Background
The last decade has witnessed a number of new treatments integrated into routine assisted reproductive technology (ART) practice, at times making public outcome reporting more challenging and less transparent [
1,
2]. A number of new IVF practice regimens have been applied differently over the world and beg for further critical evaluation including use of routine blastocyst-stage in place of cleavage-stage embryo transfer [
3] replacement of fresh embryo transfer by embryo cryopreservation (“freezing”) and subsequent thawed embryo transfer [
4], preimplantation genetic screening (PGS), [
5] single embryo transfer (SET) in place of double embryo transfer (DET) [
6] and minimal stimulation protocols [
7,
8]. At times spurred by government policies or local recommendations of professional societies, adoption of some practice changes occurred faster in some parts of the world than others. The most recent worldwide report on ART practices was authored by the International Committee for Monitoring Assisted Reproductive Technology (ICMART), and was based on data from 2008 through 2010 [
9].
Here we examined longitudinal data, reported by ART centers between 2004 and 2013 to regional registries worldwide. Our objectives were to study longitudinal changes in ART practice over the last decade of reported data, including cycle volumes, use of previously cryopreserved embryos and of SET. In addition, we assessed ART outcomes, based on live birth rates in fresh and frozen-thawed cycles. By comparing practice patterns and outcomes across regions, as well as longitudinal changes within regions, this review offers insights into the global evolution of ART practice.
Discussion
This review evaluated longitudinal data to gain insight into ART practices and outcomes, and demonstrated considerable disparities in live birth rates between regions in the world. While over the last 10 years most regions demonstrated improvement, some showed no change and others demonstrated declines in ART live birth rates.
These findings, of course, reflect major differences in patient demographics, societal norms, local laws, and economics, which impact ART practices and could not be adjusted for in this analysis. Direct comparisons of data from various regions are also subject to differences in rigor of data reporting, collection and verification outlined above. Interpretations of here presented data should, therefore, be made with caution, and in consideration of these issues. The study is also somewhat incomplete in that it does not include the Middle East, large parts of Asia, and all of Africa. One Middle Eastern country, Israel, on a per capita basis performs the highest number of ART cycles of any country in the world [
9]. We were unable to draw on appropriately published ART data in any of these regions.
The U.S. maintained throughout the decade the highest fresh cycle live birth rates compared to other regions. Continental Europe, however, for the first time appears to have caught up to the US in its fresh cycle live birth rates by 2010. The longstanding difference between US and European outcomes has been subject of scrutiny by European and US investigators [
27‐
29]. A number of reasons for the differences have been suggested, and investigators have in the past expressed strong opinions on this subject. Confluence of U.S. and Continental European fresh cycle live birth rates now suggest increasing congruity of ART practice between the two continents.
Newly integrated practices into ART to a large degree involve methods of embryo selection, including blastocyst-stage transfer, PGS, cryopreservation of all embryos and subsequent frozen-thawed transfer, and the utilization of SET. These new practices, however, may not benefit all patient groups equally and in some, particularly poor prognosis patients, may, negatively affect outcomes [
30].
Live birth rates with frozen-thawed embryos improved rapidly in the U.S. (Fig.
3b). Reported U.S. live birth rates may be somewhat inflated due to exclusion from national outcome reporting of embryo banking cycles, some PGS cycles, and cycles where no embryos are created for transfer or survive thaw to be transferred [
2,
26].
In some regions, live birth rates with frozen-thawed embryos exceed those with fresh embryos. This might be interpreted as evidence in support of routine embryo banking in lieu of fresh transfer, [
4,
31] especially since efficiency of cryopreservation has improved with vitrification. However, it more likely reflects differences in patient populations undergoing fresh versus frozen-thawed embryo transfer. For example, younger and favorable prognosis patients maybe more likely to have surplus embryos for cryopreservation or banking and therefore may be relatively overrepresented in frozen-thawed versus fresh cycles. Similarly, we have recently demonstrated how patient selection biases in US national data led CDC investigators to incorrect conclusions about PGS effectiveness [
32].
In 2004 Japan already demonstrated the lowest fresh cycle live birth rates among all regions. Over the following decade, Japan almost tripled its number of ART cycles (Fig.
1a) and lost almost two-thirds of its fresh cycle live births, dropping to a national rate of 5% by 2012 (Fig.
3a). The loss in fresh cycle live births was partially compensated for by a relatively higher live birth rate in frozen-thawed embryo transfer cycles, resulting in a relatively stable annual total live birth rate (including fresh and frozen-thawed cycles) as recently reported by Takeshima et al [
21]. Indeed, most ART infants in Japan in recent years were conceived in frozen-thawed rather than fresh ART cycles. Our findings along with those reported by Takeshima et al. suggest that a large increase in ART cycle volume was required to achieve a very modest improvement in perinatal outcomes and offset declining fresh cycle live birth rates.
No other region in the world demonstrated such rapid change in ART practice. As such, it appears that changes in national practice patterns in Japan, not equally experienced in other regions in the world, led to the observed outcome changes. For example, Teramoto and Kato proposed minimal stimulation IVF and by 2007 reported having performed 43,433 such cycles [
7]. Similar protocols have since been adopted by many Japanese ART centers, and sporadically elsewhere [
33‐
35]. The dramatic changes in Japanese ART cycle numbers and outcomes, likely represent a combination of minimal stimulation protocols, the potentially lower implantation rate in fresh versus frozen-thawed transfers due to endometrial factors, the fresh transfer of poor quality embryos that may not go on to blastocyst-stage, progressive migration to thawed embryo transfer, as well as, implementation of stringent SET regulations [
21].
Embryo selection efforts have also gained followers elsewhere: For example, SET at blastocyst- stage has also become a characteristic feature of ART in Australia/New Zealand, [
11,
36] areas of Canada [
22] and in Northern Europe [
37,
38]. Australia/New Zealand reported already in 2004 relatively low live birth rates in fresh ART cycles which gradually further declined as utilization of SET at blastocyst-stage increased (Fig.
3a). Australia and New Zealand progressively shifted from cleavage- to blastocyst- stage embryo transfers, with blastocyst-stage transfers increasing from 49.8% in 2009 to 61.1% in 2013. Concomitantly, fresh cycles that reached embryo transfer in Australia and New Zealand decreased from 76.6 to 67.5% [
11]. Likewise, data from Canada demonstrate a decline in fresh cycle live birth rates following implementation of an SET mandate in the province of Quebec in 2009 [
23]. Increased utilization of SET, therefore, appears to be temporally associated with declines in fresh cycle live birth rates. Regions with strict SET policies have been able to lower their multiple delivery rates, for example by 2012–2013 Australia/New Zealand and Japan reported the lowest multiple delivery rates of 5.6% and 4% while the US had the highest of 27%.
Utilization of elective SET is based on the premise that twin pregnancies increase maternal and neonatal outcome risks in comparison to singletons [
39]. Our group considers that a more appropriate way of framing the question in infertility patients is to compare the outcome risks of two consecutive singleton pregnancies to one twin pregnancy [
40,
41]. Moreover, it is important to recognize that the risk profiles of singleton and twin neonates vary greatly depending on whether conception occurred spontaneously or via ART [
42,
43]. Nevertheless, to minimize risks of a twin pregnancies, elective SET followed by a frozen-thawed embryo transfer has in recent years been increasingly offered to good prognosis patients, as this approach is reported to produce a similar cumulative pregnancy chance to DET [
44]. It is also important to note that we did not differentiate between elective and non-elective SET in this review because this classification is typically not discernable in registry reports. Patients undergoing non-elective SET are typically those with poor prognosis or patients undergoing minimal stimulation IVF, both groups likely experience somewhat lower live birth rates than good prognosis patients who are the best candidates for elective SET [
34,
44‐
46].
Pregnancy and live birth rates have been the traditional metrics of ART success. More recently, investigators proposed that measures of neonatal health should also be incorporated [
47,
48]. As revealed in this study, there are inconsistencies in the way perinatal outcome data are presented by various registries, making comparison between regions difficult. Further research appears indicated to develop better perinatal outcome associations with different ART practice patterns. Moreover, while most current ART registries present data per cycle of treatment, efforts are now underway in Australia/New Zealand and in the U.S. to link successive treatment cycles undertaken by each female patient. Therefore, ART registry reports may be evolving to define ART success as the birth of term, normal birthweight neonates per number of treated patients during the calendar year.
Acknowledgements
We thank Ms. Yu Kizawa for kindly assisting with Japanese translation.