Live-Birth Rates and Multiple-Birth Risk Using in Vitro Fertilization

ORIGINAL CONTRIBUTION

Live-Birth Rates and Multiple-Birth Risk Using In Vitro Fertilization

Laura A. Schieve, PhD Context To maximize birth rates, physicians who perform in vitro fertilization (IVF) Herbert B. Peterson, MD often transfer multiple embryos, but this increases the multiple-birth risk. Live-birth and multiple-birth rates may vary by patient age and embryo quality. One marker for Susan F. Meikle, MD embryo quality is cryopreservation of extra embryos (if embryos are set aside for cryo- Gary Jeng, PhD preservation, higher quality embryos may have been available for transfer). Isabella Danel, MD Objective To examine associations between the number of embryos transferred during IVF and live-birth and multiple-birth rates stratified by maternal age and whether Nancy M. Burnett, BS extra embryos were available (ie, extra embryos cryopreserved). Lynne S. Wilcox, MD Design and Setting Retrospective cohort of 300 US clinics reporting IVF transfer procedures to the Centers for Disease Control and Prevention in 1996. INCE THE GOAL OF IN VITRO FER- tilization (IVF) is pregnancy and, Subjects A total of 35 554 IVF transfer procedures. ultimately, live birth, clinical de- Main Outcome Measures Live-birth and multiple-birth rates (percentage of live cision making about IVF prac- births that were multiple). ticesS is heavily focused on maximizing Results A total number of 9873 live births were reported (multiple births from 1 preg- a woman’s chances of becoming pregnancy were counted as 1 live birth). The number of embryos needed to achieve maximum nant. One common practice that aims live-birth rates varied by age and whether extra embryos were cryopreserved. Among wom- to increase the likelihood of pregnancy en 20 to 29 years and 30 to 34 years of age, maximum live-birth rates (43% and 36%, is to transfer multiple embryos (often respectively) were achieved when 2 embryos were transferred and extra embryos were more than 3) into the uterine cavity. This cryopreserved. Among women 35 years of age and older, live-birth rates were lower overall and regardless of whether embryos were cryopreserved, live-birth rates increased if more treatment approach also presents an im- than 2 embryos were transferred. Multiple-birth rates varied by age and the number of portant drawback, however, because it embryos transferred, but not by whether embryos were cryopreserved. With 2 embryos increases the risk for multiple birth. Mul- transferred, multiple-birth rates were 22.7%, 19.7%, 11.6%, and 10.8% for women aged tiple-birth infants are at significant risk 20 to 29, 30 to 34, 35 to 39, and 40 to 44 years, respectively. Multiple-birth rates increased for a number of adverse outcomes in- as high as 45.7% for women aged 20 to 29 years and 39.8% for women aged 30 to 34 cluding preterm delivery, low birth years if 3 embryos were transferred. Among women aged 35 to 39 years, the multiple- weight, congenital malformations, fe- birth rate was 29.4% if 3 embryos were transferred. Among women 40 to 44 years of age, tal and infant death, and long-term mor- the multiple-birth rate was less than 25% even if 5 embryos were transferred. bidity and disability among survi- Conclusions Based on these data, the risk of multiple births from IVF varies by ma- vors.1-5 Twins are 5 times as likely, and ternal age and number of embryos transferred. Embryo quality was not related to mul- triplet and higher-order infants 13 times tiple birth risk but was associated with increased live-birth rates when fewer embryos were transferred. as likely, as singleton infants to die during the first year of life.2 JAMA. 1999;282:1832-1838 www.jama.com To curtail the multiple-birth risk, several countries have passed legislation remained outside the legal arena; how- tween various markers of embryo qual- that limits the number of embryos that ever, the American Society for Repro- ity and implantation, attention has fo- can be transferred to 3.6,7 Such a policy ductive Medicine has issued practice cused on whether such data can be 8 is not universally supported as it runs guidelines. The debate about embryo limits has increasingly focused on Author Affiliations: Division of Reproductive Health, counter to the expectation of au- National Center for Chronic Disease Prevention and tonomy in the patient-physician rela- whether to consider prognostic fac- Health Promotion, Centers for Disease Control and Pre- tors when setting guidelines, particu- vention, Atlanta, Ga. tionship. In the United States, the is- Corresponding Author and Reprints: Laura A. Schieve, sue of embryo transfer has thus far larly patient age, which varies in- PhD, Division of Reproductive Health, National Cen- versely with a woman’s chances for ter for Chronic Disease Prevention and Health Pro- achieving pregnancy.9,10 Additionally, motion, Centers for Disease Control and Prevention, See also p 1813 and Patient Page. Mailstop K-34, 4770 Buford Hwy NE, Atlanta, GA as studies demonstrate associations be- 30341 (e-mail: [email protected]).

Downloaded From: https://jamanetwork.com/ on 10/02/2021 BIRTH RATES AND MULTIPLE-BIRTH RISK USING IVF

translated into policy in the future. Al- ART data was 1996. In 1996, 300 US transfer, this number actually represents though current grading schemes for centers reported more than 60 000 ART 35 554 IVF transfer procedures. assessing embryo quality have limita- cycles to the CDC. Because some cen- tions, both embryo morphology grade ters did not report their data, despite the Definitions of IVF Outcomes and the ability to select embryos for federal mandate, this number does not We defined pregnancy as the presence transfer have been associated with in- represent every ART cycle performed in of 1 or more gestational sacs observed creased pregnancy and live-birth rates the United States; however, it is esti- on ultrasound (with or without the pres- in previous studies.11-17 An especially mated that data on more than 95% of all ence of a fetal heart). In rare instances provocative study that used population- cycles were reported. (Ͻ1%), the number of fetal sacs ob- based data from the United Kingdom Weselectedfresh,nondonorIVFcycles served on ultrasound was not recorded suggested that elective transfer of 2 for inclusion in the current analysis or was recorded as 0, but a pregnancy rather than 3 embryos reduced the mul- (N = 44 723). This refers to cycles in outcome was recorded (live birth, still- tiple-birth risk without affecting the which eggs were removed from a wom- birth, spontaneous abortion, therapeu- chance of live birth for any age group.11 an’s ovaries, combined with sperm, and tic abortion); these cycles were also To determine if this finding is sup- if fertilized, the resulting embryo(s) was coded as pregnancies. A total of 12 115 ported by the US IVF population, we replaced into the same woman’s uterus. pregnancies were reported. Since ART used a population-based dataset of IVF– This selection excludes cycles in which centers do not routinely treat patients assisted reproductive technology (ART) embryos derived from a woman serving beyond the first trimester, live births and cycles initiated in US clinics in 1996 to as an egg donor were transferred to the fetal losses later than the first trimester examine associations between em- patient (n = 5162); cycles in which em- were based on verbal or written reports bryo number and pregnancy, live- bryos derived from a patient were trans- from either the patient or her obstetric birth, and multiple-birth rates. The large ferred into another woman serving as a health care professional. ART centers of- sample afforded us an opportunity to gestational carrier or surrogate (n = 688); ten actively follow-up patients to ascer- more fully explore these associations by and cycles in which the embryos trans- tain pregnancy outcome. An outcome examining several important factors, in- ferred had been retrieved and fertilized (live birth, stillbirth, spontaneous abor- cluding patient age and availability of at an earlier date, frozen via cryopreser- tion, therapeutic abortion) was re- extra embryos for future ART cycles. vation, and thawed for use in the current corded for all but 457 (4%) of these preg- cycle (n = 9290). It also excludes cycles nancies. A total of 9873 live-birth METHODS in which embryos or oocytes were trans- deliveries were reported. We consid- Subjects ferred into a woman’s fallopian tubes ered each live-birth delivery as a single The Fertility Clinic Success Rate and rather than uterus (n = 4117), cycles in live birth; eg, a live-birth delivery of trip- Certification Act of 1992 requires that which embryos were transferred to both lets was counted as 1 live birth. each medical center performing IVF or the uterus and the fallopian tubes We classified a pregnancy as a mul- related ARTs report data for each ART (n = 619), and cycles in which both fresh tiple gestation if either 2 or more fetal cycle initiated to the Centers for Dis- and thawed embryos were transferred hearts were noted on an early ultra- ease Control and Prevention (CDC) an- (n = 125). Because these cycle types may sound, or 2 or more infants were born. nually for the purpose of reporting clinic- vary with respect to implantation and We defined multiple gestation based on specific pregnancy success rates.18 An pregnancy rates, and also with respect to the more stringent criterion of fetal ART cycle is considered to begin when the importance of various prognostic fac- hearts (rather than number of sacs only) a woman begins taking fertility drugs or tors, they were not combined. Separate because multiple gestations that do not starts ovarian monitoring with the in- analysisforeachcycletypewasprecluded progress to fetal hearts are generally not tent of having embryos transferred. ART by small sample sizes in many key sub- considered to be clinically relevant. We centers submit data obtained from clinic groups. Therefore, this analysis is re- classified a live-birth delivery as a mul- records for each cycle initiated during stricted to the most common type of ART tiple birth if 2 or more fetuses were born a given reporting year (January 1 treatment: fresh, nondonor IVF. and at least 1 of these was liveborn. We through December 31) in a standard- Among fresh, nondonor IVF cycles also separately examined triplet and ized format. The datafile is organized that were initiated in 1996, we excluded higher-order gestations and triplet and with 1 record per cycle. Multiple cycles cycles that did not progress to embryo higher-order births. for a single patient are not linked. Data transfer (n = 8890) and cycles for which collected include patient demograph- patient age was either missing (n = 79) Data Analysis ics, medical history and infertility diag- , younger than 20 years (n = 6), or older We categorized each IVF procedure ac- noses, clinical information pertaining to than 44 years (n = 194). Our final sample cording to 2 factors: (1) the number of the ART cycle, and information on re- included 35 554 fresh, nondonor IVF embryos transferred (1, 2, 3, 4, 5, 6, or sultant pregnancies and births. The first cycles. Because these cycles were lim- Ն7), and (2) patient age at cycle start full year for which the CDC collected ited to those that progressed to embryo (20-29, 30-34, 35-39, or 40-44 years).

Downloaded From: https://jamanetwork.com/ on 10/02/2021 BIRTH RATES AND MULTIPLE-BIRTH RISK USING IVF

Within each age-embryo number stra- cause these rates were not parallel; embryo quality, and clinic success. We tum, we present the percentages of live trends for triplet and higher-order birth evaluated trends in live-birth and mul- births per transfer procedure, mul- rates were often less pronounced than tiple-birth rates after additionally strati- tiple births per live birth, triplet or trends for triplet and higher-order ges- fying on several such factors. These in- higher-order births per live birth, and tation rates. cluded previous pregnancies, previous triplet or higher-order gestations per In addition to embryo number and live births, number of previous ART pregnancy. We present both triplet- patient age, we examined the effects of cycles, primary infertility diagnosis, use birth and triplet-gestation rates be- potential markers of patient prognosis, of intracytoplasmic sperm injection (a technique used often in male-factor in- fertilityinwhichasinglespermisdirectly Table 1. Percent Distribution of In Vitro Fertilization (IVF) Transfer Procedures by Patient Age and Factors Related to Patient Medical History or IVF Procedure* injected into the oocyte), use of assisted Patient Age at IVF Cycle Start, y hatching (use of chemicals, lasers, or me- chanical means to create an opening in 20-29 30-34 35-39 40-44 (n = 4590) (n = 12 774) (n = 13 174) (n = 5016) the zona pellucida), and whether 1 or Previous pregnancies, No. more embryos that were retrieved and 0 64.2 54.3 44.5 38.1 fertilized during the current cycle were 1 19.6 24.4 26.1 25.4 cryopreserved for use in later cycles. The 2 9.0 11.1 15.0 15.3 availability of extra embryos for cryo- Ն3 7.1 10.2 14.4 21.2 preservationindicatesthatmoreembryos Previous live births, No. were available for transfer than were ac- 0 88.0 82.0 74.8 70.3 tually transferred and therefore the em- 1 8.9 13.6 18.1 20.2 bryos transferred were electively chosen; Ն 2 3.0 4.4 7.1 9.5 this variable may be a surrogate for em- Previous IVF cycle, No. 0 70.7 59.4 53.9 48.2 bryo quality. Embryo cryopreservation 1 17.4 21.3 22.3 23.3 has been used since the early 1980s 2 7.4 10.4 11.6 12.7 and is now a standard component of Ն3 4.5 8.9 12.3 15.7 most ART programs. However, because Primary infertility diagnosis whether or not extra embryos are cryo- Endometriosis 12.2 16.3 14.9 12.2 preserved is a function of patient choice Tubal factor 32.3 32.5 34.5 27.1 as well as embryo availability and clini- Male factor 33.8 27.8 25.2 22.4 cal assessment, the cryopreserved vari- Ovulatory dysfunction 11.5 9.8 8.6 11.9 able is a nonspecific marker. Finally, we Uterine factor 0.7 1.2 1.9 3.4 examined all results after stratification Idiopathic infertility 4.3 6.5 7.6 9.9 by clinic-level characteristics. We clas- Other 5.1 5.9 7.2 13.2 sified each clinic as having a pregnancy Embryos transferred, No. rate above or below the mean pregnancy 1 3.4 4.4 6.6 10.3 rate for all clinics combined. We analo- 2 8.9 8.6 10.8 14.6 3 33.2 28.0 18.4 17.1 gously classified clinics as having high 4 33.2 34.3 34.0 20.7 or low age-specific pregnancy rates. We 5 12.0 14.3 17.4 18.3 stratified trends in live-birth rates and 6 6.4 7.4 8.6 11.4 multiple-birthratesaccordingtowhether Ն7 2.9 3.0 4.2 7.5 a cycle was performed in a clinic with Cryopreservation of Ն1 embryos low or high overall and/or age-specific retrieved during cycle pregnancy rates. No 56.4 63.0 74.5 89.3 For all analyses, the statistical signifi- Yes 43.6 37.0 25.5 10.7 cance of differences in rates between cat- Use of intracytoplasmic sperm injection ␹2 No 56.1 61.1 63.0 66.1 egories was assessed with tests. Yes, with some embryos 9.3 8.5 7.6 5.0 This research was approved by the in- Yes, with all embryos 34.6 30.4 29.5 29.0 stitutional review board at the CDC. Use of assisted hatching No 72.9 69.0 56.4 38.4 Yes, with some embryos 6.7 8.1 10.4 12.0 RESULTS Yes, with all embryos 20.4 22.9 33.2 49.6 Patient medical history and IVF pro- *PϽ.01, ␹2 to test for differences in distributions by age. Sample size was reduced for some analyses due to missing cedural factors varied significantly by values; maximum number of missing values was 3.5% for intracytoplasmic sperm injection. patient age (TABLE 1). Both previous

Downloaded From: https://jamanetwork.com/ on 10/02/2021 BIRTH RATES AND MULTIPLE-BIRTH RISK USING IVF

pregnancies and previous live births in- embryo transferred up to 3 when it 29 years and 30 to 34 years, those with creased with each successive increase reached 46%. The multiple-birth rate cryopreserved embryos achieved live- in patient age. Older women were also continued to increase beyond 3 em- birth rates of 43% and 36%, respec- more likely to have undergone IVF pre- bryos, although the magnitude of the tively, when 2 embryos were trans- viously; younger women were more increase was much smaller. Among ferred; these rates were more than likely to be diagnosed as having infer- women aged 30 to 34 years and 35 to double the rates observed among tility related to tubal or male factors. 39 years, increases in the multiple- women in these age groups for whom The distribution of the number of em- birth rate were observed with up to 4 2 embryos were transferred and no embryos transferred also varied by age; embryos transferred (45% and 38% for bryos were cryopreserved. Further, older women were more likely to be in women aged 30-34 years and 35-39 among the cryopreserved group, live- both of the outlying categories (1 or Ն7 years, respectively). Increases in the birth rates were not significantly greater embryos transferred) than women in multiple-birth rate among women aged when 3 or more embryos were trans- the younger age groups. Additionally, 40 to 44 years were less striking and did ferred than when 2 embryos were trans- use of both cryopreservation for extra not reach the high levels noted for other ferred. Among women aged 35 to 39 embryos and intracytoplasmic sperm groups until 7 or more embryos were years, live-birth rates were substan- injection declined with age, while use transferred (39%). tially increased for both 2 and 3 em- of assisted hatching procedures in- The triplet-gestation and triplet- bryo transfers (25% and 33%, respec- creased. birth rates were substantially elevated tively) when additional embryos were Pregnancy rates declined from 41% among women in the 2 youngest age cryopreserved. Among women aged 40 among women aged 20 to 29 years to groups when 3 or more embryos were to 44 years, the cryopreserved group 20% among women aged 40 to 44 years. transferred. Among women 20 to 29 achieved notably higher live-birth rates Live-birth rates declined from 35% to years of age with 3 embryos trans- when 3 embryos were transferred 13% in these same categories. In addi- ferred, the triplet-gestation and triplet- (19%); additionally, the live-birth rate tion to age, live-birth rates varied sig- birth rates both reached 10%. Among continued to increase slightly when nificantly by the number of embryos women 30 to 34 years of age with 3 em- more than 3 embryos were transferred transferred (TABLE 2). Among the 2 bryos transferred, the triplet-gestation (24% with 5 embryos). Among all age youngest age groups, the live-birth rate rate reached 10%; the triplet-birth rate groups, small sample sizes impeded increased considerably with increas- was near 7%. Among women 40 years evaluation of procedures in which 1 em- ing number of embryos transferred up of age or older, the risk for triplets was bryo was transferred and additional em- to 3; beyond 3 embryos, there was no greatly reduced; both triplet-gestation bryos were cryopreserved; among additional increase and, in fact, de- and triplet-birth rates were less than 2% women aged 40 to 44 years, there were clines were noted in a few categories. when 4 embryos were transferred and also too few procedures in which 2 em- Among women aged 35 to 39 years, the less than 5% when 5 embryos were bryos were transferred and additional live-birth rate continued to increase transferred. Triplet rates among women embryos were cryopreserved. Al- with up to 4 embryos transferred and aged 35 to 39 years were intermediate though whether embryos were cryo- then declined slightly. Among women between the younger and older age preserved had a large impact on live- aged 40 to 44 years, the live-birth rate groups. birth rates, trends in multiple-birth rates continued to increase with up to 5 em- The basic patterns in live-birth and did not vary by the cryopreserved vari- bryos transferred. However, even at this multiple-birth rates apparent in Table able. Within each age group, the trends maximum rate, the live-birth rate 2 persisted after further stratification on in multiple birth presented in Table 2 among women in the oldest age group previous pregnancies, previous live were similar to both the trends in mul- was still substantially lower than the births, previous ART cycles, infertility tiple birth among women with 1 or maximum rate observed among women diagnosis, use of intracytoplasmic more embryos cryopreserved and the in the youngest age group (20.3% vs sperm injection, use of assisted hatch- trends among women with no em- 37.7%). ing, and clinic-level pregnancy and age- bryos cryopreserved (data not shown). Similar patterns were observed for specific pregnancy rates (data not multiple births, triplet and higher- shown). For all age groups, however, COMMENT order gestations, and triplet and higher- trends in live-birth rates varied mark- Since the first successful IVF procedure order births. In general, younger edly between cycles in which 1 or more in 1978,19 the field of ART has grown rap- women were at greater risk for each of embryos had been cryopreserved and idly. In the United States alone, more these outcomes given the same num- cycles in which no embryos were cryo- than 60 000 ART cycles were initiated in ber of embryos transferred. Among preserved (TABLE 3). Women in the 1996, which resulted in more than women aged 20 to 29 years, the pro- cryopreserved group achieved higher 17 000 clinical pregnancies and more portion of live births that were mul- live-birth rates with fewer embryos than 14 000 live births.20 The majority of tiple increased substantially with each transferred. Among women aged 20 to these were achieved using fresh, non-

Downloaded From: https://jamanetwork.com/ on 10/02/2021 BIRTH RATES AND MULTIPLE-BIRTH RISK USING IVF

donor IVF treatments. We examined tiple birth. The large sample size al- tively limited to 2, as indicated by 1 or these population-based data, and in lowed us to explore the interrelation- more available embryos being cryopre- keeping with prior studies,11-17,21-23 we ships between these 3 factors. served, the live-birth rates achieved found that 3 factors—patient age, num- Although we did not have specific were comparable to those achieved with ber of embryos transferred, and the abil- laboratory data to classify embryo qual- transfer of 3 embryos; however, the ity to select embryos for transfer—had ity, we found that among women multiple-birth risk was halved and the a pronounced effect on the success of an younger than 35 years, when the num- risk for triplet and higher-order preg- IVF procedure and the risk for mul- ber of embryos transferred was elec- nancies and births was virtually elimi- nated. In contrast, women aged 35 to Table 2. Key Indicators of Live Birth and Multiple Birth by Number of Embryos Transferred 39 years appeared to receive some ben- and Patient Age efit from elective transfer of 3 rather Embryos Transferred, No. than 2 embryos; although not statisti- Age, y 12 3 4 5 6Ն7cally significant at the .05 level, the live- Live Births per In Vitro Fertilization Transfer Procedure, % birth rate increased by 8 percentage 20-29 10.4 23.7* 37.7* 37.3 36.6 37.5 31.3 points. While multiple-birth rates were 30-34 9.1 19.4* 35.1* 36.4 33.0† 34.6 28.6† also increased with 3 embryos trans- 35-39 6.3 14.0* 23.0* 33.3* 29.8* 30.1 28.3 ferred (29.4%), these risks were much 40-44 2.1 5.0* 8.3† 14.4* 20.3* 20.2 15.1† smaller than those seen in women aged Multiple Births per Live Birth, % 20 to 29 years with 3 embryos trans- 20-29 . . . 22.7† 45.7* 48.1 47.8 54.6 50.0 ferred (45.7%). There were too few pro- 30-34 . . . 19.7* 39.8* 45.4* 44.1 48.0 50.0 cedures to compare elective transfer of 35-39 . . . 11.6† 29.4* 37.5* 38.4 42.4 42.4 2 vs 3 embryos among women aged 40 40-44 . . . 10.8 11.3 20.0 24.6 24.1 38.6† to 44 years; however, we observed a Triplet or Higher Gestations per Pregnancy, % trend of increasing birth rates with elec- 20-29 ...... 13.0* 18.1† 17.9 24.0 24.5 tive transfer of up to 5 embryos. Addi- 30-34 ...... 10.8* 15.8* 17.4 23.2† 18.8 tionally, the multiple-birth rate among 35-39 ...... 4.0* 11.5* 13.4 16.3† 22.3 women aged 40 to 44 years with 5 em- 40-44 ...... 0 1.8 4.2 5.9 12.5 bryos transferred (24.6%) was compa- Triplet or Higher-Order Births per Live Birth, % rable to the multiple-birth rate seen 20-29 ...... 9.9* 12.0 10.5 16.4 7.1 among women aged 20 to 29 years with 30-34 ...... 6.7* 10.0* 8.8 9.5 10.2 only 2 embryos transferred (22.7%) and 35-39 ...... 2.2† 5.4* 6.5 8.5 11.4 the triplet-birth rate was relatively low 40-44 ...... 0 0.7 2.1 0.9 5.3 at 2.1%. *PϽ.01 for comparison between the proportion in a given embryo category to the proportion in the preceding embryo When embryos were not cryopre- category within the same age group. †PϽ.05 for comparison between the proportion in a given embryo category to the proportion in the preceding embryo served, we observed increases in the category within the same age group. live-birth rate when up to 3 embryos were transferred for women aged 20 to Table 3. Live-Birth Rate by Number of Embryos Transferred, Patient Age, and Whether Extra 29 years and 30 to 34 years, when up Embryos Were Cryopreserved for Later Use to 4 embryos were transferred for Embryos Transferred, No. women aged 35 to 39 years, and when 23456Ն7up to 5 embryos were transferred for Age 20-29 y women aged 40 to 44 years. The in- 0 embryos cryopreserved 17.9 34.3* 34.2 34.1 35.4 28.3 creased embryo number needed to Ն1 embryos cryopreserved 42.7 41.1 40.3 40.5 40.0 42.9 maximize success rates for women Age 30-34 y younger than 40 years also presented 0 embryos cryopreserved 17.2 30.4* 34.3* 30.3† 33.3 28.5 important drawbacks, however, as com- Ն1 embryos cryopreserved 36.0 41.5 38.8 38.0 37.1 28.8 Age 35-39 y mensurate increases in multiple- and 0 embryos cryopreserved 13.3 19.9* 30.8* 28.6 29.3 27.6 triplet-birth rates were noted. Ն1 embryos cryopreserved 24.7 33.0 37.6† 33.2 31.9 32.9 Our findings for patients younger Age 40-44 y than 35 years are supported by prior 0 embryos cryopreserved 5.1 7.7† 13.8* 19.6* 18.8 14.8 studies,11,12,14,15,17 most notably the Ն1 embryos cryopreserved . . . 18.8 17.5 24.0 25.9 18.4 analysis of the British IVF registry by *PϽ.01 for comparison between the proportion in a given embryo category to the proportion in the preceding embryo 11 category within the same age-cryopreserved group. Templeton and Morris. This popula- †PϽ.05 for comparison between the proportion in a given embryo category to the proportion in the preceding embryo tion-based study of British IVF cycles category within the same age-cryopreserved group. found that when more than 4 eggs had

Downloaded From: https://jamanetwork.com/ on 10/02/2021 BIRTH RATES AND MULTIPLE-BIRTH RISK USING IVF

been fertilized, the odds of a live birth the number of embryos fertilized. Our differences in the pattern of results for were no different with elective trans- definition may have been a more spe- triplet gestation and birth rates may re- fer of 2 embryos compared with elec- cific indicator of embryo quality as not flect an effect due to spontaneous or tive transfer of 3 embryos; however, the only did an excess of fertilized em- therapeutic fetal reduction. Patients and multiple-birth risk was increased when bryos need to be available, but 1 or more health care professionals may con- 3 embryos were transferred. Although of them had to be deemed acceptable for sider being faced with the choice of our findings among patients older than cryopreservation. Thus, our cryopre- therapeutic fetal reduction as an addi- 35 years are supported by a previous served group may have represented a tional undesirable consequence of a clinical study,12 our results for this age more select group of cycles. If this is true, triplet or greater gestation. The trends group are not consistent with the Brit- it also follows that our group of cycles in triplet-gestation rates provide an in- ish data. While we noted an improve- for which no embryos were cryopre- dication of the total triplet risk—the po- ment in the birth rate among women served included a heterogeneous mix of tential for having a triplet birth with as- aged 35 to 39 years with elective trans- cycles. That is, our “nonelective trans- sociated infant and maternal health fer of 3 embryos, the British study fer” group included cycles for which em- risks and the potential of being faced showed no difference in live-birth rates bryo transfer was truly limited because with the decision for a therapeutic re- between elective transfer of 2 and elec- additional embryos were not available, duction. The trends in triplet-birth rates tive transfer of 3 embryos. as well as cycles in which additional em- provide a sense of the realized public Differences in IVF practice between bryos that were available for transfer health impact of triplets in 1996. the United Kingdom and the United were neither transferred nor cryopre- This study was based on observa- States might partially explain the dis- served for any number of reasons re- tional data. Although we were able to parity between our results and the Brit- lated to clinical assessment and prac- stratify on age and availability of em- ish data. The United Kingdom limits the tice or patient choice. We do not have bryos for transfer, we cannot com- number of embryos transferred to 3,6 data on the number of embryos fertil- pletely discount the possibility that while in the United States it is not un- ized and therefore cannot subdivide this women who had 3 embryos trans- common to transfer 4, 5, or even 6 em- group further. ferred had poorer quality embryos than bryos, particularly in women aged 35 The unit of analysis for this study was those who had 2 embryos transferred years or older. Thus, even in the elec- the IVF transfer procedure. Women or differed on some other unmea- tive transfer group, differential deci- who underwent more than 1 transfer sured determinant of success. Given the sion making by US and UK practition- procedure in 1996 are therefore repre- limitations of current embryo grading ers about whether to transfer 2, 3, or sented in multiple procedures. Al- methods, only a large randomized trial more embryos in women 35 years of age though we did not have the necessary would ensure complete comparability or older may have affected comparabil- data to link cycles from the same between women with different num- ity between patients included in vari- woman, we did have medical history bers of embryos transferred. ous embryo-number groups. In fact, in data for each procedure, including Although these findings are based on the United States, there were very few whether a woman had undergone pre- observational data, they strongly sug- cycles among women aged 40 years or vious ART cycles (in 1996 or earlier). gest that embryo transfer can be lim- older in which embryo transfer had been Therefore, we repeated our analysis af- ited in many women undergoing IVF, electively limited to 2. Although this ren- ter limiting the sample to women who thereby reducing the risk of multiple dered us unable to compare elective were undergoing their first cycle and birth without reducing the chance of transfer of 2 and elective transfer of 3 em- found no difference in comparison to pregnancy and live birth. Adverse fe- bryos for this oldest age group, we were our original findings (data not shown). tal and infant outcomes associated with able to examine elective transfer of We focused this presentation on the multiple pregnancy and birth have been higher numbers of embryos and found most relevant outcomes, live birth and identified as the greatest potential haz- that live-birth rates improved when more multiple birth. We also examined the ard associated with IVF therapies. As than 3 embryos were transferred, more proximal outcomes, pregnancy technology advances, we look to de- whether or not additional embryos had and multiple-gestation pregnancy. Be- velopments in embryo culture tech- been cryopreserved. Embryo transfers cause the trends observed for live- niques, such as blastocyst culture, to beyond 3 could not be evaluated with birth rates and multiple-birth rates were eliminate the need for high-order em- the British data. parallel to the trends for pregnancy rates bryo transfers for all age groups.24 Un- A further difference between the 2 and multiple-gestation rates, respec- til then, however, persons undergoing studies is the definition of elective trans- tively, we presented only the former IVF and their physicians need to care- fer. We defined elective transfer on the here. However, because trends for trip- fully consider the trade-offs between basis of whether embryos had been cryo- let-gestation rates were in some in- benefit and risk in deciding how many preserved, whereas the comparable cat- stances more pronounced than triplet- embryos to transfer. This is particu- egory in the British study was based on birth rates, we presented both. The larly critical for younger patients.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 BIRTH RATES AND MULTIPLE-BIRTH RISK USING IVF

Funding/Support: The data used for this study were Recommendations for Public Policy. New York, NY: Steirteghem AC. The relationship between embryo collected as part of the Assisted Reproductive Tech- New York State Task Force on Life and the Law; 1998. quality and the occurrence of multiple pregnancies. nology reporting system. This system is jointly sup- 7. Jones HW. Twins or more. Fertil Steril. 1995;63: Fertil Steril. 1992;57:626-630. ported by the Centers for Disease Control and Pre- 701-702. 17. Waterstone J, Parsons J, Bolton V. Elective trans- vention, Atlanta, Ga; the Society for Assisted 8. American Society for Reproductive Medicine. fer of two embryos. Lancet. 1991;337:975-976. Reproductive Technology (SART), Birmingham, Ala; Guidelines on Number of Embryos Transferred. Bir- 18. Fertility Clinic Success Rate and Certification Act the American Society for Reproductive Medicine mingham, Ala: American Society for Reproductive of 1992 (FCSRCA), Pub L No. 102-493, October 24, (ASRM), Birmingham, Ala; and RESOLVE, the Na- Medicine; 1998. 1992. tional Infertility Association, Somerville, Mass. 9. Bustillo M. Imposing limits on the number of oo- 19. Steptoe PC, Edwards RG. Birth after reimplanta- Acknowledgment: We are grateful to SART, ASRM, cytes and embryos transferred: is it necessary/wise or tion of a human embryo. Lancet. 1978;2:336. and RESOLVE, without whose contribution this work naughty/nice? Hum Reprod. 1997;12:1616-1617. 20. Centers for Disease Control and Prevention, Ameri- would not have been possible. 10. De Jonge CJ, Wolf DP. Embryo number for trans- can Society for Reproductive Medicine, Society for As- fer should be regulated. Fertil Steril. 1997;68:784- sisted Reproductive Technology, RESOLVE. 1996 As- 786. sisted Reproductive Technology Success Rates. Atlanta, REFERENCES 11. Templeton A, Morris JK. Reducing the risk of mul- Ga: Centers for Disease Control and Prevention; 1999. 1. Kiely JL. What is the population-based risk of pre- tiple births by transfer of two embryos after in vitro 21. Elsner CW, Tucker M, Sweitzer CL, et al. Mul- term birth among twins and other multiples? Clin Ob- fertilization. N Engl J Med. 1998;339:573-577. tiple pregnancy rate and embryo number transferred stet Gynecol. 1998;41:3-11. 12. Hu Y, Maxson WS, Hoffman DI, et al. Maximiz- during in vitro fertilization. Am J Obstet Gynecol. 1997; 2. Guyer B, Martin JA, MacDorman MF, Anderson RN, ing pregnancy rates and limiting higher-order mul- 177:350-355. Strobino DM. Annual summary of vital statistics— tiple conceptions by determining the optimal num- 22. Svendsen TO, Jones D, Butler L, Muasher SJ. The 1996. Pediatrics. 1997;100:905-918. ber of embryos to transfer based on quality. Fertil Steril. incidence of multiple gestations after in vitro fertiliza- 3. Pharoah POD, Cooke T. Cerebral palsy and mul- 1998;69:650-657. tion is dependent on the number of embryos trans- tiple births. Arch Dis Child Fetal Neonatal Ed. 1996; 13. Senoz S, Ben-Chetrit A, Casper RF. An IVF fal- ferred and maternal age. Fertil Steril. 1996;65:561- 75:F174-F177. lacy: multiple pregnancy risk is lower for older women. 565. 4. Gardner MO, Goldenberg RL, Cliver SP, Tucker JM, J Assist Reprod Genet. 1997;14:192-198. 23. Widra EA, Gindoff PR, Smotrich DB, Stillman RJ. Nelson KG, Copper RL. The origin and outcome of pre- 14. Tasdemir M, Tasdemir I, Kodama H, Fukuda J, Achieving multiple-order embryo transfer identifies term twin pregnancies. Obstet Gynecol. 1995;85: Tanaka T. Two instead of three embryo transfer in in- women over 40 years of age with improved in vitro 553-557. vitro fertilization. Hum Reprod. 1995;10:2155-2158. fertilization outcome. Fertil Steril. 1996;65:103-108. 5. Spellacy WN, Handler A, Ferre CD. A case-control 15. Staessen C, Janssenswillen C, Van Den Abbeel E, 24. Gardner DK, Vella P, Lane M, Wagley L, Schlen- study of 1253 twin pregnancies from a 1982-1987 peri- Devroey P, Van Steirteghem AC. Avoidance of triplet ker T, Schoolcraft WB. Culture and transfer of hu- natal data base. Obstet Gynecol. 1990;75:168-171. pregnancies by elective transfer of two good quality man blastocysts increases implantation rates and re- 6. New York State Task Force on Life and the Law. embryos. Hum Reprod. 1993;8:1650-1653. duces the need for multiple embryo transfers. Fertil Assisted Reproductive Technologies: Analysis and 16. Staessen C, Camus M, Bollen N, Devroey P, Van Steril. 1998;69:84-88.

The precondition of any fruitful relationship between literature and science is knowledge. —Aldous Huxley (1894-1963)

Downloaded From: https://jamanetwork.com/ on 10/02/2021