Published summary: 2026-03-07
Where the evidence converges
Mega-analysis of 400 IV THC infusions found psychosis-like symptoms in 45% of healthy volunteers, with effects increasing with dose and decreasing with regular cannabis use history
Strong evidence · 23 studies
Intravenous THC administration does not replicate real-world cannabis use (smoking, edibles). Participants were healthy volunteers, not people at high risk for psychosis. The controlled setting may mo
Meta-analysis of 28 studies found blood and saliva THC levels are poor predictors of driving impairment, with negligible to weak correlations that disappear entirely in regular cannabis users
Strong evidence · 23 studies
Analyses in regular users were less robust due to fewer studies. Most studies used simulated rather than real-world driving.
Meta-analysis of 18,000+ paired samples found oral fluid THC tests detect recent cannabis use well but are unreliable for predicting whether blood THC exceeds legal driving limits
Strong evidence · 23 studies
Paired samples came from various study conditions (controlled dosing, roadside testing, clinical). The relationship between blood THC and actual driving impairment is itself imperfect. Different oral
Where questions remain
Data from 120 subjects showed cannabis tolerance develops and fades quickly, while withdrawal symptoms including insomnia, irritability, and tremor appeared after as few as 7 days of use
Moderate evidence · 148 studies
Data from controlled research settings may not reflect real-world usage patterns. The 120 subjects likely represented a relatively homogeneous population. Withdrawal was described as mild and transien
THCV-COOH, a urinary metabolite unique to natural marijuana, reliably distinguished marijuana use from prescription Marinol use in a controlled crossover study
Moderate evidence · 148 studies
Only four subjects. THCV content varies among marijuana varieties, so very low-THCV marijuana might not produce detectable THCV-COOH. The study predates the availability of many modern cannabis produc
Review finding cannabis increases heart rate with rapid tolerance development, poses minimal risk for young healthy users but significant risks for people with existing cardiovascular disease
Moderate evidence · 148 studies
As a narrative review, the evidence was synthesized without systematic methodology. Most data came from studies in young, healthy subjects and may not apply to older or medically compromised populatio
Review of animal studies found cannabinoid tolerance develops through receptor down-regulation, withdrawal symptoms are generally milder than with other drugs, and tolerant animals showed no increased vulnerability to morphine
Moderate evidence · 148 studies
All evidence comes from animal models, which may not fully translate to human experience. The review notes that results have been controversial in various aspects. The pharmacokinetic differences betw
Research gaps
- Long-term prospective studies tracking outcomes over 5+ years are largely absent from the literature.
- Research on diverse populations (different ages, ethnicities, and medical backgrounds) remains limited.
Key studies
Large meta-analysis finds regular cannabis use raises both pro-inflammatory and anti-inflammatory markers, not just one or the other
Cannabis is widely described as "anti-inflammatory," but this meta-analysis shows regular use modulates the immune system in both directions simultaneously, with implications for people using cannabis to manage inflammatory conditions.
Brief Interventions in Medical Settings Did Not Reduce Cannabis Use
Brief interventions work well for alcohol, so they have been widely adapted for cannabis. This meta-analysis shows they are not effective for cannabis in most medical settings, redirecting efforts toward settings (like EDs) where they may work.
Cannabis Use in Pregnancy Linked to Preterm Birth, Low Birth Weight, and Small Babies Even After Accounting for Tobacco
This is one of the largest and most rigorous analyses of cannabis and pregnancy outcomes. The upgrade from low to moderate certainty evidence, combined with the enormous sample size, substantially strengthens the case that prenatal cannabis use carries real risks to newborns.
Reducing Cannabis Use by 50-75% Was Enough to See Real Improvements
This provides data-driven evidence that significant reduction without complete abstinence produces meaningful clinical improvement.
Cannabis and Heart Attack/Stroke Risk: A 183-Million-Patient Meta-Analysis Finds Stroke Risk but Not Heart Attack Risk
This is the largest meta-analysis of cannabis and cardiovascular risk ever conducted. With over 183 million patients, it has statistical power that individual studies lack. The finding that stroke risk is elevated while heart attack risk is not suggests cannabis may affect the cardiovascular system
Blood and Saliva THC Levels Are Poor Indicators of Driving Impairment
Many jurisdictions use blood or oral fluid THC levels to identify impaired drivers, similar to blood alcohol. This meta-analysis shows THC biomarkers are fundamentally different from blood alcohol as impairment indicators.
How the research developed
Pre-2000
8 studies published. Includes 1 RCTs.
2000–2009
25 studies published. Includes 5 RCTs, 3 strong-evidence studies.
2010–2014
39 studies published. Includes 12 RCTs, 6 strong-evidence studies.
2015–2019
181 studies published. Includes 2 meta-analyses, 18 RCTs, 15 strong-evidence studies.
2020–present
591 studies published. Includes 14 meta-analyses, 38 RCTs, 79 strong-evidence studies.
How this summary was assembled
This consensus synthesizes 844 peer-reviewed studies: 16 meta-analyses (2%), 74 randomized controlled trials (9%), 265 observational studies (31%), 167 reviews (20%), 78 case studies (9%), 244 other study types (29%). Studies span from the earliest available research through 2025. Evidence strength ratings reflect study design, sample size, and replication across multiple research groups.
These counts describe the source summary at publication. They may differ from the current library.