Domain-specific signal, not global impairment
This systematic review of 72 observational studies finds that prenatal cannabis exposure is most consistently linked to behavioural, attention, and executive function difficulties, while global cognition, language, and motor development show little to no association after confounder adjustment.
Prenatal Cannabis Exposure: A Pattern of Selective Vulnerability, Not Global Harm
A systematic review of 72 observational studies spanning four decades finds that children exposed to cannabis in utero do not show broad cognitive impairment on the whole, but they do show a recurring pattern of difficulty with behavioural regulation, attention, and executive function, a distinction that reframes how clinicians should counsel pregnant patients about cannabis use without overstating or understating the risk.
What This Review Teaches Us About Prenatal Cannabis Exposure
The headline finding of this review is not that prenatal cannabis exposure is harmless, nor that it causes sweeping developmental damage. It is that the signal, where one exists, clusters in specific domains. Behavioural dysregulation, attention-related difficulties, and executive functioning showed negative associations in 73%, 69%, and 70% of studies in those domains respectively. Global cognition, language, and motor development told a different story: 84% of cognitive studies, 92% of language studies, and 60% of motor studies found no significant association after adjustment for confounders.
This domain-specific pattern has biological plausibility. The endocannabinoid system is deeply involved in neuronal proliferation, migration, and synaptogenesis, with particular density in prefrontal and fronto-limbic circuits that govern attention, impulse control, and emotional regulation. A disruption concentrated in those circuits, rather than diffusely across the brain, would be expected to manifest as problems with self-regulation and attention rather than as a uniform drop in IQ.
The review also surfaces a timing and dose signal worth taking seriously, though not worth overselling. Several cohorts found stronger associations with heavier or more persistent use, particularly use that continued after the mother became aware of the pregnancy. Negative behavioural findings also increased with child age, from roughly half of neonatal and early infancy studies to 93% of school-age studies. Whether this reflects an true developmental unmasking of early vulnerability or simply that behavioural and executive instruments become more sensitive and specific as children age is not something this review design can resolve.
Why This Matters
If you used cannabis during pregnancy, or are weighing whether to use it now, this review does not show that your child will have lower intelligence or language delays. It does show that researchers have repeatedly found associations with attention, behaviour, and self-regulation difficulties later in childhood, especially with heavier or ongoing use. This is not a diagnosis or a prediction for any individual child, and the review cannot tell you whether cannabis caused these differences or whether other factors tied to use, like co-occurring tobacco use or stress, played a role. It is reasonable to discuss your specific situation with your obstetric provider rather than relying on either reassurance or alarm from a single review.
This review offers a framework for counseling that avoids both extremes. Rather than warning patients broadly about global neurodevelopmental harm, clinicians can accurately describe the evidence as pointing toward a more circumscribed risk profile concentrated in behavioural regulation, attention, and executive function, with weaker and less consistent signal for IQ, language, and motor outcomes. The suggestion of dose and timing effects, with heavier and more persistent use after pregnancy awareness showing stronger associations, supports counseling that encourages cessation as early as possible rather than treating any prior use as fixed and unmodifiable risk. Confounding remains a serious unresolved issue: the review found no consistent difference in confounder adjustment between studies reporting positive associations and those reporting null findings, so the attenuation seen after adjustment in some cohorts should not be read as proof that the raw associations were purely confounded.
For research and policy audiences, the central message is methodological as much as clinical. A planned meta-analysis was abandoned because of heterogeneity in exposure definitions, outcome measures, and confounder adjustment strategies, which means that 40 years of research on this question still cannot produce a single pooled effect estimate. This is itself a policy-relevant finding: future funding and study design should prioritize standardized exposure assessment (ideally biological rather than self-report alone), harmonized outcome batteries across developmental stages, and transparent, comparable confounder models, so that the next generation of research can move beyond narrative synthesis toward answers that support firmer clinical guidance.
Study Snapshot
| Population | Pregnant women exposed to cannabis or cannabinoid-containing products during pregnancy, and their offspring followed from birth to 18 years |
| Design | Systematic review of 72 prospective and retrospective observational studies (PRISMA 2020, PROSPERO-registered) |
| Exposure assessment | Varied across studies: maternal self-report, biological testing (urine, meconium), or clinical diagnosis of cannabis use disorder |
| Comparator | Offspring with no reported prenatal cannabis exposure, or lower-exposure comparison groups |
| Primary domains assessed | Cognition, behaviour, language, attention/ADHD, motor development, executive function, visuospatial development |
| Synthesis method | Narrative synthesis by developmental stage and domain; planned meta-analysis abandoned due to heterogeneity |
| Risk of bias | 70 of 72 studies rated low risk of bias (NOS score 7 to 9); 2 rated moderate |
| Core question | Does prenatal cannabis exposure predict adverse neurodevelopmental outcomes in offspring, and if so, in which domains and at what ages? |
Prenatal cannabis exposure shows a consistent observational association with behavioural dysregulation, attention difficulties, and executive function problems, most clearly in school-age and adolescent cohorts, but does not show a consistent association with global cognitive impairment, language delay, or motor development once confounders are addressed; the evidence supports cautious counseling toward cessation, not alarmist claims of broad developmental harm.
The findings cannot establish causation, and the magnitude of risk for any individual child remains unquantified.
What This Paper Looked At
The investigators conducted a PRISMA-compliant, PROSPERO-registered systematic review searching PubMed/MEDLINE, Web of Science, and PsycINFO from database inception through October 2025, supplemented by hand-searching of review reference lists. They included observational studies, prospective or retrospective, that assessed neurodevelopmental outcomes in offspring from birth to 18 years following prenatal exposure to cannabis, THC, CBD, or diagnosed cannabis use disorder, compared against unexposed or lower-exposure groups.
Of 1560 records identified, 924 remained after duplicate removal, 112 underwent full-text review, and 72 studies ultimately met inclusion criteria, spanning publications from 1985 to 2025. Two reviewers independently extracted data on exposure assessment method (self-report, biological testing, or administrative diagnosis), outcome domain, assessment instrument, age at assessment, and confounder adjustment, using the most fully adjusted statistical model reported in each study. Risk of bias was assessed using the Newcastle-Ottawa Scale.
A quantitative meta-analysis was originally planned per the registered protocol, but the authors abandoned it because of substantial heterogeneity in how exposure was defined and measured, how outcomes were assessed, and how confounders were handled across studies. Instead, findings were organized narratively by neurodevelopmental domain (cognition, behaviour, language, attention/ADHD, motor, executive function, visuospatial) and by developmental stage (neonatal, early infancy, preschool, school age, adolescence).
What the Paper Found
Cognition showed the weakest and most inconsistent signal. Of 32 studies assessing cognitive outcomes, 84% (27/32) found no significant association with global cognition after confounder adjustment, while 16% (5/32) reported negative associations and none reported positive associations. Where effects were found, they tended to be concentrated in heavier-exposure subgroups: one cohort reported IQ-related reductions of roughly 0.9 to 8.9 points per joint per day (p<0.05) in verbal reasoning, short-term memory, and abstract/visual reasoning, and another found a Composite IQ reduction of beta=-5.04 (p<0.05), quantitative reasoning reduction of beta=-8.18 (p<0.01), and short-term memory reduction of beta=-4.51 (p<0.05) among children exposed to heavy prenatal cannabis use.
Behaviour showed the most consistent pattern of all domains studied. Of 30 studies assessing behavioural outcomes, 73% (22/30) reported a negative association between prenatal cannabis exposure and behavioural outcomes, 23% (7/30) found no association, and 3% (1/30) reported a positive association. This negative pattern strengthened with child age, rising from roughly half of neonatal and early infancy/preschool studies to 93% of school-age studies and 80% of adolescent studies.
Effects were generally modest in size and sensitive to adjustment. Across cognitive findings, reported associations typically reflected small reductions in IQ, verbal reasoning, memory, or academic achievement, concentrated among children exposed to heavier cannabis levels earlier in pregnancy. These effects were frequently attenuated after adjustment for maternal tobacco and alcohol use, socioeconomic disadvantage, and maternal mental health, though the review found no systematic difference in the extent of confounder adjustment between studies reporting positive findings and those reporting null results.
Age at assessment mattered for both direction and detection of effects. For cognitive outcomes, negative associations were essentially absent in neonatal and early infancy studies but appeared in about 10% of preschool studies, 23% of school-age studies, and roughly a third of adolescent studies. The same age-dependent emergence pattern appeared even more strongly in behavioural outcomes, suggesting that some vulnerabilities may not be detectable, or may not yet be expressed, until children reach school age and beyond.
Risk of bias was generally low across the included evidence base. Seventy of the 72 included studies scored 7 to 9 on the Newcastle-Ottawa Scale, indicating low risk of bias by that instrument’s criteria, with only 2 studies rated as moderate quality. This reflects reasonably careful design at the level of individual studies, even though it does not resolve the cross-study heterogeneity that prevented pooled quantitative synthesis.
How Strong Is This Evidence?
This is a systematic review of observational cohort, case-control, and cross-sectional studies, not a meta-analysis and not a randomized trial, a design that was never possible given the exposure under study. The consistency of the behavioural, attention, and executive function findings across 72 studies and multiple decades is notable and was corroborated by generally low risk-of-bias scores on the Newcastle-Ottawa Scale. However, the authors themselves abandoned their planned quantitative pooling because of heterogeneity in exposure definitions, outcome instruments, and confounder adjustment, which means no single effect size can be reported, and residual confounding by co-occurring substance use, socioeconomic disadvantage, and maternal mental health cannot be excluded from any individual study or from the pattern as a whole.
Where This Paper Deserves Skepticism
The authors planned a quantitative meta-analysis in their registered protocol but abandoned it due to heterogeneity in exposure definitions, outcome measures, and confounder adjustment. This means readers get a count of studies reporting positive, negative, or null associations, not a pooled estimate of how large any effect might be. A domain can show a ‘consistent’ direction of association across many small, imprecise studies without that association being clinically meaningful in magnitude.
Some included studies relied on maternal self-report, others on urine or meconium testing, others on clinical diagnosis of cannabis use disorder. These methods differ enormously in sensitivity, timing window, and susceptibility to underreporting, and pooling findings across such different exposure definitions as though they measure the same thing introduces an unquantifiable layer of noise and potential bias.
The review explicitly found no consistent difference in the degree of confounder adjustment between studies that found significant associations and studies that found none. This undermines any simple narrative that better-adjusted studies systematically show smaller effects; it suggests the field has not converged on a standard confounder model, making cross-study comparison unreliable.
Negative behavioural associations rose from about half of early studies to 93% of school-age studies. This could reflect a genuine developmental unmasking of vulnerability as executive and behavioural demands increase with age, but it could equally reflect that behavioural rating scales and diagnostic instruments become more sensitive, more numerous, and more specific as children get older. The review cannot distinguish between these explanations.
The introduction itself raises the possibility that persistent cannabis use during pregnancy may partly reflect maternal self-medication for underlying psychiatric or neurodevelopmental conditions that independently affect offspring outcomes through shared genetic or environmental pathways. This is a structural limitation of observational designs that no amount of statistical adjustment can fully resolve.
Included studies span 1985 to 2025 and are concentrated in North America and Europe, with additional but smaller contributions from Australia and other regions. Changes in cannabis potency, legal status, and patterns of use over four decades, combined with geographic concentration, limit how confidently findings generalize to today’s cannabis products or to populations outside the countries most represented in the literature.
What This Paper Does Not Show
This review does not show that prenatal cannabis exposure causes behavioural, attention, or executive function problems; all included evidence is observational and cannot establish causation.
It does not show a quantifiable overall effect size for any outcome domain, because heterogeneity across studies prevented the planned meta-analysis.
It does not show that global cognition, language, or motor development are unaffected by prenatal cannabis exposure in all children; it shows that most included studies did not detect a significant association in these domains after adjustment.
It does not show that any observed associations are independent of residual confounding, since adjustment practices varied widely and did not differ systematically between studies with positive and null findings.
It does not show outcomes specific to current cannabis products, potency levels, or routes of administration, since included studies span four decades of evolving cannabis use patterns.
It does not show individual-level risk prediction; study-level associations describe group patterns and cannot be applied to forecast outcomes for a specific pregnancy or child.
Where This Fits in the Prenatal Cannabis Counseling Debate
Cannabis use during pregnancy has risen alongside expanding legalization and growing public perception that cannabis is a relatively benign substance, including for alleviating pregnancy-related symptoms like nausea. This review arrives into a counseling landscape that has often defaulted to either blanket reassurance or blanket alarm, neither of which is well supported by the actual pattern of evidence synthesized here.
The domain-specific framing offered by this review, selective vulnerability in behavioural regulation, attention, and executive function rather than global impairment, aligns with a broader shift in developmental neuroscience away from single-number outcomes like IQ and toward more granular profiling of specific cognitive and regulatory systems. This matches the paper’s own discussion of overlapping prefrontal and fronto-limbic circuits underlying attention, impulsivity, and self-regulation.
The suggestion of dose-response and timing effects, with associations strengthening among women who continued heavier use after becoming aware of the pregnancy, supports a counseling message centered on early cessation rather than treating any lifetime history of use as a fixed, unchangeable risk factor. This is a clinically actionable distinction even though the review cannot quantify how much risk reduction accompanies earlier cessation.
The value of this review lies less in any single number and more in the discipline it imposes on the conversation. Forty years of research, 72 studies, and the investigators still could not pool the data into a meta-analysis. That is not a failure of this review; it is an honest reporting of a field that has not agreed on how to measure exposure or adjust for the factors that travel alongside it. Any clinician or advocate who cites a single effect size from this literature as though it were settled is overstating what the underlying studies can support.
What makes the behavioural, attention, and executive function findings worth taking seriously is not any one study but the consistency of direction across a genuinely large and methodologically varied set of cohorts, most of them individually rated low risk of bias. That consistency is a real signal. It is not, however, a license to claim that cannabis definitively damages a child’s capacity for self-regulation. Consistency of direction across observational studies narrows the space of plausible explanations; it does not eliminate confounding, and the review’s authors are explicit that confounder adjustment did not systematically differ between positive and null studies.
The null findings for global cognition and language deserve equal weight to the positive behavioural findings, and they are easy to lose in a public conversation that tends to remember alarming results and discard reassuring ones. A review that finds 84% of cognitive studies and 92% of language studies showing no significant association is making a real claim: whatever is happening with prenatal cannabis exposure, it does not look like the kind of diffuse neurotoxic injury that some early case reports implied.
For counseling purposes, the most defensible position sits between dismissal and alarm. Pregnant patients deserve to know that the evidence does not support claims of broad developmental harm, and they also deserve to know that a specific, recurring pattern involving behaviour, attention, and executive function has shown up across a large and reasonably well-conducted body of research, particularly with heavier and more persistent use. Neither half of that message should be dropped to make the other feel more comfortable.
What a Careful Reader Should Take Away
A careful reading of this review supports a measured conclusion: prenatal cannabis exposure is associated, in observational data, with a selective pattern of difficulty in behavioural regulation, attention, and executive function, while global cognitive, language, and motor outcomes show far less consistent association once confounders are considered. This pattern held across 72 studies conducted over four decades and multiple countries, with generally low risk of bias by Newcastle-Ottawa criteria.
The review cannot tell us whether this pattern reflects a causal effect of cannabinoid exposure on fetal brain development, a marker of maternal characteristics that independently affect child outcomes, or some combination of both. The abandonment of the planned meta-analysis because of study heterogeneity is itself an important finding, signaling that the field needs standardized exposure measurement and outcome assessment before firmer conclusions can be drawn.
For now, the most responsible use of this evidence is in shaping cautious, honest clinical counseling rather than firm risk quantification. Patients should understand that heavier and more persistent prenatal cannabis use has been associated with behavioural and executive difficulties later in childhood in multiple cohorts, without overstating this as proof of causation or extending it into claims about global intellectual harm that the data do not support.
Read This Paper Through Eight Different Lenses
This systematic review synthesizes 72 observational studies on prenatal cannabis exposure and offspring neurodevelopment, finding that global cognitive impairment is rarely supported while behavioral, attention, and executive function domains show more consistent negative associations.
Patient Takeaway
If you used cannabis during pregnancy, this review does not show that your child will have lower intelligence or global developmental delay. Across 32 studies looking at cognition, 84% found no association with IQ or general cognitive ability once other factors were accounted for. That is a meaningfully different message than the one often implied in public health messaging, and it deserves to be stated plainly rather than softened.
What the review does suggest is a different kind of signal: children exposed prenatally showed more consistent differences in behavior, attention, and executive functioning, particularly at school age and adolescence. This does not mean a diagnosis is coming, and it does not mean nothing is happening. It means the honest answer right now is uncertainty about degree and mechanism, not certainty about either harm or safety. Parents deserve that nuance instead of a reassurance that overstates what is known or a warning that overstates what is proven.
Clinician's POV
This review supports counseling patients that prenatal cannabis exposure is not clearly linked to global cognitive or language impairment, but is more consistently associated with behavioral dysregulation, attention difficulties, and executive function differences, especially at older ages. That domain-specific pattern, rather than a blanket neurodevelopmental warning, is the clinically useful takeaway and should shape how risk is framed in prenatal counseling.
This does not license treating cannabis use in pregnancy as low-risk or inconsequential. The behavioral and attentional signal was present in 69 to 73% of studies examining those domains, which is substantial even without a clean dose-response curve. It also does not license specific predictions for an individual patient's child, since the evidence is drawn from heterogeneous cohorts with variable exposure measurement and confounder adjustment, not from a tool that forecasts outcomes case by case.
A Skeptical Read
The headline finding, that cognition is largely spared while behavior and attention are affected, is intuitively appealing because it sounds precise and domain-specific. But precision in categorization is not the same as precision in causal attribution. Cannabis use during pregnancy correlates with a cluster of factors, including maternal psychiatric symptoms, tobacco and alcohol co-use, and socioeconomic disadvantage, that independently predict behavioral and attentional outcomes in children. The review itself notes that confounder adjustment varied considerably across studies without a clear pattern distinguishing positive from null findings.
A fair reading would ask whether the domain split reflects biology or measurement sensitivity: behavioral and attention measures using parent or teacher report (CBCL, TRF) may be more susceptible to reporting bias linked to maternal substance use disclosure or household instability than a standardized IQ test administered by a blinded examiner. The review does not resolve whether the domain-specific pattern is a true biological signature of endocannabinoid disruption or an artifact of which outcomes are easier to confound.
Study Critic
The review's central methodological limitation is its own stated reason for abandoning meta-analysis: heterogeneity in exposure definitions (self-report versus urine versus meconium testing), outcome instruments, developmental age at assessment, and confounder sets made quantitative pooling infeasible. The narrative synthesis that replaced it is useful for cataloguing direction of effect but cannot produce a weighted estimate of magnitude, and percentages of studies reporting negative associations (such as 73% for behavior) are not equivalent to a pooled effect size with a confidence interval.
Exposure misclassification is a structural problem across the included literature: self-reported cannabis use during pregnancy is subject to underreporting given legal and social stigma, and biological testing (urine, meconium) captures only a narrow exposure window rather than cumulative or trimester-specific dose. Because the review relies on the most adjusted model reported in each primary study rather than a standardized confounder set, comparability across studies is limited, and the appearance of a domain-specific pattern could partly reflect which domains happen to be measured in better-adjusted cohorts.
Compared to Past Research
The paper situates itself against prior literature showing that prenatal cannabis exposure is well established as a risk factor for short-term obstetric and neonatal outcomes, including low birth weight, stillbirth, and preterm delivery, citing references 5 through 7. It frames long-term neurodevelopmental evidence as comparatively underdeveloped and inconsistent, citing earlier work suggesting cognitive impairment, behavioral dysregulation, and lower intelligence scores, referenced as 8 and 9, without resolving whether those associations held after confounder adjustment.
The authors also note that existing reviews in this space combined maternal, perinatal, and child outcomes, included multiple co-exposures, or focused on narrow behavioral domains, limiting their usefulness for a comprehensive neurodevelopmental synthesis, as described with reference 18. This positions the current review as a scope expansion rather than a direct replication or refutation of specific prior effect sizes, since the prior reviews it critiques were not designed to produce comparable domain-specific estimates.
Practical Considerations
For this evidence to change clinical practice beyond cautious counseling, several things would need to be true that are not yet established. First, studies would need standardized, quantitative exposure measurement (trimester-specific dose, not binary use) paired with standardized behavioral and executive function instruments across cohorts, enabling actual pooled effect estimates rather than vote-counting of positive versus null studies.
Second, confounding by maternal psychiatric illness and poly-substance use would need to be addressed through designs capable of separating shared genetic or environmental liability from a direct pharmacologic effect, such as sibling-comparison or discordant-twin designs, neither of which appears represented in this review's included studies. Until those design features exist, clinical guidance can reasonably emphasize avoidance of cannabis in pregnancy as a precaution, but cannot specify a dose threshold, a critical exposure window, or an expected magnitude of behavioral risk for a given level of use.
Future Directions (Expected)
A disciplined next step is prospective cohorts using biological confirmation of exposure (urine or meconium) combined with trimester-specific timing data, since the review notes that heavier or persistent use after maternal awareness of pregnancy showed stronger associations in some cohorts, a signal worth testing formally rather than narratively.
Given that behavioral and executive function outcomes showed the most consistent signal, future work should prioritize mechanistic and imaging studies linking prenatal cannabinoid exposure to structural or functional changes in prefrontal and fronto-limbic circuits, building on the neurobiological plausibility the authors cite, and should pair these with confounder-robust designs such as sibling or co-twin comparisons to isolate a direct pharmacologic contribution from shared family-level risk factors.
Misreadings & Bad-Faith Takes
A common distortion would be to summarize this review as proving prenatal cannabis exposure causes ADHD or behavioral disorders. The review reports associations, predominantly from observational cohorts with variable confounder adjustment, not a demonstrated causal pathway, and it explicitly states that current evidence does not consistently support global neurodevelopmental impairment.
The opposite distortion, that cannabis in pregnancy is shown to be neurodevelopmentally safe because cognition and language were largely unaffected, is equally incorrect. The review found that 69 to 73% of studies examining behavior, attention, and executive function reported negative associations, a pattern too consistent across independent cohorts to dismiss as noise. The accurate reading sits between both extremes: domain-specific signals warranting caution and further research, not proof of harm and not evidence of safety.
Join the Conversation
If you are a clinician who counsels pregnant patients about cannabis use, how do you currently frame the distinction between global cognitive risk and more specific behavioural or attention-related risk, and does a review like this change how you would word that conversation?
10.1007/s00787-026-03126-z
Frequently Asked Questions
Does this study prove that cannabis use during pregnancy harms a child’s intelligence?
No. Most studies (84%) found no significant association between prenatal cannabis exposure and global cognitive outcomes after adjusting for confounders. A minority of studies, generally involving heavier exposure, reported modest reductions in specific cognitive measures like IQ or verbal reasoning, but this was not the consistent pattern across the literature.
What outcomes showed the most consistent association with prenatal cannabis exposure?
Behavioural dysregulation, attention-related difficulties, and executive functioning showed the most consistent negative associations, found in 73%, 69%, and 70% of the relevant studies respectively.
Can this review tell us that cannabis caused these behavioural differences?
No. All 72 included studies were observational, meaning they can show association but cannot establish causation. Factors like maternal mental health, socioeconomic circumstances, and co-occurring substance use could partly or fully explain some of the associations found.
Does the amount or timing of cannabis use during pregnancy matter?
Several cohorts suggested stronger associations with heavier or more persistent cannabis use, especially use that continued after the mother became aware of the pregnancy, consistent with a possible dose-response and timing pattern. The review cannot quantify this relationship precisely because studies measured exposure differently.
Why didn’t the researchers calculate one overall number for the risk?
The authors originally planned a meta-analysis but abandoned it because the included studies differed too much in how they defined and measured exposure, assessed outcomes, and adjusted for confounding factors. Combining such different studies into a single number would have been statistically misleading.
Were the individual studies in this review well conducted?
Most were: 70 of the 72 included studies scored 7 to 9 out of 9 on the Newcastle-Ottawa Scale, indicating low risk of bias by that measure. Only 2 studies were rated moderate quality.
Should pregnant women be reassured that cannabis is safe for the baby’s brain development based on this review?
No. The absence of a consistent global cognitive signal is not the same as proof of safety. The review found a recurring association with behavioural, attention, and executive function difficulties, and short-term obstetric risks of cannabis use during pregnancy, such as low birth weight and preterm delivery, are documented elsewhere and not evaluated by this review.
Does this review apply to all forms of cannabis and CBD products?
The review included studies assessing exposure to cannabis, THC, CBD, or clinically diagnosed cannabis use disorder, but exposure measurement varied across studies and spanned products and potency levels from 1985 to 2025, which may not reflect currently available cannabis products.
What should a clinician say to a patient who used cannabis earlier in pregnancy and has since stopped?
The review’s suggestion of dose-response and timing effects, with stronger associations tied to heavier and more persistent use, supports encouraging cessation as early as possible rather than treating any past use as a fixed, unmodifiable risk. The review does not provide data to quantify how much risk reduction accompanies early cessation.
Get the next evidence review in your inbox
Plain-language summaries of the newest cannabis research, written by a physician. Free to read, and you can unsubscribe at any time.
Join the growing cannabis community. Free. One-click subscribe. Delivered by Substack.
