A newly published Canadian study has found a strong association between frequent cannabis use during lactation and substantially higher concentrations of delta-9-tetrahydrocannabinol in breast milk, adding some of the clearest evidence yet that maternal use patterns strongly influence infant cannabinoid exposure. Researchers from McMaster University and St. Joseph’s Healthcare Hamilton reported that daily cannabis use while breastfeeding was associated with an estimated 58-fold increase in breast-milk THC concentrations compared with participants in the cannabis-using cohort who did not consume cannabis during lactation. The peer-reviewed study, titled The Hidden Route: Associations Between Maternal Patterns of Cannabis Use and Concentration of Cannabinoids in Breastmilk, was published September 9, 2026, in the Journal of Cannabis Research.
The striking “58 times higher” figure requires careful interpretation. It is a statistical fold-change based on log-transformed cannabinoid concentrations and reflects differences in both whether THC was detectable and how much was present when detected; it is not evidence that infants receive exactly 58 times more THC or experience 58 times greater health risk. Still, the raw concentrations moved in the same direction. Among participants who did not use cannabis while lactating, THC was undetectable in all seven samples. Median THC was 3.9 ng/mL among infrequent users and 53 ng/mL among daily users, with THC measurable in 93 percent of samples from the frequent-use group.
Researchers Analyzed Cannabis Patterns Rather Than Simply Use Versus Non-Use
Investigators initially enrolled 181 lactating participants in Ontario between 2023 and 2024 and collected breast-milk samples between four and 16 weeks postpartum, when mature milk was expected to be established. After follow-up losses and exclusions, 160 milk samples were available for analysis. Fifty participants reported cannabis use and 110 initially reported no use, although laboratory testing detected THC in one self-described non-user, who was subsequently reclassified. The final analysis therefore included 51 cannabis consumers and 109 controls.
Rather than looking only for THC, researchers used high-performance liquid chromatography coupled with tandem mass spectrometry to quantify five cannabis compounds: delta-9 THC, 11-hydroxy-THC, THC-COOH, CBD and cannabinol. Participants also reported when they used cannabis, how frequently they used it and whether they smoked, vaped, ate cannabis products or consumed oils. Daily consumption was classified as “frequent,” while weekly, monthly and occasional use were grouped as “infrequent.” Laboratory personnel analyzing the milk were blinded to participants’ exposure status, reducing the possibility that knowledge of reported cannabis use could influence cannabinoid measurement.
Daily Use Produced the Highest THC Concentrations
Frequency during lactation produced the strongest association observed in the study. Among 28 participants who reported daily cannabis use while breastfeeding, 26 had measurable THC in their milk, with a median concentration of 53 ng/mL. Among 14 infrequent users, THC was detected in 10 samples and the median concentration was 3.9 ng/mL. None of the seven participants who had previously used cannabis but reported no cannabis consumption during lactation had detectable THC in the analyzed milk sample. Statistical modeling translated those differences into a 6.1-fold increase with infrequent consumption and a 58-fold increase with frequent consumption relative to no lactation use.
The confidence interval surrounding the frequent-use estimate was wide—from roughly 16.5-fold to 204-fold—which highlights both the strength of the association and the statistical uncertainty created by relatively small subgroups. The study was not originally powered specifically to examine every pattern of cannabis use, and researchers explicitly describe several of their detailed comparisons as exploratory. Even so, the relationship between lactation frequency and THC concentration was highly statistically significant, and the raw data show a clear gradient from non-detectable levels among non-users to substantially higher concentrations among daily users.
THC Was Detected Far More Often Than Other Cannabinoids
THC was the cannabinoid most frequently identified among cannabis-consuming participants. It was quantifiable in 38 of 51 samples, or 74 percent, while CBN was measured in 41 percent, THC-COOH in 37 percent and both 11-hydroxy-THC and CBD in 29 percent. Among samples with measurable THC, concentrations ranged from 1 to 735 ng/mL, with a median of 42 ng/mL and a mean of 110 ng/mL. The wide range is important because it shows that two people reporting cannabis use can produce dramatically different cannabinoid concentrations in breast milk.
Patterns associated with the other cannabinoids were less consistent. Frequent cannabis use during lactation was associated with approximately 2.4 times higher THC-COOH and 2.7 times higher CBN compared with no lactation use. In contrast, the researchers did not find statistically reliable associations between consumption patterns and either 11-hydroxy-THC or CBD after correcting for multiple comparisons. THC therefore emerged as the clearest marker connecting maternal cannabis behavior with milk cannabinoid exposure.
Smoking Showed the Strongest Association Among Routes of Use
How cannabis was consumed also appeared important. Smoking during lactation was associated with a roughly 47-fold higher THC concentration compared with participants who did not use during lactation. Using two or more methods was associated with a roughly 25-fold increase, while vaping was associated with an approximately 8.6-fold increase. Each of those inhalation-related findings remained statistically significant after adjustment for multiple comparisons.
The study cannot establish that smoking inherently transfers more THC into milk than every other method, however. Only two participants exclusively used oral products such as edibles or oils, making that subgroup far too small for meaningful conclusions. The observed route differences may also reflect differences in frequency, THC potency, dose and individual metabolism rather than administration route alone. Researchers deliberately analyzed use patterns separately because frequency, timing and route overlapped heavily and the sample was too small for a fully adjusted model incorporating all of those factors simultaneously.
Continued Use From Pregnancy Through Breastfeeding Also Mattered
The study found another strong signal among participants who continued cannabis from pregnancy into the postpartum period. Those who reported cannabis use during both pregnancy and lactation had approximately 34-fold higher breast-milk THC concentrations than participants who had used cannabis during pregnancy but stopped before lactation. THC was measurable in 35 of 39 participants who continued across both periods, compared with none of the seven pregnancy-only participants used as the reference group.
Frequency during pregnancy also showed an association: daily prenatal use was linked with approximately 11.5-fold higher postpartum milk THC compared with no pregnancy use. However, the timing of cannabis exposure within pregnancy—early pregnancy, later pregnancy or throughout pregnancy—did not independently remain significant after correction for multiple comparisons. Taken together, the findings indicate that recent and continuing postpartum use appears more important to breast-milk THC exposure than simply having used cannabis sometime during pregnancy.
Earlier Studies Also Show THC Can Accumulate in Breast Milk
The new findings build on several smaller studies suggesting that THC does not behave in human milk like alcohol, where concentrations rise and then decline relatively predictably over several hours. THC is highly lipophilic, meaning it dissolves readily in fat. Human milk contains substantial lipid content, and previous work has found THC concentrations in milk can exceed those in maternal plasma. A 2021 Pediatric Research study involving 20 cannabis-using mothers reported a median milk THC concentration of 27.5 ng/mL compared with 3.7 ng/mL in plasma—a median milk-to-plasma ratio of about seven to one.
The 2024 Lactation and Cannabis Study provided additional evidence of accumulation with repeated consumption. Twenty breastfeeding participants collected multiple milk samples before and after self-directed medicinal cannabis use. THC remained measurable even after at least 12 hours of abstention, and participants who used cannabis repeatedly during the study day generally maintained higher concentrations across the eight-to-12-hour observation period. More episodes of cannabis consumption were significantly associated with greater overall THC exposure in milk.
Studies Disagree on Exactly How Long THC Remains Detectable
How long THC persists in milk remains surprisingly unsettled. A 2018 Pediatrics study detected THC in 63 percent of samples from cannabis-using mothers and found measurable concentrations as long as approximately six days after reported use. Researchers estimated that time since last consumption and number of daily uses both helped predict milk THC concentrations.
A 2021 JAMA Pediatrics study produced an even longer estimate. Among seven women with verified prenatal cannabis use who stopped postpartum, researchers calculated an average THC half-life in milk of approximately 17 days and projected that elimination could take longer than six weeks. In contrast, a smaller 2026 Spanish study of 13 postpartum women found that among participants who discontinued cannabis after delivery, breast-milk THC became non-quantifiable within the first postpartum week. Differences in cannabis frequency, potency, metabolism, milk composition, sampling schedules and analytical sensitivity may help explain these sharply divergent results.
The inconsistent clearance estimates also make simple “pump and dump” timing strategies difficult to support scientifically. Unlike alcohol, THC may remain stored in maternal fat and mammary tissue and be released gradually. The 2024 Lactation and Cannabis investigators concluded that their data did not identify a clear period after cannabis use when breastfeeding could reliably occur with zero THC exposure.
The Study Did Not Measure Effects on Infants
Perhaps the most important limitation is what the new research did not examine. Investigators measured cannabinoids in breast milk; they did not measure infant blood THC concentrations, neurodevelopment, sleep, feeding, growth, behavior or other clinical outcomes. The 58-fold association therefore describes maternal milk chemistry, not evidence of a 58-fold increase in developmental risk. The authors explicitly state that larger studies are needed to determine the implications for infant development.
Evidence on actual outcomes from THC exposure exclusively through breastfeeding remains sparse. The NIH LactMed database notes an older study in which daily or near-daily maternal cannabis use was associated with lower motor-development scores at one year, while studies of occasional use did not detect clear effects. Those studies were small, could not reliably separate prenatal from postnatal exposure and were conducted when cannabis products were generally less potent than many products available today. LactMed therefore concludes that the existing evidence cannot rule out long-term harm.
Current Medical Guidance Emphasizes Reducing Cannabis Exposure
Major medical organizations continue to recommend avoiding or reducing cannabis during lactation while also recognizing the substantial benefits of breastfeeding. The CDC advises breastfeeding parents to avoid cannabis and CBD because cannabinoids can pass into milk and infant effects remain uncertain. The American College of Obstetricians and Gynecologists similarly recommends cessation during pregnancy and lactation, but its updated 2025 guidance emphasizes that continued cannabis use is not itself a contraindication to breastfeeding and says breastfeeding should not automatically be discouraged.
The Academy of Breastfeeding Medicine takes a similar harm-reduction approach. Its clinical protocol encourages cessation or reduction of cannabis use, while recommending shared decision-making for people who continue using cannabis and want to breastfeed. It specifically advises clinicians to consider frequency, route, product type and potency—the same variables that the new Canadian study now shows can be strongly associated with breast-milk THC concentrations.
A Stronger Signal That Frequency Matters
The greatest contribution of the new Journal of Cannabis Research study may be its ability to put numbers around something previous research had repeatedly suggested: frequency of postpartum cannabis consumption matters considerably. Daily users did not merely have THC detected somewhat more often. Their milk showed a median THC concentration more than ten times that of infrequent users, and statistical modeling estimated a 58-fold difference relative to cannabis consumers who were not using during lactation. Smoking, vaping and use across both pregnancy and lactation were also strongly associated with higher THC concentrations.
At the same time, the study does not establish a toxic threshold, prove that a specific milk THC concentration harms an infant or determine whether occasional exposure is safe. Its subgroup sizes were modest, cannabis dose and potency were self-reported rather than standardized, and the statistical models were not adjusted for every potential confounder. What it does provide is increasingly persuasive pharmacological evidence that cannabis exposure through breast milk is dose-pattern dependent rather than simply present or absent. For researchers and clinicians trying to move beyond broad yes-or-no questions about cannabis and breastfeeding, that may be the study’s most useful finding.
