5
Epidemiologic Studies:
Compendium of New Publications
The continuing effort to evaluate and integrate epidemiologic studies pertinent to the possible health effects of 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), 4-amino-3,5,6-trichloropicolinic acid (picloram), and dimethyl arsenic acid (DMA or cacodylic acid)—has involved the review of thousands of publications over successive reports (the original retrospective report, nine updates prior to the current report, and three short reports on single issues, as delineated in Chapter 1). The search strategy used to identify these publications is described in Chapter 2, along with explanations of the various refinements that have been employed since the initial volume in this series was prepared.
This chapter tabulates publications of primary epidemiologic research that appeared in the period from October 1, 2012 (the closing date for inclusion in Update 2012 [IOM, 2014]), through September 30, 2014, as a compendium of the new information on human health outcomes considered by the present committee. In this chapter and later chapters, epidemiologic studies are organized into categories according to the populations being studied (Vietnam veterans, occupational populations other than Vietnam veterans, and nonoccupational populations affected by environmental exposures) or by study design (case-control). The various study designs (the most relevant being cohort, case-control, and cross-sectional) have strengths and weaknesses that influence their potential to contribute evidence considered in the health-outcomes chapters.
Design information on populations that are the subject of multiple references in this and earlier Veterans and Agent Orange (VAO) reviews—including new studies of populations that have been studied previously and studies of new
populations that had multiple health outcomes—is provided in the next chapter “Epidemiology Studies: Background on Multiply Referenced Populations,” along with committee commentary. This integrative approach has been taken to avoid repeating design information in multiple health-outcomes chapters and to make evident to the reader the extensive degree of interrelationship among many of the published analyses that have been reviewed in the course of the VAO series. (Design information on the studies of new populations that involve single health outcomes is provided in the various health-outcomes chapters.)
In addition to reviewing studies involving exposures to the specific COIs listed previously, this and earlier VAO committees have considered studies that examined compounds chemically related to the herbicides used in Vietnam, such as 2-(2-methyl-4-chlorophenoxy) propionic acid, hexachlorophene, and chlorophenols, particularly 2,4,5-trichlorophenol. Some publications did not indicate the specific herbicides or polychlorinated biphenyls (PCBs) with dioxin-like toxic actions to which study participants were exposed or the magnitude of exposure; those limitations were considered when the committee weighed the relevance of each publication, as detailed in Chapter 2. The committee considers studies of exposure to PCBs and other dioxin-like compounds (DLCs) informative if their results were reported in terms of TCDD toxic equivalents (TEQs) or concentrations of specific congeners of DLCs. The available details of the exposure assessment and the use of the resulting data in analyses are discussed in Chapter 3, which follows the same sequence to categorize the study populations.
NEW EPIDEMIOLOGIC PUBLICATIONS
The new epidemiologic publications reviewed by the committee for this update are listed in Tables 5-1, 5-2, and 5-3. The conditions listed in the “Health Outcomes Reported” columns are indicative of the chapters in which the new publications are considered. Note, however, that studies assessing the occurrence of various cancers after exposure scenarios that are temporally comparable with exposure during military service are discussed in Chapter 8, which addresses cancer outcomes as applicable to the veterans themselves. Studies of childhood cancers in relation to parental exposure to the COIs are discussed in Chapter 10, which addresses possible adverse effects in veterans’ offspring. Cancer studies that consider only childhood exposure are not considered relevant to the committee’s charge.
Publications Reporting a Single Health Outcome in New Populations
The new publications reporting a single health outcome in populations not studied previously are listed in Table 5-1 with an indication of the outcomes. Descriptions and critiques of the designs of the studies are provided in the sections of the report that discuss the results related to particular health outcomes. The
TABLE 5-1 Publications Reporting a Single Health Outcome in New Populations
| Author | Study Design | Exposure Measure(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Studies of Vietnam Veterans | ||||
| Ansbaugh et al., 2013 | Cohort | “Agent Orange” as per US Department of Veterans Affairs designation | Prostate cancer | 2,720 veterans referred to the Portland Veterans Affairs Medical Center |
| Li Q et al., 2013 | Cohort | Dioxin-TEQ levels measured in abdominal subcutaneous fat | Prostate cancer, biochemical recurrence after radical prostatectomy | 93 Vietnam veterans who underwent radical prostatectomy, median of 5.3 yrs of post-operative follow-up |
| Environmental Studies | ||||
| Delvaux et al., 2014 | Cohort | EDCs, including dioxin (TEQs) and non-dl PCBs in cord blood | Prenatal exposure to EDCs and body composition at 7–9 years of age | Flemish children; part of the Flemish Environment and Health Study |
| Ferguson et al., 2012 | Cohort | POPs in serum (including dl-PCBs 77, 105, 118, 156, 170, 180) | Reproductive hormones | Male partners (aged 18–51), in subfertile couples seeking infertility evaluation and treatment at Massachusetts General Hospital (01/2000–05/2003) |
| Gauthier et al., 2014 | Cross-sectional | OCDD and dl-PCBs (including PCBs 105, 118, 156, 157, 189) | Fasting plasma levels (pg/ml) from nondiabetic, obese, post-menopausal women | “Metabolically healthy” vs 40 “metabolically abnormal” women from Montreal, categorized on the basis of insulin sensitivity |
| Hansen et al., 2014 | Cohort | PCB congeners (including dl-PCBs 118, 156, 170, 180) | Asthma in offspring to mothers exposed to POPs | 965 women; 20-year follow-up to the Danish Fetal Origins 1988–1989 Cohort in Aarhus, Denmark |
| Author | Study Design | Exposure Measure(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Kim et al., 2013 | Cross-sectional | PCB congeners | Free t3, total T3, free T4, total T4, TSH | 138 pregnant women from 5 Korean hospitals |
| Medehouenou et al., 2014 | Cohort | OC pesticides and PCBs measured in plasma, including dl-PCBs 105, 118, and 156 | Dementia | Canadian Study of Health and Aging, a national cohort study of Canadians 65+ years of age |
| Nakamoto et al., 2013 | Cross-sectional | PCDD/Fs, dl-PCBs, and total dioxins in blood | History of disease, including asthma, atopic dermatitis, allergic rhinitis, hypertension, hyperlipidemia, diabetes, gout, thyroid and kidney disease, gastric ulcer | Japanese men (1,063) and women (1,021), aged 15–76 yrs, from general population |
| Sioen et al., 2013 | Cohort | dl-compounds (including total of PCDD/Fs and dl-PCBs | Prenatal exposure and behavior problems at 7–8 years of age | Flemish Mother–New-Born Cohort) |
| Spector et al., 2014 | Cross-sectional and longitudinal associations | PCBs (dl-PCBs 105, 118, 156 combined) | Immune function in postmenopausal women | 109 postmenopausal overweight women enrolled in the Physical Activity for Total Health study, 1998–2000 |
| Tai et al., 2013 | Cohort | Dioxin in breast milk | Neurodevelopment in infants birth to 4 months old | 216 mother-infant pairs living near the Da Nang airport in Vietnam |
| Valera et al., 2013a,b | Cohort | dl-PCB 105 | Hypertensive status | Inuit adults from Quebec and Greenland |
| Wohlfahrt-Veje et al., 2014 | Cohort | PCDDs, furans, and biphenyls in breast milk | Exposures to dl chemicals in breast milk and early growth and serum IGF1 | Copenhagen Mother Child Cohort of Growth and Reproduction; Danish children (born 1997–2001) |
| Author | Study Design | Exposure Measure(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Winneke et al., 2014 | Cohort | TEQs from maternal blood during gestation and milk within 3 weeks of birth | Behavioural sexual dimorphism in school-age children | Duisburg Cohort; 232 pregnant women 09/2000–10/2002 in Duisburg, Germany |
| Case-Control Studies | ||||
| Cocco et al., 2012 | Case-control | Pesticides (including 2,4-D, phenoxys, chlorophenols, organochlorines) | Lymphoma | Participants in the EPILYMPH case-control study in six European countries (1998–2003) |
| Glass et al., 2012 | Case-control | Phenoxy herbicides | ALL | Parental occupational exposures and ALL in Australia |
| Metayer et al., 2013 | Case-control | Pesticides, including 2,4-D | ALL | Northern California Childhood Leukemia Study |
NOTE: 2,4-D, 2,4-dichlorophenoxyacetic acid; ALL, acute lymphoblastic lymphoma; dl, dioxin-like; EDC, endocrine-disrupting chemical; OC, organochloride; OCDD, octachlorodibenzo-p-dioxin; PCB, polychlorinated biphenyl; PCDD, polychlorinated dibenzo-p-dioxin; PCDF, polychlorinated dibenzofuran; POP, persistent organic pollutant; TEQ, (total) toxic equivalent; TSH, thyroid-stimulating hormone.
publications in this table include a mix of study designs and focus principally on individual types of cancer as the primary health outcome of interest.
Publications Reporting Multiple Health Outcomes in New Populations
The new publications reporting multiple health outcomes in populations not studied previously are listed in Table 5-2 with a list of outcomes that were investigated. Comprehensive discussions of the designs of the studies are presented in Chapter 6, organized according to the type of study population. For Update 2014, six publications were identified from a new, exceptionally large epidemiological study of more than 114,000 Korean Vietnam War veterans. This study cohort is much larger in scope than all of the other published epidemiological studies conducted among Vietnam veterans. It provides results for a very large set of health outcomes, including rare conditions, as well as information on both non-fatal outcomes and cause-specific mortality. The results for new publications reporting multiple health outcomes in populations not studied previously, with
TABLE 5-2 Publications on Multiple Health Outcomes in New Study Populations
| Author | Study Design | Exposure Measures(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Studies of Vietnam Veterans | ||||
| McBride et al., 2013 | Cohort | Service in Vietnam during Vietnam War | Mortality and cancer experience (1988–2008) | New Zealand Vietnam war veterans |
| Yi, 2013 | Cohort | Service in Vietnam during the Vietnam War | Cancer incidence 1992–2003; all cancers and full spectrum individually | KVHS–Korean Vietnam veterans identified using the Korea National Cancer Incidence Database |
| Yi and Ohrr, 2014 | Cohort | AO exposure using GIS-based model | Cancer incidence 1992–2003; all cancers and full spectrum individually | KVHS–Korean Vietnam veterans identified using the Korea National Cancer Incidence Database |
| Yi et al., 2014a | Cohort | AO exposure using GIS-based model | Disease prevelance, full spectrum individually | KVHS–Korean Vietnam veterans identified using Korean National Health Insurance Health claim data, January 2000–September 2005 |
| Yi et al., 2014b | Cohort | AO exposure using GIS-based model | Morbidity and mortality from individual cancers and various diseases | KVHS–Korean Vietnam veterans with cause of death reported 1992–2005 |
| Occupational Studies | ||||
| Wang et al., 2013 | Cohort | PCDDs/PCDFs | Lung, liver, and stomach cancer, individually | Workers from an automobile foundry factory in Hubei province in China |
| Environmental Studies | ||||
| Papadopoulou et al., 2013a | Cohort | Dietary intake of dioxins and dl-compounds | Maternal diet and birth size | Mothers enrolled in Norwegian Mother and Child Cohort Study |
| Author | Study Design | Exposure Measures(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Papadopoulou et al., 2013b | Cohort | dl activity in maternal blood samples at time of delivery and anogenital distance in newborns and infants | Anogenital distance in newborns and children | Mothers and newborns enrolled in European NewGeneris Cohort |
| Papadopoulou et al., 2014 | Cohort | dl activity in maternal blood samples at time of delivery | Association between maternal diet and birth outcome | Mothers enrolled in European NewGeneris Cohort |
| Vafeiadi et al., 2013 | Cohort | In utero exposure to dioxin and dl-compounds | In utero exposure to dioxin and dl-compounds and anogenital distance | Subset of mother–Child cohorts from NewGeneris Cohort |
| Vafeiadi et al., 2014 | Cohort | dl activity in cord blood and maternal blood samples at time of delivery | dl activity in blood and birth weight, gestational age, and head circumference, | Mothers and newborns enrolled in European NewGeneris Cohort |
NOTE: AO, Agent Orange; dl, dioxin-like; GIS, geographic information system; KVHS, Korean Veterans Health Study; PCDD, polychlorinated dibenzo-p-dioxin; PCDF, polychlorinated dibenzofuran.
comments related to their reliability or limitations, appear in the appropriate outcome-specific sections of Chapters 7–13.
New Publications on Previously Studied Populations
The new publications on previously studied populations are listed in Table 5-3. The new publications are reviewed in the context of the history of publications on the same populations to take into account the fact that they are not presenting entirely new evidence, but rather enhancing a picture that has been emerging for many years.
A number of long-term studies of populations exposed to the COIs are of particular importance to the VAO project. The disease experiences of those populations are updated with the passage of time. Placing each new publication into its historical context helps the committee combine the evidence from various publications appropriately and take into consideration the interdependence of related publications. Such clusters of studies are useful in describing the course of a population’s response to an exposure, and joint consideration of an entire body
TABLE 5-3 Publications on Previously Studied Populations
| Author | Study Design | Exposure Measure(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Studies of Vietnam Veterans | ||||
| Kang et al., 2014 | Retrospective cohort | US military service in Vietnam or near Vietnam | Mortality 1965–2010; all cancers and specific (brain, breast, cervical, ovarian, pancreatic, respiratory, uterine), diabetes mellitus, heart disease, circulatory disease, respiratory disease, and nervous system disease | US women who served in Vietnam or near-Vietnam vs non-Vietnam veteran peers in the US |
| ADVA, 2014a,b | Cohort | Vietnam veteran families | Study overview, pregnancy and birth defect outcomes | Vietnam veteran sons and daughters |
| ADVA, 2014c | Cohort | Vietnam veteran families | Mortality patterns | Vietnam veteran families |
| Occupational Studies | ||||
| Goldner et al., 2013 | Cohort | 50 specific herbicides, including 2,4-D, 2,4,5-T, and 2,4,5-TP | Association between thyroid disease and use of insecticides, herbicides, and fumigants/fungicides | AHS (male private pesticide applicators) |
| Rinsky et al., 2013 | Cohort | Specific pesticides, inluding 2,4-D | Stroke mortality | AHS (male private pesticide applicators) |
| Saberi Hosnijeh et al., 2013a (same group as Saberi Hosnijeh et al., 2012a) | Cohort | TCDD | Serum metabolomics pertubations | Subcohort of IARC (Dutch phenoxy herbicide workers) |
| Saberi Hosnijeh et al., 2012b | Cohort | TCDD | Changes in lymphocyte subsets | Subcohort of IARC (Dutch phenoxy herbicide workers) |
| Starling et al., 2014 | Cohort | Pesticides (including 2,4,5-T) | Diabetes | AHS (wives of farmers) |
| Author | Study Design | Exposure Measure(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Environmental Studies | ||||
| Bouchard et al., 2014 Chevrier et al., 2014 | Cross-sectional Cohort | PCB congeners Serum TCDD concentrations | Cognitive function Thyroid hormone levels | NHANES (1999–2002) SWHS (Seveso women 0–40 yrs old at time of accident; follow-up April 2008–December 2009) |
| Eskenazi et al., 2014 | Cohort | Serum TCDD concentrations | Bone density and structure | SWHS (Seveso women 0–40 yrs old at time of accident; follow-up 2008) |
| Everett and Thompson, 2014 | Cross-sectional | Dioxins and dl-PCBs (including dl-PCBs 81, 105, 118, 126, 156, 157, 167) in blood samples | Diabetes nephropathy | NHANES (1999–2004) |
| Gallagher et al., 2013 | Cross-sectional | dl-PCBs | Serum antinuclear antibodies | NHANES (2003–2004) |
| Krieg, 2013 | Cross-sectional | Pesticide metabolites in urine after 2,4-D exposure | Cognitive function | NHANES III |
| Lin et al., 2012 | Cross-sectional | PCDDs, PCDFs, dl-PCBs (dl-PCBs 81, 105, 118, 126, 156, 157, 167, 169, 189) in serum | Cause-specific mortality through 2006 for all-causes, all cancers, and CVD | NHANES (1999–2004) |
| Lind et al., 2013 | Cross-sectional | POPs (including dl-PCBs 105, 118, 126, 156, 157, 169, 189) | Life-time weight change | PIVUS seniors (2001–2004) |
| Pahwa P et al., 2012a | Cohort | Herbicides | Chronic bronchitus | Saskatchewan Rural Health Study |
| Peters et al., 2014 | Cross-sectional | dl-PCBs (dl-PCBs 81, 118, 126, 189) | Blood pressure | NHANES (1999–2008) |
| Author | Study Design | Exposure Measure(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| ten Tusscher et al., 2014 (same cohort as Patandin et al., 1998) | Cohort | Pre-, peri-, and postnatal exposure to dioxin | Neurodevelopmental retardation | Amsterdam–Zaandam cohort; children born 1987–1991 |
| Turunen et al., 2012 Warner et al., 2013 | Cohort Cohort | PCDDs, PCDFs, PCBs Serum TCDD concentrations | C-reactive protein (an indicator of inflammation) Diabetes, metabolic syndrome, obesity | Finnish fisherman and their wives SWHS (Seveso women 0–40 yrs old at time of accident; follow-up April 2008–December 2009) |
| Wesselink et al., 2014 | Cohort | Serum TCDD concentrations | Pregnancy outcomes | SWHS (Seveso women 0–40 yrs old at time of accident; follow-up 2008–2009) |
| Case-Control Studies | ||||
| Carmichael et al., 2013 | Case-control | Pesticides (including 2,4-D, MCPA, | Hypospadias | NBDPS |
| Carmichael et al., 2014 | Case-control | cacodylic acid) Pesticides (including 2,4-D) | Selected congenital birth defects | NBDPS |
| Kachuri et al., 2013 | Case-control | Pesticides (including 2,4-D) | Multiple myeloma | CCSPH |
| Navaranjan et al., 2013 | Case-control | Herbicides, phenoxy herbicides | Hodgkin lymphoma | CCSPH |
| Pahwa P et al., 2012b | Case-control | Pesticides (including phenoxys, 2,4-D, MCPA) | Multiple myeloma | CCSPH |
| Shaw et al., 2014 | Case-control | 2,4-D | Early pregnancy and risk of gastroschisis | NBDPS |
| Author | Study Design | Exposure Measure(s) Having Results | Health Outcome(s) Reported | Study Population |
|---|---|---|---|---|
| Yang et al., 2014 | Case-control | 2,4-D | Neural tube defects and orofacial clefts | NBDPS |
NOTE: 2,4-D, 2,4-dichlorophenoxyacetic acid; 2,4,5-T, 2,4,5-trichlorophenoxyacetic acid; 2,4,5-TP, 2-(2,4,5-trichlorophenoxy) propionic acid; AHS, Agricultural Health Study; CCSPH, Cross-Canada Study of Pesticides and Health; CVD, cardiovascular disease; dl, dioxin-like; IARC, International Agency for Research on Cancer; MCPA, 2-methyl-4-chlorophenoxyacetic acid; NBDPS, National Birth Defects Prevention Study; NHANES, National Health and Nutrition Examination Survey; PCB, polychlorinated biphenyl; PCDD, polychlorinated dibenzo-p-dioxin; PCDF, polychlorinated dibenzofuran; PIVUS, Prospective Study of the Vasculature in Uppsala Seniors; POP, persistent organic pollutant; SWHS, Seveso Women’s Health Study; TCDD, 2,3,7,8-tetrachlorodibenzo-p-dioxin.
of research on a population may yield insights into relationships with potential confounding factors.
Many groups potentially exposed to the COIs have been monitored periodically, including the cohorts of the International Agency for Research on Cancer (IARC) and the National Institute for Occupational Safety and Health (NIOSH); residents of Seveso; and Ranch Hand and Army Chemical Corps personnel. For the sake of completeness, the discussions of specific health outcomes and the associated cumulative-results tables in Chapters 7–13 include references to publications discussed in previous VAO reports and to new publications. In drawing its conclusions, the committee combined the evidence in new publications and the evidence synthesized in the most recent update (Update 2012), taking into account the interdependence of related publications. For the present update, several relevant studies of dioxin-like compounds and a range of health outcomes were identified from serum samples collected from the federally funded National Health and Nutrition Examination Survey (NHANES).
Individual researchers who belong to research consortia that are evaluating cohorts in large multicenter studies (such as the IARC and NIOSH cohort studies) sometimes publish reports based on the subsets of study participants that they themselves are monitoring. The VAO committees consider all reports that have been published, including those based on entire cohorts and those based on subcohorts. In drawing its conclusions, the committee factored in both types of studies, taking into consideration the interdependence among related studies. In particular, some subcohort studies have access to information not available for the entire cohort, such as data on individual serum TCDD concentrations and personal information that can be used to adjust for confounders of concern. Furthermore, even when the analyses based on an entire cohort would include data on a subcohort as a subset, the reports on the subcohort might provide additional information on the consistency of the relationships among subcohorts, such as
whether there are important subcohort-by-exposure interaction effects, when these issues were not considered in the full-cohort studies. As long as the structures of the study populations are recognized, VAO committees have been less concerned about over-weighting unstable positive findings on small subgroups or giving “repeated consideration” to duplicative results than would be the case if a quantitative meta-analysis were being undertaken.
Many of the cohorts that have contributed to the cumulative findings of the VAO committees are no longer being followed; however, the cohorts’ histories are briefly recapitulated in the body of this report. Additional background information can be found in earlier reports in this series. The subjects of the new epidemiological studies identified include female US Vietnam veterans as well as Australian, Korean, and New Zealand veterans who served in Vietnam. These studies are augmented with a wealth of new data from civilian populations exposed to the COIs along with herbicides and pesticides with mechanistic and toxic properties similar to the COIs.
