Abstract

In this case-control study, the authors aimed to examine whether use of an electric bedding device increased breast cancer risk in African-American women. Cases were 304 African-American patients diagnosed with breast cancer during 1995–1998 who were aged 20–64 years and lived in one of three Tennessee counties. Controls were 305 African-American women without breast cancer who were selected through random digit dialing and frequency-matched to cases by age and county. Information on the use of an electric blanket or heated water bed and other risk factors was collected through telephone interviews. Breast cancer risk associated with use of an electric bedding device increased with the number of years of use, the number of seasons of use, and the length of time of use during sleep. When women who used an electric bedding device for more than 6 months per year (and therefore were more likely to have used a heated water bed, which generates lower magnetic fields) were excluded, the corresponding dose-response relations were more striking. Similar trends in dose response were shown in both premenopausal and postmenopausal women and for both estrogen receptor-positive and estrogen receptor-negative tumors. The use of electric bedding devices may increase breast cancer risk in African-American women aged 20–64 years. Such an association might not vary substantially by menopausal status or estrogen receptor status.

Received for publication November 8, 2002; accepted for publication April 9, 2003.

In 1987, Stevens (1) proposed a biologic hypothesis in which extremely low frequency electromagnetic fields (EMFs) may cause breast cancer. According to this hypothesis, EMFs can suppress production of melatonin, a hormone secreted by pineal gland cells. Melatonin can suppress production of estrogen, directly inhibit breast cancer cell growth, and boost immune function. As a result of suppression of melatonin production, estrogen levels may rise, breast cancer cells may grow faster, and the immune function that controls cell transformation may be depressed, increasing the risk of breast cancer.

Apart from Stevens’ hypothesis, the effect of EMFs on calcium homeostasis may be another potential mechanism for an association between EMFs and breast cancer. Calcium ions are messengers or effectors of many biologic processes (2). EMFs may influence the function of transmembrane proteins or the affinity of calcium-binding proteins and therefore modulate calcium fluxes (entry and extrusion) (35). The disruption of calcium homeostasis may activate oncogenes, increase oxidative stress, or facilitate tumor promotional activities via protein kinases, relating to cancer (6).

Use of electric bedding devices, especially electric blankets, may be the greatest contributor to electromagnetic exposure from residential appliances, because of high EMF intensity, prolonged exposure, and intimate contact (7). Although an EMF-breast cancer association is biologically possible, most studies on the use of electric bedding devices have not supported it, while a number of occupational studies have shown an association (810). To date, there have been six case-control studies conducted on the use of electric blankets in relation to breast cancer (1116). In most of those studies, the odds ratio estimates lay between 0.9 and 1.0, and no dose-response relation was found in terms of duration of use or mode of use (11, 1416). No association was found for both premenopausal and postmenopausal women (14, 15), although, in the studies by Vena et al. (12, 13), continuous use of an electric blanket throughout the night generated odds ratio estimates of 1.4 (95 percent confidence interval (CI): 0.9, 2.2) and 1.5 (95 percent CI: 1.0, 2.2) for pre- and postmenopausal tumors, respectively.

Previous epidemiologic studies on the use of electric bedding devices have primarily involved Caucasian women; to our knowledge, no studies have been carried out in African-American women. Two studies on occupational EMF exposures involving African-American women found a stronger association between exposure and breast cancer in African-American women than in Caucasian women (8, 17). These findings suggest that African-American women may be more susceptible to the effects of extremely low frequency EMFs and imply that it is important to conduct further research in this population, especially in residential settings. In this case-control study, we examined the relation between use of electric bedding devices and breast cancer in African-American women aged 20–64 years and assessed whether the relation varied by menopausal status or by the estrogen receptor (ER) status of the tumor.

MATERIALS AND METHODS

Study subjects

Cases were 304 African-American female patients with a first histologic diagnosis of breast cancer in 1995–1998 who lived in one of three Tennessee counties (Davidson, Shelby, and Hamilton counties), had telephone service at the time of the study, and were identified through the Tennessee Cancer Reporting System. The Tennessee Cancer Reporting System is part of the North American Association of Central Cancer Registries. The percentage of complete ascertainment for the registry was approximately 80 percent during the study period. A total of 670 eligible women with a physician’s name listed were identified from the Tennessee Cancer Reporting System. We contacted each eligible patient’s physician for permission to contact her. Thirty of the 670 patients were reported by their physicians to have died. Of the remaining 640 patients, the physician’s consent for contact was obtained for 478. A relatively large proportion of the lack of physicians’ consent (25.3 percent of the 640 patients) was due primarily to two factors: 1) a few physicians who had a relatively large number of African-American patients did not want to be involved in the study or did not respond (46.9 percent of patients without a physician’s consent) and 2) the woman’s health status was not suitable for an interview.

The patients for whom we had the physician’s consent were contacted via mail. The mailing introduced the study procedures and explained the woman’s rights as a study participant. A consent form for participation in the study was also included in the initial packet. A second packet was mailed and subsequent reminder calls were made to women who did not respond. For women who did not respond to the second mailing and did not have a telephone number listed, a home visit was made to seek the woman’s consent. Of the 478 patients with a physician’s consent, 18 were deceased and 51 could not be located. Of the remaining eligible patients whom we were able to contact, 304 (64 percent of those with a doctor’s consent, or 74 percent of those we contacted) agreed to participate in the study and were subsequently interviewed.

Controls were 305 African-American women without a history of breast cancer. The controls were frequency-matched to cases by 5-year age range and county and were selected through one-step random digit dialing (18) conducted soon after the corresponding cases were interviewed. For the selection procedures, we first grouped cases diagnosed in the same calendar year with telephone area codes from the same county and then formed the sampling frame according to the age distribution of the cases in the area, using an eligibility table. We then randomly selected one of the telephone prefixes of the cases and appended four randomly selected digits to form a telephone number. A call was made with this number to find a woman who was eligible according to ethnic background and age range. Up to nine calls were made over a 2-week period, including three daytime calls, three evening calls, and three weekend calls, for each telephone number with no answer. If an eligible woman was identified, we described the purposes and procedures of the study and then asked the woman whether she would consent to being interviewed over the telephone.

A total of 17,365 random telephone numbers were called. Based on the outcome of the last call, 4,545 of these numbers did not exist (including disconnected or changed numbers), 2,332 were business numbers, 824 were fax numbers, and 3,874 gave uncertain results, including response by an answering machine, a busy telephone line, and lack of an available respondent. Thus, the random digit dialing calls resulted in 5,790 households’ providing information on eligibility. Of these, 420 households were identified as having at least one eligible female. In households with more than one eligible female, one eligible woman was randomly chosen. For 33 eligible households, the eligible woman was never home. For the other 11 households, the eligible woman agreed to participate but preferred to be interviewed at another time, and the interview was not conducted subsequently for various reasons. Of the remaining 376 eligible women, 71 refused to participate in the study. As a result, 305 women (72.6 percent) were interviewed and were used as controls.

To compensate subjects for the time and effort involved in study participation, we paid $25 for a completed interview and provided study subjects an opportunity to enter a drawing for$200. We also paid case women $10 for agreeing to release their tumor tissue specimens. These procedures were approved by the Institutional Review Board of Meharry Medical College (Nashville, Tennessee). All participants gave informed consent prior to their inclusion in the study. Data collection We conducted telephone interviews to collect data. The interviews were conducted by trained interviewers. Information was obtained about the participant’s history of exposure to different factors on or before the reference date. The reference date was defined as the date of diagnosis for cases and the year corresponding to the diagnostic year of the matched case for controls. Information collected included data on personal habits and lifestyle (use of electric bedding devices, smoking, alcohol consumption, exercise, diet, and use of contraceptives), menstrual history (age at menarche, menopausal status, age at menopause, time from menarche to regular menses, average cycle length, and average length of period), reproductive history (age at first birth, number of pregnancies, history of infertility, and miscarriage), medical history (history of benign breast disease, history of other cancers, and use of exogenous estrogen/progesterone), family history of breast cancer, anthropometric variables (weight and height), and demographic variables. History of use of electric bedding devices was the exposure variable for this study. The study subjects were asked, “Did you ever use an electric blanket, electric mattress pad, or heated water bed on a regular basis before your breast cancer was diagnosed [or before the reference date for controls]?” Women with a history of use were further asked about the number of years of use, the number of months of use per year, and whether the device was left turned on most of the time during sleeping or was used only to warm the bed. Some other factors important for the analyses were defined as follows. First, we calculated body mass index based on weight and height at the reference date. Body mass index was calculated as weight in kilograms divided by height in meters squared (19). Second, we defined a family history of breast cancer as a history of breast cancer in either a primary relative (mother, sister, daughter) or a secondary relative (grandmother, aunt). Third, we determined menopausal status according to whether a woman was still having menstrual periods 3 months before the reference date. A woman was considered premenopausal if she had menstrual periods during the 3-month time frame or did not have periods during that time frame because of pregnancy. Postmenopausal status was defined as having had no periods during that time (except for pregnancy). Tissue collection and laboratory measurement We collected paraffin-embedded tumor tissue samples from the hospitals in which the cases had been histologically diagnosed (n = 286). Tissue samples that did not contain tumors were excluded (n = 9). ER status was measured using the immunohistochemical method (20, 21). A pathologist read all tissue slides and determined the proportion of ER-positive cells and the intensity of stained cells. The proportion was scored as follows: 1 = ≤1 percent, 2 = >1–10 percent, 3 = >10–33.3 percent, 4 = >33.3–66.6 percent, and 5 = >66.6–100 percent. Intensity was scored as follows: 0 = negative, 1 = weak, 2 = intermediate, and 3 = strong. Based on the two scores, a histologic score (the sum of the two scores) was calculated for determination of ER status. A histologic score of 3 or more was deemed ER-positive. Otherwise, the score was considered ER-negative. ER status was determined for 272 (89.5 percent) of the 304 cases. Data analysis We used unconditional logistic regression (22) to examine the relation between use of electric bedding devices and breast cancer. Odds ratios and 95 percent confidence intervals were computed. We first analyzed whether use of an electric bedding device increased the risk of breast cancer. Electric bedding devices included electric blankets, electric mattress pads, and heated water beds, as specified in the interview question. The relation was then evaluated according to the number of years for which the woman had used an electric bedding device, the average number of months of use per year, and the mode of use (device kept on most of the time or used only to warm the bed). Subsequently, we stratified study subjects according to menopausal status or ER status to examine whether the relations between use of electric bedding devices and the disease differed by menopausal status or ER status. Heated water beds generate lower magnetic fields and therefore may have a weaker association with breast cancer (15). To assess whether an association was stronger for the presumed use of electric blankets, we repeated the analysis after excluding women who reported using an electric bedding device for more than 6 months per year. This exclusion was based on the assumption that people who used a heated water bed would tend to use it for a longer period of time during the year than people who used an electric blanket. It is known that heated water beds are often used throughout the year to maintain the water at a constant temperature, while electric blankets are usually used only during certain seasons (15, 23). To control for potential confounding, we included demographic variables in the models to adjust for their complex effects. We also included as potential confounders those variables that either were associated with the risk of breast cancer in the descriptive analysis or could logically confound the relation between use of electric bedding devices and breast cancer. These variables included age, marital status, educational level, religion, household income, smoking, alcohol consumption, physical activity, body mass index, number of pregnancies, age at first birth, history of benign breast disease, use of estrogen pills, family history of breast cancer, age at first intercourse, age at menarche, average menstrual cycle length, and menopausal status. RESULTS Table 1 shows the demographic features of cases and controls. Distributions of data on age, marital status, and employment were similar between the two groups. However, cases tended to have a higher educational level and a higher family income than controls. Table 2 shows the relations between use of an electric bedding device and breast cancer. The estimated risk of breast cancer for women who used an electric bedding device as compared with those who did not was 1.4 (95 percent CI: 0.9, 2.2). The odds ratio estimates increased with increasing number of years of use; the odds ratio estimate was 4.9 (95 percent CI: 1.5, 15.6) for use for more than 10 years (p for trend = 0.011). The odds ratio estimates appeared to be higher for use in more than one season per year as compared with use in one season or less (p for trend = 0.115). The estimates for mode of use were 1.0 (95 percent CI: 0.5, 2.2) for warming the bed only and 1.7 (95 percent CI: 1.0, 3.0) for keeping the device on most of the time (p for trend = 0.074). When women who had used an electric bedding device for more than 6 months per year were excluded, the corresponding odds ratio estimates were higher and the dose-response relations were more evident for number of years used, number of seasons of use per year, and mode of use (table 3). The risk of breast cancer for women who used a device for more than 10 years relative to those who did not use one increased to 6.3 (95 percent CI: 1.5, 26.5). Table 4 shows results by menopausal status. Breast cancer risk seemed to increase with increasing levels of exposure for both premenopausal women and postmenopausal women, although p values for trend did not reach statistical significance. When data were analyzed according to the ER status of the tumor (data not shown), similar trends of higher risk with increasing exposure were found for both ER-positive and ER-negative tumors. However, the dose-response relations were significant only for ER-negative tumors. DISCUSSION In this study, use of an electric bedding device (an electric blanket, an electric mattress pad, or a heated water bed) increased the risk of breast cancer in African-American women in a dose-response fashion according to all aspects of use analyzed. When analyses were confined to women who used an electric bedding device for 6 months per year or less (and were thus more likely to use an electric blanket), the risk associated with use of an electric bedding device was higher. This coincided with possibly stronger effects due to the higher magnetic fields generated by electric blankets. A stronger association for women who used an electric bedding device for more than 10 years might result from longer cumulative exposure or from a greater probability of using an old-model electric blanket that generated stronger EMFs (24). Amid inconsistent results from previous studies, which were conducted primarily among Caucasians, the striking dose-response relation found in our study suggests that African-American women may be particularly vulnerable to the potential effects of extremely low frequency EMFs. There may be a biologic rationale for a more discernible association in African-American women, but such a conclusion is premature, because 1) the biologic interaction between EMFs and the human body is not well understood, especially at low EMF levels, and 2) there is a lack of studies on differences in sensitivity to EMF effects between different racial groups. One postulation for a possibly stronger response to EMFs among African Americans involves calcium flux and homeostasis. As we noted above, EMF exposure may change calcium fluxes and disrupt calcium homeostasis, thereby relating to cancer. African Americans may have an accelerated cellular calcium flux or turnover (enhanced calcium entry into and accelerated calcium extrusion from the cytosol) (25). As Fekete et al. (26) found, calcium fluxes were greater in lymphocytes taken from African Americans than in those taken from Caucasians. The increased cellular calcium turnover rate may result in amplified sensitivity to EMFs. Therefore, the effects of EMFs may be stronger among African Americans. This postulation may be especially plausible for breast cancer, because the breast is a calcium-handling organ (27). Another postulation is based on the hypothesis that EMFs increase cancer risk through tumor-promoting and/or copromoting effects (28). This means that EMFs might contribute to the occurrence of breast cancer by increasing the probability that previously initiated cells will proliferate and progress to neoplasticity (3, 29). Therefore, EMFs may act when tumor-initiating risk factors have been present. There are racial differences in genetic traits and nongenetic factors that may play an etiologic role. Genetically, African Americans and Caucasians might differ with regard to the mutation prevalence of some genes that may be related to cancer, such as p53 (30, 31), H-ras (32), and poly(ADP-ribose) polymerase (33). African Americans and Caucasians might also be exposed to different nongenetic factors that may influence the risk of breast cancer. For instance, African Americans are more likely to have adverse exposures (such as environmental hazards, unhealthy diets, and obesity) than Caucasians (3436). The tumor-promoting effects of EMF exposure may differ because of different distributions of genetic and nongenetic factors between different racial groups. Thus, it is possible that an association between EMFs and breast cancer appears stronger in African Americans. Our results stratified by ER status did not provide support for Stevens’ hypothesis (1). According to Stevens’ hypothesis, suppressed melatonin production due to EMF exposure can lead to increased estrogen levels and thus increased risk of breast cancer. Since estrogen executes its effects through the estrogen receptors of cells, it may be reasonable to hypothesize that EMF exposure is more likely to be related to ER-positive tumors. The observations that melatonin inhibits ER-positive breast cancer cells but not ER-negative breast cancer cells (37) and that magnetic fields enhance proliferation of ER-positive breast cells by blocking melatonin’s natural oncostatic action (38) suggest this possibility. Three (3941) of four known epidemiologic studies on residential, occupational, or electric appliance exposures (15, 3941) showed a stronger association of EMFs with ER-positive breast cancer, bolstering the laboratory results. However, our study did not show such a trend and suggests other potential mechanisms for the hypothesized association in the study population. Our study had some limitations. Nonparticipation of potential study subjects might have been a problem. A substantial proportion of eligible cases did not participate because the physician’s consent was not given or consent was not obtained from the woman. If the nonparticipating patients differed systematically from the participating patients in terms of use of electric bedding devices, the results might be biased. We were unable to assess the possible effects of selection bias through examination of differences in the use of electric bedding devices between participants and nonparticipants. Nevertheless, we were able to compare the participating and nonparticipating patients with regard to age and tumor characteristics using data obtained from the Tennessee Cancer Reporting System. Participants and nonparticipants did not differ significantly in age distribution. However, nonparticipants were more likely to have a distant metastasis of the tumor. It is well known that having a more severe form of tumor at diagnosis is related to low socioeconomic status (SES) (4244). Therefore, the nonparticipating patients might have had a lower SES. Whether and how such SES differences between nonparticipants and participants might have biased the study results depends on whether SES is related to the use of electric bedding devices and in what direction the association points. Data on the relatedness of SES and use of electric bedding devices have been sparse. While some researchers have suggested that saving money on household heating costs may be a reason for use of an electric bedding device for some women (16) and therefore women with a lower SES may be more likely to use such a device, this idea was not supported by a study carried out in young women (45). We analyzed this relation using the data obtained from the participants. We found that study women with higher educational or income levels were actually more likely to use an electric bedding device than women with lower levels. This finding implies that use of electric bedding devices might be overrepresented in the participating patients because the nonparticipants might have had a lower SES and thus a lower frequency of use. However, the possibility of systematic differences among patients in influencing the odds ratio estimates also depends on whether nonparticipating and participating controls also differed with regard to SES and whether the association between SES and use of an electric bedding device was differential by case/control status. We were unable to compare either use of electric bedding devices or SES between the participating controls and the nonparticipating women who were eligible to serve as controls. Some studies in women without breast cancer have found that low-SES people are less likely to participate in a study than their counterparts with higher SES (46). This suggests that the nonparticipating controls might also have had a lower SES. However, the mechanisms of nonparticipation among eligible controls might differ from those for eligible patients, and the evidence has not been clear enough for us to ascertain such a relation. In addition, we are unaware of what possible effects monetary compensation might have had on participation. These issues prevent us from explicitly assessing possible impacts on the odds ratio estimates, although our data showed that the association between SES and use of electric bedding devices was not differential between cases and controls. Another factor that might have influenced the odds ratio estimates is whether SES modifies the effects of electric bedding device exposure on breast cancer risk. If participants were more likely to be in a certain SES stratum and nonparticipants were more likely to be in the other stratum, an overall odds ratio estimate obtained from the participants might be misleading because of effect modification. To evaluate such a possible effect, we analyzed the relation between use of electric bedding devices and breast cancer according to educational level and income level. Our results showed that the associations between bedding devices and breast cancer were similar for different educational and income levels, meaning that SES did not modify the association. The mechanisms of nonparticipation are complex, and the effects of nonparticipation are often difficult to evaluate because of a lack of sound data on nonparticipants. We cannot exclude the possibility that some degree of nonresponse bias was present in this study, but it is not possible to ascertain with any certainty its effects on the risk estimates. Inaccuracy in recalling the history of one’s electric bedding device use is another potential problem. For example, women may not feel able to precisely quantify the average number of months per year in which they have usually used an electric bedding device. Any errors in the determination of history of device use might have influenced our results. We cannot exclude the possibility that case women were more likely to be aware of the hypothesized association between electric bedding device use and breast cancer and therefore overreported their exposure to electric bedding devices. Such a possibility might not be substantial, however, because most previous studies did not find an association between the use of electric bedding devices and breast cancer, and therefore there may be no widespread perception of a hypothesized association. Hence, any misclassification in exposure assessment might not have been very differential by case/control status and might have been more likely to dilute the association only. Furthermore, the consistent dose-response relations observed with regard to all analyzed aspects of use suggest a low likelihood of systematic recall inaccuracies in all of these aspects. History of electric bedding device use as we investigated it might not precisely reflect EMF exposure due to use. First, our original questions did not distinguish between electric blankets and heated water beds. This prevented us from accurately assessing the effect of electric blankets, which generate higher EMFs. Second, as in previous epidemiologic studies on electric blanket use, we asked only about whether the participant had ever used an electric blanket, for how many years it had been used, during how many months per year (on average) it had been used, and whether the device had been kept on most of the time or used only for warming the bed. There are actually large variations in the intensity of magnetic fields from electric blankets because of differences in types of electric blankets, the design of electric blankets, and other use-related factors. For example, overblankets and underblankets differ in design, efficiency, and length of use at night (23). Electric blankets produced after the late 1980s generate less EMF exposure (24). An individual’s heat tolerance and the heating level on which the blanket is set may influence exposure intensity (15, 23). Ignorance of these factors might lead to misclassification of EMF exposure. These factors should be evaluated in future studies. ACKNOWLEDGMENTS This work was supported by grants DAMD17-96-6270 and DAMD17-97-7287 from the US Department of Defense. The authors express their deep appreciation to Rebecca Jones and Patrick Turri for identifying patients, Kimberly Newsom-Johnson for preparing tissue slides, Cathy Everett for contributing to data collection, Tina Cantrell for administrative assistance, Dr. Scott Davis for helpful comments on the manuscript, and Drs. Robert Levine, Louis Bernard, Josiah Ochieng, and Nasar Ahmed for their support during the original study. Correspondence to Dr. Kangmin Zhu, United States Military Cancer Institute, Walter Reed Army Medical Center, Building 1, Suite A-109, 6900 Georgia Avenue NW, Washington, DC 20307-5001 (e-mail: kangmin.zhu@na.amedd.army.mil). TABLE 1. Demographic characteristics of African-American women participating in a study of electric bedding device use and breast cancer risk, Tennessee, 1995–1998  Characteristic Cases Controls No. % No. % Age (years) 20–39 33 10.9 33 10.8 40–49 105 34.5 105 34.4 50–59 106 34.9 109 35.7 ≥60 60 19.7 58 19.0 Educational level High school graduation or less 113 37.2 140 45.9 Vocational/technical school 34 11.2 29 9.5 Some college or junior college 79 26.0 77 25.2 College graduation/graduate school/professional school 77 25.3 55 18.0 Other/unclear 1 0.3 3 1.0 Missing data 0 0.0 1 0.3 Annual income <$15,000 85 28.0 107 35.1 $15,000–$29,999 67 22.0 88 28.9 $30,000–$44,999 65 21.4 56 18.4 ≥$45,000 76 25.0 42 13.8 Missing data 11 3.6 12 3.9 Marital status Married 160 52.6 167 54.8 Unmarried 144 47.4 136 44.6 Missing data 0 0.0 2 0.7 Employment status Employed 213 70.1 208 68.2 Not employed 91 29.9 95 31.1 Missing data 0 0.0 2 0.7  Characteristic Cases Controls No. % No. % Age (years) 20–39 33 10.9 33 10.8 40–49 105 34.5 105 34.4 50–59 106 34.9 109 35.7 ≥60 60 19.7 58 19.0 Educational level High school graduation or less 113 37.2 140 45.9 Vocational/technical school 34 11.2 29 9.5 Some college or junior college 79 26.0 77 25.2 College graduation/graduate school/professional school 77 25.3 55 18.0 Other/unclear 1 0.3 3 1.0 Missing data 0 0.0 1 0.3 Annual income <$15,000 85 28.0 107 35.1 $15,000–$29,999 67 22.0 88 28.9 $30,000–$44,999 65 21.4 56 18.4 ≥\$45,000 76 25.0 42 13.8 Missing data 11 3.6 12 3.9 Marital status Married 160 52.6 167 54.8 Unmarried 144 47.4 136 44.6 Missing data 0 0.0 2 0.7 Employment status Employed 213 70.1 208 68.2 Not employed 91 29.9 95 31.1 Missing data 0 0.0 2 0.7
TABLE 2.

Association between use of electric bedding devices and breast cancer risk among African-American women aged 20–64 years, Tennessee, 1995–1998

 Variable No. of cases No. of controls Odds ratio* 95% CI† p for trend Use of an electric bedding device‡ No 205 213 1.0§ Yes 73 51 1.4 0.9, 2.2 No. of years of use No use 205 213 1.0§ 1–5 41 38 1.0 0.6, 1.7 6–10 16 8 1.8 0.7, 4.7 >10 16 5 4.9 1.5, 15.6 0.011 No. of seasons in which device was used per year No use 205 213 1.0§ ≤1 38 29 1.2 0.6, 2.1 >1 35 22 1.7 0.9, 3.2 0.115 Mode of use No use 205 213 1.0§ Used to warm bed only 17 21 1.0 0.5, 2.2 Kept on most of the time 56 29 1.7 1.0, 3.0 0.074
 Variable No. of cases No. of controls Odds ratio* 95% CI† p for trend Use of an electric bedding device‡ No 205 213 1.0§ Yes 73 51 1.4 0.9, 2.2 No. of years of use No use 205 213 1.0§ 1–5 41 38 1.0 0.6, 1.7 6–10 16 8 1.8 0.7, 4.7 >10 16 5 4.9 1.5, 15.6 0.011 No. of seasons in which device was used per year No use 205 213 1.0§ ≤1 38 29 1.2 0.6, 2.1 >1 35 22 1.7 0.9, 3.2 0.115 Mode of use No use 205 213 1.0§ Used to warm bed only 17 21 1.0 0.5, 2.2 Kept on most of the time 56 29 1.7 1.0, 3.0 0.074

* Adjusted for age, marital status, educational level, religion, household income, smoking, alcohol consumption, physical activity, body mass index (weight (kg)/height (m)2), number of pregnancies, age at first birth, history of benign breast disease, use of estrogen pills, family history of breast cancer, age at first intercourse, age at menarche, average menstrual cycle length, and menopausal status. Persons with missing data on any of these variables were excluded.

† CI, confidence interval.

‡ Electric blanket, electric mattress pad, or heated water bed.

§ Referent.

TABLE 3.

Association between use of electric bedding devices and breast cancer risk among African-American women aged 20–64 years after exclusion of those who used a device in more than two seasons per year, Tennessee, 1995–1998

 Variable No. of cases No. of controls Odds ratio* 95% CI† p for trend Use of an electric bedding device‡ No 205 213 1.0§ Yes 53 34 1.5 0.9, 2.5 No. of years of use No use 205 213 1.0§ 1–5 34 28 1.0 0.5, 1.9 6–10 9 3 2.7 0.6, 11.3 >10 10 3 6.3 1.5, 26.5 0.011 No. of seasons in which device was used per year No use 205 213 1.0§ ≤1 38 29 1.1 0.6, 2.0 >1 15 5 4.1 1.3, 13.2 0.038 Mode of use No use 205 213 1.0§ Used to warm bed only 16 16 1.2 0.5, 2.8 Kept on most of the time 37 17 1.9 1.0, 3.7 0.064
 Variable No. of cases No. of controls Odds ratio* 95% CI† p for trend Use of an electric bedding device‡ No 205 213 1.0§ Yes 53 34 1.5 0.9, 2.5 No. of years of use No use 205 213 1.0§ 1–5 34 28 1.0 0.5, 1.9 6–10 9 3 2.7 0.6, 11.3 >10 10 3 6.3 1.5, 26.5 0.011 No. of seasons in which device was used per year No use 205 213 1.0§ ≤1 38 29 1.1 0.6, 2.0 >1 15 5 4.1 1.3, 13.2 0.038 Mode of use No use 205 213 1.0§ Used to warm bed only 16 16 1.2 0.5, 2.8 Kept on most of the time 37 17 1.9 1.0, 3.7 0.064

* Adjusted for age, marital status, educational level, religion, household income, smoking, alcohol consumption, physical activity, body mass index (weight (kg)/height (m)2), number of pregnancies, age at first birth, history of benign breast disease, use of estrogen pills, family history of breast cancer, age at first intercourse, age at menarche, average menstrual cycle length, and menopausal status. Persons with missing data on any of these variables were excluded.

† CI, confidence interval.

‡ Electric blanket, electric mattress pad, or heated water bed.

§ Referent.

TABLE 4.

Association between use of electric bedding devices and breast cancer risk among African-American women aged 20–64 years, by menopausal status, Tennessee, 1995–1998

 Variable No. of cases No. of controls Odds ratio* 95% CI† p for trend Premenopausal women Use of an electric bedding device‡ No 84 94 1.0§ Yes 29 18 1.4 0.6, 3.4 No. of years of use No use 84 94 1.0§ 1–5 16 14 0.8 0.3, 2.4 6–10 5 2 1.5 0.2, 12.7 >10 8 2 8.3 1.1, 64.5 0.084 No. of seasons in which device was used per year No use 84 94 1.0§ ≤1 15 10 1.1 0.3, 3.3 >1 14 8 1.9 0.6, 6.5 0.343 Mode of use No use 84 94 1.0§ Used to warm bed only 5 6 1.0 0.2, 5.6 Kept on most of the time 24 11 2.1 0.7, 5.8 0.179 Postmenopausal women Use of an electric bedding device‡ No 121 119 1.0§ Yes 44 33 1.2 0.6, 2.1 No. of years of use No use 121 119 1.0§ 1–5 25 24 0.9 0.4, 1.8 6–10 11 6 1.4 0.4, 4.8 >10 8 3 3.8 0.8, 18.4 0.193 No. of seasons in which device was used per year No use 121 119 1.0§ ≤1 23 19 1.0 0.5, 2.2 >1 21 14 1.4 0.6, 3.3 0.522 Mode of use No use 121 119 1.0§ Used to warm bed only 12 15 0.7 0.3, 1.8 Kept on most of the time 32 18 1.6 0.7, 3.4 0.376
 Variable No. of cases No. of controls Odds ratio* 95% CI† p for trend Premenopausal women Use of an electric bedding device‡ No 84 94 1.0§ Yes 29 18 1.4 0.6, 3.4 No. of years of use No use 84 94 1.0§ 1–5 16 14 0.8 0.3, 2.4 6–10 5 2 1.5 0.2, 12.7 >10 8 2 8.3 1.1, 64.5 0.084 No. of seasons in which device was used per year No use 84 94 1.0§ ≤1 15 10 1.1 0.3, 3.3 >1 14 8 1.9 0.6, 6.5 0.343 Mode of use No use 84 94 1.0§ Used to warm bed only 5 6 1.0 0.2, 5.6 Kept on most of the time 24 11 2.1 0.7, 5.8 0.179 Postmenopausal women Use of an electric bedding device‡ No 121 119 1.0§ Yes 44 33 1.2 0.6, 2.1 No. of years of use No use 121 119 1.0§ 1–5 25 24 0.9 0.4, 1.8 6–10 11 6 1.4 0.4, 4.8 >10 8 3 3.8 0.8, 18.4 0.193 No. of seasons in which device was used per year No use 121 119 1.0§ ≤1 23 19 1.0 0.5, 2.2 >1 21 14 1.4 0.6, 3.3 0.522 Mode of use No use 121 119 1.0§ Used to warm bed only 12 15 0.7 0.3, 1.8 Kept on most of the time 32 18 1.6 0.7, 3.4 0.376

* Adjusted for age, marital status, educational level, religion, household income, smoking, alcohol consumption, physical activity, body mass index (weight (kg)/height (m)2), number of pregnancies, age at first birth, history of benign breast disease, use of estrogen pills, family history of breast cancer, age at first intercourse, age at menarche, average menstrual cycle length, and menopausal status. Persons with missing data on any of these variables were excluded.

† CI, confidence interval.

‡ Electric blanket, electric mattress pad, or heated water bed.

§ Referent.

References

1.
Stevens RG. Electric power use and breast cancer: a hypothesis.
Am J Epidemiol

1987
;
125
:
556
–61.
2.
Rasmussen H, Barrett PQ. Calcium messenger system: an integrated view.
Physiol Rev

1984
;
64
:
938
–84.
3.
Lacy-Hulbert A, Metcalfe JC, Hesketh R. Biological responses to electromagnetic fields.
FASEB J

1998
;
12
:
395
–420.
4.
Walleczek J, Liburdy R. Nonthermal 60 Hz sinusoidal magnetic-field exposure enhances 45-Ca2+ uptake in rat thymocytes: dependence on mitogen activation.
FEBS Lett

1990
;
271
:
157
–60.
5.
Barbier E, Dufy B, Veyret B. Stimulation of Ca2+ influx in rat pituitary cells under exposure to a 50 Hz magnetic field.
Bioelectromagnetics

1996
;
17
:
303
–11.
6.
Stevens R. Biologically based epidemiological studies of electric power and cancer.
Environ Health Perspect

1993
;
101
:
93
–100.
7.
Preston-Martin S, Peters JM, Yu MC, et al. Myelogenous leukemia and electric blanket use.
Bioelectromagnetics

1988
;
9
:
207
–13.
8.
Dosemeci M, Blair A. Occupational cancer mortality among women employed in the telephone industry.
J Occup Med

1994
;
36
:
1204
–9.
9.
Loomis DP, Savitz DA, Ananth CV. Breast cancer mortality among female electrical workers in the United States.
J Natl Cancer Inst

1994
;
86
:
921
–5.
10.
Coogan PF, Clapp RW, Newcomb PA, et al. Occupational exposure to 60-hertz magnetic fields and risk of breast cancer in women.
Epidemiology

1996
;
7
:
459
–64.
11.
Coogan PF, Aschengrau A. Exposure to power frequency magnetic fields and risk of breast cancer in the Upper Cape Cod Cancer Incidence Study.
Arch Environ Health

1998
;
53
:
359
–67.
12.
Vena JE, Graham S, Hellmann R, et al. Use of electric blankets and risk of postmenopausal breast cancer.
Am J Epidemiol

1991
;
134
:
180
–5.
13.
Vena JE, Freudenheim JL, Marshall JR, et al. Risk of premenopausal breast cancer and use of electric blankets.
Am J Epidemiol

1994
;
140
:
974
–9.
14.
Gammon MD, Schoenberg JB, Britton JA, et al. Electric blanket use and breast cancer risk among younger women.
Am J Epidemiol

1998
;
148
:
556
–63.
15.
Zheng T, Holford TR, Mayne ST, et al. Exposure to electromagnetic fields from use of electric blankets and other in-home electrical appliances and breast cancer risk.
Am J Epidemiol

2000
;
151
:
1103
–11.
16.
McElroy JA, Newcomb PA, Remington PL, et al. Electric blanket or mattress cover use and breast cancer incidence in women 50–79 years of age.
Epidemiology

2001
;
12
:
613
–17.
17.
Cantor KP, Dosemeci M, Brinton LA, et al. Re: “Breast cancer mortality among female electrical workers in the United States.” (Letter).
J Natl Cancer Inst

1995
;
87
:
227
–8.
18.
Harlow BL, Davis S. Two one-step methods for household screening and interviewing using random digit dialing.
Am J Epidemiol

1988
;
127
:
857
–63.
19.
Pi-Sunyer FX. Obesity: criteria and classification.
Proc Nutr Soc

2000
;
59
:
505
–9.
20.
Cote RJ, Taylor CR. Immunohistochemical detection of steroid hormone receptors. In: Taylor CR, Cote RJ, eds. Immunomicroscopy: a diagnostic tool for the surgical pathologist. Philadelphia, PA: W B Saunders Company, 1994:227–87.
21.
Allred DC, Bustamante MA, Daniel CO, et al. Immunocytochemical analysis of estrogen receptors in human breast carcinomas: evaluation of 130 cases and review of the literature regarding concordance with biochemical assay and clinical relevance.
Arch Surg

1990
;
125
:
107
–13.
22.
Hosmer D, Lemeshow S. Applied logistic regression. New York, NY: John Wiley and Sons, Inc, 1989.
23.
Delpizzo V. A model to assess personal exposure to ELF magnetic fields from common household sources.
Bioelectromagnetics

1990
;
11
:
139
–47.
24.
Kaune WT. Assessing human exposure to power-frequency electric and magnetic fields.
Environ Health Perspect

1993
;
101(suppl 4)
:
121
–33.
25.
Aviv A. Cellular calcium and sodium regulation, salt-sensitivity and essential hypertension in African Americans.
Ethn Health

1996
;
1
:
275
–81.
26.
Fekete Z, Kimura M, Horiguchi M, et al. Differences between store-dependent Ca2+ fluxes in lymphocytes from African Americans and whites.
J Hypertens

1996
;
14
:
1293
–9.
27.
Cheng I, Kingensmith M, Chattopadhyay N, et al. Identification and localization of the extracellular calcium-sensing receptor in human breast.
J Clin Endocrinol Metab

1998
;
83
:
703
–7.
28.
Mevissen M, Haussler M, Lerchl A, et al. Acceleration of mammary tumorigenesis by exposure of 7,12-dimethylbenz[a]anthracene-treated female rats in a 50-Hz, 100-µT magnetic field: replication study.
J Toxicol Environ Health A

1998
;
53
:
401
–18.
29.
Goodman EM, Greenebaum B, Marron MT. Effects of electromagnetic fields on molecules and cells.
Int Rev Cytol

1995
;
158
:
279
–338.
30.
Gammon MD, Hibshoosh H, Terry MB, et al. Cigarette smoking and other risk factors in relation to p53 expression in breast cancer among young women.
Cancer Epidemiol Biomarkers Prev

1999
;
8
:
255
–63.
31.
Rose DP, Royak-Schaler R. Tumor biology and prognosis in black breast cancer patients: a review.
Cancer Detect Prev

2001
;
25
:
16
–31.
32.
Weston A, Godbold JH. Polymorphisms of H-ras-1 and p53 in breast cancer and lung cancer: a meta-analysis.
Environ Health Perspect

1997
;
105(suppl 4)
:
919
–26.
33.
Gu J, Spitz MR, Yang F, et al. Ethnic differences in poly(ADP-ribose) polymerase pseudogene genotype distribution and association with lung cancer risk.
Carcinogenesis

1999
;
20
:
1465
–9.
34.
Brown P. Race, class, and environmental health: a review and systematization of the literature.
Environ Res

1995
;
69
:
15
–30.
35.
Popkin BM, Siega-Riz AM, Haines PS. A comparison of dietary trends among racial and socioeconomic groups in the United States.
N Engl J Med

1996
;
335
:
716
–20.
36.
Cooper R. The role of genetic and environmental factors in cardiovascular disease in African Americans.
Am J Med Sci

1999
;
317
:
208
–13.
37.
Brainard GC, Kavet R, Kheifets LI. The relationship between electromagnetic field and light exposures to melatonin and breast cancer risk: a review of the relevant literature.
J Pineal Res

1999
;
26
:
65
–100.
38.
Liburdy RP, Sloma TR, Sokolic R, et al. ELF magnetic fields, breast cancer, and melatonin: 60 Hz fields block melatonin’s oncostatic action on ER+ breast cancer cell proliferation.
J Pineal Res

1993
;
14
:
89
–97.
39.
Feychting M, Forssen U, Rutqvist LE, et al. Magnetic fields and breast cancer in Swedish adults residing near high-voltage power lines.
Epidemiology

1998
;
9
:
392
–7.
40.
Forssen UM, Feychting M, Rutqvist LE, et al. Occupational and residential magnetic field exposure and breast cancer in females.
Epidemiology

2000
;
11
:
24
–9.
41.
Van Wijngaarden E, Nylander-French LA, Millikan RC, et al. Population-based case-control study of occupational exposure to electromagnetic fields and breast cancer.
Ann Epidemiol

2001
;
11
:
297
–303.
42.
Oluwole SF, Ali AO, Adu A, et al. Impact of a cancer screening program on breast cancer stage at diagnosis in a medically underserved urban community.
J Am Coll Surg

2003
;
196
:
180
–8.
43.
Bradley CJ, Given CW, Roberts C. Race, socioeconomic status, and breast cancer treatment and survival.
J Natl Cancer Inst

2002
;
94
:
490
–6.
44.
Merkin SS, Stevenson L, Powe N. Geographic socioeconomic status, race, and advanced-stage breast cancer in New York City.
Am J Public Health

2002
;
92
:
64
–70.
45.
Belanger K, Leaderer B, Hellenbrand K, et al. Spontaneous abortion and exposure to electric blankets and heated water beds.
Epidemiology

1998
;
9
:
36
–42.
46.
Blumenthal DS, Sung J, Coates R, et al. Recruitment and retention of subjects for a longitudinal cancer prevention study in an inner-city black community.
Health Serv Res

1995
;
30
:
197
–205.