A 30-year study finds that current or past marijuana use may lead to prediabetes, but not necessarily diabetes, in middle age.

What are the long-term effects of marijuana use on the body’s metabolic health?

That’s what a doctoral student, Michael Bancks, at the University of Minnesota set out to learn when he dug into the data for CARDIA, a long-term health study whose 5,115 participants were recruited by the National Heart, Lung, and Blood Institute (NHLBI) in 1985-1986. At the time, they were ages 18 to 30.

Now, in the 30th year of the study, the group has entered middle age. This is when prediabetes, a major metabolic risk factor, often starts to appear.

By using this data, researchers concluded marijuana use, both past and present, may be associated with a higher chance of developing prediabetes but not necessarily diabetes.

The results of the study were published in Diabetologia, the journal of the European Association for the Study of Diabetes.

Read More: More Than Half of U.S. Adults Have Diabetes or Prediabetes »

Each of the participants in the study gave a blood sample after 12 hours without eating so their fasting blood sugar levels could be measured. They also filled out questionnaires asking about their levels of marijuana use, if any.

The study took into account a wide range of other factors, including age, sex, race, tobacco and alcohol use, education level, medication use, psychosocial well-being, and lifestyle factors like diet, exercise frequency, and other drug use.

Although a number of people dropped out of the study over the three decades and many more were excluded for control reasons, more than 2,500 people were still included in the analysis.

Over the course of the study, more than half of the participants (who did not have prediabetes or diabetes to start) developed prediabetes.

People who were current users of marijuana were 65 percent more likely to have prediabetes compared to people who had never used marijuana, the researchers concluded.

And even among those who had stopped using marijuana as they grew older, using the drug 100 times or more in their lives was still linked to a 23 percent risk increase.

Related Reading: Cannabis-Derived Drug Shows Promise in Treating Type 2 Diabetes »

Prediabetes is a state of high blood sugar that isn’t severe enough to meet criteria for diabetes itself.

“It is a major risk factor for future development of diabetes, but if recognized, it is also an intervention point and opportunity to prevent the progression to diabetes,” said Bancks, lead author of the study, in an interview with Healthline.

About 86 million people in the United States have prediabetes. That’s more than a third of American adults.

And even for those who don’t go on to develop diabetes, prediabetes is still a problem.

“Prediabetes is [also] a risk factor for future cardiovascular disease,” said Bancks. “Damage to the heart and circulatory system may already be occurring if you have prediabetes.”

Learn More: A Short Walk After Meals Is All It Takes to Lower Blood Sugar »

There was, however, no link to diabetes itself.

This raises the question: How do we explain this discrepancy?

There are a number of possible explanations, said Bancks.

The prediabetes finding itself could be wrong. The strict controls that he imposed to clean his data may have selectively excluded people at greater risk for developing diabetes.

Or, it may simply be that marijuana use just isn’t a strong predictor of diabetes compared to other factors.

“Marijuana may have a more noticeable impact on glycemic metabolism in the prediabetes range compared with the diabetes range, when traditional diabetes risk factors are exceedingly less favorable and may dominate over any effect of marijuana,” explained Bancks, who is also a predoctoral fellow of the NHLBI Cardiovascular Disease Epidemiology & Prevention.

Another limitation of the study is that while Bancks controlled for a great many factors, he couldn’t control for everything.

Which means that instead of measuring a direct relationship, he might instead be measuring some other factor that drives up both rates of drug use and poor health outcomes, such as steep income disparities.

“Not controlling for access to nutrition and access and use of preventative healthcare would be major limitations in this study,” said Amanda Reiman, manager of marijuana law and policy for the Drug Policy Alliance and lecturer in the School of Social Welfare at the University of California, Berkeley, in an interview with Healthline. “[This is] another example of trying to point the finger at individual level behaviors rather than focus on the inequities in society.”

Read More: Will Marijuana Follow in the Footsteps of Big Tobacco? »

Whether Bancks’ findings will hold up remains to be seen.

Past research has linked marijuana consumption with reduced risk of diabetes and obesity, so his contradictory results may raise some eyebrows.

On the other hand, this research might be some of the first to find a subtler connection between marijuana use and adverse metabolic outcomes in the form of prediabetes.

For Bancks, this may represent an exciting opportunity to learn more.

“Because the results of this study do not align with the previous research on this topic, and this is one of the first studies to look at the prospective association of marijuana use and metabolic health, it strongly suggests more research is needed on the health effects of marijuana use,” he said.

Learn More: Marijuana Decreases Insulin Resistance, Improves Blood Sugar Control »