Lung Cancer

Lung Cancer Mortality among Uranium Gaseous Diffusion Plant Workers: A Cohort Study 1952–2004

 

LW Figgs

 

Environmental Health Division, Douglas County Health Department, Omaha, NE, USA

 

Correspondence to
Larry W. Figgs, PhD, MPH, REHS/RS, Environmental Health Division, Douglas County Health Department, Omaha, NE, USA

E-mail: lfiggs2@gmail.com

Received: Apr 22, 2013

Accepted: Jun 12, 2013

 

Abstract

Background: 9%–15% of all lung cancers are attributable to occupational exposures. Reports are disparate regarding elevated lung cancer mortality risk among workers employed at uranium gaseous diffusion plants.

Objective: To investigate whether external radiation exposure is associated with lung cancer mortality risk among uranium gaseous diffusion workers.

Methods: A cohort of 6820 nuclear industry workers employed from 1952 to 2003 at the Paducah uranium gaseous diffusion plant (PGDP) was assembled. A job-specific exposure matrix (JEM) was used to determine likely toxic metal exposure categories. In addition, radiation film badge dosimeters were used to monitor cumulative external ionizing radiation exposure. International Classification for Disease (ICD) codes 9 and 10 were used to identify 147 lung cancer deaths. Logistic and proportional hazards regression were used to estimate lung cancer mortality risk.

Results: Lung cancer mortality risk was elevated among workers who experienced external radiation >3.5 mrem and employment duration >12 years.

Conclusion: Employees of uranium gaseous diffusion plants carry a higher risk of lung cancer mortality; the mortality is associated with increased radiation exposure and duration of employment.

Keywords: Lung neoplasms; Mortality; Radiation; Occupational exposure; Occupational diseases; Uranium componds

 

Introduction

Worldwide, cancer has become the second leading cause of death.1 The US National Cancer Institute estimated that approximately 12 million Americans with a cancer history were alive in 2008, 1.6 million diagnosed cancers and nearly 577 190 cancer deaths in the US in 2012.2 Environmental risk factors are believed responsible for two out of every three cancers and occupational exposures may account for 40 000 incident cases and 20 000 deaths each year.3 In the 20th century's last decade lung cancer became the most common cancer associated with occupational hazards.4 Metal exposures are a common workplace concern,5-8 especially among nuclear industry workers.

The difficulty with identifying associations between work-related exposures and cancer mortality among nuclear industry workers is based on the type of cancer, the exposure assessment, and the toxic properties of confounding hazards, especially radiation. For example, exposures may be unique to a specific workplace in character, duration, and/or intensity. In other instances, the mechanisms believed to be responsible for hazard-related neoplasia are incomplete and suggest multiple pathways.5,9-12

Assessing a hazard's toxic properties, the duration of worker exposure, the intensity of that exposure, and other confounding factors may explain how past occupational cohort investigations discordantly report associations between some toxic metal exposures and lung cancer.8,11,13-17 For example, in 2004 Sorahan concluded that nickel exposure was not associated with increased lung cancer mortality.14 Concurrently, Sorahan and Esman reported that cadmium exposures do not support a carcinogenesis hypothesis15 despite earlier reports.18 A year later, Sorahan and Williams reported that nickel exposure could not be ruled out as a risk factor for increased lung cancer mortality.16 In 2009, Levy, et al, in an effort to assess lung cancer standardized mortality ratios (SMRs) among beryllium workers, observed little to support an association between lung cancer and beryllium exposure, after adjusting for tobacco smoking.11 More recently, Brusk-Hohlfeld cited 87 references in a literature review and concluded that there was sufficient epidemiological evidence supporting a link between carcinogenesis and exposure to arsenic, beryllium, cadmium, chromium and nickel.7 In a separate review citing 160 references, Wild, et al, noted that some metals (arsenic, beryllium, cadmium, chromium and nickel) were “accepted” carcinogens based primarily on International Agency for Research on Cancer (IARC) assessments published prior to 2000.8

Recent nuclear industry cohort studies have featured cancer outcomes associated with radiation or metal exposure.19,20 Godbold and Tompkins reported that the expected number of deaths derived from the US population of white males exceeded the number of all cancers observed as well as lung cancers among 814 nickel-exposed “barrier workers” employed at the Oak Ridge gaseous diffusion plant in Oak Ridge, Tennessee.21 Polednak reported no excess mortality for all-cancers (SMR: 0.88; 95% CI: 0.60–1.23), but reported mortality excesses for respiratory system cancers (SMR: 1.39; 95% CI: 0.81–2.22) and lung cancer (SMR: 1.50; 95% CI: 0.87–2.40) among 1059 white male welders exposed to uranium, fluoride, lead, nickel, mercury, chromium, and technetium at three Oak Ridge plants from 1943 to 1977.22 Frome, et al, observed no excess in all-cancer, lung, or respiratory system mortality among 106 020 nuclear industry workers in Oak Ridge, employed between 1943 and 1985.23 In another investigation, National Institute for Occupational Safety and Health (NIOSH) investigators did not observe excess all-cancer mortality (SMR: 0.82; 95% CI: 0.73–0.92), but observed statistically non-significant mortality excesses for stomach (SMR: 1.18; 95% CI: 0.65–1.94), female genital organs (SMR: 1.27; 95% CI: 0.47–2.77), bone (SMR: 1.68; 95% CI: 0.20–6.05), lympho-reticulosarcoma (SMR: 1.37; 95% CI: 0.55–2.82), and Hodgkin's disease (SMR: 1.38; 95% CI: 0.45–3.23).24 Similarly, Charles, et al, observed an increase in cancer mortality associated with occupational exposure to metals and solvents.25

Recently, Chan, et al, reported that excess mortality occurred among Paducah uranium gaseous diffusion plant (PGDP) workers who developed cancers of pancreas, myeloproliferative neoplasms and lymphomas.20 However, Chan, et al, did not observe excess mortality among lung cancer victims even though toxic metal and radiation exposures were likely higher among PGDP workers than the general population and IARC's earlier report that several of the metals were potential carcinogens.26 Aware of the potential biases and constraints associated with SMR analyses, this investigation re-examines the relationship between lung cancer mortality and uranium exposure with specific emphasis on uranium exposure and radiation applying logistic regression analysis to case-control designs and proportional hazards regression to the entire cohort.

 

Materials and Methods

Population sampling frame

The PGDP is located on 3425-acres near Paducah, Kentucky. It was built in the early 1950s to process uranium. Although owned by the US Department of Energy (DOE), since construction the facility was leased to Union Carbide (1950–1984), Martin Marietta (1984–1995), and Lockheed Martin Utilities Services (1995–2005).27 The PGDP cohort is described elsewhere in detail.28,29

Workers had to work at least six continuous months from September 1, 1952 to December 31, 2003 to be eligible for the study. Person-time accrued from the worker's initial PGDP hiring date until their death or December 31, 2004. Briefly, 6859 worker files were assembled from DOE contractors, unions, and Oak Ridge affiliated universities.30 Nosologists used state vital records agency death certificates to verify the vital status of workers dying before 1980. National Death Index (NDI) queries were used to verify post-1979 deaths. The vital status of two workers was undetermined. Thirty-nine worker files were duplicates.31 The final analysis file contained 6820 workers. Cancer morality was followed until December 31, 2004.

 

Exposure assessment

Metals

A detail description of the job-specific metal exposure matrix is reported elsewhere.32,33 Briefly, all job titles were grouped, ranked for specific metal exposures, and consolidated using worker interviews, plant production records, and job-site maps. Metal exposure rankings were based on qualitative and quantitative factors such as environmental monitoring data, location of plant processes, and interviews with long-term workers. Company representatives and long-term workers reviewed job titles and were asked to comment on whether each job title would have less, the same, or more exposure than another job title. Rankings (categories) ranged from zero to five—zero representing “no exposure expected” and five “the most exposure expected.” Rankings were categorical and unrelated to quantitative exposure intensity (concentration) or dose. Therefore, exposure rankings for a unique metal were not additive or multiplicative (i.e., a category ‘2’ exposure ranking was not twice a category ‘1’ exposure ranking). Inter-rank comparisons were invalid.

Categories ‘0’ and ‘1’ were combined for this analysis.

Arsenic, hexavalent chromium, nickel, beryllium, and uranium exposure categories were tabulated to construct a study-specific, job exposure matrix (JEM) by modifying methods described elsewhere.32 Discrete exposure ranking categories ranging from zero to five were entered into each unique metal (row)/job-title (column) cell. More than one ranking was allowed per cell in the JEM to account for changes in plant processes over time. A supplemental table provided additional ranking information.32

Categories ‘0’ and ‘1’ were combined for the analysis below.

Radiaton

External radiation exposure intensity was detremined by monitoring personal radiometric badges. Data were recorded as decimal, interval data in millirems. Millirem exposures were natural logarithm (ln) transformed to mitigate skewness and kurtosis. Tertile categories were devised in which the lowest exposure (“Low”) represented all millirem values ≤427.63 mrem. Intermediate exposures (“Intermediate”) represented all millirem values >427.63 and ≤1069.25 mrem. The highest tertile (“High”) represented all millirem values >1069.25 mrem.

Duration of exposure to radiation

Employment duration was a proxy estimate for the duration of radiation exposure. Employment duration was determined as the difference in total days between the dates last observed and initially hired. To calculate years of employment, days were divided by 365.25. Total years were stratified by tertiles. Workers employed 3.51 years or less were considered “Short” duration employees. Workers employed >3.51 years and ≤11.8 years were “Intermediate” duration workers. Those who worked >11.8 years were “Long” duration workers.

 

Case ascertainment

All death certificates with “Underlying Cause of Death” (UCD) fields containing International Classification for Disease (ICD) codes 161, 162.0-162.5, 162.8, 163 (ICD-6 and ICD-7), or codes 161, 162.0-162.5, 162.8, 163.1, 163.9 (ICD-8), or codes 162.0-162.9, 163, 164, 165, (ICD-9), or codes C33.0-C34.0-C34.3, C34.8, C34.9, C37, C38.0-C38.3, C38.8, C39, C45 (ICD-10) and dying before December 31, 2003 were considered lung cancer cases.34

 

Control selection

Case-control design

Cumulative incidence and incidence density sampling was used to select controls. Five-hundred and eighteen controls were selected by incidence density sampling (~4:1 controls per case without age-group frequency matching). A parallel case-control design is also applied because assuming that the hazard ratio (HR) is constant over time may be invalid for specific proportional hazards models.35,36

 

Statistical analysis

All statistics were estimated using
STATA™ ver 10.1 Statistics/Data Analysis Special Edition (Stata Corp, 4905 Lakeway Drive, College Station, TX 77845, USA).

Initially, all lung cancer mortality risk estimates associated with radiation exposure were derived from Cox proportional hazards regression models.35 However, test that the relative hazard (hazard in the exposed divided by hazard in the unexposed) was fixed over time often indicated that a single HR was marginally appropriate or inappropriate for some models. Since fixed proportional hazards (the assumption that the ratio of hazards between exposed and unexposed is the same at all possible survival times) is not an assumption of case-control designs, a parallel nested case-control design was pursued.

Odds ratios (ORs) were estimated by logistic regression analysis37 and adjusted using the available confounding variables. Bias introduced by ignoring smoking (subject-level tobacco smoking histories unavailable) was addressed by probabilistic sensitivity analyses, assuming that lung cancer mortality relative risk attributable to tobacco smoking was 1.5 and tobacco smoke exposure prevalence among workers ranged from 0.10 to 0.25.38

χ2 statistics with degrees of freedom and p value,39 and crude and adjusted ORs40 with 95% CI are provided where appropriate.

Cohen's k was used to assess inter-method agreement between the JEM and urine uranium concentration and external radiation.41,42 The JEM was converted to a dichotomous (exposed vs. unexposed) matrix by collapsing all values ‘1’ and ‘2’ into one “unexposed” category. JEM categories 3–5 were classified as “exposed.” Urine uranium concentration was divided into two exposure categories in which unexposed workers had values less than or equal to the median. Similarly, natural log-transformed millirem values were divided into two exposure categories in which unexposed workers had values less than or equal to the median.

Tobacco smoking effects were assessed using probabilistic sensitivity analysis.43

 

Institutional review board (IRB) approval

Only de-identified data were available for this analysis. The PDGP, University of Louisville, University of Kentucky, and University of Cincinnati IRB annually approved all data collection methods and verified investigators training to conduct ethical scientific investigations. Consent to collect worker information was obtained from employers and employee union representatives.

 

Results

There were 1674 total deaths. Four-hundred and thirty-five were classified as cancers. One-hundred forty-seven were lung cancer deaths. The first cancer death occurred April 11, 1953; the last occurred December 23, 2003. Eighteen percent (n=1223) of PGDP workers initially hired as chemical operators accounted for 29% (n=42) of all lung cancer deaths (p<0.07). Thirty-two percent (n=2190) of all PGDP workers were initially hired in maintenance categories and accounted for one-third (n=49) of all lung cancer deaths. Sixteen and a half percent (n=1129) of PGDP workers initially hired as office workers accounted for 10% (n=15) of all lung cancer deaths. Four and a half percent (n=300) of workers initially hired in a security title accounted for seven and a half percent (n=11) of all lung cancer deaths. The remainder of the cancer deaths occurred among other job title groups, however none with frequencies higher than those described above.

Table 1 compares the means or percent distribution of demographic and exposure characteristics between workers dying from lung cancer and those who did not die from lung cancer. Workers dying from lung cancer were typically older, white men compared to workers who did not die from lung cancer. Workers who died from lung cancer likely experienced higher arsenic, beryllium, chromium, nickel, uranium, and trichloroethylene (TCE) exposures compared to those who did not, and proportional differences in these exposures were significantly different for all but arsenic. Radiation exposure (film badge readings) among lung cancer cases was also significantly higher compared to other workers.

Table 1: A comparison of the mean and proportional differences of important traits and exposures between Paducah Gaseous Diffusion Plant workers dying from lung cancer and workers who did not.

Population trait

Lung cancer deaths

n = 147

All other workers

n = 6673

Person-Years (mean)

12.3

11.0

Age (mean, 95% CI)

68.4 (66.9–70.0)

59.9 (59.6–60.3)

Female gender (%)

7.5

18.4a

Race (%)

White

Non-white

Unknown

 

85

5

11

 

71b

11

17

Likely higher metal exposure based on the Job Exposure Matrix:

Arsenic exposure (%)

Beryllium exposure (%)

Chromium exposure (%)

Nickel exposure (%)

Uranium exposure (%)

TCE exposure (%)

 

 

 

24

45

50

48

53

47

 

 

 

27

36a

41a

38b

41c

38a

dRadiation exposure (ln(mrem))

Low (%)

Medium (%)

High (%)

 

54

27

22

51

 

46e

31f

31

37

 

 

Table 2 summarizes univariate (crude) and multivariate (adjusted) logistic regression analysis derived ORs for lung cancer mortality risk. From left to right are the JEM exposure categories, the number of lung cancer cases, the number of controls, a crude OR estimate, and finally the adjusted OR. ORs show a 20% increase (OR: 1.20; 95% CI: 0.54–2.63) in lung cancer deaths among PGDP workers exposed to “any metal” compared to unexposed workers after adjusting for race, age, and gender. Lung cancer mortality risk was consistently elevated for nickel, uranium, and TCE job exposure matrix categories.

Table 2: Lung cancer odds ratio (OR) estimates within Job Exposure Matrix categories

Job Exposure Matrix

Lung Ca cases

(n = 147)*

Non-cases

(n = 6673)*

Crude OR
(95% CI)

Adjusted OR
(95% CI)

Any metal

No

Yes

 

9

138

 

551

6122

 

Reference

1.38 (0.70–3.10)

 

Reference

1.20 (0.54–2.63)

Arsenic

1

2

3

4

5

 

48

21

43

6

26

 

2068

1767

1044

213

1444

 

Reference

0.51 (0.29–0.88)

1.77 (1.14–2.75)

1.21 (0.42–2.88)

0.78 (0.46–1.28)

 

Reference

0.45 (0.25–0.84)

0.91 (0.61–1.37)

0.89 (0.64–1.24)

0.81 (0.68–0.97)

Beryllium

1

2

3

4

5

 

51

20

10

5

58

 

2947

1039

288

121

2141

 

Reference

1.11 (0.62–1.91)

2.00 (0.90–4.05)

2.39 (0.73–6.09)

1.57 (1.05–2.34)

 

Reference

0.92 (0.51–1.66)

1.00 (0.64–1.57)

1.17 (0.81–1.68)

0.95 (0.83–1.09)

Chromium

1

2

3

4

5

 

64

10

22

48

 

3360

591

1183

1402

 

Reference

0.89 (0.40–1.75)

0.98 (0.57–1.62)

1.80 (1.20–2.67)

 

Reference

0.92 (0.44–1.91)

0.97 (0.80–1.67)

1.12 (0.89–1.41)

Nickel

1

2

3

4

5

 

41

31

5

9

58

 

3134

897

106

251

2148

 

Reference

2.64 (1.59–4.34)

3.61 (1.08–9.37)

2.74 (1.15–5.81)

2.06 (1.35–3.17)

 

Reference

2.16 (1.23–3.79)

1.60 (0.92–2.77)

1.12 (0.81–1.56)

1.03 (0.89–1.20)

Uranium

1

2

3

4

5

 

41

24

4

17

58

 

2804

954

146

484

2148

 

Reference

1.72 (0.99–2.93)

1.87 (0.48–5.28)

2.40 (1.26–4.36)

1.85 (1.21–2.84)

 

Reference

1.56 (0.87–2.82)

1.22 (0.66–2.25)

1.15 (0.91–1.46)

1.00 (0.87–1.16)

TCE

1

2

3

4

5

 

24

51

3

16

48

 

1652

1829

640

786

1387

 

Reference

1.92 (1.15–3.28)

0.32 (0.06–1.07)

1.40 (0.69–2.77)

2.38 (1.42–4.09)

 

Reference

1.52 (0.84–2.75)

0.62 (0.33–1.16)

1.05 (0.82–1.34)

1.17 (1.00–1.38)

 

 

Uniquely, uranium exposure was assessed using the JEM and results of urine analysis for uranium. There was substantial agreement (0.61<k<0.80; 83% agreement; k=0.68) between a modified dichotomous JEM (see Methods) and urine uranium, but only fair agreement (0.21<k<0.40; 67% agreement; k=0.34) between the dichotomous JEM and radiation badge data.41,42 Consequently, similar comparisons were not possible for the other metals and TCE.

Table 3 summarizes the relative lung cancer mortality OR estimates for employment duration tertiles (far left column). Columns two and three contain the number of cases and controls, respectively in each stratum. The last two columns represent the crude and adjusted OR estimates compared using the “Short” duration stratum as the reference worker population. The risk of lung cancer mortality among workers employed >3.51 years is not different from the reference population.

Table 3: Lung cancer odds ratio (OR) estimates by duration of employment

Employment duration*

Lung Ca cases

(n = 147)

Non-cases

(n = 6673)

Crude OR
(95% CI)

Adjusted OR

(95% CI)

Short

47

2204

Reference

Reference

Intermediate

43

2207

0.91 (0.60–1.41)

1.04 (0.67–1.67)

Long

57

2262

1.24 (0.83–1.89)

1.00 (0.77–1.19)

 

 

Table 4 summarizes the relative lung cancer mortality OR between radiation exposure—ln(mrem)—tertiles (far left column), with the lowest tertile as the reference group. Relative lung cancer mortality risk increased as the exposure level increased, compared to the reference group. However, workers receiving the highest radiation exposure were at a lower risk of dying than workers receiving the intermediate radiation exposure. In summary, lung cancer mortality risk is modestly higher among workers who were exposed to more external radiation for longer periods of time, when compared to workers who were exposed to less external radiation for shorter intervals.

Table 4: Lung cancer odds ratio (OR) estimates by external radiation exposure levels

External radiation

exposure dose* ln(mrem)

Lung Ca cases

(n = 147)

Non-cases

(n = 6673)

Crude OR
(95% CI)

Adjusted OR

(95% CI)

Low

40

2111

Reference

Reference

Intermediate

32

2120

0.80 (0.48–1.31)

1.30 (1.00–1.80)

High

75

2442

1.62 (1.08–2.45)

1.12 (0.86–1.53)

 

 

Table 5 summarizes comparisons of lung cancer mortality OR estimates stratified by radiation exposure and employment duration. In the far left column are radiation and employment duration strata. The next two columns enumerate the cases and controls within each stratum. The numbers in columns two and three were used to calculate risk estimates in columns four and five. Column four is the crude lung cancer mortality risk estimate based on a 4:1 incidence density sample of the cohort. At the very bottom of column four is the tobacco smoke adjusted estimate of the crude risk. Column five contains logistic regression analysis derived cancer mortality risk estimates for the same incidence density derived sample, adjusted for race, gender, age group, arsenic, beryllium, chromium, nickel, and TCE. The risk is elevated nearly two fold among workers who worked >3.51 years and experienced higher external radiation exposures compared to workers who worked ≤3.51 years and experienced the lowest external radiation exposure, after adjusting for race, gender, age, arsenic, beryllium, chromium, nickel, and TCE. Tobacco smoking changed the estimated precision, but not the magnitude of the mortality risk.

Table 5: A comparison of odds ratio (OR) (95% CI) estimates by radiation exposure and employment duration using a 4:1 incidence density sample

Uranium exposure by radiation and
employment duration strataa

Cases

Controls

Crude OR

(n = 406)

Adjusted ORb

(n = 406)

Low and Short

24

120

Reference

Reference

Intermediate and Intermediate

14

89

1.3 (0.6–2.8)

1.6(0.7–3.2)

High and Long

41

118

1.7 (1.0–3.2)

1.9(1.1–3.4)

Cigarette smoking adjustedc

1.7 (0.8–3.4)

 

 

Table 6 is a comparison of HRs associated with metal exposure stratified by radiation exposure and employment duration. Adjusted HRs are initially displayed for uranium and nickel exposure in the first three strata. Lung cancer mortality HRs are elevated for uranium exposed workers in the lowest radiation and short employment duration stratum (HR: 1.8; 95% CI: 0.32–9.64) as well as the high radiation and long employment duration stratum (HR: 8.4; 95% CI: 1.78–39.42). As for nickel, lung cancer mortality HRs are elevated in all three exclusive strata; the lowest radiation and short employment duration stratum (HRnickel: 1.7; 95% CI: 0.44–6.72), the intermediate stratum (HR: 1.4; 95% CI: 0.25–7.71), and the high radiation and long employment duration stratum (HRnickel: 1.2; 95% CI: 0.46–2.92). Lung cancer mortality risk associated with arsenic, beryllium, chromium and TCE were not elevated and are not shown to save space in the final Table. When “Low and Short” categories were combined with “Intermediate” radiation and employment duration categories, the risk of death associated with lung cancer is not elevated for uranium (HR: 0.5; 95% CI: 0.14–2.02), but it remains elevated for nickel (HR: 1.9; 95% CI: 0.65–5.77). Lung cancer mortality risk associated with arsenic, beryllium, chromium and TCE was not elevated. However, when “Low and Short” categories were combined with “High and Long” radiation and employment duration categories, lung cancer mortality risk is elevated for uranium (HR: 4.2; 95% CI: 1.49–11.8) and nickel (HR: 1.9; 95% CI: 0.87–3.89). Lung cancer mortality risk associated with arsenic, beryllium, chromium and TCE was not elevated.

Table 6: A comparison of proportional hazards regression lung cancer mortality risk estimates associated with metal exposure stratified by radiation exposure and employment duration.

Uranium exposure by radiation and employment duration strataa

aAdjusted Hazard Ratios

Low and Short only

Uranium

Nickel

(n = 1064)

1.8 (0.32–9.64)

1.7 (0.44–6.72)

Intermediate and Intermediate only

Uranium

Nickel

(n = 839)

0.3 (0.02–3.11)

1.4 (0.25–7.71)

High and Long only

Uranium

Nickel

(n = 1590)

8.4 (1.78–39.42)

1.2 (0.46–2.92)

Low and Short + Intermediate and Intermediate

Uranium

Nickel

Arsenic

Beryllium

Chromium

TCE

(n = 1903)

0.5 (0.14–2.02)

1.9 (0.65–5.77)

0.5 (0.24–1.09)

1.0 (0.38–2.83)

1.0 (0.60–1.57)

1.0 (0.66–1.52)

Low and Short + High and Long

Uranium

Nickel

Arsenic

Beryllium

Chromium

TCE

(n = 2654)

4.2 (1.49–11.8)

1.9 (0.87–3.98)

0.6 (0.45–0.90)

0.5 (0.27–0.98)

1.0 (0.70–1.50)

0.9 (0.64–1.20)

 

 

Discussion

These results suggests that lung cancer mortality risk among gaseous diffusion plant workers is likely elevated (Table 6) and is what one would expect based on beliefs about the health effects of radiation and uranium exposure. The lack of an elevated risk in the intermediate exposure group suggests that there may have been two groups of workers with different (unique?) sensitivity to radiation. For example, there may have been a group of workers who were most sensitive to developing radiation-induced lung cancers early on (in the first 3.5 years) (Table 6). Intermediate level exposed workers may represent a group more resistant to the effects of radiation and may not develop disease until approximately eight years later or longer (Table 6). This also may suggest that there are two etiologically specific lung tissue variants among humans—one highly susceptible to oncological radiation and another more resistant. Still, this is the first report that uranium exposed PDGP workers experienced increased lung cancer mortality risk compared to unexposed workers.29 The disparity between this and earlier efforts may be explained, in part, by inherent methodological differences between standardized mortality, case-control, and proportional hazards analyses. Key is this study's reliance on intra-cohort coworker comparisons. Coworker comparisons inherently mitigate the impact of potentially confounding or effect-modifying covariates (socioeconomic status, access to care, education, smoking, etc). Furthermore, coworker comparisons mitigate a potential healthy-worker bias (selection bias associated with the ability to perform certain work-related task) typically associated with occupational cohorts comprised of workers who were followed and worked for a long time.44

This is not the first observations that nickel is associated with lung cancer mortality in similarly occupied workers, but is the first report of this association among these workers. Notable was the observation that the risk was nearly the same across each stratum in Table 6, suggesting a risk-independent of these strata. The association (Tables 2 and 6) is consistent with reports cited above (see Introduction) and provided some assurance that these methods were at least sensitive enough to detect a previously observed association.

Still, specific weaknesses suggest some degree of caution. Typically generic, dichotomous JEMs ecologically and nondifferentially assign exposures. However, historically JEM validity has not been encouraging.45,46 The JEM applied in this investigation is based on five exposure “likelihoods.” When these categories are collapsed (see Methods), the likelihood of exposure misclassification may increase. Fortunately, bias associated with the JEM's assessment of uranium exposure was mitigated by use of available radiation and uranium exposure monitoring data, both collected at the individual level by film badge and urinalysis, respectively. A simple, linear, bivariate regression model of film badge radiation and uranium urinalysis produced a coefficient of determination of 0.49 (data not shown), suggesting that urine uranium concentration predicts about half of the variation in film badge radiation readings. This variability may be explained by workers not wearing their film badges at all times and PDGP policies that base urine uranium levels on mean values for a one-year interval. Although uranium urinalysis data strongly agreed with a dichotomous JEM assessment of uranium exposure, this analysis focuses on radiation (film badge) exposure because of the important role that radiation is believed to play in carcinogenesis and the awareness that there are potentially other sources of radiation exposure in the workplace.

A critical hurdle was the lack of individual-level tobacco smoke exposure information. To counter this, probabilistic sensitivity analyses were used to estimate the potential cigarette smoking impact on mortality estimates, assuming that cigarette smoking prevalence ranged from 10% to 25%. This suggests that cigarette smoking would have more of an impact on the estimated precision of the risk estimate than its magnitude (Table 5).

Overall, this study suggests that lung cancer is likely elevated among PGDP workers exposed to more than 1000 mrem for >3.5 years. Therefore, further attempts to reduce the duration of external radiation exposure and its intensity in the workplace may lower lung cancer mortality among similarly employed workers.

 

Acknowledgements

I thank Gail Brion, Ashley Bush and Emmanuel Otsin for their assistance with the early stages of this investigation. I also thank Dr. Adi Pour, Director of the Douglas County Health Department in Omaha, Nebraska, for supporting efforts to complete the analysis and write this manuscript.

 

Funding

This work was supported by the National Institute for Occupational Safety and Health (R01-OH-007650).

 

Conflicts of Interest: The author has no financial and personal relationships with other people or organizations that could inappropriately influence (bias) this work such as employment, consultancies, stock ownership, honoraria, paid expert testimony, patent applications/registrations, and grants or other funding.

 

References

  1. 1. Ferlay J, Parkin DM, Steliarova-Foucher E. Estimates of Cancer Incidence and Mortality in Europe in 2008. Euro J Cancer 2010;46:765-81.
  2. 2. Howlander N, Noone AM, Drapcho M, et al. SEER Cancer Statistics Review, 1975-2008. Bethesda, MD: National Cancer Institute, 2011.
  3. 3. Centers for Disease Control and Prevention. Occupational Cancer. Workplace Safety & Health Topics. Atlanta, GA: Centers for Disease Control and Prevention, 2011.
  4. 4. Doll R, Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst 1981;66:1191-308.
  5. 5. Arita A, Costa M. Epigenetics in metal carcinogenesis: nickel, arsenic, chromium and cadmium. Metallomics : Integrated Biometal Science 2009;1:222-8.
  6. 6. Beveridge R, Pintos J, Parent M, et al. Lung cancer risk associated with ocupational exposure to nickel, chromium IV, cadmium in two population-based case-control studies in Montreal. Am J Ind Med 2010;53:476-85.
  7. 7. Bruske-Hohlfeld I. Environmental and occupational risk factors for lung cancer. Meth Mol Biol 2009;472:3-23.
  8. 8. Wild P, Bourgkard E, Paris C. Lung cancer and exposure to metals: the epidemiological evidence. Meth Mol Biol 2009;472:139-67.
  9. 9. Craft ES, Abu-Qare AW, Flaherty MM, et al. Depleted and Natural Uranium: Chemistry and Toxicological Effects. J Toxicol Environ Health, Part B 2004;7:297-317.
  10. 10. Jomova K, Jenisova Z, Feszterova M, et al. Arsenic: toxicity, oxidative stress and human disease. J Appl Toxicol 2011;31:95-107.
  11. 11. Levy PS, Roth HD, Deubner DC, et al. Exposure to beryllium and occurrence of lung cancer: findings from a cox proportional hazards analysis of data from a retrospective cohort mortality study. J Occup Environ Med 2009;51:480-6.
  12. 12. Zhao J, Shi X, Castranova V, Ding M. Occupational toxicology of nickel and nickel compounds. J Environ Pathol Toxicol Oncol 2009;28:177-208.
  13. 13. Abrahim KS, Abdel-Gawad NB, Mahmoud AM, et al. Genotoxic effect of occupational exposure to cadmium. Toxicol Ind Health 2010;27:173-9.
  14. 14. Sorahan T. Morality of Workers at a Plant Manufcturing Nickel Alloys, 1958-2000. Occupat Med 2004;54:28-34.
  15. 15. Sorahan T, Williams SP. Mortality of workers at a nickel carbonyl refinery, 1958-2000. Occup Environ Med 2005;62:80-5.
  16. 16. Sorahan T, Esmen NA. Lung cancer mortality in UK nickel-cadmium battery workers, 1947-2000. Occup Environ Med 2004;61:108-16.
  17. 17. Krishnadasan A, Kennedy N, Zhao Y, et al. Nested case-control study of occupational chemical exposures and prostate cancer in aerospace and radiation workers. Am J Ind Med 2007;50:383-90.
  18. 18. Waalkes MP. Cadmium carcinogenesis in review. J Inorganic Biochem 2000;79:241-4.
  19. 19. Yiin JH, Anderson JL, Daniels RD, et al. A nested case-control study of multiple myeloma risk and uranium exposure among workers at the Oak Ridge gaseous diffusion plant. Radiat Res 2009;171:637-45.
  20. 20. Chan C, Hughes TS, Muldoon S, et al. Mortality Patterns Among Paducah Gaseous Diffusion Plant Workers. J Occup Environ Med 2010;52:725-32.
  21. 21. Godbold JH, Tompkins EA. A Long-Term Mortality Study of Workers Occupationally Exposed to Metallic Nickel at the Oak Ridge Gaseous Diffusion Plant. J Occup Med 1979;21:799-806.
  22. 22. Polednak AP. Mortality among Welders, Including a Group Exposed to Nickel Oxides. Arch Environ Health 1981;36:235-42.
  23. 23. Frome EL, Cragle DL, Watkins JP, et al. A mortality study of employees of the nuclear industry in Oak Ridge, Tennessee. Radiat Res 1997;148:64-80.
  24. 24. National Institute for Occupational Safety and Health. Final Report: Mortality Patterns among Uranium Enrichment Workers at the Portsmouth Gaseous Diffusion Plant Piketon, Ohio: U.S. Department of Health and Human Serivces, 2001.
  25. 25. Charles LE, Burchfiel CM, Fekedulegn D, et al. Occupational exposure to pesticides, metals, and solvents: the impact on mortality rates in the Honolulu Heart Program. Work 2010;37:205-15.
  26. 26. IARC. Some Inorganic and Organometallic Compounds. Lyon, IARC, 1973.
  27. 27. Alphonse Rainer et al, Petitioners v. Union Carbide Corporation, et al. Affirm judgement of the district court ed: United States Court of Appeals for the Sixth Circuit, 2005.
  28. 28. Bahr DE, Aldrich TE, Seidu D, et al. Occupational exposure to trichloroethylene and cancer risk for workers at the Paducah Gaseous Diffusion Plant. Int J Occup Med Environ Health 2011;24:67-77.
  29. 29. Chan C, Hughes TS, Muldoon S, et al. Mortality Patterns Among Paducah Gaseous Diffusion Plant Workers. J Occup Environ Med 2010;52:725-32.
  30. 30. Bahr DE, Hughes T, Aldrich TE, et al. Cohort Follow-up: the 21st Century Procedures. JRM 2009;36:16-20.
  31. 31. Bahr DE. Paducah Gaseous Diffusion Plant Cohort Follow-up: University of Kentucky, College of Public Health, Department of Epidemiology, Lexington, KY, 2009.
  32. 32. Hahn KM. Estimating Historic Exposure to Arsenic, Beryllium, Hexavalent Chromium, Nickel, and Uranium at a Uranium Enrichment, Gaseous Diffusion Plant. University of Cincinnati, 2005.
  33. 33. Figgs LW, Holsinger H, Freitas SJ, et al. Increased Suicide Risk Among Workers Following Toxic Metal Exposure at the Paducah Gaseous Diffusion Plant from 1952 to 2003. Int J Occup Environ Med 2011;2:199-214.
  34. 34. Centers for Disease Control and Prevention. National Center for Health Statistics. Atlanta, GA: National Center for Health Statistics, 2010.
  35. 35. Cox DR. Regression models and life tables (with discussion). J R Stat Soc B 1972;34:187-220.
  36. 36. Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994;81:515-26.
  37. 37. Breslow NE, Day NE. Statistical Methods in Cancer Research: Volume II- The Design and Analysis of Cohort Studies. International Agency for Research on Cancer (IARC Scientific Publication No. 82), 1987.
  38. 38. Office of the Surgeon General. Chapter 2: The Fundamentals of Mental Health and Mental Illness. In: Service PH, ed. Mental Health: A report of the Surgeon General: National Institute of Mental Health, 1999.
  39. 39. Plackett RL. Karl Pearson and the Chi-squared Test. International Statistical Review 1983;51:59-72.
  40. 40. Breslow NE, Day NE. Statistical Methods in Cancer Research: Volume 1 - The Analysis of Case-Control Studies. Lyon, International Agency for Research on Cancer (IARC Scientific Publication No. 32), 1980.
  41. 41. Cohen J. A coefficient of agreement for nomimal scales. Educ Psychol Meas 1960;20:37-46.
  42. 42. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol Bull 1968;70:213-20.
  43. 43. Orsini N, Bellocco R, Bottai M, et al. A tool for deterministic and probabilistic sensitivity analysis of epidemiologic studies. Stata J 2008;8:29-48.
  44. 44. Checkoway H, Pearce NE, Crawford-Brown DJ. Research Methods in Occupational Epidemiology. New York, NY, Oxford University Press, 1989.
  45. 45. Teschke K. Exposure surrogates: job-exposure matrices, self-reports, and expert evaluations. In: Nieuwenhuijsen MJ, ed. Exposure assessment in occupational and environmental epidemiology. New York, Oxford University Press, 2003.
  46. 46. Siemiatycki J. Exposure assessment in community-based studies of occupational cancer. Occup Hyg 1996;3:41-58.

 

TAKE-HOME MESSAGE

  • Workers of uranium gaseous diffusion plants are at higher risk of lung cancer mortality.
  • Limiting the work to 42 months (3.5 years) at high radiation levels (1096 mrem) may reduce lung cancer mortality.
  • Nuclear industry employers should biomonitor workers for metals (nickel, beryllium, arsenic, etc) other than radioactive metals.
  • Comprehensive metal biomonitoring methods need to be developed, improved, and adopted for workers similarly exposed.

 

Cite this article as: Figgs LW. Lung cancer mortality among uranium gaseous diffusion plant workers: a cohort study 1952–2004. Int J Occup Environ Med 2013;4:128-140.




 pISSN: 2008-6520
 eISSN: 2008-6814

Creative Commons LicenseThis work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License