The Centers for Disease Control and Prevention says smoking is a cause of Type 2 diabetes and makes the disease harder to control. The latest Surgeon’s General report and just about everyone else advises against this, too. But new results from the ongoing Multi-Ethnic Atherosclerosis Study doesn’t find this to be the case — not consistently, anyway.

When we say smoking, we mean tobacco in all its forms: cigars, pipes, and smokeless tobacco such as chew. Studies have suggested that exposure can increase diabetes risk up to 40 percent; it’s thought to directly cause insulin resistance. The thing is, most of these studies have been conducted on groups that were either mostly white or mostly Asian, the authors said. The researchers were curious to find out if these findings would hold up across races and ethnicities, including African Americans, Latinos, and Chinese people.

They recruited nearly 7,000 participants aged 45 to 84 and gave them urine cotinine tests. Smokers excrete nicotine through their urine, and the authors could see participants’ smoking histories. They considered those who had smoked fewer than 100 cigarettes or 20 cigars in their lifetime abstinent, while participants who had smoked more, but not in the past 30 days, were labeled former smokers. The rest were active smokers.

The study authors compared these smoking habits to insulin resistance and diabetes over a 10-year span. Results showed that current smokers had lower blood pressure and relied less on medications for high blood pressure and cholesterol than abstinent and former smokers. There was no significant difference in glucose or insulin levels, two markers of diabetes, across the groups.

Overall, there was no significant association between the intensity of participants’ tobacco exposure, as in the actual number of products consumed, and insulin resistance. There was also no consistent association between smoking and development of prediabetes.

This isn’t to say lighting up may not lead to diabetes after all, but it pokes yet another hole in the predominantly white and male scientific model clinical research largely follows. One of the reasons authors think their findings conflict with the results of others is their study’s diverse sample. Minorities bear the burden of diabetes compared to whites, according to a 2012 study, and when you break that down even further, some groups within minority populations have an even greater number of risk factors. While smoking is certainly one, it’s not as prevalent as others. For example, in this study, African-Americans were exposed to more tobacco than Latinos.

Other studies may also have inadequately adjusted for confounding variables like age, gender, ethnicity, BMI, income, and medication, the authors said. It's also possible that available data has been skewed by publication bias. “This follows from the fact that ‘positive’ studies may be as much as three times more likely to be published than negative studies,” the authors said, leaving findings like theirs out of the public eye.

That doesn’t mean this study is perfect either. The authors cite several limitations, including their loose definition of former smokers — evidence shows that it can take longer than 30 days to see any adverse effects quitters experience, which include onset of diabetes. The number of non-cigarette smokers, as well as the focus on middle-aged to elderly men and women may have affected the results as well.

Source: Keith RJ, et al. Tobacco Use, Insulin Resistance, and Risk of Type 2 Diabetes: Results from the Multi-Ethnic Study of Atherosclerosis. PLOS One. 2016.