Dementia imposes a tremendous economic burden in developed countries.1,2 Alzheimer's disease accounts for over two thirds of these dementia cases in Western society. Although the causes of dementia remain largely unclear, environmental factors are believed to be involved in the development of the disease.2 In the search for such factors, occupational exposure to extremely-low-frequency magnetic fields (ELF-MFs) has been a focus of interest in numerous epidemiologic studies.3–7 In their first case–control study, Sobel et al.3 found that a primary lifetime occupation with probable exposure to a medium-to-high level of ELF-MFs was associated with a 3-fold increased risk for sporadic Alzheimer's disease. The findings were replicated in their subsequent case–control study.4 These 2 studies constitute the main epidemiologic data supporting a possible link between ELF-MF exposure and Alzheimer's disease. Another case–control analysis of the Swedish twins register data5 indicated that ELF-MF exposure level ≥0.2 microtesla (μT) in the last occupation was related to an increased risk for dementia. By contrast, other case–control studies6,7 have shown no evidence for occupational ELF-MF exposure in association with Alzheimer's disease. In addition, a few large-scale retrospective cohort studies from the United States found a weak association between “electrical occupations” and mortality of Alzheimer's disease,8–10 whereas the Danish national inpatient register data suggested that occupational exposure to electromagnetic fields was essentially unrelated to Alzheimer's disease and other central nervous system disorders.11 Recently, 2 death certificate-based studies in Sweden12,13 provided some evidence for an association between occupational ELF-MF exposure and mortality from Alzheimer's disease.
A case–control study within the Kungsholmen project in Stockholm, Sweden, suggested that manual work was associated with late-onset Alzheimer's disease, especially in men.14 The follow-up data of the project further indicated an association between manual work involving goods production and high risk of dementia.15 We hypothesized that some work-related exposures such as ELF-MFs might contribute to the development of dementia. Using the 6-year follow-up data from the same project, we examined this hypothesis by investigating the association of lifetime occupational ELF-MF exposure with occurrence of Alzheimer's disease and dementia in late life. This population-based study allows us to overcome some limitations of previous studies such as selection bias and misclassification of outcome resulting from the use of death certificates.
The study population was derived from the Kungsholmen project, which is a population-based cohort study aimed at investigating the medical, psychologic, and sociologic aspects of the aging process with an emphasis on dementia. The project has been fully described elsewhere.16,17 As shown in Figure 1, the initial population included all registered inhabitants who were at least 75 years of age and were living in the Kungsholmen district of Stockholm in October 1987. Of the 1810 baseline participants, 1473 were diagnosed as being free of dementia by a 2-phase design (1987–1989). Of these, lifetime job history information was obtained for 931 subjects at the first follow up (1991–1993). Among the 626 subjects who were alive and remained free of dementia at the first follow up, 500 subjects underwent the second follow-up examination (1994–1996). Medical records and death certificates were available for all those who died during the first follow-up period (n = 161) or the second follow-up period (n = 126). For all participants, informed consents were requested from subjects or informants (in the case of cognitively impaired persons). All phases of the project received approval from the Ethics Committee at Karolinska Institutet.
Data Collection and Assessments
At first follow up, information on occupational history was retrospectively collected from an informant, usually a next of kin, through an interview by trained nurses with a specific questionnaire developed by an occupational hygienist. The questionnaire aimed to explore the lifespan work activities by asking about employer, job title, time period, and tasks for any kind of jobs lasting at least 6 months.15
The occupational hygienist made all assessments on occupational exposure to ELF-MFs without knowledge of an individual's disease status. The assessment aimed to quantify the individual's typical exposure level during work time. Three methods were used in the assessments.
The first method was based on job-exposure matrix. Occupations throughout the work life were first coded according to the Nordic version of the International Standard Classification of Occupations in the way it was used in the Swedish census in 1980.18 If a job description fitted 2 or more relevant codes, we chose the one that was most informative of the physical aspect of the work environment. Then, the codes were matched to a Swedish job-exposure matrix that had been developed by actual measures of individual ELF-MF exposure.19 The matrix covered the 90 most frequent male occupations according to the Swedish census in 1990 and summarized the observed ELF-MF level as arithmetic mean.
Second, we used direct measurements for numerous occupations in our population, typically for females (eg, seamstresses and switchboard operators) that were not listed in the matrix. To estimate the ELF-MF exposure levels for these jobs, we performed spot measurements on historical pieces of equipment, in the position of the operator's head, using a simple dosimeter with a bandwidth of 20 to 2000 Hz. For the assessment of electrical sewing machines, we obtained household-type sewing machines commonly used between the 1930s and 1960s. The average exposure level for these machines was measured to be approximately 1.0 μT at nearly full-speed operation. The time that tailors worked with the sewing machine was estimated to be 10% of the workday, with the rest of the day at the background level (0.1 μT); their estimated daily mean exposure level was 0.19 μT. The industrial seamstresses were estimated to work at the highest exposure level of 1.0 μT for 50% of the work time and 50% at the background level, with a daily mean exposure level of 0.55 μT. For the assessment of switchboards, we obtained a manual private branch exchange with 40 lines commonly used during the 1930s to 1970s. The measured exposure level was approximately 0.5 μT at the operation. However, the amount of time being directly exposed during a workday was very different among operators. The switchboard operators in private companies (until 1970) were estimated to be exposed 10% of the workday and the rest of the day at the background level, with a workday mean exposure level of 0.14 μT. Operators employed by telephone companies in Stockholm (until 1950) had an estimated exposure time of 25% of the workday, with a daily mean exposure level of 0.20 μT. Home telephone operators (until 1970) outside Stockholm had a work time of 12 hours a day, with the estimation of daily mean exposure level of 0.21 μT. Station telephone operators (until 1970) outside Stockholm were assigned the same exposure level as the switchboard operators in a private company (0.14 μT).
Finally, for the approximately 20% of the population who had been housewives for their longest job throughout life, we used expert estimation to assess their exposure level. We disaggregated housekeeping work into various job activities (eg, cooking, cleaning, teaching, and nursing) with ELF-MF exposure estimates available in the matrix. The time spent on each activity everyday was estimated based on personal experience. The combination of time and exposure level in each activity resulted in an estimated exposure level of 0.15 μT for housewives. We also estimated the exposure levels of several less frequent occupations through substitution. Bank tellers, for example, were assigned the same level as post office clerks.
Data on age, sex, and education were collected from subjects during the baseline interview. Education level was measured by total years of formal schooling and was divided into 3 categories according to a previous study.20 Baseline arterial blood pressure was measured by trained nurses with a standardized sphygmomanometer. Information on history of vascular disorders at baseline was derived from the inpatient register system that encompassed all hospitals in Stockholm since 1969; these disorders included heart disease (International Classification of Diseases, 8th edition [ICD-8] codes 410–414, 427, and 428), cerebrovascular disease (ICD-8 codes 430–438), and diabetes mellitus (ICD-8 code 250). Data on social and mental activities were obtained from subjects through baseline interview and then categorized like in a previous study.21 Information on alcohol use and smoking habits was collected from subjects or informants during the first follow up. Genomic DNA was prepared from peripheral blood samples that were taken at baseline, and a standard procedure was used for apolipoprotein E genotyping.22
Diagnosis of Dementia and Alzheimer's Disease
The incident cases included all demented individuals who were detected over the 2 follow-up periods. At each follow-up visit, all participants underwent a structured interview, a clinical examination, and psychologic assessments. We used the Diagnostic and Statistical Manual of Mental Disorders, revised 3rd edition criteria23 to define dementia cases following a 3-step procedure24,25; 2 examining physicians independently made a preliminary diagnosis, and a third opinion was asked in the case of disagreement. The diagnosis of Alzheimer's disease required gradual onset of symptoms, progressive deterioration, and lack of any other specific causes of dementia. Our criteria for Alzheimer's disease were similar to those of the National Institute of Neurologic and Communicative Disorders and Stroke–Alzheimer's Disease and Related Disorders Association for probable Alzheimer's disease.26 For the deceased subjects, 2 physicians made a diagnosis of dementia or Alzheimer's disease through reviewing the medical records and death certificates.
Logistic regression analysis was performed to evaluate the representativeness of the study population by comparing the baseline features between dropouts and participants. We used Cox proportional-hazards models to estimate the relative risk (RR) with 95% confidence interval (CI) of Alzheimer's disease and dementia in association with occupational ELF-MF exposure in which exposures in the principal occupation (ie, lifetime longest occupational position), the last occupation, and all occupations throughout the work life were examined. The exposure levels in principal and last occupations were dichotomized (≥0.20 μT vs. <0.2 μT) following the criteria used in a previous Swedish study.5 The lifetime occupational exposure was calculated as the time-weighted average of each job throughout the work life and was divided into 3 categories following the tertiles of ELF-MF distribution by sex. The dose-response relation was assessed by treating the 3-category factor as a continuous variable. We examined the interaction of 2 factors by including the independent variables and their cross-product term in the same model. Age, sex, education, vascular disease, apolipoprotein E genotype, alcohol use, cigarette smoking, mental activity, and social activity were treated as covariates in multiple analyses. Dementia and Alzheimer's disease were used as separate outcomes.
Of the 1473 nondemented individuals at baseline, 542 were lost to the first follow up. Multiple logistic regression analysis showed that being a dropout was associated with a lower education level (OR = 1.2; 95% CI = 1.0–1.3) and less frequent social activity (0.8; 0.7–1.0), but not with age, sex, vascular disease, or mental activity.
During the 4366 person-years of follow up (median 5.4, maximum 8.3 years), 265 subjects were diagnosed as having dementia, including 202 with Alzheimer's disease. Table 1 shows the baseline characteristics of study participants by subsequent dementia status. Subjects who developed dementia during the follow-up periods were older, more often female and apolipoprotein E ε4 carriers, less educated, and less frequent participants of mental and social activities than nondemented individuals.
The ELF-MF exposure level in the lifetime principal job ranged from 0.10 to 1.90 μT for men (median 0.18 μT) and from 0.10 to 0.69 μT for women (median 0.15 μT). These levels were based on matching the principal job to 76 occupational codes for men and 85 for women. The most frequent principal job was “commercial traveler, buyer, and salesman” for men (7.3%, 0.16 μT) and “other housekeeping and related work” for women (32.0%, 0.15 μT). Figure 2 shows the cumulative distribution of ELF-MF exposure in the lifetime principal occupation by sex. There was a similar pattern for the distribution of average ELF-MF exposure in occupations throughout the work life.
For the entire cohort, there was no association of ELF-MF exposure with the risk of dementia. However, a higher ELF-MF exposure level (≥0.2 μT) in the principal occupation was associated with an increased risk of Alzheimer's disease and dementia in men but not in women (Table 2). Further analysis showed an interaction between ELF-MF exposure (≥0.2 μT vs. <0.2 μT) and sex (women vs. men) on the risk of Alzheimer's disease (multivariate-adjusted RR for the interaction term = 0.4; 95% CI = 0.2–0.8) and dementia (0.5; 0.3–1.0). No such interaction between ELF-MF exposure and apolipoprotein E genotype on dementia risk was detected.
These results could be verified in supplementary analyses. First, to explore the possibility of distortion because of subjective exposure assessment, we repeated our analysis in subjects whose exposure was assessed according to the job-exposure matrix (n = 495; 102 Alzheimer's disease and 136 dementia cases). Among men (n = 181; 28 Alzheimer's disease and 45 dementia cases), ELF-MF exposure ≥0.20 μT was associated with multivariate-adjusted RRs of 2.0 (95% CI = 0.8–4.9) for Alzheimer's disease and 1.6 (0.8–3.0) for dementia. Among women, the corresponding figures were 0.9 (0.6–1.6) and 0.9 (0.6–1.5), respectively. Second, we restricted the analysis to the dementia-free cohort identified at the first follow up in whom information on job history was collected before dementia diagnosis (n = 626; 87 Alzheimer's disease and 111 dementia cases). For men (n = 156; 18 Alzheimer's disease and 25 dementia cases), ELF-MF exposure ≥0.20 μT was associated with multivariate-adjusted RRs of 4.7 (1.5–15.1) for Alzheimer's disease and 4.1 (1.6–10.6) for dementia, whereas among women, the figures were 0.8 (0.4–1.4) and 0.7 (0.4–1.3), respectively. Finally, similar results were also obtained from the analysis in survivors at either follow-up examination (data not shown).
A similar sex-specific pattern was seen for the association of average ELF-MF exposure throughout the work life with risk of Alzheimer's disease and dementia (Table 2). In men, there was a dose-response relation between exposure level and dementia risk, especially when several potential confounders were adjusted for. A higher ELF-MF exposure level was related to an increased risk for Alzheimer's disease, also, although the dose-response relation was less evident. In women, no association between ELF-MF exposure and dementia risk was noted. When the analyses were restricted to the dementia-free cohort identified at first follow up, exposure levels of 0.16 to 0.21 μT and >0.21 μT in men were associated with multivariate-adjusted RRs of 3.2 (0.7–14.0) and 5.4 (1.2–24.3) for Alzheimer's disease, and 1.3 (0.4–4.5) and 4.5 (1.4–14.8) for dementia compared with a level <0.16 μT. The less precise estimates of RR were largely the result of fewer cases, especially in the referent category (only 3 Alzheimer's disease and 6 dementia cases). Among women, all RR estimates were close to 1.0.
We found no association of ELF-MF exposure in the last occupation with the risk of Alzheimer's disease and dementia. For example, the multivariate-adjusted RRs related to a higher ELF-MF exposure level (≥0.2 μT vs. <0.2 μT) in the last job were 1.1 (0.5–2.4) for Alzheimer's disease and 1.3 (0.7–2.5) for dementia in men, whereas in women, the corresponding figures were 0.9 (0.6–1.2) and 0.9 (0.7–1.2).
This community-based cohort study showed that long-term occupational exposure to a higher ELF-MF level was associated with an elevated risk of Alzheimer's disease and dementia in men, but not in women. The association between average occupational ELF-MF exposure throughout the work life and risk of dementia was more evident when potential confounders such as vascular disorders, lifestyles and habits, and apolipoprotein E genotype were taken into account.
Previous case–control studies3,4 showed that Alzheimer's disease was strongly associated with the lifetime primary occupations that had likely exposure to a medium-to-high level of electromagnetic fields, particularly in occupations involving sewing machines, in which women were the predominant employees. However, the assessment of magnetic field exposure in these studies was not quantitative and the results were heavily dependent on limited number of jobs. By contrast, we carefully assessed the ELF-MF exposure levels in a wide variety of occupations. We found a moderately strong association of ELF-MF exposure in primary occupation with the risk of Alzheimer's disease and dementia in men, but not in women. This finding is in line with a few large-scale death certificate-based studies that indicated men who worked in electrical occupations (eg, electricians, electronic technicians, and power plant operators) had increased mortality of Alzheimer's disease.8–10,12,27 In the Swedish twins study, Feychting and colleagues5 showed that a higher exposure level of magnetic fields (≥0.2 μT) in the last occupation, rather than in the primary occupation and the occupation with the highest magnetic field exposure, was associated with an elevated risk for Alzheimer's disease and dementia, especially in people age 75 years and younger. No appreciable sex difference in the estimated odds ratios was noted in this twins study. In their more recent study, which was based on the Swedish nationwide death certificate data of working population, Feychting et al.12 showed that ELF-MF exposure was related to mortality of early-onset Alzheimer's disease, particularly for men who died before 75. The magnetic field exposure was believed to represent a late-acting influence on the disease process. An increased mortality risk of Alzheimer's disease related to occupational exposure to ELF-MF was also observed in the cohort of Swedish engineering industrial workers for both men and women.13 Our study, which was conducted in people at least 75 years of age, showed no association between ELF-MF exposure in the last job and risk of dementia, whereas ELF-MF exposure in the primary occupation was related to an elevated incidence of late-onset Alzheimer's disease and dementia in men. Thus, our data are not supportive of the notion that magnetic field exposure may represent a late-acting influence on late-onset Alzheimer's disease. On the contrary, the cumulative exposure throughout the work life may be more important.
Other studies6,7,28,29 provide little support for an association between magnetic field exposure and dementia. For example, a case-referent study7 based on death records of men did not show any association of Alzheimer's disease with occupational magnetic field exposure that was assessed with either a qualitative method (ie, electrical vs. nonelectrical work) or a quantitative method (ie, using a job-exposure matrix). Another case–control study29 found that neither occupational nor residential exposures to power-frequency electromagnetic fields were associated with cognitive impairment.
The biologic pathways by which the ELF-MF exposure might precipitate the Alzheimer pathologic changes are currently unknown. It has been hypothesized that ELF-MFs may be ultimately implicated in the neurodegenerative process by affecting calcium ion homeostasis at cellular level30 or by a direct reaction with specific DNA sites at gene level.31,32 Our data show no modifying effect of apolipoprotein E ε4 allele on the association between ELF-MF exposure and risk of dementia, although apolipoprotein E has been suggested to promote the role of electromagnetic fields in the pathogenesis of Alzheimer's disease.4,30
Strengths of this study include the community-based design, relatively long-term follow up, comprehensive assessment of ELF-MF exposure, and adjustment for multiple potential confounders. Nevertheless, our study has limitations. First, the data on job history were retrospectively collected from informants, which might have resulted in information bias, although a Swedish study shows that data on lifetime occupational history retrospectively collected with a questionnaire are in good agreement with the national census data.33 Second, the job-exposure matrix that covered a larger variety of occupations has been used in Sweden to evaluate a possible role of electromagnetic field exposure in cancers and neurodegenerative disorders. Although the matrix has not been formally validated with additional measurements, the defined ELF-MF exposures (≥0.2 μT) were in accord with those reported from the utility industry.34 However, the misclassification of ELF-MF exposure is probably greater for women, because the matrix had been developed specifically for male jobs35 and women could be categorized in fewer job titles. Third, we aimed to examine the relation between occupational ELF-MF exposure and risk of Alzheimer's disease and dementia. We were not able to control for the possible confounding of the residential exposure that may equal the magnitude of certain occupational exposures in some cases. Finally, although multiple factors were adjusted for, we cannot rule out the possibility that our results were affected by other potential confounders such as dietary habits and other occupational exposures.36–39 Neither can it be ruled out that some high ELF-MF exposure jobs had a low demand on mental activity that might have not been adequately controlled for by just using the information on education in early life and mental activity in late life.
Our study provides limited evidence to the potential role of long-term occupational ELF-MF exposure in the development of Alzheimer's disease and dementia in men but not in women. A generally lower ELF-MF exposure level combined with a greater misclassification of the exposure for women may be responsible for the sex difference. These findings emphasize the need of more studies to clarify the role of electromagnetic field exposure in Alzheimer's disease and dementia.
We thank the Museum of Technology and the Museum of Telecommunications in Stockholm, Sweden, for kindly providing the historical pieces of equipment (electronic sewing machines and switchboards), and the members of the Kungsholmen Project Study Group for their cooperation in data collection and management.
1. Brookmeyer R, Grey S, Kawas C. Projections of Alzheimer's disease in the United States and the public health impact of delaying disease onset. Am J Public Health
2. Fratiglioni L, Rocca WA. Epidemiology of dementia. In: Boller F, Cappa SF, eds. Handbook of Neuropsychology,
2nd ed, vol 6: Aging and Dementia
. Amsterdam: Elsevier Science BV Publisher; 2001:193–215.
3. Sobel E, Davanipour Z, Sulkava R, et al. Occupations with exposure to electromagnetic fields: a possible risk factor for Alzheimer's disease. Am J Epidemiol
4. Sobel E, Dunn M, Davanipour Z, et al. Elevated risk of Alzheimer's disease among workers with likely electromagnetic field exposure. Neurology
5. Feychting M, Pedersen NL, Svedberg P, et al. Dementia and occupational exposure to magnetic fields. Scand J Work Environ Health
6. Graves AB, Rosner D, Echeverria D, et al. Occupational exposure to electromagnetic fields and Alzheimer disease. Alzheimer Dis Assoc Disord
7. Noonan CW, Reif JS, Yost M, et al. Occupational exposure to magnetic fields in case-referent studies of neurodegenerative diseases. Scand J Work Environ Health
8. Schulte PA, Burnett CA, Boeniger MF, et al. Neurodegenerative diseases: occupational occurrence and potential risk factors, 1982 through 1991. Am J Public Health
9. Savitz DA, Checkoway H, Loomis DP. Magnetic field exposure and neurodegenerative disease mortality among electric utility workers. Epidemiology
10. Savitz DA, Loomis DP, Tse CJ. Electrical occupations and neurodegenerative disease: analysis of US mortality data. Arch Environ Health
11. Johansen C. Exposure to electromagnetic fields and risk of central nervous system disease in utility workers. Epidemiology
12. Feychting M, Jonsson F, Pedersen NL, et al. Occupational magnetic field exposure and neurodegenerative disease. Epidemiology
13. Håkansson N, Gustavsson P, Johansen C, et al. Neurodegenerative diseases in welders and other workers exposed to high levels of magnetic fields. Epidemiology
14. Fratiglioni L, Ahlbom A, Viitanen M, et al. Risk factors for late-onset Alzheimer's disease: a population-based, case-control study. Ann Neurol
15. Qiu C, Karp A, von Strauss E, et al. Lifetime principal occupation and risk of Alzheimer's disease in the Kungsholmen project. Am J Ind Med
16. Fratiglioni L, Viitanen M, Bäckman L, et al. Occurrence of dementia in advanced age: the study design of the Kungsholmen project. Neuroepidemiology
. 1992;11(suppl 1):S29–S36.
17. Fratiglioni L, Viitanen M, von Strauss E, et al. Very old women at highest risk of dementia and Alzheimer's disease: incidence data from the Kungsholmen project, Stockholm. Neurology
18. Folk och Bostadsräkningen 1980. Statistiska centralbyrån, Enheten för befolkningsstatistik
[in Swedish]. Örebro, Sweden; 1983.
19. Floderus B, Persson T, Stenlund C. Magnetic-field exposures in the workplace: reference distribution and exposures in occupational groups. Int J Occup Environ Health
20. Qiu C, Bäckman L, Winblad B, et al. The influence of education on clinically diagnosed dementia: incidence and mortality data from the Kungsholmen project. Arch Neurol
21. Wang H-X, Karp A, Winblad B, et al. Late-life engagement in social and leisure activities is associated with a decreased risk of dementia: a longitudinal study from the Kungsholmen project. Am J Epidemiol
22. Basun H, Corder EH, Guo Z, et al. Apolipoprotein E polymorphism and stroke in a population sample aged 75 years or more. Stroke
23. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders
, revised 3rd edition (DSM-III-R). Washington, DC: American Psychiatric Association; 1987.
24. Fratiglioni L, Grut M, Forsell Y, et al. Clinical diagnosis of Alzheimer's disease and other dementias in a population survey: agreement and causes of disagreement in applying Diagnostic and Statistical Manual of Mental Disorders
, revised 3rd ed criteria. Arch Neurol
25. Fratiglioni L, Wang H-X, Ericsson K, et al. Influence of social network on occurrence of dementia: a community-based longitudinal study. Lancet
26. McKhann G, Drachman D, Folstein M, et al. Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Service Task Force on Alzheimer's Disease. Neurology
27. Schulte PA, Burnett CA. EMFs and Alzheimer's disease. Neurology
28. Ahlbom A. Neurodegenerative diseases, suicide and depressive symptoms in relation to EMF. Bioelectromagnetics.
29. Li CY, Sung FC, Wu SC. Risk of cognitive impairment in relation to elevated exposure to electromagnetic fields. J Occup Environ Med
30. Sobel E, Davanipour Z. Electromagnetic field exposure may cause increased production of amyloid beta and eventually lead to Alzheimer's disease. Neurology
31. Blank M, Goodman R. Electromagnetic initiation of transcription at specific DNA sites. J Cell Biochem
32. Rao RR, Halper J, Kisaalita W. Effects of 60 Hz electromagnetic field exposure on APP695 transcription levels in differentiating human neuroblastoma cells. Bioelectrochemistry
33. Wärneryd B, Thorslund M, Östlin P. The quality of retrospective questions about occupational history: a comparison between survey and census data. Scand J Soc Med
34. Savitz DA, Ohya T, Loomis DP, et al. Correlations among indices of electric and magnetic field exposure in electric utility workers. Bioelectromagnetics
35. Deadman JE, Infante-Rivard C. Individuals estimation of exposure to extremely low frequency magnetic fields in jobs commonly held by women. Am J Epidemiol
36. Engelhart MJ, Geerlings MI, Ruitenberg A, et al. Dietary intake of antioxidants and risk of Alzheimer disease. JAMA
37. Morris MC, Evans DA, Bienias JL, et al. Consumption of fish and n-3 fatty acids and risk of incident Alzheimer disease. Arch Neurol
38. Kukull WA, Larson EB, Bowen JD, et al. Solvent exposure as a risk factor for Alzheimer's disease: a case–control study. Am J Epidemiol
39. Baldi I, Lebailly P, Mohammed-Brahim B, et al. Neurodegenerative diseases and exposure to pesticides in the elderly. Am J Epidemiol