Skip Navigation LinksHome > October 2006 - Volume 48 - Issue 10 > Employment as a Welder and Parkinson Disease Among Heavy Equ...
Journal of Occupational & Environmental Medicine:
doi: 10.1097/01.jom.0000232547.74802.d8
Original Articles

Employment as a Welder and Parkinson Disease Among Heavy Equipment Manufacturing Workers

Marsh, Gary M. PhD; Gula, Mary Jean MS

Free Access
Article Outline
Collapse Box

Author Information

From the Department of Biostatistics (Dr Marsh), Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania; and a consultant, Verona, Pennsylvania (formerly, Department of Biostatistics, University of Pittsburgh) (Ms Gula).

This research was funded through Caterpillar Inc. Gary Marsh performed this work as a consultant to Caterpillar, Inc. The research protocol was approved by the Institutional Review Board (IRB) of the University of Pittsburgh, which considers this consulting work a usual professional activity. There are pending legal cases against Caterpillar, and Dr. Marsh may be asked to serve as an expert witness given his role as director of this research project.

Address correspondence to: Gary M. Marsh, PhD, Department of Biostatistics Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261; E-mail: gmarsh@pitt.edu.

Collapse Box

Abstract

Objective: We investigated whether employment as a welder with potential exposure to manganese and other substances is associated with Parkinson disease (PD), parkinsonism or related neurological disorders, or accelerates the age of onset of PD.

Methods: We selected cases and controls from 12,595 persons ever employed at three Caterpillar Inc. (CAT) plants between 1976 and 2004 with potential to make a medical insurance claim between 1998 and 2004. Cases had filed a claim for 1) PD, 2) “secondary parkinsonism”, 3) “other degenerative diseases of the basal ganglia” or 4) “essential and other specific forms of tremor”. Cases were grouped by claims: Group 1-claims 1 and 2 and Group 2-claims 1 to 4, and as study period incident (SPI) or prevalent. Each case was matched to two series of 10 controls each on date of case’s first claim, year of birth, race and sex. Series I was also matched on plant.

Results: Odds ratios (OR) and 95% confidence intervals (CI) for the variable, “ever welder in any CAT plant” were: Group 1-SPI Cases: Series I (OR = .76, CI = .26–2.19), Series II (OR = .81, CI = .29–2.25); Group 1- Prevalent Cases: Series I (OR = .82, CI = .36–1.86), Series II (OR = .97, CI = .42–2.23); Group 2- SPI Cases: Series I (OR = 1.03, CI = .57–1.87), Series II (OR = 1.21, CI = .67–2.20) Group 2-Prevalent Cases: Series I (OR = 1.02, CI = .62–1.71), Series II (OR = .86, CI = .51–1.43). Our finding of no statistically significant associations for welding employment was maintained following adjustment for potential confounding and evaluation of possible effect modification. Employment as a welder did not accelerate the age of onset of PD.

Conclusions: Our study supported the conclusion that employment as a welder is not associated with Parkinson disease, parkinsonism or a related neurological disorder.

Parkinson disease (PD) is one of the most common neurodegenerative disorders with a prevalence of 200 to 300 cases per 100,000 persons and an overall estimated annual incidence of 12 cases per 100,000.1 PD afflicts approximately one million Americans and from 1% to 2% of persons aged 60 and older worldwide.2–4 A recent study in The Netherlands suggests that 8.5% of all men and 7.7% of all women over age 55 will develop some form of parkinsonism, of whom slightly more than one half will develop PD.5

Although the cause of PD remains unknown, it is believed that environmental factors may play a role in addition to recently identified genetic links to parkinsonism.1,6,7 Factors including head injury, family history of PD, family history of essential tremor, history of depression, age of mother at subject’s birth, exposure to general anesthesia, farming as an occupation, and well water use have been associated with an excess risk of PD.8,9 Cigarette smoking has been associated with a decreased risk of PD in several studies.8–11

Heavy occupational exposure to manganese among foundry and smelter workers and miners has been associated with a rare neurologic disorder that resembles PD, commonly termed manganism or manganese-induced parkinsonism.12 Clear distinctions in the etiologies and clinical features of PD and manganism have been detailed by neurologic specialists.13,14 Because metal welding releases low-level exposures to manganese fumes and other compounds, employment as a welder has been implicated as a risk factor for PD15–20 or as accelerating the age of onset of the disease.21 Parkinsonism or manganism among welders has been described in several case reports, but such reports do not provide evidence of associations in which the role of chance, bias, or confounding can be rigorously evaluated.22–27

Several case–control studies of PD found no association with occupational exposure to manganese or with employment as a welder or metalworker.28–38 A recent historical cohort study of 27,839 male Danish metal manufacturing workers, including 6163 welders with more than 20 years of follow up, had rates of PD and other neurologic conditions consistent with those of the general population of Denmark.11 Another recent nationwide cohort study of 49,488 Swedish welders found no support for a relation between welding and PD or any other specific basal ganglia and movement disorders.39 Other study designs, including a 1947 health survey of shipyard workers,40 mortality studies in the United States41–44 and the United Kingdom,45 and cross-sectional studies46 have also found no evidence of an association with employment as a welder and PD or parkinsonism.

In June 2004, Caterpillar Inc. (CAT) commissioned us to design and conduct an epidemiologic study of the possible association between employment as a welder with potential exposure to manganese and other substances and the risk of developing idiopathic PD, parkinsonism, or a related neurologic disorder. To study this possible association, we designed and conducted a matched case–control study of employees at three CAT heavy equipment manufacturing plants in Illinois (Decatur, Aurora, and Joliet). This is a report on the findings of our matched case-control study.

Back to Top | Article Outline

Materials and Methods

Study Design

We designed this epidemiologic investigation as a matched case–control study. Employment as a welder in one or more CAT plants was the primary study factor compared between cases and controls. Comparisons were quantified by the calculation of an odds ratio (OR), which estimates the relative risk of developing PD, parkinsonism, or a related neurologic disorder given employment as a welder. In this study, the potentially confounding factors, calendar time (date of first case claim), age (year of birth), race, sex, and plant were controlled by matching cases to controls based on these factors. Other potentially confounding factors were evaluated and controlled as needed by stratification.

Back to Top | Article Outline
Study Population

The study population, which served as the source of cases and controls, comprised 12,595 persons ever employed at one or more of three CAT plants in Joliet, Decatur, and Aurora, Illinois, between 1976 and July 2004, and who had the potential to make a CAT medical insurance claim* between July 1998 and July 2004. These three plants were chosen because a high proportion of workers were employed as welders compared with other CAT plants. The study plants became operational in 1951 (Joliet), 1956 (Decatur), and 1958 (Aurora). Since the 1950s, virtually all of the welding at these three plants has been of mild steel in four processes: electrode welding, flux cored arc welding, gas metal arc welding, or submerged arc welding. The proportion of submerged arc welding has decreased significantly over the years. In nearly all of the plant welding operations, welding products containing manganese were always used. Some study members may have also worked for some time at one or more nonstudy CAT plants, and this history was also used to identify welding employment history. The study population database used for this study was created by linking CAT work history data with corresponding medical insurance claims data files.

Back to Top | Article Outline
Case Identification

Cases were identified from members of the study population who filed a CAT medical insurance claim for idiopathic PD, parkinsonism, or a related neurologic disorder between July 1998 and July 2004. Diagnosis was made at the time of the medical condition by the employee’s private physician and coded to the International Classification of Diseases, 9th Revision (ICD-9). Specifically, we included as cases all medical insurance claims assigned to idiopathic PD (ICD-9: 332.0), secondary parkinsonism (ICD-9: 332.1), “other degenerative diseases of the basal ganglia” (ICD-9: 333.0), and “essential and other specific forms of tremor” (ICD-9: 333.1). The Appendix provides more detailed definitions of these codes. To account for potential misclassification in the diagnosis of PD and related conditions, two case groups were formed. Persons were assigned to a case group if they ever made a case claim during the claim observation period for the codes included in that case group.

Case group 1 included idiopathic PD and secondary parkinsonism. In the ICD-9, the latter condition includes manganese-induced parkinsonism, neuroleptic-induced parkinsonism, and parkinsonism due to drugs. Case group 2 contained two additional codes (ICD-9: 333.0 and 333.1) to account for persons with PD or secondary parkinsonism who potentially were misdiagnosed. Each case group was then dichotomized by type of case.

Prevalent case consisted of any study member with at least one case claim in the July 1998 to July 2004 claim identification period. Prevalent cases include persons diagnosed with a case diagnosis code before and/or during the claim identification period. A prevalent case with a case claim before July 1998 must also have made a claim during the July 1998 to July 2004 timeframe to be included as a case. Study period incident case consisted of a prevalent case who also had no history of case claims for a minimum period of 6 months before the first case claim in the claim identification period. The 6-month case claim-free period increases the likelihood that the case was newly diagnosed during the claim identification period.

Back to Top | Article Outline
Selection of Matched Controls

All members of the study population not included in case group 1 or 2 were eligible to serve as a control for one or more cases from each case group. In both control series, a case was eligible to be a control for another case until the time of his or her first case claim. Each prevalent and study period incident case was individually matched to two series of 10 controls each. Series I matching criteria were: 1) calendar time (limited to month and year to protect confidentiality) of case’s first case claim (ie, the control must have been eligible to file a case claim in the same month and year of the case’s first case claim), 2) year of birth (±7 years for all but one female case in which the caliper was expanded to 15 years), 3) race (white or nonwhite), 4) sex, and 5) plant (the plant where the employee worked the longest time [Decatur, Aurora, or Joliet plants]). Series II matching criteria were identical to series I but did not include plant. Series II controls allowed plant to serve as an independent study factor if a relevant experience or exposure is plant-related.

Back to Top | Article Outline
Identification of Available Study Data

Complete, computerized work history data were identified for all employees at the three study plants from 1976 to July 2004. This database also included work history at other CAT plants. For workers hired at CAT before 1976, only the hire date (year of hire) was available; thus, a gap in their detailed work history occurred from this earlier date of hire to the period of complete work histories. Available work history data consisted of plant location, job code, job title, and employment dates (year). Subjects were assigned to the plant where they worked the longest time. Twelve subjects were found to have no job history in any of the study plants and were excluded from the analysis. To identify persons with potential exposure as a welder, we scanned the master work history record file (case status was not included) for any mention of the terms “welder” or “fabricator” (or the abbreviated forms “weld” and “fabr”). Job codes related to employment in any CAT plant (ie, in the three study plants plus possibly other CAT plants) were used to identify welding employment. Medical claims data (month/year of diagnosis and whether the subject had a case or other claim) were linked to the work history data by an anonymous employee identification number to determine case status (case group 1 or 2 vs not in either case group).

Back to Top | Article Outline
Data Analysis

Descriptive analyses included comparing the frequency distributions of relevant study factors between the cases and the two series of matched controls (without regard to the matching). All time-related study variables for cases and matched controls were calculated as of the calendar time of the case’s first case claim. All employment-related study variables accounted for employment history in any CAT plant for which data were available on the study file. Odds ratios (ORs) and corresponding 95% confidence intervals (CI) were calculated using conventional conditional logistic regression modeling techniques. An OR is statistically significantly different from the baseline value of 1.00 if its 95% CI does not include 1.00. The primary study variable, “ever welder in any CAT plant,” and other study variables were first examined in univariate (one independent variable) models as categorical variables to identify patterns of univariate associations with the outcome variable (becoming a case) and sparse data problems. The first category of each study variable served as the arbitrary baseline of the OR calculations for the nonbaseline categories. The possible association with welding employment and case status was then further evaluated for evidence of confounding by one or more of the other study variables. Potential effect modification (interaction) was also assessed. The statistical significance of each main effect (whether a given study variable was a predictor of the risk of becoming a case) was assessed with a likelihood ratio test and expressed as a “global P value.” All statistical tests were done at the 0.05 level of significance and no adjustments were made for multiple comparisons.

Back to Top | Article Outline

Results

Characteristics of Study Population

Of the 12,595 subjects in the total study population, 3430 (27.2%) were “ever welders in any CAT plant.” Most subjects were white males (82.3%), and the highest percentage worked at the Aurora plant (38.6%). The median age at hire was similar for “ever welders” (23 years) and “never welders” (24 years). Although the majority of the total study population (57.8%) and “ever welders in any CAT plant” (69.1%) were hired before 1976, only 11.5% and 6.4%, respectively, were hired before 1965 (data not shown). Thus, for most subjects hired before 1976, the time period without detailed work history data was relatively brief.

Back to Top | Article Outline
Characteristics of Cases

Table 1 shows the distribution of study period incident and prevalent cases by case group for each code and combination of codes. Most of the 87 total cases were coded as 333.1 (benign essential tremor) alone (48.3%) or 332.0 (idiopathic PD) alone (36.7%). All of the study period incident and prevalent cases in case group 1 had at least one claim for idiopathic PD (332.0). No subjects had a claim for secondary parkinsonism (332.1) alone, and none had more than three different claim codes. Of the 87 total prevalent cases, 81 (93.1%) of the subjects were male and six (6.9%) were female. Eighty cases (92.0%) were white and seven (8.0%) were nonwhite (data not shown).

Table 1
Table 1
Image Tools
Back to Top | Article Outline
Descriptive Comparison of Cases and Controls

For descriptive purposes only, Tables 2 through 5 show for the four case groups/types, respectively, the distribution of cases and their matched series I and II controls (combined) by each study variable. Except for the factors that relate to plant (a matching factor that differed between series I and II controls), we generally observed little difference in the distribution of study factors between series I and series II controls in each of the four case groups/types. In case group 1 (PD/secondary parkinsonism), we found no study period incident or prevalent cases with an age at first claim less than 50 years. The median age at first claim for both the study period incident and prevalent cases in case group 1 was over 60 years (data not shown). Because the aggregated control series data in Tables 2 through 5 break the individual-level matching, analytically comparing cases and controls is not possible in these tables.

Table 2
Table 2
Image Tools
Table 3
Table 3
Image Tools
Table 4
Table 4
Image Tools
Table 5
Table 5
Image Tools
Back to Top | Article Outline
Analytic Comparison of Cases and Controls

Tables 6 through 9 show for the four case groups/types, respectively, ORs and corresponding 95% CIs based on series I and II controls for categories of each study variable shown in Tables 2 through 5. Key findings from each table follow.

Table 6
Table 6
Image Tools
Table 7
Table 7
Image Tools
Table 8
Table 8
Image Tools
Table 9
Table 9
Image Tools
Back to Top | Article Outline
Case Group 1: Study Period Incident Cases (Table 6).

The primary study variable “ever welder in any CAT plant” was not a statistically significant predictor of case status based on both series I controls (OR = 0.76, 95% CI = 0.26–2.19, P = 0.60) and series II controls (OR = 0.81, 95% CI = 0.29–2.25, P = 0.68). For both control series, the risk for “ever welder in any CAT plant” was approximately 20% less for cases than for controls. Among the other study variables considered, none was a statistically significant predictor of case status.

Back to Top | Article Outline
Case Group 1: Prevalent Cases (Table 7).

The primary study variable “ever welder in any CAT plant” was not a statistically significant predictor of case status based on both series I controls (OR = 0.82, 95% CI = 0.36–1.86, P = 0.63) and series II controls (OR = 0.97, 95% CI = 0.42–2.23, P = 0.94). The risk for “ever welder in any CAT plant” was 18% and 3% less for cases than for series I and II controls, respectively. Among the other study variables considered as possible confounding variables, “year of hire” and “time since first employment at any CAT plant” were statistically significant predictors of case status based on both series I and series II controls. The bottom sections of Table 7 show that although each of these study variables had some small negative confounding effect on the primary study variable based on both control series (ie, the OR for the primary study variable increased in the presence of the other variable), the primary study variable, “ever welder in any CAT plant,” remained a nonstatistically significant predictor of case status.

Back to Top | Article Outline
Case Group 2: Study Period Incident Cases (Table 8).

The primary study variable “ever welder in any CAT plant” was not a statistically significant predictor of case status as based on both series I controls (OR = 1.03, 95% CI = 0.57–1.87, P = 0.91) and series II controls (OR = 1.21, 95% CI = 0.67–2.20, P = 0.53). Persons who were “ever welder in any CAT plant” had a slightly greater risk of making a case claim than persons without this type of employment (3% and 21% greater based on series I and II controls, respectively). Among the other study variables considered as possible confounding variables, only “age at case’s first case claim” was a statistically significant predictor of case status based on series II controls only. The bottom sections of Table 8 show that although “age at case’s first case claim” had some minimal negative confounding effect on the primary study variable, the primary study variable, “ever welder in any CAT plant,” remains a nonstatistically significant predictor of case status.

Back to Top | Article Outline
Case Group 2: Prevalent Cases (Table 9).

The primary study variable “ever welder in any CAT plant” was not a statistically significant predictor of case status as based on both series I controls (OR = 1.02, 95% CI = 0.62–1.71, P = 0.92) and series II controls (OR = 0.86, 95% CI = 0.51–1.43, P = 0.55). The risk of making a case claim for persons who were “ever welder in any CAT plant” compared with those who were not is essentially the same for the series I control matched sets and 14% lower for the cases as compared with the series II matched controls. Among the other study variables considered as possible confounding variables, only “time since first employment at any CAT plant” was a statistically significant predictor of case status based on both series I and series II controls. The bottom sections of Table 9 show that although “time since first employment at any CAT plant” had some minimal negative confounding effect on the primary study variable, the primary study variable, “ever welder in any CAT plant,” remains a statistically nonsignificant predictor of case status.

Back to Top | Article Outline
Evaluation of Possible Effect Modification.

Other models were fit to evaluate possible effect modification (interaction) between the primary study variable and the other study variables (results not shown). This was done to identify possible welding employment and case associations within subgroups of the study population (eg, within specific study plants or pay types) that may have been overlooked in the aggregate level analyses presented in Tables 6 through 9. Interactions of “ever welder in any CAT plant” with “plant worked longest time,” “year of hire,” “pay type,” “duration of employment at all CAT plants,” and “time since first employment at any CAT plant” as the categorical variables used in Tables 6 through 9 were examined for all study matched sets (case groups 1 and 2, study period incident and prevalent, matched to series I and II controls).

A statistically significant interaction (P = 0.05) was found for “ever welder in any CAT plant” and “time since first employment at any CAT plant” for case group 2, prevalent cases, matched to series II controls only. The unmatched frequency distribution for these cases showed 54% of nonwelders and 59% of welders with a time since first employment less than 20 years; in controls, 61% of nonwelders and 85% of welders had a time since first employment less than 20 years. The P value for this interaction term in all of the other matched analyses ranged from 0.08 to 0.61. The interaction of “ever welder in any Caterpillar plant” with “pay type” was also statistically significant in the matched analysis of case group 1, study period incident cases, series I controls (P = 0.04) and in the case group 2, prevalent cases, series I (P = 0.04) and series II controls (P = 0.03). For both groups, the unmatched descriptive analyses revealed a higher proportion of control subjects that were ever welders with hourly pay type.

Back to Top | Article Outline

Discussion

The key finding of our matched case–control study was the primary study variable “ever welder in any CAT plant” was consistently not a statistically significant predictor of increased risk of developing idiopathic PD or a related condition as defined by the diagnostic codes included in case group 1 (ICD-9: 332.0 and 332.1) and case group 2 (ICD-9: 332.0, 332.1, 333.0, and 333.1). This finding was maintained in both series of matched controls and among prevalent and study period incident cases and after additional analytic control (beyond that achieved in the matched design) for potential confounding by the study variables “plant worked longest time,” “year of hire,” “age at hire,” “pay type,” “duration of employment at all CAT plants,” and “time since first employment at any CAT plant.” We also evaluated effect modification (interaction) to identify possible welding and case status associations with subgroups of the study population that may have been missed at the aggregate-level analyses. Although this evaluation revealed a few isolated statistically significant interactions, the pattern of these interactions did not support an association between employment as a welder and case status. Finally, our study also found no evidence that employment as a welder accelerates the age of onset of PD, as suggested by Racette.21

Relatively few epidemiologic investigations have directly examined whether welding with potential exposure to manganese and other substances is a risk factor for PD, parkinsonism, or related neurologic disorders such as manganism, and most of these have not found evidence of such associations. A 1947 health survey of 4650 shipyard workers, with approximately 75% of these being welders, found no statistically significantly higher occurrence of parkinsonism or other neurologic disease.40 A 1981 exploratory study in India reported the observation of manganese toxicity among welders, but limitations of the study preclude a meaningful interpretation of the findings.15 A small pilot study in Washington state found three of 16 male cases of PD but none of nine controls from the neurology clinic patient population had a history of employment as a welder.16 Semchuk and coworkers conducted a population-based, case–control study in Calgary, Alberta, and found no increased risk from PD among persons ever with occupational exposure to manganese.28 A 1994 Canadian study of 127 PD cases and 245 controls found no statistically significant associations between PD and exposures to heavy metals, including manganese.29 Seidler and colleagues found no significant association of PD with occupational exposure to manganese in a hospital-based, case–control study of nine German clinics.30 A significant association of PD with exposure to manganese in persons with more than 20 years of occupational exposure was found by Gorell and coworkers in a population-based, case–control study at Henry Ford Health System (HFHS) in Detroit.17–19 In this study, however, only three cases and one control subject had occupational exposure to manganese of more than 20 years and none had worked as a welder. A 2004 reanalysis of these data found no statistically significant association between occupational exposure to manganese and PD.35 A 1999 Swedish study of 113 PD cases and 263 controls found a significantly reduced PD risk for metal workers.31 A large 1999 case–control study evaluated whether occupations, including welding, were associated with an increased risk of PD among the general population of Vancouver, British Columbia. An evaluation of the occupations of 441 PD cases and 6649 controls found many, but not welding, to be associated with PD.32 In a 1999 reanalysis of case–control data from the Michigan healthcare system,17–19 the category “machine trades,” which included welders, showed a slight deficit in risk of PD compared with the control subjects.33 A 2003 case–control study involving the World War II Veteran twins cohort identified 142 twin pairs with one twin having PD. For these subjects, welding was not associated with an excess risk of PD.34 In a large 2004 hospital-based, case–control study in Korea, 367 patients with PD and 309 controls were interviewed about occupational history, lifestyle, family history, and education level. For occupations with a high potential exposure to manganese and other substances, including welding, the study found consistently negative associations with PD after adjustment for potential confounding by age, sex, smoking history, and education level.36,37 Frigerio and colleagues identified all subjects with PD in a county in Minnesota from 1976 to 1995 and matched general population controls to investigate the association of study period incident cases of PD with education and occupations. From a medical records review (97% of both cases and controls had available information), only three controls and no cases were welders. From a telephone survey that was also conducted (74% of cases and 64% of controls participated), only one control and no cases were welders.38 In 2005, Goldman and coworkers examined medical records from 2249 consecutive patients with PD or parkinsonism as diagnosed from specialty clinics. The authors did not find an association between parkinsonism and welding, yet did identify an association of increased PD risk with farming, healthcare occupations, and some other jobs.46

A mortality study from 27 states in 1982–1991 examined occupations for excess risks of PD and other neurodegenerative diseases.42 Hypotheses generated from this study were tested in a later study using death certificate information for all deaths from 1992 to 1998 in 22 states participating in the National Occupational Mortality Surveillance (NOMS) system. The later study found no association with welding and PD mortality when all (n = 540) PD deaths were evaluated, but found an elevated PD mortality risk based on 20 PD deaths under age 65.20 A recent cohort study of PD and other neurodegenerative disorders in Danish welders was conducted by Fryzek and colleagues.11 The study population consisted of 27,839 male employees, including persons both exposed and nonexposed to welding fumes. The Danish National Register of Patients was used to determine persons with a primary diagnosis of PD and the other codes of interest in this study. Company records, interviews, and questionnaire data were used to determine potentially welding-exposed persons. The authors concluded that this cohort of Danish welders had rates of PD and the other neurologic conditions examined to be consistent with those of the general population of Denmark. A large cohort study in Sweden identified 49,488 welders or flame cutters from the 1960 or 1970 Swedish National Census. Their rates of specific basal ganglia and movement disorders between 1964 and 2003 were compared with an age and geographic area matched cohort of gainfully employed men from the general population. The overall rate for basal ganglia and movement disorders combined was similar for the welder and flame cutter group compared with the population controls. No significant differences were found between the welder and flame cutter group and the general population in the analyses of PD overall and by attained age, time period of follow up, geographic area of residency, and educational level.39

A study by Racette and colleagues21 included 953 parkinsonian patients seen in their neurology specialty clinic between 1996 and 2000. Based on the observation that the mean age of the 15 welders with PD was 17 years younger (46 years) compared with patients with PD who were not welders (63 years of age), the authors suggested that welding may accelerate the onset of PD. The fact that the majority of welders (eight of 15) in the study had a family history of PD, as compared with 15% reported in unselected patients with PD limits the interpretability of these findings.47 Our case–control study did not find evidence of an accelerated age of onset of PD. In case group 1 (PD/secondary parkinsonism), we found no study period incident or prevalent cases with an age at first claim less than 50 years. The median age at first claim for both the study period incident and prevalent cases in case group 1 was over 60 years. A 2005 report by Racette and colleagues studied a group of 1423 welders from Alabama who were referred by attorneys for medical–legal evaluation.48 This group of welders had a higher prevalence of parkinsonism (5.8% with a diagnosis of “definite PD”), which was significantly higher than the estimated prevalence in the general population based on published data from a 1985 epidemiologic study of Copiah county.49 This study also has serious methodological limitations, including selection of only potential litigants, the use of an unvalidated video protocol rather than an established neurologic examination to confirm diagnosis, lack of blindness in the study methods, use of historical rather than a true control population, and performing part III of the Unified PD Rating Scale, which requires an actual physical examination, only on a “pseudorandom group” (less than 7.9%) of subjects.

Our case–control study has many strengths that overcome key limitations of earlier studies. Several studies that examined the relationship between welding and idiopathic PD or related disorders relied on subjects identified through movement disorder or specialty clinics.17,21,46,48 For example, Racette studied 1423 Alabama welders referred for medicolegal examination, giving rise to potential selection bias with respect to welding employment status.48 To mitigate potential study selection or information biases in our study, we identified subjects systematically with uniform study entrance criteria. Also, the method of data collection, assignment of job codes for welders, selection criteria for matched controls, and all other relevant study decisions were made without prior knowledge of case status or welding employment status. In examining the possible association with employment as a welder and within the context of a matched case–control study, the potentially confounding factors: time period (date of case’s first case-related claim), age (date of birth), race, sex, and plant were controlled by individually matching cases to two series of controls on the basis of these factors (series II did not match on plant). Furthermore, controls from each series were randomly sampled from large pools of eligible controls, thus eliminating any possibility of personal judgment affecting control selection. In addition to the matching, our case–control analyses accounted for additional potential confounding by stratification and evaluated possible effect modification from other study factors.

Our study was also based on detailed work history and medical insurance claims data known to be complete and accurate during well-defined time periods, thus obviating potential information bias. This detail also enabled the important differentiation of study period incident and prevalent cases that increased the likelihood that the cases in this study were newly diagnosed during the claim identification period (ie, they developed during employment at CAT). Access to medical insurance claims data allowed us to identify a subject as a case when a relevant medical claim was made rather than being limited, for example, to hospitalizations for a specific code or to a death certificate diagnosis. For example, Fryzek et al used the Danish National Health Register of Patients, a computerized listing of all hospital admissions from 1977 forward, to identify disease status,11 and several studies were based on PD mortality data gleaned from death certificates.20,41–45 Mortality data are considered an incomplete measure of PD occurrence, because only about one fourth of decedents with PD have a death certificate with that disease listed as the underlying cause.42,50

Three of the four study groups examined in our case–control study (case group 2 prevalent and study period incident and case group 1 prevalent) had good to excellent statistical power (67–90%) for detecting a true twofold or greater excess risk of becoming a case given prior employment as a welder. The generally very favorable power characteristics of this study lend considerable weight to our findings of no association with employment as a welder with potential exposure to manganese and other substances and becoming a case of PD, parkinsonism, or a related neurologic disorder.

Several other features of our study should be considered when interpreting the results. Quantitative information on specific workplace exposures to chemicals such as manganese was not available for this investigation. Information relating to change in process, protective equipment worn by study members, and potential differences among plants was also not available. Because only job title was used to identify study members with potential welding-related exposures, it was possible that not all employees with welding exposure were captured by use of the job title only. However, the designation of welder/nonwelder was done before creation of the data analysis file without knowledge of a subject’s case status to eliminate potential bias in the selection of welder job codes. Although specific workplace exposure data were not available in our study, detailed job and department work history records were available from 1976 to identify whether a subject was “ever a welder in any CAT plant.” We also had complete data on general employment variables (date of hire and termination, pay status, and so on) that we used to assess potential confounding and effect modification. Fryzek et al used company records, interviews with supervisors, and long-term workers to determine persons who worked in the welding department followed by questionnaires to obtain information on occupational history and tobacco smoking habits.11

Because smoking has been consistently linked to marked reductions in PD risk,50 we were concerned that our findings of reduced risks for PD, parkinsonism, and related neurologic disorders among persons ever employed as welders may be a reflection of negative confounding by smoking if a higher prevalence of smoking occurred among welders compared with nonwelders. Although we did not collect any smoking history data for the current study, an earlier National Institute of Occupational Safety and Health investigation of lung cancer mortality among workers from the same three study sites reported nearly identical prevalence rates for “ever a smoker” among welders and nonwelders (78% vs 79%, respectively) based on a 1985 company survey of hourly workers exposed to high noise levels.51,52 Although the extent of potential confounding by smoking is unknown in our case–control study, the results of the earlier smoking survey suggested that substantial confounding was unlikely.

Because not all CAT employees were covered by CAT-administered medical insurance during their employment history or after retirement from CAT, we were concerned about possible underascertainment of cases. For example, employees with HMO coverage may only sporadically have claims data in the CAT claims database (a claim may have been put through CAT and denied). Some employees may also have opted not to purchase any medical insurance or may have purchased insurance through another carrier. Assurance that cases have been completely ascertained is possible only for study members covered by CAT medical insurance during the entire July 1998 to July 2004 claim identification period. However, underascertainment of cases from mixed medical insurance coverage would bias the study results toward false-positive or false-negative findings only if the distribution of study members with non-CAT medical insurance during the claims identification period differed substantially between welders and nonwelders and/or between employees with and without a case claim (ie, case status).

To investigate the possibility of study bias stemming from differential underascertainment of cases, we reviewed three groups of employees from the CAT computerized personnel data files: (group 1) a 5% random sample (67 of 1352) of employees with no claims on file, (group 2) 38 of 42 employees (90.5%) with available data on type of insurance and who made a claim for idiopathic PD during the claims identification period by type of insurance (CAT vs HMO), and (group 3) a 2% random sample (225 of 11,216) of employees who made a claim other than PD during the claim identification period by type of insurance (CAT vs HMO). This review suggested that a significant disparity did not exist between welders and nonwelders who made non-PD claims either through CAT insurance or an HMO.

Another potential threat to study validity related to the possibility that an employee developing symptoms of PD or a related neurologic disorder might terminate employment (and lose medical insurance) before their disease was diagnosed and a medical claim was filed. This potential source of case underascertainment probably played a minor role in our study because most cases of PD or other neurologic disorders occurred near or after retirement age. Moreover, this potential source of case underascertainment probably had little impact among even younger workers, because PD and other related disorders can exist for several years before becoming clinically evident,53 and the early stages of these diseases may not affect one’s ability to work. In any event, if PD or a related neurologic disorder did cause people to leave work before diagnosis, it would be a potential problem only if the event of terminating employment due to such a diagnosis occurred differentially between welders and nonwelders.

We also investigated potential study bias due to possible differential underascertainment of welding employment for cases and controls. Although CAT was able to supply detailed work history records dating back to 1976, approximately 58% of the total study population of potential cases and controls was initially hired before that time period. For these persons, duration of employment and other employment-related study variables were calculated from date of hire until the case–control study stop date, defined as either the date of case’s first PD case claim or the date of last work history action if this occurred before the case’s first case claim date. It was possible, therefore, that an employee worked in a welding job only before 1976 (when complete work history records became more available) and thus would not be identified as a welder for this investigation. Such underascertainment of welding employment could bias the study results toward false-positive or false-negative findings only if the distribution of employees hired before the detailed work history period differed substantially between welders and nonwelders and/or between employees with and without a case claim (ie, case status). Furthermore, as noted previously, for most subjects hired before 1976, the time period without detailed work history data was relatively brief.

We also investigated the extent of unknown detailed work history between cases and controls. In case group 1 (study period incident and prevalent cases), we found a median difference of approximately 6 to 7 years of unknown work history between prevalent and study period incident cases and their matched controls. In contrast, for case group 2 (study period incident and prevalent cases), which included a larger proportion of younger cases relative to case group 1, the proportion and median years of unknown work history were very similar. This disparity did not imply that cases and controls in case groups 1and 2 also differed (or did not differ) with respect to the extent of unknown welding employment (the source of possible study bias), but suggested that the extent of unknown welding employment between cases and controls likely differed more in case group 1. The OR estimates for “ever welder in any CAT plant” in case group 1 are uniformly less than 1.0; in case group 2, corresponding OR estimates are generally closer to or slightly larger than 1.0. This difference may be due to missed welding employment among case group 1 cases that would cause a slight downward effect to the OR estimates or it may be due to other factors. On balance, the analysis using case group 2 may be more valid because the amount of missing work history for the cases and controls is similar.

To investigate further the potential bias due to possible differential ascertainment of welding employment, the pattern of welding employment was determined for each subject during the time period of available detailed work histories (results not shown). The corresponding case and control group distributions for case groups 1 and 2 showed that persons moved in and out of welding jobs in the course of their work history but that the extent of this movement was fairly similar between cases and controls. These findings suggest that movement in and out of welding jobs per se was not likely to lead to differential underascertainment of welding status between cases and controls (the source of potential study bias). We also investigated the feasibility of limiting study groups to persons with complete work history data and found the numbers of cases were insufficient for analysis.

Back to Top | Article Outline

Conclusions

Overall, our evaluation of potential study biases due to possible differential underascertainment of cases between welders and nonwelders and/or welding employment between cases and controls revealed no results that compromise the validity of the findings from our case–control study. The findings of our study supported the conclusion that employment as a welder with potential exposure to manganese and other substances is not associated with the risk of developing idiopathic PD, parkinsonism, or a related neurologic disorder.

Back to Top | Article Outline

Acknowledgments

Caterpillar Inc. sponsored this research, but the design, conduct, analysis, and conclusions are those of the authors. The authors acknowledge the computer programming support of Steve Sefcik. The research proposal was approved by the Institutional Review Board of the University of Pittsburgh.

Back to Top | Article Outline
Appendix
Conditions and Disorders Associated With International Classification of Diseases, 9th Revision Codes Used to Define Case Groups

332.0 Paralysis agitans

Parkinsonism or Parkinson disease

NOS (not otherwise specified)

Idiopathic

Primary

332.1 Secondary parkinsonism

Parkinsonism due to drugs

333.0 Other degenerative diseases of the basal ganglia

Atrophy or degeneration

Olivopontocerebellar (DeJerine Thomas syndrome)

Pigmentary pallidal (Hallervorden-Spatz disease) striatonigral

Parkinsonian syndrome associated with

Idiopathic orthostatic hypotension

Symptomatic orthostatic hypotension

Progressive supranuclear ophthalmoplegia

Shy-Drager syndrome

333.1 Essential and other specified forms of tremor

Benign essential tremor

Familial tremor

Back to Top | Article Outline

References

1. Tanner CM, Ben-Shlomo Y. Epidemiology of Parkinson’s disease. Adv Neurol. 1999;80:153–159.

2. Elbaz A, Bower JH, Maraganore DM, et al. Risk tables for parkinsonism and Parkinson’s disease. J Clin Epidemiol. 2002;55:25–31.

3. Bergareche A, De la Puente E, López de Munain A, et al. Prevalence of Parkinson’s disease and other types of parkinsonism. A door–to-door survey in Bidasoa, Spain. J Neurol. 2004;251:340–345.

4. Strickland D, Bertoni JM. Parkinson’s prevalence estimated by a state registry. Mov Disord. 2004;19:318–323.

5. de Lau LM, Giesbergen PC, de Rijk MC, Hofman A, Koudstaal PJ, Breteler MM. Incidence of parkinsonism and Parkinson disease in a general population: the Rotterdam Study. Neurology. 2004;63:1240–1244.

6. Huang Y, Cheung L, Rowe D, Halliday G. Genetic contributions to Parkinson’s disease. Brain Res. 2004;46:44–70.

7. Nichols WC, Pankratz N, Hernandez D, et al. Genetic screening for a single common LRRK2 mutation in familial Parkinson’s disease. Lancet. 2005;365:410–412.

8. Taylor CA, Saint-Hilaire MH, Cupples LA, et al. Environmental, medical, and family history risk factors for Parkinson’s disease: a New England-based case control study. Am J Med Genet. 1999;88:742–749.

9. Zorzon M, Capus L, Pellegrino A, Cazzato G, Zivadinov R. Familial and environmental risk factors in Parkinson’s disease: a case–control study in north-east Italy. Acta Neurol Scand. 2002;105:77–82.

10. Tanner CM, Goldman SM, Aston DA, et al. Smoking and Parkinson’s disease in twins. Neurology. 2002;58:581–588.

11. Fryzek JP, Hansen J, Cohens S, et al. A cohort study of Parkinson’s disease and other neurodegenerative disorders in Danish welders. J Occup Environ Med. 2005;47:466–472.

12. Olanow CW. Manganese-induced parkinsonism and Parkinson’s disease. Ann N Y Acad Sci. 2004;1012:209–223.

13. Jankovic J. Searching for a relationship between manganese and welding and Parkinson’s disease. Neurology. 2005;64:2021–2028.

14. Manganese. In: Air Quality Guidelines, 2nd ed. Denmark: World Health Organization; 2001.

15. Chandra SV, Shukla GS, Srivastava RS, Singh H, Gupta VP. An exploratory study of manganese exposure to welders. Clin Toxicol. 1981;18:407–416.

16. Wechsler LS, Checkoway H, Franklin GM, Costa LG. A pilot study of occupational and environmental risk factors for Parkinson’s disease. Neuro Toxicol. 1991;12:387–392.

17. Gorell JM, Johnson CC, Rybicki BA, et al. Occupational exposures to metals as risk factors for Parkinson’s disease. Neurology. 1997;48:650–658.

18. Gorell JM, Johnson CC, Rybicki BA, et al. Occupational exposure to manganese, copper, lead, iron, mercury and zinc and the risk of Parkinson’s disease. Neurotoxicology. 1999a;20:239–248.

19. Gorell JM, Rybicki BA, Johnson CC, Peterson EL. Occupational metal exposures and the risk of Parkinson’s disease. Neuroepidemiology. 1999b;18:303–308.

20. Park RM, Schulte PA, Bowman JD, et al. Potential occupational risks for neurodegenerative diseases. Am J Ind Med. 2005;48:63–77.

21. Racette BA, McGee-Minnich L, Moerlein SM, Mink JW, Videen TO, Perlmutter JS. Welding-related parkinsonism: clinical features, treatment, and pathophysiology. Neurology. 2001;56:8–13.

22. Whitlock CM Jr, Amuso SJ, Bittenbender JB. Chronic neurological disease in two manganese steel workers. Am Ind Hyg Assoc J. 1966;27:454–459.

23. Tanaka S, Lieben J. Manganese poisoning and exposure in Pennsylvania. Arch Environ Health. 1969;19:674–684.

24. Nelson K, Golnick J, Korn T, Angle C. Manganese encephalopathy: utility of early magnetic resonance imaging. Br J Ind Med. 1993;50:510–513.

25. Sato K, Ueyama H, Arakawa R, Kumamoto T, Tsuda T. A case of welder presenting with parkinsonism after chronic manganese exposure. Rinsho Shinkeigaku. 2000;40:1110–1115.

26. Sadek AH, Rauch R, Schulz PE. Parkinsonism due to manganism in a welder. Int J Toxicol. 2003;22:393–401.

27. Josephs KA, Ahlskog JE, Klos KJ, et al. Neurologic manifestations in welders with pallidal MRI T1 hyperintensity. Neurology. 2005;64:2033–2039.

28. Semchuk KM, Love EJ, Lee RG. Parkinson’s disease: a test of the multifactorial etiologic hypothesis. Neurology. 1993;43:1173–1180.

29. Hertzman C, Wiens M, Snow B, Kelly S, Calne D. A case–control study of Parkinson’s disease in a horticultural region of British Columbia. Mov Disord. 1994;9:69–75.

30. Seidler A, Hellenbrand W, Robra BP, et al. Possible environmental, occupational, and other etiologic factors for Parkinson’s disease: a case–control study in Germany. Neurology. 1996;46:1275–1284.

31. Fall PA, Fredrikson M, Axelson O, Granérus AK. Nutritional and occupational factors influencing the risk of Parkinson’s disease: a case–control study in Southeastern Sweden. Mov Disord. 1999;14:28–37.

32. Tsui JK, Calne DB, Wang, Y, Schulzer M, Marion SA. Occupational risk factors in Parkinson’s disease. Can J Public Health. 1999;90:334–337.

33. Kirkey KL, Johnson CC, Rybicki BA, Peterson EL, Kortsha GX, Gorell JM. Occupational categories at risk for Parkinson’s disease. Am J Ind Med. 2001;39:564–571.

34. Tanner CM. Is the cause of Parkinson’s disease environmental or hereditary? Evidence from twin studies. Adv Neurol. 2003;91:133–142.

35. Gorell JM, Peterson EL, Rybicki BA, Johnson CC. Multiple risk factors for Parkinson’s disease. J Neurol Sci. 2004;217:169–174.

36. Park J, Yoo C, Sim CS, et al. Occupations and Parkinson’s disease: a case–control study in South Korea. Ind Health. 2004;42:352–358.

37. Park J, Yoo CI, Sim CS, et al. Occupations and Parkinson’s disease: a multi-center case–control study in South Korea. Neurotoxicology. 2005;26:99–105.

38. Frigerio R, Elbaz A, Sanft KR, et al. Education and occupations preceding Parkinson’s disease. Neurology. 2005;65:1575–1583.

39. Fored CM, Fryzek JP, Brandt L, et al. Parkinson’s disease and other basal ganglia or movement disorders in a large nationwide cohort of Swedish welders. Occup Environ Med. 2006;63:135–140.

40. Dreessen WC, Brinton HP, Keenan RG, Thomas TR, Place EH. Health of Arc Welders in Steel Ship Construction. Federal Security Agency. US Public Health Service; 1947.

41. Walrath J, Rogot E, Murray J, Blair A. Mortality Patterns Among US Veterans by Occupation and Smoking Status. National Institutes of Health; 1985.

42. Schulte PA, Burnett CA, Boeniger MF, Johnson J. Neurodegenerative diseases: occupational occurrence and potential risk factors, 1982 through 1991. Am J Public Health. 1996;86:1281–1288.

43. Milham S. Occupational Mortality in Washington State 1950–1989. Cincinnati: US Department of Health and Human Services, National Institute for Occupational Safety and Health; 1997:288–292.

44. Kizer KW. California Occupational Mortality 1979–1981. Sacramento: California Department of Health Services, Office of Health Information and Research; 1987.

45. Coggon D, Inskip H, Winter P, Pannett B. Occupational mortality by cause of death. In: Drever F, ed. Occupational Health Decennial Supplement. London: Her Majesty’s Stationery Office; 1995:62–76.

46. Goldman SM, Tanner CM, Olanow CW, Watts RL, Field RD, Langston JW. Occupation and parkinsonism in three movement disorders clinics. Neurology. 2005;65:1430–1435.

47. Rybicki BA, Johnson CC, Peterson EL, Kortsha GX, Forell JM. A family history of Parkinson’s disease and its effect on other PD risk factors. Neuroepidemiology. 1999;18:270–278.

48. Racette BA, Tabbal SD, Jennings D, Good L, Perlmutter JS, Evanoff B. Prevalence of parkinsonism and relationship to exposure in a large sample of Alabama welders. Neurology. 2005;64:230–235.

49. Schoenberg BS, Anderson DW, Haerer AF. Prevalence of Parkinson’s disease in the biracial population of Copiah County, Mississippi. Neurology. 1985;35:841–845.

50. Checkoway H, Nelson LM. Epidemiologic approaches to the study of Parkinson’s disease etiology. Epidemiology. 1999;10:327–336.

51. Steenland K, Beaumont J, Elliot L. Lung cancer in mild steel welders. Am J Epidemiol. 1991;133:220–229.

52. Steenland K. Ten-year update on mortality among mild-steel welders. Scand J Work Environ Health. 2002;28:163–167.

53. Langston JW. Predicting Parkinson’s disease. Neurology. 1990;40(suppl 3):70–74.

*Persons covered under “CAT medical insurance” are insured through a CAT-sponsored insurance plan. Employees also had the option to chose an HMO insurance plan. Cited Here...

†The year 1976 was chosen as the earliest study population eligibility date because this was the beginning of computerized personnel record systems at CAT. The study population also includes eligible study members whose work experience began before 1976. The July 1998 to July 2004 observation period was chosen because CAT considered claim information to be complete during this time period. Cited Here...

Cited By:

This article has been cited 14 time(s).

Movement Disorders
Advances in Environmental Epidemiology
Tanner, CM
Movement Disorders, 25(3): S58-S62.
10.1002/mds.22721
CrossRef
Journal of Occupational and Environmental Hygiene
Manganese and welding fume exposure and control in construction
Meeker, JD; Susi, P; Flynn, MR
Journal of Occupational and Environmental Hygiene, 4(): 943-951.
10.1080/15459620701718867
CrossRef
Experimental Neurology
Gazing into the future: Parkinson's disease gene therapeutics to modify natural history
Maguire-Zeiss, KA; Mhyre, TR; Federoff, HJ
Experimental Neurology, 209(1): 101-113.
10.1016/j.expneurol.2007.09.030
CrossRef
Journal of Occupational and Environmental Hygiene
Welding Occupations and Mortality from Parkinson's Disease and Other Neurodegenerative Diseases Among United States Men, 1985-1999
Stampfer, MJ
Journal of Occupational and Environmental Hygiene, 6(5): 267-272.
10.1080/15459620902754703
CrossRef
International Journal of Hygiene and Environmental Health
Neurological risks associated with manganese exposure from welding operations - A literature review
Flynn, MR; Susi, P
International Journal of Hygiene and Environmental Health, 212(5): 459-469.
10.1016/j.ijheh.2008.12.003
CrossRef
Journal of Neurochemistry
Synergistic dopaminergic neurotoxicity of manganese and lipopolysaccharide: differential involvement of microglia and astroglia
Zhang, P; Lokuta, KM; Turner, DE; Liu, B
Journal of Neurochemistry, 112(2): 434-443.
10.1111/j.1471-4159.2009.06477.x
CrossRef
Environmental Health Perspectives
Meeting Report: Consensus Statement-Parkinson's Disease and the Environment: Collaborative on Health and the Environment and Parkinson's Action Network (CHE PAN) Conference 26-28 June 2007
Bronstein, J; Carvey, P; Chen, H; Cory-Slechta, D; DiMonte, D; Duda, J; English, P; Goldman, S; Grate, S; Hansen, J; Hoppin, J; Jewell, S; Kamel, F; Koroshetz, W; Langston, JW; Logroscino, G; Nelson, L; Ravina, B; Rocca, W; Ross, GW; Schettler, T; Schwarzschild, M; Scott, B; Seegal, R; Singleton, A; Steenland, K; Tanner, CM; Van Den Eeden, S; Weisskopf, M
Environmental Health Perspectives, 117(1): 117-121.
10.1289/ehp.11702
CrossRef
Movement Disorders
Transcranial sonography findings in welding-related Parkinsonism in comparison to Parkinson's disease
Walter, U; Dressler, D; Lindemann, C; Slachevsky, A; Miranda, M
Movement Disorders, 23(1): 141-145.
10.1002/mds.21795
CrossRef
Neurotoxicology
Mild steel welding fume causes manganese accumulation and subtle neuroinflammatory changes but not overt neuronal damage in discrete brain regions of rats after short-term inhalation exposure
Antonini, JM; Sriram, K; Benkovic, SA; Roberts, JR; Stone, S; Chen, BT; Schwegler-Berry, D; Jefferson, AM; Billig, BK; Felton, CM; Hammer, MA; Ma, F; Frazer, DG; O'Callaghan, JP; Miller, DB
Neurotoxicology, 30(6): 915-925.
10.1016/j.neuro.2009.09.006
CrossRef
Archives of Toxicology
Dopaminergic neurotoxicity following pulmonary exposure to manganese-containing welding fumes
Sriram, K; Lin, GX; Jefferson, AM; Roberts, JR; Chapman, RS; Chen, BT; Soukup, JM; Ghio, AJ; Antonini, JM
Archives of Toxicology, 84(7): 521-540.
10.1007/s00204-010-0525-9
CrossRef
Occupational and Environmental Medicine
Documenting neurotoxicity from occupational manganese exposure
Checkoway, H
Occupational and Environmental Medicine, 67(6): 362-363.
10.1136/oem.2009.047803
CrossRef
Clinics in Geriatric Medicine
Healthy Brain Aging: Effect of Head Injury, Alcohol and Environmental Toxins
Kumar, V; Kinsella, LJ
Clinics in Geriatric Medicine, 26(1): 29-+.
10.1016/j.cger.2009.12.006
CrossRef
Current Opinion in Neurology
Update in the epidemiology of Parkinson's disease
Elbaz, A; Moisan, F
Current Opinion in Neurology, 21(4): 454-460.
10.1097/WCO.0b013e3283050461
PDF (114) | CrossRef
Journal of Neuropathology & Experimental Neurology
The Neuropathology of Manganese-Induced Parkinsonism
Perl, DP; Olanow, CW
Journal of Neuropathology & Experimental Neurology, 66(8): 675-682.
10.1097/nen.0b013e31812503cf
PDF (101) | CrossRef
Back to Top | Article Outline

©2006The American College of Occupational and Environmental Medicine

Login

Article Tools

Images

Share