Assisted Reproductive Technology Surveillance --- United States, 2003

The number of embryos transferred in an ART procedure was categorized as 1, 2, 3, 4, or >5. The number of days of embryo culture was calculated by using dates of egg retrieval and embryo transfer and was categorized as 1, 2, 3, 4, 5, or 6. However, because of limited sample sizes, live-birth rates are presented only for the two most common days, 3 and 5. For the same reason, live-birth rates are presented for IVF with and without ICSI and not for GIFT and ZIFT. ICSI was subdivided as to whether it was used among couples diagnosed with male factor (the original indication for ICSI treatment) or couples not diagnosed with male factor.

Chi-square tests were run separately to evaluate differences in live-birth rates by select patient and treatment factors within each age group. Multivariable logistic regression was also performed to evaluate the independent effects of patient factors --- diagnosis, number of previous ART procedures, and number of previous births --- on chance to have a live birth as a result of an ART treatment. Because the patient's age is known to be a strong predictor for live birth, separate models were constructed for each of the five age groups such that these models provide an indication of the variability in live births based on patient factors within each age strata. For these analyses, the referent groups included patients with a tubal factor diagnosis, no previous ART procedures, and no previous births. Multivariable models did not include treatment factors because of multicollinearity between certain treatment factors and multiple potential effect modifications. Rather, detailed stratified analyses were performed to elucidate additional detail related to associations between different treatment factors and live birth.

In addition to presenting live-birth rates as a measure of success a second measure of success based on singleton live births is also presented according to treatment group and patient age. Singleton live births are a key measure of ART success because they have a much lower risk than multiple-infant births for adverse health outcomes, including prematurity, low birthweight, disability, and death.

Multiple birth as a separate outcome measure was also assessed. Multiple birth was assessed in two ways. First, each multiple-birth delivery was defined as a single event. A multiple-birth delivery was defined as the delivery of two or more infants in which at least one was live-born. The multiple-birth risk was thus calculated as the proportion of multiple-birth deliveries among total live-birth deliveries. Multiple birth was also assessed according to the proportion of infants from multiple deliveries among total infants (i.e., each infant was considered separately in this calculation). The proportion of live-born infants who were multiples (twins and triplets or higher order multiples) was then calculated.^§ Each of these measures represents a different focus. The multiple-birth risk, based on number of deliveries (or infant sets), provides an estimate of the individual risk posed by ART to the woman for multiple birth. The proportion of infants born in a multiple-birth delivery provides a measure of the effect of ART treatments on children in the population. Both measures are presented by type of ART treatment and by maternal age for births conceived with the patient's eggs. Multiple-birth risk is further presented by number of embryos transferred and whether additional embryos were available and cryopreserved for future use. Embryo availability (an indicator of embryo quality) has been demonstrated to have added predictive value independent of the number of embryos transferred (3,6). Proportion of infants born in a multiple-birth delivery is presented separately by patient's state of residency at time of ART treatment.

To assess the impact of ART on total births in the United States in 2003, additional analyses including all ART infants born in 2003 are presented. Because the goal of the analysis was to assess the effect of ART on the 2003 U.S. birth cohort and the Assisted Reproductive Technology Surveillance System is organized according to the date of the ART procedure rather than the infant's date of birth, a separate ART data file was created for these analyses. This data file was drawn from two different ART reporting years and was composed of 1) infants conceived from ART procedures performed in 2002 and born in 2003 (approximately 2/3 of live-birth deliveries reported to the ART Surveillance System for 2002); and 2) infants conceived from ART procedures performed in 2003 and born in 2003 (approximately 1/3 of live-birth deliveries reported to the ART Surveillance System for 2003). Data regarding the total number of live births and multiple births in the United States in 2003 were obtained from birth certificate data (U.S. natality files) from CDC's National Center for Heath Statistics (14). These data represent 100% of births registered in the United States in 2003. Data are presented in relation to the total number of infants born in the United States in 2003 by plurality of birth.

Adverse infant health outcomes, including low birthweight, very low birthweight, and preterm delivery were also evaluated. Because ART providers do not provide continued prenatal care after a pregnancy is established, birthweight and date of birth were collected via active follow-up with ART patients (83%) or their obstetric providers (17%). Low birthweight and very low birthweight were defined as <2,500 grams and <1,500 grams, respectively. Gestational age was calculated as date of birth minus date of egg retrieval (and fertilization). If date of retrieval was missing and for procedures that used frozen embryos, gestational age was calculated as date of birth minus date of embryo transfer. For comparability with the general population, date of theoretical last menstrual period (LMP) was adjusted by adding 14 days to the gestational age estimate. Preterm delivery was defined as gestational age <37 weeks. Preterm low birthweight was defined as gestational age <37 weeks and birthweight <2,500 grams. Term low birthweight was defined as gestational age >37 weeks and birthweight <2,500 grams. The rates for low birthweight, very low birthweight, preterm low birthweight, and term low birthweight among ART infants born in 2003 are presented by plurality of birth. In addition, data for each of the five outcomes are presented for ART singletons born in 2003 by type of procedure. For the most common procedure type, those using freshly fertilized embryos from the patient's eggs, the rates for each outcome are also presented according to maternal age, maternal race/ethnicity, and number of previous live births. Chi-square tests were run separately to evaluate differences in the five outcomes by type of ART procedure, maternal age, maternal race/ethnicity, and number of previous births. All analyses were performed by using the SAS^® software system (15).

Results

Of 437 medical centers in the United States and surrounding territories that performed ART in 2003, a total of 399 (91%) provided data to CDC (Figure 1). The majority of medical centers that provided ART services were located in the eastern United States, in or near major cities. Within states, the number of medical centers performing ART was variable. States with the largest number of ART centers that reported data in 2003 were California (57), New York (32), Florida (29), Texas (29), and Illinois (23). Four states and two U.S. territories had no ART medical centers (Alaska, Guam, Maine, Montana, U.S. Virgin Islands, and Wyoming).

Number and Type of ART Procedures

Overall, 122,872 ART procedures performed in 2003 were reported to CDC (Table 1). This number excludes less than 1% (n = 163) of ART procedures performed in 2003 that involved the evaluation of a new treatment procedure. The largest number of ART procedures occurred among patients who used their own freshly fertilized embryos (91,032; [74%]). Of the 122,872 procedures started, 103,017 (84%) progressed to embryo transfer. Overall, 42% of ART procedures that progressed to the transfer stage resulted in a pregnancy; 35% resulted in a live-birth delivery; and 23% resulted in a singleton live birth. Pregnancy rates, live-birth rates, and singleton live-birth rates varied according to type of ART. The highest success rates were observed among ART procedures using donor eggs and freshly fertilized embryos (59% pregnancy rate, 51% live-birth rate, and 30% singleton live-birth rate). The lowest rates were observed among procedures using the patient's eggs and thawed embryos (34% pregnancy rate, 27% live-birth rate, and 20% singleton live-birth rate).

In all, the 35,785 live-birth deliveries from ART procedures resulted in 48,756 infants (Table 1); the number of infants born was higher than the number of live-birth deliveries because of multiple-infant births. A total of 23,748 singleton infants were born as a result of ART. The largest proportion of infants born (72%; n = 35,321) were from ART procedures in which patients used freshly fertilized embryos from their own eggs.

The number of ART procedures performed among residents of each state approximately paralleled the data by medical center location (Table 2). The greatest numbers of ART procedures reported in 2003 were performed among residents of California (15,911), New York (15,534), Massachusetts (8,813), Illinois (8,676), and New Jersey (8,299). The five states with the largest number of ART procedures performed also ranked highest in terms of numbers of live-birth deliveries and infants born. ART was used by residents of certain states and territories without an ART medical center (Alaska, Guam, Maine, Montana, U.S. Virgin Islands, and Wyoming); however, each accounted for a limited percentage of total ART usage in the United States. Non-U.S. residents accounted for 1% of ART procedures, live-birth deliveries, and infants born. The ratio of number of ART procedures per million population ranged from 50 in Arkansas to 1,373 in Massachusetts, with a national average of 412 ART procedures started per million persons.

Characteristics of Patients and ART Treatments Among Women Who Used Freshly Fertilized Embryos from Their Own Eggs

Forty-six percent of ART-transfer procedures using freshly fertilized embryos from the patient's eggs were performed on women aged <35 years; 22% on women aged 35--37 years; 19% on women aged 38--40 years; 8% on women aged 41--42 years; and 4% on women aged >42 years. Patient and treatment characteristics of these women varied by age (Table 3). The most common infertility diagnoses reported among couples in which the woman was aged <41 years were male factor and tubal factor; however, diagnoses varied overall. Tubal factor and male factor were more commonly reported among younger women than women in older age categories. In contrast, diminished ovarian reserve was reported for only 2% of women aged <35 years; it was reported for 17% of women aged 41--42 years and 27% of women aged >42 years. Among all women, 10%--13% were reported as having unexplained infertility; 10%--16% were reported as having multiple female factors; and 17%--19% were reported as having both male and female factors.

Approximately 64% of women aged <35 years were undergoing their first ART procedure. The percentage of women who had undergone at least one previous ART procedure increased with age: only 42% of women aged >42 years were undergoing their first ART procedure. The percentage of women who had had a previous birth followed similar patterns. Although 21% of women aged <35 years reported at least one previous birth, this increased steadily with age: 36% of women in the oldest age group had had a previous birth.^¶

The majority of ART procedures used IVF with or without ICSI. Less than 1% of ART procedures used GIFT or ZIFT. ICSI use among couples with and without a diagnosis of male factor infertility varied by patient age. Despite variation among all age groups, the total proportion of ICSI use (i.e., combined ICSI for male factor and ICSI for other diagnoses) was greater than the proportion of in vitro fertilization with transcervical embryo transfer (IVF-ET) without ICSI.

Among all age groups, the majority of procedures included embryo culture for 3 days; the next most common procedure involved embryo culture to day 5. Culture to day 5 coincides with development of the embryo to the blastocyst stage; this technique was used more frequently among younger women.

Although limited variation existed by age, the majority of ART procedures involved transfer of more than one embryo. Among women aged <35 years, 95% of procedures involved transfer of two or more embryos, and 46% involved transfer of three or more embryos. For women aged >42 years, 83% involved transfer of two or more embryos, and 63% involved transfer of three or more embryos. The availability of extra embryos (an indicator of overall embryo quality) decreased with age. Extra embryos were available and cryopreserved for approximately 43% of women aged <35 years, whereas only 5% of women aged >42 years had extra embryos available and cryopreserved (data were not available regarding extra embryos that were not cryopreserved for future use). Overall, 1% of ART transfer procedures used a gestational carrier or surrogate. Limited variation existed by patient age.

Live-Birth Rates Among Women Who Used Freshly Fertilized Embryos from Their Own Eggs

Live-birth rates for women who underwent ART procedures using freshly fertilized embryos from their own eggs also varied by patient age and selected patient and treatment factors (Table 4). Although the average live-birth rate for ART-transfer procedures performed among women who used their own freshly fertilized eggs was 35%, live-birth rates ranged from 43% among women aged <35 years to 6% among women aged >42 years. Couples in which the woman was aged <40 years whose diagnosis of infertility was classified as ovulatory dysfunction, male factor infertility, or unexplained infertility usually had higher than average live-birth rates. Women aged <40 years with an infertility diagnosis of diminished ovarian reserve tended to have lower than average live-birth rates. The average live-birth rate for women aged 41--42 years was 15%; however, in this age category, women with an infertility diagnosis of endometriosis or uterine factor experience higher live-birth rates: 19% and 20%, respectively. The variation in success rates across diagnostic categories was not statistically significant for this age group, nor for the oldest age group (women aged >42 years). Other than women aged 41--42 years, women who had undergone a previous ART procedure had lower live-birth rates than women undergoing their first ART procedure. However, the number of previous ART procedures cannot be subdivided by whether they were successful or not because data are not available. Women in all age groups who had had one or more previous births had equal or higher live-birth rates than those with no previous births. However, the difference in live-birth rates for the number of previous births did not reach statistical significance for women aged >42 years. Multivariable adjustment for patient factors within each age strata demonstrated similar patterns to those observed in Table 4 (data not presented).

In all age groups, live-birth rates were higher among ART procedures that used IVF-ET without ICSI, in comparison with procedures that used ICSI, whether or not male factor was reported (Table 4). Among women aged <40 years, live-birth rates were particularly low among couples who used ICSI in the absence of male factor infertility. In all age groups, live-birth rates were increased among women who had extended embryo culture to day 5, transferred two or more embryos, and had extra embryos available and cryopreserved for future use. Variations in live-birth rates were statistically significant for these treatment factors within all age groups. All of the results for treatment factors need to be considered cautiously because treatment was not randomized but rather based on medical center assessment and patient choice.

Although variability among patients who used different treatment options cannot be adjusted completely, stratified analyses were used to examine associations between treatment factors and live-birth rates among more homogenous groups of patients. To address concerns that in the absence of male factor infertility ICSI might be used preferentially for women considered difficult to treat, multiple groups of patients with an indication of being difficult to treat were evaluated separately. These groups included women with previous failed ART cycles, women diagnosed with diminished ovarian reserve, and women with a low number of eggs retrieved (less than five). Within each of these groups, age-specific live-birth rates for IVF-ET with and without ICSI were examined. In all analyses, women who used IVF with ICSI had lower success rates compared with women who used IVF without ICSI (data not presented). Thus, the pattern of results remained consistent with the findings presented (Table 4). To address concerns that extended (i.e., day 5) embryo culture might be used preferentially for women with a presumed better prognosis, data regarding women deemed to have a higher likelihood of success were evaluated separately; these subgroups included women with >10 eggs retrieved, women with diagnoses other than diminished ovarian reserve, and women with extra embryos cryopreserved for future use. Again, within each of these subgroups, women who used IVF with ICSI had lower success rates compared with women who used IVF without ICSI (Table 4) (data not presented). Finally, analyses were conducted in which the data were stratified by patient age, number of embryos transferred, and number of embryos available simultaneously. These results are included with the discussion regarding multiple-birth risk.

Total live-birth rates are compared with singleton live-birth rates for women who underwent ART procedures in which freshly fertilized embryos from their own eggs were used (Figure 2). Both live-birth rates and singleton live-birth rates decreased with patient age. Across all age groups, singleton live-birth rates were lower than live-birth rates. However, the magnitude of the difference between these two measures declined with patient age. Total live-birth rates ranged from 43% among women aged <35 years to 6% among women aged >42 years, and singleton live-birth rates ranged from 27% among women aged <35 years to 5% among women aged>42 years.

Multiple-Birth Risks Associated with ART

Of 12,037 multiple-birth deliveries, 8,812 (73%) were from pregnancies conceived with freshly fertilized embryos from the patient's eggs; 1,075 (9%) were from thawed embryos from the patient's eggs; 1,836 (15%) were from freshly fertilized embryos from a donor's eggs; and 314 (3%) were from thawed embryos from a donor's eggs (Table 5). In comparison with ART procedures using the patient's eggs and freshly fertilized embryos, the risks for multiple-birth delivery were increased when eggs from a donor were used and decreased when thawed embryos were used. Among ART procedures in which freshly fertilized embryos from the patient's own eggs were used, a strong inverse relation existed between multiple-birth risk and patient age. The average multiple-birth risk (i.e., multiple-birth delivery rate) for ART procedures in which freshly fertilized embryos from the patient's eggs were used was 34%. This rate varied from 38% among women aged <35 years to 10% among women aged >42 years.

Of 48,756 infants born through ART, 51% (25,008) were born in multiple-birth deliveries (Table 5). The proportion of infants born in a multiple-birth delivery also varied by type of ART procedure and patient age.

A more detailed examination of multiple-birth risk for women who underwent ART procedures in which freshly fertilized embryos from their own eggs were used revealed that number of embryos transferred was a risk factor for multiple-birth delivery, but the magnitude of the risk varied according to patient age (Table 6). Among all age groups, transfer of two or more embryos resulted in increased live-birth delivery rates. However, the multiple-birth risk was also substantially increased. Among women aged <38 years, the percentage of multiple-birth deliveries increased with increasing number of embryos transferred from two to five or more. As a result, if success were evaluated in terms of singleton live-birth deliveries rather than total live-birth deliveries, the two youngest age groups had lower singleton success rates when three or more embryos were transferred than when two embryos were transferred. For women aged 38--40 years, transfer of three or more embryos offered a certain advantage in terms of live-birth delivery rates. However, as among younger age groups, the percentage of twin deliveries and triplets or higher order multiple-birth deliveries were increased with three or more embryos having been transferred compared with two. For women aged 41--42 and >42 years, the multiple-birth deliveries did not demonstrate a trend by number of embryos (two or more) having been transferred.

A further assessment of multiple-birth risk among patients who used freshly fertilized embryos from their own eggs and set aside extra embryos for future use is also presented (Table 6). This group can be thought of as those with elective embryo transfer because they are known to have chosen to transfer fewer embryos than the total number available. For women with elective embryo transfer who were aged <35 years, live-birth rates were >39% when only one embryo was transferred and >53% when two embryos were transferred. Whereas an increase in live-birth rates was noted among patients with single versus double elective embryo transfers, transferring two embryos posed a substantial multiple-birth risk (approximately 40%) for this group. Transferring three or more embryos posed a substantial total multiple-birth risk (46%--57%) and a substantial risk for higher order multiple births (8%--10%). For women with elective embryo transfer who were aged 35--37 years, live-birth rates were high (approximately 55%) with elective embryo transfer of a single embryo. Live-birth rates also were high (45%) among women aged 41--42 years when only two embryos were transferred. The number of cases of elective transfer of one embryo among women aged 38--40 and 41--42 years was too limited to allow adequate evaluation. Live-birth rates with elective transfer of two to five or more embryos demonstrated limited variation for women aged 38--40 years. Data are not provided for women aged >42 years, because in this age group, limited sample size precluded analysis for all number of embryos transferred categories.**

The total number and percentage of infants born in multiple-birth deliveries by maternal state of residence is presented (Table 7). The states with the highest number of ART-associated live-birth deliveries also had the highest number of infants born in multiple-birth deliveries. These include California (3,199), New York (2,940), New Jersey (1,716), Texas (1,493), Illinois (1,482), and Massachusetts (1,295). Nationwide, the percentage of infants born in multiple-birth deliveries after ART treatment was 51%; the percentage of twins and triplets or higher order multiples were 45% and 6%, respectively. The percentage of infants born in multiple-birth deliveries was >50% in the majority of states. The states with the highest proportion of infants born in multiple-birth deliveries were Puerto Rico (60%), Oregon (59%), Arizona (59%), Tennessee (58%), Louisiana (58%), North Dakota (58%), Alaska (58%), and South Dakota (58%); however, these findings should be interpreted with care because of an overall low number of live births resulting from ART in certain states.

The contribution of ART infants to the total number of U.S. infants born in 2003 is presented (Table 8). Of 4,089,950 total infants born in the United States in 2003, a total of 46,830 (1%) were conceived by ART. Infants conceived with ART accounted for 0.6% of singleton births and 18% of multiple births nationwide. Sixteen percent of all twins and 44% of infants born in triplets or higher order multiples were conceived with ART.

Perinatal Risks Associated with ART

The proportion of ART infants born in 2003 that were low birthweight, very low birthweight, preterm, preterm low birthweight, and term low birthweight are presented by plurality of birth (Table 9). The percentage of infants with low birthweight varied from 9% among singletons to 94% among triplets or higher order multiples. The percentages of very low birthweight, preterm, and preterm low birthweight followed similar patterns.

The percentages of ART singletons that were low birthweight, very low birthweight, preterm, preterm low birthweight, and term low birthweight varied by procedure type and selected maternal factors (Table 10). In comparison with singletons born after procedures using freshly fertilized embryos derived from the patient's eggs, singletons born after procedures using freshly fertilized embryos derived from donor eggs were at increased risk for four perinatal outcomes --- low birthweight, very low birthweight, preterm delivery, and preterm low birthweight. Singletons born after procedures using thawed embryos were at decreased risks for low birthweight, very low birthweight, and term low birthweight; however, they were at increased risk for preterm delivery overall. The variation in risk across procedure types did not reach statistical significance for very low birthweight.

More detailed analysis of maternal factors among singletons born after procedures using freshly fertilized embryos derived from the patient's eggs indicated limited variation in risk for any outcome according to maternal age. Lower risks were observed with a maternal race/ethnicity of non-Hispanic white. Lower risks were also observed among mother-infant pairs with one previous birth, although the difference did not reach statistical significance for very low birthweight and term low birthweight.

Discussion

According to the latest estimates of infertility in the United States from the 2002 National Survey of Family Growth, 10% of women of reproductive age (18--44 years) reported a previous infertility-associated health-care visit, and 2% reported a visit in the previous year (16). Among married couples in which the woman was of reproductive age, 7% reported they had not conceived after 12 months of unprotected intercourse. With advances in ART, couples are increasingly turning to these treatments to overcome their infertility.

Since the birth of the first infant through ART in the United States in 1981, use of ART has grown substantially. Since 1997, CDC has been monitoring ART procedures performed in the United States. During that time, a notable and consistent increase in the use of ART has occurred. The increased use of ART coupled with higher ART success rates have resulted in dramatic increases in the number of children conceived through ART each year. From 1996 (i.e., the first full year for which CDC collected data) through 2003, the number of ART procedures performed increased 90%, from 64,681 to 122,872 (1). In addition, during 1996--2003, live-birth rates for all types of ART procedures increased substantially. For the most common type of ART procedure, using freshly fertilized embryos from the patient's eggs, live-birth rates increased from 28% in 1996 to 35% in 2003. The number of infants conceived through ART procedures performed in 2003 (48,756) was more than two times higher than in 1996 (20,840).

This report documents that in 2003, ART use varied according to patient's state of residency. Residents of California, New York, Massachusetts, Illinois, and New Jersey reported the highest number of ART procedures. These states also reported the highest number of infants conceived through ART. In 2003, ART use by state of residency was not completely in line with expectations based on the total population within states (13). Whereas Massachusetts had the third highest number of ART procedures performed, it ranked thirteenth in terms of total population size.^†† Likewise, residents of District of Columbia, New Jersey, and Connecticut underwent more ART procedures than would have been expected based on their population sizes. As a result, state-specific ratios of ART procedures by population varied according to state of residency. States with the highest ratio of number of ART procedures among state residents per million population were Massachusetts (1,373), District of Columbia (990), New Jersey (961), New York (808), and Connecticut (771). This divergence is not unexpected because, in 2003, Massachusetts and New Jersey had statewide mandates for insurance coverage for ART procedures. The state variation might also be related to availability of ART services within each state. However, the relation between demand for services and availability cannot be disentangled (i.e., increased availability in certain states might reflect the increased demand for ART among state residents).

Among women who used fresh fertilized embryos from their own eggs, patient factors (e.g., infertility diagnoses, history of previous ART procedures, and previous births) varied considerably by age. The proportion of procedures in which the couple was diagnosed with male factor infertility declined with the age of the woman, while the proportion of procedures in which the couple was diagnosed with diminished ovarian reserve increased with the woman's age. History of previous ART and previous births were more common among older women. In addition, treatment factors varied considerably by the age of the woman. The proportion of procedures in which embryo transfer occurred on day 5 (i.e., the blastocyst stage) declined with the age of the woman, while the proportion of procedures in which three or more embryos were transferred increased steadily with age.

Success rates from ART use are affected by numerous patient and treatment factors; hence, considering one single measure of success in evaluating ART efficacy is not informative. At a minimum, ART treatments need to be subdivided into categories on the basis of the source of the egg (patient or donor) and the status of the embryos (freshly fertilized or thawed), because success rates vary substantially across these types. Within the type of ART treatment, further variation exists in success rates by patient and treatment factors, most notably patient age. Other factors to consider when assessing success rates are infertility diagnosis, number of previous ART procedures, number of previous births, method of embryo fertilization and transfer, number of days of embryo culture, number of embryos transferred, availability of extra embryos, and use of a gestational carrier (surrogate). Variation exists in success rates according to each of these factors.

CDC's primary focus in collecting ART data has been live-birth deliveries as an indicator of success because ART surveillance activities were developed in response to a federal mandate to report ART success rate data. This mandate requires that CDC collect data from all ART medical centers and report success rates, defined as all live births per ovarian stimulation procedures or ART procedures, for each ART clinic. Therefore, a key role for CDC has been to publish standardized data related to ART success rates, including information regarding factors that affect these rates. With these data, persons and couples can make informed decisions regarding whether to undergo this time-consuming and expensive treatment (17,18).^§§ However, success-rate data should also be balanced with consideration of effects on maternal and infant health. The data reported to CDC include information on pregnancy outcomes of public health significance and allow for monitoring of multiple-birth rates, preterm delivery, and low birthweight associated with ART.

Multiple births are associated with an increased health risk for both mothers and infants (19--21). Women with multiple-gestation pregnancies are at increased risk for maternal complications (e.g., hemorrhage and hypertension). Infants born in a multiple-birth delivery are at increased risk for prematurity, low birthweight, infant mortality, and long-term disability. The health risks associated with multiple births have also contributed to rising health-care costs. The estimated costs per live birth in 2003 ranged from $39,688--$87,788 (18).

In the United States, multiple births have increased substantially during the previous 2 decades (14,22). The rise in multiple births has been attributed to an increased use of ART and delayed childbearing (5,23,24). Although infants conceived with ART accounted for 1% of the total births in the United States in 2003, the proportion of twins and triplets or higher order multiples attributed to ART were 16% and 44%, respectively. To respond to these concerns, SART and ASRM issued voluntary guidelines on the number of embryos transferred in 1999 and recently revised them in 2004 (25).

In certain states infertility treatments (e.g., ART) might not be covered by insurance carriers, and patients might feel pressure to maximize the opportunity for live-birth delivery. In addition, anecdotal evidence suggests that certain ART providers might feel pressure to maximize their publicly reported success rates, if defined solely as total live-birth delivery, by transferring multiple embryos (26). Indeed, in the United States, high-order embryo transfer was still common practice in 2003; approximately 56% of ART cycles that used fresh, nondonor eggs or embryos and progressed to the embryo-transfer stage involved the transfer of three or more embryos; approximately 23% of cycles involved the transfer of four or more; and 18% of cycles involved the transfer of five or more embryos (1). Among women aged <35 years, the proportions of ART cycles involving four or more embryos or five or more embryos were 12% and 3 %, respectively, as women in this age category generally experience higher success rates with fewer embryos transferred. Various reports published in the scientific literature have advocated for the presentation of singleton live-birth rates as a distinct indicator of ART success (27--32). This report includes this measure and presents it with total live-birth rates. Success rates based on singleton live-birth deliveries will provide patients with a measure that more directly highlights infant outcomes with the optimal short- and long-term prognosis. Twins, albeit to a lesser extent than triplets or higher order multiples, have substantially increased risks for infant morbidity and mortality. The risks for low birthweight and preterm birth both exceed 50% for twins, and the risk for very low birthweight is 10% (14). In addition, twins are at substantially increased risk for perinatal and infant mortality (14,20,24). Thus, presentation of singleton live-birth rates is warranted.

Data regarding multiple-birth deliveries and proportion of multiple-birth infants as distinct outcomes are also provided. Data in this report indicate that 53% of infants born through ART in 2003 were multiple births; this compares with 3% in the general U.S. population during the same period (14). The twin rate was 45%, approximately 15 times higher than in the general U.S. population (3%); the triplet and higher order multiples rate was 8%, approximately 42 times higher than the general U.S. population (0.2%). Regarding the specific type of ART treatment, multiple-birth rates were among the highest for women who underwent ART procedures using freshly fertilized embryos from their own eggs (53%) or from donor eggs (60%).

In the majority of states, >50% of infants conceived through ART were born in multiple-birth deliveries. Idaho, Kentucky, Maine, New Mexico, North Carolina, Vermont, and Wyoming reported ART-associated multiple-birth rates >60%. Multiple births resulting from ART are an increasing public health concern, nationwide and for the majority of states.

For women who underwent ART procedures using freshly fertilized embryos from their own eggs, the multiple-birth risk increased when multiple embryos were transferred (two or more). However, embryo availability (an indicator of embryo quality) was also a strong predictor of multiple-birth risk and had added predictive value beyond the number of embryos transferred. When patient age, number of embryos transferred, and embryo availability were jointly considered, high live-birth rates and singleton live-birth rates were achieved, which was particularly evident among younger women as transfer of a single embryo was efficacious. Among the majority of groups, multiple-birth risk can be minimized by limiting the number of embryos transferred without compromising success rates.

In addition to the known multiple-birth risks associated with ART, singleton infants conceived from ART are at increased risk for low birthweight and preterm delivery. In this report, 9% of singleton infants conceived with ART were low birthweight, compared with 6% in the general U.S. population during the same period (14). The percentage of singleton infants conceived from ART that were very low birthweight (2%) was twice that of singletons conceived in the general U.S. population (1%), and the percentage of ART singletons born preterm (15%) was also higher than the general U.S. population (10%). Thus, adverse infant health outcomes among singletons (e.g., low birthweight and preterm delivery) should also be considered when assessing the efficacy and safety of ART. Although separating the effect of ART from the possible influence of the underlying infertility is difficult, increased ART-related risk for low birthweight among singletons has been reported in multiple subgroups of patients with different infertility patterns (10).

A comparison of perinatal outcomes among ART twins and triplets or higher order multiples with their counterparts in the general population is inadvisable. First, both ART and non-ART infertility treatments are estimated to account for a substantial proportion of multiple births in the United States, and distinguishing naturally conceived from iatrogenic multiple births is not possible. ART accounts for only 1% of the total U.S. births; however, it accounts for 16% of twins and 44% of triplets or higher order multiples in the United States. Second, the majority of multiple births conceived after ART treatment are likely dizygotic from multiple embryo transfer. Among natural conceptions, approximately one third to one half of twins might be monozygotic, depending on maternal age (33). Monozygotic twins are at increased risk for adverse outcomes in comparison with dizygotic twins (34).

This analysis was subject to certain limitations. First, ART surveillance data were reported for each ART procedure performed rather than for each patient who used ART. Linking procedures among patients who underwent more than one ART procedure in a given year is not possible. Because patients undergoing more than one procedure in a given year are most likely to be those who failed one or more treatments, the success rates reported might underestimate the true per-patient success rate. In addition, ratios of ART procedures per population might be higher than the unknown ratio of number of persons undergoing ART per population. Second, these data represent couples who sought ART services in 2003; therefore, success rates do not represent all couples with infertility who were potential ART users in 2003. Third, approximately 9% of medical centers that performed ART in 2003 did not report their data to CDC as required.

ART data are reported to CDC by the ART medical center in which the procedure was performed rather than by the state where the patient resided. In this report, ART data are presented by the female patient's state of residence. In previous reports (23), ART data were not presented by state of residence because of incomplete residency data. In 2003, residency data were missing for <9% of all live-birth deliveries reported to CDC. The range of missing residency data varied by medical center. Medical centers located in 41 states had <5% missing residency data; medical centers located in three states had 5%--10% missing residency data; and medical centers located in four states (i.e., Georgia, Massachusetts, Minnesota, and New York) had >10% missing residency data. In cases of missing residency data, residency was assigned as the state in which the ART procedure was performed. Thus, the number of procedures performed among state residents, number of infants, and number of multiple-birth infants might have been overestimated for these states. Concurrently, the numbers might be underestimated in states bordering states with missing residency data, particularly states in the Northeast region of the United States. Nonetheless, the effects of missing residency data were not substantial. Statistics were evaluated separately according to the state in which the ART medical center was located rather than the patient's state of residence. The rankings of the states in terms of total number of infants and multiple-birth infants were similar to the rankings based on patient's state of residence (data not presented).

A further consideration in reviewing the state-based statistics in this report is that the patient's state of residence was reported at the time of ART treatment. The possibility of migration during the interval between ART treatment and birth exists. Data from the U.S. Census Bureau demonstrate that annually, approximately 3% of the U.S. population move between states (35). This rate is even higher for persons aged 20--34 years.

One group with a recognized high potential for migration is members of the U.S. armed forces. Therefore, ART procedures performed among patients who attended military medical centers were evaluated separately. In 2003, a total of 739 (0.6%) ART procedures were performed in four military medical centers. These medical centers were located in California, District of Columbia, Hawaii, and Texas. In certain of these facilities, a substantial number of distinct states were listed for patient's state of residence. States and territories for which >1% of ART procedures among state residents were performed in a military medical center were Alaska, District of Columbia, Georgia, Guam, Hawaii, Maryland, Maine, North Carolina, North Dakota, New Mexico, South Carolina, South Dakota, Texas, Virginia, U.S. Virgin Islands, and Wyoming. States for which >5% of ART procedures among state residents were performed in a military medical center were District of Columbia, Guam, Virginia, and U.S. Virgin Islands. Finally, data on race/ethnicity were missing in a large proportion of procedures. Therefore, no firm conclusions can be drawn on differences between such groups.

Despite these limitations, findings from national surveillance of ART procedures performed in the United States provide useful information for patients contemplating ART, ART providers, and health-care policy makers. First, ART surveillance data can be used to monitor trends in ART use and outcomes from ART procedures. Second, data from ART surveillance can be used to assess patient and treatment factors that contribute to higher success rates. Third, ongoing surveillance data can be used to assess the risk for multiple births and adverse perinatal outcomes among singleton births. Fourth, surveillance data provide information to assess changes in clinical practice related to ART treatment.

Multiple births are one of the most important public health concerns associated with using ART. Increased use of ART treatments and the widespread practice of transferring multiple embryos during ART treatments have led to a substantial increase in multiple-birth rates in the United States (5,14,22). Balancing the chance of success with ART against the risk for multiple births is difficult in certain cases. Implementation of approaches to limit the number of embryos transferred for patients undergoing ART should reduce the occurrence of multiple births resulting from ART. Such efforts will ultimately require ART patients and providers to view treatment success in terms of singleton pregnancies and births. In addition, continued research is critical to understanding the effect of ART on maternal and child health. CDC will continue to provide updates of ART use in the United States as data become available.

Acknowledgments

The data used for this study were collected by the Society for Assisted Reproductive Technology (SART). The SART system is jointly supported by CDC, Atlanta, Georgia; SART, Birmingham, Alabama; and the American Society for Reproductive Medicine (ASRM), Birmingham, Alabama. The authors thank SART and ASRM, without whose contributions this work would not have been possible.

References

1. CDC, American Society for Reproductive Medicine, and Society for Assisted Reproductive Technology. 2003 assisted reproductive technology success rates. Atlanta, GA: US Department of Health and Human Services, CDC, National Center for Chronic Disease Prevention and Health Promotion; 2005.
2. CDC, American Society for Reproductive Medicine, Society for Assisted Reproductive Technology, and RESOLVE. 1995 assisted reproductive technology success rates. Atlanta, GA: US Department of Health and Human Services, CDC, National Center for Chronic Disease Prevention and Health Promotion; 1997.
3. Schieve LA, Peterson HB, Meikle SF, et al. Live-birth rates and multiple-birth risk using in vitro fertilization. JAMA 1999;282:1832--8.
4. Schieve LA, Meikle SF, Peterson HB, Jeng G, Burnett NM, Wilcox LS. Does assisted hatching pose a risk for monozygotic twinning in pregnancies conceived through in vitro fertilization? Fertil Steril 2000;74:288--94.
5. Reynolds MA, Schieve LA, Martin JA, Jeng G, Macaluso M. Trends in multiple births conceived using assisted reproductive technology, United States, 1997--2000. Pediatrics 2002;111(5 Part 2):1159--62.
6. Reynolds MA, Schieve LA, Jeng G, Peterson HB, Wilcox LS. Risk of multiple birth associated with in vitro fertilization using donor eggs. Am J Epidemiol 2001;154:1043--50.
7. Vahratian A, Schieve LA, Reynolds MA, Jeng G. Live-birth rates and multiple-birth risk of assisted reproductive technology pregnancies conceived using thawed embryos, USA, 1999--2000. Hum Reprod 2002;18:1442--8.
8. Wright V, Schieve LA, Vahratian A, Reynolds MA. Monozygotic twinning associated with day 5 embryo transfer in pregnancies conceived after IVF. Hum Reprod 2004;19:1831--6.
9. Kissin DM, Schieve LA, Reynolds MA. Multiple-birth risk associated with IVF and extended embryo culture: USA, 2001. Hum Reprod 2005;20:2215--23.
10. Schieve LA, Meikle SF, Ferre C, Peterson HB, Jeng G, Wilcox LS. Low and very low birthweight in infants conceived with use of assisted reproductive technology. N Engl J Med 2002;346:731--7.
11. Schieve LA, Ferre C, Peterson HB, Macaluso M, Reynolds MA, Wright VC. Perinatal outcomes among singleton infants conceived through assisted reproductive technology in the United States. Obstet Gynecol 2004;103:1144--53.
12. Schieve LA, Wilcox LS, Zeitz J, et al. Assessment of outcomes for assisted reproductive technology: overview of issues and the US experience in establishing a surveillance system. In: Vayena E, Rowe PJ, Griffin PD, eds. Current practices and controversies in assisted reproduction: report of a meeting on "Medical, Ethical and Social Aspects of Assisted Reproduction" held at WHO Headquarters in Geneva, Switzerland, September 17--21, 2001. Geneva, Switzerland: World Health Organization; 2002:361--76.
13. US Census Bureau. Annual estimates of the population for the United States and States, and for Puerto Rico: April 1, 2000 to July 1, 2005 (NST-EST2005-01). Washington, DC: US Census Bureau; 2005. Available at http://factfinder.census.gov.
14. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2003. National Vital Stat Rep 2005;54:1--116.
15. SAS^® Institute, Inc. SAS/STAT^® user's guide. Version 8. Cary, NC: SAS Institute Inc; 1999.
16. CDC. Fertility, family planning, and reproductive health of U.S. women: data from the 2002 National Survey of Family Growth. Hyattsville, MD: US Department of Health and Human Services, CDC, National Center for Health Statistics; 2005. (Vital and Health Statistics, series 23).
17. American Society for Reproductive Medicine. Frequently asked questions about infertility. Birmingham, AL: American Society for Reproductive Medicine, 2004. Available at http://www.asrm.org/Patients/faqs.html.
18. Collins J. An international survey of the health outcomes of IVF and ICSI. Hum Reprod Update 2002;8:265--77.
19. Senat MV, Ancel PY, Bouvier-Colle MH, Breart G. How does multiple pregnancy affect maternal mortality and morbidity? Clin Obstet Gynecol 1998;41:78--83.
20. ESHRE Capri Workshop Group. Multiple gestation pregnancy. Hum Reprod 2000;15:1856--64.
21. Ozturk O, Templeton A. Multiple pregnancy in assisted reproduction techniques. In: Vayena E, Rowe PJ, Griffin PD, eds. Current practices and controversies in assisted reproduction: report of a meeting on "Medical, Ethical and Social Aspects of Assisted Reproduction" held at WHO Headquarters in Geneva, Switzerland, September 17--21, 2001. Geneva, Switzerland: World Health Organization; 2002: 220--34.
22. Martin JA, Park MM. Trends in twin and triplet births: 1980--97. National Vital Stat Rep 1999;47:1--16.
23. CDC. Use of assisted reproductive technology---United States, 1996 and 1998. MMWR 2002;51:97--101.
24. Kiely JL, Kleinman JC, Kiely M. Triplets and higher-order multiple births: time trends and infant mortality. Am J Dis Child 1992;146:862--8.
25. Practice Committee, Society for Assisted Reproductive Technology and the American Society for Reproductive Medicine. Guidelines on the number of embryos transferred. Fertil Steril 2004;82:773--4.
26. Grifo J, Hoffman D, McNamee PI. We are due for a correction…and we are working to achieve one. Fertil Steril 2001;75:14.
27. ESHRE Capri Workshop Group. Prevention of twin pregnancies after IVF/ICSI by single embryo transfer. ESHRE campus course report. Hum Reprod 2001;16:790--800.
28. Cohen J, Jones HW Jr. How to avoid multiple pregnancies in assisted reproductive technologies [Review]. Semin Reprod Med 2001;19: 269--78.
29. Evers JL. Female subfertility. Lancet 2002;360:151--9.
30. Hogue CJ. Successful assisted reproductive technology: the beauty of one. Obstet Gynecol 2002;100(5 Part 1):1017--9.
31. World Health Organization. Recommendations. In: Vayena E, Rowe PJ, Griffin PD, eds. Current practices and controversies in assisted reproduction: report of a meeting on "Medical, Ethical and Social Aspects of Assisted Reproduction" held at WHO Headquarters in Geneva, Switzerland, September 17--21, 2001. Geneva, Switzerland: World Health Organization; 2002:381--96.
32. Schieve LA, Reynolds MA. What is the most relevant standard of success in assisted reproduction? Challenges in measuring and reporting success rates for assisted reproductive technology: What is optimal? Hum Reprod 2004;19:778--82.
33. Guttmacher AF. The incidence of multiple births in man and some of the other unipara. Obstet Gynecol 1953;2:22--35.
34. Derom R, Vlietinck R, Derom C, Thiery M, Van Maele G, Van den Berg H. Perinatal mortality in the East Flanders Prospective Twin Survey: preliminary results. Eur J Obstet Gynecol Reprod Biol 1991;41:25--6.
35. US Census Bureau. Geographical mobility: 2002 to 2003. Washington, DC: US Census Bureau; 2004.

* Fertility Clinic Success Rate and Certification Act of 1992 (FCSRCA), Public Law 102-493 (October 24, 1992).

^† Data regarding population size are based on July 1, 2003, estimates from the U.S. Census Bureau (13).

^§ Includes only the number of infants live-born in a multiple-birth delivery. For example, if three infants were born in a live-birth delivery and one of the three infants was stillborn, the total number of live-born infants would be two. However, these two infants would still be counted as triplets.

^¶ Data were not available to distinguish whether previous births were conceived naturally or conceived with ART or other infertility treatments.

** Results are based on total multiple-birth risk and, therefore, do not provide an indication of pregnancies that began as twins, triplets, or a higher order but reduced (either spontaneously or through medical intervention) to singletons or twins (Table 6).

^†† Data regarding population size are based on July 1, 2003,

Disclaimer   All MMWR HTML versions of articles are electronic conversions from ASCII text into HTML. This conversion may have resulted in character translation or format errors in the HTML version. Users should not rely on this HTML document, but are referred to the electronic PDF version and/or the original MMWR paper copy for the official text, figures, and tables. An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800. Contact GPO for current prices.

Date last reviewed: 5/2/2006