Psychological follow-up study of 5-year-old ICSI children

Transcription

1 Human Reproduction Page 1 of 7 Hum. Reprod. Advance Access published September 9, 2004 doi: /humrep/deh511 Psychological follow-up study of 5-year-old ICSI children I.Ponjaert-Kristoffersen 1,7, T.Tjus 2,4, J.Nekkebroeck 1, J.Squires 3, D.Verté 1, M.Heimann 4, M.Bonduelle 5 and U.-B.Wennerholm 6 on behalf of the Collaborative study of Brussels, Göteborg and New York 1 Department of Developmental- and Lifespanpsychology, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, 5 Centre for Medical Genetics, Dutch-speaking University Hospital of Brussels, Laarbeeklaan, 101, 1090 Brussels, Belgium, 2 Department of Psychology, Göteborg University, 6 Department of Obstetrics and Gynaecology, Institute for the Health of Women and Children Sahlgrenska University Hospital, Göteborg, Sweden, 3 University of Oregon, Eugene, OR and Weill-Medical College, NY, USA and 4 Center for Child and Adolescent Mental Health, University of Bergen, Norway 7 To whom correspondence should be addressed. [email protected] BACKGROUND: The developmental outcomes of children born after ICSI are still a matter of concern. The purpose of the present study was to investigate psychological outcomes for 5-year-old children born after ICSI and compare these with outcomes for children born after spontaneous conception (SC). METHODS: Three hundred singleton children born after ICSI in Belgium, Sweden and the USA were matched by maternal age, child age and gender. Outcome measures included the Wechsler Preschool and Primary scales of intelligence (WPPSI-R), Peabody Developmental Motor Scales, Parenting Stress Index and Child Behaviour Checklist. RESULTS: Regarding cognitive development, no significant differences were found on WPPSI-R verbal and performance scales between ICSI and SC children. However, some differences were noted on subtests of the Performance Scale. ICSI children more often obtained a score below 1 SD of the mean on the subtests: Object Assembly, Block Design and Mazes (all P<0.05). Significant differences by site (i.e. Belgium, Sweden and New York) were found on subtests related to parenting stress, child behaviour problems and motor development (all P<0.05). These findings can probably be explained by variables other than conception mode, such as cultural differences and selection bias. CONCLUSIONS: Although the finding that a higher proportion of ICSI children obtained scores below the cut-off on some of the visual spatial subscales of the WPPSI-R warrants further investigation, ICSI does not appear to affect the psychological well-being or cognitive development at age 5. Key words: cognitive development/emotional behavioural development/icsi/motor development/parental stress Introduction The birth of Louise Brown 25 years ago ushered in a unique population of children, the product of intensive human intervention and reproductive manipulation. As assisted reproductive technology (ART) has evolved, this manipulation and intervention has become more technical and sophisticated. ICSI, introduced a decade ago as a solution for male infertility, involves human manipulation of oocyte and sperm one spermatozoon is injected through the oocyte membrane and bypasses natural sperm selection barriers. An ongoing discussion about the potential genetic and developmental risks to ICSI offspring has resulted, due to this high degree of human intervention (Bowen et al., 1998, Sutcliffe et al., 2001; Bonduelle et al., 2003; Leslie et al., 2003). To date, a higher risk for fetal chromosomal anomalies has been documented as well as a small but statistically significant increase in congenital anomalies observed compared with spontaneously conceived (SC) children, but not compared with children born after IVF (Wennerholm et al., 2000; Ludwig and Katalinic, 2000; Bonduelle et al., 2002a,b). Furthermore, a study in which SC children (80) and IVF children (84) were compared with 89 singleton ICSI children revealed an increased risk of mild or significant developmental delay (Mental Developmental Index (MDI), 85) at 1 year, which might be predictive of an increased risk of intellectual impairment or learning difficulties. In contrast, when this group of children was reassessed at age 5 years and the sample size increased, no differences were found between ICSI, IVF and SC children in cognitive development (Bowen et al., 1998; Leslie et al., 2003). ICSI and IVF offspring were found to be developing typically on a longitudinal developmental screening questionnaire with no statistically significant differences between groups (Squires et al., 2003). Similar findings were reported when ICSI children and SC children were assessed neurodevelopmentally (Sutcliffe et al., 2003). In addition to developmental outcomes, it is often assumed that the parent child relationship may also be at risk when reproductive technologies such as ICSI are used. Human Reproduction q European Society of Human Reproduction and Embryology 2004; all rights reserved Page 1 of 7

2 I.Ponjaert-Kristoffersen et al. So far, no findings of empirical studies provide evidence for that assumption. However, ART may influence the attitudes and expectations of parents towards their child and consequently the nature of ensuing parent child attachment (Colpin et al., 1995; Golombok, 2002). Cook et al. (1997) suggest that the outcomes of family functioning and child development are to some extent dependent on the social context in which these techniques are carried out. When comparing assisted reproduction families in Eastern versus Western European countries, the findings indicate greater difficulties in parental adjustment and child behaviour in assisted conception families in Eastern Europe than in their Western counterparts. Although in-depth studies of developmental outcomes of children born after ICSI have been conducted in Europe (Wennerholm et al., 2000; Sutcliffe et al., 2001; Bonduelle et al., 2002b, 2003) and Australia (Bowen et al., 1998; Leslie et al., 2003), few have been conducted in the USA (Squires et al., 2003), and no international studies have focused on outcomes of ART children across divergent medical systems. To study the long-term effects of ICSI across diverse medical systems, a collaborative study in three countries was performed to examine psychological outcomes at the age of 5 years for ICSI children compared with SC children. Psychological data including cognitive, motor and emotional behavioural aspects of development are presented. Materials and methods Study participants Singleton children born after ICSI were compared with singleton SC children at 5 years of age (^9 months) in a multicentre controlled study in the USA, Belgium and Sweden. ICSI children were conceived using ejaculated, epididymal or testicular sperm, which was either fresh or cryopreserved. Male factor infertility was defined as, motile sperm cells after preparation. Use of donor sperm was excluded. ICSI children were recruited from the Centre for Reproductive Medicine, Vrije Universiteit Brussel (BRU); from the Fertility Centre Scandinavia, Carlanderska Hospital, and Reproductive Medicine, Sahlgrenska University Hospital of Göteborg (GOT); and from the Weill Medical College, Cornell University, New York (CNY). In BRU, SC children were recruited from schools and examined on site at their schools. In GOT, SC children were recruited from the Swedish Medical Birth Registry and examined at Sahlgrenska University Hospital. In CNY, SC children were recruited by advertisements in hospital and parent magazines in the tri-state area (New York, New Jersey and Connecticut) where ICSI families were living, and were examined at Weill Medical College. Matching for child gender, child age (^3 months per child in GOT and group matching ^ 9 months in BRU and CNY) and maternal age (^3 years in GOT and group matching in BRU and CNY) was done in the three centres. In GOT, child age and maternal age were case control matched from the Swedish Medical Birth Registry. Only in BRU was maternal educational level groupmatched and only primiparous women were included. Education was considered in six different levels depending on the length and type of formal education; adopted from the Hollingshead four factor index (Hollingshead, 1975). Exclusion criteria were multiple birth and birth at,32 weeks of gestational age, and maternal or child language different from the language of the country (Dutch, Swedish and English). Collection of these data was completed through written questionnaires in BRU and CNY and through data collected from the Swedish Medical Birth Registry and the medical records in addition to the questionnaires in GOT. Questions on demographic parameters, maternal health, pregnancy, birth and childhood history up to 5 years were asked of parents. Outcome measures At the age of 5, a psychological battery of cognitive, motor and emotional behavioural assessments was performed on children, and parenting attitudes and stress were measured. Child cognitive development was assessed individually by psychologists using the Wechsler Preschool and Primary Scale-Revised (WPPSI-R) with both verbal scales (subtests Information, Comprehension, Arithmetic, Vocabulary, Similarities and Sentences) and performance scales (subtests Object Assembly, Geometric Design, Block Design, Mazes, Picture Completion and Animal Pegs) administered (Wechsler, 1990, 1999; Vander Steene and Bos, 1997) (Table I and Ia). Motor development was assessed using the Peabody Developmental Motor Scales (PDMS) including a gross (GMQ) and fine motor quotient (FMQ) (Folio and Fewell, 2000) (Table Ib). The PDMS was administered by a psychologist in CNY and BRU and by a paediatrician in GOT. Due to different scoring systems, PDMS data from the GOT population could not be included in the study results (Table I). The Child Behaviour Checklist (CBCL), an assessment that detects both internalizing and externalizing behavioural problems in children (Achenbach and Edelbrock, 1983; Verhulst et al., 1996), was completed. Finally, stress related to parenting experiences was assessed using the Parenting Stress Index (PSI) (subscales: Parental Distress, Parent Child Dysfunctional Interaction and Difficult Child) (Abidin, 1990). Both questionnaires (CBCL and PSI) were filled out by the mothers of ICSI and SC children in all three centres (BRU, GOT and NY) (Table I). Table I. Number of children in each centre and group who were administered each of the assessment instruments BRU GOT CNY ICSI (n ¼ 100) SC (n ¼ 100) ICSI (n ¼ 98) SC (n ¼ 111) ICSI (n ¼ 102) SC (n ¼ 49) Child outcomes WPPSI-R PDMS Questionnaires (mothers) CBCL PSI WPPSI-R ¼ Wechsler Preschool and Primary scales of intelligence; PDMS ¼ Peabody Developmental Motor Scales; CBCL ¼ Child Behaviour Checklist; PSI ¼ Parenting Stress Index. Page 2 of 7

3 ICSI and psychological outcome at 5 years Table Ia. Description of the subscales of the Verbal and Performance scales of the WPPSI-R Verbal scale Information Comprehension Arithmetic Vocabulary Similarities Sentences Performance scale Object Assembly Geometric Design Block Design Mazes Picture Completion Animal Pegs Description Brief oral questions that require the child to demonstrate knowledge about events or objects in the environment. The child is asked to express in words his or her understanding of the reasons for actions and the consequences of events. The child is asked to demonstrate understanding of basic quantitative concepts by completing simple counting tasks and more difficult problems. Picture identification items on which the child is asked to name pictured objects and orally presented words for which verbal definitions have to be provided. The child is asked to demonstrate an understanding of the concept of similarity by explaining how two verbally presented objects or events are alike. A variant of the digit span task where the child has to repeat verbally the sentence read aloud by the examiner. The child is asked to assemblage a puzzle picturing a common object, taking into account the speed of performance. Requires visual recognition and discrimination skills: the child needs to find an identical design from an array of four designs; the child needs to draw a geometric figure from a printed model. The child is asked to analyse and reproduce within a specified time limit, patterns made from flat, two-coloured blocks. The child is asked to solve pencil-and-paper mazes of increasing difficulty under time constraints. The child has to identify what is missing from pictures of common objects or events. The child is invited to place pegs of the correct colours in holes below a series of pictured animals, with accuracy and speed of performance contributing to the score. Table Ib. Description of the subtests of the gross motor quotient (GMQ) and fine motor quotient (FMQ) of the PDMS GMQ Stationary Locomotion Object Manipulation FMQ Grasping Visual Motor Integration Description Ability to sustain control of the body within its centre of gravity and retain equilibrium. Ability to move from one place to another. Actions measured are crawling, walking, running, hopping and jumping forward. Ability to manipulate balls: catching, throwing and kicking. Ability to use hands, beginning with the ability to hold an object with one hand and progresses to actions involving the controlled use of the fingers of both hands. Ability to use visual perceptual skills to perform complex eye hand coordination tasks, such as reaching and grasping for an object, building with blocks and copying designs. Statistics Standardized tests with established psychometric properties including reliability and validity were selected (Achenbach and Edelbrock, 1983; Abidin, 1990; Wechsler, 1990; Folio and Fewell, 2000). Interval or ratio data were analysed using three-way ANOVAs with mode of conception, nationality and gender as independent variables. Test scores on the WPPSI-R, PDMS, CBCL and PSI were used as dependent variables. In order to account for the effects of child medical variables (e.g. birthweight and gestational age) and demographic variables (age of the mother at birth and mothers educational level), additional ANOVAs were carried out with these variables entered as independent variables. A significance level of 5% (two-tailed) was accepted throughout. Results Demographics A total of 300 children born after ICSI (100 children from BRU, 98 children from GOT and 102 children from CNY) and 260 children born after SC (100 children from BRU, 111 children from GOT and 49 children from CNY) were investigated. Dropout rates due to not reached among children were 35 out of 150 (23.3%) in BRU, none out of 112 in GOT and 88 out of 345 (25.5%) in CNY and refusals were 15 out of 150 (10%) in BRU, 14 out of 112 (12.5%) in GOT and 155 out of 345 (44.9%) in CNY. Among SC children, refusals were 46 out of 146 (31.5%) in BRU, 32 out of 143 (22.3%) in GOT and not possible to calculate due to the methodology used in CNY. All children from BRU and GOT were of Caucasian origin; in CNY, four ICSI and seven controls were non-caucasian. The gender distribution was 146 males (48.7%) and 154 females in the ICSI group and 130 males (48.9%) and 136 females in the control group. The indication for ICSI was male factor infertility in 84% and non-male factor infertility (i.e. failed IVF) in 16% of the population. Further maternal, paternal and neonatal characteristics are given in Table II. The ages of the mother and the father were significantly higher in ICSI parents, while the educational level was similar for ICSI and SC in both mothers and fathers. The rate of preterm birth as well as the rate of low birthweight was significantly higher for the ICSI group (Table II). Demographic variables and child outcome Educational level of the mother and child outcome Educational level was defined in three groups: high, medium and low. Mothers with a high educational level were those who had obtained a higher degree (n ¼ 49) or a degree (n ¼ 188). Mothers with a medium educational level had entered university (n ¼ 135) or fully passed school matriculation (n ¼ 96). Mothers with a low educational level had partially passed school matriculation (n ¼ 58) or had no qualification at all (n ¼ 11). There was an overall effect showing that, the higher the mother s educational level the higher the child s, verbal IQ (VIQ) [F(2,523) ¼ 15.17; P, 0.05 [and full-scale IQ (FSIQ) [F(2,523) ¼ 15.52; P, 0.05]. Higher scores on performance IQ (PIQ) and GMQ were only obtained if the mother had a high educational level Page 3 of 7

4 I.Ponjaert-Kristoffersen et al. Table II. Demographic and neonatal characteristics for the ICSI (n ¼ 300) and spontaneously conceived (SC; n ¼ 266) groups ICSI SC Statistical analysis b Mothers Age (median years; range) 33.0, , P, VE adj. Educational level a (median) (BRU) 3.0, , 1 6 P ¼ VE adj. Level 1 (%) Level 2 (%) Level 3 (%) Level 4 (%) Level 5 (%) Level 6 (%) Smoking during pregnancy (%) P ¼ CMH adj. Alcohol during pregnancy (%) P ¼ CMH adj. Parity (% primiparity) P, CMH adj. Chronic maternal disease (n, %) 10, 3.3 6, 2.3 P ¼ FE Pregnancy complications (n, %) 40, , 4.5 P, CMH adj. Fathers Age (median years; range) 35.2, , P, VE adj. Educational level (median) (BRU) 2.0, , 1 6 P ¼ VE adj. Neonates Gestational age (median weeks; range) 39.3, , P ¼ VE adj. Preterm birth (,37 weeks) (n, %) 33, , 4.5 P, 0.05 CMH adj Birthweight (median, range) (g) 3317, , P, 0.05 VE adj. a 1 ¼ high level; 6 ¼ low level. b VE ¼ Van Elteren; CMH ¼ Cochran Mantel Haenszel; FE ¼ Fisher s exact test. [F(2,525) ¼ 8.88; P, 0.05; F(2,327) ¼ 18.45; P, 0.05]. When the mother had a medium or low educational level, there was no difference between the child s PIQ or GMQ. Age of the mother and child outcome Mothers in the ICSI group were significantly older than mothers in the SC group (Table II). Therefore, child outcomes were re-run with age of the mother entered as a fixed factor after defining it in two groups: women who had a child at a young age (between 19 and 32 years) (n ¼ 318) and women who had a child at an older age (between 33 and 45 years) (n ¼ 209). The results indicated that the mothers who gave birth at an older age have children with a higher VIQ [F(1,523) ¼ 3.19; P, 0.05]. For FMQ, an overall effect was found indicating that women who gave birth at young age have children with a higher FMQ [F(1,330) ¼ 9,71; P, 0.05]. Birthweight and child outcome When comparing the children with a low birthweight (, 2500 g) (n ¼ 78) versus children with a normal birthweight ( g) (n ¼ 478), a trend was observed for the scores obtained on PIQ. Children with a normal birthweight tended to score higher on PIQ than children with a lower birthweight [F(1,554) 1/4 2.75; P1/4 0.98, 0.1]. Gestational age and child outcomes When comparing children up to 36 weeks of gestation (n ¼ 41) and children between 37 and 43 weeks of gestation (n ¼ 491), those born at weeks of gestation scored better on PIQ, GMQ and FMQ than children who were born at #36 weeks [F(1,532) ¼ 4.30; P, 0.05; F(1,333) ¼ 3,86; P, 0,05; F(1,337) ¼ 6.83; P, 0.05]. Cognitive development No significant differences on VIQ, PIQ and FSIQ were found between ICSI and SC children (Table III). Scores on VIQ, PIQ and FSIQ were not influenced by differences in gestational age or birthweight. However, educational level of the mother and the age of the mother at birth did play a role. A three-way ANOVA with conception mode, educational level and age of the mother at childbirth as independent variables revealed a conception mode educational level age of the mother at birth interaction effect for PIQ and FSIQ. Highly educated, young mothers and older mothers with a low educational level produced a child with a lower score for PIQ if they were conceived after ICSI [F(2,505) ¼ 3.49; P, 0.05]. On FSIQ, ICSI children of young mothers with a high educational level scored lower than SC children of a mother with the same features, while ICSI children of young mothers with a low level of education scored better than their SC counterparts [F(2,505) ¼ 5.32; P, 0.05]. Analysis on the subtest level of the Verbal Scale revealed that ICSI children obtained a significantly higher score on Table III. Mean VIQ, PIQ and FSIQ as measured by the WPPSI-R Mean SD Mean SD Verbal IQ NS Performance IQ P, 0.05 b Full-scale IQ P, 0.05 c a Three-way ANOVA (conception mode medical centre gender); (conception mode educational level age of the mother at birth). b Significant conception mode educational level age of the mother at birth interaction effect: F(2,505) ¼ 3.49; P ¼ c Significant conception mode educational level age of the mother at birth interaction effect: F(2,505) ¼ 5.32; P ¼ Page 4 of 7

5 ICSI and psychological outcome at 5 years Table IVa. Mean of the Verbal subscales as measured by the WPPSI-R Mean SD Mean SD Information P, 0.05 b Comprehension NS Arithmetic NS Vocabulary P, 0.05 c Similarities NS Sentences P, 0.05 d a Three-way ANOVA (conception mode medical centre gender); (conception mode educational level age of the mother at birth). b Significant effect of conception mode: F(1,526) ¼ 4.89; P ¼ c Significant conception mode medical centre interaction: F(2,524) ¼ 4.18; P ¼ d Significant conception mode gender interaction: F(1,401) ¼ 5.01; P ¼ Table IVb. Mean of the performance subscales as measured by the WPPSI-R Mean SD Mean SD Object Assembly P, 0.05 b Geometric Design NS Block Design NS Mazes NS Picture Completion NS Animal Pegs NS a Three-way ANOVA (conception mode medical centre gender); (conception mode educational level age of the mother at birth). b Significant conception mode medical centre gender interaction: F(2,526) ¼ 4.17; P ¼ the subtest Information in comparison with SC children. (Table IVa). A significant conception mode medical centre interaction was found, indicating that in CNY, ICSI children obtain significantly higher scores on the subtest Vocabulary than SC children (Table IVa). A significant conception mode gender interaction was found on the subtest Sentences, showing raised scores for ICSI girls in comparison with ICSI boys, while in the SC group the opposite was true (Table IVa). Analysis on the subtest level of the Performance Scale revealed a significant conception mode medical centre gender interaction on the subtest Object Assembly. Only in BRU did ICSI boys score lower than SC boys, while ICSI girls obtained a higher score in comparison with their SC counterparts (Table IVb). An additional analysis of WPPSI-R scores was performed in which the number of children scoring 1 SD below the mean (i.e. below 85) on VIQ, PIQ and FSIQ were compared, revealing no significant differences (Table V). However, Table V. Number of children scoring 1 SD below the mean on the VIQ, PIQ and FSIQ of the WPPSI-R ICSI (n ¼ 293) SC (n ¼ 265) Statistical analysis a VIQ NS PIQ NS FSIQ NS a Fisher s exact test. Table VI. Number of children scoring 1 SD below the mean on selected subtests of the WPPSI-R ICSI (n ¼ 293) SC (n ¼ 265) Statistical analysis a Information 7 17 P, 0.05 Object Assembly P, 0.05 Block Design P, 0.05 Mazes P, 0.05 a Fisher s exact test. some differences were noted when the subtests were analysed (Table VI). On Information, a subtest of the VIQ score, SC children more often scored below 1 SD of the mean than the ICSI children (not influenced by educational level, age of the mother, gestational age or birthweight). None of the other verbal subtests (Comprehension, Arithmetic, Vocabulary, Similarities and Sentences) revealed any significant differences between the two groups. The results for the Performance Scale subtests were reversed. Here, more ICSI children scored 1 SD below the mean on three different subtests (i.e. Object Assembly, Block Design and Mazes). These results could not be explained by differences in gestational age or birthweight. For Mazes, ICSI boys obtained a lower score, not explained by educational level and age of the mother [F(1,43) ¼ 9.85; P, 0.05]. ICSI children who had a mother with a low educational level more often scored below 1 SD of the mean on Block Design than their SC counterparts [F(2,43) ¼ 3.69; P, 0.05]. For Object Assembly, no demographic variables were associated with the difference between ICSI and SC children. No significant results were noted between the two groups of children for the remaining three performance subtests (Geometric Design, Picture Completion and Animal Pegs). Motor development Results from the PDMS (available in BRU and CNY) revealed differences in GMQ and FMQ, showing that ICSI children obtained significantly lower scores than the SC children (Table VII). For GMQ, these differences could not be explained by gestational age, birthweight, educational level of the mother or age of the mother at birth. However, a separate 2 2 ANOVA (conception mode gender) for each site revealed that the difference on GMQ [F(1,163) ¼ 10.07; P, 0.05] was only present in CNY. For FMQ, a trend remained after removing CNY from the analysis, indicating that ICSI children have a lower FMQ than SC children Table VII. Mean scores on the main scales of the Peabody Development Motor Scales ICSI SC Statistical analysis a Mean SD Mean SD F-value P Gross motor quotient ,0.05 Fine motor quotient ,0.05 a Main effect (conception mode) for three-way ANOVA (conception mode medical centre gender); (conception mode educational level age of the mother at birth). Page 5 of 7

6 I.Ponjaert-Kristoffersen et al. [F(1,163) ¼ 2.90; P ¼ 0.10] while gestational age, birthweight, educational level and age of the mother at birth could not account for this difference. A comparison of the number of children in both groups scoring 1 SD below the mean showed no differences. Emotional behavioural development There were no significant differences between ICSI and SC children on the CBCL scale of internalizing behaviour but for both externalizing behaviour and the total behavioural problems scores, significant differences were found (Table VIII). Children in the ICSI group scored lower on both of these scales, indicating fewer problematic behaviours. However, further analyses revealed that this effect was due to SC children in CNY scoring significantly higher than ICSI children on the externalizing scale [F(2,457) ¼ 4.75; P, 0.05] as well as for the total number of behavioural problems [F(2,458) ¼ 4.60; P, 0.05]. This effect was seen in neither the BRU nor the GOT sample. Furthermore, no significant effects due to gender, educational level of the mother, age of the mother at birth, gestational age in weeks and birthweight were noticed. Parental stress Differences were found on all subtests and the total scale of the PSI. Scores on Parental Distress, Parent Child Dysfunctional Interaction and Difficult Child were all significantly Table VIII. Mean scores on the main scales of the Child Behaviour Checklist Internalizing behaviour Externalizing behaviour Behavioural problems (total score) Mean SD Mean SD F-value P NS , ,0.05 a Main effect (conception mode) for three-way ANOVA (conception mode medical centre gender); (conception mode educational level age of the mother at birth). Table IX. Mean scores on the main scales of the Parenting Stress Inventory Mean SD Mean SD F-value P Parent Distress ,0.05 b Parent Child ,0.05 Dysfunctional Interaction Difficult Child ,0.05 Total Stress ,0.05 a Main effect (conception mode) for three-way ANOVA (conception mode medical centre gender); (conception mode educational level age of the mother at birth). b Significant conception mode age of the mother at birth interaction effect: F(1,436) ¼ 7.27; P, Page 6 of 7 higher for parents of SC children in comparison with parents of ICSI children (Table IX). These results were not influenced by educational level of the mother, gestational age or birthweight. However, an interaction effect for conception mode age of the mother at birth was found, pointing out that young mothers of SC children experience more stress than young ICSI mothers [F(1,435) ¼ 7.27; P, 0.05]. There were no differences in the number of children in each group obtaining a clinically significant score, indicating high parental stress. Likewise, no effect due to gender was found. Discussion Three hundred 5-year-old children conceived after ICSI were matched by maternal age, child age and gender and compared with 266 children born after spontaneous conception. This is the first follow-up study to include ICSI offspring from Europe and the USA and one of the first studies to follow ICSI offspring to age 5. Overall, there were few differences between groups in cognitive, motor and behaviour functioning and in levels of parenting stress, indicating that infertility treatment by ICSI does not appear to affect the psychological well-being of children at age 5. When cognitive subtest scores were analysed, however, ICSI offspring scored significantly more often below 1 SD below the mean on three subtests (i.e. Mazes, Block Design and Object Assembly). These minor differences suggest that further study of cognitive developmental outcomes is recommended for the ICSI-conceived population, especially in visual spatial areas. The lower birthweight and increased rates of prematurity in the ICSI-conceived population were not responsible for the increased number of children obtaining lower scores on visual spatial tasks in this population. A trend was observed indicating that ICSI children scored lower on Fine Motor Development than SC children. In our sample, mothers who conceived at a younger age had children with a higher FMQ in comparison with mothers who conceived at an older age. Since in our sample ICSI mothers were significantly older at the birth of their first child, this selection bias might explain the difference that was found. Subtest differences by site were found, especially in the New York population, on behavioural and motor assessments. These differences may well be attributable to cultural differences and recruitment procedures linked to the medical systems operating in the respective countries. Children were recruited from three centres Brussels (Belgium), Göteborg (Sweden) and New York (USA) each with differing medical and insurance systems, birth registries and eligibility criteria for ART procedures. For example, in the USA, there is no birth registry or central medical organization from which to recruit control (i.e. SC) children. Therefore, families responding to study advertisements in New York may have been those with concerns about their child s development; hence the greater number of externalizing behaviour problems in the SC New York population. In addition, the cost of infertility treatments is often not covered by medical insurance in the USA, resulting in higher income and bettereducated couples seeking ARTs. Although not clinically

7 ICSI and psychological outcome at 5 years relevant, parents of SC children reported higher levels of stress from parenting in comparison with parents of ICSI children. Possible explanations include that ICSI parents tended to answer questionnaires in a more socially desirable way, or that ICSI parents in this study were more tolerant of stress related to parenting after often difficult and prolonged attempts at conception. Data on ART births indicate that growing numbers of children are conceived by ICSI and IVF each year. There are an estimated children in the world who were born after ICSI (G.Palermo, personal communication), with additional centres performing the procedure each year. It is incumbent on professionals to continue rigorous follow-up of developmental outcomes of ART offspring, especially as more technical and invasive procedures are introduced. The present study benefits from the inclusion of singleton children only. This allows the effects of ICSI to be examined without the confounding effect of a multiple birth. However, the multiple birth rate is very high and thus the findings of the present study cannot be generalized for the ICSI twin and triplet population. This population should also be studied in further research. Cross-cultural multicountry studies are also needed to investigate the impact of medical systems on ICSI and other ART offspring. European countries with socialized medical systems have been more responsible in the past in investigating outcomes of children conceived after ART. This may be because of the greater access and larger numbers of families attempting ART in these medical systems, and the existence of birth registries facilitating research in many of these countries. In addition to multicountry studies, developmental evaluation of ICSI children at adolescence is needed, in order to measure both psychological and medical well-being of these children as they enter childbearing age. Continued research of child outcomes is critical to ensure that reproductive technologies are successful in the long term for all children and families. Acknowledgements We would like to thank the parents and children who participated in this study, the fertility clinics as well as the schools who assisted with recruitment. This project was funded by the Bertarelli Foundation, the Flemish Research Fund for Medical Research, the Swedish Society of Medicine and the Swedish Research Council. The protocol was approved by the ethics committee of each institution in accordance with national regulations in each country. References Abidin RR (1990) Parenting Stress Index Test Manual, 3rd edn. Paediatric Psychology Press, Charlottesville, VA. Achenbach TM and Edelbrock C (1983) Manual for the Child Behavior Checklist and Revised Child Behavior Profile. University of Vermont, Department of Psychiatry, Burlington, VT. Bonduelle M, Liebaers I, Deketelaere V, Derde MP, Camus M, Devroey P and Van Steirteghem A (2002a) Neonatal data on a cohort of 2889 infants born after intracytoplasmic sperm injection (ICSI) ( ) and of 2995 infants born after in vitro fertilization (IVF) ( ). Hum Reprod 17, Bonduelle M, Van Assche E, Joris H, Keymolen K, Devroey P, Van Steirteghem A and Liebaers I (2002b) Prenatal testing in ICSI pregnancies: incidence of chromosomal anomalies in 1586 karyotypes and relation to sperm parameters. Hum Reprod 17, Bonduelle M, Ponjaert I, Van Steirteghem A, Derde MP, Devroey P and Liebaers I (2003) Developmental outcome of children born after ICSI compared to children born after IVF at the age of two years. Hum Reprod 18,1 9. Bowen JR, Gibson FL, Garth LI and Saunders DM (1998) Medical and developmental outcome at one year for children conceived by intracytoplasmic sperm injection. Lancet 351, Colpin H, Demyttenaere K and Vandemeulebroecke L (1995) New reproductive technology and the family. The parent child relationship following in-vitro fertilisation. J Child Psychol Psychiatry 36, Cook R, Vatev I, Michova Z and Golombok S (1997) The European study of assisted reproduction families: a comparison of family functioning and child development between Eastern and Western Europe. J Psychosom Obstet Gynecol 18, Folio MR and Fewell RR (2000) Peabody Developmental Motor Scales, 2nd edn. PRO-ED, Austin, TX. Golombok S (2002) The European study of assisted reproduction families: the transition to adolescence. Hum Reprod 17, Hollingshead A (1975) Four Factor Index of Social Status. Yale University, Department of Sociology, New Haven, CT. Leslie GI, Cohen J, Gibson F, Mc Mahon C and Tennant C (2003) Children conceived using ICSI have a normal development at school age. Hum Reprod 18, Ludwig M and Katalinic A (2000) Malformation rate in foetuses and children conceived after intracytoplasmic sperm injection (ICSI): results of a prospective cohort study. RBM Online 5, Squires J, Carter A and Kaplan P (2003) Developmental monitoring of children conceived by intracytoplasmic sperm injection and in vitro fertilization. Fertil Steril 79,453. Sutcliffe AG, Taylor B, Saunders K, Thornton S, Liebermann BA and Grudzinskas JG (2001) Outcome in the second year of life after in-vitro fertilisation by intracytoplasmic sperm injection: a UK case control study. Lancet 357, Sutcliffe AG, Saunders K, McLachlan R, Taylor B, Edwards P, Grudzinskas G, Leiberman B and Thornton S (2003) A retrospective case control study of developmental and other outcomes in a cohort of Australian children conceived by intracytoplasmic sperm injection compared with a similar group in the United Kingdom. Fertil Steril 79, Vander Steene G and Bos A (1997) Wechsler Preschool and Primary Scale of Intelligence. Vlaams-Nederlandse Aanpassing, 2nd edn. Testinstructie Swets Test Publishers, The Netherlands. Verhulst C, van der Ende J and Koot HM (1996) Handleiding voor de CBCL/4-18 (Nederlandse versie). Universiteit Rotterdam, The Netherlands. Wechsler D (1999) Manual for the Wechsler Preschool and Primary Scale of Intelligence-Revised. Psykologiförlaget, Stockholm, Sweden. Wennerholm UB, Bergh C, Hamberger L, Lundin K, Nilsson L, Wikland M and Kallen B (2000) Incidence of congenital malformations in children born after ICSI. Hum Reprod 15, Submitted on November 6, 2003; resubmitted on May 7, 2004; accepted on August 11, 2004 Page 7 of 7