Can Attention-Deficit/Hyperactivity Disorder Onset Occur in Adulthood?

In this issue of JAMAPsychiatry, 2 large, longitudinal, population studies from Brazil1 and the United Kingdom2 propose aparadigmatic shift in ourunderstandingof attention-deficit/ hyperactivity disorder (ADHD). They conclude, not only that theonsetofADHDcanoccur inadulthood,but that childhoodonset and adult-onset ADHDmay be distinct syndromes. Prior to thesepublications, thediagnosisofADHDinadults hadevolvedin2directions.Ameta-analysis3oflongitudinalstudiesdocumentedanage-dependentdecline in theexpressionof ADHDsymptoms.Two-thirds ofyouthwithADHDcontinued tohaveimpairingsymptomsof ADHDinyoungadulthood,despite only 15%meeting full diagnostic criteria for the disorder. TheBrazilianandUKstudies found theexpected rateofpersistence to ages 18 to 19 years: 17.2% and 21.9%%, respectively. A longitudinal population study4 from New Zealand observed a 4.9%persistencerateofADHDtoage38years.Practitioners take heed: these lowratesof casesmeeting fulldiagnostic criteria ignorethemuchhigherpersistencerateof impairingADHDsymptoms,which are relevant in clinical practice.3 In each study, the prevalence of adult-onset ADHD (Brazil, 10.3%1; UnitedKingdom, 5.5%2; andNewZealand, 2.7%4) wasmuch larger than theprevalenceof childhood-onset adult ADHD (United Kingdom, 2.6%; Brazil, 1.5%; and New Zealand, 0.3%). These estimates should be viewedwith caution. Theadults in theBrazil andUKstudieswereaged18 to 19years. That is too small a slice of adulthood to draw firm conclusions.Moreover, theratesofchildhood-onsetadultADHDseem too low.1All rates are below3.1%,which is the lower endof the 95% CI from a 10-country population study of adults.5 Why do the prospective studies of youth and population studies of adults disagree? One answer is the potential for recall biases in the latter studies.6 Another reason for disagreement is noted in theBrazilian study1: the “false-positive paradox.” This maxim states that, even when false-positive rates are low,many, or evenmost, diagnoses in a population study canbe false. For example, if theprevalenceofADHD is 5%and the false-positive rate is 5%, thenhalf thediagnoses in apopulation studywill be false. The false-positive rate is sensitive to the method of diagnosis. The child diagnoses in the studies fromBrazil,1 theUnitedKingdom,2 andNewZealand4used reports from parents and/or teachers, but the adult diagnoses were based on self-report. Self-reports of ADHD in adults are less reliable than informant reports, which is why the heritability of adult ADHD is low using self-reports but high using informant reports. In fact, theUKstudy foundavery lowheritability for adult ADHD (35%), which could be a sign of substantialmeasurementerrorandfalse-positivediagnoses.Of further concern, another longitudinal study6 found that current symptomsofADHDwereunderreportedbyadultswhohadhad ADHD in childhood and overreported by adults who did not haveADHD inchildhood.Because these concerns suggest that the UK, Brazilian, and New Zealand studies may have underestimated the persistence of ADHD and overestimated the prevalenceofadult-onsetADHD, itwouldbeamistake forpractitioners to assume that most adults referred to them with ADHD symptoms will not have a history of ADHD in youth. Theseconcernsdonotargueagainst theexistenceofadultonset ADHD or the idea that it is a clinically relevant syndrome. In fact, as a group, the adult-onset cases showed significant functional impairment. Moreover, one of the studies ruled out the idea that adult-onset ADHD is amisdiagnosis of another disorder. Further support for the validity of adultonset ADHD comes froma study of referred adultswho retrospectively reported childhood symptoms.7 Based on clinical features and familial transmission, that study concluded that onsetofADHDin lateadolescenceandearlyadulthood isvalid.7 The new articles proffer provocative conclusions: from Brazil,1 thatchildandadultADHDare“distinctsyndromes,”from theUnitedKingdom,2 “that adult ADHD ismore complex than a straightforward continuationof the childhooddisorder,” and fromNewZealand,4thatadultADHDis“notaneurodevelopmentaldisorder.”Theseconclusionsseempremature. Ineachstudy, adult-onset ADHD was de novo only in the sense that fullthresholdADHDhadnotbeendiagnosedby the investigatorsat priorassessments.IntheNewZealandstudy,comparedwithcontrols, theadult-onsetADHDgrouphadmoreteacher-ratedsymptoms of ADHD, more conduct disorder in childhood and were more likely to have had a combined parent-teacher report of ADHDsymptomonset prior to age 12 years. Likewise, the individuals in theadult-onsetADHDgroup in theUKstudyhadsignificantlyelevatedratesofADHDsymptoms, conductdisorder, andoppositionaldefiantdisorder inchildhood. In theBrazilian study,1 only 38% of participants with adult-onset ADHD were “neurotypical”asdefinedbyADHDandconductdisordersymptoms. Likewise, in a study of referred cases, one-third of lateadolescentandadult-onsetcaseshadchildhoodhistoriesofoppositionaldefiantdisorder,conductdisorder,andschoolfailure.7 Thus,many of the “adult onsets” of ADHD appear to have had neurodevelopmental roots. Theseneurodevelopmental rootspoint toanotherparadigm changerwhich, althoughnot discussed by these investigators, can be deduced from their data: the existence of subthreshold Related articles Opinion