Smoking vs. Vaporizing Cannabis: Airways, CO, and Lung Function

The most directly studied difference concerns exposure: one particular dry-herb vaporizer produced less exhaled carbon monoxide than smoked cannabis in small controlled studies. Considerably more robust data are lacking to show fewer long-term pulmonary diseases from switching. Cough, lung function, imaging findings, and cancer are distinct endpoints and should be assessed separately. Abrams et al. (2007) Newmeyer et al. (2017) Ghasemiesfe et al. (2018)

What question do the various measurements actually answer?

Exhaled CO is a marker for a specific exposure, not a diagnosis of pulmonary disease. A questionnaire can capture cough, sputum, or wheezing. Spirometry measures airway volumes and flows. Imaging can reveal structural or functional abnormalities. None of these levels automatically substitutes for the others.

A reduction in symptoms may be clinically meaningful without yet constituting demonstrated evidence of a reduced COPD risk. Equally, an unremarkable single measurement does not render the overall exposure without concern. The question “Is vaporizing better for the lungs?” can therefore only be answered scientifically by specifying the concrete endpoint in question. Van et al. (2010) Ghasemiesfe et al. (2018)

What do the direct CO comparisons show?

Abrams studied 18 healthy cannabis users in a controlled comparison of cannabis cigarettes and the Volcano. The lower CO exposure with vaporizing was accompanied by a similar THC exposure. The second controlled study, involving 20 participants, compared smoking, vaporizing, and oral administration; CO elevation with smoking was also higher in that study. Abrams et al. (2007) Newmeyer et al. (2017)

The strength of these studies lies in the direct comparison under defined conditions. The limitations are the small sample sizes, specific devices used, and short observation periods. CO is moreover only a single marker. Its reduction cannot be used to derive a percentage figure for all inhaled substances in their entirety, let alone for lifetime disease risks.

How strong is the association between cannabis smoke and cough?

The systematic review by Ghasemiesfe covered 22 studies. The pooled analysis of two prospective studies found a relative risk for cough of 2.04 (95% CI 1.02–4.06) and for sputum of 3.84 (1.62–9.07). The authors rated the evidence on respiratory symptoms as low quality. These figures therefore do not derive from 22 comparable randomised trials. Ghasemiesfe et al. (2018)

A relative risk of 2.04 describes the ratio of two group rates. It does not mean that every person who uses cannabis will develop a cough, and without a baseline rate it provides no absolute probability. The wide confidence intervals indicate additional uncertainty. Non-uniform exposure measurement, concurrent tobacco use, and limited long-term exposure data complicate interpretation.

For permanent airway obstruction and changes in lung function, the evidence base in this review was insufficient. ‘Insufficiently established’ is not confirmation that a risk has been ruled out. At the same time, the uncertainty should not be presented as proof of a specific disease. Ghasemiesfe et al. (2018)

What should we make of the small switching study?

Van Dam and Earleywine examined 20 frequent cannabis users with at least two respiratory symptoms before and after one month of vaporizer use. Twelve individuals without respiratory illness during the trial formed the primary analysis. Symptoms and FVC improved; the change in FEV1 was not statistically significant. Van et al. (2010)

FVC refers to the total volume exhaled during a forced expiration; FEV1 refers to the proportion exhaled in the first second. This distinction prevents a single positive finding from being generalised as ‘improved lung function’. The study had no randomised control group. Expectations, changes in other habits, and the selection of the analysed subgroup may have influenced the results.

The appropriate verdict is a preliminary signal that justifies controlled research. A demonstrated treatment of respiratory conditions by a vaporizer cannot be inferred from this. Even a before-and-after comparison for a single individual would have only limited causal validity without a control.

What do CANUCK and PATH add?

CANUCK examined 139 cannabis smokers and 57 never-smokers in a cross-sectional design. Higher joint-year exposure was associated with less favourable respiratory findings, certain imaging abnormalities, and differences in airway epithelium. The combined investigation across multiple levels is a strength. However, 84 per cent of the cannabis group currently or previously smoked or vaped tobacco. This high degree of co-use limits attribution. Leung et al. (2026)

PATH followed 5.211 adults aged 18 to 24 without previously reported respiratory disease across two survey waves. For exclusive smoking, the adjusted odds ratio for new functionally relevant respiratory symptoms was 3.04 (95% CI 1.64–5.64); for smoking plus vaping it was 4.36 (2.12–8.96). For exclusive vaping it was 0.93 (0.20–4.43). After excluding tobacco users, only the association for dual use remained statistically significant. Sun et al. (2026)

The last point in particular is an immediate part of the finding. The small vaping subgroup and the very wide confidence interval preclude any reliable reassurance. Furthermore, dry-herb products were not adequately separated from other cannabis vapes. The figures therefore do not provide a precise comparison of ‘modern dry-herb vaporizers versus joints’.

Why are the new studies not a contradiction of the laboratory evidence?

Laboratory studies can compare selected substances under controlled conditions. Population studies capture real-world, variable, and often mixed patterns of use. A lower CO measurement and a still-observed respiratory risk are therefore not mutually exclusive. Different endpoints and products account for some of the apparent contradictions. Abrams et al. (2007) Leung et al. (2026) Sun et al. (2026)

For a meaningful long-term comparison, prior smoking history, tobacco use, product type, frequency of use, and transitions between modes of consumption would need to be documented repeatedly. A single self-report cannot fully capture an exposure history spanning several years. Similarly, statistical adjustment cannot entirely eliminate unobserved differences.

What can be said about COPD and cancer?

The respiratory review was unable to answer the questions on obstruction and lung function conclusively. A separate systematic evaluation of cancer found overall insufficient evidence for several specific cancer types, including lung cancer. Low long-term exposure and tobacco confounding were among the limitations. This uncertainty must be distinguished from the established fact that cannabis smoke contains potentially harmful combustion products. Ghasemiesfe et al. (2018) Ghasemiesfe et al. (2019) Moir et al. (2008)

How should the comparison be read in practice?

The most robust summary is: fewer selected combustion markers under the dry-herb vaporizer conditions studied, interesting but preliminary symptom data, and open questions regarding long-term clinical endpoints. A water filter or pleasantly cool vapour does not automatically extend this evidence base. Abrams et al. (2007) Van et al. (2010)

The health overview covers remaining THC risks. Vaporization vs. combustion explains aerosol chemistry and measurement methods. The joint comparison adds costs and practical differences without using the respiratory figures as a purchasing promise.

Studies in detail: methods, results and limitations

The following boxes make the key sources and their limits of transferability transparent. A device study or pharmacokinetics study does not automatically constitute evidence of long-term health safety. For technical articles, the studies are often relevant only to the explicitly stated distinction between health claims and yield claims.

Abrams DI et al. (2007)

Study: Randomised controlled pilot study, repeated measures
Sample: 18 healthy inpatient cannabis users
Comparison and measurement: Comparison of a standardised smoked cannabis cigarette versus the Volcano vaporizer; plasma THC, exhaled CO, physiological and neuropsychological effects.
Randomisation and blinding: Yes. Cannabis potency (1.7%, 3.4%, 6.8% THC) and delivery system were randomly assigned across six study days.
Result: THC exposure was similar between smoking and vaporizing; exhaled CO was markedly lower after vaporization; no serious adverse events.
Strengths: Direct controlled head-to-head comparison; objective biomarkers; randomised.
Limitations and potential bias: Very small sample; pilot in nature; older and lower THC potencies compared with today’s products; no long-term endpoints.

Abrams DI, Vizoso HP, Shade SB, Jay C, Kelly ME, Benowitz NL. 2007. Vaporization as a smokeless cannabis delivery system: a pilot study. Clinical pharmacology and therapeutics. DOI: 10.1038/sj.clpt.6100200 · PMID 17429350

Newmeyer MN et al. (2017)

Study: Controlled within-subject study with placebo condition
Sample: 20 participants: 11 frequent and 9 occasional cannabis users
Comparison and measurement: Placebo versus active cannabis (6.9% THC, approximately 54 mg) smoked, vaporized, or administered orally; subjective effects, heart rate, and exhaled CO.
Randomisation and blinding: Controlled study protocol; the same NIDA crossover cohort was described as randomised/double-blind in associated publications.
Result: Smoking and vaporizing increased subjective effects; heart rate rose following all active routes of administration; smoking increased CO significantly more than vaporization.
Strengths: Direct three-way comparison; objective CO endpoint; separate analysis of frequent and occasional users.
Limitations and potential bias: Very small sample; high dose; differences in tolerance; multiple publications use the same cohort and must not be counted as independent studies.

Newmeyer MN, Swortwood MJ, Abulseoud OA, Huestis MA. 2017. Subjective and physiological effects, and expired carbon monoxide concentrations in frequent and occasional cannabis smokers following smoked, vaporized, and oral cannabis administration. Drug and alcohol dependence. DOI: 10.1016/j.drugalcdep.2017.02.003 · PMID 28407543

Ghasemiesfe M et al. (2018)

Study: Systematic review and meta-analysis
Sample: 22 studies (10 prospective cohorts, 12 cross-sectional studies).
Comparison and measurement: PubMed, Embase, PsycINFO, MEDLINE, Cochrane 1973–30.04.2018; PROSPERO; 4 reviewers for extraction/RoB, 3 for strength of evidence; pooling only of similar studies with sufficiently low risk of bias.
Randomisation and blinding: Predominantly non-randomised observational studies.
Result: 22 studies in total. Pooled prospective symptom data from two studies: cough RR 2.04 (95% CI 1.02–4.06), sputum RR 3.84 (1.62–9.07). The authors rated the strength of evidence for respiratory symptoms as low; for obstruction and lung function as insufficient.
Strengths: Very rigorous methodology; PROSPERO registration; multiple reviewers; risk-of-bias and evidence-quality assessment; nuanced conclusions.
Limitations and potential bias: Primary studies mostly observational; exposure measurement and tobacco confounding; limited data at very high lifetime exposure.

Ghasemiesfe M, Ravi D, Vali M, Korenstein D, Arjomandi M, Frank J, Austin PC, Keyhani S. 2018. Marijuana Use, Respiratory Symptoms, and Pulmonary Function: A Systematic Review and Meta-analysis. Annals of internal medicine. DOI: 10.7326/M18-0522 · PMID 29971337 · PMC6231497

Van Dam NT, Earleywine M (2010)

Study: Single-arm pre/post intervention study
Sample: 20 frequent cannabis users with at least two respiratory symptoms; 12 without intercurrent respiratory illness formed the primary analysis.
Comparison and measurement: One month switching to a vaporizer; nine subjective respiratory symptoms plus spirometry (FEV1, FVC) before and after the intervention.
Randomisation and blinding: No.
Result: In the 12 participants without respiratory infection, symptoms improved markedly; FVC improved significantly; FEV1 improved numerically, narrowly falling short of significance.
Strengths: Prospective intervention; objective spirometry in addition to symptom measures.
Limitations and potential bias: Very small; no control group; selection of the 12 participants may introduce bias; short duration; no blinding.

Van Dam NT, Earleywine M. 2010. Pulmonary function in cannabis users: Support for a clinical trial of the vaporizer. The International journal on drug policy. DOI: 10.1016/j.drugpo.2010.04.001 · PMID 20451365

Leung C et al. (2026)

Study: Cross-sectional clinical multimodal study
Sample: 139 cannabis smokers stratified by joint-years (≤5, >5–20, >20) and 57 never-smokers; 48% male, median age 27; 84% of the cannabis group with current or former cigarette smoking or vaping.
Comparison and measurement: SGRQ/CAT, spirometry, plethysmography, diffusion capacity measurement, CT, hyperpolarised 129Xe MRI, bronchoscopy/airway brushings, immune/mucin gene expression and cell culture.
Randomisation and blinding: No.
Result: Cannabis-smoking groups reported more respiratory symptoms; high joint-year exposure was associated with worse flow/ratio parameters, greater emphysema/ventilation abnormalities and an altered epithelial immune/MUC5AC signature.
Strengths: Highly advanced objective phenotyping; dose-stratified groups; imaging plus physiology plus molecular biology; adjustment including for tobacco/vaping.
Limitations and potential bias: Cross-sectional design, therefore no causal inference; very high proportion with concurrent tobacco/vape exposure; residual confounding.

Leung C, Gilchrist CL, Wang CJ, Liggins JA, Li X, Yang J, Cheung CY, Gerayeli FV, Singhera GK, Hsu WJ, Lidher LS, Moo K, Leyson E, Dhillon SS, Shaipanich T, Leipsic JA, Guenette JA, Rayment JH, Kirby M, Gershon AS, Sadatsafavi M, Tan WC, Parraga G, Carlsten C, Eddy RL, Sin DD, Leung JM. 2026. Clinical, physiological, imaging and molecular responses to cannabis smoking: the Canadian Users of Cannabis Smoke (CANUCK) study. The European respiratory journal. DOI: 10.1183/13993003.01659-2025 · PMID 41198398 · PMC12805821

Sun R et al. (2026)

Study: National longitudinal observational study (PATH)
Sample: 5.211 young adults (aged 18–24) with no prior respiratory or pulmonary disease, Wave 6→7.
Comparison and measurement: Baseline 2021, follow-up to 2023; multivariable logistic regression, adjusted for demographics, other substances, health factors, obesity and cannabis frequency.
Randomisation and blinding: No.
Result: Vaping-only was not significantly associated with new functionally relevant respiratory symptoms (AOR 0.93; very wide CI); smoking-only AOR 3.04; dual use AOR 4.36. After excluding tobacco users, only dual use remained significant.
Strengths: Large longitudinal, nationally representative sample; baseline group free of respiratory disease; confounder adjustment.
Limitations and potential bias: Only 2.1% vaping-only → low statistical power and very wide CI; product type (dry-herb vaporizer vs. oil/cartridge) insufficiently granular; self-report.

Sun R, Benowitz NL, Hammond D, Mendez D, Warner KE. 2026. Cannabis Vaping, Smoking, and Dual Use and the Onset of Respiratory Symptoms Among U.S. Young Adults. American journal of preventive medicine. DOI: 10.1016/j.amepre.2026.108421 · PMID 42142848

Ghasemiesfe M et al. (2019)

Study: Systematic review and meta-analysis
Sample: 25 studies: 19 case-control, 5 cohort, 1 cross-sectional; only 2 studies with low risk of bias.
Comparison and measurement: PubMed, Embase, PsycINFO, MEDLINE, Cochrane 1973–2019; independent screening and evidence assessment; meta-analysis only where studies were sufficiently comparable.
Randomisation and blinding: No.
Result: Low strength of evidence for an increased risk of testicular germ cell tumours with regular use; evidence for lung cancer and other cancers overall insufficient.
Strengths: Methodologically rigorous; risk of bias and strength of evidence reported transparently; prevents overstating the cancer argument.
Limitations and potential bias: Long-term exposure often insufficient; strong tobacco confounding; older products of differing potency; consumption method rarely specified precisely.

Ghasemiesfe M, Barrow B, Leonard S, Keyhani S, Korenstein D. 2019. Association Between Marijuana Use and Risk of Cancer: A Systematic Review and Meta-analysis. JAMA network open. DOI: 10.1001/jamanetworkopen.2019.16318 · PMID 31774524 · PMC6902836

Moir D et al. (2008)

Study: Analytical Machine-Smoking Study
Sample: No human subjects; standardised cannabis and tobacco cigarettes smoked under two machine-smoking conditions.
Comparison and measurement: Chemical characterisation of mainstream and sidestream smoke; known toxic and carcinogenic constituents.
Randomisation and blinding: Not applicable.
Result: Qualitatively similar classes of harmful constituents to tobacco smoke, but quantitative differences; notably substantially higher ammonia as well as higher HCN/NOx and certain aromatic amines under particular cannabis-smoke conditions.
Strengths: Direct standardised chemical comparison; objective analytical methodology.
Limitations and potential bias: Machine-generated smoke does not represent real-world exposure; values are per amount of material, not per typical day of use; must not be reinterpreted as ‘cannabis is more harmful than tobacco’.

Moir D, Rickert WS, Levasseur G, Larose Y, Maertens R, White P, Desjardins S. 2008. A comparison of mainstream and sidestream marijuana and tobacco cigarette smoke produced under two machine smoking conditions. Chemical research in toxicology. DOI: 10.1021/tx700275p · PMID 18062674

Frequently Asked Questions

What question do the various measurements actually answer?

What do the direct CO comparisons show?

How strong is the association between cannabis smoke and cough?

References

Abrams DI, Vizoso HP, Shade SB, Jay C, Kelly ME, Benowitz NL. 2007. Vaporization as a smokeless cannabis delivery system: a pilot study. Clinical pharmacology and therapeutics. DOI: 10.1038/sj.clpt.6100200 · PMID 17429350
Newmeyer MN, Swortwood MJ, Abulseoud OA, Huestis MA. 2017. Subjective and physiological effects, and expired carbon monoxide concentrations in frequent and occasional cannabis smokers following smoked, vaporized, and oral cannabis administration. Drug and alcohol dependence. DOI: 10.1016/j.drugalcdep.2017.02.003 · PMID 28407543
Ghasemiesfe M, Ravi D, Vali M, Korenstein D, Arjomandi M, Frank J, Austin PC, Keyhani S. 2018. Marijuana Use, Respiratory Symptoms, and Pulmonary Function: A Systematic Review and Meta-analysis. Annals of internal medicine. DOI: 10.7326/M18-0522 · PMID 29971337 · PMC6231497
Van Dam NT, Earleywine M. 2010. Pulmonary function in cannabis users: Support for a clinical trial of the vaporizer. The International journal on drug policy. DOI: 10.1016/j.drugpo.2010.04.001 · PMID 20451365
Leung C, Gilchrist CL, Wang CJ, Liggins JA, Li X, Yang J, Cheung CY, Gerayeli FV, Singhera GK, Hsu WJ, Lidher LS, Moo K, Leyson E, Dhillon SS, Shaipanich T, Leipsic JA, Guenette JA, Rayment JH, Kirby M, Gershon AS, Sadatsafavi M, Tan WC, Parraga G, Carlsten C, Eddy RL, Sin DD, Leung JM. 2026. Clinical, physiological, imaging and molecular responses to cannabis smoking: the Canadian Users of Cannabis Smoke (CANUCK) study. The European respiratory journal. DOI: 10.1183/13993003.01659-2025 · PMID 41198398 · PMC12805821
Sun R, Benowitz NL, Hammond D, Mendez D, Warner KE. 2026. Cannabis Vaping, Smoking, and Dual Use and the Onset of Respiratory Symptoms Among U.S. Young Adults. American journal of preventive medicine. DOI: 10.1016/j.amepre.2026.108421 · PMID 42142848
Ghasemiesfe M, Barrow B, Leonard S, Keyhani S, Korenstein D. 2019. Association Between Marijuana Use and Risk of Cancer: A Systematic Review and Meta-analysis. JAMA network open. DOI: 10.1001/jamanetworkopen.2019.16318 · PMID 31774524 · PMC6902836
Moir D, Rickert WS, Levasseur G, Larose Y, Maertens R, White P, Desjardins S. 2008. A comparison of mainstream and sidestream marijuana and tobacco cigarette smoke produced under two machine smoking conditions. Chemical research in toxicology. DOI: 10.1021/tx700275p · PMID 18062674

Editorial status: 10 September 2026. Evidence-based overview based on a selected library of 34 publications; no independent systematic literature search and no medical review. Bibliographic details and available abstracts were verified; the key works by Spindle (2018) and Lanz (2016) were additionally checked in open full text. Technical selection and maintenance guidance should be distinguished from health evidence. New data may change this assessment.

More Comparisons

⚔️ Beginner vs. Expert Vaporizer: The Honest Comparison 2026 ⚔️ Butane or Battery: Power Sources in Dry-Herb Vaporizers Explained ⚔️ Butane vs Desktop Vaporizer ⚔️ Butane vs. Electric Vaporizer: Which Type Is Right for You?⚔️ E-Rig vs Dab Pen: Which Device Fits You?⚔️ Focus V Carta Sport vs Puffco Proxy: Powerhouse Meets Design Rig

View all comparisons →Still undecided? Try the Buying Guide