An Environment-Wide Association Study (EWAS) on Type 2 Diabetes Mellitus
|
|
Type 2 Diabetes (T2D) and other chronic diseases are caused by a complex combination of many genetic and environmental factors. Few methods are available to comprehensively associate specific physical environmental factors with disease. We conducted a pilot Environmental-Wide Association Study (EWAS), in which epidemiological data are comprehensively and systematically interpreted in a manner analogous to a Genome Wide Association Study (GWAS).
Methods and Findings
We performed multiple cross-sectional analyses associating 266 unique environmental factors with clinical status for T2D defined by fasting blood sugar (FBG) concentration ≥126 mg/dL. We utilized available Centers for Disease Control (CDC) National Health and Nutrition Examination Survey (NHANES) cohorts from years 1999 to 2006. Within cohort sample numbers ranged from 503 to 3,318. Logistic regression models were adjusted for age, sex, body mass index (BMI), ethnicity, and an estimate of socioeconomic status (SES). As in GWAS, multiple comparisons were controlled and significant findings were validated with other cohorts. We discovered significant associations for the pesticide-derivative heptachlor epoxide (adjusted OR in three combined cohorts of 1.7 for a 1 SD change in exposure amount; p<0.001), and the vitamin γ-tocopherol (adjusted OR 1.5; p<0.001). Higher concentrations of polychlorinated biphenyls (PCBs) such as PCB170 (adjusted OR 2.2; p<0.001) were also found. Protective factors associated with T2D included β-carotenes (adjusted OR 0.6; p<0.001).
Conclusions and Significance
Despite difficulty in ascertaining causality, the potential for novel factors of large effect associated with T2D justify the use of EWAS to create hypotheses regarding the broad contribution of the environment to disease. Even in this study based on prior collected epidemiological measures, environmental factors can be found with effect sizes comparable to the best loci yet found by GWAS.
——
Introduction
It is becoming clear that most non-communicable diseases are a result of a complex combination of genetic processes and the environment [1]. Despite the contribution of both genetics and environment to disease, many recent studies have emphasized the genetic components. For example, the Genome-wide Association Study (GWAS) is a low-cost and popular framework used by researchers to evaluate genetic factors that correlate with disease status on a genome-wide scale [2]. As of this writing, 370 publications using this method have been cataloged, with 16 just for Type 2 Diabetes Mellitus (T2D) [3]. Multiple loci markers have been found through these studies that heightens risk for T2D when present [4]. While GWAS has enabled the generation of new hypotheses regarding the relation of genetics to T2D, the genetic markers found have poor penetrance [5], [6]. Further, these genetic markers do not explain a significant portion of T2D in context of other factors [7], [8].
Perhaps the lack of impact of GWAS comes from not comprehensively considering environmental factors in disease. T2D provides an specific example: while genetics play a large role [9]–[11], specific environmental factors are also emerging as risk factors for the disease [12]. It is clear that we need to measure and assess both types of factors to better understand complex disease [1].
The current paradigm to search for the effects of multiple environmental chemicals utilizes molecular tools and model systems [13], [14]; however, there is a gap between these data and human disease. Epidemiological searches for environmental factors associated with disease have been hampered by the lack of a “chip” or standard bioassays that can broadly survey these factors. We propose borrowing the GWAS methodology to create a model Environmental-Wide Association Study (EWAS), to search for environmental factors associated with disease on a broad scale. This type of study is made possible by the use of cross-sectional epidemiological data, the National Health and Nutrition Examination Survey (NHANES), a nationally representative, biannual health survey conducted by the Centers For Disease Control and Prevention (CDC) [15]. Participants are queried regarding their health status and an extensive battery of clinical and laboratory tests are performed on a subset of these individuals. Specific environmental attributes are assayed, such as chemical toxicants, pollutants, allergens, bacterial/viral organisms, and nutrients.
The EWAS consists of two methodological steps that have analogs in a GWAS. First, we consider a panel of 266 unique environmental assays, or environmental “loci”, measured across cases of diabetics and controls, yielding several environmental factors with significantly high association with T2D while controlling for multiple hypotheses. Second, we validate the associations by taking advantage of data from other cohorts in NHANES. With EWAS, we are able to hypothesize about new associations with T2D and reconfirm others. The results from EWAS can better inform about environmental factors that need to be measured in genetic studies to begin to provide us insight in regards to disease etiology.
Ethics Statement
The NHANES is a publicly available dataset made available by the CDC and National Centers for Health Statistics and all participants have provided written consent.
We associate 266 unique environmental factors with T2D status from the NHANES. We downloaded the all of the available NHANES data for 1999–2000, 2001–2002, 2003–2004, and 2005–2006 cohorts and collated corresponding variables across them. For example, if a variable with symbol LBXVIE from 1999–2000 described “A-Tocopherol ug/dL” and variable with symbol LBXATC from 2001–2002 also described “a-tocopherol ug/dL”, we harmonized onto the single symbol for both, LBXATC.
Figure 1 presents a schematic representation of our analysis methodology. We analyzed all environmental factors from the NHANES that were a direct measurement of an environmental attribute, such as the amount of pesticide or heavy metal present in urine or blood. We did not consider internal biological system laboratory measures such as red blood cell count, triglyceride level, cholesterol level, or other physiological measures. By using direct and quantitative measures of factors, we potentially avoid issues of self-report bias.
—
Chirag J. Patel, Jayanta Bhattacharya, Atul J. Butte
Print Version
Tell-a-Friend comments powered by Disqus
