Cannabis use may affect diabetes risk in the middle-aged

"People who use marijuana may be more likely to develop pre-diabetes than those who have never smoked it," The Independent reports, after a US study found a link between long-term cannabis use and pre-diabetes.

Pre-diabetes is defined as having abnormally high blood sugar levels, but not high enough to meet criteria for diagnosis of type 2 diabetes.

The study enrolled around 3,000 healthy young US adults in the mid-1980s. Over the following years, researchers carried out regular medical assessments and questioned participants about their use of cannabis and other substances.

Cannabis use at the 25-year assessment, when the person was now in middle age, was associated with an increased risk of having pre-diabetes. However, there were no significant links between cannabis use and "full-blown" diabetes.

The main difficulty with this research is that the study design cannot prove direct cause and effect. Many other health and lifestyle factors could be linked to both cannabis use and diabetes risk, such as diet.

Cannabis is a notorious appetite stimulant – know as "the munchies", which often leads users to eat energy-rich, nutritiously poor snacks, such as crisps and sweets. If there is a link, it's possible that diet could be having an effect on diabetes risk, rather than cannabis itself.

While the short- and long-term effects of cannabis are not firmly established, the drug has been linked to mental health conditions such as psychosis and physical conditions such as lung cancer

Where did the story come from?

The study was carried out by researchers from the University of Minnesota and the University of California, San Francisco. It received various sources of financial support, including from the US National Institutes of Health.

The study was published in the peer-reviewed journal Diabetologia on an open-access basis, so it is free to read online as a PDF (384kb).

The Independent and the Mail Online's reporting of the study is accurate, although both articles could benefit from highlighting that this study cannot prove direct cause and effect. 

What kind of research was this?

This cohort study aimed to see whether cannabis use is associated with the presence or development of diabetes or pre-diabetes.

Pre-diabetes is when the person has blood glucose levels just below the threshold for meeting the criteria for diabetes. If the person doesn't make lifestyle changes, such as changing their diet, upping their physical activity and trying to lose weight, it can progress to type 2 diabetes.

Cannabis, or marijuana, has uncertain effects on a person's physical or mental health. In the US, where this study was based, it is the most frequently used illegal drug, with 18.9 million people over the age of 12 reportedly having used cannabis in 2012.

Recent studies have suggested that cannabis use may be associated with reduced odds of diabetes and other metabolic risk factors, such as a high body mass index (BMI) and waist circumference. The researchers report the possibility of bias with these studies, and the need for prospective studies to better examine these links.

In this study, researchers aimed to look at the link between self-reported cannabis use and the presence of diabetes or pre-diabetes (cross-sectional link) or the development of these conditions (prospective link).

The main limitation with this type of study is not being able to prove cannabis use has caused the diabetic conditions, as other factors may have had an influence – particularly with the cross-sectional association.   

What did the research involve?

This study involved participants of the Coronary Artery Risk Development in Young Adults (CARDIA) study. They were recruited from four urban areas in the US and aged 18 to 30 years at the time of enrolment in 1985-86.

At enrolment and each follow-up, the participants completed questionnaires and had clinical examinations, including blood tests and measurements of blood pressure and BMI. Questionnaires involved assessments of their health and lifestyle, including physical activity, alcohol, smoking and use of illegal substances.

The substance assessment asked specifically about the use of cannabis, crack or other cocaine, amphetamines or opiates in the person's lifetime or past 30 days, with frequency of once or twice, 3 to 9 times, 10 to 99 times, more than 100, or more than 500 times.

Pre-diabetes and diabetes were defined by blood glucose levels using American Diabetes Association criteria. For example, pre-diabetes was a fasting blood glucose of 5.6 to 6.9 millimole (mmol) per litre, and diabetes was a level of 7.0mmol per litre or greater.

The cross-sectional link between lifetime cannabis use and pre-diabetes or diabetes was assessed at the last follow-up assessment, around 25 years after enrolment.

The prospective link was examined between cannabis use seven years after enrolment and the later development of pre-diabetes or diabetes by year 25. The assessments included around 3,000 people.

When looking at the links between cannabis use and diabetes, the researchers took into account potential confounders – the use of other substances, smoking and alcohol, educational attainment, and examination findings, including BMI, blood pressure and cholesterol.  

What were the basic results?

Factors associated with cannabis use were being male, of white ethnicity, greater reported smoking, alcohol and other substance use, and greater physical activity.

Higher educational attainment and higher BMI were factors associated with less cannabis use. By the age of 24, 45% of the participants (1,193) had pre-diabetes and 357 had diabetes.

With full adjustment for all confounders, current use of cannabis was associated with about a two-thirds increased odds of pre-diabetes compared with never using the drug (hazard ratio [HR] 1.66, 95% confidence interval [CI] 1.15 to 2.38).

There were no significant links between pre-diabetes and former cannabis use. When broken down into frequency of use, there was a trend for increased lifetime use to be associated with an increased risk of pre-diabetes.

However, the only significant link was found for a lifetime use of 100 or more times being associated with a 40% increased risk of pre-diabetes (HR 1.40, 95% CI 1.13 to 1.72). There were no convincing links for a lower frequency use than this.

There was no statistically significant link between former, current or any lifetime use of cannabis and actual diabetes.  

How did the researchers interpret the results?

The researchers concluded that, "Marijuana [cannabis] use in young adulthood is associated with an increased risk of pre-diabetes by middle adulthood, but not with the development of diabetes by this age." 


This long-term study of healthy US adults found current cannabis use at the 25-year assessment – when the person had reached middle age – was associated with an increased likelihood of the person having pre-diabetes at this time.

Higher lifetime use of more than 100 times was also associated with an increased likelihood of pre-diabetes. However, there were no significant links between cannabis use and actual diabetes.

The main limitation of this study comes from the possibility of confounding. The researchers have attempted to take several confounders into account, including smoking and the use of alcohol and other substances.

However, various physical and mental health, lifestyle, personal and socioeconomic characteristics may be associated with both cannabis use and diabetes risk. For example, one possible factor that could be linked to both cannabis use and diabetes risk is poor diet.

Cannabis use can cause sudden and intense hunger pangs, nicknamed "the munchies". This can lead users to snack on foods with a high calorie and sugar content, but with little in the way of nutritional value. If maintained on a long-term basis, this type of diet can lead to obesity, which is a risk factor for type 2 diabetes.

This study is not able to account for the influence of all these factors, particularly as the main link was for the current use of cannabis at the 25-year assessment and pre-diabetes at the same time. This cannot prove that one thing has caused the other.

There was no link with type 2 diabetes itself. Pre-diabetes suggests the person may be on the border of developing diabetes, but they don't yet have the condition.

Another – admittedly unavoidable – limitation is that cannabis use was self-reported. This may be inaccurate, particularly when it comes to estimating the lifetime frequency of use. There is also the possibility when questioning people about their use of illegal substances that they may report never using them, when in fact they have.

This urban sample of US citizens may not be representative of everyone, particularly given they were enrolled 30 years ago. Patterns of cannabis use during the 80s and 90s may differ from use of the substance today. In particular, the strength of cannabis in terms of one of the active ingredients, tetrahydrocannabinol (THC), is thought to be much stronger than in the past.

The various possible effects of cannabis on physical and mental health – both in the immediate and longer term – are often debated. However, this study alone provides no proof that cannabis use will increase your risk of diabetes.

Cannabis remains a class B drug that is illegal to possess or distribute.

NHS Attribution