If you carry children with childhood apraxia of speech (CAS) on your caseload, there is a number worth knowing: in several of the largest genetic cohorts assembled so far, roughly one in three children with CAS carries a pathogenic variant that helps explain their motor speech disorder. That figure does not come from a single lucky study. It recurs, with real variation, across independent research groups working with different cohorts, different sequencing technologies, and different countries.
This post works through six studies that inform that number and the decisions around it. They do not all agree, and they should not be read as if they do. Only two of them measure diagnostic yield at the population level. One is a three-case mechanistic study. One reports copy number variants in a set of families without a yield denominator at all. And one looks entirely outside the CAS diagnosis, at children with idiopathic speech delay, and finds a very different, much lower yield. The disagreements here are as clinically useful as the headline number, and they are what should actually change how you think about referrals, which test to request, and who is even a candidate for testing.
Two Cohorts and a Review, Landing Near One in Three
Start with the two estimates that carry the most weight. A 2024 review of the genetic architecture of childhood speech disorder pooled data from three research cohorts ascertained specifically for CAS and found 36 pathogenic variants across 122 cases — close to one in three (Study 1). This is a synthesis of existing cohorts rather than a new independent sample, so it should be read as a summary of the field, not as a fourth data point.
The more independent confirmation comes from a tertiary hospital speech clinic that ran a clinically accredited genomics pipeline — chromosomal microarray, Fragile X testing, and exome sequencing — on 153 children ascertained for motor speech disorder and assessed by a clinical geneticist and speech pathologist. Forty-four of them, or 29%, had pathogenic variants (Study 4). That is a genuinely separate cohort — a tertiary hospital speech clinic sample assessed with a clinically accredited pipeline — and it lands almost exactly where the review predicted.
Why the Number Mostly Holds Up — With One Notable Exception
A single-center Italian study offers a useful check on how solid this is. Out of a cohort of 69 children with CAS, whole exome sequencing was performed in the families of the 34 children who had already been found to have no copy number variants (Study 3). Among those 34, high-confidence variants were identified in 7 — about one in five, not one in three. Two of the seven high-confidence variants affected KAT6A and CREBBP, confirming genes already reported elsewhere; separately, 20 variants in the low-confidence group occurred in genes not previously associated with CAS. The study's own authors describe this as a relatively high diagnostic yield, and it is, but it is meaningfully lower than the tertiary hospital and review figures. The study does not report how many of the excluded children had CNVs or what their status was, so it should not be read as evidence that the remaining 35 children were all CNV-positive — only that they were excluded from the exome analysis for that reason.
Put together, the honest summary is that diagnostic yield for children ascertained specifically for CAS runs somewhere between roughly one in five and one in three, depending on the cohort and pipeline, rather than a single fixed number every study reproduces exactly. That range is still remarkably high for a speech-language disorder, and still worth acting on — but it is a range, not a consensus figure carved in stone.
One Cohort's Predictors of a Positive Genetic Finding
The tertiary hospital cohort is also the source of the most actionable clinical checklist currently available, and it deserves to be labeled clearly as coming from one 153-child sample rather than as an established, universally validated rule. Using odds ratios, the study found that a genetic diagnosis was more likely in children with:
- Delayed walking or other gross and fine motor involvement
- Receptive language impairment
- Cognitive impairment
- Dysmorphic features
The study also found that children with both CAS and dysarthria together were more likely to receive a genetic diagnosis than children with CAS alone. These are useful, concrete flags to build into an intake conversation, but they come from a single, unreplicated cohort. Treat them as a working checklist to raise your index of suspicion, not as a validated diagnostic algorithm.
The Autism Paradox
Here is where the studies genuinely disagree, and where the disagreement matters clinically. In the same 153-child tertiary hospital cohort, autism spectrum disorder was less commonly associated with a positive genetic diagnosis, not more (Study 4). That is counterintuitive if you have absorbed the idea that autism and CAS share biological pathways.
They may well share pathways — but the evidence for that comes from a different kind of study entirely. One three-case series (n = 3) sequenced the exomes of three unrelated children with CAS and fine/gross motor involvement and found that their implicated genes, especially LAMA5 and LAMB1, converged on a dense functional network centered on the laminin-511 complex, with high expression in the developing cerebellum and phenotypic overlap with autism and other neurodevelopmental conditions (Study 2). That is a real and interesting molecular finding, but it rests on three children, not a body of replicated literature, and it is not the same kind of evidence as a population-level odds ratio from 153 children.
The clinical takeaway is not that autism and CAS are unrelated. It is that autism comorbidity, by itself, should not be treated as a strong independent trigger for a genetics referral in the way that motor delay, receptive language impairment, cognitive involvement, or dysmorphism are. In the largest clinical cohort available, it pointed the other way.
One Test Isn't Enough: What Copy Number Variants Reveal
Of the 44 positive diagnoses in the tertiary hospital cohort, 9 were reported as copy number variants, identified in a pipeline that included chromosomal microarray alongside exome sequencing and Fragile X testing (Study 4). The study does not specify that microarray was the only technology capable of catching them, and no study in this set tested an exome-only pipeline against this combined one head to head — but a pipeline without microarray or dedicated CNV analysis is less likely to detect CNVs like these.
A separate whole-genome sequencing study of 27 families with CAS (101 individuals) reinforces this. It identified 19 unique copy number variants across 17 genomic regions, validated eight of them using microarray technology, and found that many of the same CNVs also turned up in children with milder speech sound disorders, not just CAS (Study 5). This study does not report an overall diagnostic yield, so it cannot be compared directly to the one-in-three or one-in-five figures above. What it does show clearly is that CNVs are a heterogeneous but genuinely important cause of CAS. No study in this set tested exome-only sequencing against a combined exome-plus-microarray pipeline head to head, but a pipeline without microarray or dedicated CNV analysis is less likely to detect CNVs of this kind. If you are advocating for a specific test, exome sequencing combined with chromosomal microarray or CNV analysis is the stronger choice: Study 4 used exactly that combined pipeline to find its 9 CNV diagnoses, and Study 5 shows how important CNVs can be. That recommendation sits in tension with Study 1's review, which reports no apparent benefit of whole-genome over whole-exome sequencing and notes that every gene identified to date sits in coding regions — but Study 1's conclusion is about the monogenic, single-gene conditions it reviewed, not about CNVs, which is exactly the category exome sequencing alone is not designed to catch.
Beyond the CAS Label: Idiopathic Speech Delay
Every study discussed so far ascertained children who already carried a CAS diagnosis or a broader motor speech disorder diagnosis. A 2024 genome sequencing study asked a different question: what turns up in children with idiopathic speech delay — delayed speech development without a motor speech diagnosis like CAS, without a language or reading disorder, and without intellectual disability (Study 6)?
In a cohort of 23 such children, genome sequencing identified one child with a pathogenic frameshift deletion in SETD1A, a gene previously linked to a broader syndrome involving global developmental delay, mild intellectual disability, facial dysmorphism, and behavioral or psychiatric symptoms — and independently reported in CAS itself in two separate studies. Other probands in the cohort carried likely pathogenic missense variants in SPTBN1 and ARF3, novel candidate genes for future research rather than confirmed causes.
That works out to roughly 1 in 23, or about 4% — a figure this post is calculating rather than one the study reports, based on a single pathogenic case among 23 children, so the confidence interval around it is wide. Even accounting for that uncertainty, it looks far lower than the one-in-five to one-in-three range seen in CAS-ascertained cohorts. This matters for two reasons. First, it confirms that the boundary of who might benefit from genetic testing is genuinely expanding beyond a strict CAS diagnosis. Second, and just as important, it shows that the odds of a positive result drop sharply once a child's presentation is speech delay alone, without motor speech, language, or cognitive involvement. The expanding boundary does not carry the one-in-three figure with it.
A Referral Decision Guide for Your Caseload
Before applying any of this to a specific child, one caveat has to come first: every yield figure in this post comes from children who were ascertained for CAS or motor speech disorder in research cohorts or tertiary hospital speech clinics — severe, relatively rare presentations referred for specialist assessment. The review in Study 1 says explicitly that these genetic insights have not yet translated to understanding milder, far more common forms of speech or language impairment, such as phonological disorder or stuttering. Do not apply a one-in-three expectation to a general caseload of mild speech sound disorder. Reserve it for children who meet criteria for CAS, or a closely related motor speech disorder, with additional flags.
With that caveat in place, a reasonable, evidence-grounded referral guide looks like this:
- Refer with confidence: a child with diagnosed CAS plus any of delayed walking, other gross or fine motor involvement, receptive language impairment, cognitive impairment, or dysmorphic features. These are the flags associated with a positive genetic finding in the largest clinical cohort available (Study 4).
- Refer with confidence: a child with both CAS and dysarthria, which was more strongly associated with a genetic diagnosis than CAS alone (Study 4).
- Do not treat as a standalone trigger: autism spectrum disorder comorbidity by itself. It did not raise diagnostic yield in the largest available cohort, even though it is biologically interesting for other reasons.
- Advocate for the right test, not just any test: exome sequencing combined with chromosomal microarray or CNV analysis. In the tertiary hospital cohort, 9 of the 44 positive diagnoses were copy number variants that exome sequencing alone would not catch (Study 4); a separate whole-genome study confirms CNVs are a heterogeneous and clinically important cause of CAS more broadly, though it reports no overall diagnostic yield (Study 5).
- Consider, with lower expectations: children who do not meet full CAS criteria but have speech delay with no other explanation and additional syndromic features. Evidence here is thinner and yield is much lower — closer to 1 in 20 than 1 in 3, though that estimate rests on a single positive case in 23 children and should be read as illustrative rather than precise — but SETD1A, SPTBN1, and ARF3 are emerging candidates worth knowing (Study 6).
What to Tell Parents and Physicians When You Request Testing
To physicians who may not think of speech disorders as genetically tractable, the clearest framing is a comparison: in the tertiary hospital cohort, the clinical genetic diagnostic yield for motor speech disorders was 29%, comparable to epilepsy and cerebral palsy — two conditions where genetic testing is already routine (Study 4). That is a legitimate basis for requesting a referral, and it is worth stating plainly in a referral letter rather than assuming the physician will make the connection.
To families, it helps to be specific and to manage expectations honestly. A useful script:
"Because your child's speech difficulty involves [motor delay / receptive language difficulty / cognitive concerns / dysmorphic features], there's a reasonably good chance — research suggests somewhere between one in five and one in three in similar cases — that a genetic test could identify a specific cause. That doesn't mean today's therapy plan changes immediately, but it can end a long diagnostic search, help us watch for other related conditions, and sometimes open access to condition-specific supports. It's also common for the test to come back without an answer, and that doesn't mean there isn't a genetic contribution — it means current testing didn't find it."
When requesting the test itself, be specific rather than general: ask for exome sequencing paired with chromosomal microarray or CNV analysis, not exome sequencing alone, given how much a CNV-blind pipeline can miss. And if the child's presentation sits outside a strict CAS diagnosis — speech delay with syndromic features but no motor speech disorder — it's reasonable to raise testing as an option, while being clear with families that the odds of a positive result are lower than for children with a full CAS diagnosis.
None of these six studies, on its own, would justify a change in referral practice. Together, with their disagreements intact rather than smoothed over, they make a much stronger case: genetic testing for CAS and related motor speech disorders belongs in the same conversation as testing for epilepsy and cerebral palsy, with a testing pathway and a referral threshold that reflect what the evidence actually shows, not a single rounded-off number.
