Abstract
Assisted reproductive technology (ART) treatments, such as in vitro fertilization, have benefited many couples worldwide. However, ART as currently practiced, with the usual transfer of at least two embryos, poses sizable risks for multiple-gestation pregnancy and accompanying serious maternal and child health sequelae. While limiting the number of embryos transferred would appear a straightforward solution, treatments with single-embryo transfer are rare. Likely reasons for this include patient and provider desires to maximize the chance for success in a single treatment given the high cost and lack of insurance coverage, the (sometimes mistaken) belief that transferring more embryos will increase the chance for pregnancy, and the preference for multiple births among some women undergoing ART. Although recent population-based data in the USA and Europe suggest the number of embryos transferred is decreasing, the ART multiple-birth rate remains high. Comprehensive patient-education efforts and continued research on the efficacy of single-embryo transfer are needed.
Since the birth of the first infant resulting from in vitro fertilization in 1978, the field of assisted reproductive technology (ART) has advanced rapidly in both scope and use [1]. ART refers collectively to infertility treatments in which eggs and sperm are handled outside the body. A typical ART treatment includes multiple steps occurring over several weeks:
Suppression of a woman's natural follicle-stimulating hormone production with a gonadotropin releasing hormone agonist or antagonist medication
Administration of exogenous follicle-stimulating hormone to stimulate ovulation
Surgical retrieval of eggs (commonly ten or more) from the woman's ovaries
In vitro fertilization and embryo culture
Transfer of one or more embryos into the uterus
ART has grown to include an increasingly diverse set of treatment options over the past 25 years, which are now used in various combinations. These include: use of eggs retrieved from another woman serving as an egg donor (donor embryo transfer); cryopreservatation or freezing of fertilized embryos for transfer at a later date (frozen or thawed embryo transfer); fertilization using intracytoplasmic sperm injection, in which a single sperm is directly injected into the egg; transfer of unfertilized gametes or newly fertilized zygotes into the fallopian tubes rather than uterus (gamete or zygote intrafollopian transfer); culture of embryos for extended periods to the blastocyst stage; and removal of a single cell from an early embryo for preimplantaion genetic diagnosis prior to transfer.
In the USA in 2003, more than 122,000 ART procedures were performed in over 400 clinics, and over 48,000 infants were born as a result of these procedures. ART-conceived infants now represent slightly more than 1% of total annual births in the USA [2]. While the absolute numbers of ART treatments and resulting infants are greater for the USA than other countries, data from several European registers indicate ART-conceived infants represent 1.3–4.2% of births in those countries [3]. Reviews of published ART and perinatal statistics for Australia suggest that ART-conceived infants represent 2–3% of the 2003 Australian birth cohort [4,5]. In addition, data from the USA, Europe and Australia/New Zealand all indicate continued steady annual increases in ART infants [2–4]. The latest world collaborative report (based on treatments in 2000) suggests that a sizable number of ART procedures are also now performed in many Asian, Latin-American and Middle-Eastern countries [6].
Despite the great benefit ART has provided to many couples worldwide with infertility disorders, this technology has come with certain risks. Most notably, ART as currently practiced, with the transfer of multiple embryos, poses a substantial risk for a multiple-gestation pregnancy. In the USA, over 90% of ART treatments involve the transfer of two or more embryos [7]. Thus, the risk for multiple-gestation pregnancy and birth is substantial. In 2003, 34% of births conceived with ART were twins or more [2]. This has a notable impact on the total US multiple-birth rate; 17.9% of all liveborn multiple-birth infants in the USA in 2003 are estimated to have been conceived after ART treatment.
In European countries, 61–93% of ART treatments in 2002 included multiple-embryo transfer and across countries approximately 25% of births after ART were twins or more [3]. In Australia and New Zealand 72% of ART treatments in 2003 included multiple-embryo transfer and 18% of births were twins or more [4].
The multiple-birth risks associated with ART in various countries are all substantially higher than that expected from natural conception, (<2%) [8,9].
This report will summarize the scientific data on the variation and risk factors for multiple-gestation pregnancy and birth among ART conceptions and the adverse sequalae associated with multiple-gestation pregnancy, overall and for ART births in particular. In this context, data on the efficacy of limiting embryo transfer to prevent multiple-gestation pregnancy in favor of singleton pregnancies, and the current use of and barriers to single-embryo transfer on a population basis are presented and discussed.
Multiple-gestation pregnancies & ART: risk & predictive factors
Several factors have been consistently associated with multiple-gestation risk among ART pregnancies. Clearly, the number of embryos transferred is important. While it should be noted that the relative risk for monozygotic multiple gestations (those due to a single embryo splitting) is also increased with ART, the vast majority of ART multiple gestations occur in treatments in which multiple embryos were transferred. Moreover, the risk of monozygotic twinning after ART, while increased above expected, is still estimated at 1–5% [10–12], a small fraction of the total ART multiple-birth risk consistently reported. Thus, the major portion of the ART multiple-birth risk appears to be dizygotic (i.e., separate oocytes fertilized).
Among ART treatments, the multiple-birth risk varies according to several factors in addition to the number of embryos transferred. In general, the risk factors for multiple birth are also those associated with increased chance for live-birth delivery. These are the transfer of embryos deemed of high quality (based on various measures), the transfer of freshly fertilized (rather than thawed) embryos, young maternal age if the intended mother uses her own eggs, or use of donor eggs [13–17]. However, population-based studies also demonstrate that total live-birth and multiple live-birth rates are not entirely parallel. For example, the most recent analysis of US data indicates that among women aged 40 years or under, there is little difference in live-birth rates among women with two versus three or more embryos transferred; however, the multiple-birth risk, and triplet risk in particular, increases with increasing embryos transferred (e.g., in women <35 years of age in 2003 multiple-birth risks associated with two, three and four embryos transferred were 36.1, 42.3 and 46.7%, respectively, while the highest live-birth rate [46.8%] was observed with two embryos transferred) [2]. Furthermore, while there is much variation in risk by several maternal and embryo factors, in all subgroups the transfer of more than one embryo is associated with a sizable multiple-birth risk compared with natural conceptions. For example, the multiple-birth risk among women using their own eggs varies considerably by maternal age, but remains substantial even among the oldest mothers; in 2003, multiple-birth risks associated with the transfer of just two embryos were:
36.1% for women aged under 35 years
27.5% for women aged 35–37 years
17.6% for women aged 38–40 years
14.9% for women aged 41–42 years
7.3% for women aged over 42 years [2]
Adverse health consequences associated with multiple gestation & birth
Multiple-gestation pregnancy and birth presents risks for both women and infants. Women who carry multiple gestations are at elevated risk for having a serious maternal complication. The risk for pregnancy-induced hypertension or pre-eclampsia has been reported to be two- to three-times higher in multiple- than singleton-gestation pregnancies [9,18,19]. Moreover, among women who develop pre-eclampsia, multiple gestation is associated with earlier onset and increased severity [18,19]. Multiple-gestation pregnancies have also been associated with a twofold increase in risk for both physiologic and iron-deficiency anemia [9,19], a three- to fourfold increase in risk for hemorrhage [18], and a threefold increase in cesarean delivery compared with singleton gestations [18]. A study of maternal deaths in European countries indicated a significant increase in maternal mortality among women pregnant with multiple gestations compared with a singleton gestation (14.9 vs 5.2 deaths/100,000 pregnancies) [18].
The increases in maternal mortality and morbidity are particularly concerning with multiple-gestation pregnancies resulting from ART because many women using ART are already at increased risk for complications due to their advanced maternal age. In the USA, 89% of women using ART with their own eggs are aged 30 years or older and 20% are aged 40 years or over [7]. Over 90% of women using ART with donor eggs are in their late thirties or older. Indeed, several studies suggest that ART twin pregnancies are at increased risk compared with non-ART twin pregnancies for complications including pre-eclampsia or pregnancy-induced hypertension [20,21], antepartum hemorrhage [22], placenta previa [22] and cesarean delivery [21–24]. In addition, among women pregnant with twins, conception with ART has been associated with increased risk for antepartum hospitalization [20] and prolonged postpartum hospital stay [24].
Risk for adverse perinatal outcomes among infants born in a multiple-birth delivery is well documented. Risks for low birthweight (LBW) and preterm delivery both exceed 50% for twins, and the risk for very LBW is 10% [25,26]. These risks are 6–9-times greater than the risks for singletons. Moreover, twins are at greatly increased risk for perinatal and infant mortality [8,9,25] and serious neurological impairments, including cerebral palsy [9,27]. For infants in twin pregnancies reaching 20 weeks gestation, the combined risk for either fetal death, infant death, or cerebral palsy for those surviving the first year is estimated at 7–10%, five-times higher than the combined risk for singletons [28,29]. Risks for adverse outcome among triplet and higher-order multiple births are, as expected, even greater than that observed for twins. Among triplets, risks for preterm delivery and LBW exceed 90% and risks for very preterm delivery and very LBW exceed 30% [25,26]. Moreover, the risk for cerebral palsy exceeds 2% and the combined risk of fetal death, infant death or cerebral palsy among triplet pregnancies that reach 20 weeks gestation is estimated at 19% [28].
Studies comparing infant outcomes among ART twins with outcomes among non-ART twins provide mixed results. While some suggest increased risks for LBW and preterm among ART versus non-ART twins [21,23,30], others suggest comparable risks [31,32]. Studies of perinatal mortality are also incongruent; studies suggesting increased [21,23,30], comparable [33,34] and reduced mortality risks [32,35] for ART twins have been reported.
The inconsistency in study results for ART twin pregnancies stands in contrast to the numerous studies of singletons reporting moderate to strong associations between ART and preterm delivery and LBW, independent of potential confounders such as maternal age and parity [31,32,36,37]. That these associations are not clearly demonstrated among twins is not surprising. Given the great perinatal risk already posed by multiple-gestation pregnancy, it is hard to imagine much additional effect for any independent factor, including mode of conception. Moreover, approximately a third of naturally conceived twins will be monozygotic [38], which are generally at greater risk for adverse outcome than dizygotic twins; as mentioned, the vast majority of ART twins are likely to be dizygotic. While some studies restricted comparisons to known dizygotic twin pregnancies, those with unlike-sex infants, this is a very select subset of the total dizygotic twin births and results might not be entirely generalizable. The possibility of enhanced monitoring and care among women pregnant with twins via ART must also be considered a possible mitigating factor. Even studies that carefully considered sociodemographic differences between ART and non-ART twins could not account for many possible unmeasured differences in provider care among women pregnant with a high-risk twin pregnancy that resulted from ART. Finally, even in those studies suggesting a protective mortality effect for twins conceived via ART, it must be noted that the preterm, LBW and mortality rates reported for ART twins from these studies were still much higher than those among singleton births. Thus, twin pregnancy must be considered high risk regardless of mode of conception.
Reasons for multiple-embryo transfer
Despite the well-established link between multiple-gestation pregnancy and birth, and adverse outcomes for both women and their children, multiple-embryo transfer remains the norm; thus, single-embryo transfer appears to be an unacceptable or at least less acceptable treatment option to ART patients or their providers. This might be a consequence of several factors. ART procedures are quite costly and many couples in the USA must pay for the entire procedure or a sizeable portion out of pocket due to lack of or limited insurance coverage. Currently, only 15 states have laws requiring some level of infertility coverage; moreover, even in those states, ART is not always included in the legislation; even when it is included various exceptions and limitations often apply [101]. Thus, couples might feel pressure to maximize their chance for success in a single treatment. Indeed, despite the many advances over the past 25 years, embryo implantation remains the rate-limiting step of ART treatment. While 87 and 82% of ART treatments initiated progress to the oocyte-retrieval and embryo-transfer stages, respectively, on average, only 34% progress to implantation and clinically-recognized pregnancy [7]. Thus, many patients and providers might believe the potential risk for multiple-embryo implantation is balanced by the potential increased chance for pregnancy. A recent cost analysis of ART treatment in the USA highlights this dilemma. From a societal perspective, the most cost-effective strategy of achieving a birth through ART is for a woman to undergo multiple treatments, all limited to single-embryo transfer; this strategy would minimize the substantial costs associated with having an infant with adverse outcomes. However, the component of the specific cost borne by the ART patient was estimated to be highest with the single-embryo transfer strategy as more treatments might be needed to achieve success [39].
For some couples, multiple birth might be the goal. In recent surveys of fertility clinic patients in Canada [40] and the USA [41] a sizable minority of female patients (39% in the Canadian study and 20% in the US study) indicated that multiple birth was their preferred treatment outcome. The Canadian study indicated comparable findings for the male partners as well. The US study reported that while proportions varied by parity, income, maternal age, duration of infertility, and knowledge of adverse outcomes for twin gestations, in all subgroups over 10% of women desired a multiple gestation. Thus, the desire for twins by a subset of women might not be fully explained by cost issues.
Encouraging trends
Although ART remains a major risk factor for multiple birth, recent trends are encouraging. In the USA there has been a sizeable downward shift in the number of embryos transferred during ART from 1996 to 2002 [42]. While this trend indicates great progress, the data also indicate that much work remains. The most dramatic trend was a shift from three or more embryos transferred to two embryos transferred. Single-embryo transfer remained uncommon in 2002 and elective single-embryo transfer (i.e., a known choice to transfer one embryo rather than transfer of one embryo because only one was available after the fertilization stage) was 1% or less for women of all ages. In addition, even with the sizeable shift to two embryos, transfer of three or more embryos remained above 50% for all age groups.
Secular trends of a reduced number of embryos transferred are more pronounced in several other countries. By 2003, single-embryo transfer exceeded 25% in Australia and New Zealand, Finland, Slovenia and Sweden [3,4]. In addition, recent legislation in Belgium that allows for reimbursement of up to six ART treatments in women aged under 42 years if strict restrictions on embryo transfer are met, resulted in a dramatic increase in single-embryo transfer (27.2% prelegislation vs 64.8% postlegislation), with a concurrent decline in multiple-gestation pregnancy (25.9 vs 8.0%) [43].
A number of randomized trials have now demonstrated that elective single-embryo transfer is efficacious in patients identified as having a good prognosis (based on characteristics such as young maternal age and embryos assessed as ‘good quality’ prior to transfer) [44,45]. Although in several trials, the total pregnancy rate was somewhat higher in treatments with elective double-embryo transfer than elective single-embryo transfer, the singleton pregnancy rates calculated from data aggregated from four trials were comparable for single- (30.0%) and double-embryo transfer (31.5%) [45,46]. Thus, the increased pregnancy rates achieved with double-embryo transfer were almost entirely a result of much riskier multiple gestations. A few studies are particularly noteworthy. Thurin and colleagues reported comparable birth rates among women randomized to double-embryo transfer (42.9%) and those randomized to single, fresh embryo transfer followed in unsuccessful cases by transfer of a single thawed embryo (38.8%) [47]. In addition, two recent, randomized studies suggest single-embryo transfer might be particularly efficacious if embryos are cultured for 5 days to the blastocyst stage [48,49]. A limitation of all randomized studies to date is restriction to a select population of young women with a good prognosis. Recent observational studies suggest that for some women aged over 35 years (those with embryos indicated to be of high quality) single-embryo transfer is also associated with high pregnancy and live-birth rates, and, thus, might be a viable treatment option [15,50]. Further study of this group of women is needed, as is study on optimal treatment options for reducing multiple-gestation pregnancy among the many women seeking ART who fall into more ‘difficult-to-treat’ prognostic groups.
Future perspective
There is clear empiric evidence that:
ART, as currently practiced, is strongly associated with multiple-gestation pregnancy and birth
Multiple-gestation pregnancy places both women and infants at risk for numerous and serious adverse health outcomes
Single-embryo transfer would eliminate all but a small percentage of ART-associated multiple-gestation pregnancies
Despite the appearance of a simple solution, addressing the problem will be complex. Greater progress at moving toward single-embryo transfer has occurred in many European countries than in the USA. However, ART is subsidized to a much greater extent in Europe than in the USA. Indeed, the huge success in reducing multiple births in Belgium stems from a policy that links embryo transfer restrictions with healthcare reimbursement [43].
Reducing multiple births in the USA will require a multifaceted approach. Focus on singleton live-birth as the optimal outcome and preferred measure of ART success has been an important first step [51–53]. The singleton live-birth rate is now a key indicator presented annually in US ART reports [2,7]. In countries such as Belgium, or for individual practices in the USA and elsewhere that in the future adopt a global treatment strategy in which all or certain patients groups are slated to undergo multiple treatment cycles limited to single-embryo transfer (both with freshly fertilized and frozen-thawed embryos), it will also be important to present success measures as cumulative per patient live-birth rate, in which multiple treatments for the same patient are linked.
Currently, more work is needed in developing effective methods of educating prospective ART patients regarding the high risks associated with multiple-gestation pregnancy. More widespread public communication regarding the known efficacy of single-embryo transfer for certain subgroups of patients is also needed. Although it is rightly argued that current data suggesting single-embryo transfer is a viable treatment option apply only to a subgroup of ‘good-prognosis’ patients, even in these patients, single-embryo transfer remains rare. For more difficult-to-treat patients, the situation is more complex as transfer of two or more embryos might very well afford some advantage in terms of achieving a live-birth in a single ART treatment. It is important to note that it is not possible to be certain that multiple-embryo transfer offers an advantage because so few women falling into the more difficult-to-treat categories (such as those of advanced maternal age) currently choose elective single-embryo transfer. Therefore, the comparison of elective single- and double-embryo transfer, even in large population-based surveillance datasets, is severely limited. Nonetheless, even if an advantage is afforded with multiple-embryo transfer, a substantial multiple-birth risk is observed in women in all age groups who undergo ART with two or more embryos transferred. Thus, an informed and comprehensive assessment of the myriad financial and other barriers that prevent women in various prognostic groups from considering treatment plans incorporating multiple treatment cycles with single-embryo transfer (with both fresh and thawed embryos), is needed.
Public-health professionals, ART providers and scientists working in this field need to continue their focus on strategies to reduce (and hopefully eventually eliminate) ART-associated multiple births, which would reduce many adverse maternal and child health consequences currently associated with ART.
Disclaimer
The findings and conclusions in this article are those of the author and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Executive summary
Recent estimates of the multiple-birth risk among pregnancies conceived using assisted reproductive technology (ART) are 34% in the USA, 25% in Europe and 18% in Australia and New Zealand. These estimates are substantially higher than the multiple-birth rate expected from natural conception (<2%).
As the use of ART continues to increase worldwide, its impact on multiple births in populations also continues to increase.
Although ART poses a risk for both monzygotic and dizygotic multiple birth, dizygotic multiple birth due to the common practice of multiple-embryo transfer is responsible for the preponderance of current ART-associated multiple-birth risk.
Women who carry multiple gestations are at elevated risk for serious maternal complications and infants born in a multiple-birth delivery are at substantial risk for numerous adverse perinatal and child outcomes.
Despite the clear link between multiple-gestation pregnancy and adverse outcomes for women and children, multiple-embryo transfer remains the norm in ART procedures.
The reasons may include the desire to maximize the chance for pregnancy in a single treatment given the high costs and lack of insurance coverage, the belief that transferring more embryos is efficacious, and the preference for multiple births among some women undergoing ART.
Randomized trials have demonstrated that single-embryo transfer is an efficacious treatment in patients identified as having factors related to a good prognosis.
Comprehensive patient education efforts and continued research on the efficacy of single-embryo transfer, especially among those patients considered difficult to treat, are needed.