Key Takeaways
- Researchers used data from over 230,000 participants to look into the association between being diagnosed with a cannabis-related disorder and head and neck cancer.
- The results showed that the cannabis disorder group had 3.49-fold higher risk of developing head and neck cancer, with a 0.2% absolute risk increase.
- The study carries weight because of the huge sample size, but it’s important to note that the study compares people diagnosed with a cannabis disorder vs. people not diagnosed, not users vs. non-users. There was likely cannabis use in both groups.
Cannabis might not be as dangerous as other drugs or even alcohol, but that doesn’t mean it’s risk-free. Most cannabis users consume by smoking, and just like smoked tobacco, this produces a comparable range of often-harmful chemicals that researchers have long-suspected to cause health issues in long-term consumers.
A massive new study published in the Journal of the American Medical Association (JAMA) investigates this in detail, looking specifically at head and neck cancers, finding a 3.5-fold increase in risk for people with cannabis use disorders.
But should ordinary cannabis consumers be worried? What exactly does the study say? We’ve gone through the evidence and spoken to experts to find out.
The Study: What They Did and Why
Given that head and neck cancers (including cancers of the oral cavity, pharynx, larynx and salivary glands) are largely associated with tobacco carcinogens, and these same carcinogens are present in cannabis smoke, the researchers set out to find out if they create a similar risk.
The authors note that previous studies found inconsistent results, and hoped that using data from a huge de-identified network of electronic medical records (TriNetX) could offer a reliable answer to the question.
The whole network consists of data from 64 healthcare organizations with more than 90 million individual patients. The researchers found a cohort of 116,076 participants who had a diagnosis of a cannabis use disorder and had an outpatient hospital clinic visit (but no head or neck cancer at this time) between April 2004 and April 2024. These participants were paired with participants in the same situation except with no diagnosis of cannabis use disorder, and after matching, the researchers ended up with 115,865 participants per group.
The main thing the researchers were looking for were new diagnoses of head or neck cancer following what they call the “index event,” which is when the participant first visited the outpatient clinic (with or without a cannabis use disorder diagnosis, depending on the group).
Factors like alcohol use and tobacco smoking are known to affect the risks of these cancers, so the participants were paired up so that these factors didn’t vary much between groups (in fact, the difference in both cases was just 0.1%).
The Results: The Cannabis Use Disorder Group Were More Likely to Develop Head or Neck Cancer
We spoke to Niels Kokot, MD, head and neck surgeon with Keck Medicine of USC, who worked on the study, and he described the main results:
An important takeaway is that our research reveals individuals who use cannabis, particularly those with a cannabis use disorder, are significantly more likely to develop head and neck cancers compared to non-users. This finding underscores the potential health risks of cannabis use, which may be similar to those associated with tobacco smoking, as both involve inhalation of smoke containing carcinogens.
In short, the results show increased cancer risks for the cannabis disorder group, with relative risks:
- Laryngeal: 8.39
- Nasopharyngeal: 2.6
- Oral: 2.51
- Oropharyngeal: 4.9
- Salivary gland: 2.7
- Hypopharyngeal: 1.7 (not statistically significant)
- Any head or neck cancer: 3.49
These relative risks (RR) tell you how many times more likely the cannabis disorder group was to get the cancer. So for any head or neck cancer, the cannabis group was 3.49 times more likely to develop the cancer than the non-cannabis group – i.e. there were roughly 7 cases of head and neck cancer in the cannabis group for every 2 cases in the non-cannabis group.
However, just focusing on these relative risks misses a big part of the picture, and we need to look at the methodology and limitations in more detail before we draw a strong conclusion.
We spoke to Dr. Benjamin Caplan, founder and Chief Medical Officer at CED Clinic, who emphasized this point when describing what the study found:
The study reports a significant association between cannabis-related disorders and certain head and neck cancers, such as laryngeal and oropharyngeal cancers. But, several critical issues with the study’s design and how the data were interpreted raise serious questions about the reliability of these findings.
Relative vs. Absolute Risk: What Does It Really Mean for Users?
The authors of this study made much more effort to translate these figures into “real world” numbers than most researchers do. In fact, along with the relative risks listed above, they also noted the absolute risk (i.e. the risk when not compared to some other group) for the cannabis and non-cannabis group.
Focusing just on the overall risk of head and neck cancer, the absolute risks are 0.285% for the cannabis group vs. 0.091% for the non-cannabis group, equivalent to 2,850 vs. 910 cases per million individuals.
The researchers calculated the percentage increase in absolute risk as 0.2% for any head and neck cancer at any time after the initial visit, dropping to 0.14% for cancers that developed at least a year later and to 0.04% for cancers developed after five years or more.
Additionally, the authors report the “number needed to harm,” which is a measure of how many people need to be exposed to the risk factor (i.e. cannabis) to lead to one case of head or neck cancer that wouldn’t otherwise have occurred.
For any time after the initial visit, the number needed to harm was 500, increasing to 709 for cancers developing at least a year later and 2,632 for cancers developing at least five years later. Intuitively, this means that 500 people have to meet the cannabis use disorder criteria for a single extra person to develop head or neck cancer.
Dr. Caplan emphasized the importance of this difference, “In context, while the study reports an increased relative risk, the absolute risk of developing these cancers remains low for most people, especially those who use cannabis infrequently or for medical reasons.”
This is important to keep in mind alongside the main result. Yes, the cannabis group in the study had much higher risks of head and neck cancer than the non-cannabis group, but the absolute difference is still relatively small.
Understanding What “Cannabis User” Means in This Study
One very important point we’ve only briefly touched on so far concerns what the “cannabis use” group really means here. The meaning is not that they were simply users of cannabis, and the “non cannabis” group likewise does not mean that these people didn’t use cannabis. The distinction between the groups is actually whether or not they had a diagnosis of a cannabis-related disorder.
The authors explain, along with a link to the ICD code used:
“Cannabis-related disorders are defined by the excessive use of cannabis with associated psychosocial symptoms, such as impaired social and/or occupational functioning.”
The implication of this should be pretty clear if you use cannabis. Do you have a diagnosis of a cannabis related disorder? If the answer is no, you would be in the “non cannabis” group in this study. It doesn’t really matter how much you smoke; it matters if you have a diagnosed issue.
Dr. Caplan also made this point when we spoke to him, calling the reliance on ICD codes a “significant limitation,” and explaining that:
These codes are inherently subjective and often do not reflect the nuances of a patient’s cannabis use. Physicians documenting cannabis use are constrained by the limited options available, which are typically focused on diagnosing problems or disorders. This can lead to an inaccurate representation, where cannabis is marked as a problem even if it is not causing any harm to the patient.
Continuing, “Furthermore, these codes fail to capture essential details such as the amount, potency, or duration of cannabis consumption—factors that are crucial for assessing its true impact on health. As a result, these diagnoses can linger in a patient’s medical chart for years, potentially without proper context, which could skew the study’s findings. The presence of these codes, without an accurate and detailed record of cannabis use, may contribute to misleading associations between cannabis and cancer, undermining the study’s conclusions.”
This doesn’t necessarily mean the result is inaccurate or misleading, but it’s an important limitation to note. This is also mentioned in the study:
“This could decrease relative risks discovered if individuals were using cannabis in the noncannabis group, although this effect may be overcome by the high use in the cannabis use disorder group. This study was further limited by lack of information on dosage and frequency of cannabis use.”
Is the Study About Smoking or Cannabis Use Overall?
One other important point is that the cannabis use discussed in the study was just that, “use,” not specifically smoking. However, as the authors point out in the study and Dr. Kokot explained to us:
“Although our study did not distinguish between different methods of cannabis consumption, smoking is the most common, so the association we observed relates to smoked cannabis. Further research is needed to explore whether other forms of consumption, such as edibles, carry the same risk for head and neck cancers.”
In short, since most cannabis users smoke, this is mainly what the study was measuring. As Dr. Kokot points out, there is not enough information at present to draw any conclusions about the use of edibles or vaping. However, it must be said that the exposure to harmful chemicals from smoking is likely much higher than any such exposure after edibles, so it would be strange – to say the least – if the risks were equivalent.
Dr. Kokot concluded, “The next steps involve designing studies to determine the amount and type of cannabis use and how these factors impact the risk of head and neck cancer. This will help us better understand the specifics of this association.”
The Issue of Reverse Causality and What Happens When It’s Minimized
When we spoke to Dr. Caplan about the study and its limitations, he raised an important point we haven’t touched on yet: the issue of reverse causality. He commented to CBD Oracle that:
A major concern is reverse causality—the possibility that people who are developing or already have undiagnosed cancers might be using cannabis to manage symptoms like pain or discomfort. The study doesn’t adequately account for this, which means it can’t reliably tell us whether cannabis use is a cause or simply a response to these cancers.
Dr. Caplan offers a real-world example of this exact problem:
“Situation: A middle-aged person with chronic pain from a joint condition starts using cannabis to manage their symptoms. Unbeknownst to them, they’re also developing early-stage head and neck cancer.
Misinterpreted Risk: Because they use cannabis, the study might flag them as having a higher cancer risk, even though the cannabis is not the cause. Instead, the cannabis is helping them cope with the pain caused by the undiagnosed cancer.”
While Dr. Caplan didn’t make this connection himself, some of the study’s analyses may shed light on this issue.
As well as reporting the risks for developing head and neck cancer at any point following the “index event” (visiting the outpatient clinic), they also looked at what happens if you only consider cancers that developed a long time after the index event.
If you get a diagnosis of cannabis use disorder just before the “index event,” and then get your cancer diagnosis a month or two later, it could easily be that the cannabis use was a response to the symptoms of your undiagnosed cancer (as in the above example).
If your cannabis use diagnosis is a year or more before you’re diagnosed with cancer, it’s less likely that you already had cancer before your cannabis use problem was diagnosed – though it’s still not impossible.
Although the relative risk for head and neck cancer overall is increased regardless of when the diagnosis happened, the absolute risk increase is just 0.04% for cancers diagnosed more than five years later.
For specific cancer types, the relative risks generally decrease when the analysis is restricted to cancers developing more than a year after the index event. When you restrict the analysis to diagnoses received more than five years after the index event, none of the differences for specific cancer sites are statistically significant.
So in the cases most protected against reverse causality, the difference in absolute risk is very, very small and differences in relative risks eventually become non-significant.
It’s unclear whether reverse causality is the reason for this, but it is certainly a limitation of the study that needs to be kept in mind.
Conclusion – Risks Are Higher for Cannabis Users, But the Details Matter
It’s always challenging to interpret research accurately, and this study is a perfect example of the type of gray area we often operate in.
The study has a huge sample and offers clear evidence that people with diagnosed cannabis disorders are more likely to develop head or neck cancer than people without a diagnosis, but at the same time, there are several crucial limitations to keep in mind.
Firstly, the risks are mainly applicable to smokers, with non-combustion methods like edibles unlikely to carry the same risk.
Secondly, the study really relates to people diagnosed with a cannabis-related disorder (or who could be), with limited relevance to occasional users, since they were probably spread across both groups. And finally, the absolute increase in risk – 0.2% – shows the “real-world” meaning behind the more headline-friendly tripling of risk.
Dr. Caplan emphasizes this point, “This study should be a starting point for more rigorous research, not the final word. It’s crucial to consider these findings in the context of real-life use, where cannabis often plays a key role in managing symptoms rather than causing harm. Until we have more robust data, we should be careful not to overstate the risks, especially when balancing them against the benefits of cannabis, particularly in medical contexts.”
So while stoners should certainly reconsider usage methods other than smoking, it’s important to keep the limitations in mind when reading the headlines.