OBJECTIVE. The purpose of this study was to use nationwide registries to examine the health of children up to 4 years of age who were born as a result of in vitro fertilization.
METHODS. Children born after in vitro fertilization (N = 4559) from 1996 to 1999 were monitored until 2003. Two control groups were selected from the Finnish Medical Birth Register as follows: all other children (excluding children born after ovulation induction) from the same period (N = 190398, for study of perinatal health and hospitalizations) and a random sample of those children (n = 26877, for study of health-related benefits). Mortality rates and odds ratios for perinatal outcomes, hospitalizations, health-related benefits, and long-term medication use were calculated.
RESULTS. Although the health of most in vitro fertilization children was good, such children had more health problems than other children. A total of 35.7% of in vitro fertilization children and 2.2% of control children were multiple births, and the health of multiple births was worse than that of singletons. Perinatal outcomes of in vitro fertilization children were worse and hospital episodes were more common than among control children. Risks for cerebral palsy and psychological and developmental disorders were increased. Among in vitro fertilization singletons, worse results for perinatal outcomes and hospitalizations, but no increased risk for specific diseases, were found. The health of in vitro fertilization multiple births was comparable to the health of control multiple births.
CONCLUSIONS. Reducing the number of transferred embryos would improve the health of in vitro fertilization children. Additional studies are needed to explain the poorer health of in vitro fertilization singletons, as well as follow-up studies to examine the health of in vitro fertilization children from 4 years onward.
In vitro fertilization (IVF) (including intracytoplasmic sperm injections and frozen embryo transfers) is a common infertility treatment. In Finland, currently ∼2.5% of infants are born as a result of IVF,1 with women <40 years of age being able to receive 2 to 5 IVF treatment cycles within the public sector2 while paying a small fee for visits. Approximately 60% of all IVF services are supplied by private clinics, with no strict age limit. Private physicians' charges and drug costs are partly reimbursed by the Social Insurance Institution (SII). Usually, the pharmacies and IVF clinics take care of billing for the reimbursements. Approximately 76% of total IVF costs (visits, examinations, treatments, and drugs) are covered in the public sector and 50% in the private sector, with the rest being paid by women (R.K., T.S., M.G., and E.H., unpublished data, 2006).
The perinatal health of IVF singletons has been reported as being worse than that of naturally conceived singletons,3–5 with more-recent studies also showing an increased risk for preterm birth and/or low birth weight for twins.6,7 However, studies of the long-term health of IVF children are few, and their results are conflicting.
According to previous small cohort studies, the morbidity rates, growth, and development of IVF children are similar to those of control children (as reviewed by Koivurova8). Although the health of children is mainly good, large studies are needed to clarify potential health problems. Registry-based studies allow for large sample sizes and are published regularly from the Nordic countries.9–17 In the case of IVF children, research has found an excess use of hospital services, long hospitalizations, and increased risk for infections, epilepsy, and tumors,10 asthma,10,17 cerebral palsy (CP),10,11,14 sleep disturbances,14 convulsions, behavioral problems, and accidents,17 and congenital malformations.9,10,15–17 However, some of those studies were based on early IVF experience and concentrated on specific diagnoses, hospital care utilization, or singletons/twins only or did not consider multiplicity.
Results on the perinatal health of IVF twins are controversial, whereas data on the long-term health of IVF children are sparse. For this reason, our aim was to perform a large, thorough, up-to-date, registry-based study of the health of IVF children up to 4 years of age, separately for singletons and multiple births, by using several population-based registries.
Identification of IVF Children
The study is based on children born to women who received IVF between 1996 and 1998 in Finland. The women were identified, with a predesigned algorithm, from the reimbursement files of the SII.18 Data on children born as a result of IVF treatment (N = 4559) and their perinatal health were obtained from the Finnish Medical Birth Register (MBR)15,19 by using women's personal identification numbers and the children's dates of birth as the linkage keys. The MBR also includes the children's unique identification numbers. It contains information on maternal backgrounds and on infant outcomes until the age of 7 days for all infants born in Finland. The data are collected by delivery hospitals and are completed by linkage to the Central Population Register and cause-of-death statistics (compiled by Statistics Finland). The identified children were linked to 4 other nationwide registries through the children's identification numbers, namely, cause-of-death statistics, the Hospital Discharge Register (HDR) (hospital episodes, diagnoses, ie, International Classification of Diseases, 10th Revision [ICD-10] codes, and dates of admissions and discharges), the Care Register for Social Welfare (episodes in institutional care), and health-related social benefits from the SII (reimbursements for long-term medication use and child disability allowance).
As control groups, 2 groups of children were selected from the MBR. The first control group consisted of all children other than IVF children or those born as a result of ovulation induction (N = 190398) who had been conceived during the same period (1996–1998). The second control group (n = 26877) was a random sample of the first control group, selected to reduce the workload caused by large registry linkages in the SII, and was used to study the benefit payments from the SII and for the combined analysis.
The number of deaths of all children from 1996 to 2001 until the age of 2 years was obtained from cause-of-death statistics. We grouped the causes of deaths (given as ICD-10 codes) into 4 categories, namely, conditions originating from the perinatal period, congenital malformations, other medical causes, and deaths from external causes.
The HDR collects information on inpatient care and visits to outpatient clinics involving surgical or other procedures. The HDR gathers information on diagnoses (ICD-10 codes) and dates of admissions and discharges. The diagnoses include the main diagnosis and 2 secondary diagnoses for each episode. All hospitalizations until the children were 4 years of age were studied (1996–2003).
The Care Register for Social Welfare collects information on care episodes in social institutions, such as institutions for people with intellectual disabilities. For this study, we received information on the numbers of IVF children having ≥1 period of institutional care up until the end of 2004. We compared the rates of institutionalized children with the national rates for children born in 1997 or 1998, excluding the numbers of children from IVF or ovulation induction.
The SII grants child disability allowances for families who have a disabled or chronically sick child needing continuous help and surveillance at home. A child's parents applying for benefits are required to supply recent medical documents. The register of child disability allowances contains information on start and end dates, type (temporary or permanent), level (normal, increased, or special), and diagnoses. The special refund category covers ∼50 chronic diseases, entitling patients to receive higher reimbursements of long-term medication costs. Among children, the most common diseases in the special refund category are asthma, epilepsy, diabetes mellitus, and rheumatoid arthritis. The data on special refunds included the start and end dates of entitlement periods and the reasons. Information on both child disability allowance and long-term medication use was gathered from 1996 to 2001 (ie, until the children were 2 years of age).
A comparison was made between control and IVF mothers in their utilization of health care services during the pregnancy and child birth (hospitalization during the pregnancy, cesarean section, and hospitalization of ≥7 days after delivery) and in infant outcomes, first including all children and then including first births only. As health outcomes, we used very low birth weight (<1500 g), low birth weight (<2500 g), very preterm birth (<32 weeks), preterm birth (<37 weeks), low 1-minute Apgar scores (scores of 0–6), treatment in an ICU or neonatal surveillance unit, need for respiratory treatment, hospitalization of the child for ≥7 days after birth, and perinatal death.
All inpatient hospital episodes until 2 years and 4 years of age were collected separately from the HDR. The total number of hospital episodes, the length of the episodes, and the number of hospitalized children were determined. We grouped diagnoses (ICD-10 codes) into 16 categories. The 2 categories of “symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classified” (codes R00–R99) and “factors influencing health status and contact with health services” (codes Z00–Z99) were combined and renamed as “other.” Both main and secondary diagnoses were taken into account. If the child was hospitalized more than once because of the same diagnosis, then only the first hospitalization was included.
We calculated the numbers of IVF and control children who had received ≥1 child disability allowance period or reimbursement for long-term medication. The most common reasons for child disability allowance and reimbursement were counted and IVF children were compared with naturally conceived children. Finally, we combined information from the different data sources and calculated the number of children who had used services, according to any of the data sources, because of an allergic and chronic disorder and common infection-like allergy (ICD-10 codes L20–L23, L27, and L50), asthma bronchial (ICD-10 codes J45 and J46), CP (ICD-10 code G80), epilepsy (ICD-10 codes G40 and G41), diabetes mellitus (ICD-10 code E10), diarrhea (ICD-10 codes A08–A09), pneumonia (ICD-10 codes J12–J18), or disorders of psychological development and behavioral and emotional disorders usually occurring in childhood and adolescence (ICD-10 codes F80–F98).
The differences between the IVF and control groups were first tested with a χ2 test and t test for relative proportions and with logistic regression analysis, adjusting for available background characteristics. For perinatal outcomes, these characteristics were county, smoking, maternal age, socioeconomic position, and previous births. The socioeconomic position of the women was defined by using their own occupation when delivering, which is collected routinely from maternity hospitals and classified automatically by the MBR into 5 categories according to the national classification compiled by Statistics Finland, that is, upper white-collar workers, lower white-collar workers, blue-collar workers, others (entrepreneurs, students, pensioners, unemployed women, and women with an unclassified position), and unknown position.20 All analyses were made separately for singletons and multiple births. Two logit models were used, namely, an ordinary logit model in which all children were assumed to be independent and an additional model created by using the iterative, generalized, least-squares method, in which siblings born in the same delivery were assumed to be dependent.
Research Ethics and Data Protection
The study plan was approved by the National Research and Development Centre for Welfare and Health research ethics committee (September 18, 1998). For register linkages, the National Data Protection Authority was consulted, and permissions were obtained from the registry keepers.
Of the 4559 IVF children, 34.7% were twins and 1.1% were triplets. Among the 190398 control children, 2.2% were twins and only 13 sets were triplets (<0.01%). IVF mothers were older, more often married, and from a higher socioeconomic position than other mothers (Table 1).
Compared with other mothers, IVF mothers received more hospital care during pregnancy and more cesarean sections (Table 2). Adjustment for mothers' background characteristics did not change the results. Inspection of singletons and multiple births separately showed that this difference was partly, but not totally, explained by IVF children more often being twins.
Similarly, the indicators of perinatal health showed much worse health of IVF children, which was explained partly by plurality. The perinatal health of IVF multiple births was comparable to that of control multiple births; the risk for very preterm birth was increased but not statistically significantly.
Stillbirths were more common among IVF children in total, compared with other children in total (7.2 cases per 1000 vs 3.9 cases per 1000; P < .001), and among IVF singletons, compared with control singletons (6.5 cases per 1000 vs 3.7 cases per 1000; P = .014 in a test for relative proportions), but not separately for multiple births. The main causes of stillbirths were conditions originating in the perinatal period (for example, placental infarction, extreme immaturity, and abruptio placentae).
The total mortality rate up to the age of 2 years was twofold higher among IVF children, compared with control children (9.0 deaths per 1000 and 4.1 deaths per 1000, respectively). Among singletons, rates of deaths after birth until the age of 2 years were similar in all groups of children; the main causes were congenital malformations (2.4 cases per 1000 among IVF children and 1.4 cases per 1000 among control children) and conditions originating in the perinatal period (for example, extremely low birth weight and respiratory distress syndrome; 1.4 cases per 1000 and 1.3 cases per 1000, respectively). The main causes among multiple births were the same as those among singletons (malformations: 11.2 cases per 1000 and 4.6 cases per 1000; perinatal causes: 11.2 cases per 1000 and 14.8 cases per 1000, respectively), and no significant differences between the groups were found.
According to the Care Registry for Social Welfare Institutions, 2.2 IVF children per 1000 had had ≥1 period of institutional care at a social welfare institution. For other children born in 1997 to 1998, the rate was 2.7 per 1000 children. Institutional care does not include children taken into children's homes or custody but refers mainly to mentally handicapped children who stayed in an institution for people with intellectual disabilities.
Until the age of 2 years, larger proportions of IVF children and IVF singletons received child disability allowances, compared with control children (Table 3). The most common reasons (according to ICD-10 classification) for receiving child disability allowances were the same for IVF and control singletons, namely, diseases of the skin and subcutaneous tissue, diseases of the respiratory system, and conditions involving the eyes and ears. For multiple births, the most common reasons included, in addition, certain conditions originating in the perinatal period. No statistically significant differences in long-term medication use were found between IVF and control children.
When information from different data sources until the age of 2 years was combined, it was found that IVF children, singletons and multiple births taken together, had a threefold increased risk of CP and more often had disorders of psychological development or behavioral and emotional disorders, compared with control children (Table 4). This was not the case when IVF singletons and multiple births were considered separately. Of the infants with CP, 88% were preterm.
Up to the age of 4 years, a larger proportion of IVF children were hospitalized, IVF children more often had long hospital episodes, and the average length of their episodes was greater, compared with control children (Table 5). IVF children had somewhat more hospital episodes than control children at all ages, but the difference was clearest during infancy.
Compared with control children, the risk of being hospitalized was increased among IVF children for many categories of diseases (according to ICD-10 grouping), even after adjustment for the mother's socioeconomic position (data not shown). The risk among IVF singletons was increased statistically significantly for perinatal problems (ICD-10 codes P00–P96; odds ratio [OR]: 1.76; 95% confidence interval [CI]: 1.54–2.01), congenital malformations (codes Q00–Q99; OR: 1.45; 95% CI: 1.20–1.75), and problems of the genitourinary system (codes N00–N99; OR: 1.40; 95% CI: 1.11–1.77) and decreased for diseases of the respiratory system (codes J00–J99; OR: 0.86; 95% CI: 0.76–0.97). IVF multiple births had increased risk for hospitalization because of diseases originating from the perinatal period (OR: 1.34; 95% CI: 1.18–1.53) and “other” diagnoses (codes R00–R99 and Z00-Z99; OR: 1.27; 95% CI: 1.09–1.48) and decreased risk for hospitalization because of diagnoses in the categories of eye and ear (codes H00–H95; OR: 0.77; 95% CI: 0.63–0.95) and the respiratory system (OR: 0.74; 95% CI: 0.63–0.87). Otherwise, their outcomes were comparable to those of control multiple births. However, in almost every category the proportion of hospitalized children was higher among multiple births than among singletons.
In the subanalysis for first births, the results were mainly similar to the results for all children; although IVF children in total had increased risk for asthma (adjusted OR: 1.39; 95% CI: 1.08–1.79), the risk for mothers' long hospital stay for IVF singletons (OR: 1.17; 95% CI: 0.89–1.58) and the risk for cesarean sections for multiple births (OR: 1.19; 95% CI: 0.92–1.53) were not statistically significantly increased. In addition, IVF multiple births had statistically significantly decreased risk for low birth weight (adjusted OR: 0.78; 95% CI: 0.65–0.93).
There were no differences in the results of the 2 logit analyses (an ordinary logit model and an additional analysis using the iterative generalized least-squares method; see Methods). For some rare outcomes, adjustment for mother's socioeconomic position was not possible in the additional analysis because of small numbers.
We found an increased burden of disease associated with IVF, with poorer perinatal health, higher mortality rates, increased risk for hospitalization and CP, and longer hospital episodes. This burden depended in part on higher twin rates among IVF children. However, the burden of disease resulted not only from the greater number of twins but also from the poorer health of singletons, compared with naturally conceived singletons. Increased morbidity was attributable not to any specific disease but rather to small increases in many groups of diseases. In general, the health of IVF multiple births was comparable to that of other multiple births.
Are the results reliable? IVF children were identified on the basis of drugs used, laboratory and radiologic examinations, and infertility treatment procedures. We might have missed some IVF children, who would therefore be included in the control group. However, the number of missing children cannot be large18,19 and would not affect the results. The data on deaths and perinatal health received from the MBR21,22 are reliable. However, other outcome measures depend on service utilization (seeking care or applying for benefits); technically, the registers are considered to be of good quality.23
The occurrence of less-serious diseases and cases cannot be estimated from these registries, because the use of outpatient care is not registered. Our results might be biased by different thresholds for hospital admissions between IVF and control children. IVF parents, who were more often first-time parents, might have been more worried, which might have led more easily to hospital care and also longer hospital stays. It might also be that IVF children were examined more carefully by physicians, compared with naturally conceived children, if the mode of conception was known to the physicians. However, because IVF children did not have an increased rate of hospitalizations in all categories of diseases and because adjustment for parity and socioeconomic position and a subanalysis of first births did not change our results, it is unlikely that the anxiousness of parents, more-careful examinations, or lower thresholds for hospitalization alone could explain the greater frequency of visits. Rather, the greater frequency likely reflects higher morbidity rates among singleton IVF children. Furthermore, rates of almost every outcome studied were quite similar between IVF multiple births and control multiple births.
In Finland, most health care is public, financed by taxes. Private health care is covered by the national social security system, but some children are covered by additional voluntary private insurance. No private hospitals for children exist but, in 2005, ∼28% of children up to 4 years of age used private (outpatient) physicians (Social Insurance Institution of Finland, unpublished data, 2005). It is possible that, in the case of small surgical procedures, private specialist outpatient care competes with hospital outpatient clinics. If IVF children were treated more or less frequently in such private care, then a bias would result.
In Finland, health-related social benefits (child care allowance and reimbursement for long-term medication use) must be applied for. It might be that some parents are more capable of applying for the benefits. Because the adjustment for socioeconomic position did not change the results, however, there is no reason to assume that parents of IVF children with a higher socioeconomic position would receive benefits more easily than parents of control children. Informing and advising parents on these benefits is part of routine clinical practice. In addition, reimbursed diseases for long-term medication use are defined clearly, and recent medical documents are needed for receipt of both child disability allowance and support for long-term medication use. Child disability allowance is based on ICD-10 classifications and long-term medication support on defined diagnoses; therefore, it can be assumed that these are relevant in estimating disease occurrence.
Our study confirms earlier findings of poorer perinatal health,3–5,8 greater numbers of hospitalizations,9 and increased risk for congenital anomalies15,16,24 for IVF singletons, compared with naturally conceived singletons. Perinatal problems had a significant role also in hospitalizations; diseases originating from the perinatal period represented one of the most common diagnoses leading to hospitalization, among both singletons and multiple births. IVF multiple births had worse perinatal health than did IVF singletons, but IVF and control multiple births were similar with respect to perinatal health, which is largely in accordance with an earlier study (except for the finding in that study of an increased risk of admittance to a NICU and more-common longer hospitalizations after the birth).25 In contrast, a recently published Belgian study found an increased risk for preterm birth also among IVF twins, compared with naturally conceived twins, which was largely explainable by the first birth of IVF women.7 In accordance with the study by Pinborg et al,13 we did not find any excess use of hospital services among IVF multiple births.
In addition, our study confirms earlier results of higher mortality rates,8 greater numbers of hospitalizations,10,17 and increased risks for behavioral problems,17 CP,11,14 and infections10 among IVF children overall. In accordance with an earlier Finnish study based on both outpatient and inpatient visits,8 we found a slightly but not statistically significantly increased risk for diarrhea; contrary to that study, however, we did not find an increased risk for pneumonia.
Unlike previous studies,11,14 we did not find an increased risk for CP or sleeping disturbances among IVF singletons. In our study, the excess risk for CP was mainly explainable by multiplicity. In the study by Strömberg et al,11 the main reasons for the increased risk for CP were prematurity and the multiple pregnancy rate. A recent Swedish study could not confirm the increased risk for CP; the risk disappeared after adjustment for confounders.17 Furthermore, we could not find increased risk for epilepsy, tumors, or asthma among IVF children in total, as found earlier in Sweden.10 However, increased risk for epilepsy was not found in the recent Swedish study.17
A few previous studies reported about childhood morbidity for IVF multiple births. In 2 studies, no differences in neurologic sequelae were found.11,12 In our study, no increased risk for any disease among IVF multiple births was found. In general, however, IVF multiple births had higher childhood morbidity rates than did IVF singletons.
We could not find any other study examining long-term medication use, child disability allowance, and utilization of institutional care among IVF children. A slightly increased risk for child disability allowance among IVF children in our study was explainable by multiplicity, whereas no statistically significant differences in the utilization of long-term medication therapy and institutional care between the groups were found.
Potential reasons for the poorer perinatal health of IVF children include infertility itself,26–29 infertility treatments, and varying health behavior during pregnancy. Among IVF singletons, the main cause of poorer perinatal health has been suggested to be infertility itself, because of the higher incidence of preterm birth and low birth weight also among infertile women without treatment and women with infertility treatments other than IVF.30 Some modification in the gestational process induced by IVF and intracytoplasmic sperm injection has been suggested,31 as well as so-called vanishing twins (singletons originating from twin pregnancies).32 It has also been found that the risk for preterm birth increases with low-technology treatments, compared with natural pregnancy, and increases further with high-technology treatments.33
Zygosity plays a significant role when the health of IVF multiple births are compared with the health of other multiple births. In general, monozygotic twins have poorer perinatal outcomes than dizygotic twins. A larger proportion of twins are dizygotic among medically assisted pregnancies (30%), compared with naturally conceived pregnancies (1%).34 This can partly explain the results of the similar outcomes of multiple births in studies unable to take zygosity into account. In our study, 50% of IVF twins and 30% of control twins were opposite-gender twins, which suggested that more IVF children were dizygotic.
During the 1990s, the perinatal health of IVF children improved in Finland, mainly because of a decrease in higher-order multiple births.35,36 Because so many IVF pregnancies in the late 1990s were still multiple births, the health of IVF children in total was worse than that of naturally conceived children, with increased risks for CP and developmental and psychological problems. The best way to improve the health of IVF children is to favor single-embryo transfers. The way to improve the health of singletons is more problematic, because we do not know the reasons for the findings. Sufficiently large follow-up studies that consider the health of IVF children from 4 years onward are needed.
The study was supported financially by the Academy of Finland (grant 73159), the SII, the Ministry of Education (School of Doctoral Programs in Public Health), and the National Research and Development Centre for Welfare and Health.
- Accepted July 11, 2006.
- Address correspondence to Reija Klemetti, MHSc, PO Box 220, 00531, Helsinki, Finland
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Gissler M. Fertility in Finland in the past, present and future [in Finnish]. Yhteiskuntapolitiikka.2003;6 :553– 565
- Helmerhorst FM, Perquin DAM, Donker D, Keirse MJNC. Perinatal outcome of singletons and twins after assisted conception: a systematic review of controlled studies. BMJ.2004;328 :261
- ↵Verstraelen H, Goetgeluk S, Derom C, et al. Preterm birth in twins after subfertility treatment: population based cohort study. BMJ.2005;331 :1173
- ↵Koivurova S. In Vitro Fertilization in Northern Finland 1990–1995: Prenatal and Early Childhood Outcome Until Three Years of Age [doctoral thesis]. Oulu, Finland: University of Oulu; 2005. Available at: http://herkules.oulu.fi/isbn9514277228/isbn9514277228.pdf
- ↵Ericson A, Källén B. Congenital malformations in infants born after IVF: a population-based study. Hum Reprod.2001;16 :504– 509
- ↵Ericson A, Nygren KG, Otterblad Olausson P, Källén B. Hospital care utilization of infants born after IVF. Hum Reprod.2002;17 :929– 932
- ↵Pinborg A, Loft A, Schmidt L, Greisen G, Rasmussen S, Nyboe Andersen A. Neurological sequelae in twins born after assisted conception: controlled national cohort study. BMJ.2004;329 :311
- ↵Pinborg A, Loft A, Rasmussen S, Schimdt L, Rasmussen S, Nyboe Andersen A. Hospital care utilization of IVF/ICSI twins followed until 2–7 years of age: a controlled Danish national cohort study. Hum Reprod.2004;19 :2529– 2536
- ↵Lidegaard Ø, Pinborg A, Nyboe Andersen A. Imprinting diseases and IVF: Danish national IVF cohort study. Hum Reprod.2005;4 :950– 954
- ↵Statistics Finland. Classification of socio-economic groups 1989. Available at: http://tilastokeskus.fi/tk/tt/luokitukset/lk_en/sosioekon_asema_index.html
- ↵Teperi J. Multi-method approach to the assessment of data quality in the Finnish Medical Birth Registry. J Epidemiol Community Health.1993;47 :242– 247
- ↵Keskimäki I, Aro S. Accuracy and data on diagnoses, procedures and accidents in the Finnish Hospital Discharge Register. Int J Health Sci.1991;2 :15– 21
- ↵Hansen M, Bower C, Milne E, de Klerk N, Kurinczuk JJ. Assisted reproductive technologies and the risk of birth defects: a systematic review. Hum Reprod.2005;20 :328– 338
- ↵Pinborg A, Loft A, Rasmussen S, et al. Neonatal outcome in a Danish national cohort of 3438 IVF/ICSI and 10362 non-IVF/ICSI twins born between 1995 and 2000. Hum Reprod.2004;19 :435– 441
- ↵Basso O, Baird DD. Infertility and preterm delivery, birth weight, and caeserean section: a study within the Danish National Birth Cohort. Hum Reprod.2003;18 :2478– 2484
- Basso O, Olsen J. Subfecundity and neonatal mortality: longitudinal study within the Danish National Birth Cohort. BMJ.2005;330 :393– 394
- ↵Lambert RD. Safety issues in assisted reproductive technology: aetiology of health problems in singleton ART babies. Hum Reprod.2003;18 :1987– 1991
- ↵De Geyter C, De Geyter M, Steinmann S, Zhang H, Holzgreve W. Comparative birth weights of singletons born after assisted reproduction and natural conception in previously infertile women. Hum Reprod.2006;21 :705– 712
- ↵Pinborg A, Lidegaard Ø, la Cour Freiesleben N, Nyboe Andersen A. Consequences of vanishing twins in IVF/ICSI pregnancies. Hum Reprod.2005;20 :2821– 2829
- ↵Wang JX, Norman RJ, Kristiansson P. The effect of various infertility treatments on the risk of preterm birth. Hum Reprod.2002;17 :945– 949
- ↵Schachter M, Raziel A, Friedler S, Strassburger D, Bern O, Ron-El R. Monozygotic twinning after assisted reproductive techniques: a phenomenon independent of micromanipulation. Hum Reprod.2001;16 :1264– 1269
- ↵Klemetti R, Gissler M, Hemminki E. Comparison of perinatal health of children born from IVF in Finland in the early and late 1990s. Hum Reprod.2002;17 :2192– 2198
- Copyright © 2006 by the American Academy of Pediatrics