You are here

Article Text

Article menu

Download PDFPDF

Workplace
Original article
The impact of chronotype on melatonin levels among shift workers
  1. Parveen Bhatti,
  2. Dana K Mirick,
  3. Scott Davis
  1. Program in Epidemiology, Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington, USA
  1. Correspondence to Dr Parveen Bhatti, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, M4-B874, P.O. Box 19024, Seattle, WA 98109-1024, USA; pbhatti{at}fhcrc.org

Abstract

Objectives The association between shift work and cancer, which is thought to be mediated by effects on circulating melatonin levels, may be modified by chronotype (ie, the inherent preference for activity in the morning or the evening); however, few studies have examined the potential impact of chronotype on the carcinogenic effects of shift work. The authors analysed the impact of chronotype on previously reported differences in melatonin levels among healthcare workers that exclusively worked night or day shifts.

Methods The cross-sectional study included 664 men and women (310 day shift and 354 night shift workers) from which urine samples were collected throughout work and sleep periods and were assayed for 6-sulfatoxymelatonin. Participants also completed the Composite Scale of Morningness, a questionnaire used to assess chronotype.

Results Among both morning and evening-type night shift workers, 6-sulfatoxymelatonin levels were constitutively lower during daytime sleep, night-time sleep and night work compared with day shift workers during night-time sleep. However, morning-type night shift workers consistently showed 6-sulfatoxymelatonin levels that were closer to levels in day shift workers than did evening-type night shift workers. Differences in 6-sulfatoxymelatonin levels between morning-type and evening-type night shift workers relative to day shift workers were statistically significant in every instance (p<0.05).

Conclusions These results suggest that morning-type night shift workers may be better able to maintain a ‘normal’ circadian pattern of melatonin production as compared with evening-type night shift workers. The impact of this chronotype effect on cancer risk among shift workers requires further study.

https://doi.org/10.1136/oemed-2013-101730

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    What this paper adds

    • Chronotype may impact the carinogenic potential of shift work, but it has received limited attention in previous population-based studies.

    • We evaluated the impact of chronotype on the association between shift work and melatonin; melatonin disruption has been previously linked to carcinogenesis.

    • We found that morning-type night shift workers were better able to maintain normal patterns of melatonin secretion as compared with evening-type night shift workers, suggesting that morning types may be protected against the negative effects of shift work-related melatonin disruption.

    • Detailed evaluations of chronotype should be included in future studies of cancer risk as it may identify subgroups that are particularly sensitive to the carcinogenic effects of shift work.

    Introduction

    Since the International Agency for Research on Cancer's 2007 classification of shift work as a probable human carcinogen, epidemiological evidence for an association between shift work and cancer continues to be mixed.1–3 The mixed evidence has been attributed to a variety of factors including crude exposure assessment and the lack of consideration of individual characteristics that may impact adaptability to shift work schedules such as chronotype.4

    Chronotype or diurnal preference has been previously associated with tolerance to shift work. Individuals with a preference for activity during the evening (ie, evening types) have reported higher job satisfaction and better work performance than individuals with a preference for activity in the morning (ie, morning types).5 A recent study found that morning-type women who work the night shift had a higher risk of breast cancer than evening-type women when comparing subjects with high cumulative night shift work to those that never worked the night shift.6 While chronotype was assessed based only on a single item on a questionnaire in that study, it suggests that evening-type individuals may have a reduced susceptibility to the carcinogenic effects of night shift work. Thus, additional studies with more reliable assessments of chronotype could be of tremendous value to occupational disease prevention efforts.

    The potential carcinogenic effects of shift work are thought to be mediated through melatonin, which has been shown to have direct oncostatic properties and has been associated with decreased risks of breast cancer.7–9 Specifically, postsleep 6-sulfatoxymelatonin levels (6-sulfatoxymelatonin is an established urinary marker of circulating melatonin levels) were associated with decreased breast cancer risk. However, in a study that evaluated 24-hour urinary 6-sulfatoxymelatonin, no association was observed.10 Taken together, these previous studies seem to indicate that the decreased risks of cancer are not simply conferred by general increases in the levels of circulating melatonin but involve the correct timing of melatonin secretion (ie, maintaining high levels of circulating melatonin at night and low levels during the day).11

    Few studies have evaluated the potential effect of chronotype on melatonin levels in a population actually engaged in shift work. Previously, the authors reported significantly reduced urinary 6-sulfatoxymelatonin levels among exclusive night shift workers during night-time work, daytime sleep and night-time sleep periods on off nights, relative to exclusive day shift workers during night-time sleep in cross-sectional studies of female and male healthcare workers.12 ,13 To better understand the potential effect of chronotype on melatonin, an analysis of the impact of chronotype, assessed using the Composite Morningness Questionnaire14, was conducted on these previously reported differences in melatonin levels associated with shift work. In light of the previous report of lower breast cancer risk in evening-type shift workers compared with morning-type shift workers, it is hypothesised that evening-type individuals would have less disruption of their melatonin levels associated with night shift work as compared with morning-type individuals.

    Methods

    Study methods have been previously described in detail12 ,13 and are briefly summarised below.

    Study participants

    Participants were women aged 20–49 years (recruitment and data collection from November 2003 to August 2007) or men aged 20–55 years (recruitment and data collection from October 2007 to May 2011) employed as healthcare workers in the Seattle metropolitan area. Participants were also required to have a body mass index (BMI; weight (kg)/height (m)2) between 18 and 30 kg/m2 and could not be using hormones or supplements containing melatonin at least 30 days prior to enrolment. Among women, additional eligibility criteria included regular menstrual periods; no personal history of breast cancer, chemotherapy or tamoxifen therapy; no pregnancy or breast feeding within the past year; no use of supplements containing phytoestrogens or isoflavones and consumption of no more than five servings per week of soy-based foods. Male participants could not be using medications or supplements used to treat benign prostate conditions within 30 days of participation, could not have a personal history of prostate cancer or chemotherapy and could not have undergone general anaesthesia or major surgery at least 8 weeks prior to enrolment.

    Night shift workers were required to work at least 20 h per week exclusively during the graveyard shift and to sleep at night during off days. In addition, to be eligible for the study, night shift workers were required to stop work no earlier than 06:00 and work at least 8 h per shift to ensure that they were exposed to light during the late evening when melatonin levels typically rise, and that they were well into their shift around the time of typical peak melatonin secretion (01:00–02:00). Day shift workers were required to be employed at least 20 h per week and work exclusively during the day shift (ie, begin work no earlier than 06:00 and work at least 8 h per shift) and were chosen to have a similar age distribution as the night shift workers.

    Data collection

    After obtaining informed consent, a structured interview was administered to collect information about physical activity, employment history, current work and sleep schedules, reproductive and menstrual history (for female participants) and current medication use. The interviewer also assessed BMI. Subjects completed a Shift Work Questionnaire that contained a 13-item Composite Scale of Morningness. This scale was developed by Smith et al to address deficiencies observed in three widely used questionnaires to assess chronotype among shift workers.14 Though Smith et al originally validated the Composite Scale of Morningness among a group of undergraduate college students, the scale has subsequently been validated among shift workers and was shown to be stable over time even when subjects transition between day and night work schedules.15 The scale is used to assign a total chronoscore ranging from 13 to 55 points. Smith et al classified individuals with total scores of 22 and less as evening types, 23 to 43 as intermediate types and 44 and above as morning types.14 However, these cut-points are considered arbitrary as they are simply based on the 10th and 90th percentiles of the distribution of chronoscores in their student study population, which has a much different age distribution than the current study.15 As such, other cut-points were analysed including simple dichotomisation at the exact midpoint of the chronoscore scale (evening types 33 or less and morning types 34 or higher) and classification of individuals that are ±5 points from the exact midpoint of the scale as intermediate types such that evening types are those with scores of 28 or less, intermediate types are those with scores from 29 to 39 and morning types are those with scores of 40 or higher. Total sleep time and sleep efficiency (ratio of total sleep time to total time in bed) were also determined for participants during daytime sleep (night shift workers) and night-time sleep (day shift workers) using actigraphy, a method for providing electrophysiological measures of sleep in the home.16

    Urine collections were scheduled for days when at least two consecutive shifts were to be worked, followed by an off night (for the night shift workers) or at least one day shift worked, followed by a night of sleep (for the day shift workers). Just prior to each urine collection period, the participant was instructed to void his/her bladder and discard the urine; all subsequent urine excreted throughout either the work shift or the sleep period was collected, including the first void immediately following the end of the time period. Night shift workers collected all urine excreted during the daytime sleep period (following the first night shift) and the first void upon rising. During the second night shift, the participant collected all urine excreted during the shift and the first void immediately following the shift. During the following night's sleep (the ‘off’ night), the participant collected all urine excreted and the first void the next morning. Day shift workers collected all urine excreted during the day work shift and the first void immediately following the shift, as well as all urine excreted during the subsequent night of sleep and the first void the next morning. Day shift and night shift workers were asked to report alcohol consumption, tobacco use and medication use during study participation.

    Urinary 6-sulfatoxymelatonin assay

    Each sample was assayed for creatinine concentration based on a kinetic modification of the Jaffe reaction using Diagnostic Chemicals Ltd. reagents supplied by Roche Diagnostic Systems (Nutley, New Jersey) on a Roche Cobas Mira Plus chemistry analyser. Intra-assay and inter-assay coefficients of variation were 0.9–1.3% and 1.8–2.3%, respectively. Urinary concentrations of the primary metabolite of melatonin, 6-sulfatoxymelatonin, were determined with a radioimmunoassay kit (Stockgrand Ltd., Guildford, Surrey, UK). The assay was run in duplicate with low, medium, and high kit controls as well as an in-house control using a urine sample from a volunteer. Assay sensitivity was 0.5 ng/mL urine. Intra-assay and inter-assay coefficients of variation were 5.1–12.8% and 11.2–17.4%, respectively.

    Statistical methods

    Urinary 6-sulfatoxymelatonin values were approximately log-normally distributed. Log-transformed 6-sulfatoxymelatonin, normalised to creatinine concentration, was analysed as a continuous response variable. Linear regression models (SAS Proc REG, SAS Institute, Cary, NC) were employed to evaluate the impact of chronotype on differences in 6-sulfatoxymelatonin levels between the day and night shift workers. Because previous results indicated significant reductions in 6-sulfatoxymelatonin levels among the night shift workers relative to levels in day shift workers, the primary aim of this analysis was to evaluate whether such reductions among night shift workers varied between morning-type, intermediate-type and evening-type individuals. Urinary levels of 6-sulfatoxymelatonin during night-time sleep among the day shift workers were compared with levels in morning-type, intermediate-type and evening-type night shift workers at each of the following time points: daytime sleep following a night shift, night-time sleep on an off night and night-time work. In each regression model, an interaction term between night shift status and chronotype was included to determine whether there was a statistically significant difference in the effect of night shift work between the three chronotype categories. Additionally, there were two within-subject comparisons of urinary 6-sulfatoxymelatonin levels among the night shift workers: daytime versus night-time sleep and night work versus night-time sleep. These analyses employed SAS Proc MIXED (SAS Institute, Cary, NC) to fit linear regression models with correlated error structure and adjustment for time-dependent covariates. All models were adjusted for participant age, gender, hours of darkness (calculated for the Seattle area from US Naval Observatory data), BMI, number of alcoholic beverages consumed the previous 24 h and psychotherapeutic use in the previous 24 h. All statistical tests were two-sided and statistical significance was declared at p ≤ 0.05. Parameter estimates from the regression models were exponentiated to display results as percent increases or decreases in 6-sulfatoxymelatonin levels for the comparisons of interest. SEs and 95% CIs were constructed using the Delta Method.17

    Results

    Of the 914 eligible subjects identified, 869 agreed to participate. Forty-five eligible subjects declined to participate before the study commenced because of the necessary time commitment or could not be contacted by study staff. An additional 24 subjects were lost for these reasons during the study. One-hundred and thirty-nine subjects became ineligible during the study because of work schedule changes, use of melatonin supplements, irregular menstrual periods and/or pregnancy. After removing subjects with incomplete data collection, 664 participants remained for the primary analysis. Table 1 displays distributions of the primary covariates used in the analysis by shift status and gender. Male night shift workers tended to be younger (mean age=34.4 vs 36.5, night vs day shift) and reported consuming more alcohol (mean number of drinks=1.2 vs 0.6, night vs day shift) than male day shift workers. Female night shift workers tended to have a slightly higher BMI than female day shift workers (mean BMI=24.4 vs 23.6, night vs day shift). There were no other notable differences in covariates between day shift and night shift workers. In table 2, the distributions of male and female day and night shift workers by various chronotype classification schemes are provided. The Smith et al scheme resulted in very small numbers of evening-type individuals, so it was not included in any further analyses. Using the dichotomous measure of chronotype, morning types were generally more prevalent overall; a higher proportion of day shift workers were morning types, whereas night shift workers were more evenly divided between morning and evening types. Results of analyses of the impact of chronotype (dichotomised and three-level classification schemes) on melatonin differences between night shift workers and day shift workers, for men and women combined, are provided in table 3. For dichotomised chronotype, evening-type night shift workers had 53% lower 6-sulfatoxymelatonin levels during daytime sleep relative to day shift workers during night-time sleep, whereas morning-type night shift workers had 65% lower 6-sulfatoxymelatonin levels relative to day shift workers. When comparing night shift workers during night-time sleep on their off nights to day shift workers during night-time sleep, evening-type participants had 49% lower 6-sulfatoxymelatonin levels, and morning-type participants had 30% lower 6-sulfatoxymelatonin levels. During night work, evening-type night shift workers had 73% lower 6-sulfatoxymelatonin levels compared with day shift workers during night-time sleep, but morning-type night shift workers had a reduction of 54%. These differences in the effect of chronotype were statistically significant in every instance (p-interaction <0.01). Results for morning and evening types were similar for the three-level chronotype analyses (table 3).

    View this table:
    Table 1

    Descriptive characteristics of Shift Worker Study participants according to work shift status and gender

    View this table:
    Table 2

    Chronotype measures of Shift Worker Study participants according to work shift status and gender

    View this table:
    Table 3

    Results from regression analyses of melatonin levels by chronotype, night shift workers (NSW) relative to all day shift workers (DSW)

    Melatonin differences for intermediate types were consistently between those of morning types and evening types regardless of the comparison under consideration, although the difference between intermediate and evening types was only statistically significant when comparing night work melatonin levels among night shift workers to night sleep melatonin levels among day shift workers.

    As shown in table 4 for the dichotomised chronotype analysis, within the night shift workers, daytime sleep levels of 6-sulfatoxymelatonin among evening-type participants were 21% lower than levels during night-time sleep on an off night, whereas daytime sleep levels among morning-type participants were further reduced (41%) relative to night-time sleep levels. When comparing 6-sulfatoxymelatonin during night work to night-time sleep, evening types had a significantly greater reduction (52%) in levels of 6-sulfatoxymelatonin than morning types (33%). These differences by chronotype were statistically significant (p-interaction <0.01). Results for the morning and evening types in the three-level chronotype analyses were similar (table 4). Once again, melatonin differences for intermediate types fell between those of morning and evening types.

    View this table:
    Table 4

    Results from regression analyses of melatonin levels by chronotype within the night shift workers (NSW)

    Gender stratified analyses (results not shown) did not produce materially different results except when examining effects during daytime sleep. When comparing daytime sleep levels among night shift workers to night-time sleep levels among day shift workers, female evening types had 62% (95% CI −71% to −52%) lower levels of 6-sulfatoxymelatonin while male evening types had 45% (95% CI −57% to −31%) lower levels (female and male morning types were 64 and 65% lower, respectively). When comparing daytime and night-time sleep levels within night shift workers, female evening types had 36% (95% CI −54% to −18%) lower levels of 6-sulfatoxymelatonin, but males had only 5% (95% CI −3.0% to 20%) lower levels (female and male morning types were 39 and 43% lower, respectively).

    Discussion

    The results demonstrate that morning-type, intermediate-type and evening-type participants exclusively engaged in night shift work suffer from constitutively lower melatonin levels. However, as compared with evening-type night shift workers, morning-type night shift workers were able to maintain melatonin at levels that were more comparable to day shift workers at similar times of the day. Though not always statistically significant, intermediate types, as compared with evening types, also better maintained melatonin levels that were comparable to day shift workers at similar times of the day. For instance, during night work, which occurs at a time of day that is typically associated with sleep and high melatonin levels, morning and intermediate types suffered less of a disruption in melatonin levels as compared with evening types. In the case of daytime sleep, which occurs at a time of day that is typically associated with wakefulness and low melatonin levels, morning and intermediate types better suppressed their melatonin levels as compared with evening types. Gender-stratified analyses revealed that the reduced suppression of melatonin among evening types during daytime sleep was largely restricted to men.

    A previous study of light at night exposure and melatonin among 123 rotating shift nurses did not observe an association between chronotype and melatonin levels18; however, few subjects were classified as morning and evening types under the classification system that was used. Furthermore, chronotype was not assessed as a potential modifier of the null shift work–melatonin association that was observed. The lack of an association may be attributed to the rapidly rotating shift schedule (two 12-h days, two 12-h nights and 5 days off), which the authors speculated may have been insufficient to disrupt melatonin secretion.

    Among the female shift workers, we previously reported a significant effect of Asian race on melatonin differences, whereby Asian night shift workers were able to maintain melatonin levels that were closer to their day shift counterparts than white night shift workers during day sleep, night work and night sleep.19 Among female night shift workers, we observed no significant difference in chronoscore by race (mean (SD) chronoscore for white night shift workers=32(9); mean (SD) chronoscore for Asian night shift workers=32(8)). In addition, including a variable for race in our regression models did not have a substantial impact on our point estimates of interest (results not shown). This suggests that race and chronotype are having independent effects on shift-related differences in melatonin levels.

    The differences in melatonin levels between night shift and day shift workers observed in this study have been associated with tumour growth. Blask et al demonstrated that 40% decreases in melatonin among premenopausal women when exposed to 90 min of bright white light at night were associated with significant increases in tumour growth and activity in breast cancer xenografts.20 In conjunction with these findings, results suggest that relative to evening types, morning-type night shift workers may be somewhat protected against the carcinogenic effects of melatonin suppression.

    Very limited research has been completed on the impact of chronotype on cancer risk among shift workers, but based on a crude analysis of chronotype and breast cancer risk among night shift workers that found evening types to have a lower shift work-related risk of breast cancer than morning types,6 one might expect that evening types would be able to better maintain normal melatonin levels. If evening types do indeed have a lower risk of breast cancer associated with shift work as compared with morning types, then the results might suggest that pathways other than those related to melatonin are of greater importance in conferring some degree of protection to evening types against the carcinogenic effects of shift work. One potential pathway is lifestyle disturbances leading to poor diet, lack of exercise, increased alcohol consumption and increased tobacco use.21 There was no evidence for a difference in BMI between evening type (mean BMI=25.1 kg/m2) and morning type (mean BMI=25.0 kg/m2) night shift workers in this study; however, BMI may not be a good indicator since those with BMI >30 kg/m2 were not eligible to participate. There was also no significant difference between evening-type and morning-type night shift workers in the mean number of alcoholic beverages consumed in the 24 h prior to completing their night-time sleep period (1.1 vs 0.8). The limited tobacco use among study participants precluded evaluation of differences by chronotype. Another potential pathway is sleep disruption leading to stress responses and immune suppression.21 There was a marginal difference in total sleep time between evening-type (466 min) and morning-type (456 min) night shift workers. No difference in mean sleep efficiency was observed (79 vs 81% for evening-type and morning-type night shift workers, respectively).

    A previous laboratory-based study that collected hourly blood samples for the assessment of melatonin during sleep observed that chronotype was strongly related to melatonin acrophase (ie, timing of peak melatonin secretion), but not amplitude.22 Data in the current study did not allow for assessment of acrophase among study subjects. A protocol that included urine samples collected over shorter time intervals would have potentially allowed for the assessment of acrophase; however, the disruption to participants’ schedules would have likely impacted melatonin measurements (eg, interruption during sleep and light-at-night exposure to collect more frequent samples). Another limitation of this study is that it did not evaluate health effects, specifically cancer, in association with the differing reductions in 6-sulfatoxymelatonin levels. Given the large amount of time that it would take to accrue sufficient cancer cases in a prospective study of melatonin and cancer risk among shift workers, in the interim, studies such as this, involving the evaluation of biomarkers of circadian disruption and potentially cancer risk, can be useful. Despite these limitations, this is the first study to the authors’ knowledge to evaluate the impact of chronotype on differences in melatonin levels between fixed day shift and night shift workers. Though the study population consisted of healthcare workers, results should be mostly generalisable to other occupational settings. However, since the study focused on fixed day and night shift workers that were employed at least 20 h per week, results may not be applicable to rotating shift workers or part-time employees working less than two shifts per week. In addition, as a cross-sectional study, there is a possibility that the night shift workers that participated represented a self-selected group that was able to better tolerate the negative effects of working the night shift. Thus, a longitudinal study including new shift workers before they have a chance to drop out may observe even larger differences than those reported here. Detailed data collection including multiple urinary measures of 6-sulfatoxymelatonin at multiple critical time points for each study subject and the rigorous assessment of chronotype are the major strengths of this study.

    The results of this study provide evidence that, with respect to disrupted melatonin levels, morning types may be better protected from the negative effects of shift work relative to evening types. The impact of chronotype on cancer risk needs to be examined more extensively in future studies of shift workers, with detailed lifestyle and sleep quality data to investigate the contribution of other pathways underlying the association between shift work and cancer. Examination of other biomarkers in addition to 6-sulfatoxymelatonin and consideration of gender-specific effects would also be useful in this endeavour.

    References

    1. ↵
      IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Painting, firefighting, and shiftwork. IARC Monogr Eval Carcinog Risks Hum 2010;98:563–764.
      OpenUrl
    2. ↵
      1. Wang X-S,
      2. Armstrong MEG,
      3. Cairns BJ,
      4. et al
      . Shift work and chronic disease: the epidemiological evidence. Occup Med (Lond) 2011;61:78–89.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Ijaz SI,
      2. Verbeek J,
      3. Seidler A,
      4. et al
      . Night-shift work and breast cancer—a systematic review and meta-analysis. Scand J Work Environ Health 2013;39:431–47.
      OpenUrlCrossRefPubMed
    4. ↵
      1. Stevens RG,
      2. Hansen J,
      3. Costa G,
      4. et al
      . Considerations of circadian impact for defining “shift work" in cancer studies: IARC Working Group Report. Occup Environ Med 2011;68:154–62.
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Saksvik IB,
      2. Bjorvatn B,
      3. Hetland H,
      4. et al
      . Individual differences in tolerance to shift work--a systematic review. Sleep Med Rev 2011;15:221–35.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Hansen J,
      2. Lassen CF
      . Nested case-control study of night shift work and breast cancer risk among women in the Danish military. Occup Environ Med 2012;69:551–6.
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Schernhammer ES,
      2. Berrino F,
      3. Krogh V,
      4. et al
      . Urinary 6-sulfatoxymelatonin levels and risk of breast cancer in postmenopausal women. J Natl Cancer Inst 2008;100:898–905.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Schernhammer ES,
      2. Hankinson SE
      . Urinary melatonin levels and breast cancer risk. J Natl Cancer Inst 2005;97:1084–7.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Schernhammer ES,
      2. Hankinson SE
      . Urinary melatonin levels and postmenopausal breast cancer risk in the Nurses’ Health Study cohort. Cancer Epidemiol Biomarkers Prev 2009;18:74–9.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Travis RC,
      2. Allen DS,
      3. Fentiman IS,
      4. et al
      . Melatonin and breast cancer: a prospective study. J Natl Cancer Inst 2004;96:475–82.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Hrushesky WJM,
      2. Blask DE
      . Re: Melatonin and breast cancer: a prospective study. J Natl Cancer Inst 2004;96:888–9.
      OpenUrlFREE Full Text
    12. ↵
      1. Davis S,
      2. Mirick DK,
      3. Chen C,
      4. et al
      . Night Shift Work and Hormone Levels in Women. Cancer Epidemiol Biomarkers Prev 2012;21:609–18.
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Mirick DK,
      2. Bhatti P,
      3. Chen C,
      4. et al
      . Night Shift Work and Levels of 6-Sulfatoxymelatonin and Cortisol in Men. Cancer Epidemiol Biomarkers Prev 2013;22:1079–87.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Smith CS,
      2. Reilly C,
      3. Midkiff K
      . Evaluation of three circadian rhythm questionnaires with suggestions for an improved measure of morningness. J Appl Psychol 1989;74:728–38.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Greenwood KM
      . Long-term stability and psychometric properties of the Composite Scale of Morningness. Ergonomics 1994;37:377–83.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Cole RJ,
      2. Kripke DF,
      3. Gruen W,
      4. et al
      . Automatic sleep/wake identification from wrist activity. Sleep 1992;15:461–9.
      OpenUrlPubMedWeb of Science
    17. ↵
      1. Oehlert GW
      . A Note on the Delta Method. Am Statistician 1992;46:27–9.
      OpenUrlCrossRefWeb of Science
    18. ↵
      1. Grundy A,
      2. Tranmer J,
      3. Richardson H,
      4. et al
      . The influence of light at night exposure on melatonin levels among Canadian rotating shift nurses. Cancer Epidemiol Biomarkers Prev 2011;20:2404–12.
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Bhatti P,
      2. Mirick DK,
      3. Davis S
      . Racial differences in the association between night shift work and melatonin levels among women. Am J Epidemiol 2013;177:388–93.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Blask DE,
      2. Brainard GC,
      3. Dauchy RT,
      4. et al
      . Melatonin-depleted blood from premenopausal women exposed to light at night stimulates growth of human breast cancer xenografts in nude rats. Cancer Res 2005;65:11174–84.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Fritschi L,
      2. Glass DC,
      3. Heyworth JS,
      4. et al
      . Hypotheses for mechanisms linking shiftwork and cancer. Med Hypotheses 2011;77:430–6.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Gibertini M,
      2. Graham C,
      3. Cook MR
      . Self-report of circadian type reflects the phase of the melatonin rhythm. Biol Psychol 1999;50:19–33.
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Contributors All authors participated substantially in the analysis and interpretation of data and the drafting and revision of the manuscript. All authors approved the manuscript for publication.

    • Funding The study was funded by grants R01 CA097090 and R01 CA116400 from the US National Cancer Institute and an award from the Safeway Foundation.

    • Competing interests None.

    • Ethics approval Institutional Review Board of the Fred Hutchinson Cancer Research Center.

    • Provenance and peer review Not commissioned; externally peer reviewed.