| Audience | Patients who use cannabis regularly or are considering it, primary care and cannabis-medicine clinicians, and readers following population-level shifts in substance use |
| Primary Topic | the documented shift from cigarette-only to cannabis-only substance use among US adults, and what fresh national survey data confirms about that trend continuing into 2025 |
| Source | Read the full source |
Daily Cannabis Use Just Overtook Cigarettes and Alcohol: What the Peer-Reviewed Trend Data Actually Shows
A peer-reviewed analysis of NSDUH data from 42,000 to 47,000 US adults per survey period found that cannabis-only use rose from 3.9 percent in 2015 to 10.6 percent in 2021-2023, overtaking cigarette-only use, which fell from 15.0 percent to 8.8 percent over the same span. Freshly released 2025 federal survey figures reported this week show the crossover has continued: roughly 21.4 million Americans now use cannabis daily or near-daily, more than report daily cigarette or daily alcohol use. The finding is population-level substitution evidence, not a treatment claim, and the study’s own authors flag real public-health caution alongside it.
| Study Type | Cross-sectional trend analysis of national survey data (NSDUH), 2015-2023 |
| Population | US adults, unweighted sample of approximately 42,163 to 46,906 per survey period (27,001 in 2020) |
| Cannabis-Only Use, 2015 | 3.9 percent |
| Cannabis-Only Use, 2021-2023 | 7.9 percent rising to 10.6 percent |
| Cigarette-Only Use, 2015 | 15.0 percent |
| Cigarette-Only Use, 2021-2023 | 10.8 percent falling to 8.8 percent |
| Co-Use Pattern | Relatively stable across all survey periods |
| Socioeconomic Pattern | Cigarette-only use more common among lower-income, lower-education, and uninsured adults; cannabis-only use more common among college-educated, higher-income, and privately insured adults |
| 2025 Federal Follow-Up Data | Approximately 21.4 million Americans report daily or near-daily cannabis use, versus 19.9 million for cigarettes and 17.2 million for alcohol, per newly released NSDUH figures reported in national media this week |
| Journal | Addictive Behaviors |
| Published | October 2025 (online) : Read PDF @ CED clinic |
| PMID | 41109071 |
| DOI | 10.1016/j.addbeh.2025.108521 |
Using nine years of NSDUH data, the authors tracked three groups: adults who used only cigarettes in the past 30 days, adults who used only cannabis, and adults who used both. Cannabis-only use rose in almost every sociodemographic subgroup from 2015 through 2023, while cigarette-only use fell in nearly all of them over the same period.
By 2021-2023, cannabis-only use (10.6 percent) had overtaken cigarette-only use (8.8 percent) for the first time in this dataset. Co-use of both substances stayed comparatively flat across all nine years, which the authors interpret as evidence against simple additive risk and more consistent with a substitution pattern in a meaningful share of users.
This week, national outlets reported freshly released federal survey figures showing that an estimated 21.4 million Americans now use cannabis on a daily or near-daily basis, surpassing both daily cigarette use (19.9 million) and daily alcohol use (17.2 million) for the first time on record.
These are different survey years and reporting frames than the peer-reviewed 2015-2023 analysis, so the exact percentages are not directly interchangeable. But the direction is the same: cannabis has moved from a minority daily-use substance to the most commonly used one among the three, continuing the trajectory the Addictive Behaviors paper first documented in peer-reviewed form.
It would be a mistake to read this paper as a cannabis-industry win. The authors explicitly call the rising cannabis-only trend ‘a worrisome trend among adults’ and write that ‘without timely policy response, cannabis may become the next public health crisis.’ They call for targeted prevention campaigns alongside acknowledging cannabis’s emerging therapeutic role in pain management and opioid tapering.
That tension is the honest read of this data: substitution away from cigarettes is plausibly a harm-reduction signal for some adults, while the sheer scale of the increase in cannabis-only use, including daily and near-daily use, raises separate questions about cannabis use disorder, tolerance, and dependence that this study was not designed to answer.
This is cross-sectional survey data, not a randomized or prospective cohort design. It cannot establish that any individual person switched from cigarettes to cannabis, why they did, or whether their overall health risk went up or down as a result. It also cannot separate recreational, medical, and mixed-motivation cannabis use, or distinguish smoked cannabis from vaporized, edible, or tincture forms, each of which carries a different risk profile.
The study also cannot tell us anything about cannabis use disorder rates within the cannabis-only group, which is a separate and clinically important question that this dataset was not built to answer.
This finding sits inside a larger, ongoing conversation about substance substitution that CED Clinic has covered before in the context of cannabis and opioid use, where a comparable population-level substitution pattern has been documented alongside real caveats about individual variability and risk.
The pattern here is broader than opioids, though. It suggests cannabis is increasingly functioning as a general substitution substance across multiple categories, which raises the stakes for clinicians to ask specific, nonjudgmental questions about what a patient is replacing and why, rather than treating cannabis use as a single undifferentiated behavior.
What strikes me about this data is not that cannabis use is rising. That has been obvious in clinic for years. It is that cigarette-only use is falling in tandem, in a pattern too consistent across nine years and dozens of subgroups to dismiss as coincidence. Something structural is happening in how adults are choosing to use substances, and cannabis is absorbing a real share of that shift.
My practical response to this data is not celebration and not alarm. It is a screening change. When a patient tells me they use cannabis, I now ask directly what they used to use instead, whether that is cigarettes, alcohol, or something else, and whether the switch has actually reduced symptoms or risk in their specific case. The population trend is real. What it means for the person in front of me still has to be worked out one patient at a time, with honest attention to use frequency, product type, and whether tolerance or dependence patterns are emerging.
How to Read a Population Substitution Trend Without Overclaiming Individual Benefit
Population-level substitution data is easy to misread in either direction: as proof that cannabis is a safe cigarette replacement, or as proof that rising cannabis use is purely alarming.
A more careful read separates what the aggregate trend shows from what it can say about any individual patient’s risk.
A Four-Step Reading Frame
Start With Study Design
This is repeated cross-sectional survey data, useful for tracking population trends over time but not designed to trace individual substitution decisions or outcomes.
Separate the Population Trend From the Individual Case
A rising cannabis-only percentage at the population level does not tell a clinician whether a specific patient’s switch from cigarettes to cannabis improved or worsened their overall health risk.
Hold Two Facts at Once
The authors themselves flag both a plausible harm-reduction signal and a public-health concern about the scale of the increase in cannabis-only use, including daily use. Neither framing cancels out the other.
Translate the Trend Into a Screening Question
The clinically useful move is asking patients what they are substituting cannabis for and how frequently they use it, not assuming the population trend answers that question for them.
The Same Study Can Mean Different Things Depending on the Question Being Asked
Scientific papers rarely answer a single question. Patients, clinicians, researchers, policymakers, and critics often read the same data differently. The perspectives below explore how this study looks through several evidence-based lenses.
The Trend Is Real, But It Is Not a Personal Verdict
If you have already moved from cigarettes to cannabis, this data confirms you are part of a large, measurable national pattern, not an outlier. That can be reassuring in the sense that the shift is not unusual.
It does not mean the switch was automatically the healthiest choice available to you. Frequency, product type, and reason for use still matter more than which substance you technically report using.
This Changes the Intake Question
For primary care, the practical shift is in the intake conversation. Asking only ‘do you smoke cigarettes’ increasingly misses a large and growing share of substance use that has moved to cannabis instead.
A better standard question set asks about cannabis frequency, product type, and what substance, if any, cannabis has replaced, since that context changes both the risk profile and the counseling approach.
Substitution Is Not the Same as Cessation
From an addiction-medicine standpoint, substituting one substance for another can reduce certain harms, such as combustion-related respiratory exposure, while introducing others, including the risk of cannabis use disorder in adults who use daily or near-daily.
The rising daily-use figures in the 2025 data are the detail addiction specialists should watch most closely, since daily and near-daily use is the pattern most associated with dependence risk.
The Authors’ Own Caution Deserves a Careful Read
The study’s authors describe the rising cannabis-only trend as a ‘worrisome’ pattern and call for targeted prevention campaigns, even while acknowledging cannabis’s therapeutic uses in specific clinical contexts.
That dual framing is worth preserving in public messaging: a population shift away from cigarettes is not automatically a public-health win if it is accompanied by rising daily use and unaddressed dependence risk.
Legalization Correlates With the Trend, But Correlation Is Not the Whole Story
The authors note that rising cannabis-only use parallels expanding state-level legalization and increasing accessibility, while declining cigarette-only use aligns with decades of tobacco-control policy.
This survey cannot isolate how much of the shift is caused by policy change versus changing social norms, cost, product availability, or shifting harm perceptions, all of which likely contribute together.
Self-Reported, Cross-Sectional Data Has Real Limits
Self-reported substance use is vulnerable to social-desirability bias in both directions: declining cigarette stigma disclosure over time and increasing willingness to admit cannabis use as it becomes more normalized and legal.
That means part of the observed crossover could reflect reporting shifts rather than true behavior change, even though the consistency of the trend across nine years and many subgroups argues against that being the whole explanation.
The Socioeconomic Split Deserves Attention
Cigarette-only use remains more concentrated among adults with lower income, less education, or no insurance, while cannabis-only use is more common among college-educated, higher-income, privately insured adults.
That divide matters clinically: it means the substitution trend this study describes is not evenly distributed, and prevention or counseling strategies built around one group’s use pattern may not translate directly to the other.
What Would Sharpen This Picture
The clearest next step is prospective, individual-level research that follows the same adults over time to see whether cannabis initiation actually precedes cigarette cessation, and what happens to cannabis use disorder rates in that group afterward.
Better product-type breakdowns, specifically separating smoked cannabis from vaporized, edible, and tincture use, would also sharpen the risk picture well beyond what a single ‘cannabis-only’ category can show.
Join the Conversation
Have a question about how this applies to your situation? Ask Dr. Caplan
Want to discuss this topic with other patients and caregivers? Join the forum discussion
Frequently Asked Questions
Does this study prove that cannabis is a safer substitute for cigarettes?
No. It shows a population-level trend where cannabis-only use has overtaken cigarette-only use. It does not establish that switching improves any individual patient’s overall health risk.
What exactly did the study measure?
Researchers analyzed nine years of National Survey on Drug Use and Health (NSDUH) data, tracking the percentage of US adults who reported cigarette-only use, cannabis-only use, or co-use of both in the past 30 days.
When did cannabis-only use overtake cigarette-only use in this data?
By 2021-2023, cannabis-only use reached 10.6 percent of adults surveyed, compared to 8.8 percent for cigarette-only use, the first crossover documented in this nine-year dataset.
How does this relate to the newly reported 2025 daily-use figures?
Freshly released federal survey data reported this week shows about 21.4 million Americans use cannabis daily or near-daily, more than daily cigarette users (19.9 million) or daily alcohol users (17.2 million). This is a different data point from a more recent survey year, but it points in the same direction as the peer-reviewed 2015-2023 trend.
Did co-use of both cigarettes and cannabis also rise?
No. The study found co-use stayed relatively stable across all nine years, which the authors interpret as more consistent with substitution than with simple additive use of both substances.
Is this trend the same across all income and education groups?
No. Cigarette-only use remained more common among adults with lower income, less education, or no insurance, while cannabis-only use was more common among college-educated, higher-income, privately insured adults.
Do the study authors view this trend positively?
Not simply. They describe the rise in cannabis-only use as a ‘worrisome trend’ and call for targeted prevention campaigns, while also acknowledging cannabis’s emerging therapeutic uses in specific clinical contexts such as pain management.
Does the study break down cannabis use by product type, such as smoked versus edible?
No. The study measured cannabis use broadly, including smoking, vaping, edibles, and tinctures, without separating outcomes by product type, which limits how precisely its risk implications can be applied.
What does this mean for how clinicians should screen patients?
It suggests that standard tobacco-focused screening questions may miss a growing share of substance use. Asking specifically about cannabis frequency, product type, and what it may be replacing gives a more complete picture.
Should someone who smokes cigarettes switch to cannabis based on this study?
No. This is population-level survey data, not a clinical recommendation. Any decision about substituting cannabis for cigarettes should be made individually with a clinician, accounting for frequency, product type, and personal health history.
