Abstract

Cadmium, a highly persistent heavy metal, has been categorized as a probable human carcinogen by the U.S. Environmental Protection Agency. Primary exposure sources include food and tobacco smoke. We carried out a population-based case–control study of 246 women, aged 20–69 years, with breast cancer and 254 age-matched control subjects. We measured cadmium levels in urine samples by inductively coupled plasma mass spectrometry and conducted interviews by telephone to obtain information on known breast cancer risk factors. Odds ratios (ORs) and 95% confidence intervals (CIs) for breast cancer by creatinine-adjusted cadmium levels were calculated by multivariable analysis. Statistical tests were two-sided. Women in the highest quartile of creatinine-adjusted cadmium level (≥0.58 μg/g) had twice the breast cancer risk of those in the lowest quartile (<0.26 μg/g; OR = 2.29, 95% CI = 1.3 to 4.2) after adjustment for established risk factors, and there was a statistically significant increase in risk with increasing cadmium level ( Ptrend = .01). Based on this study, the absolute risk difference is 45 (95% CI = 0 to 77) per 100 000 given an overall breast cancer rate of 124 per 100 000. Whether increased cadmium is a causal factor for breast cancer or reflects the effects of treatment or disease remains to be determined.

Cadmium is a toxic, bioaccumulating, nonessential, and highly persistent heavy metal with a variety of known adverse health effects. It has been designated as a probable human carcinogen by the U.S. Environmental Protection Agency, which means that the weight of evidence of carcinogenocity based on animal studies is sufficient and the weight of the evidence based on epidemiologic studies is limited ( 1 ) . For nonoccupationally exposed women who do not smoke, food is the largest source of cadmium intake, whereas for smokers, inhalation of tobacco smoke is the predominant source of exposure ( 2 ) . Only a small fraction of inhaled or ingested cadmium is excreted, resulting in increased body burden over time ( 3 ) . In vitro studies provide provocative evidence that cadmium is associated with human breast cancer through several mechanisms ( 410 ) . For example, cadmium may act like an estrogen, forming high-affinity complexes with estrogen receptors ( 1013 ) .

Past studies suggest that values from urine samples give an indication of lifetime cadmium exposure, whereas values from serum samples represent recent exposure ( 2 ) . From September 2004 through February 2005, we collected urine samples from 246 women with breast cancer aged 20–69 years to determine whether higher body burden levels of cadmium as measured in urine are associated with risk of breast cancer. Women who were asked to contribute a urine sample were part of an ongoing population-based case–control study of breast cancer. Only women who had been diagnosed with breast cancer within the 24 months preceding the interview and whose cancers had been reported to Wisconsin's mandatory statewide cancer reporting system, the Wisconsin Cancer Reporting System, were eligible for enrollment in the parent study. Information on stage of disease and reported treatment(s) for all participating case participants was obtained from the Wisconsin Cancer Reporting System. Community control subjects were randomly selected women from State of Wisconsin driver's license lists who did not have a personal history of breast cancer. Control subjects were matched to case participants within 5-year age strata to yield an age distribution similar to that of the case participants. We collected urine samples from 254 community control subjects from September 2004 through February 2005. The participation proportions—i.e., the percentage of participants eligible to participate in the parent study who actually completed the interview—was 75% for case patients and 71% for control subjects. Of those who completed the interview and agreed to participate in the substudy, the participation proportion for the sequential sample returning the urine specimen was 90% for both case patients and control subjects. The study was approved by the University of Wisconsin–Madison Health Sciences Human Subjects Committee. Oral informed consent was obtained for the interview, and written informed consent was obtained for the urine specimen collection.

All participants were interviewed by telephone by trained interviewers. The 35-minute interview asked about physical activity, reproductive history, alcohol consumption, height and weight, use of oral contraceptives and hormone replacement therapy, personal and family medical history, demographic factors, a limited set of dietary components, and smoking history. Women collected urine samples in their homes. Urine collection kits and detailed photoessay instructions were carefully designed to minimize trace element contamination during specimen collection and handling. Samples were processed at the Wisconsin State Laboratory of Hygiene, a laboratory that is specially designed for and dedicated to trace element analysis and is subject to high-efficiency particulate air filtration. Urine samples contacted only Teflon or polyethylene materials and sample containers were exhaustively cleaned using multistep acid leachings. Cadmium was quantified by using inductively coupled plasma mass spectrometry ( 2 ) . Trace metal analyses were performed from November 2004 to August 2005 (a 2- to 6-month lag from collection time). A comprehensive quality-control program incorporating numerous methods including bottle blanks, monitoring of multiple cadmium isotopes, internal and external controls, and routine inclusion of National Institute of Standards and Technology (NIST) standard reference materials ensured high-quality data. A method reporting level of 0.010 μg/L cadmium enabled metal levels in every urine sample to be quantified. Urine creatinine level was also measured to control for kidney function ( 14 ) .

Case participants were older at first full-term pregnancy, were more likely to have a family history of breast cancer, had fewer children, reported higher income, and were more likely to have never married than similarly aged control subjects ( Table 1 ). Creatinine-adjusted cadmium levels ranged from 0.02 to 4.55 μg/g in case participants (excluding one subject whose level was 30.95 μg/g) and from 0.08 to 2.64 μg/g in control subjects. The creatinine-adjusted cadmium values for each 10-year age group in the control subjects were similar to those reported from a national age-stratified sampling of the general population of Caucasian women in the National Health and Nutrition Examination Survey (NHANES III) (data not shown) ( 15 ) .

Table 1.

Characteristics of participants with breast cancer and control subjects

  Case patients ( N = 246)
 
  Control subjects ( N = 254)
 
   
Characteristic n n  Odds ratio *  95% CI  
Parity       
    0–1 69 28 45 18 1.00  
    2 95 39 100 39 0.61 0.4 to 1.0 
    3 or more 81 33 107 42 0.45 0.3 to 0.7 
Age at first full term pregnancy (years)       
    <20 36 15 43 17 1.00  
    20–24 80 33 105 41 0.87 0.5 to 1.5 
    25–29 55 22 54 21 1.26 0.7 to 2.3 
    30+ 36 15 21 2.17 1.1 to 4.4 
    Nulliparous 39 16 31 12 1.58 0.8 to 3.0 
Family history of breast cancer       
    Absent 187 76 218 86 1.00  
    Present 52 21 30 12 2.01 1.2 to 3.3 
    Unknown 1.35 0.5 to 4.1 
Recent alcohol consumption (drinks/week)       
    None 46 19 50 20 1.00  
    1–6 173 70 178 70 1.07 0.7 to 1.7 
    7+ 27 11 25 10 1.15 0.6 to 2.3 
Body mass index (kg/m 2 )        
    <22.5 34 23 27 18 1.00  
    22.5–25.0 30 20 30 20 0.75 0.4 to 1.6 
    25.1–28.8 35 24 41 27 0.68 0.3 to 1.4 
    28.9+ 48 33 52 34 0.74 0.4 to 1.4 
Age at menarche       
    <12.0 59 24 51 20 1.00  
    12.0–12.9 51 21 58 23 0.75 0.4 to 1.3 
    13.0–13.9 60 24 72 28 0.71 0.4 to 1.2 
    ≥14.0 67 27 65 26 0.89 0.5 to 1.5 
    Unknown 1.06 0.4 to 3.0 
Menopausal status       
    Postmenopausal 147 60 152 60 1.00  
    Premenopausal 67 27 75 30 0.97 0.5 to 1.8 
    Unknown 32 13 27 11 1.24 0.6 to 2.5 
Age at menopause (years)        
    <45 25 17 43 28 1.00  
    45–49 38 26 37 24 1.84 0.9 to 3.6 
    50–54 50 34 33 22 2.69 1.4 to 5.3 
    55+ 22 15 22 14 1.64 0.8 to 3.6 
    Unknown 12 17 11 1.18 0.5 to 2.9 
Type of postmenopausal hormone therapy        
    Never 52 35 61 40 1.00  
    Estrogen only 32 22 37 24 1.01 0.6 to 1.9 
    Estrogen and progestin only 47 32 37 24 1.49 0.8 to 2.6 
    Other combination 10 1.05 0.4 to 2.8 
    Unknown 1.13 0.4 to 3.5 
Smoking status       
    Never 135 55 155 63 1.00  
    Former 77 31 65 26 1.37 0.9 to 2.1 
    Current 33 13 31 13 1.23 0.7 to 2.1 
Education       
    No high school diploma 13 0.28 0.1 to 0.9 
    High school diploma 103 42 97 38 1.00  
    Some college 60 24 69 27 0.84 0.5 to 1.3 
    College degree 77 31 72 28 1.04 0.7 to 1.6 
    Unknown 0.64 0.1 to 3.9 
Marital status       
    Never married 17 2.46 1.0 to 6.1 
    Married/living as married 199 81 201 79 1.00  
    Separated/widowed 28 11 43 17 0.64 0.4 to 1.1 
    Unknown 0.67 0.1 to 4.1 
Annual income, US dollars       
    <15 000 3.7 3.1 1.38 0.5 to 3.8 
    15 001–30 000 28 11.4 28 11.0 1.22 0.7 to 2.3 
    30 001–50 000 66 26.8 81 31.9 1.00  
    50 001–100 000 87 35.4 95 37.4 1.13 0.7 to 1.8 
    >100 000 30 12.2 19 7.5 1.94 1.0 to 3.8 
    Unknown 26 10.6 23 9.1 1.42 0.7 to 2.7 
  Case patients ( N = 246)
 
  Control subjects ( N = 254)
 
   
Characteristic n n  Odds ratio *  95% CI  
Parity       
    0–1 69 28 45 18 1.00  
    2 95 39 100 39 0.61 0.4 to 1.0 
    3 or more 81 33 107 42 0.45 0.3 to 0.7 
Age at first full term pregnancy (years)       
    <20 36 15 43 17 1.00  
    20–24 80 33 105 41 0.87 0.5 to 1.5 
    25–29 55 22 54 21 1.26 0.7 to 2.3 
    30+ 36 15 21 2.17 1.1 to 4.4 
    Nulliparous 39 16 31 12 1.58 0.8 to 3.0 
Family history of breast cancer       
    Absent 187 76 218 86 1.00  
    Present 52 21 30 12 2.01 1.2 to 3.3 
    Unknown 1.35 0.5 to 4.1 
Recent alcohol consumption (drinks/week)       
    None 46 19 50 20 1.00  
    1–6 173 70 178 70 1.07 0.7 to 1.7 
    7+ 27 11 25 10 1.15 0.6 to 2.3 
Body mass index (kg/m 2 )        
    <22.5 34 23 27 18 1.00  
    22.5–25.0 30 20 30 20 0.75 0.4 to 1.6 
    25.1–28.8 35 24 41 27 0.68 0.3 to 1.4 
    28.9+ 48 33 52 34 0.74 0.4 to 1.4 
Age at menarche       
    <12.0 59 24 51 20 1.00  
    12.0–12.9 51 21 58 23 0.75 0.4 to 1.3 
    13.0–13.9 60 24 72 28 0.71 0.4 to 1.2 
    ≥14.0 67 27 65 26 0.89 0.5 to 1.5 
    Unknown 1.06 0.4 to 3.0 
Menopausal status       
    Postmenopausal 147 60 152 60 1.00  
    Premenopausal 67 27 75 30 0.97 0.5 to 1.8 
    Unknown 32 13 27 11 1.24 0.6 to 2.5 
Age at menopause (years)        
    <45 25 17 43 28 1.00  
    45–49 38 26 37 24 1.84 0.9 to 3.6 
    50–54 50 34 33 22 2.69 1.4 to 5.3 
    55+ 22 15 22 14 1.64 0.8 to 3.6 
    Unknown 12 17 11 1.18 0.5 to 2.9 
Type of postmenopausal hormone therapy        
    Never 52 35 61 40 1.00  
    Estrogen only 32 22 37 24 1.01 0.6 to 1.9 
    Estrogen and progestin only 47 32 37 24 1.49 0.8 to 2.6 
    Other combination 10 1.05 0.4 to 2.8 
    Unknown 1.13 0.4 to 3.5 
Smoking status       
    Never 135 55 155 63 1.00  
    Former 77 31 65 26 1.37 0.9 to 2.1 
    Current 33 13 31 13 1.23 0.7 to 2.1 
Education       
    No high school diploma 13 0.28 0.1 to 0.9 
    High school diploma 103 42 97 38 1.00  
    Some college 60 24 69 27 0.84 0.5 to 1.3 
    College degree 77 31 72 28 1.04 0.7 to 1.6 
    Unknown 0.64 0.1 to 3.9 
Marital status       
    Never married 17 2.46 1.0 to 6.1 
    Married/living as married 199 81 201 79 1.00  
    Separated/widowed 28 11 43 17 0.64 0.4 to 1.1 
    Unknown 0.67 0.1 to 4.1 
Annual income, US dollars       
    <15 000 3.7 3.1 1.38 0.5 to 3.8 
    15 001–30 000 28 11.4 28 11.0 1.22 0.7 to 2.3 
    30 001–50 000 66 26.8 81 31.9 1.00  
    50 001–100 000 87 35.4 95 37.4 1.13 0.7 to 1.8 
    >100 000 30 12.2 19 7.5 1.94 1.0 to 3.8 
    Unknown 26 10.6 23 9.1 1.42 0.7 to 2.7 
*

Adjusted for age.

CI = confidence interval.

Postmenopausal women only.

We used multivariable analysis to compute odds ratios (ORs) and 95% confidence intervals (CIs) for breast cancer by quartile of cadmium level, as defined by levels in control subjects. After adjustment for age, women in the highest quartile (Q4) had twice the risk of breast cancer as those in the lowest quartile (Q1) (OR = 2.05, 95% CI = 1.2 to 3.6; P = .02, Ptrend = .07) ( Table 2 ). After adjustment for age, parity, age at first birth, family history of breast cancer, recent alcohol consumption, body mass index, age at menarche, menopausal status, age at menopause, type of postmenopausal hormone use, education, and marital status, the odds ratio for women in Q4 as compared with Q1 rose slightly to 2.29 (95% CI = 1.3 to 4.2; P = .01, Ptrend = .02). For every 0.1 μg/g increase in cadmium, we observed an 8% increase in breast cancer risk (95% CI = 1.02 to 1.14; P = .01). Adding smoking status (three categories—never, former, and current) to the models did not change the risk estimates (data not shown).

Table 2.

Multivariable-adjusted odds ratio of breast cancer associated with creatinine-adjusted cadmium level *

  Case patients ( n = 246)
 
  Control subjects ( n = 254)
 
     
Cadmium concentration level n n  OR  95% CI  OR   95% CI  
All participants         
    Q1 43 17 63 25 1.00  1.00  
    Q2 61 25 64 25 1.46 0.9 to 2.5 1.53 0.9 to 2.7 
    Q3 60 24 64 25 1.44 0.8 to 2.5 1.42 0.8 to 2.6 
    Q4 82 33 63 25 2.05 1.2 to 3.6 2.29 1.3 to 4.2 
    Continuous (per 1.0 μg/g increase)     1.59 1.0 to 2.6 2.09 1.2 to 3.8 
     P value       0.07  0.01 
Stratified by stage §         
    Localized         
        Q1 27 17 63 25 1.00  1.00  
        Q2 41 26 64 25 1.46 0.8 to 2.7 1.53 0.8 to 3.0 
        Q3 37 24 64 25 1.27 0.7 to 2.4 1.25 0.6 to 2.5 
        Q4 50 32 63 25 1.75 0.9 to 3.3 1.99 1.0 to 4.0 
        Continuous (per 1.0 μg/g)     1.48 0.9 to 2.5 2.04 1.1 to 3.9 
         P value       0.14  0.03 
    Regional/distant         
        Q1 14 18 63 25 1.00  1.00  
        Q2 16 20 64 25 1.40 0.6 to 3.2 1.57 0.6 to 3.9 
        Q3 21 26 64 25 1.94 0.9 to 4.4 1.95 0.8 to 4.9 
        Q4 29 36 63 25 3.14 1.4 to 7.1 3.92 1.6 to 9.9 
        Continuous (per 1.0 μg/g)     1.95 1.0 to 4.0 3.24 1.2 to 8.6 
         P value       0.07  0.02 
Stratified by treatment          
    Surgery and/or radiation only         
        Q1 25 18 63 25 1.00  1.00  
        Q2 33 24 64 25 1.36 0.7 to 2.6 1.65 0.8 to 3.5 
        Q3 28 20 64 25 1.14 0.6 to 2.2 1.08 0.5 to 2.3 
        Q4 51 37 63 25 2.19 1.1 to 4.2 3.18 1.5 to 6.8 
        Continuous (per 1.0 μg/g)     1.67 1.0 to 2.9 3.07 1.4 to 7.0 
         P value       0.07  0.01 
    Chemotherapy and/or hormonal therapy         
        Q1 10 17 63 25 1.00  1.00  
        Q2 15 25 64 25 1.58 0.6 to 3.9 1.54 0.6 to 4.3 
        Q3 17 28 64 25 1.82 0.7 to 4.5 2.11 0.7 to 6.0 
        Q4 18 30 63 25 2.03 0.8 to 5.1 2.08 0.7 to 6.2 
        Continuous (per 1.0 μg/g)     1.28 0.6 to 2.8 1.24 0.4 to 3.5 
         P value       0.54  0.68 
  Case patients ( n = 246)
 
  Control subjects ( n = 254)
 
     
Cadmium concentration level n n  OR  95% CI  OR   95% CI  
All participants         
    Q1 43 17 63 25 1.00  1.00  
    Q2 61 25 64 25 1.46 0.9 to 2.5 1.53 0.9 to 2.7 
    Q3 60 24 64 25 1.44 0.8 to 2.5 1.42 0.8 to 2.6 
    Q4 82 33 63 25 2.05 1.2 to 3.6 2.29 1.3 to 4.2 
    Continuous (per 1.0 μg/g increase)     1.59 1.0 to 2.6 2.09 1.2 to 3.8 
     P value       0.07  0.01 
Stratified by stage §         
    Localized         
        Q1 27 17 63 25 1.00  1.00  
        Q2 41 26 64 25 1.46 0.8 to 2.7 1.53 0.8 to 3.0 
        Q3 37 24 64 25 1.27 0.7 to 2.4 1.25 0.6 to 2.5 
        Q4 50 32 63 25 1.75 0.9 to 3.3 1.99 1.0 to 4.0 
        Continuous (per 1.0 μg/g)     1.48 0.9 to 2.5 2.04 1.1 to 3.9 
         P value       0.14  0.03 
    Regional/distant         
        Q1 14 18 63 25 1.00  1.00  
        Q2 16 20 64 25 1.40 0.6 to 3.2 1.57 0.6 to 3.9 
        Q3 21 26 64 25 1.94 0.9 to 4.4 1.95 0.8 to 4.9 
        Q4 29 36 63 25 3.14 1.4 to 7.1 3.92 1.6 to 9.9 
        Continuous (per 1.0 μg/g)     1.95 1.0 to 4.0 3.24 1.2 to 8.6 
         P value       0.07  0.02 
Stratified by treatment          
    Surgery and/or radiation only         
        Q1 25 18 63 25 1.00  1.00  
        Q2 33 24 64 25 1.36 0.7 to 2.6 1.65 0.8 to 3.5 
        Q3 28 20 64 25 1.14 0.6 to 2.2 1.08 0.5 to 2.3 
        Q4 51 37 63 25 2.19 1.1 to 4.2 3.18 1.5 to 6.8 
        Continuous (per 1.0 μg/g)     1.67 1.0 to 2.9 3.07 1.4 to 7.0 
         P value       0.07  0.01 
    Chemotherapy and/or hormonal therapy         
        Q1 10 17 63 25 1.00  1.00  
        Q2 15 25 64 25 1.58 0.6 to 3.9 1.54 0.6 to 4.3 
        Q3 17 28 64 25 1.82 0.7 to 4.5 2.11 0.7 to 6.0 
        Q4 18 30 63 25 2.03 0.8 to 5.1 2.08 0.7 to 6.2 
        Continuous (per 1.0 μg/g)     1.28 0.6 to 2.8 1.24 0.4 to 3.5 
         P value       0.54  0.68 
*

Quartile (Q) of cadmium concentration (in μg/g): Q1 = <0.263; Q2 = 0.263–0.395; Q3 = 0.396–0.579; Q4 = ≥0.580. OR = odds ratio; CI = confidence interval.

Adjusted for age.

Adjusted for age, parity, age at first birth, family history of breast cancer, recent alcohol consumption, body mass index, age at menarche, menopausal status, age at menopause, type of postmenopausal hormone use, education, and marital status.

§

For 11 case participants, stage was reported as unknown.

For 16 case participants, treatment was not reported to the cancer registry.

We carried out several exploratory stratified analyses. In one, we stratified by smoking status (never versus former or current). All of the odds ratios were higher than 1.0 for every quartile in relation to the lowest quartile, as we had observed in the full models. However, the estimates were unstable and had very large confidence intervals due to the very small numbers of smokers (data not shown). We also stratified by median age (56 years) of study participants. In age-adjusted analyses, younger women had an increased risk of breast cancer with increased cadmium level (OR for Q4 versus Q1 = 2.34, 95% CI = 1.1 to 5.0; P = .58, Ptrend = .06). However, among older women the relationship between cadmium level and breast cancer risk was not statistically significant (OR for Q4 versus Q1 = 1.36, 95% CI = 0.5 to 3.4; P = .08, Ptrend = .12). Similar patterns were observed when stratifying by menopausal status (data not shown). We also evaluated breast cancer risk associated with intake of a limited number of foods that are potentially high in cadmium (canned fish, liver, kidney, and crustaceans); however, we did not observe any associations (data not shown). The breast cancer participant with very high creatinine-adjusted cadmium levels (30.95 μg/g) had only one established risk factor (low parity) and no evidence of occupational exposure to cadmium. However, she was a heavy smoker who also reported heavy consumption of crustaceans.

Finally, we considered whether the disease process per se or treatment may have altered cadmium levels by stratifying by disease stage (localized versus regional/distant) and by treatment (surgery and/or radiation only versus chemotherapy and/or hormonal therapy). Similar results in the multivariable analysis were observed among participants with localized breast cancer (OR for Q4 versus Q1 = 1.99; 95% CI =1.0 to 4.0; P = .03, Ptrend = .10) and among participants with regional/distant breast cancer (OR = 3.92; 95% CI = 1.6 to 9.9; P = .02, Ptrend = .004) ( Table 2 ). We also analyzed differences in odds ratios by stage using the Wald test in polytomous logistic regression models and found no statistically significant differences in odds ratios (data not shown). In analyses by treatment, breast cancer risk increased with increasing cadmium level among those treated by surgery and/or radiation only (OR for Q4 versus Q1 = 3.18; 95% CI = 1.5 to 6.8; P = .01, Ptrend = .01). However, among those treated with chemotherapy the association of risk with cadmium level was not statistically significant (OR for Q4 versus Q1 = 2.08; 95% CI = 0.7 to 6.2; P = .68, Ptrend = .16) ( Table 2 ). There were no differences in odds ratios when participants were stratified by treatment (surgery/radiation only versus chemotherapy) using the Wald test in polytomous logistic regression models (data not shown).

Our finding of an association of breast cancer risk with cadmium level is provocative, although whether the association reflects an effect of cadmium on the initiation or promotion of tumor growth or possible effects of treatment or the disease itself on cadmium levels (reverse causation) is unclear. Studies of smoking—a major cadmium source—and breast cancer have yielded mixed results ( 16 , 17 ) . If cadmium were a cause of breast cancer, then a positive association of smoking with breast cancer would be expected. More research is needed to examine the relation of cadmium exposure to breast cancer risk, including additional research on potential interactions between treatment or disease status and cadmium release from tissue stores as well as the influence of smoking (with its concomitant high cadmium exposure potential) on breast cancer risk, particularly with regard to polymorphisms in detoxification enzymes.

The results of this population-based case–control study indicate a statistically significant twofold increased breast cancer risk for women in the highest quartile of cadmium level compared with those in the lowest quartile. If cadmium exposure is causal—which one epidemiology study cannot determine—the population attributable risk, using our point estimates, could account for 45 of the 124 annual breast cancer cases per 100 000. However, due to the small sample size, the absolute risk calculation has considerable uncertainty. If we used 95% confidence limits for each quartile, the population attributable risk for cadmium exposure would range from 0 to 77 of the 124 annual breast cancer cases per 100 000. Laboratory data also support a relation between high cadmium levels and increased breast cancer risk ( 48 , 1013 ) . Given the ubiquitous exposure of the general population to cadmium, the mode of the association between cadmium exposure and breast cancer risk warrants further study.

This study was supported by National Institutes of Health grants R03 CA110796 and R01 CA47147. The study sponsor had no role in the study design, data collection, analysis, or interpretation of the data.
We thank Dr. Henry Anderson and Laura Stephenson for support and assistance with the cancer data. The authors are appreciative of the staff of the Wisconsin Women's Health Study and the State Laboratory of Hygiene for data collection and laboratory support on this project. We are especially grateful to the study participants, whose generosity made this research possible.

References

(1)
Environmental Protection Agency. Cadmium. Washington (DC): Environmental Protection Agency;
1987
.
(2)
Centers for Disease Control and Prevention. Third national report on human exposure to environmental chemicals. Atlanta (GA): Centers for Disease Control and Prevention;
2005
.
(3)
Klaassen CD. Pharmacokinetics in metal toxicity.
Fundam Appl Toxicol
 
1981
;
1
:
353
–7.
(4)
Abshire MK, Buzard GS, Shiraishi N, Waalkes MP. Induction of c-myc and c-jun proto-oncogene expression in rat L6 myoblasts by cadmium is inhibited by zinc preinduction of the metallothionein gene.
J Toxicol Environ Health
 
1996
;
48
:
359
–77.
(5)
Jin P, Ringertz NR. Cadmium induces transcription of proto-oncogenes c-jun and c-myc in rat L6 myoblasts.
J Biol Chem
 
1990
;
265
:
14061
–4.
(6)
Shimada H, Shiao YH, Shibata M, Waalkes MP. Cadmium suppresses apoptosis induced by chromium.
J Toxicol Environ Health A
 
1998
;
54
:
159
–68.
(7)
Fan LZ, Cherian MG. Potential role of p53 on metallothionein induction in human epithelial breast cancer cells.
Br J Cancer
 
2002
;
87
:
1019
–26.
(8)
Meplan C, Mann K, Hainaut P. Cadmium induces conformational modifications of wild-type p53 and suppresses p53 response to DNA damage in cultured cells.
J Biol Chem
 
1999
;
274
:
31663
–70.
(9)
Nagata C, Nagao Y, Shibuya C, Kashiki Y, Shimizu H. Urinary cadmium and serum levels of estrogens and androgens in postmenopausal Japanese women.
Cancer Epidemiol Biomarkers Prev
 
2005
;
14
:
705
–8.
(10)
Johnson MD, Kenney N, Stoica A, Hilakivi-Clarke L, Singh B, Chepko G, et al. Cadmium mimics the in vivo effects of estrogen in the uterus and mammary gland.
Nat Med
 
2003
;
9
:
1081
–4.
(11)
Garcia-Morales P, Saceda M, Kenney N, Kim N, Salomon DS, Gottardis MM, et al. Effect of cadmium on estrogen receptor levels and estrogen-induced responses in human breast cancer cells.
J Biol Chem
 
1994
;
269
:
16896
–901.
(12)
Choe SY, Kim SJ, Kim HG, Lee JH, Choi Y, Lee H, et al. Evaluation of estrogenicity of major heavy metals.
Sci Total Environ
 
2003
;
312
:
15
–21.
(13)
Stoica A, Katzenellenbogen BS, Martin MB. Activation of estrogen receptor-α by the heavy metal cadmium.
Mol Endocrinol
 
2000
;
14
:
545
–53.
(14)
Elinder C-G. Normal values for cadmium in human tissues, blood, and urine in different countries. In Friberg L, Elinder C-G, Kjellstrom T, Nordberg GF, editors. Cadmium and health: a toxicological and epidemiological appraisal Vol 1: Exposure, dose, and metabolism. Boca Raton (FL): CRC Press;
1985
. p. 81–102.
(15)
Paschal DC, Burt V, Caudill SP, Gunter EW, Pirkle JL, Sampson EJ, et al. Exposure of the U.S. population aged 6 years and older to cadmium: 1988–1994.
Arch Environ Contam Toxicol
 
2000
;
38
:
377
–83.
(16)
Hamajima N, Hirose K, Tajima K, Rohan T, Calle EE, Heath CW Jr, et al. Alcohol, tobacco and breast cancer—collaborative reanalysis of individual data from 53 epidemiological studies, including 58 515 women with breast cancer and 95 067 women without the disease.
Br J Cancer
 
2002
;
87
:
1234
–45.
(17)
Li CI, Malone KE, Daling JR. The relationship between various measures of cigarette smoking and risk of breast cancer among older women 65–79 years of age (United States).
Cancer Causes Control
 
2005
;
16
:
975
–85.