ADHD PRS Review

REVIEW
Systematic Review: How the Attention-Deficit/

Hyperactivity Disorder Polygenic Risk Score Adds to
Our Understanding of ADHD and Associated Traits
Angelica Ronald, PhD, Nora de Bode, BSc, Tinca J.C. Polderman, PhD
Objective: To investigate, by systematically reviewing the literature, whether the attention-deficit/hyperactivity disorder (ADHD) polygenic risk score
(PRS) associates with ADHD and related traits in independent clinical and population samples.
Method: PubMed, Embase and PsychoInfo were systematically searched, alongside study bibliographies. Quality assessments were conducted, and a
best-evidence synthesis was applied. Studies were excluded when the predictor was not based on the latest ADHD genome-wide association study, when
PRS was not based on genome-wide results, or when the study was a review. Initially, 197 studies were retrieved (February 22, 2020), and a second
search (June 3, 2020) yielded a further 49 studies. From both searches, 57 studies were eligible, and 44 studies met inclusion criteria.
Results: Included studies were published in the last 3 years. Over 80% of the studies were rated excellent, based on a standardized quality assessment.
Evidence of associations between ADHD PRS and the following categories was strong: ADHD, ADHD traits, brain structure, education, externalizing
behaviors, neuropsychological constructs, physical health, and socioeconomic status. Evidence for associations with addiction, autism, and mental health
were mixed and were, so far, inconclusive. Odds ratios for PRS associating with ADHD ranged from 1.22% to 1.76%; variance explained in
dimensional assessments of ADHD traits was 0.7% to 3.3%.
Conclusion: A new wave of high-quality research using the ADHD PRS has emerged. Eventually, symptoms may be partly identified based on PRS,
but the current ADHD PRS is useful for research purposes only. This review shows that the ADHD PRS is robust and reliable, associating not only with
ADHD but many outcomes and challenges known to be linked to ADHD.
Key words: attention-deficit/hyperactivity disorder, genetics, neurodevelopment, comorbidity, psychiatry
J Am Acad Child Adolesc Psychiatry 2021;-(-):-–-.
ttention-deficit/hyperactivity disorder (ADHD) GWAS using comparatively smaller samples, the latest
A is a neurodevelopmental disorder that affects
approximately 5% of children and 2.5% of
adults.1 Decades of past research have established the sig-
GWAS on individuals with ADHD (n ¼ 20,183) and
controls (n ¼ 35,191) identified 12 independent loci
associated with ADHD.17 Several characteristics of the
nificant twin heritability of ADHD and family studies study suggested that these findings were robust: for
demonstrate its high familiality.2,3 More recently, signifi- example, significant SNP heritability of 22% was reported,
cant single nucleotide polymorphism (SNP) heritability the genome-wide significant loci were replicated, and no
estimates for ADHD have been reported.4 Together this marker demonstrated heterogeneity between studies.
evidence supports the hypothesis that common genetic GWAS data can be used to create a polygenic score, or,
variants acting additively play a role in the causes of as often referred to in studies of psychopathological traits, a
ADHD.3 In addition, twin, family, and molecular genetic polygenic risk score (PRS). A PRS can estimate an in-
studies suggest that these common variants may, to some dividual’s genetic liability for a particular disorder or trait,
degree, be shared with other conditions and traits, including based on current knowledge of the trait’s genetic architec-
autism and autistic traits,5-10 tobacco and alcohol use,11,12 ture. Technically, a PRS is calculated as the weighted sum of
and depressive and hypomanic symptoms.13-15 the risk alleles carried by an individual which are associated
A genome-wide association study (GWAS) is the prin- with a disorder based on a GWAS. Demontis et al.17 re-
cipal tool for identifying common genetic variants across the ported that the variance in ADHD explained by their
genome that influence complex traits.16 Following previous ADHD PRS was 5.5% in individuals of European ancestry
Journal of the American Academy of Child & Adolescent Psychiatry www.jaacap.org 1

Volume - / Number - / - 2021
RONALD et al.
(note that individuals of European ancestry were also used Deficit”[Title/Abstract] OR “Hyperactivity"[Title/Ab-
to calculate the score). In their samples, the PRS had an stract] OR “Hyperactive”[Title/Abstract] OR “attention
odds ratio (OR) of 1.56 between cases and controls and deficit hyperactivity disorder“[Title/Abstract] OR “Atten-
acted in a dose-dependent fashion: the higher the PRS, the tion problems”[Title/Abstract]) AND (“Polygenic risk
higher the OR for having ADHD. A PRS can be calculated score”[Title/Abstract] OR “Polygenic score”[Title/Ab-
in any genotyped sample, and thus the degree to which the stract]). Bibliographies of selected studies were also
ADHD PRS associates both with ADHD as well as other searched (by NB). A first search was conducted February
phenotypes can be explored. The latter is interesting, given 22, 2020, and a second search June 3, 2020. All abstracts
the reported co-occurrence and genetic overlap of ADHD were inspected by 2 reviewers (TJCP and NB). Studies
with many other conditions or traits such as autism and were excluded when the following conditions were met: (1)
substance use, as described above. the predictor was not an ADHD PRS; (2)) the PRS was
A PRS is thus a major methodological development, not not based on genome-wide results (but, for example, on a
only for the genetic field but also in terms of potential certain selection of SNPs); (3) the ADHD PRS was not
utility in a range of other research fields, because a PRS can based on the latest GWAS results of ADHD17; or (d) the
be easily calculated in any genotyped sample. The potential study was a review.
of PRS for clinical utility, screening, and personalized health Figure 1 provides a flow chart on the selection and
is currently a major topic of debate.18,19 reasons for exclusion of studies.
Here we present a systematic review of all studies using
the ADHD PRS based on the largest ADHD GWAS to The ADHD PRS
date,17 and provide a systematic quality assessment of all GWAS results allow the calculation of an individual PRS,
included studies. In our review, we structured our results by which is based on the aggregate effect of common genetic
the following outcome domains: diagnosed ADHD and variants that are associated with the trait of interest.20,21
ADHD traits (dimensional assessments of ADHD symp- The PRS can be used to test the association between the
toms or traits), addiction, autism and autistic traits, brain- aggregated common genetic risk for ADHD and other
based (imaging) measures, educational attainment, exter- human traits.
nalizing behaviors, mental health, neuropsychological con-
structs, physical health, socioeconomic variables, and other Categorization of Outcome Measures
(uncategorized) outcomes. Table S1, available online, pro- Categorization of outcomes was loosely based on ICD/
vides a complete list of outcomes per category. ICF,22,23 but not completely for the following reasons.
First, these classification systems would have meant losing
METHOD specificity. Second, these systems are not designed spe-
Details of the outline of our review and methods applied cifically with ADHD in mind. For example, we chose to
were preregistered with PROSPERO Framework (https:// categorize externalizing behaviors and addiction as 2
www.crd.york.ac.uk/PROSPERO) with registration num- specific categories because of their relevance to ADHD,
ber CRD42020176391 on April 28, 2020, and followed as rather than putting them under the umbrella category of
registered, except for the following: (1) the study by Hayden mental health. Thus, some categories were made more or
et al. (2013), on which we based our quality assessments, less specific, based on deliberation and consensus between
proposes 6 quality domains; however, given some overlap in authors AR and TJCP. Outcomes that were studied only
items of domains 1 and 2, we combined these, and thus once and did not fall readily into categories with other
used 5 domains instead of 6; and (2) given the sheer outcomes were placed in an “Other” category. Table S1,
number of included studies resulting from the latest GWAS available online, provides an overview of outcome mea-
(N ¼ 44) and the importance of an adequately powered sures in each category.
GWAS to use the PRS reliably, we decided to exclude a
systematic overview of studies based on older GWAS. Quality Assessment
In general, scientific studies may encounter various biases
Study Selection resulting in potentially reduced validity and generalization
PubMed, Embase, and PsychInfo were systematically of findings. Based on 2 studies by Hayden et al.,24,25 we set
searched for published, peer-reviewed studies written in up a series of quality assessment criteria, clustered in 5
English using the search terms: (“ADHD”[Title/Abstract] domains (listed below), to evaluate the quality of studies
OR “Attention Deficit”[Title/Abstract] OR “Attention- that we included in the current review.
2 www.jaacap.org Journal of the American Academy of Child & Adolescent Psychiatry

ADHD POLYGENIC RISK SCORES
1. Study Participation. A clear description of characteristics statistics of the largest GWAS on ADHD,17 for the first
of the sample under study is key to evaluate how adequately time, a reasonably powerful ADHD PRS became possible.
the sample represents the population of interest, and how Standard QC protocols are available26 to ensure that genetic
potential attrition may lead to selection bias affecting a data are correctly processed, and that important data checks
proper representation. are applied. Furthermore, when analyzing PRS data, a
proper correction for population stratification should be
2. The ADHD PRS. The validity and statistical power of a
applied.
PRS depends on 2 crucial conditions. The first one is a
powerful GWAS discovery sample, and the second one is 3. Assessment of Outcome Measures. The current review
proper quality control (QC) of the genetic data of the target includes multiple outcome measures that were tested for an
sample under study. With the publication of the summary association with the ADHD PRS. In the quality assessment,
FIGURE 1 PRISMA Flow Diagram of Study Selection

Idenficaon
Records idenfied through Addional records idenfied in June

databases (PubMed, PsycInfo, and bibliographies checked (N = 49)
Embase) searching in February
(N= 197)
Records aer duplicates removed

(n =151)
Screening
Records screened Records excluded

(n = 151) (n =94)
Full-text arcles excluded,

Eligibility
Full-text arcles assessed with reasons

for eligibility (n = 13)
(n = 57)
Studies included in
qualitave synthesis
(n = 44 )
Included
Studies included in
quantave synthesis
(meta-analysis)
N/A

RONALD et al.
the validity and reliability of these outcome measures, either nationality, size, sex and age ranges. Choice of SNP p value
tested in the study or as citation to earlier publication, were threshold (pT) is listed in the fourth column. Outcomes,
the focus of evaluation. along with covariates, are listed in the fifth column. Results
(sixth column) focus on the statistics, effect sizes, and their
4. Confounding Factors. Several confounding factors can
direction, for direct effects. The Results column describes
play a role in the relation between the genetic risk for
any mediation analyses in terms of percent reduction in
ADHD and the outcome measures. Given the variety of
direct effect and outlines any sensitivity/replication analyses.
outcome measures, the focus of evaluation was on the
Negative findings are reported, but statistics for negative
following generic confounders: sex, age, socioeconomic
findings are omitted for space considerations. The Results
status (SES), use of medication, and co-occurring disorders.
column also specifies the authors’ choice of significance
5. Analysis and Data Presentation. For a reader to judge threshold for testing the association between the ADHD
the quality of a study, an adequate presentation of the sta- PRS and outcome measure(s).
tistical analyses and results is required. Of importance is also
the target sample size, as sufficient statistical power is Descriptives of Outcome Measures and Samples
required to provide accurate conclusions on the relation Outcome measures were categorized in the following do-
between the ADHD PRS and outcome measures. Finally, mains (number of studies shown in parentheses): diagnosed
multiple testing correction should be applied when more ADHD (n ¼ 10), ADHD traits (n ¼ 16); substance and
than 1 outcome measure is tested for an association with the non-substance-based addiction phenotypes (n ¼ 8), autism
predictor variable (ie, ADHD PRS). spectrum disorders or autistic traits (n ¼ 5), brain-based
A checklist consisting of criteria as described above was (imaging) variables (n ¼ 8), educational attainment (n ¼
used to evaluate the quality of the 44 selected studies. Every 9), externalizing behaviors (n ¼ 8), mental health (n ¼ 11),
item was rated positive (þ), negative (), or /þ (ie, ful- neuropsychological constructs (n ¼ 6), physical health (n ¼
filling part of the criteria) by 2 independent reviewers 4), socioeconomic variables (n ¼ 4) and “other” (non-
(TJCP and NB). In case of any disagreement between the categorized outcomes) (n ¼ 9).
reviewers, consensus was achieved by discussion. Studies Across the 44 studies, a total of 48 samples were used.
were then ranked based on the number of biases. A bias was Four studies included 2 samples, and it should be noted that
present when more than 50% of the criteria of 1 domain these 48 samples are not all independent (see below). In
had a negative score. The highest quality was attained if at terms of sample characteristics, 25 of the 48 samples (52%)
least 50% of the items of each domain were rated as being were population samples, 16 (33%) were clinical samples,
positive.24,25 Of note, because item M (treatment and co- and 7 (15%) were community samples enriched for in-
morbidity) could be rated only for the clinical samples and dividuals with ADHD or mental illness. Children and ad-
was not applicable (NA) for the population samples, this olescents (<18 years of age) made up just over half the
item was excluded from the bias count. samples (n ¼ 25, 52%); 13 (27%) were adult samples; and
the remaining (n ¼ 10, 21%) included both children and
Best-Evidence Synthesis. Within each of the categories,
adults. It was most common for samples to come from
considerable variation was present in outcome measures.
Europe (n ¼ 25, 52%) followed by North America (n ¼
Therefore, we performed a best-evidence synthesis, to define
17, 35%), a mix of continents (n ¼ 4, 9%), or Asia (n ¼ 1,
the evidence for a true association between the ADHD PRS
2%), and 1 was missing the country of origin (2%). The
and each respective outcome category. The evidence for
samples used in more than 1 study were Avon Longitudinal
each category was determined by taking into account the
Study of Parents and Children (ALSPAC) (6 studies),
number of studies evaluating this association, the quality of
IMAGEN (3 studies), National Longitudinal Study of
these studies, and the consistency of findings across
Adolescent to Adult Health (3 studies), Child and Adoles-
studies.27 Based on this evaluation, 4 increasing levels of
cent Twin Study in Sweden (3 studies), Generation R (2
evidence were defined.28
studies), community-based sample recruited close to Ore-
gon Health and Science University USA (2 studies), and
RESULTS iPSYCH (2 studies).
The 44 studies29-71 that met our inclusion criteria are listed
Diagnosed ADHD. The ADHD PRS consistently associ-
in Table 1,72-113,124-131 and the results are summarized in
Figure S1, available online. Categories of outcome(s) are ated with diagnosed ADHD in all 10 studies. The odds
given in the first column for each study. Samples are ratios ranged from 1.22 to 1.76. This range omits 1 study
described in terms of name (where available), type, that associated with ADHD within a cohort with bipolar

Journal of the American Academy of Child & Adolescent Psychiatry
TABLE 1 Description of Included Studies on the Association Between the Attention-Deficit/Hyperactivity Disorder (ADHD) Polygenic Risk Score (PRS) and
Outcomes Measures
ADHD PRS p Outcome measures and
Category Study Sample threshold(s) (pT) covariates Results
ADHDt, BRAIN 1. Albaugh IMAGEN Study, France, UK, PRS calculation MRI: neuroanatomic imaging, and ADHD PRS was significantly
et al. 201929 Ireland, Germany based on imaging of white matter tract associated with ADHD traits in
N [ 1471L1597 participants, pT [ 0.05 microstructure correlates of participants with available cortical
age range [ 12L16 y ADHD symptomatology thickness data (r [ 0.125, p < .001),
52% female, 48% male ADHD traits: composite score and with available diffusion data
Population sample of the Development and Well- (r [ 0.137, p < .001).
Western European ancestry Being Assessment (DAWBA)72 ADHD PRS predicted
and the Strengths and neuroanatomic imaging, and
Difficulties Questionnaire73 imaging of white matter tract
(SDQ) microstructure, as it significantly
Covariates: age, sex, site, associated with the ADHD
socioeconomic status, pubertal dimensional symptom score
stage, total brain volume, PCs (b [ L0.044, p [ .045). Sex did
not significantly moderate the
association between PRS score and
mean FA.
Repeated analyses with the PRS
SNP threshold changed to p < .01
and <0.10 showed consistent
results, as did repeated analyses
controlling for IQ.
In voxelwise analysis within white
matter skeleton regions, the
neuroanatomic imaging, and
imaging of white matter tract
microstructure association was
significantly associated with ADHD
traits. Strongest associations (p <
.001, uncorrected) were revealed in
portions of the left inferior fronto-
occipital, superior longitudinal, and
www.jaacap.org
inferior longitudinal fasciculi.

ADHD PRS was not associated with
cortical thickness in the cortical
areas that were significantly
associated with ADHD traits
Statistical thresholds were p < .05
familywise error corrected, and
brain data were threshold-free
cluster enhancement corrected.
(continued )
5
6
RONALD et al.
TABLE 1 Continued
ADHDt, OTHER 2. Stojanovski Philadelphia PRS calculation based Mild traumatic brain injury (TBI), A significant interaction between
et al. 201930 Neurodevelopmental Cohort, on pT [ 1 and ADHD symptoms ADHD PRS and group (mild TBI vs.
USA Structured interview assessed no TBI) (t1427 [ L2.1, p [ .04).
N [ 1,233 participants with no symptoms and criteria ADHD PRS showed a positive
www.jaacap.org
traumatic brain injury (TBI); N [ corresponding to ADHD association with ADHD symptom
204 with mild TBI; 79 with high diagnostic criteria ADHD score in youths without TBI
risk TBI. Age range [ 8L21 y (DSM5-574) (t1224 [ 3.5, D R2 [ 0.009%, p [
47% female, 53% male Covariates: age, sex, parental .004) and no association with
Population sample education, PCs ADHD symptom score in those with
European ancestry mild TBI (t196 [ 20.4, D R2 [
2.004%, p [ .70).
Sensitivity analyses were run
excluding individuals with ADHD
and individuals taking medication
for emotions or behavior issues.
Both of these analyses showed a
similar interaction pattern, but the
interaction did not reach
significance.
A p < .05 significance threshold
was used because only 1
comparison was run.
ADHD, ASD 3. Jansen et al. Inside Out Sample, The PRS calculation based DSM-IV75 ADHD diagnosis, ASD ADHD PRS predicted both the
201931 Netherlands on 8 pT (0.01L1) diagnosis, and combined (ASD, combined (ADHD and/or ASD)
Clinical sample N [ 688, age ADHD, or both diagnoses) diagnoses (OR [ 1.28; p [ 1.3 3
range: 2e18 y (mean: 9.06, SD: Parent-rated Child Behavior 10L3) and ADHD-only (OR [ 1.4;
2.66) Check-List/6e18 (CBCL)76 p [ 3.6 3 10L4), but not ASD-
ADHD only sample: n [ 280 Covariates: age, PCs only. At the optimal p value
participants, 25% female, 75% threshold, R2 [ 0.02% for the
male; ASD only sample: n [ combined (ADHD and/or ASD)
295 participants, 27% female, sample and R2 [ 0.045% for the
73% male. Combined sample ADHD-only sample.
(ASD only and ADHD only Planned sensitivity analyses

samples above plus n [ 113 between ADHD symptom severity
participants with both scales and PRS were not run
ASDDADHD), 24% female, 76% because of low correlations.
male Significance threshold was p < .05
All European ancestry Bonferroni corrected for 72 tests.
Control sample from the
Netherlands, N [ 943, age
range 17L79 y, 37% male, 63%
female
(continued )
TABLE 1 Continued
ADHD, 4. Li 2019a32 National Longitudinal Study of PRS calculation based ADHD diagnosis based on ADHD PRS was associated with
ADDICTION, EA, Adolescent to Adult Health on pT [ 1. PRS retrospectively self-reported ADHD diagnosis (OR [ 1.22, p <
MH, PHYSICAL (Add Health), USA groups defined as ADHD symptoms keyed to the 0.001). In terms of probability of
N [ 7088 participants, mean low (<20th DSM-IV75 ADHD by PRS group, PRS low [
age: 29.00 y (SD: 1.74) percentile), medium Lifetime DSM-IV criteria for PRS medium < PGS high and PRS
54% female, 46% male (21ste70th alcohol abuse or dependence low < PRS high at p < .005.
Population sample percentiles), and high were assessed as the presence Overall significant group
63.6% Caucasian (including (>80th percentile) of at least 1 of the 4 items differences (comparing high,
Hispanic), 20.7% African compared on pertaining to alcohol abuse, medium, low PRS groups) were
American, 0.2% Native outcomes and/or 3 of the 7 items reported for all outcomes except
American, 5.1% Asian, and pertaining to alcohol alcohol abuse/dependence rates,
10.3% dependence occurring together hypertension, or on high blood
“other” in 12-mo period. cholesterol (at p < .005).
Educational attainment, Low and high ADHD PRS groups
measured by the question differed significantly (after
“What is the highest level of Bonferroni correction) on all
education that you have outcomes with exception of
achieved to date?” Scale alcohol abuse/dependence rates,
ranged from 1 (8th grade or hypertension, or on high blood
less) to 10 (some graduate cholesterol.
training beyond a master’s In some cases, the low PRS group
degree). differed significantly from the
Cognitive ability, measured by medium PRS group, suggesting a
Add Health Picture Vocabulary protective role for low PRS scores.
Test (AHPVT)77 Low PRS group had higher
Mental health, measured by cognition and education
diagnoses based on the DSM- attainment and lower BMI than
IV,75 the Center for medium PRS group. These same
Epidemiologic Studies variables significantly distinguished
Depression (CES-D) scale,78 and the medium and high PRS groups,
an abbreviated 4-item version as did drug abuse/dependence,
of the Cohen’s Perceived Stress ever being arrested, and perceived
www.jaacap.org
Scale.79 Also, it was asked stress.

whether the participant was Bonferroni corrected significance
“ever arrested.” threshold of p < .005 applied
Physical health determined throughout.
based on body mass index Secondary analyses demonstrated
(BMI) and patients reported consistent results in European-
whether they had hypertension ancestry subsample of total
or high blood cholesterol as sample.
reported by a doctor.
Covariates: age, sex PCs
(continued )
7
8
RONALD et al.
TABLE 1 Continued
ADHDt 5. Burton et al. Spit for Science sample, USA PRS calculation based Strengths and Weaknesses of ADHD PRS was significantly
201933 N [ 5,154 (comprising n [ on 10 pT ADHD Symptoms and Normal associated with SWAN total score
4,426 participants with parent (0.00001L 0.5) behavior rating scale (SWAN) (b [ 0.005, p [ 1.7 3 10L11, R2 [
www.jaacap.org
report; n [ 728 with self report), score80 total, inattentive, and .009), separately for parent-report
age range: 6e17 y, (mean: 11.0, hyperactive/impulsive (b[ 0.0045, p [ 9.0 3 10L9, R2 [
SD: 2.8). Of total sample, n [ subscales. .009) and self-report (b [ 0.042,
379 had community ADHD Divided sample into low, p [ 6 3 10L4, R2 [ 0.016) and
diagnosis medium, and high SWAN- separately for inattentive (b [
49% female, 51% male scoring groups (low: z 0.004, p [ 1.6 3 10L10, R2 [ 0.008)
Population sample score < L1.11, n [ 670; and hyperactive/impulsive
European Ancestry medium: z-score L0.11 to 1.11, subscales (b [ 0.004, p [ 1.3 3
n [ 3,745, and high: z- 10L9, R2 [0 .007). The association
score <1.11, n [ 739). Also with the total score was still
categorized sample using cut- significant after excluding
off identified in ROC analyses individuals with an ADHD
and published cut-off of z-score community diagnosis.
>1.65. Comparisons of ADHD PRS in the
Covariates: age, sex, array, PCs categorized SWAN-scoring groups
showed low < high, medium <
high, but low [ medium.
ADHD PRS was also significantly
higher when comparing groups

scoring above vs. below the
optimal cut-off identified in ROC
analyses for parent-reported
SWAN and using the Swanson cut-
point of z-score >1.65. The self-
rated subsample did not show a
significant difference between
groups.
Significance threshold corrected
for multiple testing throughout.
EA 6. Gialluisi Multiple samples of children with PRS calculation based Diagnoses based on school ADHD PRS was negatively associated
et al. 201934 developmental dyslexia and on 12 pT (0.01L1) history of reading problems, with WRead, Wspell, and NWRead
either unrelated controls or word reading tests, or dyslexia (R2 [ 0.004e0.007%, p w
siblings. From 8 European diagnosis. [10L5L10L7]).
countries and USA (N [ 2,562 Eight outcomes relating to ADHD PRS was not significantly
e3,468) word reading, spelling, rapid associated with the other 5
41% female, 59% male naming, and phonology that are outcomes.
considered core deficits in A significance threshold of 6.94 3
(continued )
TABLE 1 Continued
Clinical sample dyslexia: word reading (WRead), 10L5 was applied to correct for
European ancestry nonword reading (NWRead), multiple testing due to multiple
and word spelling (WSpell), other PRS being tested in parallel.
phoneme awareness (PA), digit
span (DigSpan, a measure of
verbal short-term memory), and
rapid automatized naming of
letters (RANlet), digits
(RANdig), and pictures (RANpic)
Covariates: PCs
SES 7. Rietveld and Longitudinal data from the Health No pT applied Six later-life US labor market ADHD PRS was significantly
Patel 201935 and Retirement Study (HRS), outcomes: currently working for associated with all 6 labor market
USA pay, individual earnings (gross outcomes. A 1-SD increase in
N [ 9033 including participants individual income), total ADHD PRS associated with
and spouses, age range: 50-65 y household wealth (net value of decrease in employment likelihood
54% female, 46% male total wealth, excluding second (10.15% lower odds), lower gross
Population sample home, if applicable), receiving individual income (15.80%), and
European Ancestry governmental assistance in the lower household wealth (12.98%).
form of social security disability Higher ADHD PRS associated with
insurance, receiving increased likelihood of receiving
unemployment or workers’ social security disability benefits
compensation, receiving other (20.56% higher odds), receiving
governmental transfers unemployment or worker
Educational attainment compensation (6.72% higher odds),
included as mediator and and receiving governmental
measured by years of education transfers (27.38% higher odds).
Covariates: sex, age, marital For all 6 outcomes, some of the
status, number of living association was reduced when
children, self-reported health, educational attainment was added
whether health limits work, as a mediator.
tenure in current occupation, Most results were highly consistent
www.jaacap.org
log of spousal earnings, PCs when split by sex and when split by

assessments conducted at ages
50L55 and 50L59 y.
A significance threshold of p < .05
was applied.
ADDICTION 8. Piasecki National Longitudinal Study of PRS calculation based Gambling behavior and ADHD PRS was not associated with
et al. 201936 Adolescent to Adult Health, on pT[ 1 disordered gambling either gambling behavior or
USA The 2 phenotypes were disordered gambling.
N [ 5,215 unrelated categorical: answering yes or no
(continued )
9
10
RONALD et al.
TABLE 1 Continued
participants, age range 24e34 y. to “Have you ever bought A significance threshold of p < .05
Sex ratio not provided. lottery tickets, played video was applied.
Population sample games or slot machines for
European, African, Hispanic and money, bet on horses or
www.jaacap.org
East Asian Ancestry. sporting events, or taken part in

Genetic ancestry was strongly any other kinds of gambling for
correlated (r [ 0.89) with self- money?”; and (if “yes” to the
identified previous question), answer of
race/ethnicity. The self- yes or no to: “Has your
identified race/ethnicity of 9,129 gambling ever caused serious
individuals financial problems or problems
was n [ 5754 (63%) non- in your relationships with any of
Hispanic White, n [ 1,940 (21%) your family members or
non-Hispanic Black, n [ 961 friends?”
(11%) Hispanic, 449 (5%) Asian Covariates: age, sex, PCs
and n [
23 (<1%) Native American
ASDt, 9. Torske et al. BUPGEN network, Norway PRS calculation based Diagnosis based on Autism ADHD PRS not associated with the
NEUROPSYCH 201937 N [ 176 participants referred to on pT[ 0.1 Diagnostic Observation any of the executive function
a specialized hospital unit for Schedule (ADOS), and/or the outcomes or the autistic trait scale
evaluation of autism spectrum Autism Diagnostic in a regression or when comparing
disorders (ASD), age range InterviewLRevised (ADI-R) high vs. low ADHD PRS scoring
5e22 y with Full Scale Three executive function groups (those in the top and
Intelligence Quotient (IQ) >70. outcomes from the Behavior bottom 15% of the PRS
Most (68%) had ASD. Rating Inventory of Executive distribution, respectively).
24% female, 76% male Function (BRIEF),81 an 86-item Significance threshold of p < .05
Clinical sample questionnaire. The Behavior was applied.
European ancestry Regulation Index (which
incorporates 3 subscales:
inhibit, shift, and emotional
control) and the Metacognition
Index (which incorporates 5
subscales: initiate, working

memory, plan/organize,
organization of materials, and
monitor). The Global Executive
Composite Index comprised all
8 of the above subscales.
Social function was assessed
using the Social Responsiveness
Scale, a 65-item
questionnaire.82
Covariates: age, sex, PCs
(continued )
TABLE 1 Continued
ADHD, ADHDt, 10. Nigg et al. Community recruited children, PRS calculation based A diagnostic evaluation using Using a structural equation model, it
EXTERNALISING 202038 USA on 7 pT (5 x 10-8- 1) standardized, well-normed was shown that the ADHD PRS was
ADHD sample: n [ 337 rating scales from parent and associated with ADHD severity
participants, 28% female, 72% teacher, parent semi-structured (b [ 0.171, 95% CI [ 0.085e0.258;
male clinical interview, child D R2 [ 0.029, p < .0001), irritability
Controls: n [ 177 participants, intellectual testing, and clinical (b [ 0.183, 95% CI [ 0.087e0.280;
46% female, 54% male observation. Best-estimate D R2 [ 0.034, p < .0002) and also
Age range 7L11 y research diagnoses and final with surgency/sensation seeking
Community sample enriched for eligibility were established by 2 (B [ 0.146, 95% CI [ 0.052e0.240,
children with ADHD experienced clinicians (a child D R2 [ 0.022, p [ .002). These
Northern European Ancestry psychiatrist and a child associations had adjusted for the
psychologist), who major depression PRS85 and for the
independently assigned final sadness/anxiety scores and their
diagnoses. association with ADHD. The ADHD
Dimensional score on an ADHD PRS was not associated with the
latent variable captured from sadness/anxiety latent variable.
hyperactivity and inattention In the person-centered analyses (ie,
subscales of 4 published ADHD looking at ADHD subgroups), the
scales ADHD PRS was elevated in the
Irritability captured with latent ADHD vs. not ADHD group (OR [
variable based on 2 subscale 1.43, 95% CI [ 1.17e1.75,
scores: anger and modified DR2 [.033 p [ .0004). The
soothability from the emotion dysregulation ADHD
Temperament in Middle group had elevated ADHD PRS vs.
Childhood Questionnaire other ADHD children (OR [ 1.44,
(TMCQ)83 and an oppositional 95% CI [ 1.03e2.20, Nagelkerke D
defiant disorder irritable total R2 [ 0.013, p [ .033) but the
score84 ADHD PRS did not differentiate
Latent variables were also irritable or other ADHD profiles.
created for surgency-approach All effects were independent of
and sadness-anxiety variation in ADHD severity across
A person-centered approach traits or groups. Sensitivity analysis
www.jaacap.org
compared different group suggested changes in latent

definitions of ADHD with and variable indicators or covariate
without irritability and emotion handling did not influence results.
dysregulation Significance threshold of p < .01
Covariates: sex, age, lifetime was applied.
mood disorder and PCs
EA, MH
(continued )
11
12
RONALD et al.
TABLE 1 Continued
11. Dickinson National Institute of Mental Health PRS calculation based Participants with schizophrenia The ADHD PRS did not differ
et al. 201939 Clinical Center, USA on 10 pT reduced to and their siblings were assigned significantly between schizophrenia
N [ 540 participants with DSM- a single score to 1 of 3 clusters based on patients, siblings and controls.
IV schizophrenia disorders, through principal trajectories of cognitive Within the participants with
www.jaacap.org
mean age 34.1 y (SD 10.1). components. development: cognitively stable schizophrenia, the ADHD PRS
24.6% female, 75.4% male Analyses repeated (CS), adolescent decline (AD), showed significant association with
N [ 247 siblings with no history with the 10 pT preadolescent impairment (PI). cognitive trajectory group (F [ 5.1
of psychotic disorder (limited to (0.0001-0.5) Wide-Range Achievement Test df [ 2,525, p [ .007, R2 [ 0.019%).
1 per family), 52.6% female, [WRAT] reading subtest86 and Pairwise comparisons showed PI >
47.4% male. Wechsler Adult Intelligence AD [ CS (at p < .05).
N [ 844 community control Scale [WAIS]87 used for Within the siblings, the ADHD PRS
participants, 53.8% female, cognitive assessments. did not show a significant
46.2% male Covariates: sex, age, PCs association with cognitive
Clinical sample trajectory group (F [ 0.3 df [
European ancestry 2,232), and no pairwise
comparisons were significant at
p < .05.
ADDICTION 12. Cabana- SAGE (USA) and 3 other dbGAP PRS calculation based Cocaine dependence, as ADHD-PRS was significantly
Domınguez sample datasets; on 9 pT (1 3 10L4 L measured by the DSM-IV75 associated with cocaine
et al. 201940 N [ 2,083 cases, age range 1) reduced to single dependence (pseudo- R2 [ 1.39%,
unknown, 41.6% male score with PCA p [ 4.5eL17).
N [ 4287 controls SNP threshold of p < 5.7eL04

44% female, 56.0% male applied to account for multiple
Clinical sample testing
European ancestry
MH 13. Ohi et al. The Schizophrenia Non-Affected PRS calculation based Schizophrenia (based on the ADHD PRS were not significantly
202041 Relative Project, Japan on 6 pT (0.01L1) criteria of the DSM-574 or being different between all the groups
N [ 332 participants, n [ 130 a first-degree relative of (patients with schizophrenia, their
patients with schizophrenia, someone with schizophrenia first-degree relatives, and controls)
38.2% female, 61.8% male, Covariates: PCs or between any pairwise
mean age [ 42.9, SD [ 13.1 y comparisons at p < .01.
n [ 56 unaffected first-degree Significance threshold of p < .01

relatives (41 parents/12 siblings/ applied to correct for multiple
4 offspring), 68.4% female, testing.
31.6% male, mean age: 59.7,
SD: 13.6 y
n [ 146 controls, 33.3% female,
66.6% male, mean age: 37.2,
SD: 14.1 y
Clinical sample
Japanese descent
(continued )
TABLE 1 Continued
BRAIN 14. Mooney N [ 312 Participants, age range: 7 PRS calculation based Diagnosis by Conners’ Rating ADHD PRS was negatively associated
et al. 202042 e15 y (mean age: 10.2 y), USA on pT [ 0.5 ScalesL3rd Edition short form, with TBV (b [ L0.147 [L0.27
ADHD sample: n [ 199 (30% Strengths and Difficulties to L0.03]), and this remained
female, 70% male); control Questionnaire long form significant after controlling for
sample: n [ 113 (47% female, including the impairment ADHD diagnosis.
53% male) module (SDQ), the ADHD TBV accounted for 16% of the
Community sample enriched for Rating Scale ADHD-RS association between ADHD PRS
ADHD MRI: total brain volume (TBV) and ADHD diagnosis after
Northern European ancestry and subcortical structures accounting for sex and age.
Covariates: motion during MRI ADHD PRS was not significantly
scan, PCs, age, sex, average FD associated with subcortical brain
(ie, motion during the scan structures
[average framewise Among females only, the ADHD
displacement]), sex interaction PRS was significantly associated
effect, diagnosis; TBV also a with increased putamen volume
covariate in analyses on (b [ 0.224 [0.09e
subcortical structures 0.36])
FDR correction (a [ 0.05) for the 9
volumes tested
ADHD, ADHDt, 15. Vuijk et al. Longitudinal Study of Genetic PRS calculation based DSM-IV75 Axis 1 diagnoses; a In this clinical sample including a wide
ADDICTION, 201943 Influences on Cognition on 10 pT range of parent-rated mix of psychiatric diagnoses,
ASDt, EA, (LOGIC) dimensional published scales of ADHD PRS was associated with
EXTERNALISING, N [ 433 participants, age range psychopathology broad ADHD diagnosis (OR [ 1.44,
NEUROPSYCH 7L18 y, mean age: 11.5, SD: 3.1 Somatic complaints measured 95% CI [ 1.14-1.81; Pseudo R2 [
y. Clinical sample with wide with the CBCL76 2.01; permuted p [ .0011) as well
range of diagnoses including Social cognition measured with as ADHD traits (b [ 1.46; R2 [
ADHD. ADHD participants the SRS82 2.93%; F [ 11.83, permuted p [
compared to individuals with IQ and working memory from .0007) and with Hyperactivity/
other DSM-IV Axis 1 diagnoses the Wechsler Intelligence Scale Impulsivity subscale (b [ 0.97;
37.2% female, 62.8% male for ChildreneFourth Edition for R2 [ 2.00%; F [ 8.81, permuted
www.jaacap.org
Clinical sample 7- to 16-year-olds and the p [ .0063) but not with Inattention.

Second sample for replication: Wechsler Adult Intelligence For non-ADHD outcomes, the
N [ 5,140, 19- to 60-year-old Scalee4th Edition 17- to 18- ADHD PRS predicted word reading
adult patients from a local year-olds 87,88 (b [ L2.11; R2 [ 2.05%; F [ 8.68,
health system biobank Academic achievement with the permuted p [ .0043) and
European ancestry Word Reading and Numerical numerical operations (b [ L2.20;
Operations of the Wechsler R2 [ 2.27%; F [ 9.25, permuted
Individual Achievement Test p [ .0030). ADHD PRS was also
eThird Edition89 (WIAT-III). associated with aggressive
The adult replication cohort behavior (b [ 1.58; R2 [ 2.59; F [
(continued )
13
14
RONALD et al.
TABLE 1 Continued
outcomes were ICD-10 ADHD, 10.52, permuted p [ .0019) and
whether education was working memory index
completed by age 23 y or not, (b [ L2.17; R2 [ 2.47; F [ 10.10,
and presence of substance use permuted p [ .0016). Controlling
www.jaacap.org
disorder history. for ADHD and stimulant use did

Covariates: age, sex, not change the above non-ADHD
genotyping wave (in biobank outcome findings.
analyses), PCs ADHD PRS did not significantly
predict somatic complaints
measured with the CBCL76 or social
cognition measured with the SRS,82
considered to demonstrate
discriminant validity of the ADHD
PRS.
Results are reported for the most
significant pT.
The adult psychiatric sample
showed similar results. ADHD PRS
was associated with ADHD
diagnosis (OR [ 1.21, 95% CI [
1.07e1.37, Pseudo R2 [ 0.42%,
p [ .0028) reduced likelihood of
college completion (OR [ 1.23,

95% CI [ 1.12e1.35, Pseudo R2 [
0.72%, p < .0001) and substance
use disorder (OR [ 1.18, 95% CI [
1.10e1.26, Pseudo R2 [ 0.40%, p <
.0001).
Division of youth sample into high
(>30%), medium (middle 40%), and
low (<30%) PRS scoring groups
showed that the high group had a
more severe multivariate pattern of
psychopathology compared to the

low group (b [ 0.21, p [.01). No
significant differences between the
medium and low groups were
found
Bonferroni correction for multiple
outcomes
16. Li 2019b44
(continued )
TABLE 1 Continued

ADHDt, National Longitudinal Study of PRS calculation based Latent classes were derived for ADHD PRS correlated 0.084 with
EXTERNALISING Adolescent to Adult Health on pT[ 1 externalizing behaviors (which ADHD symptoms (p < .01)
(Add Health), USA included aggressive behaviors, ADHD PRS predicted 17.0%
N [ 7,674 participants, age 7- nonaggressive rule-breaking, increased odds in the High
12 (wave 1) age range 18-32 y and substance use behaviors) Decreasing (OR [ 1.17 95% CI [
(later waves). assessed at waves 3 and 4 by in- 1.002, 1.366, p [ .05) and 8.0%
54% female, 46% male person interviews increased odds in the Moderate
Population sample 4 mediators selected from wave (OR [ 1.08, 95% CI [ 1.004, 1.163,
63.2% Caucasian, 21.2% African- 1 assessment: Supportive p [.03) externalizing trajectories,
American, parenting, school but was not associated with the
5.1% Asian, and 10.6% Hispanic. connectedness and sensation Low Increasing (95% CI [ 0.868,
seeking assessed with 1.265) trajectory, relative to the
questionnaires; Peer closeness Normal trajectory group
assessed in relation to 10 There was no longer evidence of
named friends direct associations between ADHD
ADHD assessed retrospectively PRS on externalizing trajectory
with DSM-IV items at Wave 3 groups relative to the Normal
Covariates: PCs, sex, age, trajectory group once mediators
highest level of education, were added to the models. School
income connectedness either partially or
fully mediated the effects
Significance threshold was p <.05
EXTERNALISING 17. Riglin et al. Avon Longitudinal Study of PRS calculation based Growth mixture modeling gave 5 ADHD PRS was associated with an
201945 Parents and Children (ALSPAC), on pT [ .05 in distinct irritability trajectory increased likelihood of being in
UK primary analyses; classes: low, decreasing, both the high-persistent (OR [
N [ 7924 participants, age analyses repeated on increasing, late-childhood 1.31, 95% CI [ 1.09e1.58, p [
range 7-15 y multiple thresholds limited, and high-persistent .005) and the increasing (OR [
Population sample Parent-reported data on 1.28, 95% CI [ 1.11e1.48, p [
European ancestry irritability from the oppositional .001) trajectory classes relative to
defiant disorder section of the the low irritability trajectory class.
Development and Well-Being The odds were similar for being in
Assessment (DAWBA),1 a either trajectory (high-persistent
www.jaacap.org
structured research diagnostic compared with increasing

interview, at ages 7, 10, 13 and trajectory class: OR [ 1.02, 95%

15 y CI [ 0.81e1.29, p [ .854). The
DAWBA was also used to ADHD PRS did not predict being in
diagnose ADHD, oppositional the decreasing or late childhood
defiant disorder, conduct limited trajectory groups.
disorder, generalized anxiety Results were consistent when sex
disorder, and depression. was controlled for and when
individuals with diagnoses were
excluded. PCs were not controlled
for.
Significance threshold was p < .05.
15
(continued )
TABLE 1 Continued
16
RONALD et al.
ADHD, MH, OTHER 18. Grigoroiu- Romania and UK case-control PRS calculation based Bipolar disorder in the UK sample ADHD PRS differentiated BP cases
Serbanescu samples on 10 pT (0.01-0.5) was assessed using the ICD-10, with childhood ADHD from
et al. 201946 Romanian sample: N [ 470 and in the Romanina sample controls in the meta-analysis of
bipolar disorder (BP) cases (all with DSM-IV75 criteria, based on both samples (OR [ 0.2 (0.08
www.jaacap.org
BP type 1) (60% female; 40% Diagnostic Interview for Genetic e0.32), z [3.23, FDR-corrected
male 2%); 329 controls (57% Studies (DIGS) and medical p [ 0.024).
female; 43% male); 43% of BP records. The ADHD PRS differentiated BP
cases have childhood ADHD. Childhood ADHD within BP cases with childhood ADHD from
UK sample: N [ 472 BP cases cases was assessed BP cases without childhood ADHD
with childhood ADHD data retrospectively using the in the meta-analysis, but this did
(67% BP type 1, 33% BP type 2) Wender Utah Rating Scale not survive FDR correction (OR [
(65% female; 35% male) and (WURS)90 and for some 0.18 (0.04e0.31), z [ 2.55, p [ .011
1,287 controls (34% male; 66% Romanian cases also using FDR-p [ .055).
female). 34% of the BP cases items from the Kiddie-SADS91 ADHD PRS associated with the
has childhood ADHD. clinical interview. Assessment of continuous measure of ADHD
Romanian and UK sample childhood ADHD was made by symptoms (based on WURS and
results were meta-analyzed. clinicians. Kiddie-SADS) within the BP cases in
Clinical sample Early- and late-onset BP defined the meta-analysis (b [ 1.7 (0.7
European ancestry as age of onset at >22 y or >22 e2.69), z [ 3.34, p [ 0.0008, FDR-
y of age, respectively. corrected p [ 0.024). This result
No covariates remained when sex or BP age of
onset were included as covariates.
This association was found to be

driven by BP cases with early onset
(<22 y).
ADHD PRS did not differentiate all
BP cases from controls at either
nominal or FDR-corrected
significance (OR [ 0.085 (0L0.17),
z [ 1.95, p [ .051, FDR-corrected
p [ .105). However, it did
differentiate early-onset BP cases
from controls (OR [ 2.51 (1.04
e3.97), z [3.36, p [ .0008, FDR-

corrected p [ .024) but not late-
onset cases.
ADHD PRS predicted earlier age of
onset within BP group (b [ L0.92,
(L1.61 to 0.23), z [ L2.62, p [
.009, FDR-corrected p [ .049).
Results given here for most
significant PRS pT.
FDR correction was used to adjust
significance for multiple testing.
(continued )
TABLE 1 Continued
ADDICTION, EXT, 19. Wimberley IPSYCH Sample, Denmark, born PRS calculation based At least 1 substance use disorder ADHD PRS were associated with any
OTHER, SES et al. 201947 1981L2003. on pT 0.2 (ICD-8 and ICD-10LDiagnostic SUD (OR [ 1.30, 95% CI [ 1.11
N [ 13116 participants with Criteria for Research [DCR]23) in e1.51; Nagelkerke R2[ 0.14). For
ADHD, 26% female, 74% male Danish registers after 13th types of SUD, associations were
Of these, n [ 2368 (18.1%) birthday. Categorized by type observed for alcohol (OR [ 1.26,
developed SUD (27% female, into alcohol, cannabis, and 95% CI [ 1.04e1.53), cannabis
73% male). Median age at first other illicit drugs, and second (OR [ 1.34, 95% CI [ 1.10e1.64)
SUD diagnosis was 19.4 y (IQR categorized into severity into but not illicit drugs (OR [ 1.21,
17.2e22.3 y). use, abuse, and addiction. 95% CI [ 0.99e1.50). For severity
Clinical sample from population Nicotine use not included. of SUD, associations were
cohort Other known SUD risk factors observed for use (OR [ 1.36, 95%
Due to overlap with Demontis (presence of comorbid CI [ 1.02e1.80) and addiction
et al17 discovery sample, oppositional defiant disorder/ (OR [ 1.30, 95% CI [ 1.07e1.57)
participants split into 5 groups, conduct disorder (ODD/CD), but not abuse (OR [ 1.21, 95%
with each group consecutively parental SUD, parental mental CI [ 0.88e1.65).
used as target sample, and the disorder, paternal income, Stratified by sex, the point estimate
remaining 4 groups plus other maternal education, obtained for the ADHD PRS-SUD association
Psychiatric Genomic consortium from IPSYCH and Danish was higher in females, but CIs
samples as the discovery registers overlapped with CIs for males.
sample. Covariates: observation time (to The other known SUD risk factors
European ancestry account time at risk for SUD were all themselves associated with
given varying ages of ADHD PRS (at p < .001).
participants), sex, age. and Nevertheless, the above SUD
calendar year at first ADHD associations still remained with the
diagnosis and PCs ADHD PRS when controlling for
these known SUD risk factors.
Sensitivity analyses repeated with
different pT, different assumed
prevalences of ADHD and SUD,
and variation in population
structure showed similar results.
www.jaacap.org
Significance threshold was

Bonferroni corrected to p < .007.
MH 20. Riglin et al. Avon Longitudinal Study of PRS calculation based A “general psychopathology” ADHD PRS was associated with the
202048 Parents and Children (ALSPAC), on pT <.05 in primary (“p”) factor for ages 7 and 13 y general psychopathology “p”
UK analyses; repeated Emotional, behavioral and factor at age 7 (B [ 0.087, SE [
N [ 5,518 at age 7 y and N [ on multiple neurodevelopmental problems 0.019, p < .001), and age 13 (B [
7,017 at age 13 y thresholds were determined with the 0.095, SE[ 0.020, p < .001) while
Population sample DAWBA.72 Additionally, the including the above other 3 PRS in
European ancestry Social and Communication the models.
(continued )
17
18
RONALD et al.
TABLE 1 Continued
Disorders Checklist92 (SCDC) Without other PRS in the model,
was used for social- the ADHD PRS predicted the p
communication problems factor at age 7 (B [ 0.093, SE [
related to ASD. 0.019, p < .001, R2 [ .009%) and
www.jaacap.org
No covariates age 13 (B [ 0.095, SE [ 0.019, p <

.001, R2 [ 0.009%)
Results were consistent when the
other PRS were excluded from the
model, and analyses repeated
using inverse probability weighting
to address potential bias due
missing genetic data revealed
similar results, as did analyses at
other pT.
BRAIN, 21. Barker et al. IMAGEN Study, France, UK, PRS calculation based BMI derived from height and ADHD PRS correlated with impulsivity
EXTERNALISING, 201949 Ireland, Germany on pT 0.05 weight measurements at age 19 symptoms (r [ 0.10, p [ .014 FWE
PHYSICAL N [ 604L874 participants y corrected).
Population sample Voxel-based morphometry ADHD PRS was correlated with the
European ancestry measures of whole-brain gray neural endophenotype (r [ 0.087,
matter at age 19 y p [ 0.036 FWE corrected).
Neural responses to reward In mediation analyses, the ADHD
anticipation and reward PRS associated via the

outcome from activation maps neuroimaging substrate with
from a Monetary Incentive impulsivity symptoms (b [ 0.006,
Delay fMRI task at age 19 y 90% CI [ 0.001, 0.019) and BMI
A neural endophenotype (b [ 0.009, 90% CI [ 0.001, 0.025).
created which was made up of Significance levels ascertained from
gray matter regions and regions permutation testing and 1-sided
of activation derived from the tests, and corrected for multiple
fMRI task. testing.
Impulsivity symptoms at age 19
assessed using self-reported
Barratt Impulsivity Scale (BIS)93

Covariates: sex, imaging site,
age, PCs, and total intracranial
volume
(continued )
TABLE 1 Continued
ADHDt, EA 22. De Zeeuw The Netherlands Twin Register PRS based on ADHD symptoms (CBCL and EA PRS and ADHD PRSs correlated
et al. 201950 (NTR) transmitted and non- Teacher Report Form (TRF) for both the transmitted and non-
Trios (ie, 1 offspring and both transmitted alleles for Attention Problems scale76) transmitted PRS (r [ L 0.27 and
parents). N [ 1120e2518 8 pT (0.0001L0.5) were assessed at age 10 or 12 y. r [ L 0.23, respectively).
Population sample Academic achievement was ADHD transmitted and non-
European ancestry assessed with the Cito score, a transmitted PRS were not
Dutch nationwide standardized significantly associated with
educational achievement test.94 academic achievement (R2 w
Educational attainment in 0.6%). ADHD transmitted PRS was
adults was assessed as self- associated with ADHD symptoms
reported highest degree. (R2 [ 1%e2%).
Covariates: sex, year of birth The transmitted ADHD PGS was
(only for EA), the interaction associated with ADHD symptoms
between sex and year of birth at home (b [ 0.17, CI 0.12L0.21,
(only for EA), PCs, genotyping R2 [ 2.7%, p [ 2 3 10L13) and at
platform school (b [ 0.13, CI [ 0.08L0.17,
R2 [ 1.6%, p [ 3 3 10L7) but not
with academic achievement
(b [ L 0.08, CI [ L0.14L0.01,
R2 [ 0.6%, p [ .022). In a model
that included both the EA PRS and
ADHD PRS, the above effects
remained between ADHD PRS and
ADHD symptoms at home and
school, but the association
between ADHD PRS and academic
achievement was no longer
significant.
The non-transmitted ADHD PGS
was not associated with any of the
above 3 outcomes.
Significance threshold of p <
www.jaacap.org
.01 used.

MH 23. Yao et al. Child and Adolescent Twin Study PRS calculation based Self-reported Eating Disorder (ED) ADHD PRS was associated with the
201951 in Sweden (CATSS) on pT <1 for primary symptoms were measured by 3 EDI-2 full scale (b [ 0.027, 95%
N [ 13,472 participants, analyses, and on 7 pT subscales (Drive for Thinness, CI [ 0.005, 0.049, R2 [ 0.0012%,
assessed at age 15 y (0.00001L1) for Bulimia, and Body p [ .015) and subscales Drive for
Population sample sensitivity analyses Dissatisfaction) from the Eating Thinness (b [ 0.032, 95% CI [
European ancestry Disorder InventoryL2 (EDI-2)95 0.005, 0.059, R2 [ 0.0010%, p [
at 15 y .022) and Body Dissatisfaction (b [
0.042, 95% CI [ 0.011, 0.072, R2 [
(continued )
19
20
RONALD et al.
TABLE 1 Continued
Covariates: sex, birth year, 0.0013%, p [ .007) but not the
and PCs Bulimia subscale (b [ 0.004, 95%
CI [ L0.013, 0.021, R2 [ 0.0000%,
p [ .654).
www.jaacap.org
Results were consistent at other pT;

sex differences were not significant.
Significance threshold was p < .05.
ADDICTION, 24. Rabinowitz Sample from urban school district PRS calculation based To assess past-year marijuana The ADHD PRS correlated negatively
OTHER et al. 201852 in the mid-Atlantic region of on pT < 0.05 abuse and dependence at age with parental monitoring
USA 20 y, Composite International (r [ L0.07, p < .05) but was not
N [ 1,050 participants Diagnostic significantly correlated with
56% female, 44% male InterviewLUniversity of community disadvantage
Population sample Michigan Version (CIDI-UM)96 (r [ L0.04, p > .05).
African American was used in 2 cohorts. In the ADHD PRS was not associated with
third cohort, National Survey on marijuana use disorders and the
Drug Use and Health ADHD PRS 3 community
(NSDUH)97 was used. disadvantage and ADHD PRS 3
The Structured Interview of parental monitoring interactions
Parent Management Skills and were also not significant, nor were
PracticesLYouth Version 3-way interactions involving sex,
(SIPMSP)98 was used to assess ADHD PRS, and either community
parental monitoring (proximal disadvantage or parental

contextual factor). monitoring.
The community disadvantage Significance threshold was p < .05.
score was calculated using
census-tract level items from
the 1990 and 2000 Decennial
census99 (distal contextual
factor).
Covariates: PCs
ADHDt 25. Taylor et al. Child and Adolescent Twin Study PRS calculation based ADHD traits were measured with ADHD PRS was associated with
201953 in Sweden (CATSS) on pT 0.5. Analyses The AutismLTics, ADHD and ADHD traits at ages 9 and 12 y (b
N [ 13391 participants repeated on 5 Other Comorbidities inventory [SE] [ 0.27 [0.03], R2 [ 8.4 3 10L3,
50% females, 50% male other pT (A-TAC)100 assessed by parents p [ 5.9 3 10L19) and ADHD trait
Population sample at ages 9 and 12 y subscales hyperactivity/impulsivity
European ancestry Covariates: sex, age, PCs (b [SE] [ 0.14 [0.02], R2 [ 7.7 3
10L3, p [1.9 3 10L19) and
inattention (b [SE] [ 0.13 [0.02],
R2 [ 6.0 3 10L3, p [ 2.9 3
10L15).
(continued )
TABLE 1 Continued
After excluding children with ICD-
10 diagnosed ADHD, ADHD PRS
was still associated with ADHD
traits (b [SE] [ 0.21 [0.03], R2 [
6.2 3 10L3, p [ 2.2 3 10L13) and
the ADHD subscales.
FDR-corrections was applied to
adjust for multiple testing.
ADHDt, BRAIN 26. Alemany Generation R Study, The PRS calculation based Structural MRIs; Image processing ADHD PRS was associated with
et al. 201954 Netherlands on 6 pT “priors” (0.01 using FreeSurfer to extract attention problems subscale (b [
N [ 1,053L1,139 participants, - infinitesimal) cortical and subcortical brain 0.12, SE [ 0.00, p [ 5.36 3 10L5).
the mean age: 10.16 y, SD [ volumes. Ten volumetric brain ADHD PRS was associated with
0.60, age range [ 8.72e11.9 y measures used as outcomes: smaller caudate volume (result for
49% female, 51% male total brain volume (TBV), strongest prior: (b [ L0.08, SE [
Population sample cortical gray matter (GM), total 0.03, puncorrected [7.49 3 104)
European ancestry white matter, subcortical GM, across all priors except prior 1 at
ventricular volume, cerebellum, p < .05 and 1 prior was significant
amygdalaLhippocampus after FDR correction.
complex, caudate, putamen, In subsequent mediation analyses,
and thalamus (final 3 are no evidence of caudate volume
subcortical brain volumes) acting as a mediator between
Assessed on CBCL76 attention ADHD PRS and attention problems
problems subscale at ages in full sample. Stratified by sex,
8L11 y mediation was significant for boys,
Covariates: sex, age, total indicating that 11% of the
intracranial volume (for all association between ADHD PRS
except TBV analysis), PCs (prior 0.01) and attention problems
was mediated by differences in
caudate volume.
ADHD PRS was associated with
smaller TBV (result for strongest
www.jaacap.org
prior: b [ L0.07, SE [ 0.03,

puncorrected [ .006) across all priors
except prior 0.01 at p < .05, but
none significant after FDR
correction.
FDR correction at p < .05 used as
significance threshold.
(continued )
21
22
RONALD et al.
TABLE 1 Continued
́
ADDICTION 27. Gurriaran Sample from the Addictive PRS calculation based DSM-IV75 criteria for substance ADHD PRS was not associated with
et al. 201855 Disorders Assistance Units from on 6 pT (0.001L1) use disorder substance use disorders after
Galicia health care areas, Spain Covariates: sex, age, PCs multiple testing correction (Pseudo
N[ 534 substance abuse/ R2 [ w0.4, p < .05, p > .002).
www.jaacap.org
dependence patients (mean Results were similar when MHC was

age 44.89 y, SD 9.73) included.
13% female, 87% male Permutation based p value of p <
N [ 587 Control subjects .0022 used.
recruited from blood donors at
Santiago de Compostela,
Galicia. Mean age 40.26 y
(SD [ 10.70; range [ 18L65).
Not checked for substance use
50% female, 50% male
Clinical sample
European ancestry
BRAIN 28. Szekely The LONG Cohort, USA n [ 119 PRS calculation based association between polygenic ADHD PRS was not associated with
et al. 201856 cases, n [ 339 controls on 7 pT (0.0005 e 0.5) risk for ADHD and brain growth any brain growth phenotypes (all p
Mean age at first scan [ 11.47 was determined for the LONG > 0.1).
y, SD [ 3.54; mean age at cohort Significance threshold was not
second scan [ 16.13 y, SD [ association between polygenic reported.
4.72. risk for ADHD and brain growth

41% female, 59% male was determined for the LONG
Population sample enriched for cohort
ADHD cases n [ 404 European ADHD ascertained using
Americans, clinician-administered Parent
n [ 31 African Americans, n [ Diagnostic Interview for
8 Asian Americans, and n [ 15 Children and Adolescents107
participants of mixed race Longitudinal growth in volume
across 2 time points modeled
linearly for 4 brain divisions:
cerebral cortex, basal ganglia,
cerebellum, cerebral white

matter, and 1 region of interest:
right lateral prefrontal cortex
Covariates: adjusted for age at
baseline scan, interscan interval,
sex, and PCs
(continued )
TABLE 1 Continued
ADHD, ADHDt, 29. Nigg et al. Community recruited sample, PRS calculation based (a) Stop-Go task107; (b) Identical ADHD PRS was associated with
NEUROPSYCH 201857,101-129 USA on pT 0.5 Pairs Continuous Performance ADHD diagnosis (Nagelkerke
Full sample N [ 656 Results checked for Task108; (c) Spatial span forward R2 [0.045%; b [ 0.233, SE [
European-only sample n [ 514 another 6 pT and backward109; (d) Digit span 0.053, p [ .000011) and both
(337 ADHD, 71% male; 177 non- forward and backward from the parent and teacher-rated ADHD
ADHD, 52% male) age range: Wechsler Intelligence Scale for symptom latent variables
7L11 y Children, Fourth Edition110; (e) (R2 [0.033%; b [ 0.185, SE [
22% non-European, 78% N-back, including 0-back, 1- 0.043 p [ 1.69E-05 and
European ancestry back, and 2-back conditions; (f) R2 [0.027%; b [ 0.165, SE [
Community sample enriched for Delis, Kaplan, and Kramer 0.042, p [ 8.55E-05 respectively).
children with ADHD (DKEF)111 version of the Stroop Of the 5 latent cognitive variables,
task (word, color, and color- ADHD PRS only predicted working
word); (g) DKEF Trailmaking test memory (b [ 0.227, SE [ 0.040,
(number, letter, and shifting); p [ 1.39E-08) and vigilance/arousal
and (h) a motor time (b [ 0.130, SE [ 0.049, p [ .0079).
reproduction task at fast (500- It did not predict slow output
millisecond) interval from which speed, mental clock, or response
we derived clock precision inhibition.
(Parclock variation)112. In mediation models, the ADHD
ADHD diagnoses made using PRS effect on ADHD diagnosis was
DSM-IV criteria and a best- statistically mediated by working
estimate procedure memory (indirect effect, b [ 0.101,
Separate parent and teacher- SE [ 0.029, p [ .00049, 43% of
rated ADHD symptom latent genetic effect accounted for) and
variables derived from data on 3 arousal/ alertness (indirect effect
or 4 published ADHD measures b [ 0.115, SE [ 0.041, p [ .005,
that capture inattention and 49% of genetic effect accounted
hyperactivity for). The same was found for
Cognitive latent variables were models with ADHD PRS predicting
captured using PCA models parent and teacher-rated ADHD
from data on laboratory symptom latent variables, with 43%
measures of working memory, L51% of the genetic effect
www.jaacap.org
response inhibition, executive accounted for by the latent

functioning, arousal/attention, cognitive variables.
temporal information Direct PRS tests had a Hochberg
processing, and processing correction p < .05. Mediation
speed124-129 models used p <.05.
Covariates: sex, age, PCs Analyses repeated including non-
European LONG sample
participants, and when changing
the discovery sample to be
European-only, both led to similar
conclusions.
23
(continued )
24
RONALD et al.
TABLE 1 Continued
ADHD 30. Hawi et al. Participants recruited in Australia, 1000 pT from 0.0005 The association between the ADHD PRS explained 3.25% variance
201858 UK and Ireland. to 0.5 ADHD PRS and variance in the in ADHD case-control status
N [ 480 ADHD case subjects ADHD case-control status. (Nagelkerke’ s R2 [ 0.03, p [
aged 5L18 y (mean age [ ADHD cases met the criteria of 7.6E L15)
www.jaacap.org
10.27 y, SD[ 3.03). 13% female, ADHD of DSM-IV74. Significance threshold p [.001
87% male ADHD status using DSM-IV applied.
N [ 1208 controls, age 7-60 y criteria determined with
(mean age [ 20.61 y, SD[ parental semi-structured
6.76) 51% female, 49% male interview and the Conners’
European ancestry Parent Rating Scale108
Covariates: sex, age2 , age 3
gender, PCs
OTHER 31. Taylor Avon Longitudinal Study of PRS calculation based Nine participation phenotypes ADHD PRS was negatively associated
et al. 2018 Parents and Children (ALSPAC), on 5 pT (0.0005 e 0.5) derived. Participation defined with all 9 mother and children
59
UK as well as just as responding to a participation phenotypes. For
N [ 7,486 mothers, N [ 7,508 genome-wide questionnaire or attending a example, ADHD PRS predicted
children significant SNPs clinic for which the whole cohort mother total participation score
Population sample was eligible to participate negatively (ES [ L2.18, 95%
European ancestry Continuous phenotypes CI [ L2.71 to 1.64), and it
calculated by summing the predicted the child total
number of questionnaires/ participation score negatively
clinics completed and or clinics (ES [ L2.14, 95% CI [ L2.63 to

attended 1.64).
Covariates: child sex, PCs Significance threshold not given:
results reported as effect sizes.
ADHDt, EA, 32. Selzam Twins Early Development Study, PRS calculation based Parents reported on twins’ ADHD The ADHD PRS effect was split into
PHYSICAL, et al. 201960 UK on pT 1 (using a prior) traits via the Strength and between-family and within-family
MH, SES N [ 789L2,962 dizygotic (DZ) Difficulties Questionnaire73 effects using DZ twin data.
twin pairs, assessed from 12 to hyperactivity subscale and the The between-family ADHD PRS
21 y Conners’ rating scales108 at effect, which was estimated
Population sample ages 12 and/or 16 y independent of the within-family
European ancestry Educational attainments based effect, significantly predicted more

on standardized tests taken at ADHD traits (b [ 0.11, CI [
the end of compulsory 0.08L0.14; p [ 6.8 3 10L9), higher
education in the UK (General BMI (b [ 0.07, CI [ 0.03L0.11;
Certificate of Secondary p [ .008), lower IQ (b [ L0.09,
Education [GCSE]) as obtained CI [ L.12 to L.05; p [ 4.5 3 10-
for twins at age 16 y 4), and lower GCSEs (b [ L0.18,
BMI and height were self- CI [ L0.21 to L0.15; p [ 7.3 3
reported. 10L17).
(continued )
TABLE 1 Continued
IQ involved verbal and The within family ADHD PRS effect
nonverbal ability using WISC-III showed that, within pairs, the twin
assessments with higher ADHD PRS had more
Psychotic experiences assessed ADHD traits than their co-twins
using the Specific Psychotic (b [ 0.12, CI [ 0.08L0.17, p [
Experiences Questionnaire109 at 1.50eL7). Within pairs, the twin with
age 16 y. Neuroticism assessed higher ADHD PRS also lower GCSE
using a Big Five questionnaire grades than their cotwins
110
(b [ L0.06, CI [ L0.10 to L.03,
Self-rated health assessed using p [ .001).
the RAND Short-Form Health The ADHD GPS within-family
Survey111 prediction was significantly lower
Socioeconomic status: based than between-family prediction for
on maternal age at birth of the GCSEs (b [ L0.12, CI [ L0.16
first child, maternal and paternal to L0.07, p [ 4.95e 3 L5, Diff [
highest education level, and 65.4%). The between-family ADHD
maternal and paternal PRS effect on GCSEs was
occupation. significantly reduced when
Covariates: PCs, chip, plate, and socioeconomic status was
phenotypes were corrected for controlled for (p [ 7.69 3 eL4) but
age and sex was still significant.
The ADHD PRS also significantly
predicted lower SES (b [ L0.17,
CI [ L0.21 to L0.13, p [
1.32eL13)
The ADHD PRS did not significantly
predict (either as within- or as
between-family effect): height, self-
rated health, neuroticism, psychotic
experiences.
Results were stable when analyses
were rerun on the sample split by
www.jaacap.org
same-sex/opposite-sex twins,

based on differences in CHIP, using
a prior pT of 0.1, and using PRSs
with British samples removed
Statistical significance was p < .01,
based on an Benjamini Hochberg
false discovery rate (FDR)
adjustment
OTHER
(continued )
25
26
RONALD et al.
TABLE 1 Continued
33. Schoeler Avon Longitudinal Study of PRS calculation based Exposure to bullying was assessed ADHD PRS was significantly
et al. 201961 Parents and Children (ALSPAC), on 99 pT (0.01e1) based on child reports at 8, 10, associated with bullying
UK and 13 y of age using a (standardized b [ 0.085; 95% CI [
N [ 5,028 participants modified version of the Bullying 0.056L0.113, p < .001). In a multi-
www.jaacap.org
Assessed at age 8, 10 and 13 y and Friendship Interview PRS analysis with 10 other
51% female, 49% male Schedule (BFIS).112 Mean score significant PRS predictors, ADHD
Population sample of exposure to bullying across PRS was still significantly associated
European ancestry ages was used. with bullying (standardized b [
Covariates: sex, PCs 0.062; 95% CI, [ 0.032L0.092, p <
.001).
Repeated multi-PRS analysis, which
looked at chronicity of bullying,
showed similar results. There was
no evidence of an interaction effect
of sex. The multi-PRS association of
ADHD PRS and bullying was no
longer significant when bullying
perpetration was included as a
covariate.
Permutation and false discovery
rateecorrected p values were
applied to estimate significance
thresholds.
OTHER 34. Mooney Community volunteers, USA PRS calculation based Diagnosis based on: diagnostic The ADHD PRS was associated with
et al. 202062 N [ 472 participants: n [ 302 on pT 0.5 parent interview (Kiddie reduced DNA methylation at 1
with ADHD (72.5% male), mean Schedule for Affective probe, cg15472673 at genome-
age [ 9.9 y (SD [ 1.4); n [ 170 Disorders and Schizophrenia for wide significance (p [ 6.71Ee8),
without ADHD (54.1% male), School-Age Children— and this association remained (p [
mean age [ 9.8 y (SD [ 1.4) Epidemiologic Version [KSAD- 9.76ee8) when including ADHD
Community sample enriched for S-E]), parent and teacher status in the regression model,
ADHD standardized rating forms that suggesting that the effect was not
European ancestry assessed symptoms and driven by elevated polygenic
impairment, clinician burden in ADHD cases. The probe

observations is located between the GART and
A total of 568,281 probes SON genes in a CpG island of a
assessed for DNA methylation bivariate promoter. The SNPs in the
on the MethylationEPIC ADHD PRS are not direct
BeadChip. Differential global methylation quantitative trait loci
methylation (average for cg15472673, as such the
(continued )
TABLE 1 Continued

methylation across all probes), association with the PRS is not
as well as differentially thought to be a genetic effect on
methylated positions (DMPs) DNA methylation. The ADHD PRS
derived from saliva. Cell-type was associated with DNA
adjusted b values were the methylation levels at 12 other
outcome variables. probes at p < 1.0ee5.
Covariates: sex, age, PCs, No sex interactions were significant
medication usage, maternal at the EWAS significance threshold.
smoking, number of missing In terms of differentially methylated
SNPs in the PRS calculation for regions, 1 region on chromosome 6
each patient. and a sex within the major histocompatibility
interaction term complex was identified, in which
the ADHD PRS associated with 8
probes associated with the
ADHD PRS. The association was
sex-specific: in female participants;
a higher PRS was associated with
higher methylation levels, and the
opposite was found for males.
ADHDt, BRAIN, EA, 35. Sudre et al. N [ 544 participants (mean 21 y, PRS calculation based Inattention and hyperactivity ADHD PRS predicted symptoms of
NEUROPSYCH 201863,113 212 [39%] with ADHD). on 7 pT (0.01 e 0.5) disorder symptoms measured hyperactivityeimpulsivity (b [ 0.11,
Majority European ancestry. using clinician administered SE [ 0.046, p [ .02, at FDR q <
Subpopulations with white non- Diagnostic Interview for 0.05), but not inattention (at FDR
Hispanic ancestry and African Children and Adolescents for q < 0.05).
American ancestry. parents.2 Adult symptoms of Of the neuroanatomic mediators
Clinical sample ADHD were measured by (white matter microstructure and
clinicians using the Conners’ cortical anatomy), the following
Adult ADHD Diagnostic emerged as partial or complete
Interview for DSM-IV.3 mediators: axial diffusivity within
Neuroanatomic imaging, and regions of the right anterior (29% of
imaging of white matter tract the genetic effect) and right
microstructure superior corona radiate (21% of the
www.jaacap.org
Other disorders in adults were genetic effect); for thickness, a

ascertained through the region within the left dorsomedial
Structured Clinical Interview for prefrontal cortex (24% of the
DSM-IV-TR Axis I Disorders.4 genetic effect); for surface area, a
Working memory spans region within the right lateral
assessed through number of temporal cortex (22% of the
correctly recalled digits/tapping genetic effect).
patterns Of the 6 cognitive domains, 3
Processing speed assessed emerged as significant mediators
using visual matching task (from of ADHD PRS / hyperactivity
the Woodcock Johnson III Test eimpulsivity symptoms: working
27
(continued )
28 TABLE 1 Continued
RONALD et al.
of Cognitive Abilities130) memory (28% of the genetic effect),
IQ was assessed using an age IQ (20% of the genetic effect) and
appropriate version of the focused attention (17% of the
Wechsler scales genetic effect). These mediators
Attentional processes fully explained the association
www.jaacap.org
measured using the Conners’ between ADHD PRS and

Continuous Performance hyperactivityeimpulsivity symptom.
Test,131 from which focused Sustained attention, processing
attention, perseverative/ speed, and perseverative/impulsive
impulsive responding and responding were not significant
sustained attention were mediators.
derived. In serial mediation analyses
Covariates: Age, sex. Also for (polygenic risk / brain regions /
imaging data: motion and cognition / symptoms), 2
quality control scores potential pathways emerged.
For mediation analyses of
neuroimaging data, used
permutation and voxelwise p < .05.
Results mostly held when analyses
repeated combining the 2 largest
subpopulations; with medication as
a covariate, excluding those with
comorbid disorders and confining
analyses to 1 member of each

family.
Applied a false discovery rate and
indicate the results that survived at
q < 0.05.
ASDt, OTHER 36. Serdarevic Generation R study, The PRS calculation based Neuromotor functioning assessed The ADHD PRS did not predict
et al (2020)64 Netherlands on 6 pT (0.01L1) during in-person home visits neuromotor functioning total or
N [ 1,174L1,921 participants using modified Touwen’s subscales after Bonferroni
Children were assessed in Neurodevelopmental correction; it predicted “Senses
infancy (9L20 wk) and at age 6 Examination.115 Separate and other” subscale nominally
y versions used for 9- to 15-wk- (b [0.43, CI [ 0.001L0.06;

49% female, 51% male old and 16- to 20-wk-old infants p [.04, R2 [ 0.01%).
Population sample Overall scale and Senses, ADHD PRS did not predict autistic
European ancestry Responses, Hypertone, traits in whole sample. ADHD PRS
Hypotone, Tone subscales. predicted autistic traits in boys only
Tone included both active and (pT < 0.10; b [ 0.176, CI [
passive muscle strength. 0.09L0.27, p < .001) after
Parent-rated autistic traits at correction for multiple testing, but
age 6 y using the Social not in girls.
Responsiveness Scale Models that were adjusted for the
Covariates: age, sex, PCs autism or schizophrenia PRS did
(continued )
TABLE 1 Continued
not change results.
Bonferroni-corrected significance
threshold of p < .005 applied.
ADHDt, BRAIN, 37. Shen et al IMAGEN Study, France, UK, pT <.50 Parent-rated Strengths and ADHD PRS was associated with
NEUROPSYCH (2020)65 Ireland, Germany Difficulties Questionnaire higher ADHD total trait score at
N [ 1,790 participants hyperactivityLinattention age 14 (r [ 0.14, df [ 1779, p <
Assessed at baseline at age 14 y subscale73 ages 14 and 16 y .001, 95% CI [ 0.097, 0.188),
and at follow-up at 16 y Neuropsychological variables: working memory errors (r [ 0.07,
49% female; 51% male working memory errors df [ 1779, p [ .002, 95% CI [
Population sample assessed using Cambridge 0.026, 0.121) and delay discounting
Ancestry not described Neuropsychological Testing rate (r [ 0.06, df [ 1779, p [ .007,
Automated Battery116 through a 95% CI [ 0.021, 0.109).
self-ordered searching task at For lower gray matter volume, the
age 14 y ADHD PRS associated only with the
Delay discounting assessed posterior occipital cluster
using the Monetary Choice (r [ L0.06, df [ 1777, p [ .009,
Questionnaire117 which includes 95% CI [ L0.106, L0.015).
items pitting a smaller Nonsignificant associations not
intermediate reward against a described in publication.
larger delayed reward at age 14 y Significance threshold not given.
Intrasubject variability was the
SD of reaction time in successful
go tasks in the stop signal
functional MRI task.118
Covariates: age, sex, and site;
analyses on GMV also controlled
for handedness and total
intracranial volume. nbjudifbujd
ADHD, ADHDt, 38. Hermosillo Community recruited children, PRS calculation based ADHD diagnoses were best- PRS statistically predicted ADHD
BRAIN, et al (2019)66 USA on pT 0.5 (4 other estimate research diagnoses diagnosis (b [ 0.153, SE [ 0.073
www.jaacap.org
NEUROPSYCH N [196 ADHD participants, thresholds tested in from parent semi-structured SE, p [ .038) and parent-reported

28% female, 72% male replications) clinical interviews, clinical symptoms (b [ 0.138, SE [ 0.059,
N [ 119 non-ADHD control observation, and parent/ p [ .020) but not teacher-rated
participants, 46% female, 54% teacher rating scales. symptoms. ADHD PRS did predict
male Parent-reported ADHD traits working memory (b [ 2.194, SE [
Age range [ 7L13 y, mean [ using a latent variable derived 0.060, p [ .001).
10.38 y (SD [ 1.55) from 5 commonly used scales ADHD PRS associated significantly
Community sample enriched for Teacher-reported ADHD traits with connectivity between the left
ADHD using a latent variable derived caudate nucleus and a cluster
European ancestry from 3 commonly used scales within the intraparietal sulcus (b [
(continued )
29
30
RONALD et al.
TABLE 1 Continued
Working memory assessed .467, SE [ 0.152, p [ .002), also
using digit span backward, reported as a significant correlation
spatial span backward, and N- (r [ 0.026, SE [ 0.162) and
back task. significantly associated with a
www.jaacap.org
MRI-based resting functional cluster of regions in the right

connectivity in a targeted set of nucleus accumbens with
subcortical structures. In total, 6 connectivity to cortex (b [ 0.270,
circuits involving subcortical SE [ 0.117, p [ 0.021).
regions: left and right caudate, No significant associations of the
left and right nucleus ADHD PRS with connectivity of the
accumbens, left and right right caudate nucleus; with
amygdala. connectivity between brain regions
Covariates: age, sex, PCs and either the left or the right
amygdala; or with the connectivity
of different clusters correlated with
the left nucleus accumbens.
A mediation model showed that
the PRSLADHD diagnosis
association was suppressed by 60%
when the connectivity of a circuit
(the connectivity between the left
caudate nucleus and the right
parietal cortex) was included in the

model. Effect sizes were similar for
both sexes. No other mediation
models showed a significant impact
of any of the other connectivity
circuits on the ADHD PRSLADHD
diagnosis, ADHD PRSLADHD
symptoms, or ADHD PRSLworking
memory associations.
Results reported as similar when
current or previous medication use
included in the models, when the

sample was sex-matched and with
other PRS pT.
Permutation testing was applied.
ADHD, ASD 39. LaBianca Families with multiple individuals No pT significance Diagnoses of ASD, ASD, or ADHD PRS significantly predicted
et al (2020)68 with ASD or ADHD recruited threshold combined ASD and ADHD, ASD, ADHD, and combined ASD
through adult psychiatric clinics, based on ICD-10 and ADHD. No further information
Denmark Affected status contingent on was provided.
N [ 39 multiplex families with PRS score A significant association was found
268 individuals, including first- PRS score had Danish samples between the ADHD PRS and being
(continued )
TABLE 1 Continued

and second-degree relatives of removed. a patient, and affected relatives
all ages up to 4 generations Covariates: sex, age and unaffected relatives (p [ .03)
Age range [ 7L13 y, mean [ using the KruskalLWallis ranked
10.38 y (SD [ 1.55) sum test.
Northern European ancestry
Clinical sample and family
relatives
ADHD 40. Demontis iPSYCH, a population based case- 10 pT were used (from 5 PRS prediction considered the ADHD PRS predicted ADHD across
et al (2019)17 cohort sample including all x 10-8 L 1). following: within iPSYCh, within all target samples compared to
singletons born in Denmark PRS in the iPSYCH PGC, and across all using leave- controls or pseudo-controls.
between May 1981 and sample were one-out analysis. iPSYCH cases Within iPSYCH (using 5-fold cross-
December 2005. European achieved with 5 diagnosed by psychiatrists at in- validation), mean of maximum
ancestry leave-one-out or out-patient clinics mostly variance explained by ADHD PRS
Psychiatric Genomic analyses, ie, 4 of 5 with ICD-10 identified using a using estimated PRS Nagelkerke’s
Consortium (PGC) includes trio groups used as Danish Psychiatric Register R2 was 5.5% (SE [ 0.0012), range [
and case control samples. Only training datasets for Controls randomly selected 0.047L0.06. Within iPSYCH, OR [
European ancestry individuals estimation of SNP from iPSYCH without ADHD or 1.56, 95% CI [ 1.53e1.60.
included in PRS analyses weights while moderate/severe mental Within PGC (with iPSYCH as
N [ 18,298 biologically estimating PRS for retardation discovery sample), OR [ 1.26 , 95%
independent PGC individuals the excluded target Individuals with a diagnosis of CI [ 1.22L1.31 variance explained
(n [ 5,599 cases; n [ 12,699 group. moderate-to-severe mental on liability scale 0.0103, p [ 2.4
controls) retardation were excluded from EL35)
N [ 37,076 biologically both cases and controls Across PGC and iPSYCH waves,
independent iPSYCH Diagnoses of ADHD derived average variance explained on
individuals (n [ 14,584 cases; from range of published liability scale [ 0.0371 (SE [
n [ 22,492 controls) instruments in PGC samples 0.0029)
Covariates: batch effects, Increasing deciles of ADHD PRS
genotyping wave, and PCs were associated with increasing OR
for ADHD, both for iPSYCH
and PGC.
ADDICTION, EA, 41. Du Rietz UK Biobank, UK PRS calculation based BMI using height and weight ADHD PRS significantly positively
www.jaacap.org
EXTERNALISING, et al. 201868 N [ 135,726, age [ 40L73 y; on multiple pT General cognitive ability predicted BMI (R2 [ 0.45%; p [

MH, PHYSICAL mean [ 56.79 y (SD [ 7.96) between 0 and 0.5 at obtained by 2-minute verbal- 4.5 3 10L129), cognitive ability
53% female, 47% male increments of .001 numerical reasoning test (R2 [ 0.38%; p [ 4.5 3 10L36),
European ancestry Neuroticism measured with alcohol intake frequency (R2 [
Population sample Eysenck Personality Inventory 0.09%; p [ 8.1 3 10L29), alcohol
In analyses, controls were Neuroticism ScaleeRevised119 dependency (R2 [ 0.21%; p [
individuals without ICD-10 or Anxiety and depressive 4.5 3 10L6), tobacco use (R2 [
self-reported diagnosis of disorders, bipolar disorder and 0.33%; p [ 4.2 3 10L21), risk taking
alcohol dependency, anxiety schizophrenia identified either (R2 [ 0.12%; p [ 9.3 3 10L25),
disorder, depressive disorder, through self-report or ICD-10 neuroticism (R2 [ .09%; p [ 2.2 3
31
(continued )
32
TABLE 1 Continued
RONALD et al.
BD, or schizophrenia and did codes. 10L24), depressive disorder (R2 [
not take lithium, Alcohol intake frequency (via .11%; p [ 2.2 3 10L13), and height
antidepressants, or self-report question); alcohol- (R2 [ .03%; p [ 8.7 3 10L20).
antipsychotics related diagnosis through either ADHD PRS did not significantly
self-report or ICD-10 codes. predict anxiety disorder, bipolar
www.jaacap.org
Smoking accessed through disorder, or schizophrenia.

hospital records Within neuroticism, the items were
Risk taking coded also studied. ADHD PRS
dichotomously based on yes/no significantly predicted mood
answer to “Would you describe swings (R2 [ 0.002%), fed-up
yourself as someone who takes feelings (R2 [ 0.20%), feelings of
risks?” loneliness and isolation (R2 [
Covariates: birthplace, age, sex, 0.19%), miserableness (R2 [ .13%),
batch, PCs irritability (R2 [0 .09%), being
tense/high-strung (R2 [ 0.07%),
guilty feelings (R2 [ 0.05%), and
having easily hurt feelings (R2 [
0.05%). It did not predict being a
nervous person or a worrier,
suffering from nerves, or often
worrying after embarrassment.
Secondary analyses showed there
were no significant sex 3 PRS
interaction effects.
Of 8 control phenotypes, included
to check for specificity, ADHD PRS
significantly and negatively
predicted height (R2 [ 0.03%) and
age (R2 [ 0.03%) but not the other
6 control phenotypes.
Significance threshold of p < 4.5 3
10L4 was applied.
ADHD, MH 42. Martin The Child and Adolescent Twin Primary analyses using ADHD, any anxiety disorder, any The ADHD PRS consistently
et al., 201869 Study in Sweden (CATSS), pT p < 0.1; analyses depression disorder or any predicted ADHD diagnoses using
Sweden. repeated on 4 anxiety or depressive disorder. registry clinical diagnoses (OR [
CATSS Registry diagnoses; n [ other pT CATSS had both registry-based 1.39, 95% CI [ 1.26e1.54, p [
217L443; unaffected n [ ICD-10 clinical diagnoses 7.2E-11, screening research
13,029L13,247 (captured from ages 9-22 y) and diagnoses (OR [ 1.25, 95% CI [
CATSS screening diagnoses screening-based diagnoses 1.17e1.34, p [ 2.8E-11) and
n [ 296L1,226; unaffected n [ based on parent-/self-rated algorithm-based research
2,083L12,228 items from the Autism-Tics, diagnoses (OR [ 1.76, 95% CI [
Avon Longitudinal Study of ADHD and Other 1.51e2.05, p [ 4.9E-13).
Parents and Children (ALSPAC), Comorbidities inventory (ATAC) The ADHD PRS predicted anxiety
(continued )
TABLE 1 Continued

UK (assessed at ages 9 or 12 y)100 disorders using registry clinical
ALSPAC algorithm diagnosed ALSPAC had algorithm-based diagnoses (OR [ 1.16, 95% CI [
n [ 199L724; unaffected n [ diagnoses based on a 1.02e1.32, p [ .020) and
1,728L2,732 semistructured interview, the algorithm-based research
Both population samples Development and Well-Being diagnoses (OR [ 1.20, 95% CI [
Both European ancestry Assessment (DAWBA72 at ages 1.08e1.33, p [ .00046) but not
7, 10, 13 and 15 y from parents. screening research diagnoses.
Self ratings were also obtained The ADHD PRS predicted
for anxiety and depression at 15 depressive disorders only using
and 18 y. algorithm-based research
Covariates: age, PCs diagnoses (OR [ 1.19, 95% CI [
1.06e1.33, p [.0027) and not using
registry clinical or screening
research diagnoses.
The ADHD PRS consistently
predicted any anxiety or depressive
disorder using registry clinical
diagnoses (OR [ 1.16, 95% CI [
1.04e1.29, p [ .0062), screening
research diagnoses (OR [ 1.12,
95% CI [ 1.01e1.25, p [ .031), and
algorithm-based research
diagnoses (OR [ 1.17, 95% CI [
1.07e1.27, p [ .00063).
Repeated analyses using other pT
showed similar results.
A significance threshold of p < .05
was applied.
MH 43. Rice et al. The Avon Longitudinal Study of pT < 0.50 Self-report depressive symptoms The AHDH PRS did not correlate
201970 Parents and Children (ALSPAC), using the short Mood and significantly with family history for
UK Feelings Questionnaire,120 6 major depression or schizophrenia
N [ 5416 adolescents with PRS ages (10.5, 12.5, 13.5, 16.5, 17.5, (both p > .05).
www.jaacap.org
scores and depression data on and 18.5 y) ADHD PRS predicted the early-
more than 1 assessment point Categorized individuals scoring adolescenceeonset depression

between 10 and 18 y above/below clinical cut-off of class (OR, 1.32, 95% CI [
47% male; 53% female scale 1.13L1.54, p < .001)
Population sample Family history measured as the In multi-PRS analyses including also
European ancestry number of family members with the schizophrenia and MDD PRS,
a history of depression or the ADHD PRS still predicted the
schizophrenia weighted by early (OR [ 1.27, 95% CI [
relatedness (first or second- 1.08L1.50, p [ .003)
(continued )
33
34
RONALD et al.
TABLE 1 Continued
degree relative) ADHD PRS did not predict the
Three trajectory classes later-onset depression trajectory
identified: persistently low class in either the univariate
(73.7%), later-adolescence analysis or the multi-PRS analysis.
www.jaacap.org
onset (17.3%), and early- Analyses that were re-run including

adolescence onset (9.0%). PCS, adjusting for missing
phenotypic data, and adjusting for
missing genetic data, showed
similar findings.
Significance threshold of p < .05
applied.
ADHDt, EA, 44. Zwicker Families Overcoming Risks and pT<.50. Analyses Total adversity score calculated as ADHD PRS was associated with
EXTERNALIZING, et al. 202071 Building Opportunities for Well- repeated using mean of 10 binary indicators: ADHD symptoms (b [ 0.21, 95%
OTHER, SES being (FORBOW) study, other pT biological mother’s education, CI [ 0.10L0.32, p < .001, R2 [
Canada biological father’s education, 3.0%) and externalizing behaviors
N[ 297 participants aged home ownership status, annual (b [ 0.23, 95% CI ¼ 0.12L0.34, p <
5L27 y (mean [ 13.5, SD [ household income, emotional .0001; R2 [ 4.0%; r [ 0.22, p < .05).
4.4) abuse, physical abuse, sexual ADHD PRS was associated with
53% female; 47% male abuse, neglect, exposure to adversity (b [ 0.23, 95% CI [
Sample enriched for offspring violence at home, bullying. 0.13L0.34, p < .0001; R2 [ 4.0%)
of parents with depression, Socioeconomic and as well as the socioeconomic
bipolar disorder, and victimization adversity subscales adversity (b [ 0.10, 95% CI

schizophrenia also studied 0.01L0.20, p [ .028; R2 [ 2.0%)
90% European ancestry; 10% ADHD symptoms: and victimization adversity
non-European ancestry <18 y: Kiddie Schedule for subscales (b [ 0.24, 95% CI [
Affective Disorders and 0.12L0.35, p < .0001 R2 [ 3.3%).
Schizophrenia (KSADS)e ADHD PRS did not significantly
Present and Lifetime Version; associate with IQ or with family
>18 y: Structured Clinical history for schizophrenia.
Interview for DSM-5 Mediation models to test the
Externalizing symptoms score ADHD PRS / adversity association
from KSADS interview showed that externalizing
IQ assessed with Wechsler symptoms mediated 22% of the

Abbreviated Scale of total effect of ADHD PGS on
IntelligenceeSecond Edition121 adversity. IQ did not mediate the
or Wechsler Preschool and ADHD PRS / adversity
Primary Scale of Intelligence association.
Covariates: age, sex, time in the Associations held when run
study, PCs separately in individuals with and
without ADHD; on the subset of
participants under age 17 y; after
(continued )
TABLE 1 Continued
excluding offspring of control
parents; among the subset of
participants with a biological
parent with mental illness; and on
the subset with self-reported
European descent.
Univariate PRS analysesused p <
.003 (Bonferroni significance
threshold corrected for multiple
tests).
Note: Sample N are given for genotyped polygenic risk score (PRS) sample used in analyses. Principal components (PCs) to control for population stratification. pT, single nucleotide
polymorphism p value threshold for PRS: if authors did not select a primary pT, results are reported for most significant pT. Outcome categories were labeled as: Attention-Deficit/
Hyperactivity Disorder diagnosis (ADHD); ADHD traits (ADHDt); Substance and nonsubstance-based addiction phenotypes (Addiction); Autism diagnosis (ASD); Autistic traits (ASDt);
Imaging-based assessments of brain variables including structure, function and connectivity (Brain); Educational attainment phenotypes (EA); Externalizing behaviors (Externalizing); Mental
health phenotypes (MH); Neuropsychological phenotypes (Neuropsy); Uncategorized phenotypes (Other); Physical health phenotypes (Physical); and Socioeconomic status variables (SES)
AHPVT ¼ Add Health Picture Vocabulary Test; BFIS ¼ Bullying and Friendship Interview Schedule; BRIEF ¼ Behavior Rating Inventory of Executive Function; CBCL ¼ Child Behavior Check-
List/6–18; CES-D ¼ Center for Epidemiologic Studies Depression scale; DAWBA ¼ Development And Well-Being Assessment; EDI-2 ¼ Eating Disorder Inventory2; GWAS ¼ genome-wide
association study; ICD ¼ International statistical Classification of Diseases and related health problems; MRI ¼ magnetic resonance imaging; PC ¼ principal component; PCA ¼ principal
component analysis; PRS ¼ polygenic risk score; pT ¼ p value threshold of discovery GWAS as used for ADHD PRS; SCDC ¼ Social and Communication Disorders Checklist; SDQ ¼
Strengths and Difficulties Questionnaire; SWAN ¼ Strengths and Weaknesses of ADHD Symptoms and Normal behavior rating scale; WURS ¼ Wender Utah Rating Scale.
www.jaacap.org

35
RONALD et al.
disorder,46 and 2 studies that did not provide enough in- the effect was not present for the full sample or in female
formation to calculate odds ratios.58,67 Several studies17,32 participants.64
showed, using deciles or groups based on low/medium/
Brain-Based (Imaging) Phenotypes. All but 1 of the 8
high scorers, that the ADHD PRS operated in a dose-
studies on brain structure or connectivity29,49,54,56,63,65,66
dependent manner in terms of its influence on ADHD
reported significant associations with the ADHD PRS.
status.
Five of these also conducted mediation analyses, within
In terms of ADHD and co-occurring conditions,
which there was a variety of evidence that brain structure
ADHD PRS was associated with having combined ADHD
mediates the association between the ADHD PRS and
and ASD in a multiplex family design including unaffected
ADHD. The specific brain-based outcomes are listed in
relatives and relatives with either or both conditions.68 The
Table S1, available online: 7 of the 8 studies included
ADHD PRS did not differentiate bipolar disorder cases with
structural measurements, including both gross indices such
ADHD from bipolar disorder cases without ADHD.2 In the
as gray matter volume or more detailed measurements such
context of other psychiatric disorders, ADHD PRS was
as subcortical structures; 2 studies included functional
associated with ADHD when controls were individuals with
parameters.
other psychiatric disorders.69
Educational Attainment. Seven of the 9 studies analysing
ADHD Traits. This was the most commonly studied educational attainment reported that the ADHD PRS was
outcome, and all studies found positive significant associa- associated with lower educational
tions with the ADHD PRS (16 studies). Percent variance attainment. 32,34,39,43,60,63,68
One nonsignificant finding
explained in ADHD traits by the ADHD PRS ranged from came from a study that did not test a straightforward as-
0.7% to 3.3%. These values were either directly reported or sociation but separated the PRS into transmitted and
were converted from correlations provided in the studies. nontransmitted alleles50 and thus tested 2 separate PRSs for
Five studies that reported on ADHD traits46,36,46,63,66,120 were their association with educational attainment, which re-
omitted from this range because their study designs were duces power.
different (eg, they investigated only subscales, they investigated
familial effects, the sample was bipolar disorder cases). Externalizing Behaviors. The ADHD PRS was significantly
Four of these studies investigated the ADHD trait sub- positively associated with a range of externalizing behaviors
scales separately, namely hyperactivity/impulsivity and inat- across 8 studies (cross-sectional assessments of irritability,
tention. Two studies (50%) found that the ADHD PRS was surgency, impulsivity, aggression, and risk taking), and
positively associated with higher scores on both sub- there was evidence that the ADHD PRS was also associated
scales,33,53 whereas 2 (50%) found that the ADHD PRS was with trajectories of increasing and persistent irritability and
positively associated with the hyperactivity/impulsivity sub- with high decreasing trajectories of externalizing
scale but not significantly associated with inattention.43,63 behaviors.38,43-45,47,49,68,71
Addiction. A range of addiction phenotypes were studied: 7 Mental Health. Within this category, there were 11
studies on substance-related addiction 32,40,47,52,55,58
and 1 studies32,39,41,43,46,48,51,60,68-70 with a broad range of
study on a nonLsubstance-related addiction (ie, phenotypes but not consistent significant findings. The
gambling).36 Three studies did not find the ADHD PRS ADHD PRS was significantly positively associated with
associated with their addiction phenotypes (which focused the general psychopathology factor in children (also
on gambling behaviors, substance abuse, and marijuana use referred to as the “p” factor).48 Higher ADHD PRS was
disorders). The other 5 studies reported all or some signif- associated with a bipolar disorder subtype combined with
icant positive associations, including with cocaine depen- ADHD when compared to unaffected controls but did
dence, substance use disorders, alcohol (intake frequency not associate with bipolar disorder when compared to
and alcohol-related diagnoses), smoking, cannabis use dis- unaffected controls. Four studies explored schizophrenia
order, use of illicit drugs, and severity of addiction. or subthreshold psychotic experiences, and none reported
a significant association with the ADHD PRS. In terms
Autism Spectrum Disorders and Autistic Traits. Five of anxiety, depression, and neuroticism, results were
studies investigated diagnosed autism or autistic traits. Only mixed. For example, the ADHD PRS was associated with
1 study (on autism) reported a significant positive associa- higher neuroticism in 1 study of older adults,58 and with
tion with the ADHD PRS, although full effect sizes were more perceived stress in another study,32 but was not
not provided.67 One study on autistic traits reported a associated with neuroticism in a youth sample.51 The
significant positive association in male participants only but ADHD PRS positively associated with depression in a
TABLE 2 Domains (15) and Criteria List for the Quality TABLE 3 Definitions of Levels of Evidence
Assessment of Included Studies
Level of evidence
Strong Consistent findings (75%) in at
Criteria
least 2 high-quality studies
1. Study participation; Study sample adequately represents the
population of interest Moderate Consistent findings (75%) in 1
(A) Description of the key characteristics of the study population high-quality study and at least 1
(distribution by age, gender, and ancestry/ethnicity) study of lower quality
(B) The sampling frame and recruitment are described, including Weak Findings in 1 high quality study or
characteristics of the place of recruitment or authors clearly consistent findings (75%) in at
reference where this information can be found least 3 studies of lower quality
(C) Inclusion and exclusion criteria are described or authors clearly Inconclusive Inconsistent findings irrespective
reference where this information can be found of study quality, or less than 3
(D) Information about participation at baseline and potential lower-quality studies available
attrition (for genetic data) are described, or authors clearly
reference where this information can be found Note: “(75%)”: Within a category, at least 75% of the findings of
2. Predictor measurement; ADHD PRS is adequately measured studies had to agree on the existence and direction of the relation
(E) Description of genetic data collection (eg, blood, saliva) and between the attention-deficit/hyperactivity disorder (ADHD) polygenic
genotyping (array) is provided, and target sample was not part risk score (PRS) and the outcome measure.
of ADHD GWAS
(F) Genetic data were subject to adequate quality control (minor
allele frequency, missing rate, relatedness participants, sex
mismatch, and genotype quality), an up-to-date imputation of diagnostic tool that was used.69 In terms of trajectories
method and an established reference panel was used of depression across ages 10 to 18 years in youths, the
(G) The ADHD PRS is adequately calculated (eg, pruning/clumping higher scores on the ADHD PRS associated with an
of SNPs)
early-adolescence–onset depression class but not late-onset
3. Outcome measurement; outcome of interest is measured in a
similar way for all participants
depression.70 The ADHD PRS also positively associated
(H) A clear definition of the outcome measures is provided with a range of eating disorder traits in youth.51
(I) Several indications are provided for the validity and reliability of
the outcome measure, or a reference is provided.
Neuropsychological Constructs. Of the 6 studies on
(J) The method and setting of outcome measurement is the same neuropsychological constructs,37,43,57,63,65,66 5 studies
for all study participants included working memory and all reported significant as-
4. Confounding measurement; important potential confounders are sociations between poorer working memory and higher
appropriately accounted for ADHD PRS. Other neuropsychological constructs studied
(K) Age, gender, and socioeconomic status are accounted for in the
in relation to the ADHD PRS were executive function
analysis
(L) Population stratification and potential batch effects are outcomes (all nonsignificant); vigilance/arousal (significant
accounted for in the analysis negative association); output speed, mental clock, and
(M) In the case of clinical samples, treatment and comorbidity are response inhibition (all nonsignificant); focused attention
accounted for in the analyses and delay discounting (significant). Three studies used
5. Analysis and data presentation; statistical analysis is appropriate
neuropsychological variables such as working memory as
(N) There is sufficient presentation of the data to assess the
adequacy of the analytic strategy
mediators in models of the association between the ADHD
(O) The number of participants in the target sample supports PRS and ADHD57,63,66 (Table 1).
sufficient statistical power (N > 400)
Physical Health. Of the 4 studies exploring physical
(P) The selected statistical model is adequate for the design of the
study health,32,49,60,68 3 studies included body mass index (BMI),
(Q) There is no evidence of selective reporting of results, and and all showed a significant positive association with
proper correction for multiple testing was applied ADHD PRS (albeit using different methods) (Table 1). The
Note: ADHD ¼ attention-deficit/hyperactivity disorder; PRS ¼ polygenic
other physical health phenotypes studied were height68
risk score; SNP ¼ single-nucleotide polymorphism. (mixed evidence), hypertension, and blood cholesterol32
(no associations for either in PRS group comparisons).
study of older adults.58 In a study of children, the Socioeconomic Status. Fourstudies35,41,58,66 tested
ADHD PRS was positively associated with any anxiety or whether the PRS associated with variables related to socio-
depressive disorder, but there were some nonsignificant economic status. All studies showed a significant association
associations for specific disorders dependent on the type with the ADHD PRS being negatively associated with

RONALD et al.
TABLE 4 Quality Assessment Results of Included Studies
5) Analysis,
1) Study 2) ADHD 3) 4) data
Domains participation PRS Outcomes Confounders N presentation
Criteria A B C D E F G H I J K L M N O P Q bias
Stojanovski et al. 201930 D D D D D D D D e D D D D D e D D 0
Albaugh et al. 201929 D D D D e/D D D D e/D D D D NA D D D D 0
Burton et al. 201933 D D D D D D D D D D e/D D e D D D D 0
Jansen et al. 201931 D D e/D D D D D D D D e/D D e/D D D D D 0
Li 2019a32 D D D D D D D D e D D D NA D D D D 0
Gialluisi et al. 201934 D e D e D D e/D D e e e e/D e D D D D 2
Rietveld and Patel 201935 D e e/D e e e e D e/D e/D D D NA D D D D 2
Piasecki et al. 201936 D D D D D D D D e D e/D D NA D D D e/D 0
Torske et al. 201937 D D D D D D D D e/D D e/D D e D e e/D e 1
Nigg et al. 201938 D D D D D D D D e D e/D D D D e/D D e/D 0
Dickinson et al. 201939 D D D e D D D D D D e/D D e D e/D D e/D 0
CabanaeDomınguez D D D D D D D D D e/D e D e e/D D D e/D 0
et al. 201940
Ohi et al. 202041 D D D D D D D D D D e D e D e D e/D 0
Mooney et al. 2020a42 D e/D D D D e/D D D D D e/D D e/D D e D D 0
Vuijk et al. 201943 D eD D D D D D D D D e/D D D D e D D 0
Li 2019b44 D D D D D D D D D D D D NA D D D e/D 0
Riglin et al. 201945 D D D D D D D D D D D e NA D D D e/D 0
GrigoroiueSerbanescu D e/D e/D e e/D e/D D D D e/D e e e e/D D e/D D 1
et al. 201946
Wimberley et al. 201947 D D D D D D D D D D D D e/D D D D D 0
Riglin et al. 202048 D D e/D D D D D D D D e e NA D D D e/D 1
Barker et al. 201949 D D e/D D e/D D D D D D e/D D NA D D D D 0
De Zeeuw et al. 201950 e/D D D D D D D D D D e/D D NA D D D D 0
Yao et al. 201951 D D D D D D D D D D e/D D e D D D e/D 0
Rabinowitz et al. 201852 D D e/D D D D D D D e/D e/D D e D D D D 0
Taylor et al. 201953 D D D D D D D D D D e/D D NA D D D D 0
Alemany et al. 201954 D D D D D D D D D D e/D D NA D D D D 0
́
Gurriaran et al. 201855 D D D D D D D D D e/D e/D D e D D D D 0
Szekely et al. 201856 D e/D e/D D D D D D D e e/D D NA e/D D D D 0
Nigg et al. 201857 D D D D D D D D D D e/D e/D D D D D D 0
Hawi et al. 201858 D e D D D D D D D D e/D D e D e D D 0
Taylor et al. 201859 D D D D D D D D e D e/D D NA D D D e/D 0
Selzam et al. 201960 D D D D D D D D e/D D D D NA D D D D 0
Schoeler et al. 201961 D D D D D D D D e D e/D D NA D D D D 0
Mooney et al. 2020b62 D e/D D e/D D D D D D D e/D D e/D D e/D D D 0
Sudre et al. 201863 e e D e D e/D D D D D e/D e/D D D e/D D D 1
Hermosillo et al. 202033 D D e/D D D D D D D D e/D D e/D D e/D D D 0
LaBianca et al. 202067 e D e D e e D D D D e/D e e e/D e/D D D 2
Serdarevic et al. 202064 e/D D D D D D D D D D D D NA D D D D 0
Shen et al. 202065 D D D D e/D D D D e/D e e/D e e/D D D D e/D 1
Demontis et al. 201917 D D e/D e/D D D D D D e e D e D D D D 0
Du Rietz et al. 201868 D D D D D D D D e/D D e/D D NA D D D D 0
Martin et al. 201869 D D D D D D D D D e e/D D NA D D D e/D 0
Rice et al. 201970 D D D D D D D D D D e/D D NA D D D e/D 0
Zwicker et al. 201971 D D D D D D D D e/D D e/D D NA D e/D D D 0

socioeconomic status (SES). The study by Selzam et al.60 behaviors, impaired working memory and education
showed a significant negative association with SES in both attainment, reduced brain volume, higher body mass index,
the between and within family design. and reduced socioeconomic status. These findings illustrate
that the well-known phenotypic associations between
Other (Noncategorized) Outcomes. In terms of the 9
ADHD and many of these phenotypes, stemming from
noncategorized outcomes,30,42,46,47,52,59,60,61,71 the ADHD
decades of research in epidemiology and developmental
PRS was positively associated with being bullied,61 bullying
psychology, may partly be explained by shared genetic ef-
chronicity,61 and a victimization adversity scale,71 a total
fects. There is an emerging literature, albeit not with
adversity scale,71 earlier age of onset of bipolar disorder,46
conclusive evidence according to our best-evidence synthe-
reduced participation in research studies,59 selected
sis, suggesting that outcomes beyond childhood, such as
methylation probes,62 reduced parental monitoring,52 and
addiction and adult mental health, may also associate with
risk of parental mental disorder or substance use disorder.47
the ADHD PRS. Some phenotypic outcomes are less well
The ADHD PRS did not associate with infant neuromotor
researched than others; this led to quite broad outcome
functioning,64 or community disadvantage, and did not
categories in some instances (eg, physical health), whereas
associate with ADHD traits in youths with mild traumatic
others were able to be more specific because of the larger
brain injury.30
body of literature (diagnosed ADHD, ADHD traits,
externalizing behaviors, and addiction).
Quality Assessments The ADHD PRS appears to carry a degree of specificity
Table 2 shows the items of the quality assessment (QA), and both in relation to other PRS, in terms of the wider context
Table 3 shows the levels of evidence. The results of the QA of neurodevelopment and mental health, and in its capacity
for each study are presented in Table 4. Three studies had 2 to significantly associate with only ADHD-relevant phe-
biases, and 5 studies had 1 bias, leaving 36 studies without notypes. Illustrating this, some studies used a multi-PRS
any notable bias. Studies that did have 1 or 2 biases were model and found that the signal from the ADHD PRS
randomly distributed across categories. Item K (correction remained significant when controlling for other
for age, sex, and socioeconomic status) was rated most often PRSs.48,61,70 In the wider context of neurodevelopment and
as /þ, because the majority of studies did not correct for mental health, the ADHD PRS often did not associate with
socioeconomic status and this criterion was not relevant for other conditions such as autism and schizophrenia31,41,67,68
the SES outcome category. Furthermore, sample sizes of or family history for mental health conditions,70,71 and it
target samples were, in some studies, n < 500, which we associated with bipolar disorder only when it co-occurred
considered small, although expected effect sizes may differ with ADHD.46 When studies included negative control
between outcome measures. traits, they invariably did not, as predicted, associate with
The criteria from the best-evidence synthesis (Table 3) the ADHD PRS.43,68 Yet, there were also some surprising
suggested that the evidence for an association between the and novel cross-disorder findings: for example, the ADHD
ADHD PRS and the following outcome categories was PRS was associated with eating disorder traits in adoles-
“strong”: diagnosed ADHD, ADHD traits, brain-based cents.51 However, it should be noted that the effect sizes of
imaging phenotypes, education, externalizing behaviors, these eating disorder trait associations (0.10%L0.13%)
neuropsychological constructs, physical health, and socio- were at least 5 times lower than the lowest estimated effect
economic status. The evidence was “inconclusive” for the size for ADHD PRS associating with ADHD traits (the
addiction, autism and autistic traits, and mental health range being 0.7%L3.3%). Thus, the literature supports the
categories. The “other” category was not included in the validity of the ADHD PRS: the most consistent and
best-evidence synthesis. strongest associations were with diagnosed ADHD and
ADHD traits.
As a literature, the use of the ADHD PRS is fast
DISCUSSION growing (44 studies in less than 3 years), of high quality (as
Overall, our literature review demonstrates that the ADHD indicated by our QA assessment), with both breadth (in
polygenic risk score (PRS) is reliable, robust, and operates in terms of the wide range of outcome phenotypes) and depth
a dose-dependent manner. We found strong evidence from (in terms of both replication within and between studies
our best-evidence synthesis that the common genetic vari- and extensive analytic protocols). The risk of false-positive
ants underlying ADHD, as captured by the ADHD PRS, results in PRS studies is potentially high from a combina-
associated not only with diagnosed ADHD but also with tion of authors being free to choose multiple significance
more dimensional ADHD traits, more externalizing thresholds on which to test associations and multiple
RONALD et al.
phenotypes. Most studies appeared to have clear measures ADHD. This new literature is worth highlighting in part
in place to avoid false-positive results: as noted in Table 1, because most attempts pre-GWAS to link neuroimaging
the majority of studies used some form of significance cri- data simultaneously to both genetics and behavior could be
terion correction and stated their SNP-based significance considered noble failures, beset with issues of multiple
thresholds (pT); most selected a single pT and provided a testing and low power.123,124 The studies in our review
justification for their choice; and many included sensitivity demonstrate that reduced brain volume mediates the asso-
analyses to ensure that results were robust. Common ciation between the ADHD PRS and ADHD. For example,
sensitivity analyses included repeating analyses on other pT, in 1 recent study, the ADHD PRS was negatively associated
on different ancestral groups within the sample, excluding with total brain volume, and total brain volume accounted
children on medication and in community samples by for 16% of the association between ADHD PRS and
excluding children with diagnosed ADHD. ADHD diagnosis.42 Mediation was also used successfully in
Within the studies on non-ADHD disorders, the other categories. For example, in the neuropsychological
ADHD PRS appears to be useful for predicting trajectories. category,57 the association between the ADHD PRS and
Specifically, the ADHD PRS appears to have trans- ADHD diagnosis was mediated by working memory and
diagnostic utility in characterizing subgroups of individuals arousal alertness latent variables. In the externalizing cate-
with early-onset symptoms in non-ADHD conditions. For gory, it was shown that externalizing symptoms mediated
example, although ADHD PRS did not associate with the association between the ADHD PRS and adversity.71
schizophrenia, within a schizophrenia sample, it associated The ADHD PRS can teach us about the core aspects
with cognitive trajectory from adolescence into adulthood, of ADHD and its nosology. Eventually, the ADHD PRS
being most strongly associated with the subgroup with may contribute to the clinical picture for individual patients,
(earliest) preadolescent cognitive impairment.39 The but because of the current small effect sizes, the ADHD
ADHD PRS did not associate with bipolar disorder, but it PRS is useful for research purposes only. Given the
associated with an earlier age of onset within bipolar dis- presence of the 3 presentations of ADHD in the DSM-5
order cases.46 Finally, the ADHD PRS associated with an (combined, predominantly inattentive, predominantly
early-onset depression trajectory class but not a later-onset hyperactiveimpulsive), it is perhaps surprising that only 4 of
depression trajectory class in youths assessed longitudinally the 16 studies on ADHD traits investigated associations of
at ages 10 to 18 years.70 the ADHD PRS separately by ADHD symptom
The ADHD PRS has been used in several studies to domain.33,43,53,63 Another study that touched on nosology
investigate geneenvironment correlation, namely, genetic proposed that emotional dysregulation should be considered
influences on environmental exposure. Direct effects of the a core component of ADHD, in light of their finding that an
ADHD PRS are reported on lower socioeconomic status,60 ADHD subgroup with emotional dysregulation had a higher
lower parental education and income,47 worse labor market ADHD PRS score compared to other ADHD subgroups.38
outcomes,35 adversity,71 and bullying victimization.61,71 Given the variety of outcome categories and the variety
Two studies went beyond direct genetic effects by of outcome measures within categories, a meta-analysis was
applying within-family analytic designs. De Zeeuw et al. not conducted. Still, we report the current range in effect
split the ADHD PRS into transmitted and nontransmitted sizes for ADHD and ADHD traits. Furthermore, to obtain
alleles to test for a process termed “genetic nurture.”50,122 insights into the reliability and strength of the associations,
They did not find that the parents’ nontransmitted we applied a best-evidence synthesis that was based on a
ADHD PRS (ie, the part of the ADHD PRS inherited by careful and systematic quality assessment of all studies.
parents but not transmitted to their offspring) influenced Other limitations of our systematic review include the fact
the offspring’s ADHD symptoms. The more elaborate that it is difficult to estimate the power of studies based on
design by Selzam et al. involved splitting up the covariance their target sample size without knowing the expected effect
within their sample of twin siblings into between-family and size of an association.125 We restricted our review to studies
within-family effects.60 The authors concluded that some of using PRS based on the largest and latest GWAS on diag-
the association between the ADHD PRS and educational nosed ADHD. This meant excluding studies on PRS
attainment might be due to passive geneLenvironment derived from ADHD traits or ADHD traits combined with
correlation effects. It is important to note, going forward, diagnosed ADHD,126 and studies using older ADHD PRS
that part of the signal in a PRS may be correlated with (eg, reviewed by Vuijk et al.43) as well as studies using a
socioeconomic factors. cross-disorder PRS that includes the ADHD PRS. Not all of
The reviewed literature included multiple studies the 44 studies are completely independent because of some
investigating PRSbrainbehavior pathways relevant to partially or completely overlapping samples. For most
categories, every study was based on a different sample. on different ancestral populations.32,57 Major initiatives in
However, it should be noted that 3 of the 10 studies on terms of both sample ascertainment and method develop-
mental health outcomes used the ALSPAC sample, and 2 ment are needed to ensure that the genetic architecture of
studies used the Child and Adolescent Twin Study in ADHD is understood regardless of the ancestry of the
Sweden (CATSS) sample. Nonetheless, given that the evi- population under study.128 At present, the literature on the
dence for the mental health category was mixed and ADHD PRS offers only partial insight globally, because
inconclusive, the repeated use of the ALSPAC and CATSS roughly only 1 in 20 studies on the current ADHD PRS to
sample in this category does not appear to have inflated the date uses participants with non-European ancestry.
consistency of the evidence for these categories. In terms of It is noted that some of the associations identified here
the other categories, 2 of the 16 ADHD trait studies and 3 are largely supported by studies using LD score regression as
of the 8 studies on brain-based outcomes used the well as from past twin studies. LD score regression provides
IMAGEN sample, and 2 of the 8 addiction studies used the an estimate of the degree of shared genetic effects in com-
Add Health sample. Finally, we included studies based on mon genetic architecture. PRS studies are distinguishable
clinical, enriched, and population-based samples. We found for several reasons, including the fact that they allow tests
no differences between the samples in their associations for associations between ADHD and other phenotypes that
with the outcome measures: in the outcome measures for currently lack a large GWAS. Furthermore, as seen in this
which we observed inconclusive results (ie, autism, addic- review, PRS can also easily be manipulated within more
tion, and mental health), significant associations did not complex analytic frameworks to test more complex hy-
cluster by sample type. potheses, such as analyses involving trajectory modeling or
While emphasizing the high quality of most of the mediation models.
reviewed literature and the strong evidence that has emerged In terms of individual prediction, the existing literature
for associations of the ADHD PRS with outcomes, a only goes as far as to compare groups scoring high, medium,
number of limitations and suggestions for improvements in and low on the ADHD PRS in a small number of our
this field of research are noted. Ideally, field standard ap- reviewed studies. The ADHD PRS cannot yet accurately
proaches in terms of the method of analyzing PRSs would predict individual outcomes, and a PRS is only as accurate
be devised, and pre-registration is essential. At present, there as the discovery sample from which it is computed. Anyone
are multiple approaches and methods that are only begin- who has used direct-to-consumer testing can upload their
ning to be formally compared.127 The selected pT and the genetic data on a new tool to calculate their own ADHD
justifications for selection of pT varied widely across studies: PRS.129 Most individuals who score high on the current
some selected p < .05 to avoid over-fitting, some selected ADHD PRS will not develop ADHD because the signal is
the pT that most accurately predicted ADHD in the work too weak. There is a strong need for public engagement and
by Demontis et al.,17 some used pT ¼ 1 to capture all public debate on the clinical usability of PRS.130 It is
variance, and others applied ranges of multiple pT. When possible that a more predictive ADHD PRS will be used in
studies did not specify their selected p value threshold, we the future, in combination with other known risk factors
had to select one from which to report the results, and this and clinical features, to support health services with pre-
may have exacerbated false-positive results. A reference- diction, diagnosis, and intervention.131 As pointed out
standardized approach may be needed to compare PRS elsewhere, there are some similarities between existing suc-
across different target samples, to avoid factors often specific cessful health screening practices (such as the newborn
to the target sample influencing PRS, including the variants Apgar score and neonatal blood spot screening) with how a
considered, linkage disequilibrium, and allele frequency PRS would be obtained and could work in practice.19
estimates.127 It will be exciting to see future work that In sum, our review identified 44 relevant studies and
combines the ADHD PRS with rare variation and copy demonstrates the accumulation of strong evidence that the
number variation or that incorporates the sex chromosomes. ADHD PRS associates not only with ADHD and ADHD
As shown in Table 1, the majority of this literature was traits, but also with reduced brain volume, lower education
conducted in samples of European ancestry: of the 44 attainment, more externalizing behaviors, impaired working
studies, 77% (n ¼ 34 studies) had European ancestry, 91% memory, higher body mass index, and lower socioeconomic
(n ¼ 41) had most or all European ancestry, 1 study had status. Alongside these direct effects, the ADHD PRS is
missing ancestry information, and 5% (n ¼ 2) had partic- being used to reveal more complex processes such
ipants with non-European ancestry (Japanese and African geneenvironment correlations and that the ADHD PRS
American, respectively). To maximize the value of the data, influences ADHD symptoms via effects on brain structure.
some studies ran sensitivity analyses on their samples based Genetic associations that might have been expected based
RONALD et al.
on past literature, such as between the ADHD PRS and along with Deputy Editor Samuele Cortese, MD, PhD.
addiction, autism, and mental health, are so far inconclusive Author Contributions
from the available evidence. In the context of other known Conceptualization: Ronald, Polderman
Data curation: de Bode, Polderman
risk factors for ADHD, the ADHD PRS does not have the Formal analysis: Ronald, de Bode, Polderman
largest effect size. Nevertheless, the ADHD PRS brings Investigation: Ronald, de Bode, Polderman
advantages in terms of being based on genetic variants, and Methodology: Ronald
Resources: de Bode
thus being biologically based, possessing a degree of cau- Supervision: Polderman
sality and being unchanging across the lifespan (unlike most Writing e original draft: Ronald, Polderman
Writing e review and editing: Ronald, Polderman
other risk factors). The estimated SNP heritability of
ORCID
ADHD is larger than the percent variance explained by the Angelica Ronald, PhD: https://orcid.org/0000-0002-9576-2176
current ADHD PRS. We can expect, therefore, that with a Tinca J.C. Polderman, PhD: https://orcid.org/0000-0001-5564-301X
larger GWAS of ADHD, a more accurate and predictive The authors wish to thank Prof. Martijn van den Heuvel, PhD, of the Vrije
Universiteit Amsterdam, for his assistance in evaluating the included MRI
PRS will emerge going forward. studies.
Disclosure: Drs. Ronald, Polderman, and Ms. de Bode have reported no
biomedical financial interests or potential conflicts of interest.
Accepted January 28, 2021.
Correspondence to Tinca J.C. Polderman, PhD, Amsterdam UMC, Vrije Uni-
Dr. Ronald is with Birkbeck University of London, United Kingdom. Ms. de versiteit Amsterdam, De Boelelaan 117, Amsterdam, The Netherlands; e-mail:
Bode and Dr. Polderman are with Vrije Universiteit Amsterdam, The tinca.polderman@amsterdamumc.nl
Netherlands, and Amsterdam UMC, The Netherlands.
0890-8567/$36.00/ª2021 American Academy of Child and Adolescent
The authors have reported no funding for this work. Psychiatry. Published by Elsevier Inc. This is an open access article under the
This article is part of a special series devoted to the subject of subject of child CC BY license (http://creativecommons.org/licenses/by/4.0/).
and adolescent attention-deficit/hyperactivity disorder (ADHD). The series https://doi.org/10.1016/j.jaac.2021.01.019
covers a range of topics in the area including genetics, neuroimaging, treat-
ment, and others. The series was edited by Guest Editor Jonathan Posner, MD,
REFERENCES
1. Posner J, Polanczyk GV, Sonuga-Barke E. Attention-deficit hyperactivity disorder. 15. Hosang GM, Lichtenstein P, Ronald A, Lundstr€om S, Taylor MJ. Association of ge-
Lancet. 2020;395:450-462. netic and environmental risks for attention-deficit/hyperactivity disorder with hypo-
2. Polderman TJC, Benyamin B, de Leeuw CA, et al. Meta-analysis of the heritability of manic symptoms in youths. JAMA Psychiatry. 2019;76:1150-1158.
human traits based on fifty years of twin studies. Nat Genet. 2015;47:702-709. 16. Visscher PM, Wray NR, Zhang Q, et al. 10 Years of GWAS discovery: biology,
3. Faraone SV, Larsson H. Genetics of attention deficit hyperactivity disorder. Mol Psy- function, and translation. Am J Hum Genet. 2017;101:5-22.
chiatry. 2019;24:562-575. 17. Demontis D, Walters RK, Martin J, et al. Discovery of the first genome-wide significant
4. Pettersson E, Lichtenstein P, Larsson H, et al. Genetic influences on eight psychiatric risk loci for attention deficit/hyperactivity disorder. Nat Genet. 2019;51:63-75.
disorders based on family data of 4 408 646 full and half-siblings, and genetic data of 18. Bogdan R, Baranger DAA, Agrawal A. Polygenic risk scores in clinical psychology:
333 748 cases and controlseCORRIGENDUM. Psychol Med. 2019;49:351. bridging genomic risk to individual differences. Annu Rev Clin Psychol. 2018;14:
5. Ronald A, Larsson H, Anckars€ater H, Lichtenstein P. Symptoms of autism and ADHD: 119-157.
a Swedish twin study examining their overlap. J Abnorm Psychol. 2014;123:440-451. 19. Ronald A. Editorial: Polygenic scores in child and adolescent psychiatryestrengths,
6. Grove J, Ripke S, Als TD, et al. Identification of common genetic risk variants for weaknesses, opportunities and threats. J Child Psychol Psychiatry. 2020;61:
autism spectrum disorder. Nat Genet. 2019;51:431-444. 519-521.
7. Stergiakouli E, Davey Smith G, Martin J, et al. Shared genetic influences between 20. Wray NR, Lee SH, Mehta D, Vinkhuyzen AAE, Dudbridge F, Middeldorp CM.
dimensional ASD and ADHD symptoms during child and adolescent development. Research review: polygenic methods and their application to psychiatric traits. J Child
Mol Autism. 2017;8:18. Psychol Psychiatry. 2014;55:1068-1087.
8. Taylor MJ, Charman T, Ronald A. Where are the strongest associations between 21. Martin AR, Daly MJ, Robinson EB, Hyman SE, Neale BM. Predicting polygenic risk
autistic traits and traits of ADHD? Evidence from a community-based twin study. Eur of psychiatric disorders. Biol Psychiatry. 2019;86:97-109.
Child Adolesc Psychiatry. 2015;24:1129-1138. 22. World Health Organizaton. ICF Browser. 2017; https://apps.who.int/classifications/
9. Ronald A, Edelson LR, Asherson P, Saudino KJ. Exploring the relationship between icfbrowser/. Accessed July 14, 2020.
autistic-like traits and ADHD behaviors in early childhood: findings from a community 23. World Health Organization. International Statistical Classification of Diseases and
twin study of 2-year-olds. J Abnorm Child Psychol. 2010;38:185-196. Related Health Problems 10th Revision. 2010; https://icd.who.int/browse10/2010/en.
10. Ronald A, Simonoff E, Kuntsi J, Asherson P, Plomin R. Evidence for overlapping Accessed April 6, 2020.
ˇ
genetic influences on autistic and ADHD behaviours in a community twin sample. 24. Hayden JA, Cote P, Bombardier C. Evaluation of the quality of prognosis studies in
J Child Psychol Psychiatry. 2008;49:535-542. systematic reviews. Ann Intern Med. 2006;144:427-437.
ˇ
11. Chang Z, Lichtenstein P, Larsson H. The effects of childhood ADHD symptoms on 25. Hayden JA, van der Windt DA, Cartwright JL, Cote P, Bombardier C. Assessing bias in
early-onset substance use: a Swedish twin study. J Abnorm Child Psychol. 2012;40: studies of prognostic factors. Ann Intern Med. 2013;158:280-286.
425-435. 26. Anderson CA, Pettersson FH, Clarke GM, Cardon LR, Morris AP, Zondervan KT.
12. Liu M, Jiang Y, Wedow R, et al. Association studies of up to 1.2 million individuals Data quality control in genetic case-control association studies. Nat Protoc. 2010;5:
yield new insights into the genetic etiology of tobacco and alcohol use. Nat Genet. 1564-1573.
2019;51:237-244. 27. Kuijpers T, van der Windt DAWM, van der Heijden GJMG, Bouter LM. Systematic
13. Eilertsen EM, Hannigan LJ, McAdams TA, et al. Parental prenatal symptoms of review of prognostic cohort studies on shoulder disorders. Pain. 2004;109:420-431.
depression and offspring symptoms of ADHD: a genetically informed intergenerational 28. Sackett DL, ed. Evidence-Based Medicine: How to Practice and Teach EBM. 2nd ed.
stud [published online ahead of print April 26, 2020]. J Atten Disord; https://doi.org/1 London, UK: Churchill Livingstone; 2001. reprinted.
0.1177/1087054720914386. 29. Albaugh MD, Hudziak James J, Ing A, et al. White matter microstructure is associated
14. Cross-Disorder Group of the Psychiatric Genomics Consortium. genomic relationships, with hyperactive/inattentive symptomatology and polygenic risk for attention-deficit/
novel loci, and pleiotropic mechanisms across eight psychiatric disorders. Cell. 2019; hyperactivity disorder in a population-based sample of adolescents. Neuro-
179:1469-1482. psychopharmacology. 2019;44:1597-1603.

30. Stojanovski S, Felsky D, Viviano JD, et al. Polygenic risk and neural substrates of 56. Szekely E, Schwantes-An T-HL, Justice CM, et al. Genetic associations with
attention-deficit/hyperactivity disorder symptoms in youths with a history of mild childhood brain growth, defined in two longitudinal cohorts. Genet Epidemiol. 2018;
traumatic brain injury. Biol Psychiatry. 2019;85:408-416. 42:405-414.
31. Jansen AG, Dieleman GC, Jansen PR, Verhulst FC, Posthuma D, Polderman TJC. 57. Nigg JT, Gustafsson HC, Karalunas SL, et al. Working memory and vigilance as
Psychiatric polygenic risk scores as predictor for attention deficit/hyperactivity disorder multivariate endophenotypes related to common genetic risk for attention-deficit/hy-
and autism spectrum disorder in a clinical child and adolescent sample. Behav Genet. peractivity disorder. J Am Acad Child Adolesc Psychiatry. 2018;57:175-182.
2020;50:203-212. 58. Hawi Z, Yates H, Pinar A, et al. A case–control genome-wide association study of
32. Li JJ. The positive end of the polygenic score distribution for ADHD: a low risk or a ADHD discovers a novel association with the tenascin R (TNR) gene. Transl Psychi-
protective factor? Psychol Med. 2021;51:102-111. atry. 2018;8:284.
33. Burton CL, Wright L, Shan J, et al. SWAN scale for ADHD trait-based genetic 59. Taylor AE, Jones HJ, Sallis H, et al. Exploring the association of genetic factors with
research: a validity and polygenic risk study. J Child Psychol Psychiatry. 2019;60: participation in the Avon Longitudinal Study of Parents and Children. Int J Epidemiol.
988-997. 2018;47:1207-1216.
34. Gialluisi A, Andlauer TFM, Mirza-Schreiber N, et al. Genome-wide association scan 60. Selzam S, Ritchie SJ, Pingault J-B, Reynolds CA, O’Reilly PF, Plomin R. Comparing
identifies new variants associated with a cognitive predictor of dyslexia. Transl Psy- within- and between-family polygenic score prediction. Am J Hum Genet. 2019;105:
chiatry. 2019;9:77. 351-363.
35. Rietveld CA, Patel PC. ADHD and later-life labor market outcomes in the United 61. Schoeler T, Choi SW, Dudbridge F, et al. Multi–polygenic score approach to identi-
States. Eur J Health Econ. 2019;20:949-967. fying individual vulnerabilities associated with the risk of exposure to bullying. JAMA
36. Piasecki TM, Gizer IR, Slutske WS. Polygenic risk scores for psychiatric disorders reveal Psychiatry. 2019;76:730.
novel clues about the genetics of disordered gambling. Twin Res Hum Genet. 2019;22: 62. Mooney MA, Ryabinin P, Wilmot B, Bhatt P, Mill J, Nigg JT. Large epigenome-wide
283-289. association study of childhood ADHD identifies peripheral DNA methylation associ-
37. Torske T, Nærland T, Bettella F, et al. Autism spectrum disorder polygenic scores are ated with disease and polygenic risk burden. Transl Psychiatry. 2020;10:8.
associated with every day executive function in children admitted for clinical assessment. 63. Sudre G, Frederick J, Sharp W, et al. Mapping associations between polygenic risks for
Autism Res. 2020;13:207-220. childhood neuropsychiatric disorders, symptoms of attention deficit hyperactivity dis-
38. Nigg JT, Karalunas SL, Gustafsson HC, et al. Evaluating chronic emotional dysregu- order, cognition, and the brain. Mol Psychiatry. Published online January 30, 2019;
lation and irritability in relation to ADHD and depression genetic risk in children with https://doi.org/10.1038/s41380-019-0350-3.
ADHD. J Child Psychol Psychiatry. 2020;61:205-214. 64. Serdarevic F, Tiemeier H, Jansen PR, et al. Polygenic risk scores for developmental
39. Dickinson D, Zaidman SR, Giangrande EJ, Eisenberg DP, Gregory MD, Berman KF. disorders, neuromotor functioning during infancy, and autistic traits in childhood. Biol
Distinct polygenic score profiles in schizophrenia subgroups with different trajectories Psychiatry. 2020;87:132-138.
of cognitive development. Am J Psychiatry. 2020;177:298-307. 65. Shen C, Luo Q, Jia T, et al. Neural correlates of the dual-pathway model for adhd in
40. Cabana-Domınguez J, Shivalikanjli A, Fernandez-Castillo N, Cormand B. Genome- adolescents. Am J Psychiatry. Published online May 7, 2020:appi.ajp.2020.1; https://
wide association meta-analysis of cocaine dependence: shared genetics with comorbid doi.org/10.1176/appi.ajp.2020.19020183.
conditions. Prog Neuropsychopharmacol Biol Psychiatry. 2019;94:109667. 66. Hermosillo RJM, Mooney MA, Fezcko E, et al. Polygenic risk score–derived subcortical
41. Ohi K, Nishizawa D, Shimada T, et al. Polygenetic risk scores for major psychiatric connectivity mediates attention-deficit/hyperactivity disorder diagnosis. Biol Psychiatry
disorders among schizophrenia patients, their first-degree relatives, and healthy partic- Cogn Neurosci Neuroimaging. 2020;5:330-341.
ipants. Int J Neuropsychopharmacol. 2020;23:157-164. 67. LaBianca S, LaBianca J, Pagsberg AK, et al. Copy number variants and polygenic risk
42. Mooney MA, Bhatt P, Hermosillo RJM, et al. Smaller total brain volume but not scores predict need of care in autism and/or ADHD families. J Autism Dev Disord.
subcortical structure volume related to common genetic risk for ADHD. Psychol Med. 2021;51:276-285.
Published online January. 2020;24:1-10. https://doi.org/10.1017/ 68. Du Rietz E, Coleman J, Glanville K, Choi SW, O’Reilly PF, Kuntsi J. Associ-
S0033291719004148. ation of polygenic risk for attention-deficit/hyperactivity disorder with co-occur-
43. Vuijk PJ, Martin J, Braaten EB, et al. Translating discoveries in attention-deficit/hy- ring traits and disorders. Biol Psychiatry Cogn Neurosci Neuroimaging. 2018;3:
peractivity disorder genomics to an outpatient child and adolescent psychiatric cohort. J 635-643.
Am Acad Child Adolesc Psychiatry. 2020;59:964-977. 69. Martin J, Taylor MJ, Rydell M, et al. Sex-specific manifestation of genetic risk for
44. Li JJ. Assessing phenotypic and polygenic models of ADHD to identify mechanisms of attention deficit hyperactivity disorder in the general population. J Child Psychol
risk for longitudinal trajectories of externalizing behavior. J Child Psychol Psychiatry. Psychiatry. 2018;59:908-916.
2019;60:1191-1199. 70. Rice F, Riglin L, Thapar AK, et al. Characterizing developmental trajectories and the
45. Riglin L, Eyre O, Thapar AK, et al. Identifying novel types of irritability using a role of neuropsychiatric genetic risk variants in early-onset depression. JAMA Psychi-
developmental genetic approach. Am J Psychiatry. 2019;176:635-642. atry. 2019;76:306-313.
46. Grigoroiu-Serbanescu M, Giaroli G, Thygesen JH, et al. Predictive power of the 71. Zwicker A, MacKenzie LE, Drobinin V, et al. Neurodevelopmental and genetic de-
ADHD GWAS 2019 polygenic risk scores in independent samples of bipolar patients terminants of exposure to adversity among youth at risk for mental illness. J Child
with childhood ADHD. J Affect Disord. 2020;265:651-659. Psychol Psychiatry. 2020;61:536-544.
47. Wimberley T, Agerbo E, Horsdal HT, et al. Genetic liability to ADHD and substance 72. Goodman R, Ford T, Richards H, Gatward R, Meltzer H. The Development and Well-
use disorders in individuals with ADHD. Addiction. 2020;115:1368-1377. Being Assessment: description and initial validation of an integrated assessment of child
48. Riglin L, Thapar AK, Leppert B, et al. Using genetics to examine a general liability to and adolescent psychopathology. J Child Psychol Psychiatry. 2000;41:645-655.
childhood psychopathology. Behav Genet. 2020;50:213-220. 73. Goodman R. The Strengths and Difficulties Questionnaire: a research note. J Child
49. IMAGEN ConsortiumBarker ED, Ing A, et al. Do ADHD-impulsivity and BMI have Psychol Psychiatry. 1997;38:581-586.
shared polygenic and neural correlates? Mol Psychiatry. 2021;26:1019-1028. 74. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disor-
50. de Zeeuw EL, Hottenga J-J, Ouwens KG, et al. Intergenerational transmission of ed- ders. Fifth Edition. Washington, DC: American Psychiatric Association; 2013.
ucation and adhd: effects of parental genotypes. Behav Genet. 2020;50:221-232. 75. American Psychiatric Association, American Psychiatric Association. In: Diagnostic and
51. Yao S, Kuja-Halkola R, Martin J, et al. Associations between attention-deficit/hyper- Statistical Manual of Mental Disorders: DSM-IV-TR. 4th ed., text revision. Wash-
activity disorder and various eating disorders: a Swedish nationwide population study ington, DC: American Psychiatric Association; 2000.
using multiple genetically informative approaches. Biol Psychiatry. 2019;86:577-586. 76. Achenbach TM, Rescorla L. Manual for the ASEBA School-Age Forms & Profiles: An
52. Rabinowitz JA, Musci RJ, Milam AJ, et al. The interplay between externalizing dis- Integrated System of Multi-Informant Assessment. ASEBA, Burlington, VT; 2001.
orders polygenic risk scores and contextual factors on the development of marijuana use 77. Harris KM. The Add Health Study: design and accomplishments. 2013; https://www.
disorders. Drug Alcohol Depend. 2018;191:365-373. cpc.unc.edu/projects/addhealth/documentation/guides/DesignPaperWIIV.pdf.
53. Taylor MJ, Martin J, Lu Y, et al. Association of genetic risk factors for psychiatric Accessed April 6, 2020.
disorders and traits of these disorders in a Swedish population twin sample. JAMA 78. Radloff LS. The CES-D Scale: A self-report depression scale for research in the general
Psychiatry. 2019;76:280. population. Appl Psychol Meas. 1977;1:385-401.
54. Alemany S, Jansen PR, Muetzel RL, et al. Common polygenic variations for psychiatric 79. Cohen S, Kamarck T, Mermelstein R. A Global Measure of Perceived Stress. J Health
disorders and cognition in relation to brain morphology in the general pediatric pop- Soc Behav. 1983;24:385.
ulation. J Am Acad Child Adolesc Psychiatry. 2019;58:600-607. 80. Swanson J, Posner M, Fusella J, Wasdell M, Sommer T, Fan J. Genes and attention
55. Gurriaran X, Rodrıguez-Lopez J, Florez G, et al. Relationships between substance deficit hyperactivity disorder. Curr Psychiatry Rep. 2001;3:92-100.
abuse/dependence and psychiatric disorders based on polygenic scores. Genes Brain 81. Gioia GA, Isquith PK, Guy SC, Kenworthy L. TEST REVIEW Behavior Rating In-
Behav. 2019;18:e12504. ventory of Executive Function. Child Neuropsychol. 2000;6:235-238.

RONALD et al.
82. Constantino JN. Social Responsiveness Scale. In: Volkmar FR, ed. Encyclopedia 105. Fine EM, Delis DC. Delis–Kaplan Executive Functioning System. In: Kreutzer JS,
of Autism Spectrum Disorders. New York: Springer New York; 2013: DeLuca J, Caplan B, eds. Encyclopedia of Clinical Neuropsychology. New York:
2919-2929. Springer New York; 2011:796-801.
83. Simonds J, Rothbard MK. The Temperament in Middle Childhood Questionnaire 106. Wing AM. Voluntary timing and brain function: an information processing approach.
(TMCQ): a computerized self-report instrument for ages 7-10. Published online 2004; Brain Cogn. 2002;48:7-30.
Poster presentation. 107. Reich W. Diagnostic Interview for Children and Adolescents (DICA). J Am Acad Child
84. Stringaris A, Goodman R. Three dimensions of oppositionality in youth. J Child Adolesc Psychiatry. 2000;39:59-66.
Psychol Psychiatry. 2009;50:216-223. 108. Conners CK, Sitarenios G, Parker JD, Epstein JN. The revised Conners’ Parent Rating
85. Wray NR, Ripke S, Mattheisen M, et al. Genome-wide association analyses identify 44 Scale (CPRS-R): factor structure, reliability, and criterion validity. J Abnorm Child
risk variants and refine the genetic architecture of major depression. Nat Genet. 2018; Psychol. 1998;26:257-268.
50:668-681. 109. Ronald A, Sieradzka D, Cardno AG, Haworth CMA, McGuire P, Freeman D.
86. Dumont R, Willis JO. Wide Range Achievement Test–Third Edition. In: Characterization of psychotic experiences in adolescence using the specific psychotic
Reynolds CR, Fletcher-Janzen E, eds. Encyclopedia of Special Education. Hoboken, experiences questionnaire: findings from a study of 5000 16-year-old twins. Schizophr
NJ: John Wiley & Sons; 2008. Bull. 2014;40:868-877.
87. Wechsler D, ed. Wechsler Adult Intelligence Scales. Fourth Edition. San Antonio, TX: 110. Mullins-Sweatt SN, Jamerson JE, Samuel DB, Olson DR, Widiger TA. Psychometric
Pearson; 2008. properties of an abbreviated instrument of the five-factor model. Assessment. 2006;13:
88. Wechsler D. The Wechsler Intelligence Scale for Children–. Fourth Edition. London, 119-137.
UK: The Psychological Corporation; 2003. 111. Ware JE, Sherbourne CD. The MOS 36-Item Short-Form Health Survey (SF-36). I.
89. Wechsler D. Wechsler Individual Achievement Test–. Third Edition. Pearson, New Conceptual framework and item selection. Med Care. 1992;30:473-483.
York, NY; 2009. 112. Wolke D, Woods S, Stanford K, Schulz H. Bullying and victimization of primary
90. Ward MF, Wender PH, Reimherr FW. The Wender Utah Rating Scale: an aid in school children in England and Germany: prevalence and school factors. Br J Psychol.
the retrospective diagnosis of childhood attention deficit hyperactivity disorder 2001;92:673-696.
[published erratum appears in Am J Psychiatry 1993;150:1280]. Am J Psychiatry. 113. Blackwell TL. Test review: Woodcock RW, McGrew KS, Mather N. (2001). Wood-
1993;150:885-890. cock-Johnson III Test. Riverside Publishing Company, 44. Itasca, IL: Rehabil Couns
91. Kaufman J, Birmaher B, Brent D, et al. Schedule for affective disorders and Bull; 2001:232-235.
schizophrenia for school-age children-present and lifetime version (K-SADS-PL): 114. Homack S, Riccio CA. Conners’ Continuous Performance Test (2nd ed; CCPT-II).
initial reliability and validity data. J Am Acad Child Adolesc Psychiatry. 1997;36: J Atten Disord. 2006;9:556-558.
980-988. 115. Touwen B. Examination of the Child with Minor Neurological Dysfunction. Oxford,
92. Skuse DH, Mandy WPL, Scourfield J. Measuring autistic traits: heritability, reliability UK: Heinemann; 1979.
and validity of the Social and Communication Disorders Checklist. Br J Psychiatry. 116. Sahakian BJ, Owen AM. Computerized assessment in neuropsychiatry using
2005;187:568-572. CANTAB: discussion paper. J R Soc Med. 1992;85:399-402.
93. Patton JH, Stanford MS, Barratt ES. Factor structure of the Barratt Impulsiveness Scale. 117. Kirby KN, Petry NM, Bickel WK. Heroin addicts have higher discount rates for
J Clin Psychol. 1995;51:768-774. delayed rewards than non-drug-using controls. J Exp Psychol Gen. 1999;128:78-87.
94. Cito. Centrale Eindtoets. 2020; https://www.cito.nl/onderwijs/primair-onderwijs/ 118. D’Alberto N, Chaarani B, Orr CA, et al. Individual differences in stop-related activity
centrale-eindtoets. Accessed April 6, 2020. are inflated by the adaptive algorithm in the stop signal task. Hum Brain Mapp. 2018;
95. Garner DM. Eating Disorder Inventory2: Professional Manual. Odessa, FL: Psy- 39:3263-3276.
chological Assessment Resources; 1991. 119. Eysenck H, Eysenck S. Manual for the Eysenck Personality Questionnaire (EPQ-R
96. Kessler RC. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in Adult). London, UK: Hodder & Stoughton; 1994.
the United States: results from the National Comorbidity Survey. Arch Gen Psychiatry. 120. Thabrew H, Stasiak K, Bavin L-M, Frampton C, Merry S. Validation of the Mood and
1994;51:8. Feelings Questionnaire (MFQ) and Short Mood and Feelings Questionnaire (SMFQ)
97. Substance Abuse and Mental Health Administration (SAMHSA). 2007 National in New Zealand help-seeking adolescents. Int J Methods Psychiatr Res. 2018;27:e1610.
Survey on Drug Use and Health Questionnaire. 2001; http://datafiles.samhsa.gov/ 121. Wechsler D. Wechsler Abbreviated Scale of Intelligence. London, UK: The Psycho-
study/national-survey-drug-use-and-health-nsduh-2007-nid1353. Accessed April logical Corporation; 1999.
15, 2020. 122. Kong A, Thorleifsson G, Frigge ML, et al. The nature of nurture: effects of parental
98. Capaldi DM, Patterson GR. Psychometric Properties of Fourteen Latent Constructs genotypes. Science. 2018;359:424-428.
from the Oregon Youth Study. 1989; https://doi.org/10.1007/978-1-4612-3562-0. 123. Sullivan PF. Spurious genetic associations. Biol Psychiatry. 2007;61:1121-1126.
Accessed April 15, 2020. 124. Button KS, Ioannidis JPA, Mokrysz C, et al. Power failure: why small sample size
99. United States Census Bureau. American Community Survey 5-year estimates: 2005. undermines the reliability of neuroscience. Nat Rev Neurosci. 2013;14:365-376.
2009; http://www.census.gov/programssurveys/acs/data/summary-file. 2009.html. 125. Dudbridge F. Power and predictive accuracy of polygenic risk scores. PLoS Genet.
Accessed April 15, 2020. 2013;9:e1003348.
100. Hansson SL, Svanstr€omr€ojvall A, Rastam M, Gillberg C, Gillberg C, Anckars€ater H. 126. Brikell I, Larsson H, Lu Y, et al. The contribution of common genetic risk variants for
Psychiatric telephone interview with parents for screening of childhood autismetics, ADHD to a general factor of childhood psychopathology. Mol Psychiatry. 2020;25:
attention-deficit hyperactivity disorder and other comorbidities (A-TAC): preliminary 1809-1821.
reliability and validity. Br J Psychiatry. 2005;187:262-267. 127. Pain O, Glanville KP, Hagenaars S, et al. Evaluation of polygenic prediction meth-
101. Nigg JT. The ADHD response-inhibition deficit as measured by the Stop Task: odology within a reference-standardized framework. bioRxiv. 2020. https://doi.org/10.
replication with DSM-IV combined type, extension, and qualification. J Abnorm Child 1101/2020.07.28.224782.
Psychol. 1999;27:393-402. 128. Peterson RE, Kuchenbaecker K, Walters RK, et al. Genome-wide association studies in
102. Keilp JG, Herrera J, Stritzke P, Cornblatt BA. The continuous performance ancestrally diverse populations: opportunities, methods, pitfalls, and recommendations.
test, identical pairs version (CPT-IP): III: brain functioning during perfor- Cell. 2019;179:589-603.
mance of numbers and shapes subtasks. Psychiatry Res Neuroimaging. 1997; 129. Folkersen L, Pain O, Ingasson A, Werge T, Lewis CM, Austin J. Impute.Me: an open
74:35-45. source, non-profit tool for using data from DTC genetic testing to calculate and
103. De Luca CR, Wood SJ, Anderson V, et al. Normative data from the Cantab. I: interpret polygenic risk scores. Genetics. 2019. https://doi.org/10.1101/861831.
development of executive function over the lifespan. J Clin Exp Neuropsychol. 2003; 130. Austin JC. Evidence-based genetic counseling for psychiatric disorders: a road map.
25:242-254. Cold Spring Harb Perspect Med. 2020;10:a036608.
104. Climie EA, Rostad K. Test review: Wechsler Adult Intelligence Scale. J Psychoeduc 131. Lewis CM, Vassos E. Polygenic risk scores: from research tools to clinical instruments.
Assess. 2011;29:581-586. Genome Med. 2020;12:44.


ADHD PRS Review

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ADHD PRS Review

Uploaded by

Copyright:

Available Formats

REVIEW

Systematic Review: How the Attention-Deﬁcit/

Journal of the American Academy of Child & Adolescent Psychiatry www.jaacap.org 1

2 www.jaacap.org Journal of the American Academy of Child & Adolescent Psychiatry

FIGURE 1 PRISMA Flow Diagram of Study Selection

Records idenﬁed through Addional records idenﬁed in June

Records aer duplicates removed

Records screened Records excluded

Full-text arcles excluded,

Full-text arcles assessed with reasons

Journal of the American Academy of Child & Adolescent Psychiatry www.jaacap.org 3

4 www.jaacap.org Journal of the American Academy of Child & Adolescent Psychiatry

inferior longitudinal fasciculi.

ADHD POLYGENIC RISK SCORES

(ASD only and ADHD only Planned sensitivity analyses

Scale.79 Also, it was asked stress.

ADHD POLYGENIC RISK SCORES

higher when comparing groups

ADHD POLYGENIC RISK SCORES

East Asian Ancestry. sporting events, or taken part in

subscales: initiate, working

compared different group suggested changes in latent

ADHD POLYGENIC RISK SCORES

N [ 4287 controls SNP threshold of p < 5.7eL04

n [ 56 unaffected ﬁrst-degree Signiﬁcance threshold of p < .01

ADHD POLYGENIC RISK SCORES

disorder history. for ADHD and stimulant use did

college completion (OR [ 1.23,

psychopathology compared to the

ADHD PRS p Outcome measures and

structured research diagnostic compared with increasing

ADHD POLYGENIC RISK SCORES

This association was found to be

e3.97), z [3.36, p [ .0008, FDR-

Signiﬁcance threshold was

ADHD POLYGENIC RISK SCORES

No covariates age 13 (B [ 0.095, SE [ 0.019, p <

anticipation and reward PRS associated via the

Barratt Impulsivity Scale (BIS)93

ADHD POLYGENIC RISK SCORES

Results were consistent at other pT;

parental monitoring (proximal disadvantage or parental

prior: b [ L0.07, SE [ 0.03,

ADHD POLYGENIC RISK SCORES

dependence patients (mean Results were similar when MHC was

4.72. risk for ADHD and brain growth

cerebellum, cerebral white

response inhibition, executive accounted for by the latent

ADHD POLYGENIC RISK SCORES

clinics completed and or clinics (ES [ L2.14, 95% CI [ L2.63 to

European ancestry Educational attainments based effect, signiﬁcantly predicted more

ADHD POLYGENIC RISK SCORES

impairment, clinician burden in ADHD cases. The probe

ADHD PRS p Outcome measures and

Other disorders in adults were genetic effect); for thickness, a

ADHD POLYGENIC RISK SCORES

measured using the Conners’ between ADHD PRS and

analyses to 1 member of each

y versions used for 9- to 15-wk- (b [0.43, CI [ 0.001L0.06;

ADHD POLYGENIC RISK SCORES

MRI-based resting functional cluster of regions in the right

parietal cortex) was included in the