More people in Denmark are being diagnosed with attention-deficit/hyperactivity disorder (ADHD) or autism spectrum disorder (ASD). A new study shows that those diagnosed more recently have, on average, a lower genetic predisposition to the conditions than those diagnosed earlier. The researchers believe that broader diagnostic criteria and changes in clinical practice are the most likely explanation for the trend.
The number of ADHD and ASD diagnoses has risen sharply over recent decades. Studies from several countries have reported a four- to 10-fold increase. But behind this trend lies a paradox: the people receiving these diagnoses have, on average, become less genetically predisposed to the conditions.
This has been shown by a new study of more than 37,000 people in Denmark with ADHD or ASD recently published in JAMA Psychiatry. The later in the more than 20-year period the diagnosis was made, the lower the average genetic risk among those diagnosed.
The study’s lead author was Sonja LaBianca, a doctor and postdoctoral researcher at the Department of Biological Psychiatry of the Mental Health Services in the Capital Region of Denmark. The study was carried out together with researchers from the nationwide iPSYCH consortium.
She says that the finding offers a new clue in the effort to explain why more and more people are being diagnosed.
“Our results are most consistent with changes in diagnostic criteria and clinical practice having influenced who receives an ADHD or ASD diagnosis,” says Sonja LaBianca, continuing:
“We are born with our genes, and the overall genetic make-up of the population changes only very little over 20 years. The difference between those diagnosed earlier and later therefore suggests that the route to a diagnosis itself has changed,” she adds.
Three hypotheses could explain the rise in diagnoses
When Sonja LaBianca and colleagues began the study, they started with a fundamental problem. The sharp rise in ADHD and ASD diagnoses does not inherently explain why more people are being diagnosed. Before turning to the genetic data, the researchers therefore set out three hypotheses about what might lie behind the trend.
The first was that the healthcare system has become better at identifying people with ADHD or ASD who would previously have gone without being diagnosed. The second was that people who would once have received a different psychiatric diagnosis were now more often being diagnosed with ADHD or ASD. The third was that the threshold for receiving these diagnoses had become lower so that they were also being given to people with milder symptoms.
“We formulated the hypotheses in advance so that we could translate each explanation into an expected genetic pattern. We could then investigate which pattern the actual data most closely resembled,” explains Sonja LaBianca.
To investigate whether people diagnosed later in the period had, on average, a different genetic risk profile from those diagnosed earlier, the researchers used data from the iPSYCH cohort. The cohort combines genetic information from the Danish Neonatal Screening Biobank with information from national health registries. This enabled the researchers to link anonymised DNA from dried blood spots with information about psychiatric diagnoses and when the first diagnosis was made.
Ten genetic scores put the explanations to the test
The study included 20,111 people with ADHD and 17,071 people with ASD who received their first diagnosis between 1994 and 2016.
For each person, the researchers calculated 10 polygenic scores. A polygenic score combines the contributions of many genetic variants, each of which has only a small effect, into a single number. The score captures only part of the overall genetic predisposition and can be used to describe average differences between large groups – not to determine whether an individual has or will develop ADHD or ASD.
The researchers adjusted the analyses for factors including sex, age at first diagnosis and genetic ancestry. They then simulated how the various risk scores would be expected to change over time if each of the three hypotheses were the main explanation for the rise in diagnoses. In other words, they examined not only whether genetic risk rose or fell but also whether the changes across the various scores resembled the pattern predicted by each hypothesis.
“The strength of our method is that we can compare several measures of genetic risk and see how they move together. That enables us to assess which explanation best fits the actual data,” says Sonja LaBianca.
Several genetic measures pointed in the same direction
When the researchers compared the actual data with the simulated patterns, they found a decline in several measures of genetic risk.
Among people with ADHD, the average polygenic score for ADHD fell by 0.06 standard deviations for every 10 years later the diagnosis was made. This is a modest difference at the group level, but the researchers found the same trend in the scores for ASD, bipolar disorder and schizophrenia.
Among people with ASD, the average polygenic score for ASD similarly fell by 0.07 standard deviations for every 10 years later the diagnosis was made. The researchers also saw the same trend in the scores for bipolar disorder and schizophrenia and in a genetic measure linked to educational attainment.
For Sonja LaBianca, the important point is that several measures of genetic risk point in the same direction.
The researchers would have expected a different pattern if the rise were primarily caused by people who would previously have received other psychiatric diagnoses now being diagnosed with ADHD or ASD instead. But several measures of genetic predisposition declined at the same time. Taken together, this suggests that, over time, the diagnoses have been given to a broader group of people who, on average, are less genetically predisposed to the disorders.
“We therefore believe that a lower threshold for who receives a diagnosis is the explanation that best fits the results,” notes Sonja LaBianca.
What the difference means clinically remains unclear
Those diagnosed later still had a higher average genetic risk of the disorder in question than the general population. The difference was still clear but smaller than among those diagnosed earlier in the period.
Sonja LaBianca emphasises that the study cannot determine whether more people should have been diagnosed earlier or whether some people today are being diagnosed on too weak a basis. The results describe average differences between people diagnosed at different points in time. They say nothing about the quality of an individual clinical assessment.
According to Sonja LaBianca, a lower average genetic risk could arise because those diagnosed later have milder symptoms. Another possibility is that environmental factors have become more important over time, meaning that, on average, less genetic predisposition is needed before symptoms lead to a diagnosis.
The study cannot distinguish between these possibilities. The genetic data contain no detailed information about the severity of symptoms of individual people or about how they were functioning at the time of diagnosis.
“The next step is to find out what that difference means clinically,” says Sonja LaBianca.
In her view, the study therefore shifts the discussion from the question of how many people receive a diagnosis to another: Who is being diagnosed – and why?
This question also has practical implications. An ADHD or ASD diagnosis often provides access to treatment and support, and the way society is organised and the demands it places on people may therefore help determine who is referred to mental health services and what role these diagnoses come to play.
“This needs to be investigated in studies that follow patients and include detailed information from their clinical histories. Our data only go up to 2016, so we cannot see whether the same pattern has continued since then,” concludes Sonja LaBianca.
