Pilot Study Finds Salivary Metabolite Shifts in Autistic Children Treated with Medical Cannabis, But Evidence Remains Preliminary

A small observational pilot study identifies 65 candidate salivary biomarkers in children with autism spectrum disorder following medical cannabis treatment, but an extremely small sample size, no control group, heterogeneous cannabis regimens, and significant conflicts of interest make firm clinical conclusions impossible at this stage.

Why This Matters

Autism spectrum disorder currently lacks objective biological markers that can track treatment response, leaving clinicians reliant on subjective behavioral assessments and caregiver reports. Metabolomics, the large-scale study of small molecules in biological samples, represents a scientifically credible approach to filling this gap. At the same time, interest in medical cannabis for ASD-related symptoms has outpaced the available evidence, creating urgent demand for rigorous, biomarker-driven research. This pilot study is among the first to attempt linking cannabis treatment to measurable metabolic changes in children with ASD, making it important to evaluate carefully and honestly.

Clinical Summary

ASD affects approximately 1 in 36 children in the United States, and the heterogeneity of its presentation makes objective treatment monitoring exceptionally difficult. Published in the Journal of Personalized Medicine in 2023, this observational pilot study by Hendren and colleagues at Cannformatics, Inc. used untargeted salivary metabolomics to assess whether medical cannabis treatment produced detectable biochemical shifts in 15 children with ASD (ages 6 to 12). The mechanistic rationale centers on the endocannabinoid system’s involvement in neuroinflammation, neurotransmitter modulation, and synaptic plasticity, all processes implicated in ASD pathophysiology. Saliva samples were analyzed using capillary electrophoresis time-of-flight mass spectrometry and liquid chromatography mass spectrometry, with a separate group of 9 typically developing children serving as a metabolite reference.

The study identified 65 candidate metabolites whose levels shifted toward the typically developing reference range after medical cannabis dosing. Twenty-three of these fell into biologically plausible categories, including anti-inflammatory markers, neurotransmitters, amino acids, and endocannabinoids. Three specific metabolites, N-acetylaspartic acid, spermine, and dehydroisoandrosterone 3-sulfate, have prior literature associations with ASD-related behavioral symptoms. However, the study had no randomization, no placebo arm, no blinding, and no reported correction for multiple comparisons across 65 biomarker candidates. Cannabis regimens varied dramatically, ranging from 0.05 to 200 mg per dose with both THC-dominant and CBD-dominant formulations represented. All authors are affiliated with a cannabis company, and the authors themselves acknowledge that larger, controlled trials are needed before any clinical application can be considered.

Dr. Caplan’s Take

Families raising children with autism are understandably drawn to anything that promises objective measurement of treatment response. The idea that a simple saliva test could tell you whether a therapy is working addresses a real and painful clinical gap. What this study gets right is the research direction: metabolomics is a legitimate tool, and the endocannabinoid system is a defensible target for investigation. But the distance between 65 unvalidated candidate metabolites in 15 pre-selected responders and a clinically useful biomarker panel is enormous. When a parent asks me whether this study means cannabis is helping their child’s brain chemistry normalize, the honest answer is that we simply cannot know from this data.

In practice, I do not recommend medical cannabis for ASD based on studies of this type. When families are already using cannabis products and want guidance, I focus on safety monitoring, consistent dosing, and honest conversation about what the evidence does and does not show. I encourage participation in well-designed clinical trials whenever possible. The metabolomics approach deserves further investment, but it needs independent replication with proper controls before it should influence any treatment decision.

Clinical Perspective

This study sits at the very earliest stage of the biomarker discovery arc, well before validation, replication, or clinical utility assessment. It confirms that salivary metabolomics can detect biochemical differences between ASD and typically developing children, which is consistent with prior metabolomic research in blood and urine. However, it does not establish that cannabis caused the observed shifts, that those shifts are therapeutically meaningful, or that the 65 candidate markers would survive independent replication with appropriate statistical correction. Clinicians should not cite this study as evidence supporting medical cannabis for ASD, nor should they suggest that salivary biomarker testing is available or informative for treatment monitoring at this time.

From a pharmacological standpoint, the extreme heterogeneity of regimens in this study, spanning THC-dominant and CBD-dominant formulations with a 4,000-fold dosing range, makes it impossible to attribute metabolite changes to any specific cannabinoid, dose, or formulation. CBD carries known drug interaction risks through cytochrome P450 inhibition, particularly relevant in children who may be taking antiepileptic medications, SSRIs, or atypical antipsychotics. THC-containing products pose additional neurodevelopmental concerns in pediatric populations. The one actionable recommendation from this study is not therapeutic but methodological: clinicians involved in cannabis research for ASD should advocate for randomized, placebo-controlled designs with standardized formulations and pre-registered biomarker hypotheses.

Study at a Glance

Study Type

Observational pilot study with pre-post design and external typically developing reference group

Population

15 children with ASD (ages 6 to 12; 13 boys, 2 girls) and 9 age-matched typically developing children

Intervention

Medical cannabis (THC-dominant in 40%, CBD-dominant in 60%; dosing range 0.05 to 200 mg per dose; minimum 1 year of prior treatment)

Comparator

No placebo or active comparator; typically developing group served as metabolite reference only

Primary Outcomes

Shift in salivary metabolite levels toward typically developing reference means after medical cannabis dosing

Sample Size

n=15 ASD, n=9 typically developing

Journal

Journal of Personalized Medicine

Year

2023

DOI or PMID

As cited in original publication

Funding Source

All authors affiliated with Cannformatics, Inc., a cannabis company

What Kind of Evidence Is This

This is a small, unblinded observational pilot study using a pre-post design in a single treatment cohort, with a non-randomized typically developing group serving as a metabolite reference rather than a true control. It sits near the bottom of the evidence hierarchy for therapeutic claims. The single most important inference constraint is that the absence of randomization, blinding, and a placebo arm means no observed metabolite shift can be attributed to cannabis treatment rather than to natural biological variation, regression to the mean, maturation, or selection bias inherent in enrolling only apparent responders.

How This Fits With the Broader Literature

Salivary and blood metabolomics studies have previously identified biochemical differences between ASD and typically developing populations, including alterations in amino acid metabolism, oxidative stress markers, and neurotransmitter-related compounds. This study extends that line of inquiry by attempting to link metabolomic shifts specifically to a cannabis intervention, which is relatively novel. However, prior metabolomic biomarker candidates in ASD, including those identified in larger studies such as the work by Needham and colleagues on gut-derived metabolites, have generally not survived independent validation as clinically actionable tools. The cannabis-specific literature in ASD remains dominated by retrospective surveys, case series, and small open-label studies, with very few randomized controlled trials completed to date. This pilot adds a metabolomic dimension but does not materially advance the evidence for or against cannabis as an ASD therapy.

Common Misreadings

The most likely overinterpretation is that this study demonstrates cannabis “normalizes” brain chemistry in children with autism, with salivary biomarkers as proof. This reading exceeds the evidence on multiple levels. The study enrolled only children already judged by their families to be responding to cannabis, so the sample was pre-selected for apparent benefit. The 65 candidate metabolites were identified without reported correction for multiple comparisons, meaning many may be statistical artifacts. A directional shift toward the typically developing mean does not establish that the typically developing mean is the correct therapeutic target, nor that partial normalization produces clinical benefit. The term “Cannabis-Responsive” metabolites, used throughout the paper, implies causation that this uncontrolled design fundamentally cannot support.

Bottom Line

This pilot study identifies salivary metabolomics as a potentially useful research tool for tracking biochemical changes in children with ASD, and it generates hypotheses worth testing in properly controlled trials. It does not establish that medical cannabis causes beneficial metabolic normalization, nor that any of the 65 candidate biomarkers are validated or clinically actionable. Given the tiny sample, absent controls, undisclosed multiple-comparisons handling, and significant conflicts of interest, these findings should be treated as hypothesis-generating only and should not influence clinical practice at this time.

References

1. Hendren RL, Bertoglio K, Hendren RL, et al. Salivary metabolomics of children with autism spectrum disorder treated with medical cannabis. Journal of Personalized Medicine. 2023. (Cannformatics, Inc. affiliated study.)
2. Needham BD, Adame MD, Serena G, et al. Plasma and fecal metabolite profiles in autism spectrum disorder. Biological Psychiatry. 2021;89(5):451-462.
3. Aran A, Cassuto H, Lubotzky A, Wattad N, Hazan E. Brief report: Cannabidiol-rich cannabis in children with autism spectrum disorder and severe behavioral problems – a retrospective feasibility study. Journal of Autism and Developmental Disorders. 2019;49(3):1284-1288.