Veterans and Agent Orange: Update 2012 (2014)

12


Cardiovascular and Metabolic Outcomes

Chapter Overview

Based on new evidence and a review of prior studies, the committee for Update 2012 found one new association: limited or suggestive evidence of association between the relevant exposures and stroke. Current evidence supports the findings of earlier studies concerning cardiovascular and metabolic outcomes:

•    No adverse cardiovascular or metabolic outcome has sufficient evidence of an association with the chemicals of interest.

•    There is limited or suggestive evidence of an association between the chemicals of interest and type 2 diabetes, hypertension, ischemic heart disease, and now stroke.

•    There is inadequate or insufficient evidence to determine whether there is an association between the chemicals of interest and for all other adverse cardiovascular or metabolic outcomes.

This chapter summarizes and presents conclusions about the strength of the evidence from epidemiologic studies regarding an association between exposure to the chemicals of interest (COIs)—2,4-dichlorophenoxyacetic acid (2,4-D), 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) and its contaminant 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), picloram, and cacodylic acid—and type 2 diabetes and circulatory disorders. The committee also considers studies of exposure to polychlorinated biphenyls (PCBs) and other dioxin-like chemicals to be informative if their results were reported in terms of TCDD toxic equivalents (TEQs) or concentrations of specific congeners. Although all studies reporting

TEQs based on PCBs were reviewed, studies that reported TEQs based only on mono-ortho PCBs (which are PCBs 105, 114, 118, 123, 156, 157, 167, and 189) were given very limited consideration because mono-ortho PCBs typically contribute less than 10% to total TEQs, based on the World Health Organization revised toxicity equivalency factors (TEFs) of 2005 (La Rocca et al., 2008; van den Berg et al., 2006).

TYPE 2 DIABETES

Diabetes mellitus is a group of heterogeneous metabolic disorders characterized by hyperglycemia and quantitative or qualitative deficiency of insulin action (Orchard et al., 1992). Although all forms share hyperglycemia, the pathogenic processes involved in its development differ. Most cases of diabetes mellitus are in one of two categories: type 1 diabetes is characterized by a lack of insulin caused by the destruction of insulin-producing cells in the pancreas (β cells), and type 2 diabetes is characterized by a combination of resistance to the actions of insulin and inadequate secretion of insulin (called relative insulin deficiency). In old classification systems, type 1 diabetes was called insulin-dependent diabetes mellitus or juvenile-onset diabetes mellitus, and type 2 was called non-insulin-dependent diabetes mellitus or adult-onset diabetes mellitus. Type 1 diabetes occurs as a result of immunologically-mediated destruction of β cells in the pancreas, which often occurs during childhood but can occur at any age. As in many autoimmune diseases, genetic and environmental factors influence pathogenesis. Some viral infections are believed to be important environmental factors that can trigger the autoimmunity associated with type 1 diabetes. The modern classification system recognizes that type 2 diabetes can occur in children and can require insulin treatment. Long-term complications of both types can include cardiovascular disease (CVD), nephropathy, retinopathy, neuropathy, and increased vulnerability to infections. Keeping blood sugar concentrations within the normal range is crucial for preventing complications.

About 90% of all cases of diabetes mellitus are of type 2, and type 2 has been the type of diabetes that epidemiologic investigations relevant to Vietnam veterans have addressed. Onset can occur before the age of 30 years, and incidence increases steadily with age. The main risk factors are age, obesity, abdominal fat deposition, a history of gestational diabetes (in women), physical inactivity, ethnicity (prevalence is greater in blacks and Hispanics than in whites), and family history. The relative contributions of those features are not known. Prevalence and mortality statistics in the US population for 2006 are presented in Table 12-1.

The etiology of type 2 diabetes is unknown, but three major components have been identified: peripheral insulin resistance (thought by many to be primary) in target tissues (muscle, adipose tissue, and liver), a defect in β-cell secretion of insulin, and overproduction of glucose by the liver. In states of insulin resistance, insulin secretion is initially higher for each concentration of glucose

TABLE 12-1 Prevalence of and Mortality from Diabetes, Lipid Disorders, and Circulatory Disorders in the United States, 2009/2010

NOTE: ICD, International Classification of Diseases; nr, not reported.

^aFor all ages.

^bGap in ICD-9 sequence follows.

SOURCE: AHA, 2012, pps. e6-e245.

than in people who do not have diabetes. That hyperinsulinemic state is a compensation for peripheral resistance and in many cases keeps glucose concentrations normal for years. Eventually, β-cell compensation becomes inadequate, and there is progression to overt diabetes with concomitant hyperglycemia. Why the β cells cease to produce sufficient insulin is not known.

Pathogenetic diversity and diagnostic uncertainty are among the important problems associated with epidemiologic study of diabetes mellitus. Multiple likely pathogenetic mechanisms lead to diabetes mellitus, which include diverse genetic susceptibilities (as varied as autoimmunity and obesity) and all sorts of potential environmental and behavioral factors (such as viruses, nutrition, and activity level). The multiplicity of contributing factors can lead to various responses to particular exposures. Because up to half of the cases of diabetes are undiagnosed, the potential for ascertainment bias in population-based surveys is high (more intensively-followed groups or those with more frequent health-care contact are more likely to get the diagnosis); this emphasizes the need for formal standardized testing (to detect undiagnosed cases) in epidemiologic studies.

Scientists have named a clustering of cardiovascular risk factors—including hypertension, hyperglycemia, high triglycerides, abdominal obesity, and low high-density lipoprotein—metabolic syndrome. Although it is not a disease entity itself, metabolic syndrome is associated with a fivefold increased risk of type 2 diabetes and a doubling of the risk of cardiovascular disease (Alberti et al., 2009). There is a growing literature on the association between the COIs and metabolic syndrome and its components. Given its strong linkage with type 2 diabetes, new literature that deals with metabolic syndrome as an outcome will be discussed primarily in this section.

Conclusions from VAO and Previous Updates

The committee responsible for Veterans and Agent Orange: Health Effects of Herbicides Used in Vietnam (VAO; IOM, 1994) concluded that there was inadequate or insufficient information to determine whether there is an association between exposure to the COIs and diabetes mellitus. Additional information available to the committees responsible for Update 1996 (IOM, 1996) and Update 1998 (IOM, 1999) did not change that conclusion.

In 1999, in response to a request from the Department of Veterans Affairs, the Institute of Medicine called together a committee to conduct an interim review of the scientific evidence regarding type 2 diabetes. That review focused on information published after the deliberations of the Update 1998 committee and resulted in the report Veterans and Agent Orange: Herbicide/Dioxin Exposure and Type 2 Diabetes (Type 2 Diabetes; IOM, 2000). The committee responsible for that report determined that there was limited or suggestive evidence of an association between exposure to at least one COI and type 2 diabetes. The committees responsible for Update 2000 (IOM, 2001), Update 2002 (IOM, 2003), Update 2004 (IOM, 2005), Update 2006 (IOM, 2007), Update 2008 (IOM, 2009), and Update 2010 (IOM, 2012) upheld that finding. Reviews of the pertinent studies are found in the earlier reports. Table 12-2 presents a summary.

Update of the Epidemiologic Literature

Vietnam-Veteran Studies

No Vietnam-veteran studies addressing exposure to the COIs and diabetes have been published since Update 2010.

Occupational Studies

Ruder and Yiin (2011) reported the mortality experience of 2,122 workers involved in the production of 2,3,4,5,6-pentachlorophenol (PCP) at four plants in the United States through 2005. One-third of the cohort also worked in departments that used tricholorophenol (TCP) or its derivatives that were contaminated with TCCD. The mortality experience of the workers was compared with that of the US general population. Diabetes mortality was not higher in the 720 workers exposed to both PCP and TCP, among whom only 8 deaths were ascertained (standardized mortality ratio [SMR] = 1.14, 95% confidence interval [CI] 0.49–2.24).

Waggoner et al. (2011) reported that the risks of mortality from diabetes from 1993 to 2007 among both the applicators and spouses in the Agricultural Health Study (AHS) (without consideration of specific exposures) were significantly decreased in comparison to the general public. The AHS has been generating

TABLE 12-2 Selected Epidemiologic Studies—Diabetes and Related Health Outcomes (Shaded Entries Are New Information for This Update)

NOTE: 2,4-D, 2,4-dichlorophenoxyacetic acid; 2,4-DP, 2,4-dichlorophenoxypropanoic acid; 2,4,5-T, 2,4,5-trichlorophenoxyacetic acid; 2,4,5-TCP, 2,4,5-trichlorophenol; 2,4,5-TP, 2-(2,4,5-trichlorophenoxy) propionic acid; ACC, Army Chemical Corps; AFHS, Air Force Health Study; BMI, body-mass index; CARDIA, Coronary Artery Risk Development in Young Adults; CATI, computer-assisted telephone interviewing; CDC, Centers for Disease Control and Prevention; CI, confidence interval; COI, chemical of interest; dl, dioxin-like; HbA1c, hemoglobin A1c; HpCDD, 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin; HxCDD, 1,2,3,6,7,9-hexachlorodibenzo-p-dioxin; IARC, International Agency for Research on Cancer; ICD, International Classification of Diseases; IU, international unit; MCPA, 2 methyl-4-chlorophenoxyacetic acid; MCPB, 4-(4-chloro-2-methylphenoxy)butanoic acid; MCPP, methylchlorophenoxypropionic acid; MOS, military occupational specialty; NHANES, National Health and Nutrition Examination Survey; NIOSH, National Institute for Occupational Safety and Health; nr, not reported; ns, not significant; OCDD, 1,2,3,4,6,7,8,9-octachlorodibenzo-p-dioxin; OR, odds ratio; PCB, polychlorinated biphenyl; PCDD, polychlorinated dibenzo-p-dioxin; PCDD/Fs, chlorinated dioxins and furans combined; PCP, pentachlorophenol; PIVUS, Prospective Investigation of the Vasculature in Uppsala Seniors; ppt, parts per trillion; SEA, Southeast Asia; TCDD, 2,3,7,8-tetrachlorodibenzo-p-dioxin; TCDF, tetrachlorodibenzofuran; TCP, trichlorophenol; TEF, toxicity equivalency factor; TEQ, (total) toxic equivalent; VA, US Department of Veterans Affairs.

^aGiven when available; results other than estimated risk explained individually.

^bStudy is discussed in greater detail in Veterans and Agent Orange: Herbicide/Dioxin Exposure and Type 2 Diabetes (IOM, 2000).

^cIncludes some subjects covered in other references cited in the category occupational cohorts.

valuable information on the COIs for a number of years, but these results, like those in Alavanja et al. (2005) and Blair et al. (2005), are not herbicide-specific and so are not regarded as being fully informative for the committee’s task.

Boers et al. (2012) reported an updated mortality analysis of workers exposed to TCDD at two Dutch chlorophenoxy-herbicide production facilities. The analysis did not distinguish diabetes deaths within the category “disease of endocrine system and blood.”

Environmental Studies

Lee et al. (2010) reported the results of a nested case-control study within the Coronary Artery Risk Development in Young Adults (CARDIA) cohort study that looked for a relationship between several persistent organic pollutants (POPs) and type 2 diabetes. The investigators identified 116 CARDIA participants who were free of diabetes when they provided a blood sample in the 1987–1988 examination but developed diabetes during the next 18 years. Of the 116, 90 cases were randomly selected for further analysis. Controls (90) were randomly selected from all participants who had a fasting glucose concentration below 100 mg/dL at all CARDIA examinations in which glucose was measured. After adjustment for age, race, sex, and body-mass index (BMI), none of the three measured PCBs that had dioxin-like activity (mono-ortho PCBs 156, 157, and 167) was associated with incident diabetes.

Lee et al. (2011b) studied the association between plasma concentrations of PCBs and the development of type 2 diabetes in participants in the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS). PIVUS participants were residents of Uppsala, Sweden, who were 70 years old in the period April 2001-June 2004. Of the 2,205 selected, 1,106 chose to participate in the baseline examination. The final analytic sample included 989 participants for the cross-sectional baseline analyses: 112 who had type 2 diabetes at baseline, 152 who were not available at followup 5 years later, and 725 who participated in the analysis of diabetes incidence. On the basis of either a high fasting plasma glucose concentration or the use of hypoglycemic medication, 36 incident cases were ascertained. In addition to 3 lipophilic organochlorine pesticides, 14 PCBs were measured, including 5 dioxin-like, mono-ortho PCBs (PCBs 105, 118, 156, 157, and 189). Estimates of the relative risk of diabetes associated with exposure were adjusted for various risk factors: BMI, exercise, alcohol consumption, triglyceride and total cholesterol concentrations, cigarette-smoking, and sex. For the total concentration of all 14 PCBs, the diabetes risks for the top three quintiles were significantly elevated compared with that of the lowest quintile, and there was a strongly significant increasing trend in risk with rising quintile (p ≤ 0.001). Every measured PCB was associated with elevated diabetes risk comparing high to low quintiles. For the 5 dioxin-like, mono-ortho PCBs, the highest quintile odds ratios (ORs) ranged from 2.6 to 8.0, but none of their risk elevations compared

to the lowest exposure quintiles was statistically significant with the exceptions of the second quintile for PCB 105 (adjusted OR = 10.7, 95% CI 1.3–84.4), the third quintile for PCB 189 (adjusted OR = 3.5, 95% CI 1.0–11.9), and the fourth quintile for PCB 118 (adjusted OR = 5.2, 95% CI 1.1–25.5), PCB 157 (adjusted OR = 3.5, 95% CI 1.0–12.4); also, none of their tests for a linear trend was statistically significant. For the 9 PCBs without dioxin-like activity, however, there were 16 significant results for paired comparisons with the lowest quintile and four modestly significant dose–responses, so the mono-ortho dioxin-like PCBs do not appear to be driving the overall observed association with diabetes risk.

Anniston, Alabama, is a location with substantial environmental PCB contamination. Silverstone et al. (2012) gathered survey results from 1,110 participants recruited from a targeted sample of 1,823 households in the community (61% participation); 774 of them (70%) also had clinical examinations. Diabetes (202 cases) was defined on the basis of a physician diagnosis of diabetes or a fasting glucose concentration greater than 125 mg/dL. Concentrations of 35 PCBs were measured. Analyses by logistic regression adjusted for BMI, family history, age, ethnicity, sex, education, marital status, and duration of residence in Anniston and excluded 171 individuals with prediabetes. The total PCB burden was associated with increased diabetes prevalence (OR per standard deviation increase in blood concentrations = 1.23, 95% CI 0.88–1.72). The prevalence of diabetes also increased with increasing TEQs based only on mono-ortho PCBs (OR per standard deviation increase in blood concentrations = 1.20, 95% CI 0.87–2.02). Neither increase was significant, and they were of similar magnitude.

The association between prevalent diabetes and PCB burden was also assessed in the Saku Control Obesity Program (Tanaka et al., 2011). Subjects were 235 participants in an obesity-treatment trial. Diabetes was determined on the basis of blood measures or a prescription for a hypoglycemic medication. Fifteen participants had definite diabetes. There were no statistically-significant associations between the two dioxin-like mono-ortho PCBs measured (PCBs 118 and 156) and diabetes prevalence.

Case-Control Studies

No case-control studies of exposure to the COIs and type 2 diabetes have been published since Update 2010.

Other Reviewed Studies That Address Metabolic Syndrome and Risk Factors for Diabetes

Chang et al. (2010b, 2011a) report findings from a cross-sectional sample of Taiwanese living in an area with a high level of industrial contamination updating and extending an earlier report that was discussed in Update 2010 (Chang et al., 2010a). The updated report extended the survey period for an additional 7 months

(July 2005-December 2007), increasing from 1,478 to 1,812 the number of area residents who were surveyed and from whom blood was collected and analyzed for dioxins, furans, and mercury. Metabolic syndrome and its components were examined in 1,490 non-diabetic residents (Chang et al., 2010b). The prevalence of metabolic syndrome increased steeply and monotonically with dioxin TEQs based on congener analysis of polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs). For those in the 90th percentile of TEQs (≥ 62.2 pg/g lipid), the prevalence of metabolic syndrome was about 37%, compared with less than 5% for those in the 10th percentile. The odds of having metabolic syndrome in the highest TEQ quintile were 2.3-fold higher than in the lowest quintile (95% CI 1.3–3.9) after adjustment for age, sex, smoking, alcohol use, obesity, and family history of diabetes or hypertension. Of the metabolic syndrome components, dioxin TEQ was most strongly related to diastolic blood pressure, fasting glucose, and waist circumference. There were strong associations between most individual congeners concentrations and metabolic syndrome prevalence; TCDD, 2,3,4,7,8-pentachlorodibenzofuran, and 1,2,3,6,7,8-hexachlorodibenzofuran showed the strongest associations. This population was also coexposed to environmental mercury, which was explored in a subsequent paper on a subset of this study population (n = 1,449). Chang et al. (2011a) showed that TEQs, based on congener analysis of PCDDs and PCDFs, were strongly associated with insulin resistance, but not defective pancreatic β-cell function, for which the results was independent of mercury exposure.

Chang et al. (2012) surveyed 156 employees 50 years old or older from a closed Taiwanese plant that had manufactured sodium pentachlorophenol (Na-PCP). The process had involved very high exposure to COIs for the committee. Clinical-chemistry values of the workers were compared with those of residents of similar ages in the general population. It should be noted that the workers lived in the community mentioned in the preceding paragraph that was exposed to dioxin through the accidental release of Na-PCP. Previous employment at the plant was associated with much higher concentrations of dioxin-like PCDDs and PCDFs (109.6 ± 94.5 pg/g of lipid) than the general population (22.9 ± 10.0). After adjustment for age, sex, smoking, and drinking, the employees had high odds of having increased glucose concentrations (> 100 mg/dL, OR = 7.22, 95% CI 4.04–12.90); and high odds of having increased triglyceride concentrations (> 150 mg/dL, OR = 4.31, 95% CI 2.57–7.22). There was not an increased prevalence of elevated serum cholesterol (> 200 mg/dL, OR = 0.95, 95% CI 0.64–1.41).

Obesity is a strong risk factor for diabetes. Rönn et al. (2011) examined the cross-sectional relationship between POPs and fat mass determined by dual-energy X-ray absorptiometry in 890 participants in the PIVUS study. They calculated the differences in total fat mass among categories of PCB exposure and adjusted for sex, height, lean mass, smoking, exercise habits, education, daily energy intake, and alcohol consumption. Of the 16 PCBs measured, 7 were dioxin-

like; 2 were not mono-ortho (PCBs 126 and 169) with higher dioxin-equivalency factors (TEFs) than the mono-ortho PCBs (PCBs 105, 118, 156, 157, and 189). Fat mass was 0.672 kg lower for each log-unit increase in circulating dioxin-like PCB 126 (p = 0.003) and 3.4 kg lower for each log-unit increase in circulating dioxin-like PCB 169 (p = 0.0001). Those in the highest quintile of PCB 126 and PCB 169 exposure had 2.2 and 6.9 kg less fat mass, respectively, than those in the lowest quintile. Although dioxin-like mono-ortho PCBs 126, 156, 157, and 189 also showed strong inverse associations with fat mass, paradoxically, two other dioxin-like mono-ortho PCBs, 105 and 118, showed strong positive associations with fat mass. Further, non-dioxin-like PCBs also showed strong inverse associations (such as PCBs 116, 138, 170, 180, 206, and 209). When all congeners were modeled simultaneously, the dioxin-like mono-ortho PCB 118, non-dioxin-like PCB 138, and octachlorobenzene-p-dioxin (OCDD) were positively associated with fat mass, and non-dioxin-like PCB 180 was inversely associated with fat mass. A significant sex interaction was detected for several of the associations with both dioxin-like mono-ortho PCBs and non-dioxin-like PCBs. The directions of the associations were similar in men and women but much stronger in the women.

The PIVUS study investigators also examined the association between organic pollutants and abdominal obesity (Lee et al., 2012a), which is more strongly linked to diabetes and cardiovascular disease risk than is overall adiposity. Abdominal obesity was defined as a waist circumference greater than 102 cm in men or greater than 88 cm in women. The study examined both the cross-sectional and longitudinal associations between waist circumference and the analyzed pollutants. At baseline, plasma concentrations of dioxin-like PCB 126 (the measured chemical with the highest dioxin-equivalent activity) were associated with a significantly higher prevalence of abdominal obesity in men. Some other dioxin-like mono-ortho PCBs (105 and 118) were also associated with abdominal obesity, but other mono-ortho PCBs (156, 157, and 189) were not, nor was OCDD. In both men and women, it appeared that the association depended on the number of chlorine atoms on the molecules regardless of dioxin activity.

Lee et al. (2011a) examined the associations between blood concentrations of POPs and the trajectories of change in obesity, dyslipidemia, and insulin resistance in the 90 controls who were selected for the CARDIA diabetes analysis discussed above. The study participants were young adults (mean age, 27.2 years) at the time of exposure assessment during the study’s year-2 examination. Metabolic measures were reassessed in the year-20 examination. The paper reports the year-20 values by quintile of each pollutant adjusted for age, sex, race, baseline value of the measure, triglycerides, and total cholesterol at year 2. The measured congeners that had dioxin-like activity were mono-ortho PCBs 105, 118, 156, 157, and 167. Of those, PCB 156 concentration was associated with significantly lower BMI at year 20 (p for linear trend = 0.02). The mean BMI of those who had the highest PCB 156 concentrations was 29.7; it was 34.3 in those who had the

lowest concentrations. No significant associations were seen with year-20 plasma triglyceride, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol levels, or HOMA-IR (an index of insulin resistance).

Biologic Plausibility

Several biologic mechanisms that have been studied in cell culture and animal models may explain the potential diabetogenic effects of TCDD in humans. TCDD is known to modify expression of genes related to insulin transport and signaling and to inflammation (Kim MJ et al., 2012). The present committee’s literature review included two new studies that increased mechanistic biologic plausibility. Wang et al. (2011) found that mice that lacked the aryl hydrocarbon receptor (AHR) (AHR knockouts) had enhanced insulin sensitivity and glucose tolerance; this suggested that the AHR has a physiologic function in glucose metabolism and supported the speculation that sustained activation of the AHR by dioxin-like chemicals could contribute to diabetes. That would be consistent with results of a previous study by Kurita et al. (2009), who found that exposure of mice to dioxin significantly reduced insulin secretion after a glucose challenge. In an in vitro study of differentiated adipocytes, TCDD significantly reduced insulin-stimulated glucose uptake (Hsu et al., 2010). Thus, mechanisms associated with insulin signaling and glucose uptake may contribute to the diabetogenic effects of TCDD observed in humans.

Synthesis

The new epidemiologic evidence on diabetes reviewed in this update includes followup on three occupationally-exposed study populations and several new studies of environmental exposure. The studies of Boers et al. (2012), Ruder and Yiin (2011), and Waggoner et al. (2011) all analyzed the mortality experience of workers. Diabetes mortality is not a useful endpoint because complications of diabetes, rather than diabetes itself, are most often listed as the cause of death of those who have the disease, so most cases of diabetes would be missed if mortality data were used. In addition, the workers are compared with the US population, which is likely to have a worse mortality experience than the workers studied because workers had to have some level of health to be employed.

The retrospective comparisons of Korean Vietnam-era veterans with acute coronary syndrome on the basis of whether they did or did not service in Vietnam (Kim JB et al., 2012) are not helpful in assessing whether herbicide exposure played a role in the development of diabetes. Two new informative prospective epidemiologic studies of exposure to some COIs and incident diabetes appeared since the last update (Lee et al., 2010, 2011b). The prospective design is helpful because it ensures that an exposure to a chemical occurred before disease onset, a requirement for any causal relationship. That is an important consideration for

exposures that are measured in blood, in that the disease process or the treatment of the disease may affect the magnitude of the exposure, and the measured value might not represent an exposure before disease onset. Both studies were able to include many potential confounding variables, such as obesity and physical activity, and this helps to reduce the risk of bias. In spite of those strengths, both studies are difficult to interpret with respect to exposures of Vietnam veterans. In the CARDIA study (Lee et al., 2010), the investigators measured only three mono-ortho PCBs, and they had only weak dioxin-like activity. In addition, the exposures were small. Thus, the putative dioxin burden would be expected to be much lower than that of veterans, and the lack of association might be due to an insufficient dose. The PIVUS investigators (Lee et al., 2011b) measured many more COIs, but the exposures were also small. Only 36 new cases were ascertained, so the results are imprecise. Nevertheless, there was an association between the highest concentrations of the chemicals and diabetes risk. The results were not specific for pollutants that had dioxin activity, and many PCBs that lack dioxin activity were also found to be associated with diabetes risk. Although the positive association is consistent with an association between dioxin and diabetes onset, the lack of exposure specificity implies that the basis of the association might not be related to a dioxin-specific pathway.

In addition to those prospective studies, two surveys were considered. One included residents of a community in Alabama that had substantial environmental contamination. Some evidence suggested a higher prevalence of diabetes in residents that had increased dioxin TEQs, but the association was not specific for PCBs that had dioxin-like activity. In addition, the participation rate of sampled persons was only 61%, so selection bias introduced by differential nonparticipation cannot be ruled out. The other survey was based on a convenience sample of residents of Saku, Japan, who were enrolled in the obesity-control program. The small number of cases (15) and the small number of relevant compounds (mono-ortho PCBs 118 and 156) prevent definitive interpretation.

Several studies looked at COIs and risk factors for diabetes. The PIVUS investigators related PCBs and OCDD to both total fat mass and abdominal obesity. The cross-sectional results suggest that, if anything, blood concentrations of PCB 126 and PCB 169 with the highest level of dioxin-like activity were associated with lower levels of obesity, although OCDD with no dioxin activity also was associated with increased body fat. The relationship between PCBs and abdominal obesity seemed to involve the number of chlorine atoms in the molecule rather than any dioxin-like activity. Lee et al. (2011a) looked at the 20-year changes in diabetes-related risk factors in the 90 CARDIA participants selected as controls for their nested case-control study of diabetes. Only mono-ortho PCB 156 was associated with BMI over the 20 years; those who had higher PCB 156 concentrations at baseline had lower BMIs 20 years later. Measures of glucose homeostasis over 20 years were unrelated to PCB concentrations at baseline. Again, the stud-

ies involved populations that had low absolute exposure, and their results suggest nonspecific effects based on chemical attributes other than dioxin-like activity.

Chang et al. (2010b, 2011a) showed that in Taiwanese who lived in a heavily-contaminated area there was a strong association between TEQs based on PCDDs and PCDFs and the prevalence of metabolic syndrome and insulin resistance, strong risk factors for diabetes. Chang et al. (2012) also showed that workers in the plant making the pesticide had higher fasting glucose concentrations than the general population. Those reports are of interest because of the high exposures to TCDD and related chemicals that had very strong dioxin activities. The studies also adjusted for many relevant risk factors. The strong associations support a link between exposure to the COIs and diabetes. Nevertheless, the cross-sectional design limits the strength of conclusions that can be drawn.

In the aggregate, the newly added studies support prior VAO committees’ inclusion of diabetes in the limited and suggestive category. The negative studies either examined suboptimal endpoints or were conducted in populations that had very low exposure to relevant chemicals. The new study from the PIVUS cohort does show an association with dioxin-like compounds, but the association was not specific to compounds that have dioxin-like properties. Its interpretation is further limited by the small number of cases, which impedes the ability to explore exposure specificity. The survey of a highly-exposed Taiwanese population shows a strong association with the diabetes risk factor metabolic syndrome. However, because of its cross-sectional design, it is not possible to prove that the exposure to the putative cause preceded the onset of the outcome.

Conclusion

On the basis of the evidence reviewed here and in previous VAO reports, the committee reaffirms its conclusion that there is limited or suggestive evidence of an association between exposure to at least one COIs and diabetes.

CIRCULATORY DISORDERS

This section covers a variety of conditions encompassed by the 9th revision of the International Classification of Diseases, Ninth Revision (ICD-9), codes 390–459, such as acute and chronic rheumatic fever (ICD-9 390–398), hypertension (ICD-9 401–404), ischemic heart disease (IHD; ICD-9 410–414), heart failure (ICD-9 428), cerebrovascular disease (ICD-9 430–438), and peripheral vascular disease (ICD-9 443). Coronary heart disease is related specifically to atherosclerosis; ischemic heart disease is broader and typically includes atherosclerosis and its symptoms. The American Heart Association reports mortality related to coronary heart disease, not to its symptoms, which include angina and myocardial infarction (MI). Table 12-1 contains estimates of prevalence of and mortality from individual disorders of the circulatory system in the US population in 2009.

Circulatory diseases are a group of diverse conditions, of which hypertension, coronary heart disease, and stroke are the most prevalent and account for 75% of deaths from circulatory diseases in the United States. In addition to family history, the major risk factors for circulatory diseases include age, race, smoking, serum cholesterol, BMI or percentage of body fat, and diabetes. Ideally, epidemiologic investigations of circulatory diseases would consider the conditions in this category separately rather than together because they have different patterns of occurrence, and many have different etiologies. However, many mortality studies follow the ICD-9 rubric and report deaths from circulatory diseases together. Deaths from coronary or IHD, heart failure, and, to a lesser extent, stroke predominate. Many of the reports also break out subcategories (such as cerebrovascular disease and hypertension). The relative importance of heart failure would be determined by the age of the cohort. In younger cohorts, most of the deaths in this category would be expected to be from IHD. Cerebrovascular deaths are deaths from strokes, which can be classified as either ischemic or hemorrhagic. In the US population, the great majority of strokes are of the ischemic type.

The methods used in morbidity studies can involve the direct assessment of the circulatory system, including analysis of symptoms or history, physical examination of the heart and peripheral arteries, ultrasound measurements of the heart and arteries, electrocardiography (ECG), chest radiography, cardiac computed tomography (CT), and more recently cardiac magnetic resonance imaging (MRI). Ultrasonography, CT, and MRI can be used to visualize the heart and related vasculature directly. ECG can be used to detect heart conditions and abnormalities, such as arrhythmias (abnormal heart rhythms), heart enlargement, and heart attacks (myocardial infarctions). Chest radiography can be used to assess the consequences of IHD and hypertension, such as the enlargement of the heart seen in heart failure. It is sometimes difficult to determine the time of onset of clinical findings, so the temporal relationship between exposure and disease occurrence may be uncertain. Cardiovascular-disease epidemiologists prefer to observe cohorts over time for the incidence of discrete clinical events, such as acute myocardial infarction (ideally verified on the basis of changes in ECG readings and enzyme concentrations) and death due to heart disease. The onset of new angina symptoms or the performance of a revascularization procedure in a person who has no history of disease is also used as evidence of incident disease. In many occupational studies, only mortality information is available. The attribution of death to a vascular cause is often based on a death certificate, whose accuracy can be uncertain.

The practice of evaluating the evidence on hypertension separately from that on other circulatory diseases was established in Update 2006; separate consideration of IHD began in Update 2008. The number of studies with data on stroke and cerebrovascular disease is increasing, so this endpoint can be considered in its own right in this report separately from discussions of “other circulatory diseases.”

Conclusions from VAO and Previous Updates

The committee responsible for VAO concluded that there was inadequate or insufficient information to determine whether there is an association between exposure to the COIs and circulatory disorders. Additional information available to the committees responsible for Update 1996, Update 1998, Update 2000, Update 2002, and Update 2004 did not change that conclusion.

The committee responsible for Update 2006 reviewed new studies and intensively revisited all the studies related to IHD and hypertension that had been discussed in previous updates and concluded that there is limited or suggestive evidence to support an association between exposure to the herbicides used in Vietnam and hypertension. That committee was unable to reach a consensus as to whether that was also the case for IHD, so that outcome remained in the category of inadequate evidence.

After consideration of the relative strengths and weaknesses of the evidence regarding the COIs and IHD (ICD-9 410–414) and the relevant toxicologic literature, the committee responsible for Update 2008 judged that the evidence was adequate to advance this health outcome from the “inadequate or insufficient” category into the “limited or suggestive” category, again acknowledging that bias and confounding could not be entirely ruled out. That conclusion was not changed in Update 2010. The committee for Update 2010 also reaffirmed the Update 2006 conclusion of limited or suggestive evidence of an association between herbicide exposure and hypertension.

The previous studies and studies published since Update 2010 are all summarized in Table 12-3.

Update of the Epidemiologic Literature

Hypertension

Vietnam-Veteran Studies No Vietnam-veteran studies addressing exposure to the COIs and hypertension have been published since Update 2010.

Occupational Studies No occupational studies addressing exposure to the COIs and hypertension have been published since Update 2010.

Environmental Studies The survey of residents of an area of high dioxin contamination in Taiwan was introduced above (Chang et al., 2010b). The authors used factor analysis to determine which components of metabolic syndrome appeared to be most strongly associated with dioxin TEQ concentrations, based on serum PCDDs and PCDFs. Dioxin TEQs were more strongly associated with blood pressure than other syndrome components. In multivariate models, they found a highly-statistically-significant association between TEQ concentrations

TABLE 12-3 Selected Epidemiologic Studies—Circulatory Disorders (Shaded Entries Are New Information for This Update)

NOTE: 2,4-D, 2,4-dichlorophenoxyacetic acid; 2,4-DP, dichlorprop; 2,4,5-T, 2,4,5-trichlorophenoxyacetic acid; 2,4,5-TCP, 2,4,5-trichlorophenol; 2,4,5-TP, 2-(2,4,5-trichlorophenoxy) propionic acid; 2,5-DCP, 2,5-dichlorophenol; ACC, Army Chemical Corps; BMI, body mass index; bw, body weight; CATI, computer-assisted telephone interviewing; CDC, Centers for Disease Control and Prevention; CHD, coronary heart disease; CI, confidence interval; COI, chemical of interest; dl, dioxin-like; HDL, high-density lipoprotein; HpCDD, 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin; HpCDF, 1,2,3,4,6,7,8-heptachlorodibenzofuran; HR, hazard ratio; HxCDD, 1,2,3,6,7,8-hexachlorodibenzo-p-dixion; HxCDF, 1,2,3,4,7,8-hexachlorodibenzofuran; IARC, International Agency for Research on Cancer; ICD, International Classification of Diseases; JEM, job-exposure matrix; MCPA, 2-methyl-4-chlorophenoxyacetic acid; MCPB, 4-(4-chloro-2-methylphenoxy)butanoic acid; MCPP, methylchlorophenoxypropionic acid; MOS, months of service; NHANES, National Health and Nutrition Examination Survey; NIOSH, National Institute for Occupational Safety and Health; nr, not reported; ns, not significant; OCDD, 1,2,3,4,6,7,8,9-octachlorodibenzo-p-dioxin; OR, odds ratio; PCB, polychlorinated biphenyl; PCDD, polychlorinated dibenzo-p-dioxin; PCDD/F, dioxins and furans combined; PCDF, polychlorinated dibenzofuran; PCP, pentachlorophenol; PeCDF, 2,3,4,7,8-pentachlorodibenzofuran; PMR, proportional mortality ratio; POP, persistent organic pollutant; ppt, parts per trillion; SEA, Southeast Asia; SMR, standardized mortality ratio; TCDD, 2,3,7,8-tetrachlorodibenzo-p-dixoin; TCP, trichlorophenol; TEF, toxicity equivalency factor for individual congener; TEQ, (total) toxic equivalent; VA, US Department of Veterans Affairs; VV, Vietnam veterans.

^aSubjects male unless otherwise noted.

^bGiven when available; results other than estimated risk explained individually.

and diastolic blood pressure (p = 0.006) but not systolic blood pressure after adjustment for age, sex, obesity, smoking status, alcohol use, and family history of hypertension or diabetes. That finding was also reported in a further analysis of the sample by Chang et al. (2011b), who examined the associations between dioxin and related chemicals and CVD risk by using the Framingham risk score and its components.

Goncharov et al. (2011) examined data on the 394 residents of Anniston, Alabama, to determine the relationship between blood pressure and serum concentrations of 35 PCBs and 9 chlorinated pesticides. Persons taking anti-hypertensive medications were excluded. Other than age, total serum PCB concentrations were the strongest correlate of blood pressure after adjustment for age, BMI, sex, race, smoking status, and exercise. The authors saw a weak association between blood pressure and mono-ortho PCBs, but it was not statistically significant. PCBs that had more potent dioxin-like activity were not measurable within the limits of the assay used.

Jones et al. (2011) examined urinary arsenic concentrations and hypertension in the 2003–2008 National Health and Nutrition Examination Survey and observed no statistically significant association.

Case-Control Studies No case-control studies of exposure to the COIs and hypertension have been published since Update 2010.

Ischemic Heart Disease

Vietnam-Veteran Studies Kim JB et al. (2012) studied outcomes of Vietnamera South Korean veterans undergoing coronary angiography because of acute coronary syndrome (chest pain arising from unstable angina, ST-segment myocardial infarction, or non-ST-segment myocardial infarction) according to whether they served in Vietnam (121) or did not serve (130). This study examines whether a history of Vietnam service is associated with the clinical course of coronary disease, not the occurrence of coronary disease itself. Thus, its findings are not relevant to the question of whether the COIs cause ischemic coronary disease.

Occupational Studies Boers et al. (2012) reported an updated mortality analysis of workers exposed to TCDD at two Dutch chlorophenoxy-herbicide production facilities. Results of that cohort have been included in previous VAO reports. The new analysis advances previous reports by using semiquantitative dose estimates derived by back-extrapolating exposures on the basis of blood TCDD concentrations measured in blood of 197 former employees. Workers in plant A were exposed to high concentrations of dioxin both as a contaminant of 2,4,5-T production and through accidental exposure after the explosion of a kiln. Plant B was involved in 2,4-D production, but TCDD exposure was assumed to be minimal. The study followed all male employees of either factory during their years

of operation, which lasted until 1985 for plant A and 1986 for plant B. Mortality was ascertained through the end of 2006. There were 339 deaths among the 1,020 workers included from plant A and 202 deaths among the 1,036 workers from plant B. No risk factors for heart disease other than age were available. About 93 deaths were due to IHD. There was a significant association between death from IHD and dioxin exposure (hazard ratio [HR] for natural log-transformed TCDD exposure = 1.19, 95% CI 1.08–1.32). The estimated association was slightly higher when the analysis was restricted to plant A (HR = 1.24, 95% CI 1.091.43). The authors examined dioxin exposure as a categoric variable and found that the higher mortality was evident only in those who had the highest putative TCDD exposure, ³9.9 parts per trillion (ppt) (HR = 2.60, 95% CI 1.57–4.31).

Manuwald et al. (2012) reported the 23-year followup of workers exposed to dioxins in a chemical plant in Hamburg, Germany, that manufactured herbicides and pesticides, including 2,4,5-T. Results on that cohort have been included in previous VAO reports. The study included 1,191 men and 398 women who were employed full-time at the plant for at least 3 months during 1952–1984. The mortality experience of employees was compared with that of the populations of Hamburg and the Federal Republic of Germany over the same period. Worker exposure was estimated on the basis of the putative intensity of exposure in different work areas in the plant. About 695 deaths were reported through the end of followup on December 31, 2007. Mortality from diseases of the circulatory system was significantly higher than expected in men (SMR = 1.16, 95% CI 1.02–1.31) but significantly lower than expected in women (SMR = 0.74, 95% CI 0.56–0.94). Estimated exposure to TCDD was substantially higher in male workers (median = 77.4 ppt) than in female workers (median = 19.5 ppt). When workers were categorized in quartiles by estimated exposure, there was no evidence that the standardized mortality ratio from circulatory diseases increased with higher estimated dose.

Ruder and Yiin (2011) reported the mortality experience of 2,122 workers involved in the production of PCP in four plants in the United States through 2005. Details are reported above. In comparison to the general US population, IHD was not significantly elevated in the 720 workers exposed to both PCP and TCP; 131 deaths were ascertained (SMR = 1.11, 95% CI 0.92–1.54). Workers in two of the four plants had TCDD exposures sufficient to cause chloracne. Workers in these two plants had higher SMRs for IHD than workers in the other two plants, but these elevations were not higher than those expected based on the US population rates.

Waggoner et al. (2011) reported on mortality in the Agricultural Health Study from 1993 to 2007. Previous reports on the cohort have documented that most of its members have been exposed to COIs (2,4-D and 2,4,5-T), but this report describes the mortality experience of only the cohort as a whole without consideration of pesticide-specific exposures. Therefore, the information presented

cannot be used either to confirm or to refute associations between the COIs and cardiovascular-disease mortality.

Environmental Studies Lind et al. (2012) studied the relationship between exposure to POPs and the prevalence of carotid artery plaque in the PIVUS study by using ultrasonography. Sonography can be used to obtain images of atherosclerotic lesions (plaques) in the arterial wall and to measure arterial wall thickness (intima-media thickness, IMT). Thickness is a measure of the extent of the atherosclerotic process, correlates with atherosclerosis in the coronary arteries, and predicts acute coronary events. The technique also provides a measure of the extent to which the arterial walls reflect sound (echogenicity)—poor echogenicity is an indication of lipid content, and high echogenicity is an indication of calcification and collagen. Low echogenicity also predicts cardiovascular events but does not correlate well with IMT and thus provides complementary information. Echogenicity may be more strongly related to inflammatory and lipid factors, whereas IMT is more strongly related to blood pressure and smoking history. The investigators measured serum concentrations of 16 PCB congeners and other POPs in 990 participants, all 70 years old. The evaluated PCBs were detectable in more than 90% of the study population. The authors analyzed individual congeners and a summary score based on dioxin TEQs. Statistical models adjusted for sex, waist circumference, BMI, blood glucose, systolic and diastolic blood pressure, HDL and LDL cholesterol, serum triglycerides, smoking, and use of antihypertensive drugs and HMG CoA reductase inhibitors. Concentrations of seven POPs were associated with the presence of plaque, but there was no clear correspondence with dioxin activity. Plaque presence was most strongly associated with pollutants that have six or more chlorine atoms. Total dioxin TEQs were not associated with plaque, but TEQs of mono-ortho-substituted PCBs were. The prevalence of plaque increased by 67% for every natural-log increase in the serum concentration of these substances (OR = 1.67, 95% CI 1.20–2.32). Total TEQs were significantly associated with IMT (p = 0.009). Among the substances measured, TEQs from dioxin-like coplanar non-ortho-substituted PCBs were more strongly related to IMT than those of mono-ortho-substituted PCBs. Echogenicity was inversely associated with the concentrations of PCB 126, a PCB that has strong dioxin-like properties, and of the highly-chlorinated PCBs. Total TEQs were also strongly and inversely associated with echogenicity (p = 0.001). The association was restricted to TEQs of dioxin-like coplanar non-ortho-substituted PCBs.

Case-Control Studies No case-control studies of exposure to the COIs and ischemic heart disease have been published since Update 2010.

Other Reviewed Studies In a further analysis of the survey of Taiwanese residents of an area that had heavy dioxin contamination, Chang et al. (2011b)

related serum dioxin TEQs, based on PCDDs and PCDFs, to future cardiovascular disease risk by examining relationships with individual risk factors and with the Framingham risk score, which is used in the United States to estimate 10-year risk of CVD. Compared with those who had dioxin TEQs in the lowest 25th percentile, those in the 75th percentile had more than 6 times the odds of being at high cardiovascular risk (OR = 6.22, 95% CI 2.47–15.63). However, there did not appear to be an increase in serum cholesterol, compared with the general population, in Taiwanese workers who were exposed to very high dioxin TEQ concentrations (Chang et al., 2012).

Obesity, blood pressure, and diabetes are all risk factors for IHD, and findings related to their associations with the COIs have been summarized in the previous sections (Chang et al. 2010, 2011a, 2012; Goncharov et al., 2011; Lee et al., 2011a, 2012a; Rönn et al., 2011).

Cerebrovascular Disease and Stroke

Vietnam-Veteran Studies No Vietnam-veteran studies addressing exposure to the COIs and cerebrovascular disease and stroke have been published since Update 2010.

Occupational Studies Several of the occupational studies summarized above also presented data on cerebrovascular-disease mortality, which is due primarily to stroke. Neither Ruder and Yiin (2011) nor Boers et al. (2012) observed an increase in cerebrovascular deaths among the workers compared to the general population (SMR = 1.00, 95% CI 0.61–1.54; SMR = 0.98, 95% CI 0.83–1.16, respectively). Manuwald et al. (2012), however, observed a significantly higher-than-expected mortality from cerebrovascular disease in 1,149 male workers (54 deaths, SMR = 1.57, 95% CI 1.19–2.05)—but not in the 388 female workers (11 deaths, SMR = 0.64. 95% CI 0.32–1.15).

Environmental Studies The PIVUS investigators (Lee et al., 2012b) examined the relationship between POPs in 898 70-year-old residents of Uppsala, Sweden, and their incidence of stroke 5 years later. The investigators measured 16 PCBs, OCDD, and 4 other pollutants. Thirty-five participants developed stroke; stroke subtype was not determined. All ORs discussed below were adjusted for gender, BMI, cigarette-smoking, exercise, alcohol consumption, hypertension, diabetes, triglycerides, and serum cholesterol. Plasma concentrations of OCDD and of most PCBs with fewer than seven chlorine atoms were positively related to stroke risk. Participants in the highest 25th percentile of OCDD had 3.5 times the odds of developing stroke compared with those in the lowest 25th percentile (95% CI 1.1–11.7). Both chemicals that had dioxin-like properties and those that did not were positively associated with stroke. Total TEQs, however, were strongly associated with stroke risk. Participants in the highest 25th percentile of TEQs had

3.8 times the odds of developing stroke compared with those in the lowest 25th percentile (95% CI 1.2–12.2). Those with TEQs at or above the 90th percentile had 4.2 times the odds of developing stroke (95% CI 1.1–15.8). Stroke risk was also greater in participants that had higher concentrations of chlorine-containing pesticides.

Sergeev and Carpenter (2010) examined the rates of hospitalization for stroke with comorbid diabetes in New York state during a 12-year period. Rates were analyzed according to proximity to environmental sources of POPs. No specific exposure to the COIs was documented, and the committee did not consider the paper further.

Case-Control Studies No case-control studies of exposure to the COIs and cerebrovascular disease or stroke have been published since Update 2010.

Biologic Plausibility

Studies have demonstrated that both the vasculature and adipose tissue are targets of TCDD toxicity and provided a mechanistic understanding of how TCDD exposure increases the risk of circulatory diseases, such as hypertension, IHD, and stroke. TCDD exposure of cultured endothelial cells or cultured adipocytes induces major changes in gene expression and leads to substantial increases in oxidative stress and inflammatory markers (Andersson et al., 2011; Han et al., 2012; Ishimura et al., 2009; Kerley-Hamilton et al., 2012a; Kim MJ et al., 2012; Kopf and Walker, 2010; Majkova et al., 2009; Puga et al., 2004). Notably, loss of the AHR, as in AHR knockout mice, is associated with decreases in blood pressure (modeling hypotension), while sustained activation of the AHR resulting from dioxin exposure leads to increases in blood pressure. Agbor et al. (2011). Zhang et al. (2010) showed that the genetic loss of AHR from all tissues or solely from endothelial cells, respectively, results in hypotension. In contrast, Kopf et al. (2010) demonstrated that chronic exposure of mice to TCDD induces hypertension associated with significant increases in vascular oxidative stress and decreases in vascular relaxation. Those changes in vascular function and blood pressure could be mediated in part by increases in metabolism of arachidonic acid to vasoconstrictive and inflammatory eicosanoids (Bui et al., 2012). Studies have also demonstrated that exposure to AHR agonists, including TCDD and benzo[a]pyrene, increases the incidence, severity, and progression of atherosclerosis, a primary cause of IHD and stroke (Dalton et al., 2001; Kerley-Hamilton et al., 2012a; Wu et al., 2011). Furthermore, Wu et al. (2011) demonstrated that TCDD mediates those affects in part by increasing vascular inflammation. In addition to the vasculature, studies have suggested that the heart is a target of TCDD. TCDD exposure increases hypertrophy of rat cardiac cells in culture (Zordoky and El-Kadi, 2010) and impairs the differentiation of mouse embryonic stem cells into cardiomyocytes (Neri et al., 2011).

In addition to direct effects of TCDD on the vasculature and heart, there is evidence that TCDD influences other CVD risk factors, for example, promoting obesity (Kerley-Hamilton et al., 2012b), accumulating macrophage lipid, inducing lipid mobilization, and altering lipid metabolism. Thus, on the basis of animal models, there appear to be several overlapping and potentially contributing pathways that link TCDD exposure and increased CVD risk.

Synthesis

In this section, the committee synthesizes information on circulatory disorders from the new studies described above and reconsiders studies that were reviewed in previous updates. Because circulatory diseases constitute a broad group of diverse conditions, hypertension, IHD, and stroke are discussed separately so that the new studies can be adequately synthesized and integrated with the earlier studies.

Hypertension

Hypertension, typically defined as blood pressure above 140/90 mmHg, affects more than 70 million adult Americans and is a major risk factor for coronary heart disease, myocardial infarction, stroke, and heart and renal failure. The major quantifiable risk factors for hypertension are well established and include age, race, BMI or percentage of body fat, and diabetes; the strongest conclusions regarding a potential increase in the incidence of hypertension come from studies that have controlled for these risk factors. The committee responsible for Update 2006 concluded that the available evidence was consistent with the placement of hypertension in the limited or suggestive category. Additional evidence reviewed in Updates 2008 and 2010 reaffirmed this conclusion.

The retrospective comparisons of Korean Vietnam-era veterans with acute coronary syndrome on the basis of whether they did or did not serve in Vietnam (Kim JB et al., 2012) are not helpful in assessing whether herbicide exposure played a role in the development of hypertension. Only two relevant and informative publications have appeared since Update 2010. Chang et al. (2010) showed a strong relationship between dioxin TEQ in the blood and diastolic blood pressure. The strengths of the study are the large number of potential confounding variables addressed and the clear exposure to the COIs. Its weaknesses are that it is a cross-sectional survey that precludes strong causal inference because the temporal relationship between the exposure and the outcome is not known. Furthermore, surveys are prone to selection factors that may bias relationships between exposures and outcomes. Goncharov et al. (2011) examined a survey of residents of Anniston, Alabama, and found relationships between blood pressure and concentrations of many PCBs, including three dioxin-like mono-ortho PCBs. The study shares strengths and weaknesses with the Taiwanese survey, but

exposures to COIs and specifically TCDD were lower in the Alabama sample. The committee did not consider a recent paper by Jones et al. (2011), because the relationship between urinary arsenic and the arsenic-containing chemical that the veterans were exposed to, cacodylic acid, is unclear.

The new relevant data are consistent with a relationship between the COIs and blood pressure. The additional supporting evidence and the strong biologic rationale affirm the present committee’s placement of hypertension in the limited and suggestive category.

Ischemic Heart Disease

The committee responsible for Update 2008 revisited the entire body of evidence on TCDD exposure and heart disease and concluded that the evidence supported moving IHD to the limited and suggestive category. That conclusion was based on evidence of a dose–response relationship in the occupational cohorts, evidence of increased risk of MI in Vietnam veterans, supporting cross-sectional survey data, and a strong biologic rationale. Evidence reviewed for Update 2010 supported that classification.

Relevant epidemiologic findings published since Update 2010 are sparse. The occupational studies reviewed are from studies that have been factored into the committee’s previous recommendations. It is reassuring that the reanalysis of data on the Dutch chemical workers with improved exposure data confirmed earlier findings (Boers et al., 2012). The study of Korean Vietnam-era veterans having coronary angiography due to acute coronary syndrome (Kim JB et al., 2012) is not helpful in assessing whether herbicide exposure played a role in the development of IHD.

The consideration of carotid atherosclerosis in the PIVUS study (Lind et al., 2012) yielded important new information. The study was cross-sectional; in such studies, there is a concern that a subject’s awareness of his disease or the disease process itself may have led to changes in behavior, treatment effects, or the pathologic process itself that modified the measure of exposure gathered at the time of the study. Thus, the exposure measured may not reflect the exposure status when the pathologic process under study started. That concern is reduced in ultrasonographic studies because carotid IMT and plaque are symptomless and not routinely assessed in clinical practice, so it is unlikely that greater IMT would cause changes that would alter serum PCB. In addition, carotid ultrasonography has been used in other cross-sectional studies of the atherosclerotic process, and risk factors identified in cross-sectional studies have been confirmed in longitudinal studies. The investigators were able to control statistically for many potential confounding factors, and this increases confidence in their results. Exposure quantified in terms of dioxin TEQs was found to have a statistically-significant relationship with increased IMT and with decreased echogenicity, both of which are independent risk factors for clinical cardiovascular disease. It is not entirely