Academic performance in adolescents born after ART—a nationwide registry-based cohort study

TY - JOUR

T1 - Academic performance in adolescents born after ART—a nationwide registry-based cohort study

AU - Spangmose, Anne Lærke

AU - Malchau, Sara

AU - Schmidt, Lone

AU - Vassard, Ditte

AU - Rasmussen, Steen

AU - Loft, Anne

AU - Forman, Julie Lyng

AU - Pinborg, Anja

PY - 2017/2

Y1 - 2017/2

N2 - STUDY QUESTION: Is academic performance in adolescents aged 15-16 years and conceived after ART, measured as test scores in ninth grade, comparable to that for spontaneously conceived (SC) adolescents? SUMMARY ANSWER: ART singletons had a significantly lower mean test score in the adjusted analysis when compared with SC singletons, yet the differences were small and probably not of clinical relevance. WHAT IS KNOWN ALREADY: Previous studies have shown similar intelligence quotient (IQ) levels in ART and SC children, but only a few have been on adolescents. Academic performance measured with standardized national tests has not previously been explored in a complete national cohort of adolescents conceived after ART. STUDY DESIGN, SIZE, DURATION: A Danish national registry-based cohort including all 4766 ART adolescents (n = 2836 singletons and n = 1930 twins) born in 1995-1998 were compared with two SC control cohorts: a randomly selected singleton population (n = 5660) and all twins (n = 7064) born from 1995 to 1998 in Denmark. Nine children who died during the follow-up period were excluded from the study. PARTICIPANTS/MATERIALS, SETTING, METHODS: Mean test scores on a 7-point-marking scale from -3 to 12 were compared, and adjustments were made for relevant reproductive and socio-demographic covariates including occupational and educational level of the parents. MAIN RESULTS AND THE ROLE OF CHANCE: The crude mean test score was higher in both ART singletons and ART twins compared with SC adolescents. The crude mean differences were +0.41 (95% CI 0.30-0.53) and +0.45 (95% CI 0.28-0.62) between ART and SC singletons and between ART and SC twins, respectively. However, the adjusted mean overall test score was significantly lower for ART singletons compared with SC singletons (adjusted mean difference -0.15 (95% CI -0.29-(-0.02))). For comparison, the adjusted mean difference was +2.05 (95% CI 1.82-2.28) between the highest and the lowest parental educational level, suggesting that the effect of ART is weak compared with the conventional predictors. The adjusted analyses showed significantly lower mean test scores in mathematics and physics/chemistry for ART singletons compared with SC singletons. Comparing ART twins with SC twins yielded no difference in academic performance in the adjusted analyses. Similar crude and adjusted overall mean test scores were found when comparing ART singletons and ART twins. LIMITATIONS, REASONS FOR CAUTION: Missing data on educational test scores occurred in 6.6% of adolescents aged 15-16 years for the birth cohorts 1995-1997, where all of the children according to their age should have passed the ninth grade exam at the time of data retrieval. As sensitivity analyses yielded no significant difference in the adjusted risk of having missing test scores between any of the groups, it is unlikely that this should bias our results. Adjustment for body mass index and smoking during pregnancy was not possible. WIDER IMPLICATIONS OF THE FINDINGS: As our results are based on national data, our findings can be applied to other populations. The findings of this paper suggest that a possible small negative effect of parental subfertility or ART treatment is counterbalanced by the higher educational level in the ART parents.

AB - STUDY QUESTION: Is academic performance in adolescents aged 15-16 years and conceived after ART, measured as test scores in ninth grade, comparable to that for spontaneously conceived (SC) adolescents? SUMMARY ANSWER: ART singletons had a significantly lower mean test score in the adjusted analysis when compared with SC singletons, yet the differences were small and probably not of clinical relevance. WHAT IS KNOWN ALREADY: Previous studies have shown similar intelligence quotient (IQ) levels in ART and SC children, but only a few have been on adolescents. Academic performance measured with standardized national tests has not previously been explored in a complete national cohort of adolescents conceived after ART. STUDY DESIGN, SIZE, DURATION: A Danish national registry-based cohort including all 4766 ART adolescents (n = 2836 singletons and n = 1930 twins) born in 1995-1998 were compared with two SC control cohorts: a randomly selected singleton population (n = 5660) and all twins (n = 7064) born from 1995 to 1998 in Denmark. Nine children who died during the follow-up period were excluded from the study. PARTICIPANTS/MATERIALS, SETTING, METHODS: Mean test scores on a 7-point-marking scale from -3 to 12 were compared, and adjustments were made for relevant reproductive and socio-demographic covariates including occupational and educational level of the parents. MAIN RESULTS AND THE ROLE OF CHANCE: The crude mean test score was higher in both ART singletons and ART twins compared with SC adolescents. The crude mean differences were +0.41 (95% CI 0.30-0.53) and +0.45 (95% CI 0.28-0.62) between ART and SC singletons and between ART and SC twins, respectively. However, the adjusted mean overall test score was significantly lower for ART singletons compared with SC singletons (adjusted mean difference -0.15 (95% CI -0.29-(-0.02))). For comparison, the adjusted mean difference was +2.05 (95% CI 1.82-2.28) between the highest and the lowest parental educational level, suggesting that the effect of ART is weak compared with the conventional predictors. The adjusted analyses showed significantly lower mean test scores in mathematics and physics/chemistry for ART singletons compared with SC singletons. Comparing ART twins with SC twins yielded no difference in academic performance in the adjusted analyses. Similar crude and adjusted overall mean test scores were found when comparing ART singletons and ART twins. LIMITATIONS, REASONS FOR CAUTION: Missing data on educational test scores occurred in 6.6% of adolescents aged 15-16 years for the birth cohorts 1995-1997, where all of the children according to their age should have passed the ninth grade exam at the time of data retrieval. As sensitivity analyses yielded no significant difference in the adjusted risk of having missing test scores between any of the groups, it is unlikely that this should bias our results. Adjustment for body mass index and smoking during pregnancy was not possible. WIDER IMPLICATIONS OF THE FINDINGS: As our results are based on national data, our findings can be applied to other populations. The findings of this paper suggest that a possible small negative effect of parental subfertility or ART treatment is counterbalanced by the higher educational level in the ART parents.

U2 - 10.1093/humrep/dew334

DO - 10.1093/humrep/dew334

M3 - Journal article

C2 - 28057876

SN - 0268-1161

VL - 32

SP - 447

EP - 456

JO - Human Reproduction

JF - Human Reproduction

IS - 2

ER -