Shift Work and Working at Night in Relation to Breast Cancer Incidence

In 2019, the International Agency for Research on Cancer classified night shift work as a probable human carcinogen (1). Most evidence supporting higher breast cancer risk among night shift workers comes from case–control studies; cohort studies have observed largely null associations (2).

The Sister Study is a prospective cohort of 50,884 women (2003–2009), ages 35–74 who had a sister with breast cancer but were breast cancer free themselves (3). Baseline occupational questionnaires asked about current job and past jobs held for ≥2 years. For each job, women were asked if they worked regular hours (starting/stopping at the same time every day) and if they worked at night (≥1 hour between 12:00–2:00 AM). If they reported not working regular hours, they were asked if they worked rotating shifts (number of shifts and usual start/stop times for each shift) or irregular hours. Rotating shift work (N = 3,183) included women who reported ≥1 job with rotating shifts; rotating night shift work (N = 2,275) was a subset of those women who also reported night work for the same job. Working irregular hours (N = 15,895) could include women who reported night work for ≥1 job. Any work at night (N = 13,992) included all women, regardless of schedule type, who met our definition of night work for ≥1 job. Self-reported incident breast cancer, confirmed with medical records (3), was classified as invasive breast cancer or ductal carcinoma in situ. Women were followed through September 15, 2017 (data release 7.2). The final sample included 48,451 women who completed the occupational questionnaire and were not missing covariate data.

Cox proportional hazards models were used to estimate HRs and 95% confidence intervals (CIs), with age as the timescale. For each analysis, women who reported never having that work schedule were the referent. Confounders were selected a priori and included race/ethnicity, education, marital status, and parity. We explored whether the association differed by time-varying menopausal status, timing of starting work relative to first childbirth, or time since stopping shift work.

Over follow-up (average = 9.1 years), 3,191 incident breast cancers were diagnosed. Study population characteristics are available elsewhere (3). Although ever and never rotating night shift workers were similar, ever workers were more likely to have at least some college (92% vs. 84%), be nulliparous (21% vs. 18%), or have ever smoked (47% vs. 43%).

We observed little association between breast cancer and ever working shifts, night shifts, at night, or irregular hours (Table 1). However, compared with women who never worked rotating night shifts or at night, women who reported working night shifts (adjusted HR = 1.30; 95% CI, 1.05–1.61) and who reported any work at night (adjusted HR = 1.12; 95% CI, 1.00–1.26) for ≤5 years (“short-term”) had a greater risk of breast cancer. Associations were not seen for longer durations.

The association between short-term work at night and breast cancer appeared stronger for premenopausal (adjusted HR = 1.29; 95% CI, 1.02–1.63) than for postmenopausal women (adjusted HR = 1.06; 95% CI, 0.93–1.22; Table 2). Similar associations were seen for shift and night shift work. These associations did not differ by age started working (before first childbirth) or time since stopping.

Night shift work, classified as a probable carcinogen, is hypothesized to act via light at night–related circadian disruption, by decreasing levels of melatonin (2, 4). Melatonin suppresses estrogen production, functions as an antioxidant, and has apoptotic and antiproliferative effects in breast cancer cells (4). Other potential mechanisms include oxidative stress and chronic inflammation resulting from disrupted circadian rhythm (2).

Some, but not all, case–control studies have suggested that increasing duration and frequency of night shift work may be associated with an increased breast cancer risk (2). Although the Nurses’ Health Study II observed an increased risk for long-term (20+ years) rotating night shift work (5), results from other cohort studies have been mostly null (2), consistent with our findings. The discrepancy between study types is possibly due to more comprehensive lifetime exposure data in the case–control studies, although retrospective data collection is a limitation. Despite our comprehensive lifetime exposure assessment, we observed that only short-term night shift work and work at night were associated with breast cancer risk. Although we had a wide age distribution and most women (65%) were still working at baseline, we cannot dismiss the possibility of left-truncation or healthy worker bias in our study. While these could be chance findings, one possibility is that some women who worked shifts or at night short-term may have switched to a daytime schedule sooner because of difficulties adjusting to a rotating or night schedule. There is some evidence that individuals who are “night owls” tend to tolerate shift work better than those who are “early birds” (6) and that early-bird women who work at night short-term have higher odds of breast cancer (7). We did not assess chronotype in our study.

We observed little evidence that either night or shift work was associated with breast cancer risk, except in those who participated in night work for short durations.

No potential conflicts of interest were disclosed.

Conception and design: D.P. Sandler, A.J. White

Development of methodology: A.J. White

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.P. Sandler

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.R. Sweeney, A.J. White

Writing, review, and/or revision of the manuscript: M.R. Sweeney, D.P. Sandler, N.M. Niehoff, A.J. White

Study supervision: D.P. Sandler, A.J. White

This research was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (Z01-ES044005 to D.P. Sandler and Z01-ES103332-01 to A.J. White) and a National Institute of Environmental Health Sciences training grant to the University of North Carolina (5T32ES007018-43 to M.R. Sweeney).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.