Pregnancy, Fatty Acids, and the Endocannabinoid System (ECS): One Substrate Story Behind Two Epidemics

12/03/2026 - Swiss Institute for Medical Cannabis Research
------------------------------------

The ECS was never a “cannabis system”

We need to say this out loud, because our silence has cost us decades. The endocannabinoid system, the ECS, was never a “cannabis system”. It is a human regulatory system that cannabis happens to interact with.

And because cannabis carried stigma, the ECS was quietly kept out of mainstream medical education. I think that omission is one of the most harmful things that ever happened to human healthspan.

In 2026, we still discuss non-communicable disease as if the ECS is optional trivia. Meanwhile, it is implicated across metabolism, inflammation, neurobehavior, and pain modulation.

Understanding the substrate-driven ECS, meaning the idea that membrane fatty-acid inventory is a primary controller of ECS tone, has the potential to alter neurodevelopmental and metabolic health like almost nothing else I can think of at population scale.

We also need a cultural correction. The ECS is deeply involved in the modern non-communicable diseases that plague our economies and health burden, including metabolic syndrome and immune dysregulation, yet it is basically absent from public health conversation in 2026. (Dörnyei et al., 2023; Rahaman and Ganguly, 2021)

This must change.

The ECS was never a cannabis system. We stigmatized it out of medicine, then acted surprised when chronic disease exploded.

Two epidemics. One substrate story

We have a habit in modern medicine: we split the world into organs, then split those organs into specialties, then act surprised when the same child later shows up with both neurodevelopmental differences and metabolic dysfunction.

Autism on one side. Obesity, insulin resistance, fatty liver on the other.

Different clinics, different journals, different narratives.

Pregnancy does not care how our specialties are organized. Pregnancy is one continuous developmental build. The fetus is assembled out of available materials, under a tight schedule, with no redo button.

So if we want a unifying upstream driver that can plausibly touch brain wiring and metabolic set points, we should look for a system that:

• Responds to nutrient substrates, not just genes.
• Runs in the brain and the periphery.
• Participates in synapse formation and plasticity.
• Influences appetite, reward learning, adiposity, insulin sensitivity, immune tone.
• Is active during development.

We do not have to invent such a system. We already have one:the ECS.

The thesis (stated carefully)

Here is the model I want to make falsifiable:

An omega-6 heavy substrate environment during pregnancy, especially arachidonic acid dominance relative to DHA and EPA, can load fetal membranes in a way that shifts lipid signaling tone during critical windows, plausibly programming both neurodevelopmental risk and metabolic risk.(Al Sinani et al., 2026, Broselid, 2026)

Not “omega-3 deficiency causes autism”.

Not “seed oils are the devil”. Not “one supplement fixes everything”.

Instead, a substrate model: inventory first, signaling second, phenotype third.

A pregnancy fact we under-appreciate: the placenta is a sorter, not a straw

In adult nutrition discourse, we often talk as if tissues passively reflect dietary intake. Pregnancy is different. The placenta is not a neutral pipe. It is an active sorter. (Al Sinani et al., 2026)

One of the most concrete recent demonstrations is that arachidonic acid is preferentially biomagnified over DHA in fetal erythrocytes in paired maternal-cord samples (172 dyads) (Al Sinani et al., 2026). This happens because AA is critically needed in the developing brain.

This means the fetus does not simply get “what mom eats”. The fetus gets what the placenta prioritizes, and that prioritization has downstream consequences because the fetal brain is built out of lipids and uses lipids as precursor building blocks for lipid mediators, including endocannabinoids.

Pantry, stove, smoke alarm

I keep returning to this mental model because it makes our blind spot obvious to me.

• The membrane is the pantry, meaning the stored ingredients.
• Enzymes are the stove, meaning the machinery that converts inventory into active mediators.
• In this model, the receptors are the smoke alarm, meaning the sensors that detect the mediators.

Most modern pharmacology turns knobs on burners. A large part of nutrition, especially polyunsaturated fat biology, is pantry management.

If we keep arguing about receptors while ignoring inventory, we should not be shocked when interventions underperform.

Why AA matters more than we were taught

Arachidonic acid is not “bad”. It is structural, and it is a precursor molecule for lipid mediators.

The two flagship endocannabinoids are AA-derived:

• 2-AG (2-arachidonoylglycerol)
• AEA (anandamide)

This means membrane AA content is upstream of endocannabinoid synthesis capacity and baseline signaling tone.

Also, the ECS is not just CB1 and CB2. It is an extended lipid mediator family, often called the endocannabinoidome, that shifts depending on which substrates dominate the membrane.

So when we talk about “omega-6 versus omega-3”, we are talking about what the body is physically able to synthesize on demand from its pantry ingredients (cell membrane composition).

The ECS connects “brain outcomes” and “metabolic outcomes”

If we want one mechanistic bridge between neurodevelopmental outcomes and metabolic syndrome, the ECS is one of the few systems that naturally spans:

• Appetite and reward learning
• Peripheral metabolism, including liver and adipose physiology
• Inflammatory tone and immune regulation
• Synaptic plasticity and neurobehavior

The ECS is best explained as a homeostatic suprasystem whose tone depends fundamentally on dietary PUFA composition and tissue phospholipid pools (Broselid, 2026).

Clinically, this matters because metabolic syndrome is explicitly discussed in the biomedical literature through an ECS lens, including endocannabinoid involvement in metabolic control and metabolic syndrome pathogenesis. (Dörnyei et al., 2023)

Immunology matters here too, because endocannabinoids are also discussed as regulators in immune function and immunopathologies. (Rahaman and Ganguly., 2021)

So yes, I will call it out again: in 2026, we still run national conversations about obesity, diabetes, NAFLD, inflammatory disease, and neurodevelopment without mentioning the ECS. That is a structural failure, not a knowledge gap. (Dörnyei et al., 2023; Rahaman and Ganguly., 2021)

We debate calories and willpower while ignoring a lipid signaling system that directly shapes appetite, inflammation, and metabolic set points.

The fatty-acid profile to ECS to health axis

I want to give ammunition in the form I trust most, which is multiple independent lines of evidence.

1) The endocannabinoidome shifts fast in humans

A short-term Mediterranean diet intervention (48 hours) measurably shifted circulating omega-3 derived endocannabinoid-like mediators without weight change, as summarized in your substrate-driven ECS manuscript (Castonguay-Paradis et al., 2020).

This is not about weight loss. It is about signaling plasticity. It shows that the endocannabinoidome responds to substrate patterning on a timescale clinicians can respect.

2) Linoleic acid can drive tissue ECS tone and metabolic phenotype under isocaloric conditions

Rodent work isolating linoleic acid as the variable, increasing dietary LA from 1 to 8 energy percent, increased hepatic AEA and 2-AG and drove adiposity, leptin changes, and feed efficiency under isocaloric conditions (Alvheim et al., 2012).

This is one of the clearest demonstrations that “calories” and “fatty-acid inventory” are not the same lever.

3) Genetic substrate manipulation validates causality

Fat-1 mice, Fat-2 models, and related genetic substrate manipulations show graded ECS control, where tissue AA levels determine hepatic endocannabinoid concentrations and broader phenotypes, independent of diet composition (Kaliannan et al., 2015)

When genetics can shift substrate balance and the phenotype follows, we are no longer dealing with a soft “diet story”. We are dealing with a causal substrate mechanism.

4) A practical integrator biomarker: RBC-AA as slow “substrate memory”

I propose red blood cell arachidonic acid, RBC-AA, as an integrative biomarker reflecting substrate state over the RBC lifespan, around 120 days (Broselid, 2026).

This is how we make the model clinically discussable. We stop arguing about food identity and start measuring inventory. (Broselid, 2026)

Fetal Programming through ECS Lipids

If we need a pair of studies that converts “lipids in pregnancy” from abstract to unavoidable, it is the Sakayori work from 2020 and 2025. (Sakayori et al., 2020; Sakayori et al., 2025)

In the 2020 paper, pregnant mice received a diet high in omega‑6 and low in omega‑3, with a very high n‑6 to n‑3 ratio, and the offspring later showed selectively increased intake of palatable foods, including high‑sucrose and high‑fat diets (Sakayori et al., 2020).

It was not simply “more eating”. They used a progressive ratio task and found a higher breakpoint for sucrose, a clean behavioral readout of increased motivational drive for reward (Sakayori et al., 2020).

They then traced it into the mesolimbic dopamine system. Offspring exposed in utero to the n‑6‑high/n‑3‑low diet showed higher dopamine responses in the nucleus accumbens during sucrose access, and local dopamine receptor blockade in medial accumbens reduced sucrose intake back toward control levels Sakayori et al., 2020).

The part that should make every clinician sit up is timing. The hedonic phenotype wasprogrammed during embryogenesis. In‑utero exposure was sufficient, and postnatal‑only exposure windows did not reproduce the effect Sakayori et al., 2020).

In the newer 2025 paper, the same maternal n‑6‑high/n‑3‑low exposure was taken beyond feeding and tested against a broader behavioral battery (Sakayori et al., 2025).

Here, both male and female offspring of mothers fed the LA‑high / ALA‑low pregnancy diet showed impaired social behaviors, such as fewer contacts with novel mice in social interaction tests, and at the same time enhanced recognition memory in object location tasks (Sakayori et al., 2025).

So we now have a consistent pattern from the same maternal diet model:

• hedonic overeating and higher sucrose motivation (2020)
• altered social behavior and sharpened certain forms of recognition memory (2025)

all arising from a pregnancy-only manipulation, with offspring weaned onto standard chow (Sakayori et al., 2020; Sakayori et al., 2025).

Put together, this tells us that a high omega‑6 and low omega‑3 pregnancy environment can set long‑term parameters for both reward drive and social behavior, which are the kinds of axes we care about in both ASD‑like and metabolic‑related phenotypes.

An important nuance that keeps us honest is that Sakayori did not find simple “more endocannabinoids everywhere” in adult ventral midbrain (Sakayori et al., 2020). That pushes us away from cartoon models and toward a more accurate statement:

Fatty‑acid imbalance in pregnancy can program dopamine‑sensitive circuits and social and cognitive behaviors through lipid‑mediated developmental mechanisms, with the ECS as a central bridge system, but not the only player in the story (Sakayori et al., 2020; Sakayori et al., 2025; Broselid, 2026).

Sakayori now shows that a high omega‑6, low omega‑3 pregnancy diet can program reward drive and social behavior, long after the diet is gone.

“Is that mechanism real in humans?” The best clinical matches we have right now

Human pregnancy cohorts rarely measure the full chain we want: maternal substrates, fetal substrates, endocannabinoidome mediators, and later reward-circuit function. So we triangulate, and we stay honest about limitations.

Here are the clinical and epidemiological anchors that best support “the mechanism is real in humans”, even if the measurement endpoints differ.

Prenatal omega-6 to omega-3 ratio and ADHD symptoms

A 2019 paper titled Prenatal Omega-6:Omega-3 Ratio and Attention Deficit and Hyperactivity Disorder Symptoms links prenatal omega-6 to omega-3 ratio to later ADHD symptoms (López-Vicente et al., 2019).

ADHD is not “sucrose breakpoint”, but it is a dopamine-adjacent neurobehavioral phenotype that supports the idea that prenatal PUFA balance can track later behavioral outcomes that matter (López-Vicente et al., 2019).

Pregnancy omega-3 status and early neurodevelopment

The ECLIPSES study measured maternal omega-3 long-chain PUFA concentrations in early pregnancy and assessed infant neurodevelopment (Bayley-III at 40 days), reporting associations between maternal PUFA status and infant neurodevelopment domains. (Shahabi et al., 2024)

I like this kind of paper because it is measuring neurodevelopment early, not trying to infer prenatal causality from a diagnosis a decade later. (Shahabi et al., 2024)

Prenatal omega-3 status and child metabolic health

Project Viva measured maternal second trimester plasma and cord plasma EPA and DHA and examined associations with offspring adiposity and metabolic markers across childhood (Maslova et al., 2018).

The findings are nuanced, including patterns that were stronger in early childhood and not uniformly persistent into mid-childhood, which is exactly what you expect when early programming interacts with later environment (Maslova et al., 2018).

DHA and reward-linked eating phenotypes in human brains

A Translational Psychiatry paper examined brain responses to palatable food cues and reported that serum DHA moderated external eating behavior in a vulnerable subgroup (Reis et al., 2016).

This is not pregnancy-specific, but it matters because it shows that DHA status relates to reward-linked eating phenotypes in humans, which is the kind of pattern we want to see echoing Sakayori’s findings (Reis et al., 2016; Sakayori et al., 2020).

Epidemiology anchor: fish and autism-related outcomes

A 2024 cohort consortium analysis reported associations between maternal fish consumption and omega-3 supplement use during pregnancy with child autism-related outcomes (Lyall et al., 2024).

I treat this as an anchor, not proof, because diet correlates with many exposures. Still, it supports the idea that the gestational lipid environment matters at scale (Lyall et al., 2024).

Gestational diabetes, the metabolic doorway into fetal neurodevelopment

Gestational diabetes is not just “high glucose”. It is a metabolic regime shift during pregnancy.

A recent 2026 systematic review focused on the impact of gestational diabetes mellitus on fetal neural development, which supports the premise that the gestational metabolic environment can influence fetal brain outcomes (Safwan et al., 2026).

In a substrate-driven ECS worldview, GDM is one of the states where substrate balance, lipid signaling, and inflammatory tone plausibly become higher leverage, not lower. (Broselid, 2026)

GDM is not a glucose problem. It is a pregnancy environment problem, and the fetal brain is living in it.

The claim I want us to stop avoiding in 2026

We keep talking about non-communicable disease as if it is only calories, inactivity, and “inflammation” as an unmeasured cloud. We keep talking about autism as if it is only genetics plus “unknown environment.”

At the same time, the literature is explicit that the ECS is involved in metabolic syndrome pathogenesis and metabolic control processes (Dörnyei et al., 2023).

The literature is explicit that endocannabinoids and cannabinoid receptors participate in immune regulation and immunopathologies (Rahaman and Ganguly., 2021).

Yet in 2026, the ECS is still missing from most mainstream health discussions. That gap is not benign. It shapes what gets measured, what gets funded, and what clinicians feel safe to say.

We need to stop outsourcing the ECS to the cannabis conversation. It belongs in cardiometabolic health, pregnancy care, neurodevelopment, psychiatry, rheumatology, preventive medicine (Dörnyei et al., 2023; Rahaman and Ganguly., 2021; Barrie and Manolios, 2017).

What we should measure next (not medical advice, just scientific hygiene)

If we want to test the “fatty acid profile to ECS to health” axis properly, we need to stop guessing and start measuring inventory and signaling in the same cohorts.

A practical, doable starting list:

• Maternal RBC-AA and AA to EPA early, ideally preconception or first trimester, as an integrative substrate marker.
• Paired maternal and cord fatty-acid profiles to quantify biomagnification patterns and variability (Al Sinani et al., 2026).
• Endocannabinoidome panels in subsets, to connect inventory to mediators.
• Longitudinal tracking of both neurodevelopmental metrics and early metabolic trajectories, because “two sides of the coin” is the hypothesis under test (Shahabi et al., 2024; Maslova et al., 2018).

If we do this, we will either strengthen the substrate-driven ECS model, or we will falsify it and move on. Either outcome is better than continuing to ignore the system.

The way out is not belief. The way out is measurement: inventory, mediators, outcomes, in the same pregnancy cohorts.

Author’s Details

Stefan Broselid, PhD

Scientific Educator & Endocannabinoid Specialist, Swiss Institute for Medical Cannabis Research

Founder,ECS.education

Key References

1. Al Sinani, M., Johnson, M., Crawford, M., & Al Maqbali, M. (2026). Arachidonic acid is preferentially biomagnified over DHA in fetal erythrocytes: Evidence from 172 paired maternal-cord samples in Oman. Prostaglandins, Leukotrienes and Essential Fatty Acids.https://doi.org/10.1016/j.plefa.2026.102726

2. Broselid, S. (2026).The Substrate-Driven Endocannabinoid System: Reconciling Cannabis Pharmacology with Dietary Precursor Control(manuscript in preparation).

3. Lyall K, Westlake M, Musci RJ, et al. Association of maternal fish consumption and ω-3 supplement use during pregnancy with child autism-related outcomes: results from a cohort consortium analysis. Am J Clin Nutr. 2024;120(3):583-592. doi:10.1016/j.ajcnut.2024.06.013

4. Maslova E, Rifas-Shiman SL, Olsen SF, Gillman MW, Oken E. Prenatal n-3 long-chain fatty acid status and offspring metabolic health in early and mid-childhood: results from Project Viva. Nutr Diabetes. 2018;8(1):29. Published 2018 May 25. doi:10.1038/s41387-018-0040-2

5. Dörnyei G, Vass Z, Juhász CB, Nádasy GL, Hunyady L, Szekeres M. Role of the Endocannabinoid System in Metabolic Control Processes and in the Pathogenesis of Metabolic Syndrome: An Update.Biomedicines. 2023;11(2):306. Published 2023 Jan 21. doi:10.3390/biomedicines11020306

6. Reis RS, Dalle Molle R, Ma