Skip Navigation


American Journal of Epidemiology Advance Access originally published online on November 20, 2006
American Journal of Epidemiology 2007 165(4):364-374; doi:10.1093/aje/kwk024
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
165/4/364    most recent
kwk024v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Kamel, F
Right arrow Articles by Sandler, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kamel, F
Right arrow Articles by Sandler, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology Copyright © 2006 by the Johns Hopkins Bloomberg School of Public Health All rights reserved; printed in U.S.A.

ORIGINAL CONTRIBUTIONS

Pesticide Exposure and Self-reported Parkinson's Disease in the Agricultural Health Study

F Kamel1, CM Tanner2, DM Umbach1, JA Hoppin1, MCR Alavanja3, A Blair3, K Comyns2, SM Goldman2, M Korell2, JW Langston2, GW Ross4 and DP Sandler1

1 National Institute of Environmental Health Sciences, Research Triangle Park, NC
2 The Parkinson's Institute, Sunnyvale, CA
3 National Cancer Institute, Bethesda, MD
4 Pacific Islands VA Health Care System, Honolulu, HI

Correspondence to Dr. Freya Kamel, Epidemiology Branch, MD A3-05, National Institute of Environmental Health Sciences, P.O. Box 12233, Research Triangle Park, NC 27709 (e-mail: kamel{at}mail.nih.gov).

Received for publication December 22, 2005. Accepted for publication July 13, 2006.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Previous studies based on limited exposure assessment have suggested that Parkinson's disease (PD) is associated with pesticide exposure. The authors used data obtained from licensed private pesticide applicators and spouses participating in the Agricultural Health Study to evaluate the relation of self-reported PD to pesticide exposure. Cohort members, who were enrolled in 1993–1997, provided detailed information on lifetime pesticide use. At follow-up in 1999–2003, 68% of the cohort was interviewed. Cases were defined as participants who reported physician-diagnosed PD at enrollment (prevalent cases, n = 83) or follow-up (incident cases, n = 78). Cases were compared with cohort members who did not report PD (n = 79,557 at enrollment and n = 55,931 at follow-up). Incident PD was associated with cumulative days of pesticide use at enrollment (for highest quartile vs. lowest, odds ratio (OR) = 2.3, 95% confidence interval: 1.2, 4.5; p-trend = 0.009), with personally applying pesticides more than half the time (OR = 1.9, 95% confidence interval: 0.7, 4.7), and with some specific pesticides (ORs ≥ 1.4). Prevalent PD was not associated with overall pesticide use. This study suggests that exposure to certain pesticides may increase PD risk. Findings for specific chemicals may provide fruitful leads for further investigation.

case-control studies; herbicides; insecticides; Parkinson disease; pesticides

Abbreviations: AHS, Agricultural Health Study; OR, odds ratio; PD, Parkinson's disease


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Parkinson's disease (PD) is a common neurodegenerative disease, affecting up to a million people in the United States (1). Its etiology is uncertain but probably involves both genetic and environmental components (1). Numerous studies have established the role of specific genes in familial forms of PD (2). Mutations in these genes are rare in the more common sporadic form of PD, however, and twin studies have suggested a strong environmental component to sporadic PD (3, 4).

A consistent finding of epidemiologic studies is an association of PD with rural living or farming as an occupation (1, 5). Although this association may reflect several exposures, many studies have focused on pesticides. Overall evidence suggests that PD is associated with pesticide exposure (6–10), although it may not yet support a causal inference (9, 10). Variable results may be due to differences in study design, particularly in methods used for exposure assessment (8–10). Most studies have evaluated exposure to any pesticide or to broad functional classes such as insecticides or herbicides, and little information is available for specific chemicals or factors that may affect exposure, such as use of personal protective equipment. Few investigators have assessed exposure prospectively or been able to separate prevalent and incident cases.

The Agricultural Health Study (AHS) is a large cohort study of licensed pesticide applicators and their spouses (11). Cohort members provided detailed information on lifetime pesticide use at enrollment. They reported physician-diagnosed PD either at enrollment or during a follow-up interview conducted 5 years later. We used data from these self-reports to evaluate the association of PD with pesticide exposure.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Population and questionnaires
The AHS cohort was established in 1993–1997. Persons applying for certification to use restricted-use pesticides in Iowa or North Carolina were enrolled (11). Figure 1 summarizes the process of data collection. At enrollment, 52,393 private applicators (84 percent of those eligible), mostly farmers, completed a self-administered Enrollment Questionnaire. All enrolled applicators were requested to complete a supplemental self-administered Applicator Questionnaire, and 22,915 (44 percent) complied. Applicators completed the two questionnaires a median of 1 month apart. Based on data from the Enrollment Questionnaire, persons who completed the Applicator Questionnaire were similar in most respects to those who did not, including pesticide exposure (12). Most (43,692; 83 percent) of the enrolled private applicators were married, and 32,345 (74 percent) of their spouses enrolled in the study by completing a Spouse Questionnaire, either self-administered (80 percent) or by telephone (20 percent). The Enrollment and Spouse questionnaires elicited information on pesticide use, demographic factors, lifestyle, and medical history, including physician-diagnosed PD. The Applicator Questionnaire collected additional information on pesticide exposure and medical history, including physician-diagnosed PD and neurologic symptoms.


Figure 1
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 1. Data collection in the Agricultural Health Study (AHS). At enrollment (1993–1997), licensed private pesticide applicators in Iowa and North Carolina completed a self-administered Enrollment Questionnaire. A subset (44%) also completed a self-administered supplemental Applicator Questionnaire. Their spouses completed a Spouse Questionnaire, either self-administered (80%) or by telephone (20%). At follow-up (1999–2003), cohort members were interviewed by telephone. Q, Questionnaire.

 
Five years after enrollment, we conducted follow-up telephone interviews with 57,251 cohort members (68 percent). Thirteen percent of the cohort could not be contacted, 12 percent declined the interview, 3 percent were excluded for various other reasons, 2.5 percent were deceased, and the remainder were not interviewed because of illness (<1 percent), language difficulties (<1 percent), or other reasons (1 percent). Cohort members who were interviewed, compared with those who were not, were slightly younger (16 percent vs. 18 percent were more than 60 years old) and were more likely to be spouses (42 percent vs. 31 percent), to be from Iowa (66 percent vs. 57 percent), to have more than a high school education (48 percent vs. 41 percent), and to have never smoked (62 percent vs. 56 percent). The proportions who had ever personally mixed or applied pesticides were similar (82 percent vs. 83 percent). Information from the follow-up interviews used for the present analysis was self-reported physician-diagnosed PD, age at diagnosis, age at interview, and smoking status.

The institutional review boards of the National Institutes of Health (Bethesda, Maryland), Westat, Inc. (Rockville, Maryland; Coordinating Center), the University of Iowa (Iowa City, Iowa; Iowa Field Station), and Battelle (Durham, North Carolina; North Carolina Field Station) approved the AHS. At enrollment, the study was explained to potential participants, who indicated consent by returning questionnaires; consent for follow-up interviews was indicated by participation. Questionnaires are available at the AHS website (www.aghealth.org/questionnaires.html).

Case definition
At enrollment, cohort members were asked, "Has a doctor ever told you that you had been diagnosed with Parkinson's disease?" This question was included in the Enrollment, Applicator, and Spouse questionnaires, so applicators who completed both the Enrollment Questionnaire and the Applicator Questionnaire had two opportunities to respond. At follow-up, cohort members were asked, "Has a doctor or other health professional ever told you that you had Parkinson's disease?" We excluded from the analysis any applicator who provided conflicting information on PD on the Enrollment and Applicator questionnaires (i.e., responded "yes" on one and "no" on the other; n = 12). Applicators who had missing information on one questionnaire but responded to the other were included. We also excluded applicators and spouses who responded "yes" to the question on PD at enrollment and "no" in the follow-up interview (n = 13).

We defined prevalent PD on the basis of self-report at enrollment. We first excluded persons with missing information on PD at enrollment (n = 5,073). We then defined prevalent cases as applicators who reported PD on either the Enrollment Questionnaire or the Applicator Questionnaire (or both) or spouses who reported PD on the Spouse Questionnaire. Controls were participants who responded negatively to the PD question. There were 83 prevalent cases (60 applicators and 23 spouses) and 79,557 controls.

We defined incident PD on the basis of self-report at follow-up. We first excluded prevalent cases (n = 83) and persons with missing information on PD, either because they were not interviewed or because they did not respond to the PD question (n = 28,621). We then defined incident cases as applicators and spouses who reported PD in the telephone interview; controls were those who did not. There were 78 incident cases (56 applicators and 22 spouses) and 55,931 controls. This definition included eight cases and 2,675 controls whose PD status at enrollment was unknown because of missing data but who provided information at follow-up. Excluding these participants from the analyses did not change results.

Pesticide exposure
More than 99 percent of applicators and 56 percent of spouses had personally mixed or applied pesticides. Both applicators and spouses reported duration (years) and frequency (days per year) of use of any pesticide, as well as ever use of 50 commonly used pesticides. Cumulative days of use of any pesticide was the product of duration and frequency, categorized in quartiles.

For applicators, information on pesticide-related medical attention was collected using the question, "As a result of using pesticides, how often have you seen a doctor or been hospitalized?" Responses were dichotomized as ever/never. Information on events involving high personal pesticide exposure was collected using the question, "Have you ever had an incident or experience while using any type of pesticide which caused you unusually high personal exposure?" Persons who gave positive responses were asked about the time interval between the incident and washing with soap and water. Responses were categorized as "no event," "washed within 1 hour," or "washed after 1 hour."

Applicators provided information on use of six types of personal protective equipment: fabric or leather gloves, chemical-resistant gloves, face shield or goggles, cartridge respirator or gas mask, disposable outer clothing, and other. We constructed a personal protective equipment score according to the method of Dosemeci et al. (13), by assigning points for use of each type or combination of equipment based on its potential to decrease exposure. Scores ranged from 0.1 to 1.0; a lower score indicated greater protection (lower exposure). For example, the score for use of chemical-resistant gloves was 0.4. Scores were used in two ways: as an independent variable after categorization into tertiles or to weight cumulative days of pesticide use.

Data analysis
We used logistic regression to evaluate the relation of either prevalent PD or incident PD to general pesticide variables. We constructed models with applicators and spouses combined and for the two groups separately. Models included adjustment for 1) age at enrollment, using a four-level categorical variable with persons aged 51–60 years designated the referent group; 2) state (Iowa or North Carolina); and 3) type of participant (applicator or spouse), in models including both. Most applicators were men (>99 percent) and most spouses were women (96 percent), so no further adjustment was made for gender. Neither race nor education confounded the relation of PD with pesticide exposure (<10 percent change in estimate), so these factors were not included in the final models. Additional models were created after stratification by state or adjustment for cigarette smoking. Neither strategy had a substantive effect on risk estimates; selected results from these models are presented.

We used two-stage hierarchical logistic regression to increase precision when evaluating multiple specific pesticides (14). The first-stage model contained covariates and indicators for specific pesticides with at least four exposed cases. The second-stage model included variables for functional pesticide groups (insecticides, herbicides, fungicides, fumigants) and several chemical groups (organophosphates, organochlorines, carbamates, phenoxyacetates, triazines/triazones); all groups included at least three pesticides. We assumed that the odds ratio for each pesticide would fall within a 10-fold range around its prior mean, with 95 percent certainty, by defining the prior residual variance as 0.35.

Odds ratios and 95 percent confidence intervals are reported. Two-sided p-trend values for cumulative days of pesticide use were calculated using a continuous variable defined by the midpoints of the levels in the categorical variable. We used enrollment data from the September 2002 AHS data prerelease and follow-up data from the AHS data release P2REL0312.01 (unpublished data). Analyses were performed using version 8.2 of SAS (SAS Institute, Inc., Cary, North Carolina).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Demographic and lifestyle characteristics
Both prevalent PD and incident PD were associated with older age at enrollment; the relation was stronger for prevalent PD (table 1). Incident PD was also associated with greater age at follow-up (for age ≥71 years, odds ratio (OR) = 4.4, 95 percent confidence interval: 2.4, 8.1). For both prevalent PD and incident PD, risk was lower in North Carolina than in Iowa, lower for spouses (96 percent women) than for applicators (99 percent men), and higher for non-Whites than for Whites (table 1). Risk was also lower in North Carolina in models including cigarette smoking or pesticide use (not shown). Five of the nine non-White cases (prevalent and incident) and approximately 50 percent of non-White controls were African-American. Using enrollment data, prevalent PD was inversely associated with both former and current cigarette smoking, and a gradient was observed for pack-years of smoking, but incident PD was inversely associated only with current smoking, not with former smoking (table 1). Incident PD was also inversely associated with current but not former smoking using follow-up data on smoking status (not shown). Prevalent, but not incident, PD was inversely associated with alcohol consumption at enrollment after adjustment for smoking (table 1). Similar results for demographic and lifestyle characteristics were observed in both applicators and spouses when these groups were considered separately (not shown).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Demographic and lifestyle characteristics of self-reported Parkinson's disease (PD) cases and controls identified in the Agricultural Health Study at enrollment in 1993–1997 (prevalent PD) or at follow-up in 1999–2003 (incident PD)*

 
Parkinsonian symptoms
Applicators who completed the supplemental Applicator Questionnaire provided information on several parkinsonian symptoms at enrollment—that is, at the same time prevalent cases reported PD and approximately 5 years before incident cases reported PD. Prevalent PD was associated with hand tremor, difficulty with balance, dizziness, difficulty speaking, and changes in sense of smell or taste (table 2). Increases in hand tremor and changes in smell or taste were also apparent at enrollment in persons who later reported PD at follow-up (incident PD cases) (table 2).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Symptoms reported at enrollment by self-reported Parkinson's disease (PD) cases and controls identified in the Agricultural Health Study at enrollment in 1993–1997 (prevalent PD) or at follow-up in 1999–2003 (incident PD)*

 
Exposure to any pesticide
There was a weak inverse association of prevalent PD with ever use of any pesticide and with personally mixing or applying pesticides (table 3). This relation was apparent only after adjustment for type of participant (applicator or spouse) but was seen in both groups (for ever use, OR = 0.4 for applicators and OR = 0.6 for spouses). In contrast, there was a positive association of incident PD with ever use of pesticides and with personally mixing pesticides, and there was a stronger association for those who personally applied pesticides more than half the time (table 3). Cumulative lifetime days of use was associated with a dose-related increase in incident PD but not prevalent PD (table 3).


View this table:
[in this window]
[in a new window]

  		TABLE 3. Pesticide use among self-reported Parkinson's disease (PD) cases and controls identified in the Agricultural Health Study at enrollment in 1993–1997 (prevalent PD) or at follow-up in 1999–2003 (incident PD)*

 
Similar results for cumulative lifetime days of pesticide use were seen when applicators and spouses were considered separately. For prevalent PD, odds ratios for the highest quartile were 0.8 for applicators and 0.6 for spouses; for incident PD, odds ratios were 2.2 and 2.8. Results for applicators were unchanged after exclusion of persons who never used pesticides (<1 percent); odds ratios for the highest quartile of use were 0.8 for prevalent PD and 2.1 for incident PD. Adjustment for smoking did not change results: In models including smoking status, odds ratios for the highest quartile of use were 0.8 for prevalent PD and 2.3 for incident PD. Results were similar in both states: For the highest quartile of use, odds ratios for prevalent PD were 0.9 in North Carolina and 0.7 in Iowa; odds ratios for incident PD were 2.7 in North Carolina and 2.0 in Iowa. Similar results were obtained when lifetime years of use (duration) rather than cumulative days of use (the product of frequency and duration) was used as a measure of exposure: Odds ratios for the highest quartile of years of use were 0.9 for prevalent PD and 2.5 for incident PD.

Use of personal protective equipment was inversely related to both prevalent and incident PD (table 3). Weighting cumulative days of pesticide use by the personal protective equipment score did not appreciably change the association of PD with pesticide use: Odds ratios for the highest quartile of weighted cumulative days of use were 0.7 for prevalent PD and 2.2 for incident PD. Weak positive associations of PD with pesticide-related medical care were evident for both prevalent and incident cases (table 3). No case in either group reported a diagnosis of pesticide poisoning. Delaying washing for more than 1 hour after an incident involving high personal pesticide exposure was associated with both prevalent and incident PD (table 3).

For prevalent PD, insufficient data were available to evaluate the temporal relation of pesticide exposure to disease. Among incident PD cases, 64 of 78 (85 percent) reported an age of diagnosis consistent with diagnosis after enrollment; the remaining 14 did not. Twelve of the latter had denied PD at enrollment, while the other two had missing data. Excluding these 14 cases did not substantively change results (for highest quartile of cumulative pesticide use, OR = 2.2, 95 percent confidence interval: 1.0, 4.5).

Information on AHS cohort mortality through June 2005 was available from the National Death Index. Death certificates for 61 persons listed PD as an underlying or contributing cause of death. Nineteen of these PD cases were prevalent cases, six were incident cases, and one case was excluded from both groups because of inconsistent data, leaving 35 potential PD cases identified only by death certificate. Compared with the 78 self-reported incident cases, the 35 death certificate cases were older at enrollment and more likely to have lived in North Carolina; the two groups were similar in terms of proportion of applicators versus spouses, gender, race, educational level, and smoking status. Associations with pesticide exposure for a combined group (n = 113) were similar to those for the self-reported incident cases in table 3 (for highest quartile of pesticide use, OR = 1.8). No association with smoking status was observed in the combined group (for former smokers, OR = 0.9; for current smokers, OR = 0.7).

Exposure to specific chemicals
Considering only chemicals for which there were four or more exposed cases, odds ratios for prevalent PD were elevated (≥1.4) for the herbicides pendimethalin, paraquat, and cyanazine and the fumigants carbon disulfide/carbon tetrachloride and ethylene dibromide (table 4). Odds ratios for incident PD were elevated for the herbicides dicamba, trifuralin, 2,4,5-trichlorophenoxyacetic acid (2,4,5-T), and butylate, the insecticides lindane and phorate, the fungicides chlorothalonil and benomyl, and the fumigant methyl bromide (table 4). Results were similar when the second-stage model of the hierarchical logistic regression included only variables for functional pesticide groups (not shown).


View this table:
[in this window]
[in a new window]

  		TABLE 4. Ever use of specific pesticides by self-reported Parkinson's disease (PD) cases and controls identified in the Agricultural Health Study at enrollment in 1993–1997 (prevalent PD) or at follow-up in 1999–2003 (incident PD)*

 
Applicators who completed the supplemental Applicator Questionnaire provided additional information on some pesticides, including four implicated in PD in previous studies: dieldrin, maneb, paraquat, and rotenone (8–10). When information from the Enrollment and Applicator questionnaires was combined, odds ratios for prevalent PD were 1.3 for dieldrin, 1.5 for maneb/mancozeb, 1.8 for paraquat, and 1.7 for rotenone, and odds ratios for incident PD were 0.9 for dieldrin, 2.1 for maneb/mancozeb, and 1.4 for paraquat; only one incident case had used rotenone. These results were based on 4–10 exposed cases for each pesticide.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
We found a positive association of self-reported PD with overall pesticide use among incident cases. Increasing cumulative days of use and proportion of time pesticides were personally applied were related to increasing risk for both applicators and their spouses. In contrast, prevalent PD was not associated with overall pesticide use. For both incident and prevalent PD, receiving pesticide-related medical care or experiencing an incident involving high personal pesticide exposure was associated with increased risk, while using personal protective equipment to reduce potential exposure was associated with decreased risk. Increased risk was also associated with several specific pesticides.

Previous studies have suggested that pesticide exposure may increase risk of PD (6–10). Studies published before 1999 were reviewed by Le Couteur et al. (6), who noted that 12 of 20 studies found a positive association, with 1.6- to 7-fold increases in risk. A meta-analysis published in 2000 found a combined odds ratio of 1.9 (7). Investigators in some recent studies have reported positive associations (15–22), although others have not (23–26); in general, studies with positive findings had more detailed exposure assessments. Researchers in three prospective cohort studies reported positive associations of PD with pesticide exposure (27–29). PD risk may depend on an interaction of genetic susceptibility with pesticide exposure: In two studies, persons who were either exposed to pesticides or had polymorphisms in the cytochrome P-450 2D6 (CYP2D6) gene that confer a poor-metabolizer phenotype had only a slightly increased risk of PD, but persons with both factors had a greatly increased risk (ORs = 4–8) (30, 31).

Information on the relation of PD to specific pesticides or pesticide classes is limited. Several studies found increased risk associated with exposure to either insecticides or herbicides (21, 22, 32, 33), and one study found elevated risks associated with organochlorine, organophosphate, and carbamate insecticides (34). Several studies have implicated the herbicide paraquat (22, 35, 36), which produced selective degeneration of dopaminergic nigral neurons in an animal model (37). Case reports and postmortem studies have described PD in persons exposed to organophosphate insecticides; organochlorine insecticides, including dieldrin; herbicides, including glyphosate, paraquat, and diquat; and fungicides, including maneb and other dithiocarbamates (8). Animal models have implicated rotenone (38) or a combination of paraquat and maneb (39) in the etiology of PD. Finally, certain pesticides promote synuclein fibrillation (40), a mechanism likely to be involved in PD pathogenesis (41). Thus, although some authors suggest that existing data may not yet be sufficient to infer a causal relation (9, 10), the weight of evidence suggests that pesticide exposure is associated with increased risk of PD (6–10).

Our study provides additional evidence to support the hypothesis that PD risk is associated with overall pesticide exposure, as well as new information on the relation of PD to specific pesticides. For prevalent cases, we found elevated risks associated with three herbicides and two fumigants. For incident cases, we found elevated risks associated with nine pesticides from all four functional classes. Four pesticides have previously been implicated in the etiology of PD: dieldrin, maneb, paraquat, and rotenone (8–10). We found associations of all four with prevalent PD and of maneb/mancozeb and paraquat with incident PD in the subset of applicators who completed the supplemental Applicator Questionnaire at enrollment. However, only an association of paraquat with prevalent PD was seen in the complete cohort.

Accounting for exposure to multiple pesticides presents a challenge for analysis. We used hierarchical logistic regression to evaluate exposure to multiple pesticides (14). Definition of meaningful pesticide subgroups is difficult. Our results suggest that categorization by functional group or even chemical structure may be unsatisfactory, because the relation with outcome may be variable within groups. In a recent study, Samanic et al. (42) used factor analysis to identify three groupings of pesticide use and other farm characteristics in the AHS cohort: Iowa agriculture and herbicide use; North Carolina agriculture and use of insecticides, fumigants, and fungicides; and older age and use of chlorinated pesticides. Present knowledge of mechanisms involved in PD is limited but may suggest ways to group chemicals for analysis (41, 43). Conversely, findings that chemicals affecting similar mechanistic pathways increase PD risk would provide strong evidence that these pathways are involved in the etiology of PD.

Differences in results for prevalent and incident cases are striking, but the underlying reasons are unclear. Prevalent cases may be a selected group, representing the more functional cases present in the underlying population when the cohort was enrolled. PD cases with higher pesticide exposure may have been less likely to enroll in the cohort than those with lower exposure, either because the disease course is different in the former or because other effects of pesticide exposure (44, 45) led them to stop farming at an earlier age. Our questionnaires focused on lifetime pesticide use, but prevalent PD cases may have decreased their current use as a result of disability and therefore underreported lifetime use. Sixty-eight percent of the original cohort was interviewed at follow-up. Although there were demographic differences between those interviewed and those not interviewed, pesticide use reported at enrollment was similar in the two groups. Loss of cases at follow-up due to mortality does not appear to account for differences between prevalent cases and incident cases. We lacked precise information on date of diagnosis for prevalent cases, so disease duration was unknown, as was period of pesticide use in relation to time of diagnosis. In contrast, most incident cases reported a disease duration of less than 5 years, and exposure data were collected before diagnosis.

Consistent with previous reports (1, 46), we found an inverse association of cigarette smoking with PD. The association was stronger in prevalent cases than in incident cases. A lower response rate for the follow-up interview among nonsmokers who developed PD after enrollment, compared with smokers who developed PD, might explain this difference. However, for the cohort as a whole, response at follow-up was greatest in never smokers, intermediate in former smokers, and least in current smokers, based on smoking information collected at enrollment, and smoking behavior was similar in the control groups compared with prevalent or incident cases. Furthermore, when potential PD cases identified by death certificate only were added to the incident case group, the association of PD with current smoking was weakened. Adjustment for smoking did not change associations of PD with pesticide use.

One limitation of this study was its reliance on self-reported diagnosis of PD, but the degree of misclassification may not have been large. Furthermore, misclassification is likely to result in attenuation of associations, making our estimates conservative. Information on neurologic symptoms, available for a subset of applicators, increased confidence in the diagnosis. Prevalent cases were highly likely to report several parkinsonian symptoms in comparison with controls. Interestingly, incident cases also reported high levels of hand tremor and changes in smell or taste at enrollment, 5 years before they reported a diagnosis of PD. A decline in olfaction may be an early symptom of PD (47), although changes may not be evident more than 5 years before diagnosis (48). A further limitation of the study was the use of self-reported exposure data. However, questionnaires are essentially the only way to evaluate chronic exposure to nonpersistent pesticides in nonindustrial populations. Previous studies have found that farmers in general and AHS cohort members in particular report pesticide use reliably (49, 50), although some misclassification undoubtedly occurs.

Strengths of the study included its ability to distinguish between prevalent and incident cases and its base in an agricultural population with many exposed persons. Because farming practices are considerably different in Iowa and North Carolina, the AHS cohort represents a diverse farming population (11). Cohort members provided detailed exposure data, allowing us to evaluate associations not only with overall pesticide use but also with specific chemicals and with factors that may affect exposure, such as use of personal protective equipment. We made internal comparisons of more exposed persons with less exposed persons from the same population, mitigating potential confounding. Future studies will continue to exploit the unique opportunity provided by the AHS to address the association of pesticide exposure with PD using neurologists' diagnoses and extensive exposure information.

In conclusion, the present study provides further support for the hypothesis that pesticide exposure increases PD risk, with detailed data on associations with specific pesticides and several aspects of pesticide use. Findings for specific chemicals may provide fruitful leads for further investigation.


   ACKNOWLEDGMENTS
 
This study was supported by intramural research funds from the National Institute of Environmental Health Sciences and the National Cancer Institute (Division of Cancer Epidemiology and Genetics) and by National Institute of Environmental Health Sciences grants 01-ES10803 and U54-ES12077.

The Agricultural Health Study was conducted at field stations in North Carolina (Battelle Memorial Institute, Columbus, Ohio: Charles Knott and Joy Pierce Herrington) and Iowa (University of Iowa, Iowa City, Iowa: Dr. Charles Lynch, Nyla Logsden-Sackett, Patti Gillette, and Ellen Heywood). Central coordination was provided by Westat, Inc., Rockville, Maryland (Paul Schroeder, Stanley Legum, and Marsha Dunn). Dr. Marie Richards provided statistical programming support.

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Tanner CM and Marder K. (2004) Movement disorders. In Nelson LM, Tanner CM, Van den Eeden SK (Eds.), et al. Neuroepidemiology: from principles to practice(Oxford University Press, New York, NY) pp. 131–61.
  2. Vila M and Przedborski S. (2004) Genetic clues to the pathogenesis of Parkinson's disease. Nat Med 10:suppl, S58–62.
  3. Tanner CM, Ottman R, Goldman SM, et al. (1999) Parkinson disease in twins: an etiologic study. JAMA 281:341–6.[Abstract/Free Full Text]
  4. Wirdefeldt K, Gatz M, Schalling M, et al. (2004) No evidence for heritability of Parkinson disease in Swedish twins. Neurology 63:305–11.[Abstract/Free Full Text]
  5. Priyadarshi A, Khuder SA, Schaub EA, et al. (2001) Environmental risk factors and Parkinson's disease: a metaanalysis. Environ Res 86:122–7.[Medline]
  6. Le Couteur DG, McLean AJ, Taylor MC, et al. (1999) Pesticides and Parkinson's disease. Biomed Pharmacother 53:122–30.[CrossRef][Medline]
  7. Priyadarshi A, Khuder SA, Schaub EA, et al. (2000) A meta-analysis of Parkinson's disease and exposure to pesticides. Neurotoxicology 21:435–40.[ISI][Medline]
  8. Kamel F and Hoppin JA. (2004) Association of pesticide exposure with neurologic dysfunction and disease. Environ Health Perspect 112:950–8.[ISI][Medline]
  9. Brown TP, Rumsby PC, Capleton AC, et al. (2006) Pesticides and Parkinson's disease—is there a link? Environ Health Perspect 114:156–64.[ISI][Medline]
  10. Li AA, Mink PJ, McIntosh LJ, et al. (2005) Evaluation of epidemiologic and animal data associating pesticides with Parkinson's disease. J Occup Environ Med 47:1059–87.[CrossRef][ISI][Medline]
  11. Alavanja M, Sandler D, McMaster S, et al. (1996) The Agricultural Health Study. Environ Health Perspect 104:362–9.[ISI][Medline]
  12. Tarone RE, Alavanja MC, Zahm SH, et al. (1997) The Agricultural Health Study: factors affecting completion and return of self-administered questionnaires in a large prospective cohort study of pesticide applicators. Am J Ind Med 31:233–42.[CrossRef][ISI][Medline]
  13. Dosemeci M, Alavanja MC, Rowland AS, et al. (2002) A quantitative approach for estimating exposure to pesticides in the Agricultural Health Study. Ann Occup Hyg 46:245–60.[Abstract/Free Full Text]
  14. Witte JS, Greenland S, Kim LL, et al. (2000) Multilevel modeling in epidemiology with GLIMMIX. Epidemiology 11:684–8.[CrossRef][ISI][Medline]
  15. Fall PA, Fredrikson M, Axelson O, et al. (1999) Nutritional and occupational factors influencing the risk of Parkinson's disease: a case-control study in southeastern Sweden. Mov Disord 14:28–37.[CrossRef][ISI][Medline]
  16. Ritz B and Yu F. (2000) Parkinson's disease mortality and pesticide exposure in California 1984 –1994. Int J Epidemiol 29:323–9.[Abstract/Free Full Text]
  17. Engel LS, Checkoway H, Keifer MC, et al. (2001) Parkinsonism and occupational exposure to pesticides. Occup Environ Med 58:582–9.[Abstract/Free Full Text]
  18. Herishanu YO, Medvedovski M, Goldsmith JR, et al. (2001) A case-control study of Parkinson's disease in urban population of southern Israel. Can J Neurol Sci 28:144–7.[ISI][Medline]
  19. Baldi I, Cantagrel A, Lebailly P, et al. (2003) Association between Parkinson's disease and exposure to pesticides in southwestern France. Neuroepidemiology 22:305–10.[CrossRef][ISI][Medline]
  20. Baldereschi M, Di CA, Vanni P, et al. (2003) Lifestyle-related risk factors for Parkinson's disease: a population-based study. Acta Neurol Scand 108:239–44.[CrossRef][ISI][Medline]
  21. Gorell JM, Peterson EL, Rybicki BA, et al. (2004) Multiple risk factors for Parkinson's disease. J Neurol Sci 217:169–74.[CrossRef][ISI][Medline]
  22. Firestone JA, Smith-Weller T, Franklin G, et al. (2005) Pesticides and risk of Parkinson disease: a population-based case-control study. Arch Neurol 62:91–5.[Abstract/Free Full Text]
  23. Kuopio AM, Marttila RJ, Helenius H, et al. (1999) Environmental risk factors in Parkinson's disease. Mov Disord 14:928–39.[CrossRef][ISI][Medline]
  24. Taylor CA, Saint-Hilaire MH, Cupples LA, et al. (1999) Environmental, medical, and family history risk factors for Parkinson's disease: a New England-based case control study. Am J Med Genet 88:742–9.[CrossRef][ISI][Medline]
  25. Behari M, Srivastava AK, Das RR, et al. (2001) Risk factors of Parkinson's disease in Indian patients. J Neurol Sci 190:49–55.[CrossRef][ISI][Medline]
  26. Nuti A, Ceravolo R, Dell'Agnello G, et al. (2004) Environmental factors and Parkinson's disease: a case-control study in the Tuscany region of Italy. Parkinsonism Relat Disord 10:481–5.[CrossRef][ISI][Medline]
  27. Petrovitch H, Ross GW, Abbott RD, et al. (2002) Plantation work and risk of Parkinson disease in a population-based longitudinal study. Arch Neurol 59:1787–92.[Abstract/Free Full Text]
  28. Baldi I, Lebailly P, Mohammed-Brahim B, et al. (2003) Neurodegenerative diseases and exposure to pesticides in the elderly. Am J Epidemiol 157:409–14.[Abstract/Free Full Text]
  29. Ascherio A, Chen H, Weisskopf MG, et al. (2006) Pesticide exposure and risk for Parkinson's disease. Ann Neurol 60:197–203.[CrossRef][ISI][Medline]
  30. Elbaz A, Levecque C, Clavel J, et al. (2004) CYP2D6 polymorphism, pesticide exposure, and Parkinson's disease. Ann Neurol 55:430–4.[CrossRef][ISI][Medline]
  31. Deng Y, Newman B, Dunne MP, et al. (2004) Further evidence that interactions between CYP2D6 and pesticide exposure increase risk for Parkinson's disease. (Letter). Ann Neurol 55:897.[CrossRef][ISI][Medline]
  32. Semchuk K, Love EJ, Lee RG. (1992) Parkinson's disease and exposure to agricultural work and pesticide chemicals. Neurology 42:1328–35.[Abstract/Free Full Text]
  33. Butterfield PG, Valanis BG, Spencer PS, et al. (1993) Environmental antecedents of young-onset Parkinson's disease. Neurology 43:1150–8.[Abstract/Free Full Text]
  34. Seidler A, Hellenbrand W, Robra BP, et al. (1996) Possible environmental, occupational, and other etiologic factors for Parkinson's disease: a case-control study in Germany. Neurology 46:1275–84.[Abstract/Free Full Text]
  35. Hertzman C, Wiens M, Bowering D, et al. (1990) Parkinson's disease: a case-control study of occupational and environmental risk factors. Am J Ind Med 17:349–55.[ISI][Medline]
  36. Liou HH, Tsai MC, Chen CJ, et al. (1997) Environmental risk factors and Parkinson's disease: a case-control study in Taiwan. Neurology 48:1583–8.[Abstract]
  37. McCormack AL, Thiruchelvam M, Manning-Bog AB, et al. (2002) Environmental risk factors and Parkinson's disease: selective degeneration of nigral dopaminergic neurons caused by the herbicide paraquat. Neurobiol Dis 10:119–27.[CrossRef][ISI][Medline]
  38. Betarbet R, Sherer TB, MacKenzie G, et al. (2000) Chronic systemic pesticide exposure reproduces features of Parkinson's disease. Nat Neurosci 3:1301–6.[CrossRef][ISI][Medline]
  39. Thiruchelvam M, Richfield EK, Baggs RB, et al. (2000) The nigrostriatal dopaminergic system as a preferential target of repeated exposures to combined paraquat and maneb: implications for Parkinson's disease. J Neurosci 20:9207–14.[Abstract/Free Full Text]
  40. Uversky VN, Li J, Bower K, et al. (2002) Synergistic effects of pesticides and metals on the fibrillation of alpha-synuclein: implications for Parkinson's disease. Neurotoxicology 23:527–36.[CrossRef][ISI][Medline]
  41. Di Monte D. (2003) The environment and Parkinson's disease: is the nigrostriatal system preferentially targeted by neurotoxins? Lancet Neurol 2:531–8.[CrossRef][ISI][Medline]
  42. Samanic C, Hoppin JA, Lubin JH, et al. (2005) Factor analysis of pesticide use patterns among pesticide applicators in the Agricultural Health Study. J Expo Anal Environ Epidemiol 15:225–33.[CrossRef][ISI][Medline]
  43. Greenamyre JT, Betarbet R, Sherer TB. (2003) The rotenone model of Parkinson's disease: genes, environment and mitochondria. Parkinsonism Relat Disord 9:suppl, S59–64.
  44. Kamel F, Engel LS, Gladen BC, et al. (2005) Neurologic symptoms in licensed private pesticide applicators in the Agricultural Health Study. Environ Health Perspect 113:877–82.[ISI][Medline]
  45. Hoppin JA, Umbach DM, London SJ, et al. (2002) Chemical predictors of wheeze among farmer pesticide applicators in the Agricultural Health Study. Am J Respir Crit Care Med 165:683–9.[Abstract/Free Full Text]
  46. Hernan MA, Takkouche B, Caamano-Isorna F, et al. (2002) A meta-analysis of coffee drinking, cigarette smoking, and the risk of Parkinson's disease. Ann Neurol 52:276–84.[CrossRef][ISI][Medline]
  47. Ponsen M, Stoffers D, Booij J, et al. (2004) Idiopathic hyposmia as a preclinical sign of Parkinson's disease. Ann Neurol 56:173–81.[CrossRef][ISI][Medline]
  48. Marras C, Goldman S, Smith A, et al. (2005) Smell identification ability in twin pairs discordant for Parkinson's disease. Mov Disord 20:687–93.[CrossRef][ISI][Medline]
  49. Blair A, Tarone R, Sandler D, et al. (2002) Reliability of reporting on life-style and agricultural factors by a sample of participants in the Agricultural Health Study from Iowa. Epidemiology 13:94–9.[CrossRef][ISI][Medline]
  50. Hoppin JA, Yucel F, Dosemeci M, et al. (2002) Accuracy of self-reported pesticide use duration information from licensed pesticide applicators in the Agricultural Health Study. J Expo Anal Environ Epidemiol 12:313–18.[CrossRef][ISI][Medline]
  51. Compendium of pesticide common names Wood A. (2005) (Justis Publishing Ltd, London, United Kingdom) (http://www.alanwood.net/pesticides/).

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 NeurologyHome page
L. Brighina, R. Frigerio, N. K. Schneider, T. G. Lesnick, M. de Andrade, J. M. Cunningham, M. J. Farrer, S. J. Lincoln, H. Checkoway, W. A. Rocca, et al.
{alpha}-Synuclein, pesticides, and Parkinson disease: A case-control study
Neurology, April 15, 2008; 70(16_Part_2): 1461 - 1469.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
165/4/364    most recent
kwk024v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Kamel, F
Right arrow Articles by Sandler, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kamel, F
Right arrow Articles by Sandler, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?