Association of assisted reproductive technology (ART) treatment and parental infertility diagnosis with autism in ART-conceived children

Are assisted reproductive technology (ART) treatment factors or infertility diagnoses associated with autism among ART-conceived children? Our study suggests that the incidence of autism diagnosis in ART-conceived children during the first 5 years of life was higher when intracytoplasmic sperm injection (ICSI) was used compared with conventional IVF, and lower when parents had unexplained infertility (among singletons) or tubal factor infertility (among multiples) compared with other types of infertility. Some studies found an increased risk of autism among ART-conceived infants compared with spontaneously-conceived infants. However, few studies, and none in the USA, have examined the associations between types of ART procedures and parental infertility diagnoses with autism among ART-conceived children. Population-based retrospective cohort study using linkages between National ART Surveillance System (NASS) data for 1996-2006, California Birth Certificate data for 1997-2006, and California Department of Developmental Services (DDS) Autism Caseload data for 1997-2011. All live born ART-conceived infants born in California in 1997-2006 (n = 42 383) with 5-year observation period were included in the study. We assessed the annual incidence of autism diagnosis documented in DDS, which includes information on the vast majority of persons with autism in California, and the association of autism diagnosis with ART treatment factors and infertility diagnoses. Among ART-conceived singletons born in California between 1997 and 2006, the incidence of autism diagnosis remained at ∼0.8% (P for trend 0.19) and was lower with parental diagnosis of unexplained infertility (adjusted hazard risk ratio [aHRR]; 95% confidence interval: 0.38; 0.15-0.94) and higher when ICSI was used (aHRR 1.65; 1.08-2.52), when compared with cases without these patient and treatment characteristics. Among ART-conceived multiples, the incidence of autism diagnosis between 1997 and 2006 remained at ∼1.2% (P for trend 0.93) and was lower with parental diagnosis of tubal factor infertility (aHRR 0.56; 0.35-0.90) and higher when ICSI was used (aHRR 1.71; 1.10-2.66). Study limitations include imperfect data linkages, lack of data on embryo quality and possible underestimation of autism diagnosis cases. Limitations of the observational study design could affect the analysis by the possibility of residual confounders. Since information about ICSI use was missing for most frozen/thawed embryo transfer cycles, our findings of association of ICSI use and autism diagnosis can only be generalizable to fresh embryo transfer cycles. Our study provides additional evidence of the association between some types of ART procedures with autism diagnosis. Additional research is required to explain the increased risk of autism diagnosis with ICSI use, as well as studies on the effectiveness and safety of ICSI.