Based on new evidence and a review of prior studies, the current committee did not find any new associations between the relevant exposures and neurologic disorders. Current evidence supports the findings of earlier updates that:
- There is limited or suggestive evidence of an association between the chemicals of interest (COIs) and Parkinson disease and diseases that present with Parkinson-like symptoms.
- There is inadequate or insufficient evidence to determine whether there is an association between the COIs and any of the other adverse neurologic outcomes.
This chapter considers the possible effects of toxic exposure to the herbicides used during the Vietnam War and specific clinical conditions associated with the central nervous system (CNS) and the peripheral nervous system (PNS), primarily brain dysfunctions. The COIs are 2,4-dichlorophenoxyacetic acid (2,4-D), 2,4,5-trichlorophenoxyacetic acid (2,4,5-T), picloram (4-amino-3,5,6-trichloropicolinic acid), cacodylic acid (dimethyl arsenic acid [DMA]), and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), a contaminant of 2,4,5-T. As described in Chapter 3, studies of the effects of exposure to polychlorinated biphenyls (PCBs) and other dioxin-like chemicals were also considered informative if their results were reported in terms of TCDD toxic equivalents (TEQs) or concentrations of specific congeners of dioxin-like chemicals. Studies that report TEQs
based only on mono-ortho PCBs (which are PCBs 105, 114, 118, 123, 156, 157, 167, and 189) are considered even though their TEQs are several orders of magnitude lower than those of the non-ortho PCBs (77, 81, 126, and 169), based on the revised World Health Organization (WHO) toxicity equivalency factor (TEF) scheme of 2005 (La Rocca et al., 2008; van den Berg et al., 2006). The lower TEQs of the mono-ortho PCBs, however, may be counterbalanced by their abundance, which is generally many orders of magnitude higher than the non-ortho PCBs (H.-Y. Park et al., 2010).
Examples of the diseases that result from the degeneration of specific brain areas are Parkinson disease (PD), Alzheimer disease (AD), spinocerebellar degeneration, and amyotrophic lateral sclerosis (ALS). These diseases may occur in the absence of any toxicant exposure, but all may be triggered by environmental factors, including toxicant exposure (Bronstein et al., 2009; Chin-Chan et al., 2015; de la Monte and Ming, 2014; H. Kang et al., 2014b; Tanner et al., 2014; M. D. Wang et al., 2014).
Disorders of the PNS are generally referred to as neuropathies. Neuropathies can be purely motor, presenting as deficits in strength, but most often they present with the involvement of both motor and sensory fibers. Neuropathies are often symmetric and start with symptoms related to dysfunction of fibers that travel the greatest distance to their target organ. For that reason, the symptoms of neuropathy often start in the digits and travel toward the torso. Many neuropathies also affect autonomic fibers and thus can result in changes in blood pressure and heart rate and in symptoms related to the control of digestion. Toxicant exposure can induce immediate (i.e., acute) damage to peripheral nerves, and previous updates found limited or suggestive evidence that dioxin exposure can cause such short-term effects. However, the overall focus of this chapter is on delayed adverse effects on both the PNS and the CNS.
The immediate effects of toxicants may involve all regions of the nervous system, whereas delayed effects are likely to be related to focal deficits. Diffuse damage to the CNS may cause alterations in thinking, consciousness, or attention, sometimes in combination with abnormalities of movement, while focal dysfunction can cause myriad syndromes, depending on which area of the brain is involved and the extent and severity of damage. For the purposes of this review, neurologic deficits associated with Vietnam service are distinguished from psychiatric and psychologic conditions—such as posttraumatic stress disorder (PTSD), depression, and anxiety—and from chronic fatigue syndrome.
In the original Veterans and Agent Orange (VAO) report (IOM, 1994), attention was focused on persistent neurobehavioral disorders. That focus was maintained through Update 2002 (IOM, 2003c). A slight change in emphasis toward chronic neurodegenerative disorders was reflected in the name change of this chapter to “Neurologic Disorders” in Update 2004 (IOM, 2005), which was carried forward in all subsequent updates. In this update, the chapter covers data pertinent to persistent neurologic disorders of all types.
Many studies have addressed the possible contribution of various chemical exposures to neurologic disorders. However, for the purposes of the VAO committees, several of these studies have been limited by the use of nonspecific or mixed exposures and by a lack of adequate lag times between the exposure and the neurologic outcome of interest. Case identification of neurologic disorders is also an important consideration and is often difficult because there are few disorders for which there are specific diagnostic tests. Because the nervous system is not readily accessible for biopsy, pathologic confirmation is usually not feasible. However, identifiable neurologic disorders always result in objective abnormalities that are reflected in anatomic or functional tests or discovered via clinical examination.
This chapter reviews the association between exposure to the COIs and neurobehavioral disorders, neurodegenerative disorders, and chronic PNS disorders. The scientific evidence supporting the biologic plausibility of each category of disorders is also reviewed here. More complete discussions of the categories of association and of this committee’s approach to categorizing health outcomes are presented in Chapter 3. For citations new to this update that revisit previously studied populations, the relevant details on the experimental design can be found in Chapter 5.
Experimental data regarding the biologic plausibility of a connection between exposure to the COIs and various neurologic disorders continue to accrue. This section summarizes in a general way some of the information reviewed in the current update and, for completeness, includes pertinent information from prior updates.
Several studies have dealt with mechanisms of neurotoxicity that might be ascribed to the COIs, notably 2,4-D and TCDD. The molecular effects of the COIs are described in detail in Chapter 4. Some aspects of the biochemical activity of the COIs suggest pathways by which there could be effects on the neural systems. A number of studies suggest that the COIs, primarily 2,4-D, have neurologic effects, both neurochemical and behavioral, in animal models if exposure occurs during development or in cultured nerve cells (Konjuh et al., 2008; Pasandi et al., 2017; Rosso et al., 2000a,b; Stürtz et al., 2008); older references described the behavioral effects of a developmental exposure of rodents to a 2,4-D–2,4,5-T mixture (Mohammad and St. Omer, 1986; St. Omer and Mohammad, 1987). The exposure of zebrafish during development to the dioxin-like chemical PCB 126 leads to behavioral deficits in adults and substantial gene expression changes in the adult brain (Aluru et al., 2017). Juvenile TCDD exposure in zebrafish has been shown to lead to lesions in the olfactory neuroepithelium and to dramatic changes in the expression of genes important for neurological function (Q. Liu et al., 2014). Perinatal exposures to TCDD and to coplanar, dioxin-like PCBs
have reportedly caused deficits in learning behavior in rats (Curran et al., 2011; Haijima et al., 2010; Hojo et al., 2008; Kakeyama et al. 2014). However, those studies should be interpreted with caution because the developing nervous system is different from the mature nervous system and may not be an appropriate model for the possible consequences of exposure to the COIs by adults, as was the case for Vietnam veterans.
Some studies further support suggestions that the concentration of reactive oxygen species could alter the functions of specific signaling cascades and be involved in neurodegeneration (Drechsel and Patel, 2008). Such studies do not specifically concern the COIs but they are potentially relevant to these chemicals inasmuch as TCDD and herbicides have been reported to elicit oxidative stress (Byers et al., 2006; Celik et al., 2006; J. Kumar et al., 2014c; D. Shen et al., 2005; Wan et al., 2014). TCDD has been shown to affect phosphokinase C biochemistry in nerve cells and so could affect the integrity and physiology of nerve cells (S. Y. Kim et al., 2007; H. G. Lee et al., 2007). TCDD has also been shown to affect signaling pathways that regulate nitric oxide synthesis in neural and glial cells, leading to neurotoxicity, senescence, and cell death (Duan et al., 2014; Jiang et al., 2014; Y. Li et al., 2013; Nie et al., 2015; Wan et al., 2014). Pellacani et al. (2014) found that PCB 126 reduced neuroblastoma cell viability. Zhao et al. (2016) exposed C6 glioma cell culture to TCDD and found dose- and time-dependent down-regulation of glutamate transporter-1 expression, which could contribute to neurotoxicity. Cytochrome P450 1A1, the aryl hydrocarbon receptor (AHR), and the AHR nuclear transporter occur in the brain, so TCDD may exert effects in the brain (P. Huang et al., 2000). In addition, earlier studies in hepatocytes indicated that 2,4-D affects aspects of mitochondrial energetics and mitochondrial calcium flux (Palmeira et al., 1994a,b, 1995a,b); if these effects occur in mitochondria of nervous-system cells, the energy balance and energy pathways of cells in the nervous system could be affected and disrupt nervous system function. This is supported by the work of Morales-Hernández et al. (2012), which showed that TCDD in cultured neuronal cells induces cell death by the disruption of intracellular calcium levels.
Laboratory-based studies have emphasized the importance of alterations in neurotransmitter systems as potential mechanisms underlying TCDD-induced neurobehavioral disorders (Jiang et al., 2014; H. Q. Xie et al., 2013). Neuronal cultures treated with 2,4-D exhibited decreased neurite extension associated with intracellular changes, including a decrease in microtubules, the inhibition of the polymerization of tubulin, disorganization of the Golgi apparatus, and the inhibition of ganglioside synthesis (Rosso et al., 2000a,b). Those mechanisms are important for maintaining the connections among nerve cells, which are necessary for neuronal function and are involved in axon regeneration and recovery from peripheral neuropathy. Early animal experiments have demonstrated that TCDD treatments affect the fundamental molecular events that underlie the neurotransmission initiated by calcium uptake (Hong et al., 1998). Mechanistic studies
have demonstrated that 2,4,5-T can alter cellular metabolism and the cholinergic transmission necessary for neuromuscular transmission (Sastry et al., 1997).
TCDD treatment of rats at doses that do not cause general systemic illness or wasting produces electric changes in peripheral nerves that are associated with altered functions and pathologic findings that are characteristic of toxicant-induced axonal peripheral neuropathy (Grahmann et al., 1993; Grehl et al., 1993). In cultured cells, Jung et al. (2009) explored the mechanism by which TCDD inhibits neurite outgrowth and found it to be associated with a reduction in glutaminase.
As discussed in Chapter 4, extrapolating observations of cells in culture or in animal models to humans is complicated by differences in sensitivity and susceptibility among animals, strains, and species; by the lack of strong evidence of organ-specific effects occurring consistently across species; and by differences in the route, dose, duration, and timing of chemical exposures. Thus, although the toxicologic observations themselves cannot establish a conclusion that the COIs produced neurotoxic effects in humans, they establish biologic plausibility and point to potential mechanisms that might have come into play.
The literature search for this update identified publications on populations with relevant exposures that examined overall mortality from any nervous system disorder (Collins et al., 2016) or hospitalization, again with all nervous system disorders combined, except epilepsy which was presented separately (Cox et al., 2015). However, while the results of grouping all nervous system disorders together are presented, they are not considered informative for assessing whether specific nervous system disorders may be due to an exposure to the COIs. Therefore, such results were not considered by the committee when weighing the evidence for specific conclusions.
Cox et al. (2015) used hospital discharge records from 1988 to 2009 to report the prevalent health conditions among a cohort of 2,783 male New Zealand Vietnam veterans, who served during 1964 to 1972 and were presumed to be exposed to dioxin. For participants 30 years of age or older, person-years of follow-up were calculated by 5-year age categories. Age-specific hospitalization rates were calculated using the total number of annual hospitalizations published by the Ministry of Health and the average annual resident population. Standardized hospitalization rates were calculated for both the veteran cohort and the general population, and from these estimates a standardized hospitalization ratio (SHR) was calculated with 99% confidence limits as a means to address multiple tests performed for various outcomes. Results were presented for eight categories of mental and neurologic disorders and a ninth category for “other nervous disorders.” Statistically significant increases in the hospitalization rates were observed for alcohol-related disorders (n = 89; SHR = 1.91, 99% confidence interval [CI] 1.39–2.43), which the authors note was often associated with PTSD, and other nervous disorders
(n = 135; SHR = 1.32, 99% CI 1.02–1.61). “Other nervous disorder” was the largest single category among mental and neurologic disorders, but it was unclear which disorders were included in that category with the exception of those disorders listed under the general category of mental and neurological disorders; by a process of elimination, this category also likely included PD and ALS. However, hospitalization rates are not a good measure of PD and several other neurologic disorders because people with those conditions are not necessarily hospitalized for the symptoms. No difference in the hospitalization rates was seen for the category of “senility or organic mental illness” (n = 28; SHR = 0.95, 99% CI = 0.49–1.41); the authors do not list specific neurologic disorders of interest, and senility may have included both vascular and non-vascular or Alzheimer-like dementia. Overall, the results of this study for hospital admissions due to all causes combined showed a small increase in rates for Vietnam veterans compared with the population of New Zealand. However, exposure to the COIs was not validated through serum measurements, and the study did not control for smoking or ethnicity or other potentially important risk factors.
Collins et al. (2016) provides additional follow-up time to a retrospective analysis of a cohort of 2,192 workers (only 5 of whom were female) exposed to dioxins during trichlorophenol (TCP) and pentachlorophenol (PCP) production at a chemical manufacturing plant in Michigan. The U.S. population was used as the comparator to generate standardized mortality ratios. Work history records were used to determine the length of exposure. Serum samples to measure the levels of six types of dioxins were collected for 431 workers who had been exposed to TCP or PCP. Historic concentrations for each dioxin congener were calculated from the median concentrations of serum samples and the known half-lives associated with each congener. A job-exposure matrix was created for both the TCP and the PCP production facilities based on measured concentrations for workers in different jobs. A pharmacokinetic model was applied to job-specific concentrations and with the work history of each member of the study group to estimate each worker’s time-dependent serum concentration profiles for each dioxin congener (i.e., TCDD as well as Hexa-CDD, Hepta-CDD, and Octa-CDD). Complete vital status follow-up through December 2011 was achieved for the cohort, and there were 1,198 decedents through the entire study period (1979–2011); 1,615 deaths were among TCP workers and 773 deaths occurred among PCP workers (196 workers were exposed to both TCP and PCP and are counted in each of those groups). Compared with the U.S. population, no difference in mortality from all diseases of the nervous system was found for all workers (n = 21; SMR = 0.74, 95% CI 0.46–1.13), and the SMRs were identical for TCP workers (n = 17; SMR = 0.79, 95% CI 0.46–1.26) and PCP workers (n = 7; SMR = 0.79, 95% CI 0.32–1.63).
A second occupational study reported on pesticide applicators from Argentina who completed a questionnaire to document self-reported use of pesticides and general symptoms or consultations (signs of irritation, fatigue/tiredness,
headache, nervousness or depression, medical consultation, and hospitalization) (Butinof et al., 2015). Included among the pesticides were the herbicides 2,4-D, dicamba, and picloram—three of the COIs for Vietnam veterans. The percentage of completed questionnaires indicating the use of 2,4-D was 93.5%, with lower percentages for dicamba (69.4%) and picloram (46.0%). However, the survey respondents reported the use of 11 different pesticides on average, and effect estimates were based on grouped exposures or did not report any of the COIs separately, making this analysis of little utility for the purposes of the committee.
This section summarizes the findings of VAO and previous updates on neurobehavioral disorders and incorporates information published since Update 2014 into the evidence database.
Conclusions from VAO and Previous Updates
On the basis of the data available at the time, the committees responsible for VAO and all subsequent updates through Update 2014 concluded that there was inadequate or insufficient evidence to determine whether there is an association between exposure to the COIs and neurobehavioral disorders. The data that informed that conclusion were mostly from the Air Force Health Study (AFHS, 1991a,b, 1995, 2000; Barrett et al., 2001, 2003). Urban et al. (2007) confirmed that acute neurologic symptoms experienced shortly after an acute exposure to TCDD could be sustained more than 30 years after the exposure, but this study did not address delayed effects. In general, many of the studies reviewed by previous VAO Updates were found to be methodologically flawed (Dahlgren et al., 2003; Pazderova-Vejlupkova et al., 1981; Pelclová et al., 2001, 2002) or uninformative (ADVA, 2005c; Decoufle et al., 1992; Kamel et al., 2007a; R. M. Park et al., 2005; Sapbamrer and Nata, 2014; Solomon et al., 2007; Visintainer et al., 1995). Difficulties in case identification and diagnosis, misclassification of exposures because of a lack of quantitative measures, subject ascertainment and selection bias, and uncontrolled confounding from many comorbid conditions are common weaknesses in the studies reviewed. The variability of the test results over time, the weak and inconsistent associations, and a lack of consistent dose–response relationships, also prevent those studies from supporting an association between the exposures of interest and neurobehavioral disorders.
More recent analyses using National Health and Nutrition Examination Survey (NHANES) data found no overall association between serum concentrations of PCBs with dioxin-like activity (mono-ortho PCBs 118 and 156 and non-ortho PCBs 126 and 169) and cognitive function scores in 708 adults aged 60–84 years, but a subgroup analysis demonstrated lower cognition with increasing serum
concentrations of dioxin-like PCBs (mono-ortho and non-ortho substituted PCBs) (Bouchard et al., 2014). Also using NHANES data, Krieg (2013) performed a limited assessment of cognition in 700 adults, aged 20–59 years, and associated scores with 12 pesticide metabolites measured in the urine of subjects, including two chemicals found in the urine after 2,4-D exposure: unmetabolized 2,4-D and 2,4-dichlorophenol. No urinary marker of 2,4-D was associated with any deficit in any of the domains of neurobehavior that were tested. In contrast, increased urine levels of the trace metabolite 2,4-dichlorophenol were associated with an improved performance on the serial digit learning test (p = 0.0002).
Update of the Epidemiologic Literature
Since Update 2014, no new studies of Vietnam veterans or case-control studies that examined neurobehavioral, cognitive, or neuropsychiatric outcomes in relation to exposure to the COIs have been published.
In a study of workers exposed to PCBs from the Dortmund transformer and capacitor recycling plant in Germany, Fimm et al. (2017) examined a broad range of cognitive functions covering attention, executive processing, reasoning, memory, and motor performance. The cohort consisted of 237 individuals, half of whom were workers, and the other half were family members of the workers, employees of surrounding companies, and area residents. The mean age was 44 years, and 17% had completed high school, that is, 12–13 years of education, with the vast majority having completed 8–10 years and a few having completed less than 8 years of education. Subjects were excluded primarily for lack of German fluency, leading to 187 individuals with complete data on the neuropsychological battery. Blood plasma PCB levels were determined and analyzed as three mutually exclusive classes: low chlorinated (PCB 28, 52, 101), high chlorinated (PCB 138, 153, 180), and dioxin-like PCBs, which included mono- and non-ortho PCBs. However, fewer than 10% of the participants had detectable levels (limit of detection = 0.01 ug/L blood plasma) of the non-ortho (coplanar) PCBs (77, 81, 126, and 169), and hence these PCBs were excluded from the analysis, leaving the mono-ortho PCBs 105, 114, 118, 123, 156, 157, 167, and 189. Exposure was dichotomized into high and low, with the 95th percentile for each congener serving as the cutpoint. For each class (low, high, and dioxin-like), an individual was placed in the high category if his or her level of at least one congener in that class was elevated. With this definition, 72% of the original 237 people were classified as having high dioxin-like PCB exposure, though many of them also had high exposures to low- or high-chlorinated PCBs, or both. Forty of the participants had elevated levels for all of the dioxin-like PCB congeners. A structural equations model was fit to identify correlations among the various tests
and PCB burdens. In multiple linear regression models, adjusting for education, dioxin-like PCBs showed no associations with any of the cognitive (vocabulary and other achievement, non-verbal reasoning), visuo-spatial processing, word fluency, flexibility, verbal and visual memory, nonverbal (geometric) learning, and several tests of attention and alertness. However, in the four sensorimotor tasks designed to assess fine motor coordination and the precision of arm/hand movement, dioxin-like PCBs were associated with poorer line tracking; specifically, a significant decrease (β = –0.206, p = 0.002) in performance on this test of fine motor control was associated with a 10-fold increase in the sum of 8 dioxin-like PCBs. The high-chlorinated PCBs were associated with poorer aiming, another fine motor task; in contrast, the low-chlorinated PCBs were associated with reduced scores on verbal tests.
Przybyla et al. (2017) conducted an analysis of the associations of whole-blood concentrations of four dioxin-like PCBs (118, 126, 156, and 169), six non-dioxin-like PCBs, and two metals (lead and cadmium) with cognitive impairment in adults aged 60–84 years as measured by the Digit Symbol Coding Test of the Wechsler Adult Intelligence Scale. The study sample consisted of 498 men and women who participated in the 1999–2000 and 2001–2002 cycles of NHANES and had not had a stroke. The final model included only those five neurotoxins (including PCB 118) that were significantly (p ≤ 0.05) associated with cognitive functioning. Estimates were adjusted for poverty–income ratio, education, race, age, sex, and smoking status. Lower cognitive scores were found for older adults who had higher concentrations of PCB 146 (β = −0.16, 95% CI −0.29− –0.02; p = 0.02), whereas higher scores were observed for increased concentrations of PCB 153, both of these being non-dioxin-like PCBs. Lower cognitive scores were also found for older adults who had higher concentrations of PCB 118, but the association was not statistically significant (β = −0.06, 95% CI −0.20–0.08; p = 0.41). The results are difficult to interpret because PCBs are generally strongly correlated with one another (the substantial change in the beta coefficient comparing the fully adjusted model with the model with only the single PCB 153 provides evidence of such correlations in these data). Adjusting the model for highly correlated variables can introduce a large bias in a non-predictable direction. Thus, it is difficult to interpret the findings of this study, although it should be noted that the cross-sectional nature is not a weakness, given that the half-lives of these compounds are generally a decade or longer.
Ames et al. (2018) reported on neurocognitive and physical functioning in the Seveso Women’s Health Study, which enrolled women who were between newborn and 40 years of age at the time of the explosion of the Seveso chemical plant in 1976 and who lived in the two closer exposure zones (A and B). The women were evaluated for their physical function in 1996 (n = 154) and with
neurocognitive tests in 2008 (n = 459). All the women assessed for physical function were post-menarche at the time of the explosion and post-menopausal at the time of testing. Archived serum samples collected in 1976 that were stored at −20° C were used for an analysis of TCDD. Serum TCDD levels were reported in pg/g lipid or parts per trillion. The physical function tests were a 10-foot walking test of functional mobility, a coin-flipping test of manual dexterity, a grip strength test, and a reach down test of lower-body mobility. Working memory was assessed on the Wechsler Adult Intelligence Scale digit span and spatial span tests, each with both backward and forward tests. Multiple linear regression models were used to assess the dose–response relationship between TCDD concentration and the outcomes of interest, with a squared term to assess consistency with a linear relationship; they also conducted further sensitivity analyses using semi-parametric methods to make fewer assumptions about functional forms. The potential confounding variables included educational attainment, smoking, alcohol consumption, age at interview, age at explosion, menarche status at explosion, menopause status (pre versus post) at interview (> 12 months without a menstrual cycle or surgical menopause), body mass index (BMI) category (referent: BMI < 25 kg/m2, overweight: BMI ≥ 25 kg/m2 and < 30 kg/m2, and obese: BM ≥ 30 kg/m2), and marital status. Directed acyclic graphs were used to inform covariate selection into the initial adjusted model, which was pared down following a change-in-estimate approach, retaining variables if their exclusion caused > 10% change in the TCDD coefficient. Results showed that serum TCDD was not associated with walking speed, upper-body mobility, or manual dexterity (p > 0.05 for these outcomes). An inverted U-shaped association was observed for grip strength, with poorer strength at both lowest and highest TCDD exposure levels, with significance for both the linear and quadratic terms in the non-dominant arm, and borderline significance in the dominant arm.1 There was no association between TCDD and the Wechsler digit and spatial span tests (p > 0.05). Neither menopause status at assessment nor age at exposure modified the associations between TCDD and working memory, that is, the lack of association was observed in both the pre- and post-menopausal women and at all ages of exposure following the Seveso explosion.
1 p = 0.01 for both linear (positive) and quadratic (negative) beta coefficients in the non-dominant arm; in the dominant arm, p-values were reported as < 0.05 for both the linear and quadratic terms in the table of the paper, but 0.07 and 0.08, respectively, in the figure; this discrepancy was due to use of the robust variance estimates for the figure but not the table (personal communication with the first author, Jennifer Ames).
Other Identified Studies
One additional study in this area was identified by the committee, but it examined biologic markers of effect of neurotransmission pathways that do not relate to a diagnosable health outcome, and therefore it was given limited consideration. Putschogl et al. (2015) used health information and blood and urine samples collected from people who either worked at a transformer and capacitor recycling plant in Dortmund, Germany, or else who lived in the immediate area and might have been exposed to dioxin-like (PCBs 105, 114, 118, 156, 157, 167, and 189) and non-dioxin-like PCBs as a result of contamination of the area by the facility. The focus of the study was on determining associations with neurotransmitter metabolites for dopamine (homovanillin acid) and norepinephrine (vanillylmandelic acid) in urine as markers of targeted effects on these specific neurotransmission pathways. All 13 dioxin-like and non-dioxin-like PCB congeners examined were significantly associated with lower urinary metabolite levels of homovanillin acid after adjustment for creatinine. Moreover, highly chlorinated congeners were more strongly associated with increased concentration of homovanillin acid but significantly reduced concentration of vanillylmandelic acid, after adjustment for creatinine. These metabolites, however, are not specific to neuronal sources, as the dietary consumption of foods with high monoamine content (e.g., cheese, orange juice, and bananas) can also contribute to homovanillin. Only 12% of peripheral homovanillin acid derives from the brain, and vanillylmandelic acid is heavily influenced by hepatic function. Neither information on diet nor diagnoses of hypertension were collected, which may confound the association. Nevertheless, these measured metabolites are a poor indicator of CNS neurotransmission, and the problems with the overall design and analysis of the study (detailed in Chapter 5) seriously limit its contribution to the scientific evidence on effects of exposure to the COIs.
Some toxicologic studies have suggested a possible involvement of the COIs in the occurrence of neurobehavioral effects. Akahoshi et al. (2009) produced a mouse neuroblastoma cell line that overexpressed the Ahr, a TCDD-induced protein hypothesized to be important in the synthesis of dopamine, whose perturbation has been implicated in a number of neurobehavioral syndromes. An elevated expression of Ahr in these cells was associated with an increased production of neurotransmitters and augmented dopamine expression, but the implication of that finding is not clear. In vitro exposure of human CD34+ cells to TCDD has been associated with a modulation of gene expression of the GABAergic pathway, which may be associated with altered synaptic transmission, visual perception, and other neurologic conditions (Fracchiolla et al., 2011). Lensu et al. (2006) treated rats with 50 μg/kg TCDD or with leptin, a chemical with
well-recognized effects on food consumption. When certain brain areas of the two treated animal groups were compared 24 hours later, the results were not consistent with a primary role for the hypothalamus brain region in TCDD-induced wasting syndrome.
There is not consistent epidemiologic evidence of an association between exposure to the COIs and neurobehavioral or cognitive disorders. One occupational study (Fimm et al., 2017) and two environmental studies (Ames et al., 2018; Przybyla et al., 2017) examined an array of neurocognitive tests, and none showed evidence of deficits in cognitive function or attention or alertness. An analysis of the Seveso Women’s Health Study cohort found reductions in grip strength, one of four measures of motor function (Ames et al., 2018). More research on the COIs and these endpoints may be warranted, especially for adverse effects of the COIs on motor function.
On the basis of the evidence reviewed here and in previous VAO reports, the committee concludes that there is inadequate or insufficient evidence to determine whether there is an association between exposure to the COIs and neurobehavioral (cognitive or neuropsychiatric) disorders.
This section summarizes the findings on neurodegenerative diseases—specifically PD, ALS, and AD—discussed in previous VAO reports. While multiple sclerosis would also be considered in this section, until Update 2014 no published epidemiologic studies for it in relation to exposure to the COIs were identified. A publication by H. K. Kang et al. (2014a) found an elevated risk of death from multiple sclerosis among female Vietnam-era veterans; however, the estimate was imprecise due to the small number of cases. Among Korean veterans, Yi et al. (2014a) found no association between prevalent multiple sclerosis and herbicide exposure, but this very large study with substantial statistical power did identify associations for a number of neurological conditions (paroxysmal disorders, nerve/plexus disorders, and paralytic syndromes), which are also considerably more specific than the outcomes evaluated in previous VAO updates.
Parkinson Disease and Parkinsonism
PD is a progressive neurodegenerative disorder that affects an estimated 6.3 million people worldwide (EBC, 2018), and it is the second-most common neurodegenerative disease (after AD). Its primary clinical manifestations are bradykinesia, resting tremor, cogwheel rigidity, and gait instability. These signs were first described as a single entity in 1817 by James Parkinson. Many non-motor manifestations of PD have been described, and they can precede the motor symptoms or present together with the motor symptoms of the disease. These include cognitive dysfunction that often progresses to frank dementia, sleep disturbances, hallucinations, psychosis, mood disorders, fatigue, and autonomic dysfunction affecting gastrointestinal, urinary, and heart function (Langston, 2006).
In the nearly two centuries since its initial description, much has been learned about some genetic predispositions and the pathophysiology of PD, but its etiology in most patients remains largely unknown, with environmental risk factors being largely understudied. The diagnosis of PD is based primarily on a clinical examination, although in recent years magnetic resonance imaging and functional brain imaging have become increasingly useful. Idiopathic PD at its onset may be difficult to distinguish from a variety of Parkinsonian syndromes, including drug-induced Parkinsonism, and neurodegenerative diseases, such as multiple systems atrophy, that present with Parkinsonian features but also develop additional brain abnormalities. Ultimately, a diagnosis of idiopathic PD can be confirmed with a postmortem pathology examination of brain tissue showing the characteristic loss of neurons from the substantia nigra and telltale protein aggregates known as Lewy body intracellular inclusions. Pathology findings in other forms of Parkinsonism show different patterns of brain injury and protein aggregation. Mortality and hospitalization records are systems with poor diagnostic standards for PD, and tend to under-report it. For example, on death records, PD is usually mentioned as a contributing cause instead of an underlying cause of death, and as such, leads to under-reporting. For PD, studies have shown that at best 60% of PD patients have PD mentioned on their death certificates (Benito-León et al., 2014; Désesquelles et al., 2014; Moscovich et al., 2017). Hospitalizations for PD do not occur until very late in the disease process, if at all, and thus even though the accuracy of hospital records in terms of diagnosis might be better than death certificates, they may miss the less severe cases. However, unless the exposure also contributes to differential under-reporting of PD on death certificates or admission to hospitals (and there is no obvious reason why this would be the case, unless exposed subjects are followed for disease more carefully by their health care system) one would expect disease misclassification to be non-differential. Therefore, as long as the specificity of the diagnosis is close to perfect (those who are listed as PD truly have the disease, which is what would be expected from death certificates and hospital records) the effect estimate would not be biased. Although the gold standard of diagnosis is pathology of the protein aggregates
in the brain (Lewy-bodies), this standard is rarely, if ever, achieved in an epidemiologic investigation due to the low rate of autopsies or brain collection. On the other hand, the longer the disease durations, the more likely it is that the diagnosis is accurate (Adler et al., 2014; Wermuth et al., 2012, 2015). Diagnosis is more accurate for patients who develop PD when they are younger than 80 years, and the youngest onset patients are likely the most accurately diagnosed. Clinical accuracy also is much higher if patients are diagnosed in specialty clinics of tertiary care facilities (by movement disorder specialists).
Several studies have attempted to estimate the incidence and prevalence of PD, but methodological differences among the studies make direct comparisons of such estimates difficult. Overall prevalence estimates of PD range from 100 to 200 per 100,000 people (von Campenhausen et al., 2005), but a recent meta-analysis of 47 studies published between 1985 and 2014 estimated an overall PD prevalence of 315 (95% CI 113–873) per 100,000 people (Pringsheim et al., 2014). Restricting the analysis to the highest-quality studies, the estimated prevalence was 571 (95% CI 243–1,339) per 100,000 people, and the authors suggested that higher-quality studies may provide a more precise estimate of disease prevalence. Stratifying on age, the prevalence estimates clearly increase with increasing age: 41 per 100,000 in individuals 40–49 years; 107 per 100,000 in individuals 50–59 years; 428 per 100,000 in individuals 60–69 years; 1,087 per 100,000 in individuals 70–79 years; and 1,903 per 100,000 in individuals over age 80 years (Pringsheim et al., 2014). Meta-analyses and other data summaries suggest a slight male preponderance in the incidence and prevalence of PD, with PD starting earlier in men (Georgiev et al., 2017).
Similarly, the estimates of PD incidence in published studies vary considerably, possibly because of methodological differences in case ascertainment and in which diagnostic criteria were applied. In a 2017 review, Tysnes and Storstein (2017) estimated that the annual incidence of PD per 100,000 inhabitants ranges from less than 10 to more than 20. A meta-analysis by age group and sex found that males had a higher incidence of PD in all age groups and that the overall incidence rate of PD for people 40 years and older was 37.55 per 100,000 person-years (95% CI 26.20–53.83) for females compared with 61.21 (95% CI 43.57–85.99) in males (Hirsch et al., 2016). The incidence rate for both males and females increases with age. For females the incidence rates by age group were calculated as: 3.26 per 100,000 in individuals 40–49 years, 8.43 per 100,000 in individuals 50–59 years, 30.32 per 100,000 in individuals 60–69 years, 93.32 per 100,000 in individuals 70–79 years, and 103.48 per 100,000 in individuals over age 80 years. For males, the incidence rates by age group were: 3.57 per 100,000 in individuals 40–49 years, 14.67 per 100,000 in individuals 50–59 years, 58.22 per 100,000 in individuals 60–69 years, 162.58 per 100,000 in individuals 70–79 years, and 258.47 per 100,000 in individuals over age 80 years.
Research on the genetic, epigenetic, and environmental causes of PD suggests multiple risk factors, including aging, environmental exposure, and genetic
predisposition (Gao and Hong, 2011; Kwok, 2010). The peak incidence and prevalence of PD are consistently found in people 60–80 years of age. A consensus statement from a 2007 meeting of PD experts (Bronstein et al., 2009) concluded that, in addition to firm evidence that the toxicant 1-methyl-4-phenyl-1,2,4,6-tetrahydropyr-idine (MPTP) can induce PD, there is substantial evidence that men are at greater risk than women, while smoking and coffee consumption are associated with a reduced risk. Of note, it has been proposed that the latter factors—especially smoking—may not be protective but rather a case of reverse causation (Ritz et al., 2014). Specifically, the higher rate of quitting smoking and reduction in caffeine intake observed among PD patients may be prodromal behavior changes in the long premotor phase of PD when many non-motor symptoms occur, such as the loss of smell, digestive problems, and sleep disturbances. Further evidence of environmental exposures playing a role in the development of PD has continued to accrue (Chin-Chan et al., 2015; Mostafalou and Abdollahi, 2017; Ritz et al., 2016; Tanner et al., 2014).
Heredity has long been suspected as an important risk factor for PD. Although risk estimates of PD for first-degree relatives of a person with PD vary from study to study and from country to country, among large studies in the United States first-degree relatives of an affected individual are 2.7–3.5 times more likely to develop PD than an individual without a family history of it. For these first-degree relatives, the cumulative lifetime risk of developing PD is between 3% and 7% (Farlow et al., 2014). An estimated 5–10% of people with PD have monogenic forms of it, which exhibit a classical Mendelian type of inheritance, but the majority of PD cases are sporadic (Kalinderi et al., 2016). All known monogenic forms of PD combined explain only about 30% of familial and 3–5% of sporadic cases (K. R. Kumar et al. 2011). At least 13 gene mutations have been identified in autosomal dominant PD, including mutations in parkin and α-synuclein (Klein and Lohmann-Hedrich, 2007). Mutations associated with an autosomal recessive inheritance pattern have also been described; however, these disease genes are found in only a handful of familial cases worldwide. More complex genetics that act to increase susceptibility to PD in conjunction with environmental risk factors may play a much more important role, and this has been the focus of much study in the recent decade (Ritz et al., 2016, 2017).
Conclusions from VAO and Previous Updates
The committees responsible for VAO, Update 1996, Update 1998, Update 2000, Update 2002, Update 2004, and Update 2006 all concluded that there was inadequate or insufficient information to determine whether there is an association between exposure to the COIs and PD. Five case-control studies reviewed by those committees had investigated the association between PD and “herbicide” exposure without providing further specificity and had reported mixed findings.
Two studies reviewed in Update 2008 examined the association specifically with chlorophenoxy acid and ester herbicides and found increased odds ratios
(Brighina et al., 2008; Hancock et al., 2008). The doubling in risk observed by Hancock et al. (2008), however, did not achieve formal statistical significance. Increases in the PD risk for the chemical class of chlorophenoxy acids or ester herbicides were reported by Brighina et al. (2008) in that patients with PD were 52% more likely than control subjects to have used these pesticides. In the Agricultural Health Study (AHS), incident PD was related in a dose–response manner to increasing days of pesticide use in general (Kamel et al., 2007b). In analyses of single pesticide exposures, 2,4-D did not increase PD incidence risk, but 2,4,5-T exposures did increase risk by 80% in adjusted hierarchical logistic regression models. On the basis of that evidence, the committee for Update 2008 concluded that there was limited or suggestive evidence associating exposure to the COIs with PD. Additional studies considered by the committees responsible for Update 2010 and Update 2012 led them to affirm this conclusion.
The committee responsible for Update 2014 was charged specifically with determining whether various diagnoses with Parkinsonian symptoms should be included in the presumptive service-related category for PD. Because diagnostic specificity is improbable in both the studies on which the conclusion of limited or suggestive association with exposure to military herbicides was based and in the documentation for the claims submitted to the Department of Veterans Affairs (VA) by Vietnam veterans, the committee clarified that the finding for PD should be interpreted by VA to include all diseases with Parkinson-like symptoms unless those symptoms can be definitively shown to be secondary to an external agent other than the herbicides sprayed in Vietnam.
Update 2014 included three studies of Vietnam veterans: one U.S. and two from the Korean Veterans Health Study. Among three cohorts of U.S. Vietnam-era military women—4,734 deployed to the theater of the war, 2,062 who served in countries near Vietnam, and 5,313 who were not deployed and served primarily in the United States—PD mortality, adjusted for age, race, duration of military service, officer status, and nursing status, was not different in those deployed to Vietnam from the non-deployed cohort, and there was no suggestion of an increase when this comparison was made for the subsets of only nurses (H. K. Kang et al., 2014a). In the Korean Veterans Health Study, 180,639 Korean veterans were followed for vital status and cause of death (Yi et al., 2014b). An exposure opportunity index (EOI) score was assigned to each veteran based on the proximity of his unit to the herbicide-sprayed areas. No association was found between PD mortality and the individual EOI scores or when the high-exposure group was compared with the low-exposure group. A second analysis compared the high- and low-exposure groups with respect to the prevalence of primary PD (International Classification of Diseases, 10th Revision [ICD-10] G20) and secondary Parkinsonism (ICD-10 G21). Effect estimates that were adjusted for age, rank, smoking, drinking, physical activity, the domestic use of herbicides, education, income, and BMI were less suggestive of an association with herbicide exposure than were the unadjusted results (Yi et al., 2014a).
Environmental (Weisskopf et al., 2012) and occupational (van der Mark et al., 2014) exposure studies reviewed in Update 2014 did not support a statistically significant association between exposure to the COIs and PD.
Update of the Epidemiologic Literature
One new study of Parkinson disease or Parkinsonism among Korean veterans who served in Vietnam was identified. No other studies of populations that specifically assessed exposure to the COIs have been published since Update 2014. Table 39, which can be found at www.nap.edu/catalog/25137, summarizes the results of studies related to Parkinson disease and Parkinson-like conditions.
Vietnam-Veteran Studies Y. S. Yang (2016) conducted an analysis of the effect of exposure to Agent Orange on Korean veterans who served in Vietnam and had been diagnosed with PD. The study consisted of 143 PD patients with exposure and 500 PD patients without exposure to Agent Orange, as determined by self-report and a military record–verified “perceived exposure index.” Comparing these patients’ clinical characteristics and their radiolabeled 18F-FP-CIT PET uptake, this study found differences in clinical profile (p < 0.05) using motor subscales from the Unified Parkinson’s Disease Rating Scale III: tremor at rest, rigidity, finger taps, and rapid alternating movement. Findings from the use of FP-CIT PET showed lower uptake in basal ganglia and higher asymmetry in exposed (p < 0.05) versus unexposed patients. The authors suggested the possibility that the PD in patients exposed to Agent Orange has a different pathophysiology from idiopathic PD.
Other Identified Studies The committee identified several additional studies that examined PD mortality or prevalence among occupational cohorts or in different populations with environmental exposures. However, each of them lacked the necessary exposure specificity to be considered further as contributing to the evidence base of the potential effect of the COIs. For example, although Ruder et al. (2014) examined U.S. workers exposed to mixed PCBs, the specific dioxin-like PCBs were not investigated separately, and no TEQs or other quantification of relevant exposures was presented. Likewise, a case-control study of clinically confirmed PD among French male farmers assessed occupational pesticide and herbicide use, duration, intensity, and cumulative exposure but did not report the specific chemicals or classes (Moisan et al., 2015). Two environmental studies were also identified—one among the residents of rural central California (Narayan et al., 2015) and the other conducted in the Netherlands (Brouwer et al., 2015), but neither specified the herbicides used in enough detail or used objective measures of exposure, such as serum concentrations, to contribute to the evidence base of exposure to the COIs and PD.
McDowell and Chesselet (2012) reviewed the literature on the ability of both toxicant-induced (6-hydroxydopamine, MPTP, rotenone, cycad) and genetically based animal models to reproduce the non-motor symptoms of PD. The very clear PD-like toxicity resulting from human exposure to MPTP indicates that select chemicals can produce the same type of damage to dopaminergic neurons as occurs in classical PD, and MPTP has become an important toxicant in studies that use animal and in vitro models. It is notable that MPTP’s bioactive metabolite, MPP+, is similar in chemical structure to paraquat (a commonly used herbicide, but not one that was used in Vietnam), but structurally unrelated to any of this report’s COIs. Pesticides shown to produce PD-like toxicity in animal models include paraquat, rotenone, maneb, and dieldrin. Substantial research has gone into understanding the molecular mechanisms responsible for the toxicity, especially in connection with paraquat and rotenone (Blandini and Armentero, 2012; Di Monte et al., 2002; Drechsel and Patel, 2008; Duty and Jenner, 2011; Hatcher et al., 2008; Moretto and Colosio, 2013; Nunomura et al., 2007; Sherer et al., 2002; Yadav et al., 2012). The damage done to dopaminergic neurons in PD is probably caused by oxidative stress and inflammation and may well also involve damage to the mitochondria in the target cells (Anderson and Maes, 2014: Janda et al., 2012; Liang et al., 2007; Littleljohn et al., 2011; Sarnico et al., 2008).
The COIs are known to be distributed to the CNS. Cholanians et al. (2016) showed that arsenic exposure leads to an accumulation of α-synuclein in both cultured cells and adult rats. The accumulation of α-synuclein plays a key role in the pathogenesis of PD, suggesting that exposure to arsenical herbicides may have an influence on PD progression. Bongiovanni et al. (2007) found that rat cerebellar granule cells in culture (an in vitro model using cells not involved in PD pathology) produce increased concentrations of reactive oxygen species when exposed to 2,4-D. González-Barbosa et al. (2017) showed that TCDD exposure causes dysregulation of the UbcH7–parkin complex in the ventral midbrain of the mouse, but further studies are needed to characterize the consequences for dopaminergic cells. The COIs have not been investigated in experimental systems such as those that have shown that paraquat and other compounds cause inflammation and oxidative stress, so it is not known whether any of the COIs could produce these responses.
Research on the neurotoxicity of 2,4-D has been going on for a number of years, but most of it has focused on its effects on the developing rodent nervous system. The studies have often used high doses of 2,4-D that have resulted in adverse changes in the developing nervous system—both neurochemical (such as changes in D2 receptors, tyrosine hydroxylase, and dopamine beta-hydroxylase) and behavioral (for example, Bortolozzi et al., 1999, 2002, 2003, 2004; Duffard et al., 1996; Evangelista de Duffard et al., 1990, 1995; Garcia et al., 2004, 2006; Rosso et al., 2000a,b). The injection of 2,4-D directly into the rat brain results in
toxicity in the basal ganglia (Bortolozzi et al., 2001), but this route of administration is highly artificial. The postpartum dietary exposure of females to 2,4-D results in adverse alterations in maternal behavior and neurochemical changes, including increases in dopamine and its metabolites 3,4-dihydroxyphenylacetic acid and homovanillic acid (Stürtz et al., 2008). Such an increase in dopamine is the reverse of what is seen in PD, in which a degradation of the dopaminergic system occurs. In addition, a study of mice and 2,4-D yielded no evidence of neurochemical damage to the dopaminergic system (Thiffault et al., 2001). Because most of the studies were on the developing nervous system, not the mature nervous system, and some studies yielded evidence of a lack of a role of 2,4-D in the development of PD, the existing studies do not support a role for the COIs in the etiology of PD.
Previously reviewed studies of PD in Vietnam veterans have not shown increased mortality or incidence of PD in U.S. or Korean Vietnam veterans. One new epidemiologic study of PD in Korean veterans who served in Vietnam was identified. Findings from the study suggested the possibility of a different pathophysiology for PD in patients exposed to Agent Orange than for idiopathic PD in terms of its clinical characteristics and brain ligand measures (fluororidopa PET). A biologic mechanism by which the COIs may cause PD has not been demonstrated. Nevertheless, the overall epidemiologic evidence continues to support an association between herbicide exposure and PD and to be consistent with an association with exposure to the phenoxy herbicides specifically.
On the basis of the lack of new evidence reviewed here supporting or refuting an association with PD, and given the evidence presented in previous VAO reports, the committee maintains the conclusion that there is limited or suggestive evidence of an association between exposure to the COIs and PD, including Parkinson-like conditions such as Parkinsonism, in the setting of dementia, multiple system atrophy, and progressive supranuclear palsy.
Amyotrophic Lateral Sclerosis
ALS is a progressive, adult-onset, motor neuron disease that presents with muscle atrophy, weakness, and fasciculations and with signs that indicate the involvement of motor neuron pathways in the CNS. The incidence of sporadic ALS is 1–2 per 100,000 person-years, and it peaks at the ages of 55–75 years (Brooks, 1996). The incidence of ALS in European populations is 2–3 people per 100,000 person-years (Johnston et al., 2006). The diagnosis of ALS is made through
clinical examination and electrodiagnostic testing and has a high degree of accuracy when performed by experienced neurologists (Rowland, 1998; Rowland and Shneider, 2001). ALS is generally more common in men than in women by a factor ranging from 1.2 to 1.5, depending on the age group (Manjaly et al., 2010). Established risk factors to date are older age, male gender, and a family history of ALS (Couratier et al., 2016).
In most cases the cause of ALS is unknown, but about 5–10% of cases are recognized as resulting from the inheritance of autosomal dominant or recessive genes (Wood, 2014); 20% of familial-ALS patients have mutations in the gene that encodes for superoxide dismutase-1 (Rosen et al., 1993). Many other possible etiologic factors have been investigated (Breland and Currier, 1967; Gallagher and Sander, 1987; Hanisch et al., 1976; H. Kang et al., 2014b; Kurtzke and Beebe, 1980; Mitchell and Borasio, 2007; Roelofs-Iverson et al., 1984; Sutedja et al., 2009a,b; M. D. Wang et al., 2014), including military service (Weisskopf et al., 2005), but no conclusive evidence has been found of an association with any of the environmental exposures addressed. Pesticides have also been suggested as a risk factor for ALS, but the results of individual studies and meta-analyses have been mixed depending how exposure was defined, and herbicides were seldom specified and not reported as a separate category (Kamel et al., 2012; H. Kang et al., 2014b; Malek et al., 2012). H. Kang et al. (2014b) expanded inclusion criteria for their meta-analysis of pesticide exposures to include the occupation of “farmer” and a rural residence. Although some of the pesticide exposures included herbicides, they were not reported as separate estimates, and thus although an increased risk of ALS was associated with pesticide exposure and with having an occupation of farmer, these results are of limited usefulness.
Conclusions of VAO and Previous Updates
ALS was first evaluated as a disease that might be associated with the COIs by the committee for Update 2002. Pesticide or herbicide exposure has been associated with an increased risk of ALS, including a doubling of the risk after long-term occupational exposure to pesticides (Deapen and Henderson, 1986) and a tripling after exposure to agricultural chemical products (Savettieri et al., 1991) and herbicides (McGuire et al., 1997), but none of the risk estimates was statistically significant. A population-based case-control study demonstrated associations between an exposure to agricultural chemical products and ALS in men, with an OR of 2.4 and a trend with duration of exposure that were both statistically significant (McGuire et al., 1997). A mortality study of Dow Chemical Company employees exposed to 2,4-D found three deaths from ALS, which resulted in a statistically significant—but imprecise—positive association (C. J. Burns et al., 2001). Weisskopf et al. (2005) followed the vital status of subjects in the American Cancer Society’s cohort for the Cancer Prevention Study II and found an increased risk of ALS in those who served in any of the armed services
during times of conflict. They adjusted for a variety of confounding variables in their model, including exposure to herbicides, and found that none of them significantly altered their conclusions; thus, this large study indirectly suggests the lack of a strong effect of herbicide exposure on ALS risk. A latter analysis of the American Cancer Society’s Cancer Prevention Study II also found no association between self-reported pesticide or herbicide exposure and ALS, but the lack of exposure specificity and the possibility of exposure estimation error limit the weight of this evidence (Weisskopf et al., 2009). Malek et al. (2014) compared 66 ALS patients to 66 controls, administering a questionnaire on occupational, vocational, and avocational exposures, and found self-reported pesticide exposure to be associated with ALS. The association was even stronger after controlling for smoking and education in a multivariate model but quite imprecise. Additional analyses conducted on occupational exposure to insecticides, to herbicides, or to fungicides and fumigants, individually, found no associations with ALS, but the sample sizes were very small. None of the results is based on sufficiently specific exposure metrics to be fully informative for VAO purposes.
Studies of veterans who served in Vietnam have reported weak and generally not statistically significant associations with the risk of ALS. A case-control study of Australian Vietnam veterans reported an association between deployment in Vietnam and ALS (ADVA, 2005c) but did not specifically study exposure to pesticides or herbicides. In the Korean Veterans Health Study, Yi et al. (2014b) calculated individual EOI scores based on the proximity of the veteran’s unit to given areas when herbicides were sprayed and partitioned the veterans into high- and low-exposure groups. No association was found between spinal muscular atrophy and the EOI scores or when the high-exposure group was compared with the low-exposure group. However, in the analysis of disease prevalence among the same Korean Vietnam veteran cohort, after adjusting for age, rank, smoking, drinking, physical activity, domestic use of herbicides, education, income, and BMI, Yi et al. (2014a) found the risk for spinal muscular atrophy to be slightly elevated in both the analysis of the scores comparing high- and low-exposure group and as a continuous variable. The more specific diagnosis of motor neuron disease (ICD-10 G12.2), which includes ALS, had nearly the same risk estimate, but because these cases represented only about one-third of those in the entire spinal muscular atrophy grouping, estimates were more imprecise.
Update of the Epidemiologic Literature
Two studies of ALS in Vietnam veterans and one case-control study of occupational exposures and environmental toxicants on the odds of developing ALS have been identified since Update 2014. Table 40, which can be found at www.nap.edu/catalog/25137, summarizes the results of studies related to ALS.
Vietnam-Veteran Studies Beard et al. (2016) used the National Registry of Veterans with ALS to identify medical record–confirmed ALS cases from 2005 to 2010 for veterans who consented at the time of enrollment in the registry to participate in further studies. Controls were selected from the Veterans Benefits Administration’s Beneficiary Identification and Records Locator System database. A total of 958 controls who were located, screened, and found to be eligible to participate were frequency-matched to 621 cases based on diagnosis age within 5 years and on socioeconomic status, as roughly estimated by the veteran’s use of the VA health care system before the date of diagnosis. All participants were followed until 2013. Clinical characteristics were extracted from medical records, and standardized telephone interviews were used to collect self-reported information on military service (branch of the longest service, number of branches, rank, the total service time, and the end of the most recent service), deployments (number; locations; time in theater; ever received imminent danger pay, hardship duty, or combat zone tax exclusion benefits for deployment to 17 foreign countries or five sea regions [plus fill-in options]), and exposures that occurred before the diagnosis date (cases) or the interview date (controls) as well as on potential confounders. Self-reported information on 39 specific military exposures was also collected, some of which were conflict-specific (e.g., Agent Orange exposure during the Vietnam War). For 31 of 32 war-specific exposures, participants were asked whether they had ever been exposed, days exposed (not exposed, ≤ 5, 6–30, > 30), and whether they felt ill after exposure (not exposed, no, yes). Inverse probability weighting was used to adjust for potential bias from confounding, missing covariate data, and selection arising from a case group that disproportionately included long-term survivors and a control group that may or may not have differed from U.S. military veterans at large. The study population consisted of 302 cases and 3,793 veterans who were aged 18–25 years during the Vietnam War. War-specific exposures for these veterans were analyzed separately, although the authors did not present all such results independently from the analysis that included all veterans and not all veterans served in that war. For those results presented, 58 cases and 77 controls reported exposure to Agent Orange in the field (OR = 2.80, 95% CI 1.44–5.44), and 8 cases and 13 controls reported mixing and application of Agent Orange (OR = 1.15, 95% CI 0.38–3.44). Vietnam veterans diagnosed with ALS reported greater exposure to pesticides on clothing or bedding (OR = 1.83; 95% CI 0.99–3.40). Overall, ALS was positively associated with an exposure to herbicides for military purposes, nasopharyngeal radium, personal pesticides, exhaust from heaters or generators, high-intensity radar waves, contaminated food, explosions within 1 mile, herbicides in the field, mixing and application of burning agents, burning agents in the field, and Agent Orange in the field.
In a second study using the same registry population of ALS veteran cases, Beard et al. (2017) examined associations between military-related factors and
ALS survival. A total of 616 medical record–confirmed cases were followed from enrollment in the registry (2005–2010) until death or July 25, 2013, whichever came first. Vital status information was obtained from several sources within VA, and information regarding military service, deployments, and 39 related exposures was collected from self-report via a standardized telephone interview (see details above in Beard et al. ). Inverse probability weights were used to adjust for potential confounding and missing covariate data biases as well as to adjust for potential selection bias among a case group that included a disproportionate number of long-term survivors at enrollment. In the study population, 137 veterans diagnosed with ALS had reported service in Vietnam. The war-specific exposures for these veterans were analyzed separately, although the authors did not present all such results independently from the analysis that included all veterans. A total of 446 deaths occurred during 24,267 person-months of followup (median follow-up: 28 months). For exposures constrained to deployment in Vietnam, longer survival was associated with exposure to Agent Orange in the field (n = 39 deaths; hazard ratio [HR] = 0.66, 95% CI 0.42–1.05) and mixing and application of Agent Orange (n = 6 deaths; HR = 0.62, 95% CI 0.32–1.20); however neither exposure estimate is statistically significant.
Case-Control Studies F. C. Su et al. (2016) conducted a case-control study to evaluate the association of occupational exposures and environmental toxicants with the odds of developing ALS in Michigan. ALS cases (n = 156) with a diagnosis of definitive, probable, probable with laboratory support, or possible ALS by revised El Escorial criteria were recruited from a tertiary referral center for ALS. Controls (n = 128) were recruited from postings and the University of Michigan clinical research volunteer database and were excluded if they were diagnosed as having ALS or another neurodegenerative condition or if they had a family history of ALS in a first- or second-degree blood relative. Participants completed a self-administered written survey that collected information on demographics, occupational and residential exposures, military service, and smoking history. Specifically, there were 58 identified exposure risk factors, 20 occupational groups, and 20 industrial groups queried for each job. Blood samples were collected from 129 cases and 119 controls and measured for concentrations of 122 persistent environmental pollutants, including organochlorine pesticides, PCBs (and dioxin-like PCB 118, specifically), and brominated flame retardants. Odds ratios and 95% CIs were calculated using standardized chemical concentrations from the imputed sample (n = 284, 10 imputations) and adjusted for age, sex, and educational levels. An effect estimate for PCB 118 was calculated only for the model that used individual compounds as covariates, which showed exposure to be associated with a decreased risk of ALS (OR = 0.69, 95% CI 0.48–1.01).
Other Identified Studies One additional study of an occupationally exposed cohort and ALS was identified (Ruder et al., 2014), but it lacked the necessary
exposure specificity or quantification to be considered further as contributing to the evidence base of the potential effect of exposure to the COIs on ALS.
Several studies have addressed the mechanisms of neurotoxicity that might be ascribed to COIs, notably 2,4-D and TCDD. Some of those effects suggest possible pathways by which the COIs could disrupt neuronal systems. A number of the studies suggest that the COIs have had neurologic effects in animal models when exposure occurred during development. There also are studies that suggest reactive oxygen species could alter specific signaling cascades and be involved in neurodegeneration. Although they do not specifically concern the COIs, such studies are potentially relevant inasmuch as TCDD and herbicides have been reported to elicit oxidative stress (Celik et al., 2006; Shen et al., 2005). The mechanistic studies suggest avenues that might be pursued to determine linkages between the COIs and the neurologic outcomes that could result in adult humans. No toxicology studies concerning exposure to the COIs and ALS have been published since Update 2006.
There is overall inadequate or insufficient evidence of an association between pesticides and ALS in broad terms, but the studies published to date have had low power to detect associations, which has resulted in numerous studies with wide CIs and non-significant ORs for exposure to the COIs. Three studies have been published since Update 2014. Two analyses of Vietnam veterans used data from the National Registry of Veterans with ALS (Beard et al., 2016, 2017). Exposure to Agent Orange in the field was statistically significantly associated with increased odds of ALS, but mixing and application of Agent Orange was not associated with ALS. Overall, ALS was positively associated with exposure to herbicides for military purposes, contaminated food, explosions within 1 mile, herbicides in the field, mixing and application of burning agents, burning agents in the field, and Agent Orange in the field. The second analysis of this population examined associations between military-related factors and ALS survival; longer survival was associated with exposure to Agent Orange in the field and mixing and application of Agent Orange, but neither of these exposure estimates was statistically significant. The third study examined occupational exposures and 122 environmental toxicants, including organochlorine pesticides and PCB 118 as measured in serum, on the odds of developing ALS. Exposure to PCB 118 was associated with a decreased risk of ALS in this population. The many epidemiologic studies reviewed in the series do not indicate a consistent association between ALS and herbicides as a class or, specifically 2,4-D or 2,4,5-T or dioxin-like PCBs, and the toxicology studies do not support an ALS mechanism specifically.
On the basis of the evidence reviewed here and in previous VAO reports, the committee concludes that the evidence of an association between exposure to the COIs and ALS remains inadequate or insufficient.
AD is a progressive, neurodegenerative form of dementia that is characterized by memory loss, confusion, mood changes, social withdrawal, and deteriorating judgment. The course of the disease is divided into four stages—pre-dementia, early, moderate, and advanced—depending on the level of cognitive and functional impairment. Diagnosis typically occurs in people over 60 years old as the symptoms develop, although pre-dementia and early AD are occasionally seen in people as young as 30 years old. AD is the sixth leading cause of death in the United States and the fifth leading cause of death in people over 65 years old (S. Singh et al., 2012). In 2017 an estimated 5.5 million Americans were living with the diagnosis. Mean life expectancy is 7 years after an AD diagnosis; about 3% of people who receive the diagnosis live 14 years or more (Alzheimer’s Association, 2018). Although the etiology of the disease remains elusive, suspected risk factors for AD include diet, exposure to aluminum or solvents, and genetics (Chin-Chan et al., 2015; de la Monte and Tong, 2014; Tanner et al., 2014).
Similar to PD diagnosis, mortality and hospitalization records are poor instruments for identifying AD cases (Désesquelles et al., 2014; Romero et al., 2014). AD is usually mentioned as contributing cause instead of the underlying cause of death, and therefore, is more prone to under-reporting. Death certificates for AD are further limited for diagnostic and research purposes because many times the type of dementia (vascular, AD, other) is not specified on death certificates. Like PD, hospitalizations for AD do not occur until very late in the disease process, if at all, and thus even though the accuracy of hospital records in terms of diagnosis might be better than death certificates, they may miss most of the less severe cases or those institutionalized in long-term care facilities who might avoid hospitalizations (Wilson and Truman, 2004). However, as long as the specificity of the AD diagnosis is close to perfect, the effect estimate would not be biased. The gold standard of AD diagnosis is pathology of the protein aggregates in the brain (amyloid beta and tau aggregates) and this standard is rarely, if ever, achieved in an epidemiologic project due to the low rate of autopsies or brain collection. However, one can assume that the longer the disease duration the more accurate the AD diagnosis. Also it is known that patients are diagnosed more accurately when they develop the diseases before the age of 80 years and the youngest onset patients are likely the most accurately diagnosed. Clinical accuracy also is much higher if patients are diagnosed in specialty clinics of tertiary care facilities (by dementia specialists), which are not common outside of urban areas.
Conclusions from VAO and Previous Updates
AD was first addressed directly in Update 2012. Until that time, literature searches had not identified epidemiologic studies that assessed the possible association of AD with exposure to the specific COIs. However, an association with exposure to the broad classification of “pesticides” had been investigated. Because AD is a condition of considerable interest to aging Vietnam veterans, studies with peripherally related available information (including those with nonspecific exposure assessment) on this outcome were included. On the basis of two studies of nonspecific herbicide exposure or mixed exposure to herbicides and other pesticides and AD (Baldi et al., 2003; Gauthier et al., 2001) that found inconsistent associations, the Update 2012 committee concluded that there was inadequate or insufficient evidence to determine whether there is an association between exposure to the COIs and AD.
Two studies of Vietnam veterans and one environmental study of AD were reviewed in Update 2014. The analyses of mortality in the Korean Veterans Health Study found no association between AD (ICD-10 G30) and the EOI scores analyzed as a continuous variable or between the high- and low-exposure groups (Yi et al., 2014b). However, a separate investigation of disease prevalence among the Korean Vietnam veterans found the risk for AD to be elevated in both the analysis of the scores as a continuous variable and the high- versus low-exposure comparison after adjusting for age, rank, smoking, drinking, physical activity, domestic use of herbicides, education, income, and BMI (Yi et al., 2014a). A subanalysis using the Canadian Study of Health and Aging examined 2,023 subjects for whom blood was available and who had a firm diagnosis of having dementia (n = 574, of which 399 were specifically diagnosed with AD) or not (n = 1,449) (Medehouenou et al., 2014). Among the 10 PCB congeners measured in the serum analyses were the mono-ortho PCBs 105, 118, and 156, which exhibit dioxin-like activity only to a modest extent. Two adjusted models were applied to the data, and of the 10 PCBs analyzed, only PCBs 105 and 118 showed an inverse association with dementia overall in the first model, but adjustment for additional confounders in the second model eliminated that negative association. Thus, no relationship with AD specifically was seen for any of the PCBs using either model. The Update 2014 committee reiterated that the evidence for an association between exposure to at least one of the COIs and AD is inadequate or insufficient.
Update of the Epidemiologic Literature
No studies of AD among Vietnam veterans or other populations that specifically assessed exposure to the COIs have been published since Update 2014.
Environmental Studies D. H. Lee et al. (2016) measured PCBs and organochlorine pesticides in the participants of the PIVUS (Prospective Investigation
of the Vasculature in Uppsala Seniors) cohort of residents 70 years and older in Uppsala, Sweden, to examine cognitive impairment. Of 2,025 eligible subjects, only half (n = 1,016) agreed to participate and completed a questionnaire to assess medical history, smoking history, and medication use. In a subcohort of 989 men and women, the investigators determined whether PCB or organochlorine exposures were associated with cognitive impairment during 10 years of follow-up, relying on medical record review and death certificate information to ascertain disease status. General exposure to 16 PCBs and 3 organochlorine pesticides (p,p′-DDE, trans-nonachlor, and hexachlorobenzene) was quantified using high-resolution gas chromatography/mass spectrometry from stored serum samples collected at the time of entry into the PIVUS study, and these measure were normalized using individual lipid levels. The authors analyzed exposures using a compound summary score for the three organochlorine pesticides and the PCBs, grouping them according to the < 25th, 25th–75th, and > 75th percentile of exposure. Adjusted hazard ratios for the higher summary score PCB exposures compared with exposure < 25th percentile showed no associations with cognitive impairment, but the study did not distinguish between dioxin-like and non-dioxin-like PCBs. For the highest exposure (> 75th percentile) of hexachlorobenzene, a chemical with dioxin-like activity, the weight-loss-adjusted HR was 2.4 (95% CI 1.0–5.8) in those who developed cognitive impairment after age 75, and the p-trend was statistically significant. These results suggest that exposure to hexachlorobenzene may predispose to the development of cognitive impairment or AD.
Other Identified Studies Two additional studies of occupational exposures and mortality from AD and dementia (Ruder et al., 2014) or mortality from non-vascular dementia (Koeman et al., 2015) were identified. However, both studies lacked the necessary exposure specificity or quantification to be considered further as contributing to the evidence base of the potential effect of exposure to the COIs on AD or dementia.
There has been little toxicologic investigation of adult exposure to the COIs and endpoints relevant to AD.
The findings in the Korean Veterans Health Study (Yi et al., 2014a,b) remain the first in which the risk of AD has been investigated in association with a fully relevant exposure, but the findings were not consistent. Except for the analysis of the PIVUS cohort (D. H. Lee et al., 2016), which only suggested an association with cognitive impairment for hexachlorobenzene and only in subgroup analyses,
there are no new relevant epidemiologic studies on AD to add to the evidence base since the last update, and the toxicologic data remains sparse.
On the basis of the lack of new evidence reviewed here and that reviewed in previous VAO reports, the committee concludes that there is inadequate or insufficient evidence to determine whether there is an association between exposure to the COIs and AD.
Peripheral neuropathies are an array of disorders caused by damage to nerve fibers (axonal neuropathies) or to the myelin sheath that surrounds many fibers (demyelinating neuropathies). The manifestations of neuropathy can include a combination of sensory changes, weakness, and autonomic instability. Clinically, various forms of peripheral neuropathy can be characterized by the distribution of nerve abnormalities and their patterns of progression.
Peripheral neuropathy resulting from toxic exposure usually affects nerve fibers in a symmetric pattern, beginning distally in the longest fibers (in the toes) and moving proximally (toward the spine). This kind of neuropathy is called symmetric axonal sensorimotor polyneuropathy. Sensory deficits begin at the toes, progress above the ankles, and only later affect the hands. Motor symptoms show the same general pattern. Physiologically, various forms of peripheral neuropathy can be characterized by the results of electrodiagnostic testing to indicate which neural structures are affected. Most toxicant-induced neuropathies involve injury to the nerve-cell bodies or the axons, giving rise to changes in the amplitude of a nerve’s response to an electric stimulus. The clinical manifestations of most symmetric axonal neuropathies are similar except for variations in the rates of progression and in whether pain is prominent. No specific signature distinguishes a toxicant-related neuropathy from one induced by other causes. As many as 30% of neuropathies are “idiopathic,” that is, no etiology is determined despite exhaustive clinical evaluation.
The most common toxicant-induced neuropathy occurs as a result of chronic alcohol exposure. Peripheral neuropathy also occurs commonly as a complication of diabetes; its reported prevalence in people who have chronic diabetes is up to 50%. Thus, it is important to include an assessment of alcohol use and diabetes as covariates in epidemiologic studies because the neuropathies that are related to these conditions are clinically and physiologically indistinguishable from other toxicant-induced neuropathies.
Toxicant exposure can result in early onset (immediate) peripheral neuropathy or delayed-onset peripheral neuropathy, which occurs years after the external exposure has ended. For classification purposes, the committee considers a neuropathy early onset if abnormalities appear within 1 year after external exposure
ends and delayed-onset if abnormalities appear more than 1 year after external exposure ends. Because the exposures of interest for Vietnam veterans are long past (more than 40 years ago), the immediate effects of the COIs are no longer pertinent for this cohort. Such outcomes would include early onset peripheral neuropathy and porphyria cutanea tarda. Because early onset peripheral neuropathy is not necessarily a transient condition, it may become a chronic condition that should be distinguished from delayed-onset peripheral neuropathy. The focus of this section is on data related to delayed-onset peripheral neuropathy.
Conclusions from VAO and Previous Updates
Several studies of Vietnam veterans have examined peripheral neuropathy. A study by the Centers for Disease Control and Prevention (CDC, 1988b) reported a slight excess in the signs or symptoms of peripheral neuropathy among deployed versus non-deployed Vietnam-era veterans. Decoufle et al. (1992) reported no association between self-reported exposure to herbicides in Vietnam and peripheral neuropathy. There was no indication of an increased incidence of peripheral neuropathy in the first examination of the Air Force Health Study (AFHS), which established the baseline for Ranch Hand veterans (AFHS, 1984a). Michalek et al. (2001c) described peripheral neuropathy in the AFHS cohort. In a primary analysis, the investigators had included diabetes as a potential confounder in the statistical model. In a secondary analysis, the subjects who had conditions that were known to be associated with neuropathy were excluded, and the subjects who had diabetes were enumerated. In both analyses, there were strong and significant associations between serum dioxin concentrations and possible and probable neuropathy, and significant trends were found with increasing concentrations of dioxin. However, there were too few nondiabetic subjects to produce useful estimates of risk in the absence of the contribution of diabetes. Thus, questions remained about the specific association between an exposure to the COIs and peripheral neuropathy in the absence of any effect of diabetes. The large veteran studies are limited by the confounding nature of concurrent diabetes and alcohol exposure, both of which are also related to neuropathy.
In the study of disease prevalence among the Korean Vietnam veterans, after adjusting for age, rank, smoking, drinking, physical activity, the domestic use of herbicides, education, income, and BMI, Yi et al. (2014a) found that the risk for polyneuropathies of the PNS (ICD-10 G60–G64) was slightly elevated in both the analysis of the EOI scores as a continuous variable and in the high- versus low-exposure comparison.
D. H. Lee et al. (2008) evaluated the association of exposure to a variety of toxicants to the presence of neuropathy in subjects who had either frank diabetes or impaired glucose tolerance. No evidence of an increased incidence of neuropathy or of a dose–response relationship was found for either group when stratified by high- versus low-hemoglobin A1c level that suggested a concentration-dependent
risk of neuropathy. However, this study was limited by the small sample size and a lack of information regarding the duration of diabetes.
Update of the Epidemiologic Literature
No studies of exposure to the COIs and chronic peripheral system disorders were identified for the current update.
No new toxicity studies directly pertinent to the COIs and peripheral neuropathy were identified for the present update. However, it is worth reiterating findings from earlier updates. Neuronal cell cultures treated with 2,4-D showed decreased neurite extension associated with intracellular changes, including a decrease in microtubules, an inhibition of the polymerization of tubulin, disorganization of the Golgi apparatus, and an inhibition of ganglioside synthesis (Rosso et al., 2000a,b). The normal activity of those target processes is important for maintaining synaptic connections between nerve cells and for supporting the mechanisms involved in axon regeneration during recovery from peripheral neuropathy. Grahmann et al. (1993) and Grehl et al. (1993) reported observation of, respectively, electrophysiologic and pathologic abnormalities in the peripheral nerves of rats treated with TCDD. When the animals were sacrificed 8 months after exposure, there was pathologic evidence of persistent axonal nerve damage and histologic findings typical of toxicant-induced injury. These results constitute evidence of biologic plausibility for an association between exposure to the COIs and peripheral neuropathy.
The epidemiologic studies relating industrial or individual exposure to acute neuropathy were judged by the committee for Update 1996 and later updates to provide limited or suggestive evidence of an association between exposure to the COIs and early onset transient peripheral neuropathy. Beginning with Update 2010, these acute outcomes were removed from this section to keep the focus on chronic and delayed-onset conditions. That committee concluded that no data suggest that exposure to COIs can lead to the development of delayed-onset chronic neuropathy many years after the termination of exposure of those who did not originally complain of early onset neuropathy and concluded the evidence to be inadequate or insufficient. Since Update 2010, only one relevant study was identified and reviewed, and that study was reviewed in Update 2014. Among Korean Vietnam veterans, the prevalence of polyneuropathies was slightly elevated in high- versus low-exposure comparisons. Given the lack of new epidemiologic studies and toxicologic information, the committee agrees with the prior two
committees and concludes that the evidence does not support an association between exposure to the COIs and the development of delayed-onset chronic neuropathy.
On the basis of the lack of new evidence to date, the committee concludes that there is inadequate or insufficient evidence to determine whether there is an association between exposure to the COIs and delayed-onset chronic neuropathy.
Hearing loss increases markedly with age, from approximately 3% among adults aged 20–29 years to an estimated 49% among people 60–69 years of age and more than 80% among people 85 years and older (NASEM, 2016b). These estimates are based on NHANES data and are likely to underestimate the true population prevalence of hearing loss because the NHANES sampling frame does not include people living in assisted care facilities, group homes, or nursing homes or those unable to come to the mobile examination center. Hearing loss is somewhat higher in men than in women (NASEM, 2016b). The most common forms of hearing impairment in adults are presbycusis and tinnitus. Heritable factors may influence the susceptibility to hearing loss, but external agents can also contribute. Aspirin at high doses can cause reversible tinnitus, and permanent hearing loss may be induced by certain pharmaceuticals (particularly antibiotics and antineoplastic drugs) and by some environmental and industrial chemicals (primarily solvents and metals) (Cannizzaro et al., 2014). In occupational medicine, hearing loss is most often regarded as noise induced. Farmers and migrant or seasonal agricultural workers have also been found to have high rates of hearing loss compared with those who had the lowest levels of pesticide exposure (Crawford et al., 2008; Rabinowitz et al., 2005).
Conclusions from VAO and Previous Updates
Epidemiologic results on hearing loss in relation to service in Vietnam or to herbicide exposure more generally were first discussed in Update 2010, when two citations that addressed this health outcome were identified. O’Toole et al. (2009) re-examined the health status of a cohort of Australian Vietnam veterans; as for almost every health endpoint surveyed in that group, the incidences of self-reported complete or partial deafness and of tinnitus showed statistically significant increases compared to the general population, but these results are likely compromised by methodologic problems with the study. Excesses in self-reported hearing loss were also found among licensed pesticide applicators in the AHS at the time of the 5-year follow-up interview (Crawford et al., 2008), but this effect was associated with insecticide exposure, not with herbicide use.
Update of the Epidemiologic Literature
No epidemiologic studies addressing exposure to the COIs and hearing loss have been published since Update 2014.
Although no studies of hearing loss in adult animals directly exposed to the COIs were found, Crofton and Rice (1999) reported that perinatal maternal exposure to PCB 126 (a dioxin-like PCB) resulted in low-frequency hearing deficits in the offspring of exposed maternal rats. Increased auditory thresholds occurred in the group treated at 1.0 μg/kg/day for 0.5- and 1-kHz tones, but higher frequencies were not significantly affected. The frequency-specific deficit was hypothesized to be secondary to a postnatal hypothyroxinemia that occurred during a sensitive period for the development of the low-frequency regions of the cochlea, which was consistent with the finding that the pups had decreased T4 concentrations in their sera on postnatal day 21.
Two prior studies observed an increased prevalence of hearing loss in Vietnam veterans and pesticide applicators, but neither was able to examine exposure specifically to the COIs or to confirm hearing loss clinically. Furthermore, the report from the AHS (Crawford et al., 2008) observed an association only in insecticide applicators, not in herbicide applicators. Hearing loss among Australian Vietnam veterans was compared with the general Australian population, not veterans from the same era who were not deployed to Vietnam, so it could not distinguish between hearing loss that may be associated with noise that was related to military service and hearing loss potentially associated with exposures to toxic chemicals. In the absence of new studies, the conclusion remains unchanged since Update 2010.
On the basis of the evidence reviewed here, the committee concludes that there is inadequate or insufficient evidence to determine whether there is an association between exposure to the COIs and hearing loss.