OBJECTIVE. Our aim was to assess the incidence of cerebral palsy among children conceived with in vitro fertilization and children conceived without in vitro fertilization.
METHODS. A population-based, cohort study, including all live-born singletons and twins born in Denmark between January 1, 1995, and December 31, 2000, was performed. Children conceived with in vitro fertilization (9255 children) were identified through the In Vitro Fertilization Register; children conceived without in vitro fertilization (394713) were identified through the Danish Medical Birth Register. Cerebral palsy diagnoses were obtained from the National Register of Hospital Discharges. The main outcome measure was the incidence of cerebral palsy in the in vitro fertilization and non-in vitro fertilization groups.
RESULTS. Children born after in vitro fertilization had an increased risk of cerebral palsy; these results were largely unchanged after adjustment for maternal age, gender, parity, small-for-gestational age status, and educational level. The independent effect of in vitro fertilization vanished after additional adjustment for multiplicity or preterm delivery. When both multiplicity and preterm delivery were included in the multivariate models, preterm delivery remained associated strongly with the risk of cerebral palsy.
CONCLUSIONS. The large proportions of preterm deliveries with in vitro fertilization, primarily for twins but also for singletons, pose an increased risk of cerebral palsy.
In vitro fertilization (IVF), including intracytoplasmic sperm injection (ICSI), is a common service within the Danish health care system, with the highest availability in Europe.1 Methodologic shortcomings in the evaluation of risks from IVF are well known, because the situation is complex and large sample sizes are needed to study infrequent events. It is known that IVF is associated with preterm delivery, an acknowledged complication for assisted reproduction with multiple births2,3 but also for IVF singletons.2,4,5 Uncertainty remains regarding whether any risk of IVF stems from the underlying cause of infertility or from the treatment.6 Only subfertility (defined as prolonged time to pregnancy or seeking fertility advice) has been explored; subfertility itself was associated with low birth weight and preterm delivery.7–10
Cerebral palsy (CP) represents a group of disorders that result from damage to the immature brain, with permanent motor disorders and frequently impaired mental development.11 Almost one half of all children with CP are delivered preterm,11 and twins have a fourfold higher risk of developing CP than do singletons.12,13 The increased risk of CP among twins has been associated with features specific to twin pregnancies, such as monochorionic pregnancies, twin-twin transfusion syndrome, and discordant birth weights in twin pairs,14,15 which suggests that the cause of CP for some twins may differ from that for singletons. Finally, being small for gestational age (SGA) is also associated with IVF2 and is a risk factor for CP, especially among children born at term or near term.16 The associations among IVF, multiple births, preterm delivery, and SGA status suggest that IVF may result in an increased risk of CP.
There are only a few population-based, cohort studies addressing the question of long-term neurologic sequelae, such as CP, among children born after IVF.17–19 A population-based, Swedish study by Stromberg et al17 found an increased risk of CP among 5680 children born after IVF between 1982 and 1995. In a population-based, Danish study, Pinborg et al18 found no difference in risk of CP among IVF twins, compared with non-IVF twins, with adjustment for gender, year of birth, maternal age, and gestational age. Lidegaard et al19 found a statistically significantly increased risk of CP among IVF singletons, of all Danish singletons born between 1995 and 2001, although no statistical adjustment was made for gestational age.
The aim of this population-based, cohort study was to assess the overall risk of CP among IVF children, compared with non-IVF children, including all live-born singletons and twins born in Denmark between 1995 and 2000. Unlike other Danish studies,18,19 we considered the risk of CP both for singletons and twins combined and for singletons and twins separately. A key goal was to clarify the effects of preterm delivery, multiple births, and SGA status on the risk of CP after IVF. To add new information and to address the shortcomings of some previous studies, we also estimated associations between the risk of CP, the underlying cause of infertility, and the treatment method for all children born after IVF in this 6-year period.
This cohort study was based on data from Danish national registers, and linkage between the registers was achieved with the civil registration number, which is a unique number given to all citizens in Denmark. Our study included all singletons and twins born alive between January 1, 1995, and December 31, 2000, as identified through the Danish Medical Birth Register, which contains information on all births in Denmark.20 Data on IVF exposure for study infants were obtained through the IVF Register.21 The IVF Register contains information on underlying causes of infertility and types of fertility treatment, namely, conventional IVF or ICSI, egg donation, number of embryos transferred, and whether the embryos transferred were fresh or frozen and thawed (frozen embryo replacement). Since 1994, it has been mandatory for fertility clinics (public and private) to report each treatment cycle to the register. The register does not record intrauterine insemination or induction of ovulation without IVF. Permission to conduct this study was obtained from the Danish Data Protection Agency.
Children diagnosed as having CP were identified through the National Register of Hospital Discharges (NRHD), on the basis of hospital inpatient discharge codes and outpatient diagnoses.22 CP was defined with an International Classification of Diseases, 10th Revision, code of G80.0 to G83.9 in the NRHD. Follow-up monitoring of the cohort for a CP diagnosis was through December 31, 2001, corresponding to a range of 1 to 7 years of age.
Additional covariate information was obtained from the Medical Birth Register, NRHD, and Statistics Denmark, including gestational age, multiplicity, infant gender, maternal age and educational level, birth weight, and parity. Maternal age was categorized as <30 years (reference category), 30 through 34 years, or ≥35 years. Gestational age was treated as a continuous variable (with weeks as the analytic unit); preterm delivery was defined as delivery before 37 completed weeks of gestation. Educational level was categorized as basic school (9–10 years of education), vocational training or intermediate-length education (11–16 years of education; reference category), or university education (≥17 years of education). Parity was dichotomized as primiparous or multiparous (reference category). Multiplicity was categorized as singletons (reference category) or twins, excluding triplets or higher multiple births because they occur for only 2% of IVF children (191 children). The IVF legislation in Denmark was changed in 1997 to recommend that only 2 fresh or 3 frozen and thawed embryos be transferred per cycle, as a rule23; consequently, the number of multiple births higher than twins actually decreased throughout the study period. SGA was defined as birth weight <2 SDs from the expected birth weight at the specific gestational age, with the method described by Marsal et al.24
We attempted to assess selected risk factors for CP specific to twin pregnancies. We estimated discordant birth weights in twin pairs, defined as birth weight variation between twins of ≥20% (relative to the heavier twin).14 Monochorionicity is associated with CP, and monochorionicity is more frequent among non-IVF twins. We had no information on chorionicity; therefore, we calculated the risk of CP for twins of opposite gender only (thereby removing the effect of a monochorionic placenta). We had no information on twin-twin transfusion syndrome.
The risk of receiving a CP diagnosis was measured as the incidence rate, with the child as the unit of analysis. The incidence rate was defined as the number of children who received a CP diagnosis in a given time frame (1995–2001) divided by the total number of person-years the children in the cohort were at risk for the diagnosis during that time. The follow-up period varied from 1 to 7 years of age; therefore, some children with CP might not have been diagnosed before the end of the follow-up period. Also, because the number of IVF treatments increased during the study period, a disproportionate number of IVF children had shorter follow-up times, compared with non-IVF children. To deal with these aspects of the data, we used Cox regression models in the statistical analysis and the incidence rate ratio was estimated as the hazard rate ratio (HRR), with 95% confidence intervals (CIs). It was verified graphically that the proportional-hazards assumption was met.
To account for correlations between twins in the cohort, robust variance estimates were calculated with the method described by Lin and Wei.25 Robust variance estimates were not applied to subsequent siblings because, although subsequent siblings would share the same mother, they would not share several of the crucial confounders included in the analyses, such as maternal age, parity, gender, and perhaps even maternal educational level. To gauge the extent to which the correlation between siblings might cause variance to be underestimated, we fit a model in which the robust variance estimates were calculated by taking into account the siblings; the results were no different from those with the model taking into account only the twin correlation. Finally, we also estimated CP risk in a stratified analysis that included only 1 pregnancy per woman.
Stratified analyses and Cox regression models were used to adjust for confounding factors, and the results were reported as HRRs with 95% CIs. The basic model included maternal age, parity, maternal educational level, and infant gender as covariates. Preterm delivery, multiplicity, and SGA status were then added to the basic model, separately and together, for evaluation of the effects of these parameters on the association between IVF and CP. All variables were tested for mutual interaction; if results were statistically significant, then an interaction term was included in the model. Additional stratified analyses were performed for singleton, twin, term, and preterm children.
Additional subanalyses (for IVF children only) evaluated the effects of the underlying cause of infertility and the treatment method on the risk of CP and the risk of preterm delivery (ie, preterm delivery was considered the outcome of interest in a separate analysis). For these analyses with Cox regression models, the CP risk estimates were adjusted for maternal age, preterm delivery, infant gender, maternal educational level, parity, and multiplicity. The risk of preterm delivery within this subcohort was estimated by using a general linear model with logarithmic link function, with adjustment for maternal age, infant gender, maternal educational level, parity, and multiplicity. Stata software (version 8; Stata Corp, College Station, TX) was used in the analyses.
A total of 403968 singletons and twins were born in Denmark between 1995 and 2000, from 307960 mothers; 9255 (2.3%) of the children were born after IVF, from 7000 mothers. Of the IVF women, 6535 (93%) contributed to the cohort with only 1 pregnancy, whereas 219529 (73%) of the non-IVF women contributed to the cohort with 1 pregnancy and 76646 (25%) of the non-IVF women contributed to the cohort with 2 pregnancies. Less than 2% of all women contributed with >2 pregnancies. The percentage of IVF singletons and twins increased from 1.4% in 1995 to 3.0% in 2000.
In comparison with mothers of non-IVF children, mothers of IVF children were older (P < .001), had more education (P < .001), and were more often primiparous (P < .001). IVF children were more often delivered preterm (18%, compared with 5% among non-IVF children; P < .001) and were more frequently of multiple births (P < .001). The SGA rate was 8.8% among IVF children and 3.6% among non-IVF children (P < .001) (Table 1). In this study, 38.6% of the IVF children were twins and 2.7% of the non-IVF children were twins. The preterm rate among IVF twins was 35.9%, similar to the preterm rate among twins in the non-IVF group (33.6%). The preterm rate among IVF singletons was 6.5%, compared with 3.7% among non-IVF singletons (P < .001). IVF twin pairs were more often of different gender (871 of 1785 twin pairs, 49%), compared with non-IVF twins (1865 of 5397 twin pairs, 35%; P < .001), and IVF twin pairs had higher rates of discordant birth weight (393 of 1732 twin pairs, 23%), compared with non-IVF twins (946 of 5157 twin pairs, 18%; P < .001).
During the follow-up period, 1048 singletons and twins were diagnosed as having CP. Forty IVF singletons and twins received a CP diagnosis (40 of 9255 infants, 0.43%) and 1008 non-IVF singletons and twins received a CP diagnosis (1008 of 394713 infants, 0.26%; P < .001).
The risk of receiving a CP diagnosis for IVF children was estimated with a crude HRR of 1.79 (95% CI: 1.28–2.50). Adjustment for maternal age, educational level, gender, and parity (the basic multivariate model) reduced the estimate to HRR of 1.61 (95% CI: 1.13–2.30); a HRR of 1.57 (95% CI: 1.10–2.23) was obtained when we included only 1 pregnancy per woman. Maternal age, parity, gender of the child, and educational level also had statistically significant independent associations with the risk of CP (Table 2).
With stratification with respect to multiplicity and application of the basic model, the HRR for CP was 1.28 (95% CI: 0.80–2.03) for IVF singletons, compared with non-IVF singletons, whereas IVF twins showed a HRR of 1.08 (95% CI: 0.57–2.05), compared with non-IVF twins. We identified 2 pairs of IVF twins (gestational ages: 28 and 30 weeks) in which both twins had CP and 3 pairs of non-IVF twins (gestational ages: 32, 35, and 40 weeks) in which both twins had CP. The risk of having >1 twin with CP was increased among IVF mothers, although not statistically significantly, with a recurrence risk ratio of 2.15 (95% CI: 0.18–19). No IVF mothers with a child with CP (singleton or twin) had a subsequent singleton with CP, but 3 non-IVF mothers with a singleton with CP had a subsequent singleton with CP, all born at term.
Not surprisingly, multiplicity, preterm delivery, and SGA status had large and statistically significant independent effects on the risk of receiving a CP diagnosis, when added separately to the basic multivariate analysis (Table 2). In the cases of preterm delivery and multiplicity, however, the risk of CP from IVF vanished. When preterm delivery was added to the basic multivariate model, the risk associated with IVF was HRR of 1.07 (95% CI: 0.76–1.52); when multiplicity was included, the HRR was 1.11 (95% CI: 0.76–1.63). In contrast, when SGA status was added to the basic multivariate model, the risk of CP in IVF remained statistically significant (HRR: 1.51; 95% CI: 1.08–2.16). We then added multiplicity, preterm delivery, and SGA status, in all possible combinations, to the basic multivariate model and observed that preterm delivery seemed to be the key factor affecting the risk of CP from IVF. In any model that included preterm delivery, the risk of CP from IVF disappeared. If a model included multiplicity without preterm delivery, then the risk of CP from IVF disappeared but, if preterm delivery was included with multiplicity, then the risks for CP from IVF or from multiplicity both disappeared. The inclusion of SGA status in any model did not seem to affect greatly the risk of CP from IVF (beyond the effects observed for preterm delivery or multiplicity), and SGA status remained a statistically significant independent risk factor for CP in all model combinations. There was statistically significant interaction between SGA status and both preterm delivery and multiplicity in the risk of CP, but the inclusion of the interaction terms also did not affect markedly the risk of CP from IVF. In the models that included preterm delivery and/or multiplicity and/or SGA status, the independent effects of the other covariates on the risk of CP were largely unchanged from the basic multivariate model (data not shown). When we added gestational age (as a continuous variable) to the basic multivariate model, the risk of CP for IVF children disappeared (HRR: 0.97; 95% CI: 0.69–1.38).
Apparently, preterm delivery is a step on the causal path to CP for IVF children. To investigate more completely whether IVF had a residual effect on the risk of CP, we performed a stratified analysis (basic multivariate model) with strata of term/preterm and multiplicity. For term singletons, the risk estimate for CP from IVF was a HRR of 0.84 (95% CI: 0.43–1.63); for term twins, the risk estimate was a HRR of 0.83 (95% CI: 0.27–2.61). For preterm singletons, the risk estimate for CP from IVF was a HRR of 1.91 (95% CI: 1.00–3.66); for preterm twins, the risk estimate was a HRR of 1.15 (95% CI: 0.54–2.45). Adding SGA status in the analyses did not change these results appreciably, and the risk estimates for CP from IVF remained nonsignificant.
The importance of preterm delivery as an independent risk factor for CP was also apparent in the population-attributable fraction for CP among IVF children due to preterm delivery, which we estimated to be 59.4% (95% CI: 58.8%–60.0%) on the basis of these study data. That means that, if we assume that the association between preterm delivery and CP is causal, then ∼59% of CP cases among IVF children may be attributable to preterm delivery. In comparison, we estimated the attributable fraction for CP among non-IVF children due to preterm delivery to be 19% (95% CI: 18.9%–19.2%). This difference between IVF and non-IVF children in attributable fraction for CP due to preterm delivery is largely attributable to the larger proportion of IVF children born preterm (18%), compared with non-IVF children (5%), rather than a difference between the 2 groups in the risk of CP associated with preterm birth.
In considering twin-specific risk factors and using the basic multivariate model, we did not find a significantly increased risk of CP for IVF twins of opposite gender, compared with non-IVF twins of opposite gender (HRR: 1.49; 95% CI: 0.52–4.31). IVF twins with discordant birth weight, compared with non-IVF twins with discordant birth weight, did not have increased risk of CP (crude HRR: 1.32; 95% CI: 0.48–3.66).
For the subanalyses among IVF children, Table 3 provides details on the underlying causes of infertility and the treatment methods. The most common underlying causes of infertility were tubal factors (4243 of 9255 cases, 46%) or male factors (2796 of 9255 cases, 30%); 18% of cases (1634 of 9255 cases) had unknown or unspecified causes. Because an individual case could have >1 underlying cause, we included only cases with a single underlying cause of infertility for the multivariate analyses (6054 of 9255 cases, 65%). Although 23 (58%) of 40 cases of CP among IVF children were associated with a tubal factor only (underlying cause in 3596 of 9255 IVF cases, 39%), we found no statistically significant association between this or any other single underlying cause of infertility and risk of CP in the IVF group (Table 3). With the same analytic approach, we did not find an increased risk of preterm delivery associated with any single underlying cause of infertility in models adjusting for maternal educational level, age, parity, gender, and multiplicity (results not shown).
The vast majority of treatments were conventional IVF, but the use of ICSI increased over the years, from 4.7% of treatments in 1995 to 31.0% in 2000. The ICSI group had a smaller proportion of children with CP, compared with the conventional IVF group, although this was not statistically significant (HRR: 0.80; 95% CI: 0.31–2.00) (Table 3). Five hundred sixty children were born after frozen embryo replacement; they had no statistically increased risk of CP (HRR: 2.19; 95% CI: 0.77–6.28), compared with children born after fresh embryo transfer, but numbers were small (Table 3). Only 111 children were born after egg donation and none of them had a CP diagnosis, as would be expected from the small sample size. In assessment of the risk of preterm delivery associated with a specific type of treatment, only children born after egg donation had an increased risk of preterm delivery (HRR: 1.69; 95% CI: 1.14–2.50), in comparison with other types of treatment and with adjustment for maternal educational level, age, parity, gender, and multiplicity.
Children born after IVF have an increased risk of CP, primarily because of preterm delivery. Low gestational age is a major step on the causal path to CP among IVF children, for singletons as well as twins.
In the basic multivariate model, the risk of CP for IVF children was increased, compared with non-IVF children (adjusted HRR: 1.61; 95% CI: 1.13–2.30). When multiplicity and preterm delivery were included in the multivariate models, the risk of IVF disappeared, which indicates that the increased risk of CP for IVF children is largely attributable to the large proportion of IVF children who are born preterm (especially because a large proportion of IVF children are both born as multiple births and born preterm) and not to the treatment itself (Table 2). Moreover, there is no apparent significant residual risk from IVF for CP among term singletons or twins; in fact, the direction of the association was somewhat protective, underscoring the importance of preterm delivery in the risk of CP after IVF. The increased risk of CP from IVF among preterm singletons approached statistical significance (HRR: 1.91; 95% CI: 1.00–3.66), which warrants additional exploration. Preterm births represent a heterogeneous group, and additional work is needed to understand the IVF-preterm birth-CP pathway, considering the different conditions leading to preterm birth.
Because SGA rates are increased with IVF and SGA status is a risk factor for CP, SGA status also might be expected to be a step on the causal pathway from IVF to CP. In this study, inclusion of SGA status in the multivariate analyses did not have a marked effect on the risk of CP among IVF children, unlike preterm delivery. A test for interaction showed that preterm delivery and SGA status had mutually modifying effects, and the association between SGA status and CP was strongest for term children. The association between preterm delivery and CP was much stronger than the association between SGA status and CP, and it is most likely that being born preterm overpowers the effect of being SGA among preterm children. However, the method used to classify SGA status in this study was based on a simplified model of fetal growth with few parameters (weight, gestational age, and gender), and more-sophisticated measures may be needed to clarify the associations among IVF, fetal growth, gestational age, and CP.
In our study, the increased risk of CP for IVF children was lower than the risk found by Stromberg et al17 but, in accordance with their findings, the risk attributable to IVF decreased when preterm delivery was included in the analyses. Lidegaard et al,19 using a slightly different Danish cohort for their study (1995–2001, compared with 1995–2000 in this study), found a statistically significantly increased risk of CP for singletons, but they did not adjust for preterm delivery. In our multivariate analysis, the risk of CP among IVF singletons was elevated but not statistically significantly.
More IVF singletons were born preterm (6.5%), compared with non-IVF singletons (3.7%; P < .001), and the etiology of preterm delivery among IVF singletons still needs to be explained. Subfertility has been associated with preterm delivery7–10 and might be part of the explanation, but we lacked the data (eg, prolonged time to pregnancy) to investigate this pathway. Another explanation for preterm delivery for IVF singletons could be the vanishing embryo syndrome. In IVF, commonly ≥2 embryos are transferred, which produces a potential risk of losing a twin or triplet in early pregnancy. Some authors found an increased risk of preterm delivery and low birth weight among IVF children born after vanishing of a co-embryo.6,26 Others hypothesized that this event might damage the surviving fetus and that CP of unknown etiology could result from the vanishing embryo syndrome, but the literature in this area is sparse.27,28 An association between CP and IVF pregnancies in which the number of embryos transferred originally was higher than the number of infants at delivery was indicated.29
Like the studies by Pinborg et al18 and Stromberg et al,17 we found no difference in the risk of CP between IVF twins and other twins. In stratified analyses for twins only, we excluded same-gender twins (thus monochorionic twins), and the risk estimate for CP from IVF for different-gender twins was elevated but did not reach statistical significance. We did not find any difference in the risk of CP between IVF twins with birth weight discordance, compared with non-IVF twins with birth weight discordance. Data on twin-twin transfusion syndrome were not available. Additional data on the unique features of twin pregnancies will be needed to draw conclusions regarding potential differences in the etiology of CP after IVF for some twins, compared with singletons.
Approximately 59% of CP cases among IVF children may be attributable to preterm delivery, which underscores the importance of minimizing the number of children born preterm with IVF. One way to reduce the overall preterm delivery rate with IVF may be to reduce the high rates of twins (and higher multiple births), because twins accounted for 78% of the preterm IVF children in this study. More-widespread use of single-embryo transfer might reduce this problem and prevent some of the adverse outcomes associated with multiple and preterm births. Evidence regarding pregnancy rates and neonatal outcomes after transfer of ≥1 embryo is still limited,30,31 but studies from Finland and Sweden, where single-embryo transfer is used to a greater extent than in many other countries, show promising results of high pregnancy rates with single-embryo transfer.32,33
Within the IVF group, we found no independent association between any single underlying cause of infertility and the risk of CP or the risk of preterm delivery. Although most cases of CP were associated with a tubal factor only, this was also the most common underlying cause. Therefore, the relationship between IVF and CP does not seem to be influenced significantly by the underlying cause of infertility, although studies with larger sample sizes for individual causes are needed to confirm these results.
In accordance with the results of the study by Pinborg et al,18 no difference in the risk of CP between children born after conventional IVF and those born after ICSI treatment was seen. For children born after transfer of frozen and thawed embryos, we found no statistically significantly increased risk of CP (HRR: 2.19; 95% CI: 0.77–6.28), but numbers were small. Additional studies on the outcomes of frozen embryo replacement are needed, because the numbers of frozen embryo replacement treatments are increasing. Frozen embryo replacement accounted for 6% of IVF treatments overall in this study, but the number in Denmark increased to 14% in 2003.34
The present study included all live-born singletons and twins born in Denmark in a 6-year period, which reduced any selection bias in the study population to a minimum. All data in the registers were collected prospectively and were independent of the conduct of the study. As in the studies by Pinborg et al18 and Lidegaard et al,19 the diagnosis of CP was retrieved from the NRHD, and the validity and completeness of CP diagnoses in the NRHD have been questioned.35 In a validation study, it was shown that the NRHD includes a number of false-positive CP diagnoses.35 Furthermore, CP caused by postnatal events is included in the NRHD, but such cases account for only 3.2% of CP cases.35 Because neither of these sources of misclassification is likely to be associated systematically with IVF exposure, the result in our data is to bias the estimate of the effect of IVF on CP risk toward the null hypothesis. A national Danish CP register is presently in the process of construction,36 with expert review of diagnoses, and reproduction of this study with the CP register, when it is established in a few years, is needed.
The findings considered only fertility treatments performed in vitro (IVF); therefore, children born after in vivo fertilization (hormone stimulation followed by insemination or intercourse) were placed in the nonexposed group. In vivo fertilization, like IVF, results in a large proportion of multiple births.3 It is possible that the risk of CP after in vivo fertilization is also increased, because of a high proportion of multiple births and preterm deliveries or for other reasons; therefore, the risk of CP with artificial reproduction in general may be even higher than presented here.
Children born after IVF have an increased risk of CP because of the association between IVF and preterm delivery, which was the most powerful predictor of CP found in this study. The risk was not associated with an underlying cause of infertility, but it was associated strongly with preterm delivery, primarily for twins but also for singletons. These findings, which are consistent with other studies, are of great public health importance and call for prevention of the high rate of multiple births and preterm deliveries in IVF.
Funding was provided by the Centers for Disease Control and Prevention (Atlanta, GA), the Ludvig and Sara Elsass Fund, the Aase and Ejnar Danielsens Fund, and the Else and Mogens Wedell-Wedellsborgs Fund. The funding sources did not participate in any part of the performance of the study.
We thank Lone Mortensen, IVF Register, National Board of Health, and Søren Leth-Sørensen, Statistics Denmark, for their assistance during data retrieval. We thank Vibeke Holsteen, MD, for supervision and advice.
- Accepted March 6, 2006.
- Address correspondence to Dorte Hvidtjørn, MPH, NANEA, Department of Epidemiology, University of Aarhus, Paludan-Müllers vej 17, 8000 Aarhus C, Denmark. E-mail:
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Andersen AN, Gianaroli L, Felberbaum R, de Mouzon J, Nygren KJ. Assisted reproductive technology in Europe, 2001: results generated from European registers by ESHRE. Hum Reprod.2005;20 :1158– 1176
- ↵Helmerhorst FM, Perquin DA, Donker D, Keirse MJ. Perinatal outcome of singletons and twins after assisted conception: a systematic review of controlled studies. BMJ.2004;328 :261
- ↵ESHRE Capri Workshop Group. Multiple gestation pregnancy. Hum Reprod.2000;15 :1856– 1864
- ↵Gissler M, Malin SM, Hemminki E. In-vitro fertilization pregnancies and perinatal health in Finland 1991–1993. Hum Reprod.1995;10 :1856– 1861
- ↵Basso O, Baird DD. Infertility and preterm delivery, birthweight, and caesarean section: a study within the Danish National Birth Cohort. Hum Reprod.2003;18 :2478– 2484
- ↵Surveillance of Cerebral Palsy in Europe. SCPE Surveillance of Cerebral Palsy in Europe 1976–1990. Grenoble, France: Surveillance of Cerebral Palsy in Europe; 2002. Report No. ISRN TIMC/RS-02-01-FR+SCPE
- ↵Pinborg A, Loft A, Schmidt L, Greisen G, Rasmussen S, Andersen AN. Neurological sequelae in twins born after assisted conception: controlled national cohort study. BMJ.2004;329 :311
- ↵Lidegaard O, Pinborg A, Andersen AN. Imprinting diseases and IVF: Danish national IVF cohort study. Hum Reprod.2005;20 :950– 954
- ↵Sandegaard, National Board of Health. Register declaration for NRHD, 2004. Available at: www.sst.dk/upload/informatik_og_sundhedsdata/sundhedsstatistik/registre/registerdeklarationer/registerdeklaration_lpr_jls.pdf. Accessed June 20, 2006
- ↵Lov 460 af 10.juni 1997. Law of artificial reproductive techniques regarding medical treatment, diagnostics and research [in Danish].
- ↵Pinborg A, Lidegaard O, la Cour Freiesleben N, Andersen AN. Consequences of vanishing twins in IVF/ICSI pregnancies. Hum Reprod.2005;20 :2821– 2829
- ↵Landy HJ, Keith LG. The vanishing twin: a review. Hum Reprod Update.1998;4 :177– 183
- ↵Hvidtjorn D, Grove J, Schendel D, et al. “Vanishing embryo syndrome” in IVF/ICSI. Hum Reprod.2005;20 :2550– 2551
- ↵Gerris JM. Single embryo transfer and IVF/ICSI outcome: a balanced appraisal. Hum Reprod Update.2005;11 :105– 121
- ↵Martikainen H, Orava M, Lakkakorpi J, Tuomivaara L. Day 2 elective single embryo transfer in clinical practice: better outcome in ICSI cycles. Hum Reprod.2004;19 :1364– 1366
- ↵Dansk Fertilitetsselskab. Annual report from the Danish Fertility Society: 2003 [in Danish]. Available at: www.fertilitetsselskab.dk/aarsrapport.htm#2003. Accessed June 20, 2006
- ↵Uldall P. Success and development for the CP registries in Europe [in Danish]. Spastikeren.2004;54 :22– 23
- Copyright © 2006 by the American Academy of Pediatrics