A unique analysis of environmental contributors to type 2 diabetes has confirmed a link between several pollutants and the disease, while also pointing toward a form of vitamin E as a possible risk factor.
A unique analysis of 266 potential environmental contributors to type 2 diabetes has confirmed a link between several pollutants and the disease, while also pointing researchers for the first time toward a form of vitamin E as a possible risk factor.
The analysis, led by Atul Butte, an HHMI Physician-Scientist Early Career awardee, drew on government datasets to examine a wide range of environmental influences on type 2 diabetes. Butte conceived of an “environment-wide association study,” a parallel to genome-wide association studies that have become a tool of choice among geneticists in recent years. In genome-wide association studies, researchers scan the entire genome of individuals with a particular disease, searching for genetic variations that increase disease risk.
Genetics is certainly a hot field, but so far genome-wide association studies have not explained a large amount of the risk for type 2 diabetes and other diseases. So we decided to borrow the concept of the genome-wide association study and apply it to the environment.
The time has come to put an equal focus on environmental risk factors, says Butte, whose study appears online May 20, 2010, in the journal PLoS One. “Genetics is certainly a hot field, but so far genome-wide association studies have not explained a large amount of the risk for type 2 diabetes and other diseases,” he says. “So we decided to borrow the concept of the genome-wide association study and apply it to the environment.”
While the current study examines only a small number of environmental factors, Butte considers it a “first step” toward broader examinations of environmental influences on disease, including big contributors to diabetes like amount of physical activity and diet and a broad range of pollutants, toxins, and infectious agents. If proven useful, Butte, an assistant professor of medical informatics and pediatrics at the Stanford University School of Medicine, thinks environment-wide association studies could point scientists toward previously unknown risks for many diseases for further study.
Type 2 diabetes affects around 23.6 million people in the United States, or 7.8% of the population, according to the American Diabetes Association. Its prevalence is rising so quickly that the Centers for Disease Control and Prevention (CDC) has labeled it an epidemic. The condition occurs when the body stops responding to insulin, which leads to high concentrations of blood sugar and, if not controlled, a host of health problems, including heart disease, blindness, kidney failure, and lower-extremity amputations.
Previous studies of environmental influences on type 2 diabetes have focused on single factors, such as particulate air pollution or a single type of nutrient. But Butte and graduate student Chirag Patel wanted to analyze as many potential risk factors as possible to shake out previously unknown associations. To do so, they examined data collected by the CDC as part of the ongoing National Health and Nutrition Examination Survey (NHANES). Every two years, the CDC takes a snapshot of the nation’s health via questionnaires and blood and urine samples from a cross-section of Americans. The agency tests each sample for hundreds of pollutants and nutrients, and if they are present, measures their concentrations. Butte wanted to examine every environmental factor in the NHANES data collected in four cohorts between 1999 and 2006.
Each cohort included between 507 and 3,318 people who had their blood and urine tested. Butte divided these participants into a “case” group, who had high concentrations of blood sugar, and a “control” group, who had normal blood sugar. Ideally, each participant’s diabetes status would have been confirmed by a physician, he said, but NHANES does not include physician exams or reviews of medical records. So he chose high blood sugar as a surrogate marker for diabetes, even though using the marker may have excluded diabetics who exhibited good blood sugar control. “If anything, we're missing diabetics, and they're in the control group instead of case group,” he says. “We figured that was a safer way to go.” That way they would be underestimating possible environmental associations rather than overestimating them.
From this mass of data, Butte and Patel performed advanced statistical analysis of 266 individual environmental factors, including a range of pollutants, nutrients, and infectious agents. To guard against spurious associations, Butte, Patel, and their Stanford colleague Jayanta Bhattacharya excluded the influence of body mass index, age, gender, ethnicity, and an estimate of socio-economic status, which all influence the risk for type 2 diabetes. They only considered a risk factor significant if it was associated with diabetes in two or more of the four separate NHANES cohorts they examined.
By comparing the blood or urine concentrations of each factor between the case group and the control group, Butte and Patel identified four factors linked to the disorder. Their analysis confirmed previous findings that high blood levels of industrial pollutants called polychlorinated biphenyls (PCBs) were highly associated with the disease. The prevalence of Type 2 diabetes was two to three times higher for those with the higher levels of the pollutant compared to those with the lowest. Type 2 diabetes prevalence among those with high levels of heptachlor epoxide, a break down product of a previously common pesticide, was at about two times higher than those with low levels of the compound. (The United States banned the manufacture of PCBs in the United States in 1979 and banned heptachlor for most uses in 1988, but the compounds persist in the environment, especially near former industrial sites or contaminated soil.)
The analysis also indentified a factor never before linked to type 2 diabetes: a form of vitamin E called gamma-tocopherol. Vitamin E appears in eight different molecular forms; gamma-tocopherol is the most common form in the American diet. Prevalence of type 2 diabetes among study participants with high blood levels of gamma-tocopherol—which, like other forms of vitamin E, is an antioxidant—was two times greater compared to people with low levels of the nutrient. Butte says that much additional research is needed to sort out how this form of vitamin E is related to type 2 diabetes. “This finding, in particular, shows the value of surveying across as broad a range of environmental factors as possible,” Butte says.
There was also good news. Butte and his colleagues confirmed previous studies showing the protective association of the vitamin beta-carotene. The prevalence of type 2 diabetes among people with high amounts of beta-carotene, a form of vitamin A, was about 40% lower than those with lowest amounts of the vitamin.
Butte hopes the study will spur further examinations of a wide range of environmental factors on disease. While technological advances—including “gene chips” that allow quick scans of variation across the entire human genome—have spurred the adoption of genome-wide association studies, Butte says, “We have to realize the environment plays a huge role as well. One reason the environment gets disrespected is because we don't have a nice, clean chip with which to study it. If this paper inspires researchers or companies to come up with better ways to measure the environmental side of disease risk, we will have done some good.”