
Human breast milk contains naturally occurring compounds that interact with the body’s endocannabinoid system, but describing them simply as “cannabinoids in breast milk” can easily create confusion. The compounds found naturally in milk are endocannabinoids—lipid signaling molecules produced by the human body—not THC, CBD, or other phytocannabinoids produced by the cannabis plant. Among the most important are 2-arachidonoylglycerol (2-AG) and anandamide (AEA, or arachidonoylethanolamide). These molecules participate in the same biological signaling network that responds to THC, which is why the discovery of endocannabinoids in human milk has attracted considerable interest from researchers studying infant nutrition, feeding, metabolism, and neurodevelopment.
The distinction is essential. Human beings possess an endocannabinoid system regardless of whether they have ever consumed cannabis. The system includes endogenous signaling molecules such as AEA and 2-AG, cannabinoid receptors—primarily CB1 and CB2—and enzymes that synthesize and break down these molecules. Research has established that this signaling system is active very early in life, including during embryonic and neonatal development. Endocannabinoids have been detected in maternal milk, and animal research suggests that endocannabinoid signaling is particularly important during the transition from fetal nutrition to independent suckling after birth.
What Endocannabinoids Have Been Found in Human Milk?
Modern analytical chemistry has confirmed that human milk contains an entire network of endocannabinoids and chemically related lipid mediators. A 2016 study led by Junfang Wu used liquid chromatography–tandem mass spectrometry to investigate endocannabinoids, related compounds, and oxylipins in human milk. Researchers detected 12 endocannabinoid or endocannabinoid-related compounds, with 2-AG occurring at the highest concentration among the measured endocannabinoids. The study was also important because it demonstrated that measured concentrations could change significantly depending on how milk samples were stored, meaning seemingly small differences in laboratory handling can influence reported endocannabinoid levels.
A more detailed 2018 study by Adriana Gaitán and colleagues analyzed transitional and mature human milk from 24 lactating women and characterized what the researchers called the endocannabinoid metabolome. They detected AEA and 2-AG alongside a wider family of related compounds, including palmitoylethanolamide, oleoylethanolamide, docosahexaenoylethanolamide, eicosapentaenoylethanolamide, several monoacylglycerols, and their fatty-acid precursors. Interestingly, most of the compounds measured did not differ significantly between transitional and mature milk in that cohort, suggesting that this lipid-signaling network may remain an ongoing component of human milk rather than appearing only during the earliest days of lactation.
2-AG Appears to Be Especially Abundant
Of the classical endocannabinoids identified in milk, 2-arachidonoylglycerol appears to be especially prominent. A 2021 study analyzed breast milk collected from 36 women between four and eight weeks postpartum. The researchers measured milk during feeds occurring throughout a 24-hour period and found an average 2-AG concentration of about 59 ng/mL in the normal-BMI group. AEA was below the analytical limit of detection in that particular study. The researchers also observed a daily rhythm: 2-AG concentrations were significantly higher during the daytime than at night.
A larger longitudinal study published in 2025 adds further evidence that milk endocannabinoids change during lactation. Researchers followed 92 women and measured AEA and 2-AG at approximately 2–8, 28–47, and 88–119 days postpartum. AEA peaked during the middle sampling period, while 2-AG concentrations were higher during the later two periods than during the first days after delivery. In that study, 2-AG concentrations were approximately 1,000 times greater than AEA concentrations. The absolute numbers differed substantially from those reported in some earlier studies, which highlights an important issue in this field: collection techniques, sample handling, storage, analytical methods, maternal characteristics, and timing can all influence measured endocannabinoid concentrations.
Why Would Endocannabinoids Be Present in Breast Milk?
One possibility is that milk-borne endocannabinoids participate in the regulation of feeding and energy balance during a developmental period when reliable food intake is essential for survival. The adult endocannabinoid system is strongly involved in appetite, metabolism, energy storage, and reward-related feeding. Researchers therefore began investigating whether a similar biological pathway might help newborn mammals initiate and maintain suckling. Early animal experiments produced striking results.
In a landmark 2001 study, Ester Fride and colleagues blocked CB1 cannabinoid receptors in newborn mice. When CB1 signaling was inhibited immediately after birth, milk ingestion fell dramatically, growth was severely disrupted, and some animals failed to survive. Administration of cannabinoids capable of activating the system partly reversed these effects. Later experiments using another CB1 antagonist again produced dose-dependent interference with milk intake, growth, body temperature, and survival. These studies led researchers to propose that endocannabinoid activation of CB1 receptors is important for normal initiation of suckling in newborn mice.
The Endocannabinoid System and the Newborn Suckling Response
The proposed mechanism appears to involve more than simply making a newborn “hungry.” Fride and other researchers suggested that CB1 signaling may participate in the neural and muscular processes necessary for effective suckling, including activation or coordination of oral motor structures and tongue musculature. Reviews of developmental endocannabinoid biology have consequently described the early postnatal period as one of several developmental windows in which cannabinoid signaling appears particularly important.
These results are biologically compelling, but they need to be interpreted carefully. Much of the strongest evidence connecting endocannabinoids directly to the initiation of suckling comes from mouse experiments, not controlled studies in human infants. Detecting 2-AG in human breast milk supports the plausibility of a related biological role, but it does not establish that a specific concentration of milk 2-AG directly controls human infant appetite or feeding behavior. Researchers studying human milk continue to describe the precise physiological functions of these compounds as incompletely understood.
Endocannabinoids and Early Brain Development
The endocannabinoid system also participates in nervous-system development. CB1 receptors and endogenous cannabinoid signaling appear during prenatal and postnatal life, where experimental research has connected them with processes including neural progenitor development, neuronal migration, axonal growth, synapse formation, and the organization of developing neural circuits. This does not mean that breast-milk endocannabinoids have been proven to control each of these processes directly. Rather, it shows that infants are developing within a biological environment in which endocannabinoid signaling already performs multiple regulatory functions.
This distinction is particularly important when discussing breast milk. Scientists know that AEA, 2-AG, and related lipids are present in milk, and they know independently that the endocannabinoid system contributes to early development. What remains much less certain is how much of an infant’s endocannabinoid activity is influenced specifically by compounds consumed in human milk, how these compounds are absorbed and metabolized in the gastrointestinal tract, and whether differences in milk concentrations have measurable long-term effects on neurological development. These questions remain active areas of nutritional and developmental research.
Human Milk Contains More Than AEA and 2-AG
Another important discovery is that human milk contains a much broader lipid-signaling system than the two best-known endocannabinoids alone. The 2018 metabolomic study identified palmitoylethanolamide (PEA) and oleoylethanolamide (OEA) along with several ethanolamides and glycerol esters derived from long-chain fatty acids. Some of these compounds do not activate cannabinoid receptors in exactly the same manner as AEA and 2-AG, but they participate in interconnected signaling pathways involving metabolism, inflammation, satiety, and lipid physiology.
The raw materials for these molecules are also significant. AEA and 2-AG are derived from arachidonic acid-containing membrane lipids, connecting the endocannabinoid system with the broader biology of dietary and tissue fatty acids. Human milk contains long-chain polyunsaturated fatty acids that are important during infant development, including arachidonic acid and DHA. Researchers are increasingly studying milk as a complex signaling system rather than simply a mixture of calories, protein, fat, vitamins, and minerals. Endocannabinoids belong to a much larger collection of hormones, immune factors, lipids, enzymes, and other biologically active molecules delivered during breastfeeding.
Endocannabinoid Levels Can Change Over Time
Breast milk is not chemically identical from one feeding to the next or from the first week postpartum to several months into lactation. Endocannabinoids appear to be part of this dynamic composition. The 2025 study of 92 women found significant changes in both AEA and 2-AG across approximately the first four months postpartum. Researchers also reported associations between milk endocannabinoid concentrations and maternal characteristics such as pre-pregnancy body mass index, gestational weight gain, and milk triglycerides.
The 2021 research additionally found evidence that 2-AG concentrations vary across the day. Together, these studies raise the possibility that the endocannabinoid composition of milk responds to maternal physiology, metabolism, milk composition, circadian rhythms, or changing infant requirements. Researchers cannot yet say precisely why these variations occur or whether they represent deliberate biological regulation of infant feeding. What they do show is that talking about a single fixed “endocannabinoid level” in breast milk is probably unrealistic. Human milk is a changing biological fluid whose chemistry reflects both maternal and infant physiology.
Natural Endocannabinoids Are Not the Same as THC or CBD
The discovery of natural endocannabinoids in human milk has sometimes been misrepresented as evidence that THC or cannabis consumption during breastfeeding is biologically equivalent to normal lactation. It is not. AEA and 2-AG are endogenous human signaling molecules. THC and CBD are phytocannabinoids produced by cannabis. Although THC can activate some of the same cannabinoid receptors as endocannabinoids, it has different pharmacokinetics, concentrations, persistence, receptor effects, and patterns of exposure.
When a breastfeeding person consumes cannabis, THC can enter breast milk because it is highly lipophilic and associates readily with fatty tissues and milk. A 2018 Pediatrics study involving 50 women found measurable THC in 63% of 54 milk samples, in some cases up to approximately six days after the last reported cannabis use. A more recent prospective study of 20 frequent cannabis users found measurable THC even after at least 12 hours of abstinence and observed higher milk concentrations with repeated use during the day.
Current LactMed guidance likewise notes that THC is excreted into human milk and that studies have reported detection lasting from days to, in some cases, more than six weeks depending on use patterns and study methodology. Long-term developmental effects from exposure through milk remain inadequately defined. The existence of normal endocannabinoids in breast milk therefore should not be interpreted as evidence that exposure to plant-derived THC during breastfeeding is harmless or physiologically equivalent.
What Scientists Still Do Not Know
The evidence that naturally produced endocannabinoids occur in human milk is now strong. Multiple analytical studies have independently detected 2-AG, AEA, and numerous endocannabinoid-related compounds. What remains uncertain is their precise biological importance after ingestion by a human infant. Researchers still need to determine how much survives digestion, whether these lipids act mainly within the infant gastrointestinal tract or enter systemic circulation, how infant age affects their metabolism, and whether differences between mothers produce meaningful developmental consequences.
Human studies are also relatively small. The 2018 metabolomic investigation involved 24 mothers, the 2021 circadian study included 36, and even the larger 2025 longitudinal study followed fewer than 100 women. Differences in collection and storage add another challenge because 2-AG and related lipid concentrations can change after milk has been expressed. Wu and colleagues showed that storage conditions significantly altered measured concentrations, emphasizing the need for carefully standardized methods before investigators can confidently compare results among populations.
Final Thoughts on Endocannabinoids in Breast Milk
One of the more fascinating discoveries in modern breast-milk research is that the human body’s own cannabinoid signaling system extends into lactation. Human milk naturally contains 2-AG, anandamide, and a larger family of endocannabinoid-related lipid mediators. Evidence from animal models suggests that cannabinoid signaling plays a critical role in initiating neonatal suckling, while human studies show that milk endocannabinoid concentrations vary with lactation stage, time of day, milk lipid content, and maternal metabolic factors.
The discovery also provides an instructive example of why the word “cannabinoid” extends far beyond marijuana. Cannabis did not create the human endocannabinoid system; rather, compounds in cannabis happen to interact with an ancient biological signaling network already present in humans and many other animals. The body’s own cannabinoids participate in normal physiology from the earliest stages of life, and breast milk appears to be one component of that system.
At the same time, endogenous cannabinoids such as AEA and 2-AG should never be conflated with THC entering milk after cannabis consumption. They are different molecules originating from different sources and producing different patterns of biological exposure. The most scientifically accurate conclusion is therefore both intriguing and restrained: human breast milk naturally contains endocannabinoids, these molecules may contribute to feeding and early development, and scientists are only beginning to understand the full significance of this remarkable signaling system.






