Exposure to extremely low-frequency electromagnetic fields (EMF) has been hypothesized to increase the risk of breast cancer by inhibiting the normal nocturnal rise in melatonin levels.1-3 Suppression of melatonin may, in turn, lead to elevated estrogen levels, thus increasing breast cancer risk.2,3 Information on electric blanket use may provide insight into the association of exposure to EMF with disease, because regular use could entail substantial exposure to EMF as a result of the proximity of the blanket to the body and the habitual nature of the exposure. Eight published studies have addressed the association of exposure to electric blankets and breast cancer.4-12 These studies show either no association7-12 or a slight association with continuous use throughout the night.4-6 Several of the studies had limited information on electric blanket use,7 or they combined electric blankets with other bed-warming devices, such as mattress pads and heated waterbeds.7,8,11
We report here on the association of electric blanket use and breast cancer in a population-based investigation of breast cancer on Long Island. A brief questionnaire ascertaining information on electric blanket use and other selected appliances had been included as part of the Long Island Breast Cancer Study Project (LIBCSP) interview. The primary purpose of the LIBCSP was to assess the association of breast cancer with exposure to organochlorine compounds and other environmental pollutants.13 A companion study, the Electromagnetic Fields and Breast Cancer on Long Island Study (EBCLIS), was conducted specifically to evaluate the association between exposure to EMF and breast cancer.14 It included the subset of LIBCSP participants under 75 years of age who lived in their current home for at least 15 years. In this second study, a more detailed interview was conducted to ascertain electric blanket use, exposure to other appliances, and occupational exposures; in addition, extensive measurements of EMF were made in the homes.
Long Island Breast Cancer Study Project
The background, design, and recruitment of the LIBCSP have been described previously.13 The study was approved by the institutional human subject review boards of participating institutions. Briefly, cases were Nassau and Suffolk county women newly diagnosed with a first primary in situ or invasive breast cancer between August 1, 1996, and July 31, 1997. Physicians of potentially eligible case women were contacted to confirm the diagnosis and the date of diagnosis and for permission to contact the women. Control women were a sample of current residents of Nassau and Suffolk counties who were frequency-matched to the expected distribution of cases by 5-year age groups. Control subjects under 65 years of age were identified by random-digit dialing; control subjects aged 65 and over were selected from Health Care Financing Administration (HCFA) rosters.
Potential study participants were first recruited by mail to introduce them to the study. This was followed by a telephone call to answer questions and to arrange for an in-home visit. At this visit, women provided signed informed consent and then participated in a comprehensive, 2-hour questionnaire. Women also provided blood and urine specimens. At the time of the LIBCSP interview, women were given a brochure informing them of EBCLIS.
Electromagnetic Fields and Breast Cancer on Long Island Study
The background, design, and recruitment of EBCLIS have been described in detail.14 The study was approved by the institutional human subjects review boards of participating institutions. Women eligible for EBCLIS were those LIBCSP participants who had lived in their current residences for 15 years or more (long-term residents). Cases had to be under 75 years of age at diagnosis and control subjects had to be under 75 at their date of identification (reference date). The upper age limit for eligibility was selected because of the relatively small number of women recruited by the LIBCSP at ages 75 and over.13 The target sample size for EBCLIS was 600 cases and 600 control subjects. To meet these targets, recruitment was based on eligible cases diagnosed between August 1, 1996, and June 20, 1997, and a random sample of LIBCSP control subjects frequency-matched by age. Of 663 cases and 702 control subjects who were eligible for the study, 576 cases (87%) and 585 control subjects (83%) participated. The EBCLIS interviewers visited the homes of participants where they obtained signed informed consent, administered a 30-minute interview and took EMF measurements in 2 rooms of the home and at the front door.14 At a second visit, the electrical wiring configuration of each residence was diagrammed from outside the home.15
LIBCSP participants under 75 years of age included 1354 cases and 1426 control subjects, of whom 86% and 69% participated, respectively. Because EBCLIS cases and control subjects were a subset of the LIBCSP population, the EBCLIS participation rates are dependent on the participation rates of the LIBCSP. It is difficult to calculate a net response rate for EBCLIS based on the LIBCSP participation rates, because the residential history of nonparticipants in the LIBCSP was not determined; as such, it was not possible to determine how many were potentially eligible for EBCLIS. It is also difficult to determine the true impact of the lower control participation rates in the LIBCSP on the EBCLIS results, especially among older women, who would also be more likely to have lived in their homes for 15 years or longer.
Exposure to Electric Blankets and Related Appliances
The LIBCSP questionnaire included separate questions about “ever use” (defined as use at least 3 times in any 1 year) of an electric blanket, an electric mattress pad, a waterbed with a heater, or an electric heating pad. For each item, respondents were asked about their age at first use, years of use, frequency of use, whether the device was used to “warm the bed only” or used “directly on” (directly on the body), and how many hours per day it was used. The LIBCSP questionnaire also included questions about other electric appliances used in the home (eg, personal computers, microwave ovens, hair dryers). The EBCLIS questionnaire asked additional electric blanket questions of women who had answered yes to “ever use” on the LIBCSP questionnaire. These additional questions included whether the electric blanket was used “throughout the night” and the usual setting (high, medium, low, or no temperature setting). We also asked whether the heat was on or off or the window open or closed in the room while the blanket was in use because if the heat was off or the window open, the blanket would be more likely to cycle on and off and thus might produce higher magnetic fields. We also asked whether the electric blanket was purchased before or after 1989 because blankets purchased before 1989 produced stronger magnetic fields than those purchased afterward.16 Information was collected on other household and occupational appliances not included in the LIBCSP questionnaire (eg, electric stoves/oven, televisions, photocopiers). Analyses of household and occupational appliance use are ongoing.
Responses to the LIBCSP questionnaire were available for the total LIBCSP population under the age of 75 (N = 2780), whereas additional questions were asked only of the subgroup included in EBCLIS (N = 1161). For this reason, we analyzed the 2 groups separately. Some analyses were stratified by menopausal status, because it has been suggested that EMF exposure might affect premenopausal women differently as a result of their higher circulating estrogen levels.6,17
Menopausal status was defined using information provided by participants on date of last menstrual period, prior hysterectomy or oophorectomies, cigarette smoking, and hormone replacement therapy use. We defined a participant as postmenopausal if her last menstrual period was more than 6 months before the reference date or if both ovaries had been removed before the reference date (88% of participants). The remaining women either were taking hormone replacement therapy or had undergone a hysterectomy without removal of both ovaries. Lacking the usual signs and symptoms of menopause, these women were assigned menopausal status based on their reference age and smoking status. We considered a smoker to be postmenopausal if her age at reference was ≥54.8 years (90th percentile for natural menopause among smoking control subjects), and a nonsmoker was considered postmenopausal if her age at reference was ≥55.4 years (90th percentile for natural menopause among nonsmoking control subjects). Imputation of menopausal status based on age and smoking status reduced the percentage of women with unknown menopausal status to 3%.13
Because women who used the blanket only to warm the bed would incur minimal magnetic field exposure, we defined electric blanket use as having the blanket directly on the body. This subset of electric blanket users formed the basis for analyses of frequency and duration of use among LIBCSP and EBCLIS women, as well as more detailed analyses of use for the EBCLIS subpopulation (throughout the night or not, temperature setting, heat on/off in the bedroom, window open/closed in the bedroom, and year blanket was purchased).
Frequency of electric blanket use was defined as a continuous variable and dichotomized as less than 90 times per year or 90-365 times per year to distinguish less frequent users from daily users during the winter or daily users year round. We did not ascertain seasonal use; therefore, daily users throughout a season could not be distinguished from daily users year round. Years of electric blanket use was defined as a continuous variable and categorized into quartiles with cut-points based on the distribution in the control subjects. It has been proposed that EMF might act as a cancer promoter, and thus continued use up to the time of diagnosis might be an important aspect of exposure.18 We collected information on age at first use and duration of use up to the LIBCSP interview date, but did not specifically ask if the woman was currently using the blanket. To define an approximate measure of “current use” as of the reference date, we defined a woman as a “current user” if her age at interview was within 3 years of the age at first use of an electric blanket plus the number of years of use. We chose a 3-year window to take into account the time between the interview date and the reference date (average of 96 days for cases and 167 days for control subjects13), and to allow for any inaccuracy in recalling the age at first use and number of years of use. Similarly, “former users” were women whose age at interview was at least 3 years greater than the sum of the age at first use of the blanket plus the number of years of use.
In univariate analyses, we examined associations between breast cancer status and potential breast cancer risk factors using the Pearson χ2 test. Independent variables for multivariate analyses were selected from variables associated with breast cancer in the univariate analyses (P ≤0.25). We performed unconditional logistic regression analyses for all women and for premenopausal and postmenopausal women separately to estimate adjusted odds ratios (ORs) and 95% confidence intervals. For the models including all women, we included the following covariates: age at reference date (date of diagnosis for cases and similar date assigned to control subjects, continuous), parity (number of live-births, continuous), education (defined as less than high school or high school graduate as the referent vs. some college, college graduate, or postcollege), first-degree family history of breast cancer (mother, sister, or daughter with breast cancer), and history of benign breast disease. The premenopausal models included the same variables as the model for all women, with the addition of body mass index (weight in kilograms/height in meters squared) at reference date (continuous). The postmenopausal models included the same variables as the premenopausal model with the addition of body mass index at age 20 (continuous). Tests for trend in breast cancer risk with increasing years of use were computed using the Wald χ2 statistic in unconditional logistic regression models.
Table 1 presents results for women in the LIBCSP stratified by menopausal status. For all indicators of exposure, the point estimates were close to the null and all confidence intervals included 1.0. There was no trend in the ORs with increasing years of electric blanket use and no association with greater frequency of use. When women were cross-classified by frequency and duration of use, there was no association in women with both greater frequency and longer duration of use (data not shown).
As evident in Table 2, analyses among women in the EBCLIS stratified by menopausal status yielded similar results. All ORs, including those for “ever use,” for use “directly on,” and for use “directly on throughout the night,” were close to 1.0. Furthermore, there was no suggestion of an association in subgroups likely to have higher exposure (those who reported using higher setting, having the heat off, having the window open, or having used an older electric blanket). Again, there was no trend with increasing years of use, no association with greater frequency of use, and no association in women with both longer duration and greater frequency of use (data not shown).
Because some women could not be classified as premenopausal or postmenopausal (31 cases and 62 control subjects), we repeated the main analyses categorizing the age at reference date into the strata <50 and ≥50 years, and into decades of age at reference date, rather than stratifying by menopausal status. Additional analyses included all participants with menopausal status as a covariate. The results of these analyses for all LIBCSP participants and for the EBCLIS subpopulation (data not shown) were similar to those presented in Tables 1 and 2.
When combined use of electric blankets, electric mattress pads, and heated waterbeds was compared with never use of any of these appliances for the LIBCSP population, the results were essentially identical to those presented previously. Taking those never exposed to any of the bed-warming devices (rather than just electric blankets) as the referent group also had no effect on the results.
When all women with breast cancer in the LIBCSP were partitioned according to whether their cancer was in situ (16%) or invasive (84%), no association with electric blanket use was found, although the number of in situ cases was small.
The association of electric blanket use with breast cancer by estrogen receptor (ER) and progesterone receptor (PR) status in the LIBCSP is presented in Table 3. Hormone receptor data were not available for approximately one third of cases. The number of cases who were positive for one receptor and negative for the other was small, particularly among premenopausal women. For cases who were positive or negative for both receptors, there was no indication of an association between hormone receptor status and breast cancer.
We collected detailed information on use of electric blankets and other bed-warming devices to explore their possible association with breast cancer risk. Our results provide no support for an association of electric blanket use with breast cancer in premenopausal women, postmenopausal women, or all women combined. Of the 8 previous studies,4-12 only the 2 studies by Vena et al.4,5 show any indication of increased risk. The remaining studies show ORs close to 1.0 for use of electric blankets or combined use of electrical heating devices with little variation in risk by mode of use, duration of use, menopausal status, or hormone receptor status.7-12 Our results are consistent with these findings.
Premenopausal women have higher circulating estrogen levels and, for this reason, could be more vulnerable to the EMF effects that have been hypothesized to increase the risk of breast cancer.6 One study found a stronger association of occupational EMF exposure with breast cancer among premenopausal women.17 However, previous studies of electric blankets that stratified by menopausal status do not suggest a stronger association among premenopausal women.5,7-10 In the present study, the overall ORs in premenopausal women did not differ from 1.0, and there was no increased breast cancer risk with indices of higher exposure.
It is also plausible that an association between EMF and breast cancer might be more apparent in cases with hormone receptor-positive cancers (ER+/PR+), because this group has been postulated to be more susceptible to hormone-related risk factors, whereas hormone receptor-negative cancers (ER−/PR−) may be related to risk factors independent of hormonal exposure.19-21 In the present study, the OR for ever-use was close to 1.0 for ER+/PR+ and ER−/PR− cases among both premenopausal and postmenopausal women. Previous studies have similarly found no evidence of an association with hormone receptor status.8,9 Two studies in Sweden showed elevated risks in ER-positive breast cancers in women under the age of 50 with higher residential22 or occupational23 EMF exposure, but with very wide confidence intervals.
LIBCSP participation rates in women under 75 were 86% for cases and 69% for control subjects. The largest difference in participation rates was in the 65- to 74-year age group, with 80% in cases and 52% in control subjects.13 We examined the associations between ever use of electric blankets and breast cancer risk in a 10-year age strata and found no association in any stratum, suggesting that the higher nonresponse rate in older control subjects did not overly affect our findings.
A number of limitations in the assessment of EMF exposure from electric blankets should be pointed out. Self-reported exposures, especially those covering long time periods, are inherently imprecise. The questionnaires used in this study elicited age at first use and duration of use, but did not inquire about different periods of use through the respondent’s life, current use, or years since stopping. Current and former use were inferred from age at first use and duration of use. Therefore, if a woman discontinued use and then resumed use up to the time of interview, she might be misclassified as a former user instead of a current user. However, because the OR for current users was below 1.0 and that for former users was slightly above 1.0, the ORs would most likely both be closer to unity without this misclassification. Finally, it was not possible to distinguish between daily electric blanket use during the winter only and daily use throughout the year because we did not specifically ask about season of use. For this reason, frequency of use was categorized as never, fewer than 90 times per year, and 90-365 times per year (daily use year round or daily use in the winter months).
Acknowledging these limitations, we conclude that the results of this comprehensive study do not support an association between electric blanket use and breast cancer. Our results are consistent with those of previous investigations.
We thank the study participants and also the Scientific Advisory Committee: David A. Savitz (chair), Louise Brinton, Fred Dietrich, Norma Logan, Richard Stevens, and Sholom Wacholder. We appreciate the assistance of their Local Advisory Committee: Mary Dowden, Miriam Goodman, and Mary Joan Shea.
The Electromagnetic Fields and Breast Cancer on Long Island Study (EBCLIS) Group included: at Stony Brook, M. Cristina Leske (PI), Sang Ahnn, Judith M. Greene, Roger Grimson, Kevin Henderson, Geoffrey C. Kabat, Erin S. O’Leary, and Elinor Randi Schoenfeld; at WESTAT, Carol Haines and Jacqueline Slattery-Telonidis; and at EM Factors, William T. Kaune. The Long Island Cancer Study Project (LIBCSP) team was composed of Marilie D. Gammon (PI), Julie A. Britton, Alfred I. Neugut, and Susan L. Teitelbaum.
1.Caplan LS, Schoenfeld ER, O’Leary ES, et al. Breast cancer and electromagnetic fields—a review. Ann Epidemiol
2.Stevens RG. Electric power use and breast cancer. A hypothesis. Am J Epidemiol
3.Stevens RG, Davis S. The melatonin hypothesis: electric power and breast cancer. Environ Health Perspect
. 1996;104(suppl 1):135–140.
4.Vena JE, Graham S, Hellman R, et al. Use of electric blankets and risk of postmenopausal breast cancer. Am J Epidemiol
5.Vena JE, Freudenheim JL, Marshall JR, et al. Risk of premenopausal breast cancer and use of electric blankets. Am J Epidemiol
6.Vena JE, Marshall JR, Freudenheim JL, et al. The authors reply [Letter]. Am J Epidemiol
7.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
8.Gammon MD, Schoenberg JB, Britton JA, et al. Electric blanket use and breast cancer risk among younger women. Am J Epidemiol
9.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
10.Laden F, Neas LM, Tolbert PE, et al. Electric blanket use and breast cancer in the Nurses’ Health Study. Am J Epidemiol
11.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
12.Davis S, Mirick DK, Stevens RG. Residential magnetic fields and the risk of breast cancer. Am J Epidemiol
13.Gammon MD, Neugut AI, Santella RM, et al. The Long Island Breast Cancer Study Project: description of a multi-institutional collaboration to identify environmental risk factors for breast cancer. Breast Cancer Research and Treatment
14.Schoenfeld ER, O’Leary ES, Henderson K, et al. for the EBCLIS group. Electromagnetic fields and breast cancer on Long Island: a case-control study. Am J Epidemiol
15.O’Leary ES, Schoenfeld ER, Henderson K, et al. for the EBCLIS group. Wire coding in the EMF and Breast Cancer on Long Island Study—relationship to magnetic fields. Journal of Exposure Analysis and Environmental Epidemiology
16.Florig HK, Hoburg JF. Power frequency magnetic fields from electric blankets. Health Physics
17.Coogan PF, Clapp RW, Newcomb PA, et al. Occupational exposure to 60 Hertz magnetic fields and risk of breast cancer in women. Epidemiology
18.Stevens RG. Risk of premenopausal breast cancer and use of electric blankets. Am J Epidemiol
19.Potter JD, Cerhan JR, Sellers TA, et al. Progesterone and estrogen receptors and mammary neoplasia in the Iowa Women’s Health Study: how many kinds of breast cancer are there? Cancer Epidemiol Biomarkers Prev
20.Yasui Y, Potter JD. The shape of age-incidence curves of female breast cancer by hormone-receptor status. Cancer Causes Control
21.Huang W-Y, Newman B, Millikan RC, et al. Hormone-related factors and risk of breast cancer in relation to estrogen receptor and progesterone receptor status. Am J Epidemiol
22.Feychting M, Forssen U, Rutquist LE, et al. Magnetic fields and breast cancer in Swedish adults residing near high-voltage power lines. Epidemiology
23.Forssen UM, Feychting M, Rutquist LE, et al. Occupational and residential magnetic field exposure and breast cancer in females. Epidemiology