg collection in a 5-year
period. This will account for all live birth episodes (including second
and further live births) from one egg collection.
Long term. For long-term outcomes the CLBR could be reported as live
birth episodes per woman per three egg collections over 10 years. Since
the number of patients having three or more complete treatment cycles
in each individual centre will be small, this will need to be reported at a
national level to help policy makers make decisions about the number
of cycles that should be supported by public funding. The assumption
is made that most couples would have completed treatment within
this time duration.
It is worth noting that this time based approach is based on number of
egg collections and is unable to take into account the opportunities for
conception (e.g. embryo transfers) over a given period. As cumulative
live birth rates after 2, 5 or 10 years could be the same irrespective
of the number of embryo replacements, this method of reporting
cumulative outcomes will not be able to provide data on number of treatments needed to conceive or identify clinics where patients require fewer
cycles to achieve a pregnancy leading to live birth.
Our suggestions, summarized in Table III relate to CLBR reporting
only. We acknowledge that other relevant outcomes include measures,
which determine the burden of treatment, risks of OHSS, patient experience and long-term safety for a woman and her baby. These are issues for
consideration in order to capture a global view of the full impact of IVF
treatment. The challenge lies in the practicality of achieving this—both
in terms of tools and definitions used as well as the extended time
horizon required for some of these. Clinics do not normally evaluate
the patient experience and indeed, given that two individuals are
involved, these experiences may be different for each of them. Longterm follow-up is ideal, but, as discussed earlier in the manuscript,
truly meaningful long-term follow-up would mean reporting outcomes
many years after treatment. While this is important, the two serve separate functions and cannot necessarily be combined in the same report
and for the same purpose.
Conclusion
Patients beginning a course of IVF treatment need information about
their chances of having a baby. This can be measured as the chance for
live birth per treatment (fresh or frozen IVF) but it can also be presented
as the likelihood of live birth following repeated treatments.
Although, it is possible to report on cumulative live birth rates, several
different definitions are used. This makes it difficult to generate a single
comprehensible headline figure, which captures the success rate of
IVF/ICSI. Until agreement on this is achieved, we recommend a triple
outcome strategy for reporting short-, medium- and long-term results.
Authors’ roles
A.M. conceived the idea and wrote the initial draft. D.M. contributed
to the manuscript, especially in terms of elaborating on the statistical
concepts. S.B. contributed to the development of the initial concept
and provided intellectual input in structuring the arguments in the text.
All authors contributed to drafting and revising the manuscript. They
all approved the final version.
Funding
No external funding was obtained for this manuscript
Cumulative live birth
Conflict of interest
None declared.
References
Bodri D, Kawachiya S, De Brucker M, Tournaye H, Kondo M, Kato R,
Matsumoto T. Cumulative success rates following mild IVF in unselected
infertile patients: a 3-year, single-centre cohort study. Reprod Biomed
Online 2014;28:572– 581.
Daya S. Life table (survival) analysis to generate cumulative pregnancy rates in
assisted reproduction: are we overestimating our success rates? Hum
Reprod 2005;20:1135 – 1143.
De Brucker M, Haentjens P, Evenepoel J, Devroey P, Collins J, Tournaye H.
Cumulative delivery rates in different age groups after artificial
insemination with donor sperm. Hum Reprod 2009;24:1891 – 1899.
De Mouzon J, Goossens V, Bhattacharya S, Castilla JA, Ferraretti AP,
Korsak V, Kupka M, Nygren KG, Nyboe Andersen A, European IVFmonitoring (EIM) Consortium, for the European Society of Human
Reproduction and Embryology (ESHRE). Assisted reproductive
technology in Europe, 2006: results generated from European registers
by ESHRE. Hum Reprod 2010;25:1851– 1862.
Doody MC. Drop-out behaviour and fertility table analysis of pregnancy
rates. Hum Reprod 1993;8:886 –889.
Engmann L, Maconochie N, Bekir JS, Jacobs HS, Tan SL. Cumulative
probability of clinical pregnancy and live birth after a multiple cycle IVF
package: a more realistic assessment of overall and age-specific success
rates? Br J Obstet Gynaecol 1999;106:165 – 170.
Garrido N, Bellver J, Remohi J, Simon C, Pellicer A. Cumulative live-birth rates
per total number of embryos needed to reach newborn in consecutive in
vitro fertilization (IVF) cycles: a new approach to measuring the likelihood
of IVF success. Fertil Steril 2011;96:40 – 46.
Germond M, Urner F, Chanson A, Primi MP, Wirthner D, Senn A. What is
the most relevant standard of success in assisted reproduction?: The
cumulated singleton/twin delivery rates per oocyte pick-up: the
CUSIDERA and CUTWIDERA. Hum Reprod 2004;19:2442 – 2444.
Gnoth C, Maxrath B, Skonieczny T, Friol K, Godehardt E, Tigges J. Final ART
success rates: a 10 years survey. Hum Reprod 2011;26:2239– 2246.
Haan G, Bernardus RE, Hollanders HM, Leerentveld BA, Prak FM,
Naaktgeboren N. Selective drop-out in successive in-vitro fertilization
attempts: the pendulum danger. Hum Reprod 1991;6:939 – 943.
Hull MG. Effectiveness of infertility treatments: choice and comparative
analysis. Int J Gynaecol Obstet 1994;47:99 – 108.
Hull MG, Eddowes HA, Fahy U, Abuzeid MI, Mills MS, Cahill DJ, Fleming CF,
Wardle PG, Ford WC, McDermott A. Expectations of assisted conception
for infertility. BMJ 1992;304:1465– 1469.
Ke H, Chen X, Liu YD, Ye DS, He YX, Chen SL. Cumulative live birth rate
after three ovarian stimulation IVF cycles for poor ovarian responders
2707
according to the bologna criteria. J Huazhong Univ Sci Technolog Med Sci
2013;33:418 – 422.
Li HW, Lee VC, Lau EY, Yeung WS, Ho PC, Ng EH. Role of baseline antral
follicle count and anti-Mullerian hormone in prediction of cumulative live
birth in the first in vitro fertilisation cycle: a retrospective cohort analysis.
PLoS One 2013;8:e61095.
Li HW, Lee VC, Lau EY, Yeung WS, Ho PC, Ng EH. Cumulative live-birth rate
in women with polycystic ovary syndrome or isolated polycystic ovaries
undergoing in-vitro fertilisation treatment. J Assist Reprod Genet 2014;
31:205– 211.
Lintsen AM, Braat DD, Habbema JD, Kremer JA, Eijkemans MJ.
Can differences in IVF success rates between centres be explained
by patient characteristics and sample size? Hum Reprod 2010;
25:110– 117.
Luke B, Brown MB, Wantman E, Lederman A, Gibbons W, Schattman GL,
Lobo RA, Leach RE, Stern JE. Cumulative birth rates with linked
assisted reproductive technology cycles. N Engl J Med 2012;366:
2483 – 2491.
Malizia BA, Hacker MR, Penzias AS. Cumulative live-birth rates after in vitro
fertilization. N Engl J Med 2009;360:236 – 243.
Malizia BA, Dodge LE, Penzias AS, Hacker MR. The cumulative probability of
liveborn multiples after in vitro fertilization: a cohort study of more than
10,000 women. Fertil Steril 2013;99:393 – 399.
Missmer SA, Pearson KR, Ryan LM, Meeker JD, Cramer DW, Hauser R.
Analysis of multiple-cycle data from couples undergoing in vitro
fertilization: methodologic issues and statistical approaches. Epidemiology
2011;22:497 – 504.
Olivius K, Friden B, Lundin K, Bergh C. Cumulative probability of live birth
after three in vitro fertilization/intracytoplasmic sperm injection cycles.
Fertil Steril 2002;77:505 –510.
Roque M, Lattes K, Serra S, Solà I, Geber S, Carreras R, Checa MA. Fresh
embryo transfer versus frozen embryo transfer in in vitro fertilization
cycles: a systematic review and meta-analysis. Fertil Steril 2013;
99:156– 162.
Stern JE, Brown MB, Wantman E, Kalra SK, Luke B. Live birth rates and birth
outcomes by diagnosis using linked cycles from the SART CORS database.
J Assist Reprod Genet 2013;30:1445– 1450.
Stolwijk AM, Hamilton CJ, Hollanders JM, Bastiaans LA, Zielhuis GA. A more
realistic approach to the cumulative pregnancy rate after in-vitro
fertilization. Hum Reprod 1996;11:660– 663.
Thurin-Kjellberg A, Olivius C, Bergh C. Cumulative live-birth rates in a trial
of single-embryo or double-embryo transfer. N Engl J Med 2009;
361:1812– 1813.
Tiitinen A, Hydén-Granskog C, Gissler M. What is the most relevant standard
of success in assisted reproduction?: The value of cryopreservation on
cumulative pregnancy rates per single oocyte retrieval should not be
forgotten. Hum Reprod 2004;19:2439 – 2441.
Walters DE. A plea for a more fastidious and objective use of statistics. J Assist
Reprod Genet 1994;11:63– 73.