How do ADHD and autism overlap?

For years, diagnostic rules made the research question harder than it needed to be. DSM-IV did not allow an ADHD diagnosis when autism was present. Studies often recruited an autism group or an ADHD group, excluded people with both, and then tried to explain the similarity between the groups.

DSM-5 removed that exclusion in 2013. The change did not create AuDHD. It allowed clinical records to catch up with people whose attention, activity, sensory, communication, and repetitive-pattern traits had never respected the boundary.

“AuDHD” is not a third formal diagnosis. It is community shorthand for having both. The research question underneath it has several layers: how often the diagnoses co-occur, which trait dimensions overlap, whether some underlying influences are shared, and what changes when both patterns are present in one life.

Rommelse and colleagues’ 2010 review gathered the family, twin, candidate-gene, linkage, and early genome-wide evidence available while the diagnoses were still officially treated as mutually exclusive. Family and twin studies repeatedly found that ADHD traits and autistic traits travelled together more than chance would predict.

One young-adult twin study in the review included 674 people and estimated a genetic correlation of .72 between self-reported ADHD symptoms and autistic traits. Another large child sample found that shared inherited influences varied across trait combinations. The review’s durable proposal was not that a single gene or brain pathway explained both. It was that the overlap was partly familial and dimensional, and that the diagnostic wall was scientifically misleading.

Its molecular findings were early and inconsistent. Lists of candidate genes from that era should not be read as a current explanation for one person. The family signal survived better than the specific-gene story.

As adult datasets grew, researchers could ask about dimensions rather than only labels. Ghirardi and colleagues studied 6,866 Swedish twins aged 20 to 28. Inattention was similarly associated with repetitive/restricted traits and with social-communication traits. Hyperactivity/impulsivity was more strongly associated with repetitive/restricted traits than with social-communication traits.

Genetic and non-shared environmental components each accounted for substantial portions of those correlations. The authors also warned that “non-shared environment” in a twin model includes measurement error. The result is not a recipe dividing AuDHD into a percentage genetic and a percentage environmental. It shows that the amount and source of overlap depend on which traits are compared.

The study used self-report traits in a population sample, not clinical diagnoses. People with more severe difficulties were less likely to respond. It is strong evidence about covariance in young adults and weaker evidence about the experience of clinically diagnosed AuDHD adults.

Two different kinds of adult evidence made co-occurrence harder to dismiss. First, Solberg and colleagues linked nationwide Norwegian registers. The dataset included 38,636 adults recorded with ADHD, 7,528 with autism, 1,467 with both, and more than 1.6 million in the remaining population.

Adults with both diagnoses were visible at population scale, not only in specialist clinics. Their patterns of additional psychiatric diagnoses were often more pronounced than those of either single-diagnosis group. Yet the ADHD and autism groups also had different patterns: for example, substance-use diagnoses were relatively more elevated with ADHD, while schizophrenia-spectrum diagnoses were relatively more elevated with autism. Shared influences had not erased distinction.

Second, Krakowski and colleagues’ 2021 scoping review looked at the hidden structure of combined ADHD and autistic symptoms. Four of five factor-analysis studies found separate ADHD and autism dimensions. All three latent-class studies found groups of people with high levels of both.

That apparent tension is useful. At the variable level, the two sets of traits often separate. At the person level, the traits often co-occur. AuDHD does not require every ADHD trait to blend with every autistic trait.

Recent health-record evidence adds scale but needs a firm boundary around interpretation. In 2025, Yerys and colleagues analysed US Medicaid claims for more than 3.5 million adults from 2008 to 2019. ADHD was recorded for 26.7% of autistic adults without intellectual disability and 40.2% of autistic adults with intellectual disability, compared with 2.7% in the general Medicaid comparison group.

Those are rates in Medicaid records, not universal prevalence estimates. Insurance eligibility, access to clinicians, coding practice, and the study’s algorithms all shape who appears in each group. Claims can show that diagnosed co-occurrence is common and consequential in a large US health system. They cannot show how many autistic adults have unrecognised ADHD, how traits feel, or whether medication caused the better outcomes associated with prescriptions.

Together, the later work also challenges a simple severity ladder. People with both diagnoses may face more barriers or co-occurring health needs, but “more severe autism” is not a sufficient explanation. Two interacting patterns can create contradiction as well as addition: needing routine and novelty, seeking stimulation and becoming overloaded by it, or having deep focus that is hard to direct on demand.

The mutual-exclusion rule has aged worst. It encouraged diagnostic overshadowing: an attention or activity pattern could be treated as “part of the autism,” while social, sensory, or repetitive patterns could be overlooked once ADHD was recognised.

Single-cause language has also aged poorly. Early molecular work sometimes searched for a small set of genes that would explain the overlap. Modern results fit many small inherited influences, development, environment, measurement, and several trait dimensions. Shared heritability is a population statistic, not a personal origin story.

Finally, prevalence ranges quoted without a denominator have become a source of confusion. Rates change depending on whether researchers start with an autism clinic, an ADHD clinic, a twin cohort, an insurance database, children, or adults. “How common is AuDHD?” has no honest single number without saying who was counted and how.

The strongest conclusion is straightforward: ADHD and autism are distinct, correlated, and able to co-occur in the same person. Family and twin studies support shared inherited influences. Trait studies find both separable dimensions and combined profiles. National registers and health claims show substantial groups of adults with both diagnoses.

The next conclusion is more practical. Once one condition is recognised, the other should not be ruled out by default. Assessment still needs evidence that each set of criteria is met and that the pattern is not better explained by another condition, a temporary state, or the measurement tool itself.

Research is much less developed on what support looks like when both are present. Most studies in this guide were designed to explain overlap, count diagnoses, or model traits. They did not test integrated support for an AuDHD life.

Adult research still inherits childhood definitions and clinic samples. We need longitudinal work that follows varied profiles across adulthood, including late-diagnosed people, people with intellectual disability, people who communicate without speech, and people whose race, gender, or economic position changes access to diagnosis.

Measurement overlap remains a live problem. Restlessness can reflect sensory regulation, anxiety, or hyperactivity. Missing a social cue can arise from attention, processing speed, or a difference in communication style. A score can record the same outward behaviour without identifying its meaning.

The field also needs intervention studies built for co-occurrence. A support that helps one demand may intensify another. The relevant outcome is not whether someone looks less autistic or less ADHD. It is whether daily life becomes more workable, chosen goals become more reachable, and the cost of constant compensation falls.

AuDHD is useful language because it lets people discuss the interaction, not only list two labels. Research writing often says “comorbidity,” a term that can make the person sound like a pile of problems. We use “co-occurring” where the technical meaning does not require the older word.

The evidence should not become a stereotype. Not every AuDHD person has the same push-pull between routine and novelty, the same sensory profile, or the same executive pattern. Group averages cannot decide whether a person’s self-understanding is credible.

Good assessment keeps the categories available without forcing every experience into them. Good support starts with the actual pattern: what drains capacity, what creates traction, what conflicts, and what the person wants to be different.

Then / Now / Still unknown

• Then: Diagnostic rules treated autism and ADHD as mutually exclusive even while family and twin studies showed their traits travelling together.
• Now: Adult twin, registry, latent-structure, and health-record studies support both real overlap and meaningful distinction.
• Still unknown: Which integrated supports work best for different AuDHD profiles across adult life.

Rommelse et al. 2010. The foundational review that assembled family and twin evidence before dual diagnosis was permitted. Its shared-heritability argument held up better than its early candidate-gene discussion.

Ghirardi et al. 2019. A population-based study of 6,866 young-adult twins. It showed that overlap changes by trait dimension and that the model cannot be reduced to one shared cause.

Solberg et al. 2019. A nationwide Norwegian register study with separate ADHD, autism, and combined groups. It demonstrated adult co-occurrence at scale and showed distinct patterns of additional diagnoses.

Krakowski et al. 2021. Eight studies of combined symptom structure. Most factor analyses kept the dimensions separate; every person-centred latent-class study found a co-occurring profile.

Yerys et al. 2025. More than 3.5 million US Medicaid-enrolled adults. ADHD diagnoses were much more common in autistic groups than in the comparison group, but the estimates belong to claims data and should not be treated as population prevalence.

We searched PubMed, PubMed Central, and the NeuroDiversion citation manifest on 5 October 2026. We selected the early review already cited across our AuDHD pages, an adult twin study, a national adult registry comparison, a scoping review of symptom structure, and the largest recent adult health-record cohort we found. All five were open and read in full. This is a curated evidence guide, not a systematic review.