Marijuana Use Is Rising, But Is It a Cancer Risk?

The trends are clear: Americans are in the midst of a marijuana high. Over the past 30 years, daily or near-daily marijuana use soared 15-fold, surpassing daily alcohol use for the first time in 2022. That same year, marijuana use reached historic levels among Americans aged 19-50 — with 11% of 19- to 30-year-olds saying they used the drug every day.

A key reason for the surge is that more states are legalizing both medical and recreational marijuana use. Another driver, which is closely tied to legalization, is the changing public perceptions around marijuana: Many people just don’t see much harm in the habit, or at least view a daily marijuana joint as safer than smoking cigarettes. 

And they’re not necessarily wrong: Although it’s obvious marijuana use can have consequences — including intoxication, dependence, and respiratory symptoms such as chronic bronchitis — there is little, or not enough, evidence to definitively conclude that it’s a cancer risk.

But that also doesn’t mean marijuana is completely in the clear. 

“Insufficient evidence doesn’t mean the risk isn’t there,” said Nigar Nargis, PhD, senior scientific director of tobacco control research, American Cancer Society (ACS).

‘The Crux of the Problem’

Marijuana smoke does contain many of the same carcinogens found in tobacco smoke, so it seems logical that a cannabis habit could contribute to some cancers. Yet studies have largely failed to bear that logic out. 

In 2017, the National Academies of Sciences, Engineering, and Medicine (NASEM) published a comprehensive research review on cannabis smoking and cancer risk. It found modest evidence of an association with just one cancer: a subtype of testicular cancer. In the cases of lung and head and neck cancers, studies indicated no significant association between habitual cannabis use and risk for these cancers. When it came to other cannabis-cancer relationships, the evidence was mostly deemed insufficient or simply absent. 

However, the overarching conclusion from the NASEM review was that studies to date have been hampered by limitations, such as small sample sizes and survey-based measurements of cannabis use that lack details on frequency and duration of use. In addition, many marijuana users may also smoke cigarettes, making it difficult to untangle the effects of marijuana itself.

“That’s the crux of the problem,” Nargis said. “We have a huge knowledge gap where existing evidence doesn’t allow us to draw conclusions.”

That long-standing gap is becoming more concerning, she said, because legalization may now be sending a “signal” to the public that cannabis is safe.

This concern prompted Nargis and her colleagues to explore whether studies conducted since the 2017 NASEM report have lifted the marijuana-cancer risk haze at all. Their conclusion, published in February in The Lancet Public Health: not really.

“Unfortunately, the evidence base hasn’t improved much,” Nargis said. However, she added, some studies have hinted at links between cannabis use and certain cancers beyond testicular. Although these studies have their own limitations, Nargis stressed, they do point to directions for future research.

Head and Neck Cancers

While the NASEM report cited reassuring data on head and neck cancers, a study published last year in JAMA Otolaryngology-Head & Neck Surgery reached a different conclusion. The researchers tried to overcome some limitations of prior research — including small sample sizes and relatively light and self-reported marijuana use — by analyzing records from patients diagnosed with cannabis use disorder at 64 US healthcare organizations.

The study involved over 116,000 patients with cannabis use disorder, matched against a control group without that diagnosis. Head and neck cancers were rare in both groups, but the overall incidence over 20 years was about three times higher among patients with cannabis use disorder (0.28% vs 0.09%).

After propensity score matching — based on factors such as age and tobacco and alcohol use — patients with cannabis use disorder had a 2.5-8.5 times higher risk for head and neck cancers, especially laryngeal cancer: any type (risk ratio [RR], 3.49), laryngeal cancer (RR, 8.39), oropharyngeal cancer (RR, 4.90), salivary gland cancer (RR, 2.70), nasopharyngeal cancer (RR, 2.60), and oral cancer (RR, 2.51).

But although the study was large, “it’s not particularly strong evidence,” said Gideon Meyerowitz-Katz, MPH, PhD, an epidemiologist and senior research fellow at the University of Wollongong, Australia.

Meyerowitz-Katz pointed to some key limitations, including the focus on people with cannabis use disorder, who are not representative of users in general. The study also lacked information on factors that aren’t captured in patient records, such as occupation — which, Meyerowitz-Katz noted, is known to be associated with both head and neck cancer risk and cannabis use. 

Beyond that, the risk increases were generally small, even with extensive use of the drug. 

“If we assume the study results are causal,” Meyerowitz-Katz said, “they suggest that people who use cannabis enough to get a diagnosis of cannabis use disorder get head and neck cancer at a rate of around 3 per 1000 people, compared to 1 per 1000 people who don’t use cannabis.”

Cannabis and Childhood Cancers

As marijuana use has shot up among Americans generally, so too has prenatal use. One study found, for instance, that the rates almost doubled from about 3.4%-7% of pregnant women in the US between 2002 and 2017. Many women say they use it to manage morning sickness.

Given the growing prenatal use, however, there is a need to better understand the potential risks of fetal exposure to the drug, said Kyle M. Walsh, PhD, associate professor in neurosurgery and pediatrics, Duke University School of Medicine, Durham, North Carolina.

The fortunate rarity of childhood cancers makes it challenging to study whether maternal substance use is a pediatric cancer risk factor. It’s also hard to define a control group, Walsh said, because parents of children with cancer often have difficulty recollecting their exposures before and during pregnancy.

To get past these limitations, Walsh and his colleagues took a different approach. Instead of trying to track cannabis use and tie it to cancer risk, Walsh’s team focused on families of children with cancer to see whether prenatal substance use was associated with any particular cancer subtypes. Their study, published last year in Cancer Epidemiology, Biomarkers & Prevention, surveyed 3145 US families with a child diagnosed with cancer before age 18. The study, however, did not focus on just marijuana; it looked at illicit drug use during pregnancy more generally. Although the authors assumed that would mostly mean marijuana, it could include other illicit drugs, such as cocaine. 

Overall, 4% of mothers reported using illicit drugs during pregnancy. Prenatal use of illicit drugs was associated with an increased prevalence of two tumor types: intracranial embryonal tumors, including medulloblastoma and primitive neuroectodermal tumors (prevalence ratio [PR], 1.94), and retinoblastoma (PR, 3.11).

“Seeing those two subtypes emerge was quite interesting to us, because they’re both derived from a cell type in the developing fetal brain,” Walsh said. That, he added, “aligns in some ways” with research finding associations between prenatal cannabis use and increased frequencies of ADHD and autism spectrum disorders in children.

Interestingly, Walsh noted, prenatal cigarette smoking — which was also examined in the study — was not associated with any cancer subtype, suggesting that smoking might not explain the observed associations between prenatal drug use and central nervous system tumors. But, he stressed, it will take much more research to establish whether prenatal marijuana use, specifically, is associated with any childhood cancers, including studies in mice to examine whether cannabis exposure in utero affects neurodevelopment in ways that could promote cancer.

Testicular Cancer

Testicular cancer is the one cancer that has been linked to cannabis use with some consistency. But even those findings are shaky, according to Meyerowitz-Katz. 

A 2019 meta-analysis in JAMA Network Open concluded that long-term marijuana use (over more than a decade) was associated with a significantly higher risk for nonseminomatous testicular germ cell tumors (odds ratio, 1.85). But the authors called the strength of the evidence — from three small case-control studies — low. All three had minimal controls for confounding, according to Meyerowitz-Katz.

“Whether this association is due to cannabis or other factors is hard to know,” he said. “People who use cannabis regularly are, of course, very different from people who rarely or never use it.”

In their 2025 Lancet Public Health review, Nargis and her colleagues pointed to a more recent study, published in 2021 in BMC Pharmacology and Toxicology, that looked at the issue in broader strokes. The study found parallels between population marijuana use and testicular cancer rates, as well as higher rates of the cancer in US states where marijuana was legal vs those where it wasn’t.

However, Nargis said, observational studies such as this must be interpreted with caution because they lack data on individuals.

If regular cannabis use does have effects on testicular cancer risk, the mechanisms are speculative at best. Researchers have noted that the testes harbor cannabinoid receptors, and there is experimental evidence that binding those receptors may alter normal hormonal and testicular function. But the path from smoking weed to developing testicular cancer is far from mapped out.

Risk for Other Cancers?

The recent Lancet Public Health overview also highlights emerging evidence suggesting a relationship between cannabis use and risks for a range of other cancer types. 

A handful of observational studies, for instance, showed correlations between population-level cannabis use and risks for several cancers, such as breast, liver, thyroid, and prostate. The observational studies, mostly from a research team at the University of Western Australia, made headlines last year with a perspectives piece published in Addiction Biology, claiming there is “compelling” evidence that cannabis is “genotoxic” and raises cancer risk.

But, as Meyerowitz-Katz pointed out, the paper is only a perspective, not a study. And the human data it cites are from the same limited evidence base critiqued in the NASEM and ACS reports.

Meyerowitz-Katz does not discount the possibility that marijuana use contributes to some cancers. “I wouldn’t be surprised if we find that extensive cannabis use — particularly smoking — is related to cancer risk,” he said. But based on the existing evidence, he noted, the risk, if real, is “quite small.” 

Where to Go From Here?

What’s needed, Nargis said, are large-scale cohort studies like those that showed cigarette smoking is a cancer risk factor. For the ACS, she said, the next step is to analyze decades of data from its own Cancer Prevention Studies, which included participants with a history of cannabis use and cancer diagnoses verified using state registries.

Nargis also noted that nearly all studies to date have focused on marijuana smoking, and “almost nothing” is known about the long-term health risks of newer ways to use cannabis, including vaping and edibles.

“What’s concerning,” she said, “is that the regulatory environment is not keeping up with this new product development.”

With the evolving laws and attitudes around cannabis use, Nargis said, it’s the responsibility of the research community to find out “the truth” about its long-term health effects.

“People should be able to make their choices based on evidence,” she said.