Morbidity in a Danish National cohort of 472 IVF/ICSI twins, 1132 non-IVF/ICSI twins and 634 IVF/ICSI singletons: health-related and social implications for the children and their families

Anja Pinborg1,3, Anne Loft1, Lone Schmidt2 and Anders Nyboe Andersen1

1 The Fertility Clinic, University of Copenhagen, Rigshospitalet, Denmark, and 2 Institute of Public Health, University of Copenhagen, Denmark

3 To whom correspondence should be addressed: The Fertility Clinic, Rigshospitalet, Copenhagen University Hospital, Blegdamsvej 9, DK-2100 Copenhagen Ø, Denmark. e-mail: apinborg{at}rh.dk


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: There is a lack of knowledge on child health as well as family well-being in IVF/ICSI twins. METHODS: These data originated from questionnaires completed by mothers taking part in a national cohort study of twin and singleton births occurring in Denmark in 1997. The overall response rate was 83%. The three cohorts consisted of all IVF/ICSI twin children (n = 472), all IVF/ICSI singletons (n = 634) and all non-IVF/ICSI twin children (n = 1132) born in Denmark in 1997. RESULTS: No major differences in physical health were observed between IVF/ICSI twins and non-IVF/ICSI twins. Compared with IVF/ICSI singletons, more IVF/ICSI twins were admitted to a neonatal intensive care unit (NICU) (P < 0.01) and more had surgical interventions (P = 0.03) and special needs (P = 0.02), moreover they had poorer speech development (P < 0.01). Correspondingly, IVF/ICSI twin mothers rated their infant’s general health poorer than IVF/ICSI singleton mothers did. All discrepancies between IVF/ICSI twins and singletons disappeared after stratification for birthweight except for NICU admissions and speech development. Multiple logistic regression analyses showed that both IVF/ICSI and non-IVF/ICSI twin parents experienced more marital stress [odds ratio (OR) 2.9, 95% CI 2.2–3.8] and that twins had more impact on the mother’s life (OR 1.7, 95% CI 1.2–2.4) compared with singletons. Nevertheless, the only predictor of low divorce/separation risk was IVF/ICSI treatment. CONCLUSION: Our study indicates that physical health of IVF/ICSI twins is comparable with that of non-IVF/ICSI twins. However, physical health of IVF/ICSI twins is poorer and the implications for the families stronger compared with IVF/ICSI singletons.

Key words: follow-up/IVF/morbidity/mortality/twins


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
During the last decades the twinning rate has increased in many European countries. Data from European Registers collected by the European Society of Human Reproduction and Embryology for the year 1999 showed that 23.9% of IVF or ICSI deliveries in Europe were twin deliveries (Nygren and Andersen, 2002Go). Although studies on the outcome of IVF pregnancies show that singleton IVF pregnancies have a slightly increased perinatal morbidity and mortality (defined as number of intrauterine or intrapartum deaths and neonatal deaths <7 days per 1000 children born with a gestational age of >=24 weeks), twin pregnancies are the main course of the overall poorer perinatal outcome in IVF pregnancies (Bergh et al., 1999Go; Dhont et al., 1999Go; Westergaard et al., 1999Go). Reports published on IVF/ICSI twin pregnancies found comparable perinatal outcome between IVF/ICSI and spontaneously conceived twins (Olivennes et al., 1996Go; Bernasko et al., 1997Go).

Two recent studies have shown that obstetric outcome for induced dizygotic (DZ) twin pregnancy is less optimal than in natural DZ twin pregnancy (Koudstaal et al., 2000Go; Lambalk and van Hooff, 2001Go). Furthermore, reports on spontaneous twins have shown higher perinatal mortality rates among same-sex and monozygotic (MZ) twins; this higher rate was limited to monochorionic MZ twins and there was no significant difference between DZ and dichorionic MZ pairs (Loos et al., 1998Go). It is generally accepted that congenital malformations [conditions registered in the International Classification of Diseases and Health Related Problems, 10th Revision (ICD-10; Danish National Board of Health, 1993), as a congenital malformation or chromosome abnormality (ICD-10: Q00– Q99)] are more common in twins than singletons, mainly in MZ twins (Little and Bryan, 1988Go).

IVF/ICSI children have an increased risk of developing cerebral palsy (Strömberg et al., 2002Go) and a higher hospitalization rate, mainly due to the high twinning rate (Ericson et al., 2002Go). Further, spontaneously conceived twins have an increased risk of cerebral palsy (Petterson et al., 1993Go). Few reports in the literature specifically address the morbidity in IVF/ICSI twins. One survey has shown a lower risk of adverse outcome in assisted reproduction treatment twins compared with spontaneously conceived twins (Minakami et al., 1998Go).

The present study assessed the morbidity in IVF/ICSI twins, reported by the mothers. The questionnaire design enabled us to gain information on parameters that are normally inaccessible in register or case–control studies, i.e. special needs, development, mother-estimated child health and social consequences for the families. We hypothesized that IVF/ICSI twins have lower morbidity and mortality rates than non-IVF/ICSI twins due to the higher rate of dichorionic IVF/ICSI twins. Secondly, we assumed that morbidity in IVF/ICSI twins is increased compared with IVF/ICSI singletons due to the poorer neonatal outcome.

The aim of this study was to test the above hypotheses in a nationwide cohort of twins born in Denmark in 1997; IVF/ICSI twins constituted the study population and non-IVF/ICSI twins the control group. All IVF/ICSI singletons born in Denmark in 1997 constituted the second control group.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
This study is part of a larger national register-based cohort study designed to assess morbidity and mortality in all twin children born after IVF or ICSI in Denmark from 1995 to 2000. To supplement and validate the register data on morbidity we generated a national survey by questionnaire, which was sent to all twin mothers (IVF/ICSI and non-IVF/ICSI) and all IVF/ICSI singleton mothers, who delivered in Denmark in 1997. This paper presents the data from the questionnaires completed by the mothers.

The Danish Medical Birth Registry (MBR), recording all pregnancies and births in Denmark, was used to identify all women who delivered twins from January 1, 1997 to December 31, 1997. Since January 1, 1994 it has been compulsory to report each initiated IVF/ICSI cycle to the Danish IVF Registry in the National Board of Health (Andersen et al., 1999Go). The personal identification number (Central Person Registry, CPR number) of the women from MBR was cross-linked with the Danish IVF registry to identify the records on the women, who had undergone IVF/ICSI treatment prior to delivery (defined as liveborn and stillborn babies delivered after 24 completed weeks of gestation). This enabled a separation of the twin mothers into the study population of IVF/ICSI twin mothers and the control group of non-IVF/ICSI twin mothers. The control group of IVF/ICSI singleton mothers was identified in the same way. The non-IVF/ICSI twin mothers, with a history of other types of assisted reproduction treatment (ovarian stimulation with or without intrauterine insemination) prior to their delivery, could not be identified through the IVF registry and were ascertained by an item in the questionnaire. Addresses were found through The Danish Central Person Registry (CPR Registry). Eleven women had emigrated, one had died, and addresses on five women were not found in the CPR Registry.

In October 2001, the questionnaires (n = 1769) were mailed to the study group consisting of all women, who delivered IVF/ICSI twins in Denmark in 1997 (n = 266), and to the two control groups; all IVF/ICSI singleton mothers (n = 764) and all non-IVF/ICSI twin mothers (n = 739) who delivered in Denmark in 1997. Also women with stillborn children or children who died later in life received a questionnaire. After two requests, 1436 questionnaires were received resulting in an overall response rate of 83% [IVF/ICSI twin mothers 89% (n = 236), IVF/ICSI singleton mothers 83% (n = 634) and non-IVF/ICSI twin mothers 77% (n = 566)]. The study thus included 472 IVF/ICSI twins, 634 IVF/ICSI singletons, and 1132 non-IVF/ICSI twin children, all 3–4 years old. Data presented in this paper originate from the questionnaires filled in by the mothers and from discharge reports. If a mother reported that her child had a malformation or was not healthy at birth or later in life, discharge reports from the relevant departments were received to verify the diagnosis (ICD-10 code). Malformations identified up to 3 years of age were reported by the mothers. In this study, a distinction was made between major and minor congenital malformations based on the severity of the malformation. (Major congenital defects were defined as abnormalities that significantly impaired normal body function or reduced life expectancy; minor anomalies were those of primarily cosmetic significance.) All children with mental retardation, independent of severity, were allocated to one group, and all cerebral palsy syndromes were combined. If a child had more than one diagnosis, the most severe diagnosis was chosen. In case of doubt, gestational age was controlled in the hospital files.

Demographic questions were addressed towards maternal age at delivery date, duration of infertility, parity, social position and divorce/separation rate.

The local ethics committee approved the study [journal number: (KF) 01-179/01].

Questionnaire
The survey comprised several types of questions. The first part ascertained demographic information and infertility history. In addition to age, parity and social position, the women were asked about chronic diseases, their history of infertility and fertility treatment in relation to their delivery in 1997. Social position was measured in a standardized way including seven items about school education, vocational training, and job position (Hansen, 1984Go). We categorized the respondents into six social classes (I, II, III, IV, V, VI). In the analyses the groups were merged into three different groups: high (I + II), medium (III + IV), and low social class (V + VI). A minor group was outside classification due to maternity leave, sickness, or unemployment.

The second part of the survey covered the pregnancy (prenatal diagnosis, hospitalization and delivery mode), perinatal outcome and admission to a neonatal care unit. The third part regarded morbidity of the children including questions on common and chronic diseases, growth, impairments and disabilities, hospital care utilization and special needs for children, for example rehabilitation services. Moreover, the mothers were asked to estimate the general health of their children, their motor function and speech development. The fourth and last part consisted of eight items on marital relationship and on the child’s/children’s impact on the mothers’ life. The ‘marital benefit’ items assessed the extent to which the child/children had ‘strengthened the marital relationship’ or ‘brought the couple closer together’. The ‘marital stress’ items assessed the extent to which the child/children had ‘caused crises in the marital relationship’, ‘caused thoughts about divorce’, or whether the couples had divorced/separated. The item on divorce/separation had a response key of ‘yes’ or ‘no’. In the remaining four items on marital relationship, the response key was a 5-point Likert Scale from (1) strongly disagree to (5) strongly agree. In the three items on ‘impact on mothers’ life’ the response key was a 4-point scale from (1) none at all to (4) a great deal.

The questionnaire was pilot-tested before use. This test showed good distribution of scores across the different response categories. (Questionnaire in English is available from A.Pinborg.)

Non-respondents
Data on all women, who did not return the questionnaire, were drawn from The Danish Medical Birth Registry and The Danish Registry of Courses of Death. Records on maternal age, parity, birthweight and mortality rate, were obtained. Age of the women was calculated on the date of delivery. For all three cohorts separate comparisons of respondents versus non-respondents were done. There were no significant differences in age, parity, rates of infants with low birthweight (LBW) or mortality rates between respondents and non-respondents in the IVF/ICSI twin group. In the IVF/ICSI singleton group, mean birthweight in respondents (3416 g) was significantly higher than in non-respondents (3253 g, P = 0.02), also in the non-IVF/ICSI twin group, mean birthweight was higher in respondents (2598 g) compared with non-respondents (2509 g, P = 0.03). Mortality rates differed between respondents and non-respondents in the two control groups; mortality rate in the IVF/ICSI singleton group was 1.3% for respondents and 5.9% for non-respondents (P = 0.001) the corresponding rates in the non-IVF/ICSI twin group were 1.2 and 5.8% (P < 0.001).

Statistics
The results were analysed using Statistical Package for Social Sciences version 10.0. Statistical significance was defined as P < 0.05. Differences of the means of continuous parametric data were analysed with the use of Student’s t-test. {chi}2-Analyses were used to compare differences of frequencies between groups. All analyses were carried out separately for all three groups. Odds ratios were calculated with the Mantel–Haenszel estimate after stratification for maternal age, parity and birthweight. The participants were divided into 5 year age groups: <30, 30–34, 35–39 and >=40 years, parity was entered as a binary variable, either nulliparae or multiparae, and birthweight was divided into three groups: <1500, 1500–2499 and >=2500 g.

Multiple logistic regression analyses were performed separately for each of the eight items measuring the extent to which the child/children had an impact on the marital relationship and on the mother’s life. The following predicting variables were entered into each model: twins versus singleton, IVF/ICSI treatment versus no IVF/ICSI treatment, nulliparae versus multiparae, birthweight <1500, 1500– 2499, >=2500 g, and maternal age <30, 30–34, 35–39, >=40 years. The outcome measures were divorce/separation, influence on marital relationship and on the mothers’ life. All items were entered as binary dependent variables. Divorce/separation was entered as ‘yes’ or ‘no’. In the four items on influence on marital relationship ‘agree’ plus ‘strongly agree’ were summed and ‘disagree’ and ‘strongly disagree’ were summed. The women, who responded ‘neither ... nor’, were excluded from the analyses. In the three items concerning influence on the mother’s life ‘a great deal’ plus ‘some’, and ‘very little’ plus ‘none at all’ were summed. We used backwards elimination and excluded predicting variables, which were not significantly associated with the outcome measure. All three groups of women were studied together.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Demographic data
Maternal characteristics
Table I shows maternal characteristics of the respondents in the three cohorts. Age and parity distribution of the women who delivered IVF/ICSI twins differed strikingly from those of the women who delivered non-IVF/ICSI twins. As expected, IVF/ICSI twin mothers were older and of lower parity; in contrast, IVF/ICSI twin mothers were younger and with higher parity compared with IVF/ICSI singleton mothers. No trend for social position was observed after stratification for maternal age and parity between women in the three cohorts. No difference was seen in duration of infertility and treatment method between IVF/ICSI twin and IVF/ICSI singleton mothers. In the non-IVF/ICSI twin group, 17.3% had a history of ovulation induction with or without insemination prior to their delivery in 1997. The proportion of boys among IVF/ICSI twins was 53.8%, IVF/ICSI singletons 53.1 and 53.0% among non-IVF/ICSI twins, no significant differences were observed between the groups. The frequency of like-sexed twin pairs in IVF/ICSI twins (50.0%) was lower compared with control twins (65.7%, P < 0.001), indicating a higher rate of dichorionic IVF/ICSI twins.


View this table:
[in this window]
[in a new window]
 
Table I. Demographic characteristics of the respondents in the three cohorts
 
Obstetric outcome
Data on respondents and non-respondents in each of the three cohorts were pooled in the analyses of birthweight and gestational age due to the unequal distribution in the two control groups. After stratification for maternal age and parity, no differences between IVF/ICSI twins and control twins were observed in the proportion of children with LBW (<2.5 kg) (OR 0.9, 95% CI 0.7–1.2), very low birthweight (VLBW, <1.5 kg) (OR 1.6, 95% CI 1.0–2.4) or in the proportion of children born before 37 completed weeks (OR 1.0, 95% CI 0.8–1.3) or before 32 completed weeks (OR 1.4, 95% CI 0.6–1.9).

IVF/ICSI twins had an increased risk of LBW (OR 7.5, 95% CI 5.5–10.2) and VLBW (OR 7.4, 95% CI 3.7–14.8) compared with IVF/ICSI singletons. Also the risk of duration of gestation <37 weeks (OR 5.0, 95% CI 3.9–6.4) and duration of gestation <32 weeks (OR 6.0, 95% CI 3.6–10.1) was strongly increased for IVF/ICSI twins compared with IVF/ICSI singletons.

No significant differences in frequencies of malformations were observed between IVF and ICSI children. Of the IVF children (twins plus singletons), 4.5% had malformations compared with 3.8% of the ICSI children (twins plus singletons). Malformations and abnormal karyotypes are listed in Appendix A.

Morbidity
Neonatal intensive care unit (NICU)
Of the IVF/ICSI twins, 39.9% were admitted to NICU; the corresponding percentages for control twins and IVF/ICSI singletons were 37.7% (P = 0.4) and 18.9% (P < 0.001), respectively. Sixteen IVF/ICSI twins and 42 control twins with admissions as healthy co-twins were not included in the analyses. Of the IVF/ICSI twins, 29.7% were admitted to NICU >7 days the values for IVF/ICSI singletons and control twins were 7.7% (P < 0.001) and 26.1% (P = 0.2) respectively. NICU admissions >28 days were seen in 11.2, 1.6% (P < 0.001) and 8.7% (P = 0.1) of IVF/ICSI twins, IVF/ICSI singletons and control twins respectively. In separate analyses for children with birthweight >=2500 g, 22.9% of the IVF/ICSI twins, 15.6% of the IVF/ICSI singletons (P < 0.01) and 24.1% of the non-IVF/ICSI twins (P = 0.7) were admitted to NICU. The average number of days spent in NICU (only children who were admitted to NICU were included) was 22.8 for IVF/ICSI twin children and 19.1 for non-IVF/ICSI twin children; an excess period of hospitalization amounting to 3.7 days was seen (P = 0.03). A similar comparison between IVF/ICSI twins and singletons showed an excess hospitalization of 12.4 days in IVF/ICSI twins (P < 0.001). Hospitalized IVF/ICSI singletons spent on average 10.4 days in NICU.

Disabilities
Disabilities are shown in Table II. There were no differences between IVF/ICSI twins and the two control groups regarding severe neurological disabilities. The four IVF/ICSI twins with severe neurological disabilities were born before 33 completed weeks, while all seven IVF/ICSI singletons with severe neurological disabilities were delivered at term. Of the control twins, five were born before 33 completed weeks, five between 33 and 37 completed weeks and four were born at term. No differences were observed between the groups regarding hearing, vision, and speech impairments or less severe neurological disorders.


View this table:
[in this window]
[in a new window]
 
Table II. Disabilities, allergic and chronic disorders and cancer in liveborn IVF/ICSI twins, IVF/ICSI singletons and control twins
 
Allergic and chronic disorders and common infections
Allergic and chronic disorders are listed in Table II. The occurrence of allergic disorders (asthma, infantile eczema and ingestion allergy) did not differ between IVF/ICSI twins and control groups. Chronic disorders and cancer occurred too rarely to make statistical comparison meaningful. The proportion of IVF/ICSI twins, IVF/ICSI singletons and control twins who had experienced otitis media was (34, 31, 31%), acute tonsillitis (21, 23, 19%), pseudo-croup (11, 8, 11%), pneumonia (10, 8, 10%), urinary infection (3, 3, 2%) and diarrhoea (39, 40, 44%). No significant differences were seen between IVF/ICSI twins and the two control groups regarding the frequency of common infections.

Surgical procedures, admissions and ambulatory visits
Among the IVF/ICSI twins, 9.3% had at least one surgical intervention during the 4-year follow-up period. The corresponding figures for IVF/ICSI singletons and non-IVF/ICSI twins were 5.8% (P = 0.03) and 8.7% (P = 0.7) respectively. IVF/ICSI twins had a higher risk of surgical interventions (OR 1.7, 95% CI 1.1–2.6) than IVF/ICSI singletons; after stratification for birthweight this difference disappeared (OR 1.6, 95% CI 1.0–2.7). The types of surgical interventions are listed in Table III.


View this table:
[in this window]
[in a new window]
 
Table III. Total number of surgical interventions in liveborn IVF/ICSI twins, IVF/ICSI singletons and control twins. One child could be registered more than once
 
We studied hospitalizations excluding delivery or NICU admissions. Of the IVF/ICSI twins, 36.7% had been hospitalized in the follow-up period, the corresponding rates for IVF/ICSI singletons and control twins were 32.1% (P = 0.1) and 33.8% (P = 0.3) respectively. IVF/ICSI twins had on average 1.60 admissions to the hospital (calculated only for hospitalized children), IVF/ICSI singletons 1.49 and control twins 1.69; no significant differences were observed. In a separate analysis made on infants with birthweight >=2500 g, no differences between IVF/ICSI twins and the control groups were found. Of the IVF/ICSI twins, IVF/ICSI singletons and non-IVF/ICSI twins, 13.5, 12.2% (P = 0.6) and 13.0% (P = 0.8) had at least one ambulatory visit during the last year.

Special needs
Special needs were defined as speech therapy, physiotherapy, occupational therapy, or educational support (auxiliary remedial teacher). The crude percentages of children (IVF/ICSI twins, IVF/ICSI singletons, non-IVF/ICSI twins) with special needs were 9.9, 6.1 and 10.7% respectively. Table IV shows odds ratios for special needs and speech therapy in IVF/ICSI twins versus control groups. A higher risk of special needs in IVF/ICSI twins compared with IVF/ICSI singletons disappeared after stratification for birthweight. Of the IVF/ICSI twins, 6.4% had the need of speech therapy; the corresponding figures for IVF/ICSI singletons and non-IVF/ICSI twins were 3.2 and 7.8% respectively. The IVF/ICSI twins more often needed speech therapy compared with IVF/ICSI singletons; after stratification for birthweight this difference remained. The need for an auxiliary remedial teacher was 0.9% among IVF/ICSI twins, 1.9% (P = 0.2) among IVF/ICSI singletons and 2.1% (P = 0.09) among control twins.


View this table:
[in this window]
[in a new window]
 
Table IV. Odds ratios (95% CI) with birthweight stratification for special needs, speech therapist, speech development and general health
 
Speech development and motor function
The women were asked to rate their infant’s speech development and their ability of walking/running on a 5-point scale from ‘much better’ to ‘much worse’ compared with children at the same age level. The IVF/ICSI twin mothers were more likely to assess their children’s speech development better than other children at the same age compared with the control twin mothers, but less likely compared with the IVF/ICSI singleton mothers (Table IV). The differences in speech development remained after adjustment for birthweight. We observed no differences in children’s ability of walking/running between IVF/ICSI twins and the control groups.

General health condition
The children’s general health condition during the last year before completing the questionnaire was evaluated in an item with four response categories from ‘healthy all the time’ to ‘sick most of the time’. In the analyses the first two and the last two categories were pooled. IVF/ICSI twin mothers estimated the general health condition of their children to be poorer than IVF/ICSI singleton mothers did; after stratification for birthweight this difference disappeared (Table IV). No differences were observed between IVF/ICSI twins and non-IVF/ICSI twins.

Mortality rates
Perinatal mortality and the number of children, who died before 4 years of age, are listed in Table V. Due to the differences in mortality rates between respondents and non-respondents in the two control groups, the numbers of dead children in the complete cohorts (respondents plus non-respondents) were analysed. None of the mortality rates differed significantly between IVF/ICSI twins and the two control groups, although there was a tendency towards a lower mortality rate in IVF/ICSI singletons (P = 0.08) and control twins (P = 0.09) compared with IVF/ICSI twins.


View this table:
[in this window]
[in a new window]
 
Table V. Mortality rates until 4 years of age in the study and the control groups, including both respondents and non-respondents
 
Social implications
Multiple logistic regression analyses were done separately for each of the items on social implications as the outcome variables (Table VI). Twins were a predictor of more marital stress (i.e. the children had caused crises in the relationship and caused thoughts about divorce) as well as less marital benefit. However, ‘no IVF/ICSI treatment’ and age >30 years were the only predictors of a high risk of divorce/separation. Twins, nulliparity, birthweight <1500 g and age <30 years were associated with infant’s having a high impact on mothers’ personal and social life. Twins and nulliparity were the only predictors of infant’s having a high impact on women’s professional life.


View this table:
[in this window]
[in a new window]
 
Table VI. Multiple logistic regression analysis with odds ratio and 95% confidence intervals on the child’s/children’s influence on marital relationship and mother’s personal life
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Data in the present national cohort study were based on questionnaires completed by the mothers. This design enabled a description of aspects on morbidity apart from diagnoses and admissions, and it enabled a gain of knowledge on implications for the families. Finally, it was a controlled follow-up study on all IVF/ICSI twins born in Denmark during the same year with a high response rate. All reported data on malformations, disabilities and chronic disorders were reviewed in the discharge reports to achieve the correct ICD-10 diagnoses. The limitations of the present study were the relatively small sample size compared with those obtainable in register studies and the lack of information from non-respondents. However, we performed a number of non-respondent analyses with register data. To our knowledge, this is the first study addressing most aspects of morbidity in IVF/ICSI twins including implications for the families.

In contrast to our first hypothesis, concerning lower morbidity in IVF/ICSI twins versus control twins due to the higher dichorionic rate, we found comparable crude odds ratios between IVF/ICSI twins and non-IVF/ICSI twins in all the investigated areas of morbidity besides speech development. IVF/ICSI twin mothers assessed their infant’s speech development better than non-IVF/ICSI twin mothers did. This may be explained by the fact that IVF/ICSI mothers were more emotionally involved and interacted more with their children compared with non-IVF/ICSI mothers, thereby facilitating the speech development (Golombok et al., 1996Go). On the other hand, the IVF/ICSI twin mothers may have rated speech development more positively due to the lack of older siblings to compare with.

Consistent with previous controlled studies (Bergh et al., 1999Go; Dhont et al., 1999Go; Westergaard et al., 1999Go), we found obstetric outcome in twins conceived after IVF/ICSI treatment comparable with non-IVF/ICSI twins. The higher frequency of children born before 32 completed weeks and of children with VLBW in IVF/ICSI twin deliveries disappeared after adjustment for age and parity.

Although we observed no discrepancies between IVF/ICSI twins and singletons regarding common infections, chronic disorders and allergy, our results confirmed the second hypothesis that IVF/ICSI twins have higher morbidity compared with IVF/ICSI singletons, i.e. a higher frequency of NICU admissions, surgical interventions and special needs including poorer speech development. Correspondingly, IVF/ICSI twin mothers rated general health of their offspring poorer than IVF/ICSI singleton mothers did. These results are consistent with two earlier reports addressing growth and physical outcome in 38 and 94 IVF twins, respectively (Brandes et al., 1992Go; Saunders et al., 1996Go).

Two recent register studies, including >1000 IVF twins, have shown an increased risk of cerebral palsy and a higher hospitalization rate in IVF twins compared with IVF singletons (Ericson et al., 2002Go; Strömberg et al., 2002Go). These findings were not confirmed in the present study, probably due to the fact that fewer subjects were included. Although we observed no cases of cancer in the study population, our study is consistent with a previous survey on cancer incidence in IVF children, where no difference between IVF and non-IVF children was observed (Bruinsma et al., 2000Go).

Most previous reports on perinatal mortality rates between IVF and control twins have shown no significant differences (Bergh et al., 1999Go; Westergaard et al., 1999Go), while studies with separate analyses for dizygotic twins have shown diverging results (Dhont et al., 1999Go; Koudstaal et al., 2000Go; Lambalk et al., 2001Go). In the present study, we observed no significant difference (P = 0.12) in perinatal mortality rates between IVF/ICSI twins and control twins. A Danish register study showed that the number of stillborn children and children who died within the first year of life was only slightly and not significantly higher in IVF/ICSI twins compared with IVF/ ICSI singletons (Westergaard et al., 1999Go). This finding was confirmed in our study, although the difference only reached a significance level of P = 0.08.

Obstetric outcome and mortality rates did not differ between respondents and non-respondents in the IVF/ICSI twin group. By contrast, non-respondents in the two control groups had a lower mean birthweight and a higher mortality rate compared with respondents. Presumably, the explanation is that IVF/ICSI twin mothers felt an obligation to complete the questionnaire due to the interest in the future of their offspring. It is possible that the adverse obstetric outcome of non-respondents in the control groups reflects higher morbidity rates, so we may perhaps underestimate the general degree of morbidity in the control groups. Stratification for birthweight or exclusion of children with birthweight <2500 g from the analyses compensated for the unequal distribution between respondents and non-respondents in the control groups. Furthermore, data on respondents and non-respondents were pooled in each of the three cohorts regarding perinatal outcome and mortality rates.

Twins were a predictor of increased marital stress and less marital benefit compared with singletons, nevertheless the only predictors of divorce/separation were ‘no IVF/ICSI treatment’ and age >30 years. The crude numbers showed that 7.3, 6.9 and 13.3% of the couples with IVF/ICSI twins, IVF/ICSI singletons and non-IVF/ICSI twins had divorced/separated 4 years after the delivery. Despite twins causing increased marital stress, IVF/ICSI twin parents had a low divorce/separation risk, indicating strong marital relationships in IVF/ICSI parents. Presumably, IVF/ICSI parents cope better with the increased marital stress, thus avoiding divorce/separation. As expected, twins were a predictor of children having high influence on all indicators of personal life of the mothers. Nulliparity was also associated with child/children having high impact on all the assessed aspects of mother’s life, as previously shown (Colpin et al., 1999Go). Maternal age <30 years and delivery of a child with VLBW was associated with infants having high impact on the mother’s life, but not on her professional life. Altogether, these results suggest that twins have more impact on their families than singletons.

Two reports comparing health economic costs in single embryo transfer (SET) versus double embryo transfer (DET) have shown that SET, if long-term morbidity is taken into account, is more cost-efficient (Wølner-Hanssen and Rydhstroem, 1998Go; De Sutter et al., 2002Go). Moreover, prospective SET studies have shown pregnancy rates of 40% in younger women with high quality embryos (Gerris et al., 1999Go; Martikainen et al., 2001Go; De Neuborg et al., 2002Go). Our study indicates that physical health of IVF/ICSI twins is comparable with that of non-IVF/ICSI twins and not, as hypothesized, better. As expected, IVF/ICSI twins have poorer physical health and cause more strain on their families compared with IVF/ICSI singletons. The impact of this study may imply a reduction of the twin rate, thereby minimizing the individual health risk of each IVF/ICSI child as well as the strain on their families. Twin rate reduction requires a change in the embryo transfer policy into SET in selected patients.


    Acknowledgements
 
The Danish Medical Research Council, Queen Elisabeth’s Children’s Hospital Research Fund and The Medical Research Fund of Copenhagen, The Faroe Islands and Greenland approved the study.


View this table:
[in this window]
[in a new window]
 
Appendix A. The total number of major and minor congenital malformations in intrauterine or intrapartum death and liveborn children with gestational age of >=24 weeks
 

    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Andersen, A.N., Westergaard, H.B. and Olsen, J. (1999) The Danish in vitro fertilisation (IVF) register. Dan. Med. Bull., 46, 357–360.[ISI][Medline]

Bergh, T., Ericson, A., Hillensjö, T., Nygren, K-G. and Wennerholm, U-B. (1999) Deliveries and children born after in-vitro fertilisation in Sweden 1982–95: a retrospective cohort study. Lancet, 354, 1579–1585.[CrossRef][ISI][Medline]

Bernasko, J., Lynch, L., Lapinski, R. and Berkowitz, L. (1997) Twin pregnancies conceived by assisted reproductive techniques: maternal and neonatal outcomes. Obstet. Gynecol., 89, 368–372.[Abstract/Free Full Text]

Brandes, J.M., Scher, A., Itzkovits, J., Thaler, I., Sarid, M. and Gershoni-Baruch, R. (1992) Growth and development of children conceived by in vitro fertilization. Pediatrics, 90, 424–429.[Abstract]

Bruinsma, F., Venn, A., Lancaster, P., Speirs, A. and Healy, D. (2000) Incidence of cancer in children born after in-vitro fertilization. Hum. Reprod., 15, 604–607.[Abstract/Free Full Text]

Colpin, H., De Munter, A., Nys, K. and Vandemeulebroecke, L. (1999) Parenting stress and psychosocial well-being among parents with twins conceived naturally or by reproductive technology. Hum. Reprod., 14, 3133–3137.[Abstract/Free Full Text]

De Neuborg, D., Mangelschots, K., Van Royen, E., Vercruyssen, M., Ryckaert, G., Valkenburg, M., Barudy-Vasquez, J. and Gerris, J. (2002) Impact of patients’ choice for single embryo transfer of a top quality embryo versus double embryo transfer in the first IVF/ICSI cycle. Hum. Reprod., 17, 2621–2625.[Abstract/Free Full Text]

De Sutter, P., Gerris, J. and Dhont, M. (2002) A health-economic decision–analytic model comparing double with single embryo transfer in IVF/ICSI. Hum. Reprod., 17, 2891–2896.[Abstract/Free Full Text]

Dhont, M., Sutter, P.D., Ruyssinck, G., Martens, G. and Bekaert, A. (1999) Perinatal outcome of pregnancies after assisted reproduction: a case–control study. Am. J. Obstet. Gynecol., 181, 688–695.[ISI][Medline]

Ericson, A., Nygren, K.G., Otterblad Olauson, P. and Källén, B. (2002) Hospital care utilization of infants born after IVF. Hum. Reprod., 17, 929–932.[Abstract/Free Full Text]

Gerris, J., De Neubourg, D., Mangelschots, K., Van Royen, E., Van de Meerssche, M. and Valkenburg, M. (1999) Prevention of twin pregnancy after in-vitro fertilization or intracytoplasmic sperm injection based on strict embryo criteria: a prospective randomized clinical trial. Hum. Reprod., 14, 2581–2587.[Abstract/Free Full Text]

Golombok, S., Brewaeys A., Cook, R., Giavazzi, M.T., Guerra, D., Mantovani, A., van Hull, E., Crosignani, P.G. and Dexeus, S. (1996) The European study of assisted reproduction families: family functioning and child development. Hum. Reprod., 11, 2324–2331[Abstract]

Hansen, E.J. (1984) Socialgrupper i Danmark [Social Class in Denmark]. The Danish National Institute of Social Research, Copenhagen, 189 pp.

Koudstaal, J., Bruinse, H.W., Helmerhorst, F.M., Vermeiden, J.P.W., Willemsen, W.N.P. and Visser, G.H.A. (2000) Obstetric outcome of twin pregnancies after in-vitro fertilization: a matched control study in four Dutch University hospitals. Hum. Reprod., 15, 935–940.[Abstract/Free Full Text]

Lambalk, C.B. and van Hooff, M. (2001) Natural versus induced twinning and pregnancy outcome: a Dutch nationwide survey of primiparous dizygotic twin deliveries. Fertil. Steril., 75, 731–736.[CrossRef][ISI][Medline]

Little, J. and Bryan, E. (1988) Congenital anomalies. In MacGillivray, I., Campbell, D.M. and Thompson, B. (eds), Twinning and Twins. Wiley, Chichester, chap. 11, pp. 207–240.

Loos, R., Derom, C., Vlietinck, R. and Derom, R. (1998) The East Flanders Prospective Twin Survey: a population-based register. Twin Res., 1, 167–175.[Medline]

Martikainen, H., Tiitinen, A., Tomás, C., Tapainen, J., Orava, M., Tuomivaara, L., Vilska, S., Hydén-Granskog, C., Hovatta, O. and the Finish ET Study Group (2001) One versus two embryo transfer after IVF and ICSI: a randomized study. Hum. Reprod., 16, 1900–1903.[Abstract/Free Full Text]

Minakami, H., Sayama, M., Honma, Y., Matsubara, S., Koike, T., Sato, I., Uchida, A., Eguchi, Y., Momoi, M. and Araki, S. (1998) Lower risk of adverse outcome in twins conceived by artificial reproductive techniques compared with spontaneously conceived twins. Hum. Reprod., 13, 2005–2008.[Abstract]

Nygren, K.G. and Andersen, A.N. (2002) Assisted reproductive technology in Europe, 1999. Results generated from European registers by ESHRE. Hum. Reprod., 17, 3260–3274.[Abstract/Free Full Text]

Olivennes, F., Fanchin, R., Kadhel, P., Fernandez, H., Rufat, P. and Frydman, R. (1996) Perinatal outcome of twin pregnancies obtained after in vitro fertilization: comparison with twin pregnancies obtained spontaneously or after ovarian stimulation. Fertil. Steril., 66, 105–109.[ISI][Medline]

Petterson, B., Nelson, K.B., Watson, L. and Stanley, L. (1993) Twins, triplets and cerebral palsy in births in Western Australia in 1980s. Br. Med. J., 307, 1239–1243.[ISI][Medline]

Saunders, K., Spensley, J., Munro, J. and Halasz, G. (1996) Growth and physical outcome of children conceived by in vitro fertilization. Pediatrics, 97, 688–692.[Abstract]

Strömberg, B., Dahlquist, G., Ericson, A., Finnström, O., Köster, M. and Stjernquist, K. (2002) Neurological sequelae in children born after in-vitro fertilisation: a population based study. Lancet, 359, 461–465.[CrossRef][ISI][Medline]

Westergaard, H.B., Johansen, A.M.T., Erb, K. and Andersen, A.N. (1999) Danish National In-Vitro Fertilization Registry 1994 and 1995: a controlled study of birth, malformations and cytogenetic findings. Hum. Reprod., 14, 1896–1902.[Abstract/Free Full Text]

Wølner-Hanssen, P. and Rydhstroem, H. (1998) Cost-effectiveness analysis of in-vitro fertilization: estimated costs per successful pregnancy after transfer of one or two embryos. Hum. Reprod., 13, 88–94.[Abstract]

Submitted on December 4, 2002; accepted on February 28, 2003.