Sensationalist headlines have suggested that smoking weed can raise your risk of diabetes, but that wasn’t quite what this latest research found. What the authors actually described was an association between a condition known as prediabetes and marijuana use.

The study found that adults who smoked a lot of the drug throughout their lives were significantly more likely to demonstrate poor blood sugar control – which can progress to type 2 diabetes – compared with those who had never dabbled. But unexpectedly, despite these observations, the team couldn’t find any links between marijuana use and the development of diabetes. The study has been published in Diabetologia.

Marijuana is a popular drug, used by around 19 million people in the U.S. alone. And given its increasing popularity, as both a recreational and a medicinal drug, scientists are pretty keen to probe its potential effects on the body and how these could impact health.

As a psychoactive drug, the literature has consequently been flooded with studies examining what it could do to the brain, which nobody seems to agree on. Metabolic health, on the other hand, seems to have been neglected in comparison, but what has been found so far is nonetheless interesting.

For example, scientists have observed a paradox in which weed users eat more (we all know about the “munchies”) yet have a decreased risk of various indicators of poor metabolic health, like a large waist size. An analysis of eight studies even found marijuana users were 30% less likely to develop diabetes. But not everyone is sold on that – issues have been raised with the methodology in some of the studies.

Hoping to offer some clarity, researchers from the University of Minnesota designed a study to examine the relationship between marijuana use and the incidence of prediabetes and type 2 diabetes. The investigation involved more than 3,000 individuals who were between the ages of 18-30 at the time of recruitment in 1985-1986, meaning they have so far been followed for 30 years.

Alongside gathering data on self-reported marijuana use and incidence of prediabetes and diabetes, the team also looked at BMI and waist size as possible confounding factors. And what they observed over time was rather intriguing.

After accounting for various lifestyle choices and physiological characteristics, they found that current users were 65% more likely to have prediabetes than those who reported they had never used it, whereas those who said they’d used it 100 times or more during their lifetimes were almost 50% more likely to currently have the condition. Interestingly, though, they couldn’t find any links between marijuana use and type 2 diabetes.

Taking this one step further, they looked at individuals who didn’t have diabetes or prediabetes at year seven and assessed them during an 18-year follow-up period. Within these individuals, those who smoked weed a minimum of 100 times throughout their lives had a 40% greater risk of developing prediabetes, but once again not diabetes, compared with those who abstained. They also found no links between use of this drug and BMI.

So not only do these results contradict earlier work, but they are also confusing since you might also expect an increase in diabetes if you see a greater risk of prediabetes. The researchers can only speculate at this stage, but the latter could be due to exclusion of potential participants who were most likely to develop diabetes because in order to qualify, they had to be diabetes-free at the start of the follow-up.

They’re also unsure why marijuana could reduce the ability to control blood sugar, but they postulate that the use of illicit drugs could indicate an overall increased propensity for drug use or activities that are generally bad for your health. Needless to say, more research is warranted.