Genetic Testing History in Adults with Autism Spectrum Disorder

Eligibility Criteria

This was a retrospective chart review study using a previously characterized patient cohort from our autism specialty clinic described in Thom et al. (2022). In brief, eligibility criteria included: (1) past or current patient at the Lurie Center; (2) documented developmental history; (3) comprehensive clinical evaluation by a Lurie Center or MGH developmental pediatrician, psychiatrist, psychologist, neuropsychologist, or neurologist; and (4) support for an ASD diagnosis. In the current study, patients were included who had at least one visit with a qualifying clinician from our hospital between 5/1/2010 and 12/15/2020.

Information sources

Potentially eligible adults were originally identified for an ongoing study investigating co-occurring medical conditions in adults with an ASD diagnosis (the “original” study). The identification process and confirmation of ASD diagnosis are described by Thom et al. (2022). Medical records from adults excluded from the original study due to genetic conditions were reviewed for ASD diagnosis to determine eligibility for the current study. Updated demographic and encounter history information for adults with a confirmed ASD diagnosis were obtained from the Mass General Brigham Research Patient Data Registry (RPDR). For the current study, data were most recently requested from the RPDR on May 12, 2021. Data on the presence and severity of intellectual disability were available for a subset of adults included in a study of cardiometabolic risk factors (Thom et al., 2022). Further treatment history information and clinical and genetic testing data were collected using manual review of the electronic health record (EHR) from December 2020 to September 2022, with a medical record cutoff date of December 15, 2020.

Selection process

Figure 1 illustrates the study flow for the selection process of the original study and the current study. Most potentially eligible adults had a confirmed ASD diagnosis from the original study. For 33 patients who were ineligible for the original study due to genetic conditions, medical records pertaining to developmental and clinical history were reviewed by the same expert psychiatrists participating in the original study to confirm ASD diagnosis. For all patients with a confirmed ASD diagnosis, data from the Research Patient Data Registry (RPDR) and the patients’ electronic health records (EHRs) were reviewed for qualifying visits.

• Download figure
• Open in new tab

Figure 1: Study Flow

Measures

Demographic characteristics of patients including age, sex, race, and ethnicity, were classified using data returned from the RPDR. Data on clinical and family history of the patient were coded based on retrospective manual review of developmental pediatrics, psychiatry, neurology, and primary care notes in the EHR. A keyword search of the full EHR for “seizure” was used to verify the presence or absence of seizure history. Presence or absence of intellectual disability was assigned based on full scale IQ, when documented in the EHR. If full scale IQ was not documented, presence or absence of intellectual disability was assigned based on qualitative assessment in the clinical notes or documentation of services through the Department of Developmental Services.

Genetic testing history was determined based on a keyword search of the full EHR for “genetic” and subsequent review of relevant clinical documentation and lab testing. For each occurrence of genetic testing, the date of sampling, type of test, and testing results were recorded. Testing type history was summarized over time as one or more tests using a recommended method (chromosomal microarray, autism ID panel, or exome sequencing), karyotype testing only, or other non-recommended method only. Testing results were classified as positive, negative, or variant of unknown significance (VUS) based on the lab report associated with the testing or, if lab results were unavailable in the EHR, documentation of testing results by a clinician. Final decisions on the categorization of type of testing and designation of testing as ASD-related were made by the study’s senior author (AMN).

Statistical Approach

Demographic and clinical characteristics of adults in the sample were summarized by genetic testing history using means, standard deviations, and ranges for continuous variables and frequencies and percentages for categorical variables. Timing of earliest ASD clinic encounter and most recent ASD clinic encounter were categorized based on their observed distributions in the sample to achieve adequate distribution of adults across categories. Robust linear regression (continuous variables) and chi-square tests (categorical variables) assessed the statistical significance of differences in characteristics among those with testing information in the EHR, those without testing history in the EHR, and those who declined genetic testing. Ninety-five percent confidence intervals (CIs) for percentages were calculated using Wilson’s method for binomial confidence intervals (specific genetic tests) and Goodman’s method for multinomial confidence intervals (testing history summary, testing results; Goodman, 1965). Associations of candidate demographic and clinical predictors with frequency of testing using a recommended method among those adults with testing history documented were quantified using relative risk regression models fit with the modified Poisson approach (Zou, 2004). Due to the modest number of adults with a history of testing using a recommended method and the potential for strong associations between candidate predictors, associations with candidate clinical predictors were initially assessed individually in single predictor models controlling for age and sex. Clinical predictors associated with p-values less than 0.10 in the single predictor models were then entered together into a multiple predictor model, from which statistical significance was assessed.

Adults with missing values for demographic, clinical, and genetic testing variables were excluded from corresponding data analyses; details on frequency of missing data are provided in table and figure footnotes. Data analysis was conducted using version 9.4 of SAS (SAS Institute, Cary, NC) and version 4.3.2 of R (R Core Team, Vienna, Austria). Multinomial confidence intervals were calculated using the MultinomCI function from the R DescTools package (Signorell 2024). All statistical tests were two-sided and conducted at the test-wise alpha=0.05 significance level.

Study selection

The 630 adults with confirmed ASD diagnoses from the original study were combined with 31 adults, who although ASD diagnosis was confirmed were excluded from the original study because they had a known genetic diagnosis. Of the 661 patients who met the eligibility criteria, records for 652 patients were available in the research database. Six hundred thirty of these patients were seen by a qualifying provider from our hospital between May 1, 2010 and December 15, 2020, and thus were included in the current study.

Study characteristics

The demographic and clinical characteristics of the 630 adults included in this study are described in Table 1.

Results

Main findings

This study shows that, even at a highly specialized ASD clinic, the percentage of adults with documented ASD-related genetic testing history in records from 2010-2020 was only 41% (261 of 630 adults). Of the adults with documented ASD-related genetic testing history, only 54% (95% CI: 46%, 61%) had received one or more of the recommended methods (chromosomal microarray, autism/intellectual disability gene panel, or exome sequencing). Importantly, 28% (95% CI: 19%, 39%) of adults who received one or more of the recommended methods received a genetic diagnosis with an additional 10% (95% CI: 5%, 19%) for whom new gene variants were identified that may cause ASD. Although few adults received testing with the autism/intellectual disability gene panel or exome sequencing between 2009-2020, pathogenic or unknown significance variants in ASD-related genes were identified in 53% (95% CI: 26%, 79%) with the gene panel and 60% (95% CI: 31%, 83%) with exome sequencing. Together, these findings highlight the potential benefit of reviewing genetic testing history in adults with ASD and performing recommended ASD-related genetic testing to identify, where possible, the genetic cause of ASD.

Relationship to other reports

There is very limited information in the medical record on the percentage of adults with ASD who receive ASD-related genetic testing as part of their routine clinical care. A retrospective chart review study (2013-2019) of 1280 children and adults diagnosed with ASD based on the ADOS-2 reported that only 16.5% had received any ASD-related genetic testing (Moreno-De-Luca et al. 2020). Only 10% (132) of the patients were 20 years or older in this cohort from the Rhode Island Consortium for Autism Research and Treatment (RI-CART) study. The percentage of these adults with any genetic testing was very low: Fragile X (∼12%), karyotype (∼7%), and chromosomal microarray (∼2.5%). Interestingly, genetic testing was more likely to be ordered by pediatric subspecialists than psychiatrists or psychologists. This study is consistent with our findings in our larger population of adults with confirmed ASD that many adults lack the recommended ASD-related genetic testing. It also suggests that the percentage of adults with ASD-related genetic testing may be much lower in the community than in ASD specialty clinics.

A 2021 study in Sweden found that only 2.8% of the 213 adolescents and adults had received a referral for genetic testing after an ASD diagnosis (Hellquist and Tammimies 2022). Interestingly, the prevalence of ASD in 18-24 years (older ages not reported here) was 2.4%, similar to the percentage of adults with ASD in Massachusetts, and the recommendation in Sweden by several medical societies, similar to here in the U.S., is that clinical genetic testing be offered to everyone with ASD. Recommendations in Sweden included chromosomal microarray and exome sequencing. The autism/intellectual disability gene panel that has been commonly used in American clinical genetic testing was not mentioned. There are multiple reviews advocating for adults with ASD to receive recommended ASD-related genetic testing (e.g., Schaefer 2016 and Kreiman and Boles 2020) consistent with our experience that many adults still need to be offered ASD-related genetic testing.

A recent meta-analysis by leading investigators in the field of neurodevelopmental (NDD) genetics on the outcomes of genetic testing found that the yield of exome sequencing was 30-43% compared to 15-20% with chromosomal microarray (Srivastava et al. 2019). In isolated NDD, the yield was 31% (CI: 25-38%) and in NDD plus other conditions was 53% (CI: 41-64%), for example ASD with ID. The percentage of our patients with a positive result on chromosomal microarray was much higher than reported in this meta-analysis. This could be due to differences in the diagnoses included in the studies reviewed (and method for confirming the diagnosis); our study included only adults with an expert confirmed ASD diagnosis. It could also be that positive results on chromosomal microarray were more likely to be recorded in the medical record of adults with autism for the time period in our study (2010-2020). The yield of positive results on the exome sequencing in this study was consistent with our study (47%). This increases the confidence in the reliability of our results for exome sequencing, given that the number of patients with exome sequencing in our study was very small (15 adults).

Limitations

This study highlights the need to review and document ASD-related genetic testing in adult patients. This study has the advantage of including a large cohort (630 adults) that range in age from 20- to 66-year-olds, as the majority of studies in ASD focus on ages 22 and younger despite the fact that most people with ASD are adults. However, the mean age in the study was 29.6 (SD 7.1); thus, the lack of genetic testing and/or documentation could be even lower in older adults with ASD. Although autism prevalence is lower among non-Hispanic White children than other racial and ethnic groups (Maenner et al., 2023), our patient cohort has the limitation of being primarily White (88%) and non-Hispanic (98%). Findings from our study may not generalize to other populations of adults underrepresented among patients of our clinic, and in particular to Black, Hispanic, and Asian adults. It is possible that the rates of ASD-related genetic testing performed and documented may be even lower in the general population than the population represented in our clinic’s cohort.

Importantly, all the patient records in this study were reviewed by multiple psychiatrists specialized in working with adults with ASD in our clinic, and the ASD diagnosis was confirmed by consensus. However, there were patients seen in our clinic during 2010-2020 for whom the records were insufficient to confirm the diagnosis of ASD (Figure 1). Thus, we do not know what percentage of the patients with insufficient records had an accurate ASD diagnosis and were excluded from the study. Notably, there are only 26% women included in this study. This 3:1 ratio of males to females with ASD, compared to the often cited 4:1 ratio, highlights an improvement in the inclusion of women with ASD in this study. However, community studies here in the Boston area highlight that the true male-to-female ratio may be closer to 1.8:1 men to women with ASD due to the delayed age of diagnosis for many girls and women with ASD and well-documented gender bias inherent to common diagnostic and research tools such as the Autism Diagnostic Observation Scale (ADOS; (D’Mello et al. 2022). Our ASD specialty clinic serves a range of ASD severity levels; however, we see a larger number of moderately-to-severely affected adults with ASD (Level 2 and 3) because of the difficulties they often face receiving medical care in other clinical settings. Our specialty clinic is designed to accommodate the special sensory and safety needs of these patients and is staffed, including receptionists and medical assistants, by people sensitive to needs of children and adults with severe forms of ASD.

There is also a risk of reporting biases due to data not covered in the EHR. We are unable to determine if adults without documentation of testing in their records never received testing or if they received testing that was undocumented. For adults with undocumented testing, it may be that less than 54% received the recommended testing, or that the rate of positive results was lower in the patients in this group than in those with documented genetic testing. We also relied on the accuracy of clinical characterization documented in provider notes, and on the completeness and accuracy of genetic testing data available in the record. Despite these limitations, our findings that many adults lack the recommended ASD-related genetic testing which could reveal a genetic cause is still valid and highly relevant for clinicians caring for adults with ASD.

The time period of the record reviews is also relevant. We included patients who were seen between 2010 to 2020 by a qualified specialist in our hospital. Since 2020 (the latest date in this record review), the use of the autism/intellectual disability gene panel has become more common due to better insurance coverage (including for the majority of our patients who receive state-funded health insurance). Moreover, the likelihood of receiving a genetic diagnosis, particularly for our moderate-to-severely affected adults with ASD (Level 2 and 3), is likely higher than reported here due to the increasing number of confirmed ASD-related genes included in the panel. Previously categorized VUS have also been recategorized as pathogenic based on the findings of large population-level ASD-related genetic studies such as SPARK (Wang et al. 2022). As of 2023, access to whole exome sequencing is also increasing with many insurance companies now covering the costs.

Implications

Identification of the genetic cause of ASD in adults can provide a powerful tool for improving quality and access to health care and other ASD-related services. For clinicians, the genetic cause provides critical information about prognosis and associated medical problems. For adults with ASD, the genetic cause informs care and facilitates connection to support groups for specific genetic conditions. For researchers, the genetic cause aids mechanistic and novel therapeutic studies. Our study reveals that more than half of adults with ASD are likely yet to have recommended genetic testing. As ASD-related genetic testing is now more widely available and the number of ASD-related genes is expanding, there is an opportunity now for clinicians to offer ASD-related genetic testing to their adult patients with ASD and for adults with ASD to receive the benefits of identifying the genetic cause.

Notably, there may be an education gap for patients as well as providers when considering ASD-related genetic testing. A Canadian survey of 461 autistic individuals found that only 27% would have wanted genetic testing during childhood and only 35% felt that testing should be routinely offered to autistic adults (Byres et al. 2023). Here we advocate for increasing access to and documentation of genetic testing for adults with ASD. There is a concern among many adults who identify as autistic that ASD-related genetic testing could lead to a negative selection for embryos, for example, that have an ASD-related gene. It is important to distinguish this concern from the current study. All of the patients in this study are adults who have a confirmed diagnosis of autism spectrum disorder.

Knowledge of a genetic cause of ASD can inform medical care, particularly for patients in which the genetic variant affects multiple organ systems (Kreiman and Boles 2020). This can be lifesaving for adults with language impairment and/or reduced ability to report pain or discomfort. Moreover, in multiple cases in our clinic, identification of a genetic cause lifts decades of guilt parents have carried that something they did (or did not do) during pregnancy, or at the time of birth, led to the severe ASD-related challenges their adult son or daughter faces. Identification of an ASD-related gene has also enabled multiple women with ASD and co-occurring ID to receive specialized services for ASD from the government, where the ASD diagnosis was not made until adulthood because of widespread and longstanding gender bias in ASD diagnosis.