Abstract
Background:
Dream enacting behavior is a characteristic feature of rapid eye movement sleep behavior disorder, the most specific prodromal marker of synucleinopathies. Pesticide exposure may be associated with dream enacting behaviors, but epidemiological evidence is limited.
Objectives:
To examine high pesticide exposure events in relation to dream enacting behaviors among farmers in the Agricultural Health Study.
Methods:
We conducted multivariable logistic regression analyses to examine high pesticide exposure events reported in 1993–1997 in relation to dream enacting behaviors assessed in 2013–2015 among 11,248 farmers (age 47±11 years).
Results:
A history of dream enacting behaviors was reported by 939 (8.3%) farmers. Compared with farmers who did not report any high pesticide exposure event, those who reported were more likely to endorse dream enacting behaviors two decades later (odds ratio =1.75 (95% confidence interval: 1.49–2.05)). The association appeared stronger when there was a long delay in washing with soap and water after the event (2.63 (1.62–4.27) for waiting >6 hours vs. 1.71 (1.36–2.15) for washing within 30 minutes), and when the exposure involved the respiratory or digestive tract (2.04 (1.62–2.57) vs. 1.58 (1.29–1.93) for dermal contact only). In the analyses of specific pesticides involved, we found positive associations with two organochlorine insecticides (DDT and lindane), four organophosphate insecticides (phorate, ethoprop, terbufos, and parathion), two herbicides (alachlor and paraquat), and fungicides as a group.
Conclusions:
This study provides the first epidemiological evidence that high pesticide exposures may be associated with a higher risk of dream enacting behaviors.
Introduction
Rapid eye movement sleep behavior disorder (RBD) is a parasomnia characterized by a loss of muscle atonia during rapid eye movement sleep1 with the presence of dream enacting behaviors. A clinical RBD diagnosis requires episodes of motor behaviors or vocalization during rapid eye movement sleep evidenced by video-polysomnography or a history of dream enacting behaviors, coupled with polysomnography evidence of rapid eye movement sleep without atonia1. In epidemiological studies, as polysomnography is often infeasible, the presence of dream enacting behaviors has been commonly used as a surrogate for probable RBD2,3.
Although polysomnography-confirmed RBD is rare in the general population4–6, it is very common among patients with α-synuclein disorders, including Parkinson’s disease (PD), dementia with Lewy bodies, and multiple system atrophy7. Importantly, idiopathic RBD develops in prodromal synucleinopathies, many of which will eventually convert to neurodegenerative diseases8. Therefore, research on RBD may provide an unprecedented opportunity to understand the decades-long prodromal development of synucleinopathy.
Studies have investigated potential risk factors of polysomnography-confirmed RBD9–11or probable RBD based on symptomatic screeners12–16. While the findings are not entirely consistent, they generally suggest higher prevalence of RBD associated with low educational level11–13,16, smoking11,13,15, head injury11,12,15,16, depression9, use of antidepressants9,13,16, and pesticide exposures11,14–16. Notably, the findings of RBD on head injury and pesticide use are consistent with those for PD17, suggesting that head injury and pesticides may contribute to synucleinopathies in their early stages of development.
By leveraging unique data from the Agricultural Health Study, we investigated whether high exposures to pesticides might be associated with dream enacting behaviors. In the study, about 14% of the eligible pesticide applicators (hereafter referred to as farmers) in Iowa and North Carolina reported at enrollment ever having had at least one high pesticide exposure event 18. We examined whether such high exposure events were associated with dream enacting behaviors, and whether the association varied based on the decade when the highest exposure event occurred, the body parts exposed, the time delay until cleaning with soap and water, and the specific chemicals involved during the highest exposure event.
Methods
Study Population
The Agricultural Health Study aims to investigate how pesticides and other agricultural exposures may affect the health of farming populations19. Briefly, 52,394 farmers from Iowa and North Carolina enrolled between 1993 and 1997 by completing a questionnaire about demographics, lifestyle, farming practices, and lifetime use of pesticides. Of these, 22,916 (44%) farmers also completed a take-home questionnaire to provide additional details on pesticide use, including a history of unusually high pesticide exposure events. Following the baseline surveys (Phase 1), participants updated their exposure and health status every 5–6 years via telephone interviews or mailed surveys at Phase 2 in 1999–2003, Phase 3 in 2005–2010, and Phase 4 in 2013–2015. Participants consented to the study by returning the enrollment questionnaires or participating in the telephone or mailed follow-up surveys. The study protocol was approved by the institutional review boards of the National Institute of Environmental Health Sciences and the National Cancer Institute, and other relevant institutions. This specific secondary data analysis was exempted by the institutional review board of Michigan State University.
High Pesticide Exposure Events
In the Phase 1 take-home survey, farmers reported whether they had ever had an incident or experience while using any type of pesticide which caused unusually high personal exposure. Those who answered yes were further asked details about the highest exposure incident they self-identified, including the name of the pesticide involved, the decade it occurred, the body parts exposed, and the time delay between the event and washing with soap and water.
High pesticide exposure events were subsequently updated in Phases 2 and 3 with slightly different questions. At Phase 2, participants were asked, “since enrollment, did you have any incidents with fertilizers, herbicides or other pesticides that caused you unusually high personal exposure?” and at Phase 3 “Since the year of the last interview, have you had any incidents or spills that resulted in unusually high exposure to pesticides from contact with your skin, from breathing fumes, or dust, or from accidental ingestion?” On each occasion, details of the most recent incident were obtained.
Dream Enacting Behaviors
We assessed dream enacting behaviors at the Phase 4 survey using the validated one-item screening question for probable RBD3. We asked farmers: “Have you ever been told, or suspected yourself, that you seem to “act out your dreams” while sleeping? For example, punching or flailing your arms in the air, shouting, or screaming while asleep.” This question was designed for use in large epidemiological studies to screen for probable RBD and has shown high sensitivity (93.8%) and specificity (87.2%) in a clinical validation study3. To those who answered ‘yes’ to this question, we further asked about the time-period it first occurred (<1, 1–5, 5–10, >10 years ago) and the frequency of occurrence (< 3 times in life, <1/month, 1–3/month, once a week, or >1/week). We reported previously that the presence of dream enacting behaviors was associated with ~8 times higher odds of having PD in this cohort20.
Statistical analysis
The primary analysis examined high pesticide exposure events reported at baseline in relation to dream enacting behaviors reported at Phase 4, using multivariable logistic regression adjusting for baseline age, sex, race, state, education level, smoking status, alcohol drinking, and histories of head injury and depression as defined in Table 1. Of the 13,991 farmers who participated in both surveys, 11,248 were included in the primary analyses after excluding proxy-respondents at Phase 4 (n=1,458), farmers with missing data on dream enacting behaviors (n=177), high pesticide exposure events (n=284), or covariates (n=824) (Figure S1).
Table 1.
Baseline characteristics (1993–1997) by dream enacting behaviors reported in the Phase 4 survey (2013–2015)a.
| Characteristics | Dream enacting behaviors |
|
|---|---|---|
| Never (n = 10,309) |
Ever (n = 939) |
|
|
| ||
| Age (year) | ||
| ≤45 | 4,913 (47.7) | 432 (46.0) |
| 46–55 | 2,817 (27.3) | 250 (26.6) |
| 56–65 | 2,122 (20.6) | 201 (21.4) |
| ≥66 | 457 (4.4) | 56 (6.0) |
| Sex | ||
| Men | 10,039 (97.4) | 924 (98.4) |
| Women | 270 (2.6) | 15 (1.6) |
| Race | ||
| White | 10,218 (99.1) | 928 (98.8) |
| Others | 91 (0.9) | 11 (1.2) |
| State | ||
| Iowa | 7,584 (73.6) | 663 (70.6) |
| North Carolina | 2,725 (26.4) | 276 (29.4) |
| Marital status | ||
| Married/living as married | 8,747 (84.9) | 827 (88.1) |
| Never married | 1,096 (10.6) | 62 (6.6) |
| Divorced/widowed | 466 (4.5) | 50 (5.3) |
| Education | ||
| High school or lower | 5,318 (51.6) | 483 (51.4) |
| 1–3 y beyond high school | 2,702 (26.2) | 260 (27.7) |
| College graduate or more | 2,289 (22.2) | 196 (20.9) |
| Smoking status | ||
| Never | 5,999 (58.2) | 490 (52.2) |
| Former | 3,248 (31.5) | 323 (34.4) |
| Current | 1,062 (10.3) | 126 (13.4) |
| Drinking alcohol in the past year | ||
| No | 3,341 (32.4) | 290 (30.9) |
| Yes | 6,968 (67.6) | 649 (69.1) |
| Ever diagnosed with head injury | ||
| No | 8,999 (87.3) | 781 (83.2) |
| Yes | 1,310 (12.7) | 158 (16.8) |
| Ever diagnosed with depression | ||
| No | 9,863 (95.7) | 861 (91.7) |
| Yes | 446 (4.3) | 78 (8.3) |
| Cumulative days of any pesticide use (days)b | ||
| 0–64 | 2,439 (23.7) | 214 (22.8) |
| >64–225 | 3,357 (32.6) | 282 (30.0) |
| >225–457 | 2,181 (21.2) | 205 (21.8) |
| >457 | 1,929 (18.7) | 207 (22.0) |
| Missing | 403 (3.9) | 31 (3.3) |
| Number of non-specific symptomsc | ||
| 0 | 5,684 (55.1) | 398 (42.4) |
| 1 | 1,904 (18.5) | 190 (20.2) |
| 2–3 | 1,496 (14.5) | 173 (18.4) |
| >3 | 1,179 (11.4) | 173 (18.4) |
| Missing | 46 (0.5) | 5 (0.5) |
| Number of health conditionsd | ||
| 0 | 6,068 (58.9) | 481 (51.2) |
| 1 | 2,837 (27.5) | 301 (32.1) |
| >1 | 1,404 (13.6) | 157 (16.7) |
Numbers and percentages are provided.
Cumulative days of any pesticide use was obtained as a product of years of use (based on enrollment question: “How many years did you personally mix or apply pesticides?”) and days of use (based on enrollment question: “During those years, how many days per year did you personally mix or apply pesticides?”).
Number of 23 non-specific symptoms (dizziness, tension, nausea/vomiting, fatigue, excessive sweating, poor night vision, absentmindedness, headache, loss of appetite, fast heart rate, poor balance, blurred/double vision, difficulty concentrating, numbness in hands or feet, loss of consciousness, irritability, tremor in hands, insomnia, difficulty speaking, weakness in your arms or legs, changes in smell or taste, feeling depressed, and twitches in arms or legs), defined as experienced once a week and more during the year before enrollment.
Number of 12 major health conditions (cardiovascular diseases and stroke, high blood pressure, diabetes, asthma, chronic obstructive pulmonary disease, pneumonia, farmer’s lung disease, kidney diseases, kidney stones, ulcerative colitis, nervous system diseases, and cancer), defined as ever diagnosed at or before enrollment.
In the analyses of details of the highest exposure event, we analyzed the association by the decades when the event occurred, the time delay between the event and washing with soap and water, and the route of exposure, as defined in Table 2. Among farmers who reported high pesticide exposure events, we further examined the linear trend for associations between the time delay in cleaning and dream enacting behaviors by defining the median of each exposure category as a continuous variable. Finally, a total of 130 specific pesticides and active ingredients were reported for the highest exposed events. We first grouped these specific pesticides into functional or chemical groups and then analyzed the associations of dream enacting behaviors with each pesticide group and specific pesticide. We limited these analyses to the pesticides that had at least 5 exposed cases. In all analyses, we used farmers without any high pesticide exposure events as the reference group.
Table 2.
High pesticide exposure events at baseline (1993–1997) in relation to dream enacting behaviors reported in the Phase 4 survey (2013–2015).
| High pesticide exposure events | Dream enacting behaviors |
Odds ratio (95% confidence interval)a |
|
|---|---|---|---|
| Never (n = 10,309) |
Ever (n = 939) |
||
|
| |||
| Ever exposed | |||
| No | 8,698 (84.4) | 703 (74.9) | 1.00 (Ref) |
| Yes | 1,611 (15.6) | 236 (25.1) | 1.74 (1.49, 2.05) |
| Decade when the highest exposed event occurred | |||
| No exposure | 8,698 (84.4) | 703 (74.9) | 1.00 (Ref) |
| 1990s | 244 (2.4) | 45 (4.8) | 2.41 (1.72, 3.36) |
| 1980s | 616 (6.0) | 78 (8.3) | 1.53 (1.19, 1.97) |
| 1970s | 438 (4.3) | 57 (6.1) | 1.48 (1.11, 1.98) |
| 1960s or beforec | 171 (1.7) | 27 (2.9) | 1.78 (1.17, 2.71) |
| Missingd | 142 (1.4) | 29 (3.1) | -- |
| Time delay between the eventb and washing with soap and water | |||
| No exposure | 8,698 (84.4) | 703 (74.9) | 1.00 (Ref) |
| <30 min | 686 (6.7) | 97 (10.3) | 1.71 (1.36, 2.15) |
| 30–59 min | 265 (2.6) | 32 (3.4) | 1.40 (0.96, 2.05) |
| 1–3 h | 312 (3.0) | 44 (4.7) | 1.67 (1.20, 2.32) |
| 4–6 h | 202 (2.0) | 33 (3.5) | 1.91 (1.31, 2.80) |
| >6 h | 93 (0.9) | 21 (2.2) | 2.63 (1.62, 4.27) |
| Missingd | 53 (0.5) | 9 (1.0) | -- |
| Exposure routeb | |||
| No exposure | 8,699 (84.4) | 703 (74.9) | 1.00 (Ref) |
| Dermal onlye | 1,023 (9.9) | 132 (14.1) | 1.58 (1.29, 1.93) |
| Respiratory/digestive tracte | 564 (5.5) | 101 (10.8) | 2.04 (1.62, 2.57) |
| Missingd | 24 (0.2) | 3 (0.3) | -- |
Numbers and percentages are provided.
Adjusted for age, sex, race, state, education, marital status, smoking status, alcohol consumption, history of head injury, and history of depression.
The highest exposure event.
1940s to 1960s.
Numbers with missing data among those who reported high pesticide exposure events.
Respiratory or gastrointestinal tract exposure refers to participants having breathed fumes, ingested or swallowed pesticides during the event no matter whether they had dermal exposure or not. Dermal only exposure refers to participants reported exposure on head, face, arms, hands, back, abdomen, groin area, legs, or feet but did not report breathing/ingesting/swallowing the pesticide.
We conducted multiple sensitivity analyses to examine the robustness of study results. First, because farmers who used pesticides heavily were more likely to experience high pesticide exposure event18, we further adjusted for a) the cumulative lifetime days of use of any pesticide at baseline, or b) the ever-use of 14 specific pesticides that showed associations with dream enacting behaviors in a previous analysis15. Second, to account for potential confounding from comorbidity and symptoms that may relate to both high pesticide exposure events and dream enacting behaviors, we further adjusted for the number of non-specific symptoms21 and the number of major health conditions at baseline. Third, we redefined cases as those who reported symptom onset within 10 years before the Phase 4 survey to minimize potential impacts from prevalent cases. Fourth, we repeated the analyses by excluding those who reported <3 episodes of dream enacting behaviors to reduce potential false positives. Fifth, we applied inverse probability weighting to account for potential selection biases due to attrition and proxy response, and to make inferences to all farmers who returned the take-home questionnaire at baseline and were still alive and eligible for the Phase 4 survey, using an approach published previously22. Finally, in the analyses of specific pesticides or pesticide groups that were involved in the highest exposed event, we further adjusted for a) the cumulative lifetime days of use of any pesticide as a surrogate for overall exposure or b) the cumulative lifetime days of use of that specific pesticide or pesticide group as a surrogate of exposure to that specific chemical(s).
In addition to the analyses using baseline high pesticide exposure events, we conducted two secondary analyses combining updated exposures. Specifically, we first combined high pesticide exposure events reported at Phase 2 with those reported at baseline and examined the association with cases with onset within 10 years before the Phase 4 survey. Using the same strategy, we also updated exposures through Phase 3 and examined their associations with cases with onset within 5 years before Phase 4. In both analyses, we can reasonably assume the cases were incident after the exposure. Sample sizes for these secondary analyses are also presented in Figure S1. We did not analyze details of the high exposure events reported at Phase 2 and 3 due to small sample sizes. Analyses were based on data releases P1REL201701.00, P2REL201701.00, P3REL201808.00, P4REL201809.04, and AHSREL201909.00. We conducted analyses using the Statistical Analysis System, version 9.3 (SAS Institute, Inc., Cary, North Carolina), with two-sided tests and α of 0.05.
Role of the funding source
The funding agency had no role in the study design, data collection, data analysis, result interpretation, or manuscript preparation and submission. The authors had full access to the analytic data and had the final responsibility for the decision to submit the manuscript for publication.
Data sharing:
No additional data available.
Results
A total of 939 (8.3%) of the 11,248 eligible participants reported dream enacting behaviors at the Phase 4 survey. Compared to farmers without dream enacting behaviors, those who reported such behaviors were more likely to be current smokers, married or living as married, had histories of head injury and depression, and reported more non-specific symptoms and major chronic conditions at baseline (Table 1).
In the analytic sample, 1,847 farmers (16.4%) reported a history of high pesticide exposure events at baseline. Compared to farmers without such an exposure, those who reported a history of high pesticide exposure events had 74% higher odds of endorsing dream enacting behaviors two decades later (odds ratio (OR)=1.74; 95% confidence intervals (CI): 1.49–2.05; Table 2). In the analyses considering when the highest exposed event occurred, incidents in the 1990s showed the strongest association (OR=2.41, 95%CI: 1.72, 3.36), followed by moderate associations for events in earlier decades (ORs ranging from 1.48 to 1.78). In the analysis examining the time interval between the highest exposed event and washing with soap and water, compared with those without such an exposure, the OR ranged from 1.40 (95%CI: 0.96–2.05) for cleaning within 30–59 minutes to 2.63 (95%CI:1.63, 4.27) for waiting >6 hours. Among farmers who reported high pesticide exposure events, the P-value for linear trend was 0.09. Finally, with no high exposure events as the reference, the OR was 2.04 (95%CI 1.62–2.57) for exposure events involving the respiratory or digestive tract versus 1.58 (95%CI 1.29–1.93) for events via dermal exposure only (P-value for between-group difference = 0.07). Similar results were obtained in all sensitivity analyses (Table S1).
Table 3 presents data for specific pesticides involved in the highest exposed event. We observed positive associations of dream enacting behaviors with two organochlorine insecticides (lindane and DDT (i.e., dichlorodiphenyltrichloroethane)), four organophosphate insecticides (phorate, ethoprop, terbufos, and parathion), and two herbicides (alachlor and paraquat) with OR ranging from 1.61 (95%CI: 1.05–2.48) for alachlor to 5.88 (95%CI: 1.95–17.84) for parathion. For pesticide groups, we found associations of dream enacting behaviors with organochlorines (OR=2.58, 95%CI:1.61–4.14), organophosphates (OR=2.33, 95%CI: 1.72–3.15), chloroacetanilides (OR=1.46, 95%CI: 1.01–2.13) and fungicides (OR=2.75, 95%CI: 1.12–6.75). Similar results were found in the sensitivity analyses that further adjusted for the cumulative lifetime days of use of any pesticides or the cumulative lifetime days of use of that specific pesticide or pesticide group (Table S2). Notably, most of these pesticides have been banned in the European Union, but many are still in use in the US (Table S3).
Table 3.
Specific pesticides involved in the highest exposed event at baseline (1993–1997) in relation to dream enacting behaviors reported in the Phase 4 survey (2013–2015).
| Pesticides involved in the highest exposed event | Dream enacting behaviors |
Odds ratio (95% confidence interval) a |
|
|---|---|---|---|
| Never | Ever | ||
|
| |||
| No exposure | 8,698 | 703 | 1.00 (Ref) |
| Pesticides by chemical/functional groupsb | |||
| Organochlorine insecticides | 98 | 22 | 2.58 (1.61, 4.14) |
| Organophosphate insecticides | 270 | 55 | 2.33 (1.72, 3.15) |
| Carbamate insecticides | 76 | 11 | 1.71 (0.90, 3.26) |
| Chloroacetanilide herbicides | 273 | 33 | 1.46 (1.01, 2.13) |
| Triazine herbicides | 223 | 22 | 1.21 (0.77, 1.89) |
| Phenoxy herbicides | 191 | 19 | 1.15 (0.71, 1.86) |
| Fungicides | 27 | 6 | 2.75 (1.12, 6.75) |
| Fumigants | 55 | 6 | 1.17 (0.50, 2.77) |
| Individual pesticideb | |||
| Lindane | 22 | 6 | 3.15 (1.26, 7.88) |
| DDT | 23 | 6 | 2.96 (1.19, 7.34) |
| Phorate | 68 | 18 | 2.90 (1.70, 4.95) |
| Ethoprop | 20 | 6 | 3.10(1.22, 7.88) |
| Terbufos | 39 | 9 | 2.99 (1.43, 6.25) |
| Parathion | 9 | 5 | 5.88 (1.95, 17.84) |
| Malathion | 46 | 7 | 1.72 (0.77, 3.85) |
| Alachlor | 187 | 25 | 1.61 (1.05, 2.48) |
| Propachlor | 37 | 5 | 1.62 (0.63, 4.16) |
| Atrazine | 150 | 15 | 1.21 (0.71, 2.08) |
| Cyanazine | 56 | 7 | 1.60 (0.72, 3.53) |
| 2,4-D | 182 | 17 | 1.08 (0.65, 1.79) |
| Paraquat | 20 | 6 | 3.48 (1.37, 8.81) |
| Butylate | 55 | 9 | 2.01 (0.98, 4.10) |
| Trifluralin | 180 | 20 | 1.37 (0.85, 2.19) |
| Glyphosate | 36 | 5 | 1.72 (0.67, 4.42) |
Adjusted for age, sex, state, education, marital status, smoking status, alcohol consumption, history of head injury, and history of depression.
Participants reported high pesticide exposure events but not the specific pesticide involved in the highest exposed event were excluded from all analyses (n= 104). Estimates are reported only for pesticide groups or individual pesticides with at least five exposed cases.
We also found similar results in analyses using updated event exposures at Phase 2 and 3 (Table 4). Compared with farmers without a high pesticide exposure event, those with the events before the Phase 2 survey had a 74% higher odds of reporting dream enacting behaviors first noticed within 10 years before the Phase 4 survey (OR=1.74; 95%CI 1.36–2.22). Similarly, a history of high pesticide exposure events before the Phase 3 survey was associated with a 92% higher odds of reporting dream enacting behaviors first noticed within 5 years before the Phase 4 survey (OR=1.92; 95%CI 1.39–2.65).
Table 4.
High pesticide exposure events up to the Phase 2 (1999–2003) and Phase 3 (2005–2010) surveys in relation to dream enacting behaviors reported in the Phase 4 survey (2013–2015)
| High pesticide exposure events | Dream enacting behaviors |
Odds ratio (95% confidence interval) a |
|
|---|---|---|---|
| Never | Ever | ||
|
| |||
| Events up to Phase 2 and dream enacting behaviors onset within 10 years before the Phase 4 survey (n=17,030) | |||
| Neverb | 6,719 (40.8) | 206 (35.7) | 1.00 (Ref) |
| Everc | 2,039 (12.4) | 104 (18.0) | 1.74 (1.36, 2.22) |
| Uncertaind | 7,695 (46.8) | 267 (46.3) | 1.14 (0.95, 1.38) |
|
| |||
| Events up to Phase 3 and dream enacting behaviors onset within 5 years before the Phase 4 survey (n=17,823) | |||
| Neverb | 3,823 (22.0) | 71 (16.8) | 1.00 (Ref) |
| Everc | 2,411 (13.9) | 84 (19.9) | 1.92 (1.39, 2.65) |
| Uncertaind | 11,166 (64.2) | 268 (63.4) | 1.27 (0.98, 1.66) |
Adjusted for age, sex, race, state, education, marital status, smoking status, alcohol consumption and history of head injury, and history of depression.
Reference group, farmers who consistently reported no high pesticide exposure events across surveys.
Exposed, farmers who had reported high pesticide exposure event at least once in the surveys.
Exposure uncertain, farmers who reported never in at least one survey but missing on others, so impossible to classify them as no or exposed.
Discussion
To the best of our knowledge, this is the first epidemiological study to examine the association between unusually high pesticide exposures and dream enacting behaviors. We found that, among U.S. farmers, a history of high pesticide exposure events was associated with an increased odds of reporting dream enacting behaviors two decades later. Further, our data are consistent with the possibility that a longer delay between the event and subsequent washing may further increase the risk, and that exposures via inhalation and ingestion may be associated with higher risk than dermal exposures only. Finally, we identified that multiple individual pesticides involved in the highest exposed event might contribute to dream enacting behaviors in farmers. We conducted several sensitivity analyses which uniformly support the robustness of our primary findings.
RBD is by far the most specific prodromal marker of α-synucleinopathies23, and RBD onset can predate the clinical diagnosis of these diseases by years if not decades. In a meta-analysis of 3,262 polysomnography-confirmed RBD patients, about 30% developed neurodegenerative diseases at five years of follow-up, 82.4% at 10.5 years, and 96.6% at 14 years8. This relatively well-defined temporal relationship offers a window of opportunity for a better understanding of the prolonged and complex prodromal neurodegeneration, during which many environmental factors may come into play by initiating neurodegenerative pathology or modifying its progression. This is especially true for pesticides. Occupational pesticide exposure is among the most strongly suspected environmental risk factors for neurodegenerative synucleinopathies such as PD17; however, we have limited knowledge on the critical windows of exposure, or the specific pesticides that contribute to the pathogenesis. Findings from this study, together with our recent findings that both high and occupational exposures to pesticides are associated with olfactory impairment22,24, suggest that pesticides may contribute to PD pathogenesis even in its earliest stages. Besides PD, RBD is also a prodromal feature of less common synucleinopathies such as dementia with Lewy bodies and multiple system atrophy; however, the potential roles of pesticides in these diseases have been rarely investigated.
To our knowledge, besides our analyses in the Agricultural Health Study, only three studies have explored the potential role of pesticides in RBD, and data are suggestive of a positive association with some inconsistency. In a hospital-based case-control study of 347 polysomnography-confirmed RBD cases and 347 controls, RBD cases were more likely to report occupational use of insecticides (OR=3.67, 95%CI: 1.42–9.30) and herbicides (OR=2.54, 95%CI: 1.05–6.16), but not non-occupational uses of pesticides (OR=1.04, 95%CI: 0.63–1.70)11. Two other population-based cross-sectional analyses were conducted based on probable RBD identified by screening tools. The first (n=3,635) reported a non-statistically significant association between ever-use of pesticides and probable RBD (OR=2.31, 95%CI: 0.70–7.61)16 whereas the second study (n=7,225) reported a positive association of probable RBD with non-occupational pesticide exposure (OR=2.21, 95% CI: 1.40–3.50) but not with occupational use (OR=0.57, 95%CI: 0.26–1.26)14. Importantly, in all three studies, pesticide exposures were queried with simple “yes or no” questions and at the same time as the outcome was assessed. Exposure misclassification and recall bias thus cannot be excluded.
In contrast, the current study is much larger, and pesticide exposures were assessed about two decades before the outcome ascertainment in the primary analyses and thus were less likely to be differentially reported. More importantly, study participants were licensed pesticide applicators who have been shown to be able to provide reliable and detailed information on pesticide use25. With these attributes, we were able to comprehensively assess the association of pesticide exposures with probable RBD in much more detail than previous studies. It is, however, important to note that we used the presence of dream enacting behaviors as a surrogate for probable RBD. Although we used a validated screener for probable RBD3 and similar approaches have been widely used in other epidemiological studies2,12,16, measurement errors are inevitable23. Nevertheless, because the outcome was assessed about two decades after the exposure of interest, we expect this misclassification would be biased towards the null and thus might have underestimated the true association of pesticides with RBD.
In this cohort, we previously reported that a history of dream enacting behaviors was modestly associated with ever use of 14 specific pesticides, for example, DDT and permethrin used on poultry/livestock15. In the current study, we investigated the potential effects of unusually high exposures to pesticides. While these exposures are rare in the general populations, they were reported by 16.4% of farmers in our analysis, and most were able to recall the event details. This exposure has been analyzed previously and showed associations with neurologic symptoms21, depression26, asthma27, and olfactory impairment21. Therefore, this data collection presents a unique opportunity to examine the role of high pesticide exposure in dream enacting behaviors and RBD.
In this study, we demonstrated a robust association between a history of high pesticide exposure events and dream enacting behaviors. Two specific findings deserve attention. First, compared with the reference group, the association appears to be stronger for longer delays between the exposure and washing with soap and water, supporting a protective effect of good hygiene practices following high exposures to hazardous chemicals. The only exception was a moderate OR of 1.71 for those who washed within 30 mins of the accident. How long one waits to wash after a high exposure event could reflect the severity of the exposure, and thus immediate washing might indicate a more severe exposure. This could explain the higher OR associated with washing within 30 minutes vs. 30 minutes to 1 hour. Another interesting observation is that the association appears to be modestly stronger for exposure events involving the respiratory/digestive tract than for dermal-only exposure events. These preliminary findings warrant further investigations.
For most of the analyses on specific pesticides involved in the highest exposed event, we had a small number of exposed cases. Nevertheless, we found notable associations of dream enacting behaviors with 4 of the 8 pesticides groups and 8 of the 16 specific chemicals that we evaluated. Of all chemicals, organophosphate insecticides and chloroacetanilide herbicides were most frequently reported for such events, and both were associated with higher odds of reporting dream enacting behaviors 20 years later. Organochlorine insecticides and fungicides were also associated with higher odds of reporting dream enacting behaviors. In contrast, the other two commonly named chemical groups (i.e., triazine and phenoxy) involved in the highest exposed events, along with carbamate insecticides and fumigants, were not statistically associated with dream enacting behaviors.
Both of the organochlorines we analyzed, DDT and lindane showed positive associations with dream enacting behaviors. Previous studies also reported associations of organochlorine insecticides with PD28,29 and dementia30,31. In our analysis, 4 out of the 5 organophosphates demonstrated strong associations with ORs ≥ 2.90. Organophosphate insecticides are still widely used in the U.S., but compared to organochlorines, they are less likely to persist in the environment and to bioaccumulate. They are, however, known for delayed neurotoxicity following acute high exposures like poisoning32. Occupational use of organophosphates has been linked to neurological impairment33 and dementia31,32,34, but the evidence on PD is mixed32. Of the two chloroacetanilide herbicides we analyzed, both alachlor and propachlor had an OR above 1.60, but one was statistically significant while the other was not. The epidemiological evidence on chloroacetanilide and neurological diseases is sparse, but two ecological studies linked PD to levels of alachlor/metolachlor in groundwater35 and occupational use of alachlor36. Further, in our recent analysis, we found that high exposures to alachlor or metolachlor were both associated with olfactory impairment22, supporting roles of chloroacetanilides in the early stages of neurodegeneration. Besides those chemical groups, we also found a statistically significant association with paraquat, an herbicide that has been related to PD in most, albeit not all, of the available studies37–42. In summary, our findings on specific pesticides are provocative and generally consistent with existing knowledge about their potential roles in neurodegenerative diseases. The analyses, however, were based on limited sample sizes and thus need independent confirmation.
The present study has several limitations. First, using dream enacting behaviors symptomatic screening could have yielded substantial false positives for RBD. We thus cautiously used the term dream enacting behaviors rather than probable RBD in this study, and we encourage further research to reduce false positives in population-based studies43. Second, although we asked about the time of onset of dream enacting behaviors and the frequency of occurrence, the accuracy of this information is uncertain. Third, although the high pesticide exposure events have been studied in multiple previous analyses, there is no standardized way to assess unusual pesticide exposures in population-based studies. Therefore, participants’ responses to our exposure question were subject to their interpretations. Fourth, only ~61% of eligible participants at enrollment completed the Phase 4 survey that asked about dream enacting behaviors 20 years later. Nevertheless, we conducted sensitivity analyses using inverse probability weighting to account for potential selection biases and found similar results. Fifth, we have made efforts to adjust for cumulative lifetime days of pesticide use and other potential confounders. However, research on potential risk factors for RBD is limited, and we cannot exclude the possibility of residual confounding. Finally, as our study participants were predominantly male farmers, the results may not be readily generalizable to populations of other demographics or those with less pesticide exposure.
In conclusion, the current study suggests that pesticide exposure may be associated with a higher risk of dream enacting behaviors and RBD. If these findings are confirmed, future studies should investigate how pesticides contribute to RBD and synucleinopathies in their earliest stages.
Supplementary Material
Funding sources:
This study was supported by the National Institute of Environmental Health Sciences (R01ES029227), and in part by the intramural research program of the National Institutes of Health, the National Institute of Environmental Health Sciences (Z01-ES-049030), and the National Cancer Institute (Z01-CP-010119).
Footnotes
Declaration of interests: The authors have no conflicts of interest to disclose. All authors have declared no financial relationships with any organizations that might have an interest in the submitted work in the previous three years and no other relationships or activities that could appear to have influenced the submitted work.
Financial Disclosures of all authors: none
Ethical approval: This study was approved by the institutional review boards of the National Institute of Environmental Health Sciences and the National Cancer Institute. This specific secondary data analysis was exempted by the institutional review board of Michigan State University.
Contributor Information
Yaqun Yuan, Department of Epidemiology and Biostatistics, College of Human Medicine, Michigan State University, East Lansing, MI, USA.
Srishti Shrestha, The Memory Impairment and Neurodegenerative Dementia Center, University of Mississippi Medical Center, Jackson, MS, USA.
Zhehui Luo, Department of Epidemiology and Biostatistics, College of Human Medicine, Michigan State University, East Lansing, MI, USA.
Chenxi Li, Department of Epidemiology and Biostatistics, College of Human Medicine, Michigan State University, East Lansing, MI, USA.
Brenda L. Plassman, Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, NC, USA.
Christine G. Parks, Epidemiology Branch, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, USA.
Jonathan N. Hofmann, Occupational and Environmental Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Rockville, MD, USA.
Laura E. Beane Freeman, Occupational and Environmental Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Rockville, MD, USA.
Dale P. Sandler, Epidemiology Branch, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, USA.
Honglei Chen, Department of Epidemiology and Biostatistics, College of Human Medicine, Michigan State University, East Lansing, MI, USA.
Reference
- 1.Sateia MJ. International Classification of Sleep Disorders-Third Edition. Chest. 2014;146(5):1387–1394. doi: 10.1378/chest.14-0970 [DOI] [PubMed] [Google Scholar]
- 2.Mahlknecht P, Seppi K, Frauscher B, et al. Probable RBD and association with neurodegenerative disease markers: A population-based study. Movement Disord. 2015;30(10):1417–1421. doi: 10.1002/mds.26350 [DOI] [PubMed] [Google Scholar]
- 3.Postuma RB, Arnulf I, Hogl B, et al. A single-question screen for rapid eye movement sleep behavior disorder: A multicenter validation study. Movement Disord. 2012;27(7):913–916. doi: 10.1002/mds.25037 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kang SH, Yoon IY, Lee SD, Han JW, Kim TH, Kim KW. REM Sleep Behavior Disorder in the Korean Elderly Population: Prevalence and Clinical Characteristics. Sleep. 2013;36(8):1147–1152. doi: 10.5665/sleep.2874 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Haba-Rubio J, Frauscher B, Marques-Vidal P, et al. Prevalence and determinants of rapid eye movement sleep behavior disorder in the general population. Sleep. 2017;41(2). doi: 10.1093/sleep/zsx197 [DOI] [PubMed] [Google Scholar]
- 6.Pujol M, Pujol J, Alonso T, et al. Idiopathic REM sleep behavior disorder in the elderly Spanish community: a primary care center study with a two-stage design using video-polysomnography. Sleep Med. 2017;40:116–121. doi: 10.1016/j.sleep.2017.07.021 [DOI] [PubMed] [Google Scholar]
- 7.Howell MJ, Schenck CH. Rapid Eye Movement Sleep Behavior Disorder and Neurodegenerative Disease. Jama Neurol. 2015;72(6):707–712. doi: 10.1001/jamaneurol.2014.4563 [DOI] [PubMed] [Google Scholar]
- 8.Galbiati A, Verga L, Giora E, Zucconi M, Ferini-Strambi L. The risk of neurodegeneration in REM sleep behavior disorder: A systematic review and meta-analysis of longitudinal studies. Sleep Med Rev. 2019;43:37–46. doi: 10.1016/j.smrv.2018.09.008 [DOI] [PubMed] [Google Scholar]
- 9.Frauscher B, Jennum P, Ju YES, et al. Comorbidity and medication in REM sleep behavior disorder. Neurology. 2014;82(12):1076–1079. doi: 10.1212/wnl.0000000000000247 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Dauvilliers Y, Postuma RB, Ferini-Strambi L, et al. Family history of idiopathic REM behavior disorder. Neurology. 2013;80(24):2233–2235. doi: 10.1212/wnl.0b013e318296e967 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Postuma RB, Montplaisir JY, Pelletier A, et al. Environmental risk factors for REM sleep behavior disorder. Neurology. 2012;79(5):428–434. doi: 10.1212/wnl.0b013e31825dd383 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wong JC, Li J, Pavlova M, et al. Risk factors for probable REM sleep behavior disorder. Neurology. 2016;86(14):1306–1312. doi: 10.1212/wnl.0000000000002414 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Yao C, Fereshtehnejad SM, Keezer MR, Wolfson C, Pelletier A, Postuma RB. Risk factors for possible REM sleep behavior disorder: A CLSA population-based cohort study. Neurology. 2018;92(5):e475–e485. doi: 10.1212/wnl.0000000000006849 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Zhang H, Gu Z, Yao C, et al. Risk factors for possible REM sleep behavior disorders: A community-based study in Beijing. Neurology. 2020;95(16):e2214–e2224. doi: 10.1212/wnl.0000000000010610 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Shrestha S, Kamel F, Umbach DM, et al. Factors associated with dream enacting behaviors among US farmers. Parkinsonism Relat D. 2018;57:9–15. doi: 10.1016/j.parkreldis.2018.07.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ma JF, Qiao Y, Gao X, et al. A community-based study of risk factors for probable rapid eye movement sleep behavior disorder. Sleep Med. 2017;30:71–76. doi: 10.1016/j.sleep.2016.06.027 [DOI] [PubMed] [Google Scholar]
- 17.Ascherio A, Schwarzschild MA. The epidemiology of Parkinson’s disease: risk factors and prevention. Lancet Neurology. 2016;15(12):1257–1272. doi: 10.1016/s1474-4422(16)30230-7 [DOI] [PubMed] [Google Scholar]
- 18.Alavanja MC, Sandler DP, McDonnell CJ, et al. Characteristics of persons who self-reported a high pesticide exposure event in the Agricultural Health Study. Environ Res. 1999;80(2):180–186. doi: 10.1006/enrs.1998.3887 [DOI] [PubMed] [Google Scholar]
- 19.Alavanja MC, Sandler DP, McMaster SB, et al. The Agricultural Health Study. Environ Health Persp. 1996;104(4):362–369. doi: 10.1289/ehp.96104362 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Shrestha S, Kamel F, Umbach DM, et al. Nonmotor symptoms and Parkinson disease in United States farmers and spouses. Plos One. 2017;12(9):e0185510. doi: 10.1371/journal.pone.0185510 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kamel F, Engel LS, Gladen BC, Hoppin JA, Alavanja MC, Sandler DP. Neurologic symptoms in licensed private pesticide applicators in the agricultural health study. Environ Health Persp. 2005;113(7):877–882. doi: 10.1289/ehp.7645 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Shrestha S, Kamel F, Umbach DM, et al. High Pesticide Exposure Events and Olfactory Impairment among U.S. Farmers. Environ Health Persp. 2019;127(1):17005. doi: 10.1289/ehp3713 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Dauvilliers Y, Schenck CH, Postuma RB, et al. REM sleep behaviour disorder. Nat Rev Dis Primers. 2018;4(1):19. doi: 10.1038/s41572-018-0016-5 [DOI] [PubMed] [Google Scholar]
- 24.Shrestha S, Umbach DM, Freeman LEB, et al. Occupational pesticide use and self-reported olfactory impairment in US farmers. Occup Environ Med. 2021;78(3):179–191. doi: 10.1136/oemed-2020-106818 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Blair A, Tarone R, Sandler D, et al. Reliability of Reporting on Life-Style and Agricultural Factors by a Sample of Participants in the Agricultural Health Study from Iowa. Epidemiology. 2002;13(1):94–99. doi: 10.1097/00001648-200201000-00015 [DOI] [PubMed] [Google Scholar]
- 26.Beseler CL, Stallones L, Hoppin JA, et al. Depression and pesticide exposures among private pesticide applicators enrolled in the Agricultural Health Study. Environ Health Persp. 2008;116(12):1713–1719. doi: 10.1289/ehp.11091 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Keim SA, Alavanja MC. Pesticide use by persons who reported a high pesticide exposure event in the agricultural health study. Environ Res. 2001;85(3):256–259. doi: 10.1006/enrs.2000.4224 [DOI] [PubMed] [Google Scholar]
- 28.Hancock DB, Martin ER, Mayhew GM, et al. Pesticide exposure and risk of Parkinson’s disease: A family-based case-control study. Bmc Neurol. 2008;8(1):6. doi: 10.1186/1471-2377-8-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Elbaz A, Clavel J, Rathouz PJ, et al. Professional exposure to pesticides and Parkinson disease. Ann Neurol. 2009;66(4):494–504. doi: 10.1002/ana.21717 [DOI] [PubMed] [Google Scholar]
- 30.Richardson JR, Roy A, Shalat SL, et al. Elevated Serum Pesticide Levels and Risk for Alzheimer Disease. Jama Neurol. 2014;71(3):284–290. doi: 10.1001/jamaneurol.2013.6030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hayden KM, Norton MC, Darcey D, et al. Occupational exposure to pesticides increases the risk of incident AD. Neurology. 2010;74(19):1524–1530. doi: 10.1212/wnl.0b013e3181dd4423 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sánchez-Santed F, Colomina MT, Hernández EH. Organophosphate pesticide exposure and neurodegeneration. Cortex. 2016;74:417–426. doi: 10.1016/j.cortex.2015.10.003 [DOI] [PubMed] [Google Scholar]
- 33.Richardson JR, Fitsanakis V, Westerink RHS, Kanthasamy AG. Neurotoxicity of pesticides. Acta Neuropathol. 2019;138(3):343–362. doi: 10.1007/s00401-019-02033-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Parrón T, Requena M, Hernández AF, Alarcón R. Association between environmental exposure to pesticides and neurodegenerative diseases. Toxicol Appl Pharm. 2011;256(3):379–385. doi: 10.1016/j.taap.2011.05.006 [DOI] [PubMed] [Google Scholar]
- 35.James KA, Hall DA. Groundwater Pesticide Levels and the Association With Parkinson Disease. Int J Toxicol. 2015;34(3):266–273. doi: 10.1177/1091581815583561 [DOI] [PubMed] [Google Scholar]
- 36.Wan N, Lin G. Parkinson’s Disease and Pesticides Exposure: New Findings From a Comprehensive Study in Nebraska, USA. J Rural Heal. 2016;32(3):303–313. doi: 10.1111/jrh.12154 [DOI] [PubMed] [Google Scholar]
- 37.Manning-Bog AB, McCormack AL, Li J, Uversky VN, Fink AL, Monte DAD. The Herbicide Paraquat Causes Up-regulation and Aggregation of α-Synuclein in Mice PARAQUAT AND α-SYNUCLEIN*. J Biol Chem. 2002;277(3):1641–1644. doi: 10.1074/jbc.c100560200 [DOI] [PubMed] [Google Scholar]
- 38.McCormack AL, Atienza JG, Johnston LC, Andersen JK, Vu S, Monte DAD. Role of oxidative stress in paraquat-induced dopaminergic cell degeneration. J Neurochem. 2005;93(4):1030–1037. doi: 10.1111/j.1471-4159.2005.03088.x [DOI] [PubMed] [Google Scholar]
- 39.Purisai MG, McCormack AL, Cumine S, Li J, Isla MZ, Monte DAD. Microglial activation as a priming event leading to paraquat-induced dopaminergic cell degeneration. Neurobiol Dis. 2007;25(2):392–400. doi: 10.1016/j.nbd.2006.10.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Tanner CM, Kamel F, Ross GW, et al. Rotenone, Paraquat, and Parkinson’s Disease. Environ Health Persp. 2011;119(6):866–872. doi: 10.1289/ehp.1002839 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Shrestha S, Parks CG, Umbach DM, et al. Pesticide use and incident Parkinson’s disease in a cohort of farmers and their spouses. Environ Res. 2020;191:110186. doi: 10.1016/j.envres.2020.110186 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Lee PC, Bordelon Y, Bronstein J, Ritz B. Traumatic brain injury, paraquat exposure, and their relationship to Parkinson disease. Neurology. 2012;79(20):2061–2066. doi: 10.1212/wnl.0b013e3182749f28 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Postuma RB, Pelletier A, Berg D, Gagnon JF, Escudier F, Montplaisir J. Screening for prodromal Parkinson’s disease in the general community: a sleep-based approach. Sleep Med. 2016;21:101–105. doi: 10.1016/j.sleep.2015.12.016 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.