Grantee Research Project Results

Aspects of Aim 1, the environmental soil sampling and analysis, were conducted during the first year of the project. Soil samples from two regions in the state of Mississippi, the Delta region and a non-Delta region, were collected and analyzed for select OC compound levels. The objective of this aim was to provide the environmental data that would constitute part of the EPHI. The proposed work was to collect 40 soil samples in each of the two locations. We decided to increase that number to 60 in each of the two regions to represent more locations in the two regions.

 

The additional years of the grant activities concentrated upon the human subjects research by maintaining the Institutional Review Board (IRB) approvals from both Mississippi State University and the Veterans Administration, maintaining approval of these approvals by the U.S. Environmental Protection Agency, utilizing a randomized patient list from which the subjects were recruited, and conducting the organochlorine compound analysis of the serum samples. The procedures at the Sonny Montgomery Veterans Administration (VA) Hospital in Jackson, Mississippi, involved an IRB application, which was approved, followed by a review by the hospital’s Research and Development Committee, which approved the protocol, followed by the construction of a randomized patient list of individuals who met the criteria for the research project (males who lived in the 18-county Delta region or who lived in a select part of Mississippi that was the non-Delta region and who were not service-connected to possible TCDD [“dioxin”] exposure during the Vietnam conflict). (Note: TCDD has been implicated in the literature for association with T2D, so this potential confounder was eliminated.) As patients who were on the randomized list were noted to be scheduled for an appointment, they were telephoned, apprised of the research project and asked if they were interested in participating. Those who provided a positive response were contacted when they came into the hospital for their scheduled appointment; provided the information about the project, including the Informed Consent Form; and had their questions answered by the Investigator at the VA (Dr. Dana Jones) or her Research Associate (both of whom were trained and approved for human subjects research). Those subjects still interested in participating signed the Informed Consent Form, were provided a copy of the form, and allowed the collection of extra tubes of blood for the purposes of this project. The blood was centrifuged and the serum collected and frozen. A code was established to provide to the subject samples and information for the purpose of deidentification of subjects. An information sheet was filled out on each subject with clinical and demographic information. The code identifying the subjects was retained at the VA under lock. The deidentified information sheet and serum samples were supplied to Mississippi State University. All 300 of the approved subjects were recruited and their samples analyzed.

 

The information from the patient information sheet was entered into a spreadsheet. The information sheets are maintained at Mississippi State University under lock.

 

Analysis of OC compounds in serum was performed by gas chromatography/mass spectrometry (GC/MS) following organic solvent extraction. The methodology, developed initially by DPX Labs (Columbia, South Carolina) for analysis of pesticides in fruits and vegetables, was modified in our laboratories for OC extraction from human plasma or serum. An internal standard containing C¹³ p,p’-DDT and C¹³ trans-nonachlor in hexane was added to 1 mL of serum. Acetonitrile was added to the sample to precipitate proteins. Following centrifugation, the resultant supernatant was mixed with deionized water added to the supernatant and the mixture was aspirated into a DPX disposable pipette solid phase extraction column (DPX). OC compounds were eluted from the sorbent matrix, concentrated and resuspended. Concentrations of the target OC compounds were determined by isotope dilution GC/MS. Extracts were analyzed using an Agilent Technologies 6890N gas chromatograph connected to a 5975C triple-axis mass spectrometer. A targeted mass analysis was performed in electron ionization (EI+) mode using single ion monitoring (SIM) for the analytes described above. Quantification and confirmation ions were monitored for each analyte and its respective isotopically labeled internal standard. Limits of quantitation were 100 pg/mL serum (14.8 ng/g serum lipid) for trans-nonachlor, oxychlordane, and p,p’-DDE. Areas under the curve were converted to pg/mL utilizing a standard curve. Values were adjusted for serum lipid content from the clinical data provided by the VA.

 

Data analysis consisted of calculation of descriptive statistics for the clinical and demographic information obtained from the data sheets on the subjects and for the organochlorine analytes quantified in serum samples from the subjects. In addition, these data were analyzed statistically by construction of a multivariable logistic regression model of the organochlorine analytes and clinical and demographic factors in relationship to T2D.

 

The outputs for the project were various organochlorine analyte concentrations in soil samples and serum samples, as well as the clinical and demographic data obtained from the subject information sheets. The outcomes of the project are the statistical associations calculated for factors related to the outcome of T2D in the sample of subjects in this project.

 

Results

 

Out of the 60 soil samples from each region, the number of soil samples with detectable levels of DDE were much higher in the Delta region than the non-Delta region, 40 and 14 respectively. Among the samples with detectable levels, samples from the Delta counties contained approximately 10-fold higher levels of DDE than non-Delta samples (Table 1).

 

Table 1. Soil Levels of DDE From Samples From Delta or Non-Delta Regions of Mississippi, n = 60 for Each Region

 

DDE value is for detectable samples only.

 

Serum samples from both Deltans and non-Deltans displayed a wide range of DDE levels. Seventy-eight percent of Deltan samples contained detectable levels of DDE compared to 71 percent of non-Deltan samples. Mean DDE from detectable serum samples were about 1.5-fold higher in samples from Deltans than non-Deltans, as anticipated. African Americans had about 2.1-fold higher levels of DDE than Caucasians, regardless of region (Table 2). Samples for Delta non-diabetics had three-fold higher DDE levels than samples from non-Delta non-diabetics. However, Delta and non-Delta diabetics had similar serum DDE levels (Table 3).

 

Table 2. Serum Levels of DDE From Subjects Residing in Mississippi

 

 

Table 3. Serum Levels of DDE From T2D and Normal Subjects Residing in Delta and Non-Delta Regions of Mississippi

 

 

The core statistical model used for data analysis was constructed using the clinical and demographic explanatory variables shown in Table 4. DDE and all possible two-way interaction terms were added to the core model and the model was refit. No association of DDE was found with diabetes. The interaction term DDE*Delta was added to the model and was found to be associated with diabetes. The variables age, race and BMI, all previously reported in the literature to be associated with diabetes, were added back and the model was refit. The final resulting model is shown in Table 5. The relationship between DDE and Delta/non-Delta was further explored (Table 6) and there was a significant association with diabetes and increasing DDE concentration in the non-Delta population. The lack of association in the Delta population may be due to the high levels of DDE in the sampled subjects and/or the high level of T2D risk factors in the Delta population, which overrides any environmental factor association with T2D in a highly vulnerable population.

 

Table 4. Clinical and Demographic Explanatory Variables

 

 

Table 5. Core Multivariable Logistic Regression Model With Specific Interaction Terms, n = 300

 

Maximum-rescaled r² = 0.2384

Hosmer Lemeshow goodness of fit = 0.2045

 

Table 6. Relationship of DDE and Delta and Their Association With Type 2 Diabetes Mellitus, n = 300

 

 

The sample collection and chemical analyses were conducted consistent with the quality assurance description of the proposal. Internal standards were used and appropriate blanks, spikes and standard curves were conducted.

 

Summary of Findings

 

While the expectation was that several organochlorine residues would be found in the soils and the sera sampled, this was not the case. Only DDE was present in the soil in a sufficiently large number of soil samples to be reliably quantified. DDE was the major analyte in the serum samples, with trans-nonachlor and oxychlordane quantifiable in occasional serum samples.

 

The levels of DDE in Delta soils and Delta subjects were substantially higher than in the corresponding non-Delta soils and subjects. The statistical analysis indicated that DDE and known risk factors for T2D (i.e., age, race, hypertension, LDL levels and BMI) all are associated with the prevalence of T2D in non-Delta subjects but not Delta subjects.

The Delta region had 10 times higher soil DDE residues than the non-Delta region, as expected. Subjects who resided in the Delta region had higher serum DDE levels than subjects from the non-Delta region; however, both means were higher than the national average, based on the Centers for Disease Control and Prevention National Health and Nutrition Examination Survey data. In retrospect, this was not really surprising, since the OC insecticides, while heavily used in the state of Mississippi and particularly in the Delta, were banned in the early 1970s. Therefore, several decades have passed for these residues to degrade or become non-bioavailable in deeper soil and sediments and thus were not found in the surface soils sampled, nor were they available to continue to contaminate people in recent years. African-Americans had higher levels of DDE than Caucasians.

 

The data analysis using a multivariable statistical model indicated that age, race, BMI, hypertension and high LDL levels were associated with diabetes as was expected. The model also indicated that DDE was associated with T2D in the non-Delta subjects. The lack of association in the Delta population may be due to the very high levels of DDE detected in most all of the subjects. Additionally, if DDE is indeed a risk factor for T2D development it might have been overshadowed in the Delta population, despite its high levels, by the high T2D risk levels in the Delta population (e.g., obesity, high African-American population). The results suggest that DDE may play a role in development of T2D, but it may be masked in populations, such as Mississippi Deltans, who have multiple risk factors and exceptionally high exposure levels.

 

The project’s first hypothesis was supported by the project’s data: Soil levels of OC compounds and serum levels of these compounds in people residing in a region of intense agriculture are greater than levels in soil and people from a less intensely farmed region. The data generated did not support the second hypothesis: A quantitative relationship exists between serum levels of OC compounds and the prevalence of endocrine disorders of glucose metabolism (T2D/prediabetes/increased insulin resistance). Therefore, an EPHI could not be estimated from these data across all the subjects included in this study. However, it is possible that in such a high-risk population as the Mississippi Delta population is an outlier and is obviously non-typical because of its health disparities; the levels of DDE might be predictive of risk in less vulnerable populations. If the latter is true, then DDE levels might serve as an EPHI for more general low exposure/average risk populations; this possibility would require considerably more documentation to verify its use as an EPHI.