A chart review from a multidisciplinary neurology practice in Orlando, Florida, found that autistic patients with a history of seizures had higher adjusted odds of having an abnormal genetic finding recorded than patients without a seizure history. Female patients also had higher adjusted odds than male patients after researchers accounted for other factors.
The findings come from records of 1,884 patients with an autism spectrum disorder diagnosis seen at Kids Neuro Care, which works with the University of Central Florida College of Medicine. The study was a retrospective review of clinical charts, not a prospective trial, and it cannot show that seizures or sex caused an abnormal genetic result.
What the records showed
Genetic results were documented for 1,011 patients. Of those, 371—36.7%, with a 95% confidence interval of 33.7% to 39.8%—had a result recorded as abnormal.
Among patients with documented testing, 82 of 172 people with a seizure history had an abnormal finding, compared with 289 of 839 people without seizures. After adjustment, seizure history was associated with 1.63 times the odds of an abnormal finding.
Abnormal findings were recorded for 106 of 259 female patients, or 40.9%, and 265 of 752 male patients, or 35.2%. The female-versus-male difference was not statistically significant before adjustment, but it became significant after accounting for age, seizures and EEG results. The adjusted odds were 1.48 times higher for female patients.
In the main adjusted model, age was not linked to abnormal genetic findings. Abnormal EEG results also were not significantly associated with the findings. A smaller analysis that included MRI results still found associations with seizure history and female sex, while abnormal MRI was not independently linked to an abnormal genetic result.
“Abnormal” covered several types of findings
The researchers counted any result recorded in the chart as abnormal and did not reclassify findings according to current genetics standards. The records also did not reliably identify which type of genetic test each patient had received.
Among 368 abnormal results with a written description, gene or sequence variants accounted for 127, or 34.5%. Deletions and duplications each accounted for 77, or 20.9%, while runs of homozygosity accounted for 61, or 16.6%. A patient could have more than one finding.
Only 65 of all 371 abnormal results, or 17.5%, named a specific syndrome or condition. Twenty-seven results, or 7.3%, included a Fragile X-related note. Those notes included affected patients as well as carrier, premutation and borderline “grey-zone” results, so an abnormal chart label did not necessarily mean that a genetic cause of autism had been established.
Why the findings need careful interpretation
Only about half of the patients had a recorded genetic test, and testing was not random. Across the full group, patients who had an EEG had about 2.4 times the odds of having genetic testing, while those who had an MRI had about 2.2 times the odds. Patients with a seizure history were less likely to have a recorded genetic test, although the records could not explain why.
The researchers could not determine the order of seizures, EEGs, MRIs and genetic tests because test dates were unavailable. The records also lacked information about intellectual disability, unusual physical features and developmental regression. As a result, the association with female sex may reflect differences in clinical complexity rather than sex itself.
The analysis had limited ability to distinguish patients with abnormal and normal results: model AUC values ranged from 0.59 to 0.61, only slightly better than chance. Because the review came from one specialty practice, its patients may have had more complex neurological features than people seen in community or primary care settings.
What the study adds to the testing discussion
The researchers said the results support broad, guideline-based genetic testing in autism rather than restricting testing to patients with seizures or to girls. That conclusion is important because the study’s associations do not identify a reliable group in which testing would be useful only on the basis of seizure history or sex.
A recorded abnormal result can have different meanings: it may help explain autism, remain uncertain, or be unrelated. The study therefore provides information about patterns in clinical records, not a measure of how well a particular test detects an explanation for autism.
A separate 2025 study of 523 children with autism at a U.S. center reported epilepsy in 35.4% of children whose test results explained their autism, compared with 16.4% of those without a genetic explanation. In that study, exome sequencing identified a cause in 23.0% of children tested, compared with 8.3% for chromosomal microarray. Those results come from a different study population and do not establish that one testing approach will produce the same results in the Florida chart review.
AI tools were used to assist with the preparation of this article under our editorial standards.
Leave a Reply