• lead;
  • occupation;
  • brain neoplasms;
  • cohort studies


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We evaluated the association between potential occupational lead exposure and the risk of brain cancer mortality in the National Longitudinal Mortality Study (NLMS), which is a prospective census-based cohort study of mortality among the noninstitutionalized United States population (1979–1989). The present study was limited to individuals for whom occupation and industry were available (n = 317,968). Estimates of probability and intensity of lead exposure were assigned using a job-exposure matrix (JEM). Risk estimates for the impact of lead on brain cancer mortality were computed using standardized mortality ratio (SMR) and proportional hazards and Poisson regression techniques, adjusting for the effects of age, gender and several other covariates. Brain cancer mortality rates were greater among individuals in jobs potentially involving lead exposure as compared to those unexposed (age- and gender-adjusted hazard ratio (HR) = 1.5; 95% confidence interval (CI) = 0.9–2.3) with indications of an exposure–response trend (probability: low HR = 0.7 (95% CI = 0.2–2.2), medium HR = 1.4 (95% CI = 0.8–2.5), high HR = 2.2 (95% CI = 1.2–4.0); intensity: low HR = 1.2 (95% CI = 0.7–2.1), medium/high HR = 1.9 (95% CI = 1.0–3.4)). Brain cancer risk was greatest among individuals with the highest levels of probability and intensity (HR = 2.3; 95% CI = 1.3–4.2). These findings provide further support for an association between occupational lead exposure and brain cancer mortality, but need to be interpreted cautiously due to the consideration of brain cancer as one disease entity and the absence of biological measures of lead exposure. © 2006 Wiley-Liss, Inc.

Exposure to lead compounds is predominantly due to anthropogenic activity,1, 2 and has long been suspected to result in chronic health effects.1 The greatest potential for exposure has been experienced by industrial workers, and lead exposure is currently generally well controlled in major lead-using industries such as smelting and battery manufacturing industries.1, 2, 3 A 5- to 10-fold decline in median and 75th percentile of lead exposure in general industry has been reported between 1979 and 1997.4 Nevertheless, little to no decreases in lead exposure levels have been observed in certain work environments such as the construction industry,4 and cases of clinical lead poisoning in certain industries still occur.1 Historically, the largest source of environmental lead exposure in the United States was through inhalation and ingestion of air, dust, soil, water and food contaminated from the use of lead in pipes, paints, food and drink cans and gasoline. These uses have been phased out in many developed countries, and geometric mean blood lead levels among adults in the United States have declined from 13.1 μg/dL (0.63 μmol/L) in the late 1970s to 1.64 μg/dL (0.08 μmol/L) in 1999–2002.5 However, sections of the general population continue to be exposed to excessive amounts of lead, especially from lead-based paints and contaminated soil in urban settings with an older housing stock.5, 6, 7, 8, 9, 10, 11, 12, 13 Additionally, lead accumulates in the body which may become biologically available long after the occupational or environmental exposure has ceased.2, 14, 15, 16 Therefore, lead exposure is still a public health concern.

Although the etiology of brain cancer remains largely unknown,17, 18, 19, 20, 21 there are several clues that exposure to lead may impact brain cancer risk. Lead has been shown to pass the blood–brain barrier,22 which may result in elevated lead levels in brain tissue.23 Lead is thought to play a facilitative role in carcinogenesis, involving inhibition of DNA synthesis and repair, oxidative damage and interaction with DNA-binding proteins and tumor suppressor proteins.2, 24, 25 Additionally, brain tissues are reported to be relatively susceptible to oxidative stress and lipid peroxidation,17 suggesting that the brain may be sensitive to the carcinogenic effects of lead. Experimental studies reporting an increased incidence of brain tumors in rats fed lead salts support this hypothesis.26, 27, 28 On the other hand, the epidemiological literature for an association between lead exposure and brain cancer is inconclusive. Nonetheless, several studies evaluating brain tumor subtypes or relying on (semi-) quantitative measures of exposure reported findings indicative of an association.29, 30, 31, 32, 33, 34

We assessed whether employment in occupations potentially involving exposure to lead compounds is related to an increased risk of mortality from brain cancer in the National Longitudinal Mortality Study (NLMS). The NLMS is a prospective census-based cohort study of mortality among the noninstitutionalized United States population, conducted by the National Heart, Lung, and Blood Institute in collaboration with the National Center for Health Statistics and the United States Bureau of the Census.35, 36

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Study population

The NLMS public-use data file consists of a national sample of the United States population (n = 637,162), as identified from the Current Population Survey (CPS) of the Bureau of the Census of March 1979, April 1980, August 1980, December 1980 and March 1981 ( The full study is larger, involving 1.3 million persons,35, 36 but for confidentiality reasons, a subset of 5 samples closely reflecting the full NLMS database of the original 12 CPSs is provided for the limited access data set which is publicly available. All personal identifiers, geographical references, specific NLMS cohort references and specifically identifiable time components were removed from this public-use file. The present study is limited to individuals for whom occupation or industry codes were available, leaving a total of 317,968 individuals for the current analysis. Occupation or industry codes were missing for individuals who were not in the labor force (i.e., age < 14 (45% of missing occupation or industry), home maker (27%), retired (16%), student (9%), inability to work (2%) or unemployed but actively looking for work (1%)). Records from the 5 CPS cohorts were matched to the national death index (NDI) to identify the occurrence and cause of death for each individual cohort member. Mortality follow-up in the public-use dataset was restricted to the period 1979–1989. Causes of death were classified according to ICD-9 codes; our analysis was limited to mortality from brain cancer (ICD-9 191).

Exposure assessment

Participants in the CPS employed at the time of the survey were asked about the job worked during the week preceding the survey. For persons unemployed but actively looking for work within the 4 week period prior to the survey, information was obtained for the most recent job held (if any) within 5 years of the survey.37 Occupation and industry reported in the CPS were assigned 3-digit codes according to the 1970 US Bureau of the Census classification system of jobs and industries. On the basis of these codes, levels of exposure to lead were assigned to the current (for employed individuals) or most recent (for job seekers) job held, using a job-exposure matrix (JEM) for lead.31 This JEM was previously developed by an industrial hygienist (M.D.) on the basis of information from the published literature, computerized exposure databases, unpublished industrial hygiene reports and the industrial hygienist's personal experience.31 An estimate of intensity level (none = 0, low = 1, medium = 2, high = 3) and probability (none = 0, low = 1, medium = 2, high = 3) was assigned to each individual's 3-digit occupation and industry code. Intensity estimates reflected average blood lead levels of less than 0.9 (low intensity), 0.9–1.4 (medium intensity) and greater than 1.4 μmol/L (high intensity).31 The probability of exposure was estimated on the basis of the proportion of exposed workers within a given job title or industry and the number of occupations or industries coded likewise.38 If exposure was determined by the occupation itself regardless of industry, final intensity and probability scores were obtained by squaring the occupational scores. On the other hand, if exposure was determined by both occupation and industry, then the final probability and intensity score was based on multiplying the scores of occupation and industry.31 Finally, the final probability and intensity scores were further grouped on the basis of 4 a priori selected categories (none = 0, low = 1–2, medium = 3–4, high = ≥6).31 Almost 19% of cohort members (n = 59,352) were considered potentially exposed to lead in their jobs. An overview of occupations most commonly assigned possible lead exposure in this cohort is presented in Table I. The distribution of exposed jobs across levels of probability and intensity is shown in Table II.

Table I. Percentage of Common Occupations and Industries among Individuals with Possible Occupational Lead Exposure1: National Longitudinal Mortality Study 1979–1989
Exposure levelCommon occupationsCommon industries
1970 CodeTitle%21970 CodeTitle%2
  • 1

    Total individuals with potential lead exposure = 59,352 (18.7% of all 317,968 individuals in cohort).

  • 2

    Percent of exposed individuals within each probability or intensity category.

 Low (N = 12,763)961Firemen, fire protection5.2937Local public administration5.9
153Electrical and electronic engineering technicians6.469Special trade contractors7.8
753Freight and material handlers19.4
 Medium (N = 27,718)964Policemen and detectives5.1937Local public administration5.4
962Guards and watchmen6.5417Trucking service9.3
705Deliverymen and routemen6.7
715Truck drivers22.0
 High (N = 18,871)623Garage workers and gas station attendants6.7639Motor vehicle dealers5.3
522Plumbers and pipe fitters7.5648Gasoline service stations6.7
321Estimators and investigators, n.e.c.8.0757Automobile repair and related services8.3
510Painters, construction and maintenance8.669Special trade contractors12.5
680Welders and flamecutters12.1
481Heavy equipment mechanics, incl. diesel15.3
473Automobile mechanics16.9
 Low (N = 37,219)705Deliverymen and routemen5.069Special trade contractors5.2
430Electricians5.1937Local public administration6.0
753Freight and material handlers6.6417Trucking service8.2
715Truck drivers16.3
 Medium (N = 13,625)492Miscellaneous mechanicsand repairmen5.8648Gasoline service stations6.7
640Mine operatives, n.e.c.7.3
623Garage workers and gas station attendants9.3
321Estimators and investigators, n.e.c.11.0
680Welders and flamecutters16.7
481Heavy equipment mechanics, incl. diesel21.2
 High (N = 8,508)530Pressmen and plate printers, printing5.7639Motor vehicle dealers7.7
422Compositors and typesetters6.5339Printing, publishing and allied industries, except newspapers8.3
644Painters, manufactured articles6.6757Automobile repair and related services13.6
190Painters and sculptors7.369Special trade contractors20.8
522Plumbers and pipe fitters16.6
510Painters, construction and maintenance19.0
473Automobile mechanics37.6
Table II. Distribution of Lead-Exposed Jobs by Levels of Probability and Intensity: National Longitudinal Mortality Study 1979–1989
Probability levelIntensity levelTotal
 Observed number of brain cancer deaths3003
 Population at risk12,096664312,763
 Observed number of brain cancer deaths130013
 Population at risk24,6561,3911,67127,718
 Observed number of brain cancer deaths010313
 Population at risk46711,5706,83418,871
 Observed number of brain cancer deaths1610329
 Population at risk37,21913,6258,50859,352

Potential confounding factors

Several variables were considered potential confounders on the basis of previously identified risk factors, including gender, age, race, living in an urban area, marital status and educational level.31 Age was categorized into 6 groups (<35, 35–44, 45–54, 55–64, 65–74 and >75 years of age), and race was classified as white or nonwhite. Residential location was classified as urban or rural according to the 1970 Census definition, where persons living in urbanized areas and places with a population of 2,500 or more outside urbanized areas were classified as urban dwellers; others were considered to be living in rural areas. Marital status was grouped into ever (currently married, separated, widowed or divorced) or never married, whereas 3 education levels were considered on the basis of the highest grade completed (less than high school, some high school or high school graduate and some college). Annual family income (<$15,000, $15,000–$24,999, >$25,000) was also considered as a potential confounder, but information on this variable was missing for almost 5% of individuals (n = 15,258). Therefore, we did not include income as a covariate in our analyses. However, results from analyses limited to individuals with information on all covariates (including income) were very similar to the findings presented here (data not shown).

Statistical analysis

The design of the public-use file is that the follow-up time starts for all records at the same hypothetical date and continues as indicated to mortality or censorship.39 To be censored for this file means that the record was determined to be of a person alive at the end of the 9 years of follow-up. To assess the robustness of our findings to statistical assumptions inherent in the use of regression models, we evaluated the association between potential occupational lead exposure and brain cancer mortality with a variety of techniques to estimate the relative risk and corresponding 95% confidence interval (CI), including proportional hazards regression and grouped data analysis methods. Grouped data analyses included external adjustment methods using standardized mortality ratios (SMRs), and internal comparison rate ratios (RRs) using Poisson regression techniques. Findings from nonparametric Mantel–Haenszel RR analyses40 were very similar to those from Poisson regression, and are therefore not reported here. Analyses were performed using SAS version 8.2 (SAS Institute, Cary, NC).

Proportional hazards analysis

Initially, hazard ratios (HRs) and 95% CI comparing the risk of brain cancer among lead-exposed and unexposed individuals were estimated with the Cox proportional hazards model41 (exact method) using the SAS PHREG procedure. Follow-up time was treated as the fundamental time variable, adjusting for the effect of age by covariate modeling on the basis of individuals' age at entry into the study.42 A reduced regression model adjusted for the effects of age (continuous) and gender to facilitate comparison with the results obtained from the analysis of grouped data (see below). Furthermore, a full model was employed to control for age (continuous), gender (male or female), race (white or nonwhite), urban status (urban or rural), marital status (ever or never married) and education level (<any high school, some high school or some college). The full model only included individuals with complete information on all covariates; therefore, 2,106 subjects (0.7% of the eligible cohort) were excluded due to missing data. Results from the grouped data analysis and the reduced Cox proportional hazards regression model were very similar; therefore, only findings from the full Cox model are presented. The proportionality assumption for lead exposure and potential confounders was checked graphically by inspecting the log of the negative log of survival (i.e., ln[−ln|S(t)|]) against survival time for the covariate categories after adjustment for other covariates.39, 43 If the proportionality assumption holds, the difference in ln[−ln|S(t)|] over any 2 or more levels of the covariate should be approximately constant over the follow-up time period.39, 43 Although interpretation was limited due to sparse data, the log–log plots suggested that the proportional hazards assumption was violated for lead exposure and several other covariates. Therefore, we performed additional analyses of grouped data to assess the robustness of the association.

Analysis of grouped data

Several basic time indicators are needed to compute person-time for the analysis of grouped data in cohort studies, including date of birth, date of study entry and date of last observation.44 This information was not available for the cohort members in the public-use file; therefore, these dates were created on the basis of the reported age at the time of survey, a hypothetical start of follow-up (assumed to be July 1, 1980 for all cohort members) and the number of days of follow-up (with a maximum of 9 years or 3,288 days). Date of birth was created by subtracting the age from the hypothetical start date, and by subtracting an additional 6 months to account for the variability in day of birth throughout a calendar year. For example, the birth date of an individual aged 20 at the start of follow-up (i.e., July 1, 1980) was assumed to be January 1, 1960. The end of follow-up was determined by adding the number of days of follow-up to the hypothetical start date. That is, for a person alive at the end of 9 years of follow-up the end date was July 1, 1989.

Person-year data were generated according to the method described by Wood et al. using SAS in which time-dependent variables are accurately classified at each interval of observation.44 Subsequently, SMRs were computed as the ratio of observed over expected brain cancer deaths, where the number of expected deaths was based on 1980–1989 average mortality rates in the general United States population reported by the Centers for Disease Control and Prevention ( The 95% CIs were based on the Poisson distribution of the observed numbers of deaths.45 Additionally, ratios of SMRs (RSMR) and corresponding 95% CIs were computed.46 Finally, RRs and associated 95% CI were estimated with multivariate Poisson regression models using the SAS GENMOD procedure.47, 48 SMRs and RRs were adjusted for age (<35, 35–44, 45–54, 55–64, 65–74, >75) and gender (male or female).


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Descriptive characteristics of the current study cohort are shown by lead exposure status in Table III. Exposed and unexposed subjects were similar in most respects at baseline, i.e., the majority was less than 35 years of age (51%), white (88%), lived in an urban setting (67%) and were married at baseline or had been married previously (74%). The distribution of gender and education differed somewhat by exposure status with exposed individuals more likely than unexposed subjects to be men (85% vs. 49%) and less educated (22.5% vs. 38.4% more than high school). Analysis of crude brain cancer rates among the unexposed showed an association with age (elevated rates in older people), marital status (married individuals at higher risk) and education (more educated subjects at lower risk). Crude RRs by gender, race and urban status demonstrated that men, whites and rural dwellers may be at increased risk (Table III).

Table III. Descriptive Characteristics of Cohort: National Longitudinal Mortality Study 1979–1989
Demographic variableUnexposedExposed
Population-at-risk at baseline (%)Person-yearsObs1Crude rate ratio (95% CI)Population-at-risk at baseline (%)Person-yearsObsCrude rate ratio (95% CI)
  • 1

    Observed number of brain cancer deaths.

  • 2

    This age group was chosen as the referent category because it comprises the largest number of observed brain cancer deaths.

  • 3

    HS, high school.

Total258,616 (100)2,288,835.419059,352 (100)522,835.3329
 Men127,757 (49.4)1,122,964.24481.0   (ref)50,150 (84.5)440,816.00261.0   (ref)
 Women130,859 (50.6)1,165,871.17420.84 (0.56–1.28)9,202 (15.5)82,019.3330.62 (0.19–2.05)
 <35132,141 (51.1)919,305.7860.05 (0.02–0.11)30,527 (51.4)209,119.1920.06 (0.01–0.28)
 35–4447,343 (18.3)511,682.2570.10 (0.05–0.23)11,424 (19.3)123,323.2050.26 (0.09–0.75)
 45–5439,886 (15.4)373,033.21120.24 (0.13–0.46)9,055 (15.3)86,959.6460.44 (0.16–1.19)
 55–64229,554 (11.4)315,170.75421.0   (ref)6,675 (11.3)70,392.44111.0   (ref)
 65–748,301 (3.2)140,968.11180.96 (0.55–1.66)1,469 (2.5)28,742.5540.89 (0.28–2.80)
 75+1,391 (0.5)28,675.3251.31 (0.52–3.31)202 (0.3)4,298.3211.49 (0.19–11.53)
 White228,631 (88.4)2,024,076.00831.0   (ref)52,379 (88.3)461,512.00291.0   (ref)
 Non-white29,985 (11.6)264,759.4170.64 (0.30–1.39)6,973 (11.8)61,323.330
Urban status
 Urban174,822 (67.6)1,547,160.58551.038,093 (64.2)335,203.38151.0
 Rural83,794 (32.4)741,674.83351.33 (0.87–2.03)21,259 (35.8)187,631.96141.67 (0.80–3.45)
Marital status
 Never67,068 (25.9)598,974.97101.013,360 (22.5)119,096.0011.0
 Ever189,573 (73.3)1,672,181.31802.87 (1.48–5.53)45,861 (77.3)402,571.22288.28 (1.13–60.9)
 <Some HS21,639 (8.37)186,577.49181.07,013 (11.8)60,076.9051.0
 Some HS137,641 (53.2)1,219,557.79370.31 (0.18–0.55)38,984 (65.7)344,258.91150.52 (0.19–1.44)
 >HS99,238 (38.4)881,826.95350.41 (0.23–0.73)13,330 (22.5)118,296.8490.91 (0.31–2.73)

Brain cancer mortality rates were greater among individuals in jobs potentially involving lead exposure, with a crude RR of 1.41 (95% CI = 0.93–2.14) comparing workers with any exposure to those without exposure. Adjustment for potential confounders either by proportional hazards or Poisson regression did not greatly impact this association (Table IV). The HR for the reduced and full model were 1.46 (95% CI = 0.94–2.26; data not shown) and 1.56 (95% CI = 1.00–2.43), respectively, whereas the Poisson regression RR was 1.42 (95% CI = 0.91–2.20). Regression analyses showed an exposure–response relationship between brain cancer risk and lead exposure, with rates increasing as probability and intensity of exposure increased (Table III). HRs or RRs were strongest when individuals with the highest levels of exposure (i.e., high probability and medium/high intensity) were compared to those employed in jobs unlikely to involve lead exposure (reduced model HR = 2.28; 95% CI = 1.25–4.18, full model HR = 2.39; 95% CI = 1.29–4.41). As an alternative presentation of the exposure–response relationship, we also computed risk estimates for individuals in jobs with low probability and intensity (full model HR = 0.76; 95% CI = 0.24–2.41); low probability and medium/high intensity (no observed deaths); medium/high probability and low intensity (full model HR = 1.61; 95% CI = 0.89–2.94); and medium/high probability and intensity (full model HR = 2.06; 95% CI = 1.12–3.77).

Table IV. Adjusted Risk Estimates and 95% Confidence Intervals (CI) for the Association between Levels of Occupational Lead Exposure and Brain Cancer Mortality: National Longitudinal Mortality Study 1979–1989
Exposure levelPopulation-at-riskPerson-yearsObs1HR (CI)2Analysis of grouped data3
  • 1

    Observed number of brain cancer deaths.

  • 2

    HR, hazard ratio; adjusted for age (continuous), gender (male or female), race (white or nonwhite), urban status (urban or rural), marital status (ever or never married), and education level (<any high school, some high school, some college)—complete case analysis including subject with complete information on all covariates (excluded subjects n = 2,106).

  • 3

    SMR, standardized mortality ratio; RR, Poisson regression rate ratio; risk estimates adjusted for age (<35, 35–44, 45–54, 55–64, 65–74, >75) and gender (male or female).

Not exposed258,6162,288,835901.0 (ref)0.87 (0.70–1.06)1.0 (ref)
Any exposure59,352522,835291.56 (1.00–2.43)1.11 (0.74–1.59)1.42 (0.91–2.20)
 Low12,763112,44830.72 (0.23–2.30)0.50 (0.10–1.47)0.65 (0.20–2.06)
 Medium27,718243,885131.47 (0.81–2.68)1.06 (0.56–1.81)1.34 (0.74–2.43)
 High18,871166,502132.35 (1.28–4.32)1.64 (0.87–2.80)2.12 (1.17–3.87)
Intensity (any probability)
 Low37,219327,667161.33 (0.77–2.31)0.95 (0.54–1.54)1.21 (0.70–2.09)
 Medium/high22,133195,169131.99 (1.09–3.66)1.39 (0.74–2.38)1.81 (0.99–3.29)
 Medium13,625120,181102.50 (1.27–4.92)1.77 (0.85–3.25)2.28 (1.17–4.44)
 High8,50874,98831.19 (0.37–3.80)0.82 (0.17–2.39)1.07 (0.33–3.40)
Intensity (probability > low)
 Low25,123221,094131.61 (0.88–2.92)1.16 (0.62–1.99)1.48 (0.82–2.67)
 Medium/high21,466189,294132.05 (1.12–3.76)1.44 (0.77–2.46)1.87 (1.03–3.42)
Intensity (probability > medium)
 Low4674,17300.00 (0.00–46.1)
 Medium/high18,404162,330132.39 (1.29–4.41)1.66 (0.88–2.83)2.21 (1.21–4.04)

Findings using SMRs instead of regression-based risk estimates were less clear-cut, with little indication of an elevated brain cancer risk among exposed subjects relative to the general United States population (SMR = 1.11; 95% CI = 0.74–1.59). Nevertheless, patterns of risk with increasing levels of exposure were similar to those observed with regression analyses. The ratio of SMRs comparing workers with the highest levels of exposure (i.e., probability > medium and medium/high intensity) to those unexposed was only slightly lower than corresponding HR or RR estimates (RSMR = 1.66/0.87 = 1.91; 95% CI = 0.98–3.43).


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Despite decades of active epidemiological research, little progress has been made in conclusively identifying preventable risk factors for tumors of the brain and nervous system.17, 18, 19, 20, 21, 22, 49 The prevalence of established risk factors (i.e., therapeutic radiation, certain inherited genes17, 18, 19, 20, 21, 22, 49) in the general population is rare; therefore, these factors are of limited public health relevance to explain the majority of brain cancers or suggest preventive measures.21, 22 However, it has been suggested that moderate risks cannot be excluded for most occupational and environmental exposures.18

The epidemiological evidence for the carcinogenicity of lead has been reviewed on multiple occasions,2, 50, 51, 52, 53, 54 and is summarized in Table V. Steenland and Boffetta in 2000 considered 6 occupational cohort studies29, 30, 58, 60, 61, 62, 63 particularly informative with respect to brain cancer risk because of high documented exposures.51 Some of these studies reported an elevated risk for brain cancer, in particular in the highest exposed subgroups,29, 61, 62 whereas others found little evidence for an association.58, 60, 63 They computed an overall relative risk of 1.06 (95% CI = 0.8–1.4) using meta-analytic techniques, and concluded that the evidence for excess brain cancer risk is weak despite some animal evidence. However, they pointed out that support for an association was provided by a death certificate-based study31 and a case–control study of gliomas nested within an occupational cohort,29 and concluded that “brain cancer remains a concern.”51 Several additional population-based studies reported an increased meningioma risk of about 2-fold or greater among subjects possibly exposed to lead.32, 34, 69

Table V. Selected Epidemiological Studies of (Occupational) Lead Exposure and Brain Cancer1
First author (Year)CountryStudy populationFollow-upOutcomeExposure measureObsRR (CI)
  • 1

    Obs, observed number of exposed cases; NR, not reported;

  • 2

    Cohort studies: measure of association = standardized mortality or incidence ratio (SMR, SIR), rate ratio (RR) or Odds ratio (OR; for nested case–control studies); Case–control studies: measure of association = Odds ratio (OR); CI = 95% confidence interval.

Cohort studies2
 Cooper (1985)55The United States2,300 lead production workers and 4,519 battery plant workers1947–1980Mortality (CNS)Industry (SMR)  
 Lead battery plants81.1 (0.5–2.1)
 Lead production facilities30.9 (0.2–2.6)
 Sweeney (1986)56The United States2,510 male tetraethyl lead workers1952–1977Mortality (brain)Industry (SMR)42.1 (0.7–4.9)
 Sankila (1990)57Finland3,749 glass factory workers1953–1986Cancer incidence (CNS)Industry (SIR)60.6 (0.2–1.3)
 Steenland (1992)51, 58The United States1,990 male hourly lead smelter workers1940–1988MortalityIndustry (SMR)NRAbout 1.2
 Cocco (1994)59Italy4,740 male lead and zinc miners1960–1988Mortality (CNS)Industry (SMR)81.2 (0.5–2.3)
 Gerhardsson (1995)60Sweden664 male lead battery workers1969–1989Mortality and cancer incidence (CNS)Industry (SIR)10.8 (0.0–4.2)
 Anttila (1996)29, 30FinlandNested case–control study of26 CNS tumors (16 gliomas)and 200 controls within a cohort of 20,741 workersbiologically monitored forblood lead 1973–19831973–1988Cancer incidenceHighest blood lead (μmol*l−1; OR)  
CNS  0.8–1.391.4 (0.5–4.1)
   1.4–4.3102.2 (0.7–6.6)
Glioma  0.8–1.386.7 (0.7–347)
   1.4–4.3711 (1.0–626)
 Cocco (1997)61Italy1,388 workers and laborers in production/maintenance departments in lead smelting plant1950–1992Mortality (brain)Industry (SMR)  
 National reference41.3 (0.3–3.2)
 Regional reference42.2 (0.6–5.6)
 Lundstrom (1997)62Sweden3,979 primary smelter workers including 1,992 lead-exposed workers1955–1987Mortality (CNS)Lead-exposed workers (SMR), highest exposed subgroup41.6 (0.4–4.2)
 Wong (2000)63The United States4,518 workers at lead battery plants and 2,300 at lead smelters1947–1995Mortality (CNS)Industry (SMR)150.7 (0.4–1.2)
 Englyst (2001)64Sweden3,979 lead smelter workers1958–1987Cancer incidence (CNS)Lead subcohort 1 (SIR)10.6 (0.0–3.6)
 Lead subcohort 2 (SIR)00.0 (0.0–6.5)
 Jemal (2002)65The United States3,592 white males and females from NHANES II 1976–19801976–1992Mortality (brain)Blood lead levels (RR)>50 percentile50.5 (0.1–5.8)
 Navas-Acien (2002)33, 34Sweden1,779,646 men and 1,066,346 women gainfully employed in 19701971–1989Morbidity(glioma and meningioma)Lead exposure (SIR)  
 Possible exposure (JEM)  
 Glioma101.1 (0.6–2.0)
 Meningioma72.4 (1.1–5.0)
 Wesseling (2002)66Finland413,877 women with blue collar occupations in 19701971–1995Cancer incidence (CNS)Lead (SIR)  
 Low exposureNR1.3 (1.0–1.6)
 Med/high exposure (JEM)NR1.3 (0.9–2.0)
Case–control studies2
 Mallin (1989)67The United States1,212 deceased cases and 3,198 deceased noncancer controls1979–1984BrainGlassworkers (white males)83.0 (NR)
 Cocco (1998)31The United States27,060 deceased cases and 108,240 deceased noncancer controls1984–1992BrainCaucasian men (JEM): high probability, high intensity lead exposure142.1 (1.1–4.0)
 Hu (1998)68China218 cases and 436 controls1989–1995GliomaLead exposure (self-report)0
 Cocco (1999)32The United States12,980 female deceased cases and 51,920 female deceased noncancer controls1984–1992CNSAny lead exposure (JEM)  
 CNS cancer3661.1 (1.0–1.2)
 Meningioma91.9 (1.0–3.9)
 Hu (1999)69China183 cases and 366 controls1989–1996MeningiomaLead exposure (self-report)  
 Men67.2 (1.0–52)
 Women105.7 (1.4–23)
 Carozza (2000)70The United States476 cases and 462 controls1991–1994GliomaFoundry/smelter workers62.6 (0.5–13)
 Painters101.6 (0.5–4.9)

The results of this study were little affected by the analysis approach, and provide additional support for an association between occupational lead exposure and brain cancer risk. Brain cancer mortality rates were greater among those potentially exposed as compared to unexposed subjects, with indications of an exposure–response trend. Findings based on external comparison analysis were somewhat less indicative of an association, which may be due to a slightly favorable brain tumor mortality experience in our NLMS cohort as compared to the general United States population (SMR = 0.92; 95% CI = 0.76–1.10 based on 119 observed and 130 expected brain cancer deaths). However, SMR-based analyses also showed a positive exposure–response trend, and RSMR estimates were similar (albeit somewhat lower) to HR or RR estimates.

Analysis of crude brain cancer mortality rates indicated elevated rates in older people, and possible differential risk by marital status, education, gender, race and urban status although many of these associations were closer to the null after applying the full Cox proportional hazards model including all covariates and lead exposure status (data not shown). These risk factors were taken into account in the analysis to reduce the effect of a mixture of socioeconomic, lifestyle, environmental and occupational factors as well as diagnostic bias on the results.31 Previous studies in the United States, Europe and China have been inconsistent regarding associations with rural dwelling31, 71 and educational level,68, 70, 72, 73, 74 but findings have been indicative of an increased brain tumor risk among married individuals.31, 73

Almost 19% of cohort members were considered potentially exposed to lead in their jobs for all probabilities and intensities combined, which is in line with prevalence estimates reported in other studies. For instance, death certificate-based investigations have reported prevalence estimates of ∼20% among men31 and 3% among women32 in the United States using the same JEM. Furthermore, a lifetime prevalence of 47% of exposure to lead compounds was found based on expert review of work history questionnaires in a population-based case–control study conducted in Montreal, Canada from 1979–1985.75 Other case–control studies carried out in the United States during the 1980s and 1990s reported prevalence estimates ranging from 4 to 36% based on self-report.76, 77, 78, 79, 80

Findings from this study (and many previous studies) must be interpreted in light of several limitations, including the consideration of all brain tumors as one entity, the absence of biological measures of exposure, and a small number of exposed subjects.

The classification of brain and central nervous system (CNS) tumors is complex, and consideration of histopathological characteristics is considered elementary.81, 82, 83 The current WHO classification assigns morphology codes to 8 histological groups,83, 84 but meningioma and gliomas comprise more than 70% of all tumors. Meningioma is a predominantly benign tumor that arises from tissues surrounding the brain and spinal cord, and accounts for over 29% of all tumors.83, 84, 85, 86 Gliomas are tumors spanning a wide range of neoplasms with distinct clinical, histopathological and genetic features that arise from glial cells with a structural or supportive function.83, 84, 85, 86 Gliomas account for 42% of all tumors and 77% of malignant tumors.84 Since different classes of brain tumors arise from distinct cell types, they may have different etiologies.18 Consequently, real effects may be masked when diseases with different etiologies are studied as one disease, and future studies of brain tumors should focus on biologically distinct tumor types.18 Our analysis addressed the association between occupational lead exposure and malignant brain tumors, since the number of deaths from malignant (ICD-9 192.1 and 192.3; n = 1) and benign (ICD-9 225.2 and 225.4; n = 5) meningiomas in the NLMS public-use data file was small. In addition to the complexity of classifying brain and CNS tumors, death certificates indicating brain cancer may reflect metastases from other sites87 for which the exposure under study would not be relevant.

Exposure assessment was based on linking the occupation and industry reported by the subjects at baseline with a JEM previously developed based on information from the published literature, computerized exposure databases, unpublished industrial hygiene reports and personal experience.31 This approach enabled us to elucidate exposure–response relationships by evaluating associations across strata of probability and intensity. However, these exposure estimates can only be considered crude surrogates for biological measures of exposure, such as bone lead levels,15 since the current job at baseline may not be representative of the subjects' work history. Nonetheless, Gomez-Marin et al. recently found that current occupation can be used as a surrogate for longest-held job for many occupational subgroups,88 including skilled jobs potentially involving elevated levels of lead exposure.

The possibility of confounding by occupational exposures other than the one under study is another concern when relying on job titles or JEMs. For example, jobs assigned high-probability and medium/high-intensity lead exposure included gas station attendants, painters, welders, plumbers and automobile mechanics. Furthermore, 8 out of 13 brain cancer deaths in the medium/high intensity category occurred among those employed as automobile mechanics (2 deaths; 28,302 person-years), heavy equipment mechanics (3 deaths; 25,331 person-years) and welders and flamecutters (3 deaths; 20,177 person-years). These job titles are individually possibly associated with brain cancer mortality with a full model HR of 2.30 (95% CI = 0.56–9.56), 3.15 (95% CI = 0.97–10.20) and 5.12 (95% CI = 1.58–16.61), respectively. On the other hand, the remaining occupations in the medium/high intensity category showed little evidence for an increased brain cancer risk (full model HR = 1.02; 95% CI = 0.32–3.26). These 3 occupations may also involve exposure to aromatic hydrocarbons, metal fumes and electromagnetic fields.31, 89 Therefore, we cannot confidently rule out potential confounding although the impact on the observed associations is likely to be small because these co-exposures have only been inconclusively linked with brain cancer risk.19, 21, 90

Finally, past epidemiological studies generally reported only on a small number of exposed brain cancer cases (e.g., n < 10), thereby resulting in statistically unreliable risk estimates (Table IV). Because of the large size of the NLMS cohort eligible for the current study (n = 317,968), we observed 29 exposed brain cancer deaths, and 13 deaths were assigned high-probability and medium/high-lead exposure. This number is larger than that in many previous studies, and yielded risk estimates that were generally quite precise. Nevertheless, we considered the number of lead-exposed brain cancer deaths insufficient to reliably assess effect modification with other potential occupational risk factors, such as exposure to extremely low-frequency electromagnetic fields.33

In conclusion, this study provides further suggestive evidence for a role of lead exposure in the development of brain cancer. Future studies evaluating this association should focus on different brain tumor subtypes and biological measures of lead exposure.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The authors thank Dr. R. Hoover for helpful suggestions. This paper uses data supplied by the National Heart, Lung, and Blood Institute, NIH, DHHS from the National Longitudinal Mortality Study. The views expressed in this paper are those of the authors and do not necessarily reflect the views of the National Heart, Lung, and Blood Institute, the Bureau of the Census or the National Center for Health Statistics.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • 1
    Gidlow DA. Lead toxicity. Occup Med (Lond) 2004; 54: 7684.
  • 2
    Inorganic and organic lead compounds. In: Proceedings of the IARC working group on the evaluation of carcinogenic risks to humans, Lyon, France, Feb 10–17, 2004. IARC Monographs on the Evaluation of Carcinogenic Risks to Human, vol. 87. Lyon, France: IARC, 2004.
  • 3
    Landrigan PJ. Strategies for epidemiologic studies of lead in bone in occupationally exposed populations. Environ Health Perspect 1991; 91: 816.
  • 4
    Okun A, Cooper G, Bailer AJ, Bena J, Stayner L. Trends in occupational lead exposure since the 1978 OSHA lead standard. Am J Ind Med 2004; 45: 55872.
  • 5
    Muntner P, Menke A, DeSalvo KB, Rabito FA, Batuman V. Continued decline in blood lead levels among adults in the United States: the National Health and Nutrition Examination Surveys. Arch Intern Med 2005; 165: 215561.
  • 6
    Maas RP, Patch SC, Pandolfo TJ, Druhan JL, Gandy NF. Lead content and exposure from children's and adult's jewelry products. Bull Environ Contam Toxicol 2005; 74: 43744.
  • 7
    Roberts JR, Hulsey TC, Curtis GB, Reigart JR. Using geographic information systems to assess risk for elevated blood lead levels in children. Public Health Rep 2003; 118: 2219.
  • 8
    Mielke HW, Reagan PL. Soil is an important pathway of human lead exposure. Environ Health Perspect 1998; 106 ( Suppl. 1): 21729.
  • 9
    Leighton J, Klitzman S, Sedlar S, Matte T, Cohen NL. The effect of lead-based paint hazard remediation on blood lead levels of lead poisoned children in New York City. Environ Res 2003; 92: 18290.
  • 10
    Lanphear BP, Burgoon DA, Rust SW, Eberly S, Galke W. Environmental exposures to lead and urban children's blood lead levels. Environ Res 1998; 76: 12030.
  • 11
    Gasana J, Chamorro A. Environmental lead contamination in Miami inner-city area. J Expo Anal Environ Epidemiol 2002; 12: 26572.
  • 12
    Meyer PA, Pivetz T, Dignam TA, Homa DM, Schoonover J, Brody D. Surveillance for elevated blood lead levels among children—United States, 1997–2001. MMWR Surveill Summ 2003; 52: 121.
  • 13
    Chadha A, McKelvey LD, Mangis JK. Targeting lead in the multimedia environment in the continental United States. J Air Waste Manag Assoc 1998; 48: 315.
  • 14
    Hu H, Payton M, Korrick S, Aro A, Sparrow D, Weiss ST, Rotnitzky A. Determinants of bone and blood lead levels among community-exposed middle-aged to elderly men. The normative aging study. Am J Epidemiol 1996; 144: 74959.
  • 15
    Hu H, Rabinowitz M, Smith D. Bone lead as a biological marker in epidemiologic studies of chronic toxicity: conceptual paradigms. Environ Health Perspect 1998; 106: 18.
  • 16
    Rabinowitz MB. Toxicokinetics of bone lead. Environ Health Perspect 1991; 91: 337.
  • 17
    Ohgaki H, Kleihues P. Epidemiology and etiology of gliomas. Acta Neuropathol (Berl) 2005; 109: 93108.
  • 18
    Batchelor TT, Dorfman MV, Hunter DJ. Epidemiology, pathology, and imaging of brain tumors. In: BlackPM, LoefflerJS, eds. Cancer of the nervous system. Philadelphia, PA: Lippincott Williams & Wilkins, 2005. 113.
  • 19
    Wrensch M, Minn Y, Chew T, Bondy M, Berger MS. Epidemiology of primary brain tumors: current concepts and review of the literature. Neuro Oncol 2002; 4: 27899.
  • 20
    El-Zein R, Minn AY, Wrensch M, Bondy M. Epidemiology of brain tumors. In: LevinVA, ed. Cancer in the nervous system, 2nd ed. New York: Oxford University Press, 2002. 25266.
  • 21
    Savitz D, Trichopoulos D. Brain cancer. In: AdamiHO, HunterD, TrichopoulosD, eds. Textbook of cancer epidemiology. New York: Oxford University Press, 2002. 486503.
  • 22
    Inskip PD, Linet MS, Heineman EF. Etiology of brain tumors in adults. Epidemiol Rev 1995; 17: 382414.
  • 23
    Gerhardsson L, Englyst V, Lundstrom NG, Nordberg G, Sandberg S, Steinvall F. Lead in tissues of deceased lead smelter workers. J Trace Elem Med Biol 1995; 9: 13643.
  • 24
    Silbergeld EK. Facilitative mechanisms of lead as a carcinogen. Mutat Res 2003; 533: 12133.
  • 25
    Silbergeld EK, Waalkes M, Rice JM. Lead as a carcinogen: experimental evidence and mechanisms of action. Am J Ind Med 2000; 38: 31623.
  • 26
    Zawirska B, Medras K. The role of the kidneys in disorders of porphyrin metabolism during carcinogenesis induced with lead acetate. Arch Immunol Ther Exp (Warsz) 1972; 20: 25772.
  • 27
    Hass GM, McDonald JH, Oyasu R, Battifora HA, Paloucek JT. Renal neoplasia induced by combinations of dietary lead subacetate and N-2-fluoenylacetamide. In: KingJSJ, ed. Renal neoplasia. Boston, MA: Little Brown Co., 1967. 377412.
  • 28
    Oyasu R, Battifora HA, Clasen RA, McDonald JH, Hass GM. Induction of cerebral gliomas in rats with dietary lead subacetate and 2-acetylaminofluorene. Cancer Res 1970; 30: 124861.
  • 29
    Anttila A, Heikkila P, Nykyri E, Kauppinen T, Pukkala E, Hernberg S, Hemminki K. Risk of nervous system cancer among workers exposed to lead. J Occup Environ Med 1996; 38: 1316.
  • 30
    Anttila A, Heikkila P, Pukkala E, Nykyri E, Kauppinen T, Hernberg S, Hemminki K. Excess lung cancer among workers exposed to lead. Scand J Work Environ Health 1995; 21: 4609.
  • 31
    Cocco P, Dosemeci M, Heineman EF. Brain cancer and occupational exposure to lead. J Occup Environ Med 1998; 40: 93742.
  • 32
    Cocco P, Heineman EF, Dosemeci M. Occupational risk factors for cancer of the central nervous system (CNS) among US women. Am J Ind Med 1999; 36: 704.
  • 33
    Navas-Acien A, Pollan M, Gustavsson P, Floderus B, Plato N, Dosemeci M. Interactive effect of chemical substances and occupational electromagnetic field exposure on the risk of gliomas and meningiomas in Swedish men. Cancer Epidemiol Biomarkers Prev 2002; 11: 167883.
  • 34
    Navas-Acien A, Pollan M, Gustavsson P, Plato N. Occupation, exposure to chemicals and risk of gliomas and meningiomas in Sweden. Am J Ind Med 2002; 42: 21427.
  • 35
    Rogot E, Sorlie PD, Johnson NJ, Glover CS, Treasure DW. A mortality study of one million persons by demographic, social and economic factors: 1979–1981 follow-up. First data book. National Institutes of Health, Public Health Service, Department of Health and Human Services, 1988.
  • 36
    Rogot E, Sorlie PD, Johnson NJ, Schmitt C. A mortality study of 1.3 million persons by demographic, social and economic factors: 1979–1985 follow-up. Second data book. National Institutes of Health, Public Health Service, Department of Health and Human Services, 1992.
  • 37
    Muntaner C, Sorlie P, O'Campo P, Johnson N, Backlund E. Occupational hierarchy, economic sector, and mortality from cardiovascular disease among men and women. Findings from the National Longitudinal Mortality Study. Ann Epidemiol 2001; 11: 194201.
  • 38
    Cocco P, Ward MH, Dosemeci M. Risk of stomach cancer associated with 12 workplace hazards: analysis of death certificates from 24 states of the United States with the aid of job exposure matrices. Occup Environ Med 1999; 56: 7817.
  • 39
    Kposowa AJ. Suicide mortality in the United States: differentials by industrial and occupational groups. Am J Ind Med 1999; 36: 64552.
  • 40
    Honda Y, Macaluso M, Brill I. A SAS program for the stratified analysis of follow-up data. J Occup Health 1998; 40: 1547.
  • 41
    Katz MH, Hauck WW. Proportional hazards (Cox) regression. J Gen Intern Med 1993; 8: 70211.
  • 42
    Breslow NE, Lubin JH, Marek P, Langholz B. Multiplicative models and cohort analysis. J Am Stat Assoc 1983; 78: 112.
  • 43
    Klein JP, Rizzo JD, Zhang MJ, Keiding N. Statistical methods for the analysis and presentation of the results of bone marrow transplants. Part 2: regression modeling. Bone Marrow Transplant 2001; 28: 100111.
  • 44
    Wood J, Richardson D, Wing S. A simple program to create exact person-time data in cohort analyses. Int J Epidemiol 1997; 26: 3959.
  • 45
    Sun J, Ono Y, Takeuchi Y. A simple method for calculating the exact confidence interval of the standardized mortality ratio with an SAS function. J Occup Health 1996; 38: 1967.
  • 46
    Morris JA, Gardner MJ. Calculating confidence intervals for relative risks (odds ratios) and standardised ratios and rates. Br Med J (Clin Res Ed) 1988; 296: 13136.
  • 47
    Frome EL, Checkoway H. Epidemiologic programs for computers and calculators. Use of Poisson regression models in estimating incidence rates and ratios. Am J Epidemiol 1985; 121: 30923.
  • 48
    Frome EL, Morris MD. Evaluating goodness of fit of Poisson regression models in cohort studies. American Statistician 1989; 43: 1447.
  • 49
    Davis FG, McCarthy BJ. Epidemiology of brain tumors. Curr Opin Neurol 2000; 13: 63540.
  • 50
    Landrigan PJ, Boffetta P, Apostoli P. The reproductive toxicity and carcinogenicity of lead: a critical review. Am J Ind Med 2000; 38: 23143.
  • 51
    Steenland K, Boffetta P. Lead and cancer in humans: where are we now? Am J Ind Med 2000; 38: 2959.
  • 52
    Hayes RB. The carcinogenicity of metals in humans. Cancer Causes Control 1997; 8: 37185.
  • 53
    Fu H, Boffetta P. Cancer and occupational exposure to inorganic lead compounds: a meta-analysis of published data. Occup Environ Med 1995; 52: 7381.
  • 54
    Vainio H. Lead and cancer—association or causation? Scand J Work Environ Health 1997; 23: 13.
  • 55
    Cooper WC, Wong O, Kheifets L. Mortality among employees of lead battery plants and lead-producing plants, 1947–1980. Scand J Work Environ Health 1985; 11: 33145.
  • 56
    Sweeney MH, Beaumont JJ, Waxweiler RJ, Halperin WE. An investigation of mortality from cancer and other causes of death among workers employed at an east Texas chemical plant. Arch Environ Health 1986; 41: 238.
  • 57
    Sankila R, Karjalainen S, Pukkala E, Oksanen H, Hakulinen T, Teppo L, Hakama M. Cancer risk among glass factory workers: an excess of lung cancer? Br J Ind Med 1990; 47: 8158.
  • 58
    Steenland K, Selevan S, Landrigan P. The mortality of lead smelter workers: an update. Am J Public Health 1992; 82: 16414.
  • 59
    Cocco PL, Carta P, Belli S, Picchiri GF, Flore MV. Mortality of Sardinian lead and zinc miners: 1960–88. Occup Environ Med 1994; 51: 67482.
  • 60
    Gerhardsson L, Hagmar L, Rylander L, Skerfving S. Mortality and cancer incidence among secondary lead smelter workers. Occup Environ Med 1995; 52: 66772.
  • 61
    Cocco P, Hua F, Boffetta P, Carta P, Flore C, Flore V, Onnis A, Picchiri GF, Colin D. Mortality of Italian lead smelter workers. Scand J Work Environ Health 1997; 23: 1523.
  • 62
    Lundstrom NG, Nordberg G, Englyst V, Gerhardsson L, Hagmar L, Jin T, Rylander L, Wall S. Cumulative lead exposure in relation to mortality and lung cancer morbidity in a cohort of primary smelter workers. Scand J Work Environ Health 1997; 23: 2430.
  • 63
    Wong O, Harris F. Cancer mortality study of employees at lead battery plants and lead smelters, 1947–1995. Am J Ind Med 2000; 38: 25570.
  • 64
    Englyst V, Lundstrom NG, Gerhardsson L, Rylander L, Nordberg G. Lung cancer risks among lead smelter workers also exposed to arsenic. Sci Total Environ 2001; 273: 7782.
  • 65
    Jemal A, Graubard BI, Devesa SS, Flegal KM. The association of blood lead level and cancer mortality among whites in the United States. Environ Health Perspect 2002; 110: 3259.
  • 66
    Wesseling C, Pukkala E, Neuvonen K, Kauppinen T, Boffetta P, Partanen T. Cancer of the brain and nervous system and occupational exposures in Finnish women. J Occup Environ Med 2002; 44: 6638.
  • 67
    Mallin K, Rubin M, Joo E. Occupational cancer mortality in Illinois white and black males, 1979–1984, for seven cancer sites. Am J Ind Med 1989; 15: 699717.
  • 68
    Hu J, Johnson KC, Mao Y, Guo L, Zhao X, Jia X, Bi D, Huang G, Liu R. Risk factors for glioma in adults: a case–control study in northeast China. Cancer Detect Prev 1998; 22: 1008.
  • 69
    Hu J, Little J, Xu T, Zhao X, Guo L, Jia X, Huang G, Bi D, Liu R. Risk factors for meningioma in adults: a case–control study in northeast China. Int J Cancer 1999; 83: 299304.
  • 70
    Carozza SE, Wrensch M, Miike R, Newman B, Olshan AF, Savitz DA, Yost M, Lee M. Occupation and adult gliomas. Am J Epidemiol 2000; 152: 83846.
  • 71
    Auvinen A, Hietanen M, Luukkonen R, Koskela RS. Brain tumors and salivary gland cancers among cellular telephone users. Epidemiology 2002; 13: 3569.
  • 72
    Schlehofer B, Hettinger I, Ryan P, Blettner M, Preston-Martin S, Little J, Arslan A, Ahlbom A, Giles GG, Howe GR, Menegoz F, Rodvall Y et al. Occupational risk factors for low grade and high grade glioma: results from an international case control study of adult brain tumours. Int J Cancer 2005; 113: 11625.
  • 73
    Inskip PD, Tarone RE, Hatch EE, Wilcosky TC, Fine HA, Black PM, Loeffler JS, Shapiro WR, Selker RG, Linet MS. Sociodemographic indicators and risk of brain tumours. Int J Epidemiol 2003; 32: 22533.
  • 74
    Zheng T, Cantor KP, Zhang Y, Keim S, Lynch CF. Occupational risk factors for brain cancer: a population-based case–control study in Iowa. J Occup Environ Med 2001; 43: 31724.
  • 75
    Parent ME, Siemiatycki J, Fritschi L. Occupational exposures and gastric cancer. Epidemiology 1998; 9: 4855.
  • 76
    Kamel F, Umbach DM, Munsat TL, Shefner JM, Hu H, Sandler DP. Lead exposure and amyotrophic lateral sclerosis. Epidemiology 2002; 13: 3119.
  • 77
    Gracia CR, Sammel MD, Coutifaris C, Guzick DS, Barnhart KT. Occupational exposures and male infertility. Am J Epidemiol 2005; 162: 72933.
  • 78
    Jackson LW, Correa-Villasenor A, Lees PS, Dominici F, Stewart PA, Breysse PN, Matanoski G. Parental lead exposure and total anomalous pulmonary venous return. Birth Defects Res A Clin Mol Teratol 2004; 70: 18593.
  • 79
    Kerr MA, Nasca PC, Mundt KA, Michalek AM, Baptiste MS, Mahoney MC. Parental occupational exposures and risk of neuroblastoma: a case–control study (United States). Cancer Causes Control 2000; 11: 63543.
  • 80
    Gorell JM, Johnson CC, Rybicki BA, Peterson EL, Kortsha GX, Brown GG, Richardson RJ. Occupational exposures to metals as risk factors for Parkinson's disease. Neurology 1997; 48: 6508.
  • 81
    De Girolami U, Smith TW. Neuropathology of central nervous system tumors. In: BlackPM, LoefflerJS, eds. Cancer of the nervous system. Philadelphia, PA: Lippincott Williams & Wilkins, 2005. p. 1545.
  • 82
    DeAngelis LM, Posner JB. Cancer of the central nervous system and pituitary gland. In: LenhardRE, OsteenRT, GanslerT, eds. Clinical oncology. Atlanta, GA: American Cancer Society, 2001. p. 653703.
  • 83
    Kleihues P, Louis DN, Scheithauer BW, Rorke LB, Reifenberger G, Burger PC, Cavenee WK. The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol 2002; 61: 21525; Discussion 26–9.
  • 84
    CBTRUS. Statistical report: primary brain tumors in the United States, 1997–2001. Hinsdale, IL: Central Brain Tumor Registry of the United States, 2004.
  • 85
    Doolittle ND. State of the science in brain tumor classification. Semin Oncol Nurs 2004; 20: 22430.
  • 86
    Hill CI, Nixon CS, Ruehmeier JL, Wolf LM. Brain tumors. Phys Ther 2002; 82: 496502.
  • 87
    McLaughlin JK, Lipworth L. Brain cancer and cosmetic breast implants: a review of the epidemiologic evidence. Ann Plast Surg 2004; 52: 1157.
  • 88
    Gomez-Marin O, Fleming LE, Caban A, Leblanc WG, Lee DJ, Pitman T. Longest held job in U.S. occupational groups: the National Health Interview Survey. J Occup Environ Med 2005; 47: 7990.
  • 89
    van Wijngaarden E, Stewart PA. Critical literature review of determinants and levels of occupational benzene exposure for United States community-based case–control studies. Appl Occup Environ Hyg 2003; 18: 67893.
  • 90
    Blair A, Stewart WF, Stewart PA, Sandler DP, Axelson O, Vineis P, Checkoway H, Savitz D, Pearce N, Rice C. A philosophy for dealing with hypothesized uncontrolled confounding in epidemiological investigations. Med Lav 1995; 86: 10610.