Background: Melatonin shows potential oncostatic activity and is acutely suppressed by light exposure. Some evidence suggests an association between night work and breast cancer risk, possibly through the melatonin pathway.
Methods: In a cohort of premenopausal nurses, we prospectively studied the relation between rotating night shift work and breast cancer risk. Total number of months during which the nurses worked rotating night shifts was first assessed at baseline in 1989 and periodically updated thereafter. We used Cox proportional hazards models to calculate relative risks (RRs) and 95% confidence intervals (CIs).
Results: Among 115,022 women without cancer at baseline, 1,352 developed invasive breast cancer during 12 years of follow up. Women who reported more than 20 years of rotating night shift work experienced an elevated relative risk of breast cancer compared with women who did not report any rotating night shift work (multivariate RR = 1.79; 95% CI = 1.06–3.01). There was no increase in risk associated with fewer years of rotating night work.
Conclusion: Our results suggest a modestly elevated risk of breast cancer after longer periods of rotating night work. Additional studies are warranted to rule out small sample size or uncontrolled sources for confounding as alternative explanations.
From the *Channing Laboratory, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts; †Ludwig Boltzmann-Institute for Applied Cancer Research, KFJ-Spital, and Applied Cancer Research–Institute for Translational Research Vienna (LBI-ACR VIEnna & ACR-ITR VIEnna), Austria; and the Departments of ‡Epidemiology and §Environmental Health, Harvard School of Public Health, Boston, Massachusetts.
Submitted 4 March 2005; accepted 10 August 2005.
Correspondence: Eva S. Schernhammer, Channing Laboratory, 181 Longwood Avenue, Boston, MA 02115. E-mail: email@example.com.
Environmental lighting powerfully influences the circadian system in humans.1 In particular, light exposure may have an adverse effect on breast cancer risk through suppression of melatonin, a hormone that is intimately linked to the circadian system and demonstrates cancer-protective capability in experimental models.2 Observational studies have consistently associated night work with an increase in breast cancer risk,3 but these studies were mostly retrospective in design. The only prospective study4 on the possible link between night shift work and breast cancer was the original Nurses’ Health Study, which comprised primarily postmenopausal women.
The Nurses’ Health Study II is a prospective cohort study that began in 1989, when 116,671 registered female U.S. nurses of ages 25 to 42 years were enrolled. Since 1989, they have completed biennial mailed questionnaires that include items about their health status and known or suspected risk factors for cancer.5 Response rates to questionnaires are at 90%. Further details of the Nurses’ Health Study II are described elsewhere.6
Assessment of Night Shift Working Status
The 1989 questionnaire included detailed questions on total months during which study participants had worked on rotating night shifts for at least 3 nights per month in addition to having worked days or evenings in that month. This information was updated in 1991, 1993, and 1997. The prespecified categories for total numbers of months working on rotating night shifts were none, 1–4, 5–9, 10–14, 15–19, and 20 or more. Because the 1995 and 1999 questionnaires did not assess night work information, women were queried retrospectively in 2001 for the time periods 1993–1995 and 1997–1999. Questions were asked regarding both rotating night shifts, as previously described, and permanent night shifts for 6 or more months.
Documentation of Breast Cancer and Deaths
We identified breast cancer cases as cases occurring between 1 June 1989 and 1 June 2001. Nurses who reported breast cancer were asked for permission to review their medical records, and breast cancer was confirmed through review of these records. In addition, approximately two thirds of the deaths among cohort members were reported to us by next of kin or the postal system in response to follow-up questionnaires. We also searched the National Death Index to identify deaths among the nonrespondents to each 2-year questionnaire.7 A total of 1,352 cases of breast cancer were reported in the base population during 12 years of follow up, and pathology records were obtained for 98%. We based our analyses on the total, because the accuracy of the self-reporting was extremely high in this cohort (>99%).
The population for this study consisted of 116,087 women who returned the 1989 questionnaire and answered the question on night work (99.5% of the total). Women with a history of cancer (except nonmelanoma skin cancer) at baseline were excluded. A total of 115,022 women remained to form the baseline population for this analysis, and 1,313,203 person years of follow-up were accrued from 1989 to 2001.
We computed person-years of follow up to the date of diagnosis of breast cancer, death from any cause, or the end of the study period (1 June 2001), whichever came first. Information on lifetime history of rotating night shift work obtained from the 1989 questionnaire was updated according to the biennial follow-up questionnaires. Gaps in the prospectively collected information for 1993–1995 and 1997–1999 were filled with retrospectively assessed information on months of rotating night work during these 2-year periods, as ascertained on the 2001 questionnaire. To evaluate the validity of this approach, we conducted sensitivity analyses either carrying forward night work information from the previous cycle or setting number of months worked on rotating night shifts to zero for these 2 time periods.
Women were categorized according to the total number of years they worked rotating night shifts at least 3 nights per month in addition to days or evenings in that month (“never,” “1–9 years,” “10–19 years,” or “≥20 years”). Information about breast cancer and established risk factors for breast cancer was updated biennially. We calculated Mantel-Haenszel summary relative risks (RRs) and associated 95% confidence intervals (CIs) adjusting for age in 5-year categories.8 Cox proportional hazards models were used to calculate RRs and 95% CIs with adjustment for age (in years) and a number of known breast cancer risk factors.
At baseline, in 1989, women who had never worked on rotating night shifts accounted for 38% of the person-years of follow up among all women. Women who worked rotating shifts were similar in their characteristics to those who had never done such work (Table 1). However, night shift workers were older and tended to smoke more than those who had never worked on rotating nightshifts.
Table 2 shows the relationship between total number of years on rotating night shifts and breast cancer. Higher duration of shift work was associated with an increased breast cancer risk; women who had worked for 20 or more years on rotating night shifts had a 79% greater risk of breast cancer compared with women who had never worked this schedule (multivariate-adjusted RR =1.79; 95% CI = 1.06–3.01; P for trend = 0.65). Moreover, results were similar when the same analysis was limited to rotating night shifts at baseline (2.79; 1.38–5.66). However, this analysis was based on only 8 cases exposed at baseline who reported 20 or more years of night work in 1989.
In our main analyses, we used retrospective information from 2001 to create a composite night work variable. We also conducted 2 types of sensitivity analyses. In the first, rather than using the 2001 data retrospectively, we conservatively assumed that nurses had not worked a single month on rotating night shifts during the time periods in question, which produced essentially no change in the results (2.01; 1.17–3.44). Second, we carried forward the last reported value (ie, from the previous questionnaire cycle), an approach commonly applied for missing information in updated analyses. In this analysis, we saw a slightly weaker association (1.56; 0.91–2.66) when compared with our main approach.
In stratified analyses, although limited by small case numbers, we observed no effect modification by body mass index or smoking history. The exclusion of current smokers did strengthen our estimates but only modestly (1.92; 1.10–3.35).
It is possible that women stopped working night shifts due to early symptoms associated with breast cancer. If true, such a “healthy worker effect” would have biased our results toward the null. To address this hypothesis, we repeated our analyses excluding the first 4 years of follow up after the baseline report on night work (1989–1993). We analyzed the relation between number of nights worked as assessed in 1989 and the risk of breast cancer from 1993 through 2001. This 4-year latency period was associated with an increase in breast cancer risk (1.79; 0.88–3.63), similar to the risk we observed in our main analyses.
In this large prospective cohort study with detailed and updated information on night shift work, the risk of breast cancer was found to be modestly elevated in women who worked for 20 or more years on rotating night shifts compared with those who never worked rotating night shifts. Our finding, which is consistent with previous studies suggesting a link between night work and breast cancer risk, is the first prospective report on this association in premenopausal women.
Several retrospective studies9–15 and one prospective study4 of night workers have found an elevated breast cancer risk associated with occupational exposure to light at night. The only prospective cohort study to date used data from the original Nurses’ Health Study.4 In that study, the RR associated with extended periods (30 or more years) of rotating night work was 1.36 (95% CI = 1.04–1.78) after controlling for known breast cancer risk factors. This risk increased with numbers of years of shift work (P for trend = 0.02). Studies of flight attendants also consistently find that female cabin crew members are at increased risk for breast cancer,14 although the incomplete assessment of possible confounding factors remains an important limitation of these retrospective studies.
Although we did not validate self-reported duration of rotating night shifts, it is likely that our results are accurate, because other self-reports have been highly accurate in a similar cohort,16 and previous validations of similar questions (eg, electric blanket use)17 have shown reasonable reproducibility. Moreover, the prospective design of our study eliminates recall bias. We may have missed women who were on permanent nightshifts and thus perhaps did not classify themselves as working on rotating shifts; assuming that the average serum melatonin levels would likely be higher than those of rotating shift workers (because of better entrainment of their circadian rhythm), we anticipate that such misclassification would bias our results toward the null. In secondary analyses, we excluded women who had worked 6 or more months of permanent night work, with similar findings, thus providing support for our theory.
A potential limitation in our study is that women who work most frequently on night shifts may differ from women with no night shift experience in ways that influence risk of breast cancer and in ways we were not able to control. For example, there was a marked difference in smoking history among women who never had worked night shift and those with 20 or more years of night work; moreover, smoking duration rather than current smoking status was most strongly related to breast cancer risk in this cohort.18 However, neither the additional adjustment for pack-years smoked nor the exclusion of smokers from our analyses altered our results. Nonetheless, although we controlled for known potential confounding factors, uncontrolled confounding may stem from differences in socioeconomic status or other differences in lifestyle that we were not able to consider.
Another limitation of our study is that few women had worked rotating night shifts for 20 or more years, leading to small case numbers in the longest exposure categories. Thus, although we have seen similarly increased risks of breast cancer with 30 or more years of night work based on much larger case numbers in our previous study4 of older women (Nurses’ Health Study), larger studies of premenopausal women are still needed to confirm our findings. In addition, we had no information on the actual number of nights worked per month. It is conceivable that women with the longest duration of night work are also those who work the highest number of night shifts per month. Thus, frequency of night shifts, as well as number of years of such work, should be considered in future studies.
In conclusion, working on rotating night shifts was associated with a modestly increased breast cancer risk among the female nurses in our cohort. The findings from our study, in combination with the results of earlier work, reduce the likelihood that this association is due solely to chance. Because breast cancer constitutes an enormous disease burden in the United States, and because a significant portion of workers engage in shift work, further studies examining the relationship between light exposure at night and cancer risk through the melatonin pathway are warranted.
We are indebted to the participants of the Nurses’ Health Study II for their continuing outstanding dedication to the study. We would also like to express our deep gratitude to Richard Stevens for providing stimulus to assess night work in the Nurses' Health Study II.
1. Czeisler CA, Dijk DJ. Human circadian physiology and sleep-wake regulation. In: Takahashi JS, Turek FW, Moore RY, eds. Handbook of Behavioral Neurobiology: Circadian Clocks
. New York: Plenum Publishing Co; 2001:531–569.
2. Vijayalaxmi Thomas CRJ, Reiter RJ, et al. Melatonin: from basic research to cancer treatment clinics. J Clin Oncol
3. Schernhammer ES, Hankinson SE. Light at night: a novel risk factor for cancer in shift workers? Clin Occup Environ Med
4. Schernhammer ES, Laden F, Speizer FE, et al. Rotating night shifts and risk of breast cancer in women participating in the Nurses’ Health Study. J Natl Cancer Inst
5. Colditz GA, Manson JE, Hankinson SE. The Nurses’ Health Study: 20-year contribution to the understanding of health among women. J Womens Health
6. Rockhill B, Willett WC, Hunter DJ, et al. Physical activity and breast cancer risk in a cohort of young women. J Natl Cancer Inst
7. Rich-Edwards JW, Corsano KA, Stampfer MJ. Test of the National Death Index and equifax nationwide death search. Am J Epidemiol
8. Rothman KJ, Boice JD. Epidemiologic Analysis With a Programmable Calculator.
Washington, DC: Government Printing Office; 1979. NIH publication no. (PHS) 79–1649.
9. Pukkala E, Auvinen A, Wahlberg G. Incidence of cancer among Finnish airline cabin attendants. BMJ
10. Tynes T, Hannevik M, Andersen A, et al. Incidence of breast cancer in Norwegian female radio and telegraph operators. Cancer Causes Control
11. Rafnsson V, Tulinius H, Jonasson JG, et al. Risk of breast cancer in female flight attendants: a population-based study (Iceland). Cancer Causes Control
12. Hansen J. Increased breast cancer risk among women who work predominantly at night. Epidemiology
13. Davis S, Mirick DK, Stevens RG. Night shift work, light at night, and risk of breast cancer. J Natl Cancer Inst
14. Ballard T, Lagorio S, De Angelis G, et al. Cancer incidence and mortality among flight personnel: a meta-analysis. Aviat Space Environ Med
15. Kliukiene J, Tynes T, Andersen A. Follow-up of radio and telegraph operators with exposure to electromagnetic fields and risk of breast cancer. Eur J Cancer Prev
16. Colditz GA, Martin P, Stampfer MJ, et al. Validation of questionnaire information on risk factors and disease outcomes in a prospective cohort study of women. Am J Epidemiol
17. Laden F, Neas LM, Tolbert PE, et al. Electric blanket use and breast cancer in the Nurses’ Health Study. Am J Epidemiol
18. Al-Delaimy WK, Cho E, Chen WY, Colditz G, Willet WC. A prospective study of smoking and risk of breast cancer in young adult women. Cancer Epidemiol Biomarkers