Multiplicity and early gestational age contribute to an increased risk of cerebral palsy from assisted conception: a population-based cohort study

Transcription

1 Human Reproduction, Vol.25, No.8 pp , 2010 Advanced Access publication on June 16, 2010 doi: /humrep/deq070 ORIGINAL ARTICLE Reproductive epidemiology Multiplicity and early gestational age contribute to an increased risk of cerebral palsy from assisted conception: a population-based cohort study D. Hvidtjørn 1,*,J.Grove 1, D. Schendel 2, C. Sværke 1, L.A. Schieve 2, P. Uldall 3,4, E. Ernst 5, B. Jacobsson 6, and P. Thorsen 1 1 Institute of Public Health, Department of Epidemiology, University of Aarhus, 8000 Århus, Denmark 2 Centers for Disease Control and Prevention, National Center on Birth Defects and Developmental Disabilities, Atlanta, GA, USA 3 National Institute of Public Health, Copenhagen, Denmark 4 Pediatric Clinic, Rigshospitalet, University Hospital of Copenhagen, Denmark 5 Reproductive Laboratory, Skejby University Hospital, 8200 Århus, Denmark 6 Perinatal Center, Department of Obstetrics and Gynecology, Institute for the Health of Women and Children, The Sahlgrenska Academy at Göteborg University, Göteborg, Sweden *Correspondence address. dh@soci.au.dk Submitted on January 5, 2010; resubmitted on February 16, 2010; accepted on February 18, 2010 background: This paper assesses the risk of cerebral palsy (CP) in children born after assisted conception compared with children born after natural conception (NC). methods: This population based follow-up study included all 588,967 children born in Denmark from 1995 to Assisted conception was defined as IVF, with or without ICSI, and ovulation induction (OI), with or without subsequent insemination. results: There were (5.6%) children born in Denmark from 1995 to 2003 as a result of assisted conception and through to June 2009, 1146 (0.19%) children received a CP diagnosis. Children born after assisted conception had an increased risk of a CP diagnosis, crude hazard rate ratio (HRR) 1.90 (95% CI: ) compared with NC children. Divided into IVF and OI children compared with NC children, the risk was HRR 2.34 (95% CI: ) and HRR 1.55 (95% CI: ), respectively. When we included the intermediate factors multiplicity and gestational age in multivariate models, the risk of CP in assisted conception disappeared. In general, children with CP born after assisted conception had similar CP subtypes and co-morbidities as children with CP born after NC. conclusion: The risk of CP is increased after both IVF and OI. The increased risk of CP in children born after assisted conception, and in particular IVF, is strongly associated with the high proportion of multiplicity and preterm delivery in these pregnancies. A more widespread use of single embryo transfer warrants consideration to enhance the long-term health of children born after IVF. Key words: assisted reproduction / cerebral palsy / multiple births / preterm births Introduction As use of assisted conception has continued to rise in the past two decades (ASRM, 2004; Nyboe et al., 2009), a growing body of evidence has accumulated demonstrating the risks after assisted conception of adverse perinatal sequela including multiple birth, preterm delivery (PTD) and low birthweight (Olivennes et al., 1993; Wang et al., 2002; Gaudoin et al., 2003; Helmerhorst et al., 2004; Kallen et al., 2004; Ombelet et al., 2006; Nyboe et al., 2009). A number of Scandinavian studies have assessed the risk of cerebral palsy (CP) in children born after in vitro fertilization (IVF) compared with children born after natural conception (NC) (Ericson et al., 2002; Stromberg et al., 2002; Pinborg et al., 2003; Pinborg et al., 2004; Kallen et al., 2005; Lidegaard et al., 2005; Hvidtjørn et al., 2006; Klemetti et al., 2006). These population-based studies show a statistically significant increased risk of CP in IVF children considering singletons and multiples combined, and a recent meta-analysis including IVF children showed a summary odds ratio (OR) of 2.18 & The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 2116 Hvidtjørn et al. (95% confidence interval (CI): ) (Hvidtjørn et al., 2009). However, results from four studies that additionally analyzed the impact of gestational age (GA), suggest that the risk of CP in IVF children is largely attributable to a high proportion of PTD (Stromberg et al., 2002; Kallen et al., 2005; Hvidtjørn et al., 2006; Klemetti et al., 2006). In some previous studies the risk of CP in IVF children was also assessed in strata of multiplicity and a meta-analysis including IVF singletons showed a summary OR of 1.82 (95% CI: ) (Hvidtjørn et al., 2009). Also in singletons, the role of PTD seemed to be a key factor in the causal pathway to CP as showed by decreasing effect estimates when GA was included in the analyses (Stromberg et al., 2002; Hvidtjørn et al., 2006). In contrast, IVF twins and NC twins seemed to have similar risk of CP (Stromberg et al., 2002; Pinborg et al., 2003; Pinborg et al., 2004; Hvidtjørn et al., 2006; Klemetti et al., 2006). Thus, it seems evident that children born after IVF have an increased risk of CP. However, in all but one (Stromberg et al., 2002) of the studies assessing the risk of CP in IVF children, the CP diagnosis was retrieved from hospital discharge registers, but the validity and completeness of the CP diagnosis in these sources have been questioned (Topp et al., 1997). Additionally, we were not able to locate any studies on the risk of CP in children born after ovulation induction (OI). OI also results in PTD and a large proportion of multiples, although to a lesser extent than IVF (Olivennes et al., 1993; Wang et al., 2002; Gaudoin et al., 2003; Kallen et al., 2004; Ombelet et al., 2006). Children diagnosed with CP can be impaired to various degrees, from only minor motor dysfunction to severe and comprehensive impairment including mental retardation and other co-morbidities. To our knowledge, no studies have described the impairment associated with CP in children born after assisted conception. The purpose of this population-based cohort study was to assess the risk of CP in children born after a broad spectrum of assisted conception techniques and further to assess the differences in prevalence of CP subtype, mobility and presence of co-morbidity in children with CP born after NC or assisted conception. We identified children with a CP diagnosis from the National Register of Cerebral Palsy (NCPR) in Denmark, comprising detailed information on children with a CP diagnosis confirmed by a child neurologist (Uldall et al., 2001). In this study, the term-assisted conception refers to IVF, with or without intracytoplasmic sperm injection (ICSI), or OI, with or without insemination. Materials and Methods This cohort study was based on data from Danish national registers and linkage between the registers was achieved via the unique civil registration number given to all citizens in Denmark. All children were identified through The Danish Medical Birth Register (MBR), which contains information on all births in Denmark (Knudsen and Olsen, 1998). Our study comprised all children live born from 1 January 1995 to 31 December Children exposed to IVF were identified through the IVF Register holding data from all private and public fertility clinics including underlying causes of infertility (ovulation factor, male factor or tubal factor) and type of fertility treatment (conventional IVF or ICSI, the number of embryos transferred, and whether the embryos transferred were fresh or frozen and thawed). Children exposed to OI were identified through the Danish Drug Prescription Register (DDPR) holding information on all prescription drugs sold at pharmacies in Denmark. The medications used during OI are prescription drugs bought at the pharmacy enabling identification of the women who were prescribed OI treatment in the DDPR. Drugs used in assisted conception were identified by the authors through cross-checking with: the official Danish Pharmaceutical Classified Catalogue the US website and Books of Instruction of Danish fertility clinics from the time period in question and evaluated by clinical experts. To identify women who had OI in relation to the index pregnancy, we set a time window for the date of dispatch of 12 weeks before and 4 weeks after the last menstrual period (LMP). As the drugs used in OI are also used in IVF, we excluded women from the OI group whose pregnancy with the index child was included in the IVF register. Children diagnosed with cerebral palsy were identified through the NCPR. A Cerebral Palsy Register has existed in eastern Denmark since 1967 (Uldall et al., 2001), but was recently enlarged to national coverage. The NCPR holds detailed information on CP diagnosis validated by child neurologists, CP subtypes and levels of impairments on children in all of Denmark with a CP diagnosis from birth year 1995 through to Additional covariate information was obtained from the MBR and Statistics Denmark (SD): gestational age, multiplicity, sex, maternal age, educational level, smoking during pregnancy and parity. As there are no mothers less than 20 years of age in the IVF Register, children (1.7%) born from mothers of less than 20 years were excluded. Maternal age was dichotomized into years of age (referent category (ref.)) and 35 years or older; educational level was categorized as basic school (9 10 years of education), intermediate length education (11 16 years of education, ref.) and university education (17 or more years of education); parity was dichotomized into primipara and multipara (ref.); multiplicity was divided into singletons (ref.) and twins or more; gestational age was grouped into 20 27, 28 31, 32 36, weeks (ref.) and 42+ weeks. For some analyses, GA was dichotomized into preterm (before 37 weeks of gestation) and term delivery (ref.). The IVF legislation in Denmark was changed in 1997 recommending that only two fresh or three frozen and thawed embryos were transferred per cycle as a rule (Lov nr. 460, 1997) and consequently the number of gestations higher than twins actually decreased throughout the study period. Cox regression calculating hazard rate ratios (HRR) with 95% CI was used to measure the risk of a CP diagnosis in children born after assisted conception combined and in IVF and OI separately relative to the risk in NC children. The risk was measured by the incidence rate, using the child as the unit of analysis. The incidence rate was defined as the number of children receiving a CP diagnosis in a given time frame (1 January June 2009) divided by the total number of person-years the children in the cohort were at risk for the diagnosis during that time. Follow-up started at the time of birth and ended when the child died, emigrated, received a CP diagnosis or the end of follow-up, whichever came first. It was verified that the proportional hazards assumption was met. To account for correlations between siblings in the cohort, additional analyses were included employing robust standard errors. To ensure that the findings were not influenced by a difference in survival rate among children born after natural or assisted conception, we performed additional analyses excluding children who died before 1 year of age. The basic multivariate model included maternal age, parity, educational level, smoking during pregnancy and sex as covariates. Gestational age and multiplicity were then added to the basic model, separately and together, to evaluate the effect of these factors on the association between assisted conception and CP. In secondary analyses within the group of children with a CP diagnosis, we performed bivariate analyses of characteristics of children with CP,

3 Risk of cerebral palsy after assisted conception 2117 comparing the NC group to assisted conception combined and to IVF and OI separately. We considered crucial factors for mother and child (multiplicity and PTD) from the MBR, and from the NCPR, we used information about complications during pregnancy (pre-eclampsia, threatening abortion and urinary tract infection), complications during delivery (fever) or post-partum (admission to neonatal ward, hyperbilirubinemia, surfactant use, respirator use, anemia, adrenocortical hormones, sepsis, cerebral affects (irritation, depression), convulsions), cerebral palsy characteristics as subtype (unilateral, bilateral, ataxic, hyperkinetic and dystonic), mobility (walk alone, walk with device and unable to walk) and co-morbidity (mental retardation based on 3 levels of developmental quotient, eye function and epilepsy). As many of these factors are linked to PTD, we repeated the analyses in strata of preterm and term children. In supplemental analyses, we further subdivided IVF births, to assess the risk of CP based on several IVF-related factors: 1) whether the number of embryos transferred was higher than the number of children born ( vanishing embryos ) (showing proportions with exact CI) 2) underlying cause of infertility (showing HRR) and 3) type of embryo transferred (fresh, frozen and thawed or from donated oocytes) (showing HRR). This information was only available for IVF children. As the risk of CP in twins is associated with monozygocity (Pharoah et al., 2006), we also assessed the risk of CP in IVF pregnancies where the number of embryos transferred was smaller than the number of children born. We compared risk estimates from the present study with validated CP diagnoses to risk estimates from an earlier study using CP diagnoses from hospital discharge registers (Hvidtjørn et al., 2006) to assess the effects on risk estimates of non-validated CP diagnoses. Stata 8.2 was used for the analyses, StataCorp Stata Statistical Software: release 8. College Station, TX: statacorp LP. Approval for this study was obtained from the Danish Data Protection Agency. Results From January to December , children were born alive in Denmark from mothers of 20 years or more. In this period (5.6%) children resulted from assisted conception; (2.5%) from IVF and (3.1%) from OI. Compared with NC mothers, mothers having assisted conception were statistically significantly older, more often primipara and non-smokers, Table I Characteristics of women who used assisted conception in Denmark ; natural conception compared with the combined assisted conception and IVF compared with OI. Natural conception Assisted conception P IVF OI P... Maternal age,0.001, (86.2) (71.8) 9819 (65.5) (76.9) (13.8) 9356 (28.2) 5172 (34.5) 4184 (23.1) Parity,0.001,0.001 Primipara (41.0) (66.9) (74.0) (61.0) Multipara (59.0) (33.1) 3892 (26.0) 7076 (39.0) Educational level (years), Lower (9 10) (23.1) 4641 (16.7) 2175 (17.5) 2466 (16.1) Intermediate (11 16) (68.7) (72.7) 8975 (72.3) (73.1) Higher (17 + ) (8.2) 2930 (10.6) 1261 (10.2) 1669 (10.8) Smoking,0.001,0.001 Non-smoker (77.7) (83.2) (82.4) (84.8) Smoker or stopped in pregnancy (22.3) 5386 (16.8) 2631 (17.6) 2755 (15.2) Multiplicity,0.001,0.001 Singletons (97.6) (70.9) 8501 (56.8) (82.6) Twins (2.4) 9120 (27.5) 6234 (41.6) 2886 (15.9) Triplets 215 (0.0) 495 (1.5) 239 (1.6) 256 (1.4) Quadruplets 4 (0.0) 17 (0.1) 5 (0.0) 12 (0.1) Gestational age (weeks),0.001, (0.2) 397 (1.2) 278 (1.9) 119 (0.7) (0.6) 868 (2.6) 529 (3.5) 339 (1.9) (4.8) 5177 (15.7) 3108 (20.9) 2069 (11.4) (85.9) (75.3) (71.2) (78.6) (8.5) 1728 (5.2) 377 (2.5) 1351 (7.4) Birthweight (grams),0.001, (0.7) 1116 (3.5) 672 (4.6) 444 (2.5) (3.6) 4680 (15.0) 2864 (19.4) 1816 (10.1) (95.7) (81.5) (76.0) (87.4)

4 2118 Hvidtjørn et al. more highly educated and morely likely to deliver multiples, or children with low BW and PTD (Table I). Furthermore, mothers having IVF were statistically significantly different from mothers having OI in all of these characteristics (Table I). Through June 2009, 1146 (0.19%) children from the cohort received a CP diagnosis validated by the NCPR. The follow-up time was from 5 to 13 years. The children with CP differed with statistical significance from other cohort children in the following characteristics; they were more often multiples and premature and their mothers were more often primipara, smokers and of lower educational level (Table II). For children born after assisted conception, the risk of getting a CP diagnosis was increased, HRR 1.90 (95% CI: ) compared with NC children. Divided into IVF and OI children compared with NC children, the risk was a HRR of 2.34 (95% CI: ) and a HRR of 1.55 (95% CI: ), respectively (Table III). Excluding children who died before 1 year of age did not change the results. In the multivariable analyses including sex, maternal age, smoking, parity and educational level (basic analysis), the risk of CP remained statistically significant in assisted conception compared with NC, HRR 1.72 (95% CI: ), in IVF separately, HRR 2.00 (95% CI: ) and in OI separately, HRR 1.47 (95% CI: ) (Table III). When we included the intermediate factors multiplicity and gestational age, first one by one and then combined in the same model, the risk of CP in assisted conception disappeared (Table III). Furthermore, in the analysis including both multiplicity and gestational age, the risk of CP from multiplicity also disappeared. Finally, we made the crude and the basic analyses in strata of multiplicity and found no statistically significant increased risk of CP in either singletons or multiples, although the estimates tended to be higher for singletons than for twins (Table III, bottom). In the bivariate models (Table IV) for characteristics of children with CP diagnoses, comparing the NC group to the assisted conception group and to the IVF or OI group separately, we found that fewer IVF mothers with a CP child had a severe disease. The proportion of multiple births was higher in children with CP born after assisted conception and these children were more often born preterm than NC children with CP. However, in strata of multiplicity, there was no difference in PTD rates. Threatening abortion was more often seen in IVF children with CP than in NC children with CP. A number of short-term post-partum complications (use of surfactant or respirator, anemia with blood transfusion, sepsis and convulsions) in children with CP were statistically significantly associated with IVF. Admission to neonatal ward and hyperbilirubinemia were associated with both IVF and OI, while use of adrenocortical hormones was higher in OI children with CP than in NC children with CP. There was no difference in subtypes of CP and level of motor function between CP children born after NC or assisted conception. There was a decreased prevalence of epilepsy in IVF children with CP. There was an association between mental retardation and severe eye dysfunction in OI children with CP. When repeating the analyses in Table IV in strata of gestational age, no associations were statistically significant in the term stratum, but in the preterm stratum fewer IVF mothers with a CP child had a severe disease, fewer assisted conception mothers had pre-eclampsia, more IVF children with CP used surfactant or were on a respirator and more OI children had mental retardation or severe eye dysfunction (data not shown). Table II Characteristics of children born in Denmark who recieved a CP diagnosis. No CP, N (%) CP, N (%) P... Study population Maternal age (85.4) 958 (83.6) (14.6) 188 (16.4) Parity,0.001 Primipara (42.5) 578 (50.4) Multipara (57.5) 568 (49.6) Educational level, years (22.7) 284 (27.0) years (68.9) 712 (67.7) 17+ years (8.4) 56 (5.3) Smoking,0.001 Non-smoker (78.0) 841 (73.6) Smoker or stopped (22.0) 301 (26.4) Multiplicity,0.001 Singletons (96.1) 1007 (88.0) Twins (3.8) 123 (10.8) Triplets or more 717 (0.1) 14 (1.2) Gestational age (weeks), (0.3) 43 (3.8) (0.7) 171 (15.0) (5.3) 192 (16.9) (85.4) 654 (57.5) (8.3) 78 (6.8) Birthweight (grams), (0.8) 173 (15.5) (4.2) 237 (21.3) (95.0) 705 (63.2) Within the IVF children we assessed the risks of CP within strata of underlying causes of infertility. Since many women have more than one underlying cause or unspecific causes, we only used data from women with one single underlying cause which were available for 9855 IVF children (65.6% of the IVF children). Compared with NC children we found the following risk estimates; ovulation factor HRR 1.79 (95% CI: ), male factor HRR 2.32 (95% CI: ) and tubal factor HRR 2.49 (95% CI: ). When we added multiplicity or gestational age to the model, the statistically significant risks by underlying cause of infertility disappeared. Compared with NC children we found comparable results when we separately evaluated IVF using freshly fertilized embryos, HRR 2.27 (95% ) and IVF using previously frozen, thawed embryos (HRR 3.66 (95%CI: ). Sample size was insufficient to assess IVF using donor oocytes; however, our estimate was in the expected direction for this select group of children, HRR 3.40 (95% CI: ) compared with NC children. Within the IVF children only, we calculated the distribution of CP in pregnancies where the number of embryos transferred was higher than

5 Risk of cerebral palsy after assisted conception 2119 Table III Risk of CP in children born after assisted conception and IVF and OI separately compared with NC children, hazard risk ratios (HRR) and 95% CI. Assisted conception IVF OI... Crude 1.90 ( ) 2.34 ( ) 1.55 ( ) Basic* 1.72 ( ) 2.00 ( ) 1.47 ( ) Basic* and multiplicity 1.12 ( ) 1.07 ( ) 1.13 ( ) Basic* and GA 0.96 ( ) 0.90 ( ) 1.01 ( ) Basic* and multiplicity and GA 0.96 ( ) 0.91 ( ) 1.01 ( ) Twins and more** 0.96 ( ) 0.99 ( ) 1.00 ( ) GA weeks 20 27*** ( ) ( ) ( ) GA weeks 28 31*** ( ) ( ) ( ) GA weeks 32 36*** 4.44 ( ) 4.39 ( ) 4.47 ( ) GA weeks reference reference reference GA weeks 42+*** 1.58 ( ) 1.19 ( ) 1.17 ( ) In strata of multiplicity Singletons, crude 1.31 ( ) 1.44 ( ) 1.24 ( ) Twins and more, crude 1.19 ( ) 1.22 ( ) 1.13 ( ) Singletons, basic* 1.21 ( ) 1.21 ( ) 1.21 ( ) Twins and more, basic* 1.04 ( ) 1.07 ( ) 1.00 ( ) *Multiple analyses including sex, maternal age, education, smoking and parity. **Twins and more compared with singletons and controlled for assisted conception, sex, maternal age, education, smoking, GA and parity. ***Strata of GA compared with term born children (GA 37 41) and controlled for assisted conception, sex, maternal age, education, smoking, multiplicity and parity. the number of children born ( vanishing embryos ) compared with pregnancies where the number of embryos transferred was the same as the number of children born. Due to the 1997 recommendation that only two fresh or three frozen and thawed embryos be transferred per cycle, it was not surprising that the number and proportion of singletons from pregnancies in which.1 embryo had been initially transferred (87.7%) was remarkably higher than the number and proportion of twins from pregnancies in which.2 embryos had been transferred. (16.2%) (Table V). Thus, it was not feasible to reliably estimate the risk for CP in multiple gestations with vanishing embryos because of the relatively low number of exposed pregnancies. There were 0 cases of CP in singletons conceived from single embryo transfer (SET) (also in a relatively small number of exposed pregnancies), in contrast to 21 CP cases (0.28%) in singletons conceived after transfer of two or more embryos, but the confidence intervals for these two proportions overlapped; a larger sample of SET pregnancies are needed to assess the trend for an increased CP risk more reliably. To assess the risk associated with monozygocity, we identified 38 twins resulting from transfer of only one embryo and none of these had a CP diagnosis. We further identified 71 triplets where only two embryos had been transferred and 1 of them had a CP diagnosis. We refrained from further analyses due to small numbers. In the current study with CP diagnoses obtained from the NCPR validated by child neurologists, we observed similar results to earlier studies using CP diagnoses from hospital discharge registers. However, compared with our previous study from 2006 (Hvidtjørn et al., 2006) with children born and 1008 children diagnosed with CP in the hospital discharge registers, we find higher risk estimates on the risk of CP in IVF children. For better comparison, we made separate analyses of the current study cohort among children born in that time period ( ): 397,511 children (fewer children than in the 2006 study because mothers,20 years of age were excluded in the current study) with 872 children diagnosed with CP in the NCPR. In the combined group (singletons and twins), we here find a crude HRR of 2.33 (95% CI: ) in contrast to a crude HRR of 1.61 (95% CI: ) in the former study. For singletons the crude HRR is 1.41 (95% CI: ) in the present study in contrast to a crude HRR of 1.28 (95% CI: ) in the former. Discussion In a large population-based follow-up study with validated CP diagnoses, we were able to confirm previous findings of an increased risk of CP in children born after IVF. Furthermore, we also found an increased risk of CP in children born after OI in the crude analyses and when we adjusted for sex, maternal age, educational level, smoking and parity. The increased risk for CP in children born after assisted conception was largely associated with multiplicity and PTD as seen in Table III; nevertheless, the proportion of multiples in assisted conception played a large role in the overall risk as 54.5% of CP children resulting from assisted conception were multiples compared with only 7.2% of the NC children with CP (Table IV). From a public health perspective this is of great importance because the number of multiple births in IVF can clearly be reduced to a large extent by use of SET. We found no statistically significantly increased risk of CP in IVF singletons, HRR 1.45 (95% CI: ). Previous studies report inconsistent findings; some find an association between CP and IVF

6 2120 Hvidtjørn et al. Table IV Characteristics of children born in Denmark and diagnosed with CP, children born after NC compared with children born after assisted conception, IVF and OI, respectively. Natural conception Assisted conception P IVF P OI P... Basic factors, mother No severe maternal disease 880 (88.6) 104 (93.7) (96.8) (89.8) Severe maternal disease 113 (11.4) 7 (6.3) 2 (3.2) 5 (10.2) Basic factors, child GA weeks (2.9) 13 (11.2), (12.7), (9.4) GA weeks (14.0) 28 (24.1) 22 (34.9) 6 (11.3) GA weeks (15.7) 32 (27.6) 18 (28.6) 14 (26.4) GA weeks (59.9) 42 (36.2) 15 (23.8) 27 (50.9) GA weeks (7.5) 1 (0.9) 0 (0.0) 1 (1.9) Singletons 954 (92.8) 53 (45.7), (33.3), (60.4),0.001 Twins 70 (6.8) 53 (45.7) 40 (63.5) 13 (24.5) Triplets 4 (0.4) 10 (8.6) 2 (3.2) 8 (15.1) Singletons born at term 675 (71.3) 35 (66.0) (57.1) (71.9) Singletons born preterm (,37 GA weeks) 272 (28.7) 18 (34.0) 9 (42.9) 9 (28.1) Multiples born at term 14 (19.2) 8 (12.7) (7.1) (23.8) Multiples born preterm (,37 GA weeks) 59 (80.8) 55 (87.3) 39 (92.9) 16 (76.2) Pregnancy No pre-eclampsia 903 (91.1) 100 (92.6) (96.8) (87.0) Pre-eclampsia 88 (8.9) 8 (7.4) 2 (3.2) 6 (13.0) No threatening abortion 878 (91.8) 89 (84.0) (80.7) (88.6) Threatening abortion 78 (8.2) 17 (16.4) 12 (19.3) 5 (11.4) No urinary tract infection 736 (95.3) 86 (95.6) (100.0) (90.0) Urinary tract infection 36 (4.7) 4 (4.4) 0.(0.9) 4 (10.0) Delivery No fever 921 (95.4) 104 (95.4) (93.4) (97.9) Fever (4.6) 5 (4.6) 4 (6.6) 1 (2.1) Post-partum No admission to neonatal ward 317 (30.8) 15 (12.9), (11.1) (15.1) Admission to neonatal ward 711 (69.2) 101 (87.1) 56 (88.9) 45 (84.9) No hyperbilibinarumi 728 (72.3) 52 (46.9), (40.0), (54.9) Hyperbilibinarumi 279 (27.7) 59 (53.1) 36 (60.0) 23 (45.1) No surfaktant 931 (91.4) 84 (75.7), (68.9), (84.0) Surfaktant 88 (8.6) 27 (24.3) 19 (31.1) 8 (16.0) No respirator 875 (86.2) 84 (75.7) (71.2) (80.8) Respirator 140 (13.8) 27 (24.3) 17 (28.8) 10 (19.2) No anaemia 928 (91.3) 93 (82.3) (80.3) (84.6) Anaemia with blood transfusion 88 (8.7) 20 (17.7) 12 (19.7) 8 (15.4) No adrenocortical hormone 948 (93.5) 99 (88.4) (90.3) (86.0) Adrenocortical hormone 66 (6.5) 13 (11.6) 6 (9.7) 7 (14.0) No sepsis 638 (62.8) 58 (51.3) (43.6) (60.8) Sepsis 378 (37.2) 55 (48.7) 35 (56.4) 20 (39.2) No cerebral affects 669 (66.0) 77 (67.5) (77.4) (55.8) Cerebral affects 344 (34.0) 37 (32.5) 14 (22.6) 23 (44.2) No convulsions 759 (74.6) 93 (81.6) (88.7) (73.1) Convulsions 258 (25.4) 21 (18.4) 7 (11.3) 14 (26.9) Continued

7 Risk of cerebral palsy after assisted conception 2121 Table IV Continued Natural conception Assisted conception P IVF P OI P... Cerebral palsy characteristics subtype Unilateral 411 (39.9) 36 (31.0) (33.3) (28.3) Bilateral 489 (47.5) 59 (50.9) 31 (49.2) 28 (52.8) Ataxic 35 (3.4) 5 (4.3) 2 (3.2) 3 (5.7) Hyperkine 15 (1.5) 0 (0.0) 0 (0.0) 0 (0.0) Dyston 67 (6.5) 13 (11.2) 7 (11.1) 6 (11.3) Not classified 12 (1.2) 3 (2.6) 2 (3.2) 1 (1.9) Level of motor function Walk alone 539 (52.6) 55 (48.7) (53.2) (43.1) Use device 246 (24.0) 28 (24.8) 16 (25.8) 12 (25.5) Immobile 239 (23.4) 30 (26.5) 13 (21.0) 17 (33.4) Co-morbidity mental retardation Normal 458 (45.6) 46 (41.8) (57.6) (23.5) DQ (28.0) 31 (28.2) 14 (23.7) 17 (33.3) DQ, (26.3) 33 (30.0) 11 (18.6) 22 (43.2) Eye function Normal 846 (84.7) 90 (81.8) (91.7) (70.0) Severe eye dysfunction 153 (15.3) 20 (18.2) 5 (8.3) 15 (30.0) Epilepsia No 721 (70.3) 84 (73.7) (85.5) (59.6) Yes 305 (29.7) 30 (26.3) 9 (14.5) 21 (40.4) singletons (Stromberg et al., 2002; Lidegaard et al., 2005) and others do not (Hvidtjørn et al., 2006; Klemetti et al., 2006). A meta-analysis including IVF singletons from studies without overlapping cohorts (Stromberg et al., 2002; Lidegaard et al., 2005; Klemetti et al., 2006) find an increased risk of CP in IVF singletons with an OR of 1.82 (95% CI: ). Thus, our non-significant results based on analysis of 8501 IVF singletons of whom 21 (0.25%) received a CP diagnosis might be due to insufficient sample size in the singleton stratum. The proportion of PTD was much higher in CP children born after assisted conception, 73 (62.9%) compared with CP children born after NC, 333 (32.6%), also resulting in more post-partum complications linked to PTD (Table IV). Generally, this did not give rise to more co-morbidities or more severe subtypes among CP children born after assisted conception. This is in accordance with studies showing lower rates of disability and better survival in preterm children with CP compared with term born children with CP (Hutton, 2006). However, OI children with CP had a higher prevalence of mental retardation and severe eye dysfunction, also in the stratum of preterm OI children and these findings need further investigation. There was no increased risk of mental retardation in children with CP born after IVF, likely due to the large proportion of PTD in these children as premature children with bilateral CP have higher IQ than mature born children with bilateral CP, perhaps reflecting the brains ability to compensate better when it is less well developed (Hemming et al., 2008). Based on CP diagnoses validated by child neurologists, we confirmed previous findings, although risk estimates were somewhat higher for CP in IVF children compared with studies using CP diagnoses from hospital discharge registers. Accordingly, this suggests that possible non-systematic misclassification of CP diagnoses in the Danish hospital discharge register contributed to the lower risk estimates reported from the previous studies. Therefore, it is important to validate diagnoses in complex diseases, when these diagnoses are retrieved from national registries collected for administrative purposes, in order to avoid possible attenuation of exposure risk estimates due to non-systematic misclassification of disease outcome. Table V Distribution of children with CP born after NC or IVF (with or without vanishing embryos) with exact 95% CI. Total CP CP percent numbers with 95% CI... Singletons NC ( ) Singletons IVF, no ( ) vanishing embryos Singletons IVF, vanishing ( ) embryos Twins NC ( ) Twins IVF, no vanishing ( ) embryos Twins IVF, vanishing ( ) embryos

8 2122 Hvidtjørn et al. In contrast with our previously reported findings from a Danish cohort study ( ) ((Hvidtjørn et al., 2005), in the current study we did not find a marked increased risk of CP in IVF children born after vanishing embryos. One possible explanation might be the trend in our study cohort for decreasing the number of embryos transferred following Danish IVF legislation in 1997 restricting the number of embryos that could be transferred (National Board of Health, 2007), as the present study additionally included children born Thus, in the current study cohort, transfer of more than two embryos was less frequent than in the previous study so singletons were more likely to have had a vanishing embryo than were twins; on the other hand SET was also relatively uncommon. What we observed, however, was that in the group of 1042 IVF singletons born after SET, no children received a CP diagnosis compared with 21 (0.28%) of the IVF singletons born after two or more embryo transfers and 954 (0.18%) of the NC singletons (Table V); a larger sample of SET pregnancies is needed to assess this trend for an increased CP risk more reliably. Further, there was a significantly higher risk of CP in twins than in those singletons that had a vanished embryo, probably due to the increased risk for CP in twins per se. The possible association between vanishing embryos and CP clearly needs further study, preferably in studies including zygocity. Our sources of data were Danish national health registers. These registers are maintained for administrative or monitoring purposes and will therefore not always include specific details desirable for research. For an exposure to IVF, we used the IVF Register and according to the Danish legislation, it is mandatory for all public and private fertility clinics in Denmark to report all initiated IVF and ICSI treatments, successful or not, to the IVF Register and the register is therefore considered close to complete (Andersen et al., 1999). For an exposure to OI, we used the DDPR with data on all dispatched prescriptions for hormones used in OI. The main purpose of the DDPR is entirely administrative, that is to ensure that the prescription is not used twice and therefore it is expected that all dispatches are reported to the DDPR. However, it is not possible to establish the date the woman actually administered the medicine. Hormones used in OI can be prescribed for the next 3 cycles and consequently, we had to select a time-window of prescriptions within 3 months before LMP to define the use of OI in the index pregnancy. Therefore the hormones could have been administered 1 or 2 months earlier than the index pregnancy. Additionally, we cannot be sure that the woman ever administered the medicine she purchased and the findings regarding OI have to be interpreted with caution. Unfortunately we lack data about time to pregnancy, a factor that has been associated with PTD (Henriksen et al., 1997; Basso and Baird, 2003). In conclusion, similar to children born after IVF, children born after OI are at increased risk of CP. The increased risk of CP in children born after assisted conception and in particular IVF is strongly associated with the high proportion of multiplicity and PTD in these pregnancies. A more widespread use of SET may warrant recommendation to enhance the long-term health of children born after IVF. Apart from an observed increased risk of mental retardation and severe eye dysfunction in OI children with CP that needs confirmation in further study, children with CP born after assisted conception appear to have a similar diagnostic and co-morbidity profile as children with CP born after NC. Funding The study was funded as a co-financed PhD project by The Danish Agency for Science, Technology and Innovation, University of Aarhus and The Elsass Foundation. Further funding was applied by The Health Insurance Foundation, The Augustinus Foundation, Julie von Müllens Foundation, Direktør Jacob Madsen & Hustru Olga Madsens Fond and Aase and Ejnar Danielsen Foundation. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. References Andersen AN, Westergaard HB, Olsen J. The Danish in vitro fertilisation (IVF) register. Dan Med Bull 1999;46: Assisted reproductive technology in the United States: 2000 results generated from the American Society for Reproductive Medicine/ Society for Assisted Reproductive Technology Registry. Fertil Steril 2004;81: Basso O, Baird DD. Infertility and preterm delivery, birthweight, and Caesarean section: a study within the Danish National Birth Cohort. Hum Reprod 2003;18: Ericson A, Nygren KG, Olausson PO, Kallen B. Hospital care utilization of infants born after IVF. Hum Reprod 2002;17: Gaudoin M, Dobbie R, Finlayson A, Chalmers J, Cameron IT, Fleming R. Ovulation induction/intrauterine insemination in infertile couples is associated with low-birth-weight infants. Am J Obstet Gynecol 2003; 188: 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. Hemming K, Colver A, Hutton JL, Kurinczuk JJ, Pharoah PO. The influence of gestational age on severity of impairment in spastic cerebral palsy. J Pediatr 2008;153:203 8, 208. Henriksen TB, Baird DD, Olsen J, Hedegaard M, Secher NJ, Wilcox AJ. Time to pregnancy and preterm delivery. Obstet Gynecol 1997; 89: Hutton JL. Cerebral palsy life expectancy. Clin Perinatol 2006;33: Hvidtjørn D, Grove J, Schendel D, Vaeth M, Ernst E, Nielsen L, Thorsen P. Vanishing embryo syndrome in IVF/ICSI. Hum Reprod 2005;20: Hvidtjørn D, Grove J, Schendel DE, Vaeth M, Ernst E, Nielsen LF, Thorsen P. Cerebral palsy among children born after in vitro fertilization: the role of preterm delivery a population-based, cohort study. Pediatrics 2006;118: Hvidtjørn D, Schieve L, Schendel D, Jacobsson B, Svaerke C, Thorsen P. Cerebral palsy, autism spectrum disorders, and developmental delay in children born after assisted conception: a systematic review and meta-analysis. Arch Pediatr Adolesc Med 2009;163: Kallen B, Olausson PO, Nygren KG. Neonatal outcome in pregnancies from ovarian stimulation. Obstet Gynecol 2002;100: Kallen B, Finnstrom O, Nygren KG, Olausson PO. In vitro fertilization in Sweden: child morbidity including cancer risk. Fertil Steril 2005; 84: Klemetti R, Sevon T, Gissler M, Hemminki E. Health of children born as a result of in vitro fertilization. Pediatrics 2006;118: Knudsen LB, Olsen J. The Danish Medical Birth Registry. Dan Med Bull 1998;45: Lidegaard O, Pinborg A, Andersen AN. Imprinting diseases and IVF: Danish National IVF cohort study. Hum Reprod 2005;20:

9 Risk of cerebral palsy after assisted conception 2123 Lov nr.460. Lov om kunstig befrugtning i forbindelse med lægelig behandling, diagnostik og forskning m.v. af 10.juni National Board of Health. IVF-Behandlinger i Danmark , Nyboe Andersen A, Goossens V, Bhattacharya S, Ferraretti AP, Kupka MS, de Mouzon J, Nygren KG. Assisted reproductive technology and intrauterine inseminations in Europe, 2005: results generated from European registers by ESHRE. Hum Reprod 2009;24: Olivennes F, Rufat P, Andre B, Pourade A, Quiros MC, Frydman R. The increased risk of complication observed in singleton pregnancies resulting from in-vitro fertilization (IVF) does not seem to be related to the IVF method itself. Hum Reprod 1993;8: Ombelet W, Martens G, De SP, Gerris J, Bosmans E, Ruyssinck G, Defoort P, Molenberghs G, Gyselaers W. Perinatal outcome of 12,021 singleton and 3108 twin births after non-ivf-assisted reproduction: a cohort study. Hum Reprod 2006;21: Pharoah PO, Price TS, Plomin R. Cerebral palsy in twins: a national study. Arch Dis Child Fetal Neonatal Ed 2006;87: Pinborg A, Loft A, Schmidt L, Andersen AN. 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. Hum Reprod 2003;18: 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;7:311. Stromberg B, Dahlquist G, Ericson A, Finnstrom O, Koster M, Stjernqvist K. Neurological sequelae in children born after in-vitro fertilisation: a population-based study. Lancet 2002;359: Topp M, Langhoff-Roos J, Uldall P. Validation of a cerebral palsy register. J Clin Epidemiol 1997;50: Uldall P, Michelsen SI, Topp M, Madsen M. The Danish Cerebral Palsy Registry. A registry on a specific impairment. Dan Med Bull 2001;48: Wang JX, Norman RJ, Kristiansson P. The effect of various infertility treatments on the risk of preterm birth. Hum Reprod 2002; 17: