Vol. 14 Issue 1, Winter 2009
By Suzanne M. Snedeker, Ph.D., Associate Director for Translational Research, BCERF
Since the late 1990s research studies have explored whether exposure to environmental chemicals affects the risk of diabetes. A recent review by Dr. David Carpenter, professor at the Institute for Health and Environment at the University of Albany, traces the evidence and possible role of various environmental chemicals in the etiology of diabetes (1). In Dr. Carpenter’s review, he relates that while earlier studies focused on exploring linkages between environmental contaminants such as dioxin and polychlorinated biphenyls (PCBs) and the risk of diabetes, in more recent years evidence has accumulated suggesting that various persistent chlorinated pesticides may be related to diabetes. In addition, data from the Agricultural Health Study support a role for other less persistent pesticides (such as triazines and organophosphates) in the risk of diabetes in farm workers who are pesticides applicators (5), as well as linkages to gestational diabetes in their spouses (6). This Research Commentary will review the emerging evidence and the data gaps that need to be filled to answer the important public health question of whether general or occupational exposures to environmental chemicals affect the risk of diabetes.
The studies that support a role for persistent organochlorines and diabetes risk include studies published by several different groups of investigators that analyzed data from the 1999-2002 National Health and Nutrition Examination Survey (NHANES) (2-3), and a Hispanic version of this survey (4). The NHANES studies, conducted by the Centers for Disease Control and Prevention (CDC), collect health status data and analyze blood and urine samples for a variety of environmental contaminants. Both the general US population and ethnic subgroups, including Hispanic populations, have been included in these surveys. These surveys suggested that self-reported diabetes was related to a variety of persistent organic pollutants (POPs), including dioxins, PCBs, and pesticides such as DDT and a metabolite of chlordane called oxychlordane (2, 3, 4). The strengths of these studies include that researchers assessed health risks in a cross-section of the American population and blood levels of the various chemicals were measured to approximate past exposures. Disadvantages of these studies include that the diabetes was “self-reported” rather than obtaining diagnosis information from medical records.
As these studies have emerged, questions still remain about their interpretation. It has been hard to determine if any of these associations are truly “causal.” Did exposure to the specific chemicals measured in the NHANES and earlier studies cause the diabetes in the subjects studied, or are there alternative explanations? For instance, as has been hypothesized in Carpenter’s review (1), many of the persistent organochlorines are fat-soluble and the primary source of exposure is from animal fat (meat and dairy) and fatty fish. Therefore, it is possible that the particular chemical that was measured may be a surrogate for another chemical that was not measured, but that may occur in the same food source. Many fat-soluble chemicals tend to co-migrate as they travel up the food chain. Hence, it has been very difficult to determine the role of individual POPs as causal factors in inducing diabetes (1).
Other questions include, do the chemicals act directly to affect diabetes risk, or do they affect other risk factors for diabetes? For example, do the chemicals play a role in the onset of obesity? The answers to these questions are still being explored. There are other trends that do not have an explanation. The blood levels of many of the POPs have been dropping in Americans in the last 20 years (most were banned by the late 1970s), and yet the incidence of diabetes has been increasing in the American population during the same time period. This is the opposite of what would be predicted if organochlorines alone were the driving force for causing diabetes in the last two decades. While studies done to date do suggest that certain persistent chlorinated chemicals may play some role in determining diabetes risk, more work is needed to evaluate a wider range of environmental chemicals.
The NHANES studies provided evidence that exposures to chlorinated chemicals from common sources, e.g. contaminated food, could affect diabetes risk in the US population. But, few studies have determined if occupational exposures to these and other chemicals affect the risk of diabetes. And no studies have explored whether chemical exposures could affect the risk of gestational diabetes in women. Some of these questions have been explored by investigators from the Agricultural Health Study, a long-term, large-scale study of how pesticide use affects a variety of health endpoints in pesticide applicators and their family members (subjects are from Iowa and North Carolina; see http://aghealth.nci.nih.gov/).
Unlike the NHANES studies, blood levels of the pesticides were not measured in the Agricultural Health Study. Rather, detailed questionnaires approximated past use of many different types of pesticides, including major classes of pesticides (insecticides, herbicides, and fungicides). The team of researchers, led by Dr. Dale Sandler, an epidemiologist at the National Institute of Environmental Health Sciences (NIEHS), obtained information on lifetime exposure to pesticides and self-reported diabetes self-reported in 33,457 licensed pesticide applicators (5). The researchers looked at two measures of pesticide use: whether the applicators reported ever using the pesticide and cumulative lifetime days of use.
Of the 50 pesticides included in the survey, seven pesticides were associated with an increased incidence of diabetes in the applicators (5). This included five insecticides; three were organochlorines that persist in the environment but are no longer in use (aldrin, chlordane, and heptachlor) and two were organophosphate pesticides that are still used in agricultural settings (dichlorvos and trichlorfon). Dichlorvos is used to control insects in barns and on livestock. Trichlorfon was used in the past to control soil insects to protect food crops and parasites on livestock, but currently its major use is to control grubs and other soil insects in ornamental plants used in landscaping, and in lawns and golf courses (6, 7). This study is one of the first to report that use of trichlorfon is associated with a higher risk of diabetes in occupationally exposed individuals (use was significantly higher in diabetics compared to non-diabetics in the Agricultural Health study subjects). The two herbicides associated with a higher risk of diabetes included alachlor (one of the most highly-used herbicides in the US), and cyanazine, a triazine herbicide that has been phased out of use in recent years.
There is other evidence that organophosphate pesticide exposure is linked to diabetes. There is evidence from animal studies demonstrating that this class of insecticides can change how the sugar glucose is processed, as well as evidence that in pesticide poisonings, glucose is higher in the blood or urine when people were exposed to high levels of certain organophosphate insecticides (as cited in 5). However, the Agricultural Health Study is the first study to show that long-term occupational exposure to organophosphate pesticides may increase the risk of diabetes in applicators. These findings should be confirmed in other occupations that have used organophosphate pesticides for pest control, including structural pest control operators, those working in horticulture and greenhouses, golf-course managers, and other turf pesticide professionals.
What is also remarkable is what this study did not show. Exposure to pesticides that are known to be contaminated with dioxin, such as the phenoxy herbicides 2,4,5-TP, were not associated with a significantly higher risk of diabetes in the Agricultural Health Study applicators (5), which is in contrast to findings from the NHANES studies (2-4) and other earlier studies that did show a relationship between dioxin exposure and diabetes (1). Past use of DDT, an insecticide associated with higher diabetes risk in the NHANES studies (2-4), also was not associated with a higher risk of diabetes in the pesticide applicators enrolled in the Agricultural Health Study (5). The reason for the differences between these outcomes is not known.
Despite the interest in whether environmental chemicals affect the risk of diabetes in the general US population and through occupational exposures from agricultural use in predominantly male applicators, few studies have investigated whether there are chemical linkages to gestational diabetes in women. This gap was addressed in a study of pesticide exposures and gestational diabetes in female spouses of applicators in the Agricultural Health Study (8). Female spouses of applicators enrolled in the study completed questionnaires that included information on past pregnancies, self-reported gestational diabetes, and exposure to pesticides. Information on four types of pesticide exposure were assessed: 1) agricultural exposure (mixing or applying pesticides or repairing application equipment as a part of farm work); 2) indirect exposure (planting, pruning, weeding, picking or harvesting crops); 3) residential use (applying pesticides inside the home or the garden); or, 4) a report of no exposure. A significantly higher risk of gestational diabetes was associated with seven different pesticides used only in the first use category, agricultural exposures. This included two organophosphate insecticides (diazinon and phorate), one carbamate insecticide (carbofuran), and four herbicides (2,4,5-T, 2,4,5-TP, atrazine, and butylate). There was no increased risk of diabetes in women who were indirectly exposed to pesticides or who were exposed through residential use.
None of the specific pesticides associated with gestational diabetes (8) were the same as the pesticides seen to increase diabetes in the applicators (predominantly male) in the study discussed previously (5). However, in both studies there were pesticides in the same class that were related to diabetes risk. For instance, atrazine is a triazine herbicide related to the risk of gestational diabetes (8). Atrazine is similar in structure to the triazine herbicide cyanazine, which was associated with a higher risk of diabetes in the applicator study (5). Also, in the gestational diabetes study and the applicator study, certain organophosphate pesticides were associated with a higher risk of diabetes, though the specific pesticides were different in the two studies. Unlike the applicator study, the gestational diabetes study did observe an association with pesticides known to be contaminated with dioxin (the herbicides 2,4,5-T and 2,4,5-TP), which is consistent with other studies that have seen a higher risk of diabetes with dioxin exposure (1-4).
There are limitations in the gestational diabetes study. The number of women exposed to the different pesticides evaluated was small, and for this reason, the results of this study need to be confirmed. Further studies on whether there are chemical linkages to gestational diabetes are extremely important. As mentioned in the conclusion of the gestational diabetes study (8), 4% of women in the US develop gestational diabetes (9) and women with gestational diabetes have a 20-50% chance of developing adult onset diabetes within ten years (10). Hence, finding factors that may contribute to gestational diabetes and subsequent development of adult onset diabetes is important in understanding causes and prevention of diabetes in women.
There are many ongoing studies to determine how environmental chemicals affect diabetes risk, and I recommend Dr. Carpenter’s review (1) for one of the best overviews of all the different mechanisms and genes affected by exposures to various chemicals that may play a role in the etiology of diabetes. It is evident that far more mechanisms than changes in glucose metabolism and insulin resistance are affected by chemical exposures. Specific genes and sets of genes can be activated or deactivated by various chemicals, and these actions at a cellular and molecular level may play an important role in the development of diabetes. But, while many genes have been identified in studies evaluating how environmental chemicals affect diabetes risk, the exact mechanisms have yet to be identified.
Given the rise of diabetes in recent years, and its public health consequences, it is very important to establish whether there are linkages to specific chemicals or classes of chemicals. Once established, strategies to decrease exposures to these chemicals in everyday and occupational settings will be important to decrease the risk of this chronic disease.