The birth of the first infant conceived through in vitro fertilization(IVF)occurred35yearsago.1Theearlyfocusof IVFprograms wasonenhancingsuccessratesandoptimizingclinicalandlaboratory techniques inaneffort to improvechancesofpregnancy. Currently,morethan40%2ofyoungcouplesconceiveonasingle trial of IVF comparedwith less than 20% in the 1980s.3 In 2012, thenumberof infantsbornworldwideas a result of these technologieswas5million.Thesenumberswill continueto increase with 350 000 infants born annually as a result of an estimated 1.5million IVF cycles.4 The percentage of total births resulting from IVF is currently 1.4% (>60 000 infants/y) in the United States5and0.5%to4.4%(>100 000infants/y)inwesternEurope.6 Therefore, thesuccessof IVFsuggests it is important toshift the focus to the safety of the procedure for children. Pregnancies resulting from IVF differ from spontaneous conceptions in a number of ways that may affect health and welfare of children.Adverseoutcomeshavebeenmost clearly documentedfor risks relatedtoobstetricalandperinatalhealth, with increases in preterm delivery, low birth weight, neonatal admissions, and perinatal death.7,8 Additionally, data suggest an increased risk of congenital anomalies.8,9 However, it is unclear whether the increased risks are the result of the underlying infertility, specific parental factors, pregnancy multiplicity, oruniqueaspectsof theprocedures.10,11 Thishasbeen particularly true for neurodevelopmental disorders including autism spectrum disorders and mental retardation. Over the last 2decades, theprevalenceof autismspectrum disordershasincreased,yetthepathogenesisremainslargelyunknown. Increasingly, both genetic abnormalities and environmental stresshavebeen identifiedaspossible contributing factors. Increasedparentalage,12multiplebirth,pretermbirth,and lowbirthweight13,14areknownriskfactors forautism.ThesefactorsalsofrequentlycomplicateIVFpregnancies. Inaddition,concernshavebeen raisedabout theepigenetic risk thatmayoccur withboththeovarianstimulationandprolongedinvitroculture thatarepartof theIVFprocess.15Givenanevolvingtheoryabout theroleofepigenetics innormalbraindevelopment,16,17whether IVFmaybeplacing children at risk of neurodevelopmental abnormalities is an issue. Prior studieshavebeen largely reassuringregarding therisk ofautismfollowing IVF.18Mostof thesestudieshavecomefrom theScandinaviancountrieswherecountry-widedatabasesand registriesprovidepopulations forstudyandlinkagewith importantaspectsofboththeexposuresandoutcomesofinterest.However, these studies havebeen compromisedby relatively small samplesizes,poorlycharacterizedoutcomesof interest,andlimited specific details about IVF and the associated procedures. The studybySandinandcolleagues19 in this issueof JAMA hasseveralstrengthsthataddresstheseconcerns.Thestudy,conducted in Sweden, was population based and included a large number of children (2.5million, ofwhom6959had thenarrow diagnosisofautisticdisorderand15 830hadthediagnosisofmental retardation). Power therefore shouldhavebeenadequate to address thepotential associationbetween IVFandautistic disorderandmentalretardationandtoidentifywhetherspecific IVF procedures, particularly those that involvemoremanipulation of the egg andmale factor infertility, were associated with additional risk. The inclusionof all births in thepopulationeliminated the risk of selection bias. The linkagewith the IVF registry allowed identification of specific IVF procedures including intracytoplasmic sperm injection (ICSI) and identification of spermsource,alongwithwhether theembryosweretransferred onday3orday5inafreshorfrozencycle.Allof thesefactorsmay affectpregnancyandchildhoodhealth.Therewasexcellentconfirmationof theoutcomeof interest (autisticdisorderandmental retardation)becauseall children inSwedenundergoadevelopmentalassessmentandanyabnormalities requireevaluation bya specialized team.Despite these strengths, andeven in this largepopulation,as theanalyses focusedonspecific techniques andpatientpopulations, suchassurgicallyextractedsperm(the groupwiththehighestrisk),thenumbersofchildrenweresmall— particularly for affected individuals. The results of the study support the absence of an associationbetweenanyIVFprocedureandautisticdisordercompared with spontaneous conception (relative risk [RR], 1.14 [95% CI, 0.94-1.39]; 19.0vs 15.6per 100 000person-years).Althoughthe authors found a small risk ofmental retardation following IVF comparedwithspontaneouspregnancies (RR,1.18[95%CI, 1.011.36]; 46.3 vs 39.8 per 100 000 person-years), this increase becamenonsignificantwhenonlyevaluatingsingletonbirths (RR, 1.01 [95%CI,0.83-1.24]).Thiswasalso true for the identified increased risk of autistic disorder with the use of surgically extracted sperm (RR, 3.29 [95% CI, 1.58-6.87, 110.1 vs 30.9 per 100 000person-years] in all births;RR,0.73 [95%CI, 0.10-5.30, 18.3vs 24.3per 100 000person-years] in singletons). The identified riskswithsurgicallyextractedspermforbothautisticdisorderandmental retardationwere increasedfurtherbypreterm birth(forautisticdisorder,RR,8.06[95%CI,2.97-21.85]; formental retardation, RR 3.31 [95%CI, 1.18-9.31]). The association of multiple birth and preterm birth with these outcomes is particularly important because decreasing the number ofmultiple births is a primary goal of assisted reRelated article page 75 Opinion