Background

ADHD is a neurodevelopmental disorder characterized by symptoms of inattention, hyperactivity and impulsiveness that lead to impairments of everyday life. According to current diagnostic criteria, a diagnosis of ADHD requires symptom manifestations before the age of 12 years, and that symptoms are not better accounted for by other mental disorders. ^, Although some inattention, hyperactivity, or impulsiveness can be experienced by most people during their lifetime, the prevalence of the clinical ADHD disorder is much lower and relatively similar across different countries, that is, 5.3% in children and 2.5% in early adulthood. The essential features of the clinical syndrome(s) currently recognized as ADHD were described more than a century ago. It was early observed that ADHD may run in families, indicating some shared environmental or genetic risk. Since these early descriptions, the role of nature versus nurture in ADHD has been debated in scientific literature and popular media and many attempts have followed to identify specific genetic and environmental risk factors and biological mechanisms.

An Introduction to Genetics and Genomics

Genetics is usually defined as studies of how characteristics can be inherited. This involves investigations of genes, genetic variations, and mechanisms of inheritance. The fundamental principles of inheritance were proposed by Gregor Mendel in 1865. Forty years later, the Danish botanist Wilhelm Johansson introduced the term “gene” to describe the Mendelian units of heredity, “genotype” as the genetic material in an organisms and “phenotype” as the observable characteristics in the organism. Alternative forms of genetic units are termed alleles. For every trait, humans inherit 2 alleles, one from each parent. A person who has 2 identical alleles for a gene is said to be homozygous for that gene. The genome is defined as the complete set of genetic information in an organism, and from the late 1980s, the term “genomics” was introduced to describe (simultaneous) studies of all genes in an organism. In his investigations of binary morphologic characteristics, Mendel discovered certain principles that apply to different types of inheritance. However, today the concept “Mendelian” diseases or traits are used to describe conditions where their inheritance clearly follows Mendel’s principles, that is, the trait depends on a single locus, and where the alleles are either dominant or recessive.

ADHD symptoms have been observed in rare Mendelian inherited neurometabolic diseases and other rare genetic conditions, such as Fragile X syndrome, neurofibromatosis type 1, DiGeorge syndrome, tuberous sclerosis complex, Turner syndrome, Williams syndrome, and Klinefelter syndrome. However, most ADHD cases are not related to such syndromes. Like other common psychiatric and somatic disorders, personality characteristics and other observable traits, nonsyndromic ADHD is considered a complex disorder and subject to complex heredity. Many types of DNA variants can cause disease. These include chromosomal aberrations, structural DNA variants such as copy number variants (CNVs) and single nucleotide variants (SNVs). As every nucleotide the genome may exist in 4 versions (A, T, G, or C), there are billions of naturally occurring SNVs in the human genome, distributed among the earth’s 8 billion people. More than 99% of the nucleotides in the human genome are nevertheless common to all people, while approximately every 100 to 300 nucleotide are different. The most common SNVs (>1% allele frequency) are termed single nucleotide polymorphisms (SNPs).

Heritability of Attention Deficit Hyperactivity Disorder

Classic (quantitative) genetic methods, such as family studies, adoption studies, and twin studies have demonstrated a familial risk of ADHD and that this familial risk of ADHD is mainly caused by genetic factors, rather than shared environments. ^, Twin studies make use of the fact that monozygotic twins are nearly 100% genetically identical, while dizygotic twins on average share 50% of their segregating genes (alleles). In a comparison of 37 different twin studies on ADHD or ADHD symptoms, the mean heritability across all studies was 74% and 77% to 88% when only studies using categorical measures of ADHD were included. This shows that heritability estimates vary depending on whether ADHD is measured as a categorical construct or using symptom counting. Heritability measures of ADHD also depend on whether symptoms are rated by parents and teachers or using self-rating of symptoms. Same-teacher and same-parent ratings tend to give higher heritability estimates (70%–80%) than self-ratings and different parent and different teacher ratings (30%–40%). ^, This also explains why heritability estimates in adults that used self-reporting of symptoms gave lower heritability (30%–40%) than studies using composite ratings from several sources (72%–80%). The twin-based heritability of ADHD is similar to other psychiatric disorders such as schizophrenia, bipolar disorder, and autism spectrum disorder (70%–80%), but higher than anxiety disorders, major depression, and personality traits (typically 30%–50%) ^, However, it is important to emphasize that heritability values are not suited to predict individual risk. Such estimates are always measured at a group level, typically in one geographic area and at a particular time. If the environment is altered, heritability may also change. Twin studies also show that genetic predisposition to ADHD interacts with environmental influences to explain the variability in hyperactivity and inattention symptoms.

Recurrence Risk of Attention Deficit Hyperactivity Disorder

Persons with ADHD and prospective parents may be concerned about recurrence of this condition in future generations. Given the well-established twin-based heritability estimates of ADHD (70%–80%), this percentage has also been interpreted as an estimate of familial recurrence risk. Such misconceptions have been propagated by popular media and “influencers” and may add to the burden of concern of affected persons and families. The recurrence risk of any condition is determined by its population prevalence, as well as heritability and environmental factors. In a register-based study of 2.5 million Norwegians born 1967 to 2011 and their parents, the overall prevalence of ADHD was 2.4% in adults and 3.3% in children and adolescents. Among mothers with ADHD, 18% of daughters and 30% of sons had been diagnosed with ADHD. In comparison, only 13% of daughters and 22% of sons of men with ADHD had ADHD. The highest recurrence risk was observed when both parents had ADHD (25% of daughters and 41% of boys). Offspring of parents with bipolar disorder, major depression, and other psychiatric also have an increased risk of ADHD and other neurodevelopmental disorders. The high recurrence risk of ADHD is explained by shared genetic and environmental factors, and probably some confirmation bias related to the fact that it may be more likely to consider this condition if ADHD already has been diagnosed in a close relative. Still, it is important to emphasize that the observed recurrence rates are much lower than the reported heritability of ADHD.

The unequal risk of transmitting a condition to the offspring from affected fathers or mothers, as indicated in the study cited earlier, is indicative of a parent-of-origin effect. Such effects are observed in many neurodevelopmental disorders and may result from genomic imprinting, specific effects of the maternal genome on the intrauterine microenvironment and fetus (maternal effects), and mitochondrial genome and sex chromosome effects. The intrauterine microenvironment is influenced by environmental and dietary exposures, as well as the maternal genotype. Such genetic differences can affect prenatal development and long-term health of the offspring. For instance, it has been observed that rare coding variants in mothers that interfere with serotonin production are associated with an increased risk of ADHD in the offspring. Using genomic data from pedigrees with ADHD and statistical tools, it is possible to calculate the contribution from the different parent-of-origin effects to ADHD genetic risk. ^,

Early Molecular Genetic Studies

As mentioned earlier, ADHD symptoms may be observed in rare genetic syndromes with Mendelian patterns of inheritance. Linkage studies are commonly used to identify the chromosomal regions (loci), causative genes, and genetic variants in families suffering from highly penetrant genetic diseases. However, most linkage studies in families with ADHD provided inconclusive results, probably due to the polygenic nature of nonsyndromic ADHD. One notable exception is a linkage study performed in multigenerational families in a population isolate in Colombia, which showed tentative linkage to chromosomal regions 4q13.2, 5q33.3, 11q22, and 17p11.2. In combination with case–control studies, this led to the identification of the LPHN3 risk gene at the 4q13.2 locus. LPHN3 encodes an adhesion G-protein coupled receptor (GPCR) that is expressed in many brain regions and may be involved in memory, attention, and activity functions. The latrophilin proteins (LPHN1–3) are also termed adhesion GPCRs L1 to 3. Such receptor complexes are potential drug targets for cardiovascular, immune, neuropsychiatric, and neoplastic diseases.

Candidate Gene Studies

ADHD symptoms are routinely treated with stimulant drugs that increase brain levels of the catecholamines dopamine and norepinephrine (noradrenaline), or the alpha-2A adrenergic agonists guanfacine and clonidine. Some animal models used in ADHD research show dopamine-related dysfunctions and certain brain regions suspected to be involved in ADHD-related traits contain catecholaminergic synaptic terminals. Thus, it has long been speculated that ADHD symptoms could be related to genetic variants affecting catecholamine signaling. ^,

Candidate gene studies have explored variants in monoamine-related genes, including dopamine receptors (DRD1–5), subtypes of α and β-adrenergic receptors and the transporters of dopamine (DAT, encoded by the SLC6A3 gene [also known as DAT1 ]) and the norepinephrine transporter, encoded by SLC6A2. ^{, ,} Among the dopamine-related genes, DRD4 has received much attention. This receptor has an intracellular domain of variable size, encoded by a 48 base pair variable number of tandem repeats (VNTRs). Two to 11 repeats are found in different frequencies in different human populations. Some studies have indicated an association of the 7 repeat variant (or other long alleles) and an increased risk of obsessive-compulsive disorder, Tourette’s syndrome, substance use disorder, schizophrenia, autism, ADHD, as well as certain personality traits. Early suggestions that the 7 repeat allele of the DRD4 VNTR has been subject to positive selection were not confirmed in later studies. ^, A recent meta-analysis of allelic variants in the DRD4 receptor, including the exon 3 VNTR, showed some evidence of association with ADHD cases of European-Caucasian descent, despite negative findings in well powered genome-wide association (GWA) studies. The authors argued that since the DRD4 VNTRs were poorly tagged (not in linkage disequilibrium) with SNPs typically included in GWA studies (GWAS), it is difficult to exclude an association with ADHD. Common genetic variants in the DRD2 locus have also been associated with ADHD, but even stronger with other neuropsychiatric disorders and personality traits (see later discussion). The short form of DRD2 receptors are presynaptic auto receptors that can modulate neurotransmitter synthesis and release and are established targets of antipsychotic drugs that influence many brain functions. This diversity of neurobiological effects may be consistent with the pleiotropic effects of common variants in this receptor. In addition to the monoamine neurotransmitter receptor and transporter genes mentioned earlier, dozens of other proteins are implicated in the synthesis, metabolism, and signal transduction of dopamine and related monoamines. In a comprehensive analysis, no association was observed between genetic variants in core dopamine-related genes and ADHD, but an association was reported when many other monoamine-related variants were included in the analysis. In conclusion, recent molecular genetic studies suggest that ADHD is rarely caused by alterations in dopamine-related genes.

Rare Mendelian syndromes provide information about the consequences (phenotypes) of dopamine or norepinephrine deficiency in humans. The autosomal recessive neurologic syndrome tyrosine hydroxylase deficiency is caused by inactivating mutations in the rate-limiting enzyme in dopamine production. These patients have extremely low brain levels of dopamine, severe motor problems from an early age that are responsive to l -DOPA replacement therapy, but no obvious psychiatric phenotype. Likewise, persons with loss-of-function variants in the enzyme dopamine β-hydroxylase that converts dopamine into norepinephrine have low levels of epinephrine and orthostatic hypotension, but normal cognitive development. Thus, despite popular media claims that brains of persons with ADHD are “lacking” dopamine or norepinephrine, there is little evidence to support this belief.

Neurometabolic Disorders and Attention Deficit Hyperactivity Disorder

Some neurometabolic disorders are known to interfere with neurotransmitter metabolism. Phenylketonuria and tyrosinemia are inborn errors of metabolism characterized by a defective catabolism of phenylalanine and tyrosine, and extremely elevated tissue levels of these amino acids. A high prevalence of ADHD has been observed in both conditions. Biochemical analyses and cellular models indicate that dopamine synthesis may be impaired in these conditions, possibly linking ADHD symptoms to altered neurotransmitter functions. Other early molecular genetic studies in ADHD examined atypical forms like ADHD with generalized resistance to thyroid hormone (GRTH). GRTH features decreased responsiveness to thyroid hormone, elevated T3 and T4 levels, and ADHD-like cognitive symptoms. Mutations in the thyroid receptor gene ( THRB ) have been found in some patients with ADHD, but due to the rarity of GRTH, such mutations account for a very small proportion of sporadic ADHD cases. Since some metabolic disorders can lead to irreversible and life-threatening disease if not recognized and treated at an early age, they are included in newborn screening programs in many countries. As screening programs are expanded and whole exome sequencing (WES) or whole genome DNA sequencing becomes routine, more potentially treatable genetic conditions can be detected at an early age. As ADHD symptoms may be among the first clinical manifestations of genetic syndromes, and many conditions may be treated if diagnosed early, clinicians should be able to recognize such cases among children and adults with atypical ADHD symptoms. ^, Such “red flags” include atypical or progressive symptoms, a combination of neurologic/somatic and psychiatric symptoms, including intellectual disability and ADHD and a family history of similar problems.

Genome-Wide Association Studies

GWA studies (GWAS) compare allele frequencies in different groups, for example, persons with ADHD and control persons without ADHD. In contrast to candidate gene studies focusing on a few or a single predefined gene or genetic variants, GWAS are exploring many genetic markers (SNPs) spanning all chromosomes, theoretically capturing genetic variation across the whole genome. Microarray chips used for GWAS are usually designed to analyze around 1 million different allelic variants, or roughly 1 SNP for every 3000 base pairs of DNA, typically present in greater than 1% of the population. Comparison of many different markers has a statistical downside: to assert genome-wide statistical significance, it is usually considered that associations should have P -values less than .00000005 (5 x 10 ⁻⁸ ). For complex conditions with many contributing genetic variants of small effects, this stringent P value requires very large samples. Obviously, the application of this conservative, but still arbitrary significance threshold is no guarantee against chance findings, or that associations with P >10 ⁻⁸ are not reflecting “real” differences in allele frequencies.

Most genetic markers detected by GWAS are intergenic or located in introns, but assuming that functional variants are in LD with the SNP genetic markers, GWAS can direct attention to possible risk genes and coding or regulatory risk variants. Results from GWAS can also be used to study CNVs and by imputation rare variants that are in LD with common variants. However, identification of causative genetic variants, genes, and biological mechanisms responsible for the GWAS findings is not straightforward, as the causative variants may be physically distant from tagging SNPs and recently emerged, rare variants are poorly captured is such studies.

The first GWAS of ADHD had modest samples sizes and were inconclusive. However, in a meta-analysis published in 2019 combining 12 studies with 20,183 children and adults with ADHD and 35,191 controls, 12 loci reached genome-wide significance ( P < 5 x 10 ⁻⁸ ). In an updated analysis from 2023 with 38,691 individuals with ADHD and 186,843 controls, 27 genome-wide significant loci were reported, implicating 76 possible risk genes. Most cases and controls were from Denmark and Iceland, supplemented with 10 smaller ADHD cohorts with European ancestry, collected by the Psychiatric Genomics Consortium. Based on these GWAS findings, the authors concluded that ADHD is very polygenic, implicating thousands of risk variants, of which the majority also influence other psychiatric disorders with concordant or discordant effects. All 12 loci implicated in the 2019 analysis remained significant ( P -values <8 × 10 ⁻⁴ ) in the new analysis, indicating some robustness of the initial findings, at least in European populations.

Most of the GWAS signals in the 2023 ADHD meta-analysis were found deep in noncoding intergenic and intronic chromosomal regions and many of the implicated loci, genes, and genetic variants in this GWAS have also been reported in other psychiatric disorders. For most of the 27 significantly associated signals, it was possible to identify credible functional variants and implicated risk genes, typically encoding transcription factors and regulatory proteins involved in synaptic functions, neuronal migration, and differentiation. For example, the implicated transcription factor genes FOXP1 and FOXP2 regulate expression of many different proteins in brain tissues. Highly penetrant variants in these genes have been implicated in rare neurodevelopmental diseases. MEF2C belongs to the MEF2 family of transcription factors and is involved in neuronal differentiation, proliferation, survival, and synapse formation. MEF2C-related disorders include neurodevelopmental syndromes with developmental delay, intellectual disability, lack of verbal language, motor symptoms, and seizures. SORCS3 encodes a receptor implicated in control of neuronal viability and function and is highly expressed in the CA1 region of the hippocampus. Altogether, these 76 nominated genes represent multiple biological functions involved in neuronal cell growth, migration, communication, and development. By comparing the ADHD-associated genetic variants to databases of gene expression patterns, it was observed that many of the ADHD risk genes are upregulated during early embryonic brain development and are enriched in brain-specific neuronal subtypes and midbrain dopaminergic neurons. Altered functions of transcription factors and other genes involved in maturation of the nervous system may explain why many children with ADHD have a delayed maturation of cerebral cortex and reduced volume of some cortical and subcortical brain regions.

GWAS indicates that ADHD is very polygenic, with thousands of risk variants, most of which are also implicated in other psychiatric disorders. From a collection of 772 different GWAS results, including 514 phenotypes from the UK Biobank project, ADHD showed a significant genetic correlation ( r _g ) with 56 phenotypes, including negative correlations with educational attainment and longevity and positive correlations with body mass index, insomnia, and psychiatric disorders such as ASD, schizophrenia, major depressive disorder, and cannabis use disorder. Although common variants (SNPs) in the dopamine and norepinephrine receptor and transporters were not among the most prominent findings in recent GWAS studies, rare variants and VNTRs in these genes might be implicated in ADHD susceptibility. As discussed earlier, more studies are needed to fully evaluate the full spectrum of genetic variation in these neurotransmitter systems.

Polygenic Risk Scores

PRS, or simply polygenic scores, may be defined as quantitative measures of an individual’s genetic liability for a certain trait. PRS are typically calculated by summing up common risk alleles weighed by their effect sizes as provided by a GWAS. Published summary statistics from large GWAS data sets can be used for this purpose, eliminating the need to share large amounts of sensitive data. PRS have several potential clinical applications, such as in screening for disease, risk stratification, treatment optimization, and disease prognosis prediction. PRS can also be used to compare genetic relationships among different traits and to disentangle environmental and genetic risk factors. PRS obtained from GWAS on clinically diagnosed ADHD cases predict attention problems in the general population, indicating that these are correlated traits, where the clinical condition is at an extreme end of a continuous distribution. Using data from population based cohorts, such as the UK Biobank and the Estonian Biobank, ADHD PRS have been shown to correlate with many common traits and conditions, such as anxiety, depression, risk-taking, alcohol intake, chronic obstructive pulmonary disease, obesity, and type 2 diabetes. ^, However, as these ADHD PRS are mainly based on GWAS data from White Europeans, predictions of health outcomes are more precise among participants of European descent than in persons with African ancestry. GWAS and PRS have also been used to study the relationship between childhood, adult persistence, and late onset ADHD. GWAS findings in children and adults with ADHD showed very similar results, ^, although the PRS for ADHD was higher in persistent ADHD compared with the late-diagnosed ADHD. ^, A subgroup of ADHD cases appears to have adult-onset clinical ADHD. However, it is possible that many of these persons have early vulnerabilities and that their subclinical symptoms convert to a clinical presentation when facing challenges later in life.

In many fields of medicine, including psychiatry, it has been suggested that PRS can be used to supplement existing diagnostic procedures. ADHD PRS predicts clinical ADHD and ASD in clinical child and adolescent target samples and can reveal clinical heterogeneity within groups of ADHD cases. ADHD PRS might also be used for screening purposes, as it provides independent information to ADHD rating scales and family history of ADHD to distinguish between cases and population controls. However, the variance explained by the ADHD PRS is low, illustrating that currently available PRS have limited diagnostic utility. The patients and protocols used in research settings may also not be representative of routine clinical practice. Thus, the practical value of PRS ultimately needs to be formally demonstrated in clinical representative trials.

Rare Variants

CNVs are common structural DNA variants present in all humans, but rare, large (>100 kB) CNVs are often associated with altered brain structure and function and psychiatric disorders. Large CNVs can encompass many genes, result in reduced fertility (fecundity) and are, therefore, often observed in single individuals. One can investigate how such genetic variants arise by comparing genotypes of carriers and their parents. Some CNV duplications are functional in the sense that they contain functional genes, resulting in an increased expression of a protein encoded in the region in question. Depending on which part of the genome is involved, DNA variations can give rise to variable clinical pictures, ranging from no symptoms to severe disease or lethality, and they may also result in miscarriage. Several studies have documented an increased prevalence of CNV duplications and deletions in cases with ADHD, schizophrenia, ASD, and other neuropsychiatric disorders, showing that many CNVs are enriched in genes implicated in neurodevelopmental genes that also are related to brain development and structural brain alterations. In a study of Icelandic and Norwegian ADHD cases and controls, 8 CNVs were significantly enriched in the ADHD cases: deletions at 2p16.3, 15q11.2, 15q13.3 and 22q11.21, and duplications at 1q21.1 distal, 16p11.2 proximal, 16p13.11, and 22q11.21. When analyzed together, 19 CNVs were associated with ADHD, even when comorbid ASD and schizophrenia were excluded from the sample.

WES and exome chip genotyping have shown that rare coding variants are associated with ADHD in case–control studies and family studies. When GWAS and WES data were combined, some rare and common variants converged on the same genes and biological pathways, demonstrating a spectrum of different genetic risk factors with different population frequencies and effect sizes, for example, SORCS3 risk gene was implicated in ADHD by both common and rare variants.

People do not Inherit DSM-5 Disorders

Psychiatric diagnoses are artificial constructs designed to capture important common features of human symptoms and behaviors. Criteria listed in diagnostic manuals, such as the Diagnostic and Statistical Manual of Mental Disorders (DSM) are selected based on practical utility and reliability, not to capture discrete biological entities, subjective experiences, or individual variation. Every clinician knows that there are fuzzy borders between diagnostic categories and that patients may fulfill criteria for multiple disorders. Not surprisingly, this comorbidity is also reflected in the genetic findings. The ADHD genetic findings discussed earlier show strong overlap with multiple psychiatric disorders, somatic conditions, and personality traits. Genetic findings indicate that people do not inherit DSM-5 disorders, but rather certain traits that in combination with specific environmental exposures increase the likelihood of being diagnosed with ADHD. ^{, , ,} A high novelty seeking or impulsivity trait might be beneficial in a context of creating original art or being a successful entrepreneur. However, novelty and sensation seeking might also lead to risk taking or behaving antisocial, testing illegal drugs or traffic speeding, as observed in many clinical ADHD studies.

Pharmacogenomics and Genetic Testing

An important motivation for performing genetic studies is to study disease mechanisms and targets for diagnosis, prevention, and treatment. Stimulants and other medications targeting catecholaminergic neurotransmission are the most widely used and best documented treatment options for ADHD. ^, Genetic studies on ADHD have been systematically explored to find druggable targets, showing that many established and potentially new druggable targets are among the associated loci and candidate genes. More recently, machine learning algorithms have been developed that can explore large data sets and systematically search for druggable targets.

Most of the credible genetic findings in ADHD as reviewed here come from GWA studies and targeted genotyping. However, as molecular genetic technology has evolved and costs have plummeted, WES, whole genome sequencing, as well as genome wide DNA methylation studies are rapidly being introduced into genetic research and clinical testing. Some clinicians and consumer directed companies advocate routine whole genome sequencing for clinical, and recreational purposes. Whole genome sequence data can easily be linked to published genetic studies to calculate individual PRS and provide statistical predictions for genetic risk (eg, percentiles) of rare diseases and common traits and conditions, such as ADHD. For Mendelian diseases, with well-documented risk variants, such predictions can be fairly accurate. However, for complex, polygenic traits, such as ADHD, these statistical predictions, for example, using PRS, are of little value. Predictions based on GWAS typically only estimate part of the genetic component to the total risk of disease. As discussed earlier, an individual’s total risk of ADHD is a function of genetic risk, environmental factors, and lifetime stochastic events. Although genetic testing may provide essential and actionable information, it also raises important ethical and legal issues. Genetic data are not private. For example, genetic information also reveals private genetic information about relatives, who may not have consented to the sharing of such data. For this reason, customer-directed genetic testing is forbidden by law in many countries.