Shift Work and Prostate Cancer Incidence in Industrial Workers
A Historical Cohort Study in a German Chemical Company
; ; ; ;
Background: There is inconsistent evidence for a possible carcinogenic effect of shift work. In particular, little is known about the putative association of shift work with prostate cancer.
Method: We studied a cohort of 27 828 male industrial production workers residing in the German federal state of Rhineland-Palatinate who worked for at least one year in a chemical company in the period 1995–2005. We obtained data on shift work and potential confounders including age, occupational task, and duration of employment from personnel files and from the records of the occupational health service. New cases of cancer in the period 2000–2009 were ascertained from the state cancer registry. Differences in risk between shift workers and daytime workers were analyzed with Cox regression, stratified by stage of cancer, and adjusted for potential confounding effects.
Results: There were 146 new cases of prostate cancer in 12 609 rotating shift workers and 191 in 15 219 daytime workers. The median year of birth was 1960 in the first group and 1959 in the second. The shift workers did not have an elevated hazard ratio for prostate cancer in comparison to the daytime workers (HR = 0.93, 95% confidence interval [CI] 0.73–1.18). Some differences were seen depending on tumor stage. Both groups of workers had a higher incidence of prostate carcinoma than the general population (standardized incidence rate [SIR] = 1.44, 95% CI 1.22–1.70 for daytime workers; SIR = 1.51, 95% CI 1.30–1.74 for shift workers).
Conclusion: In this well-documented, large-scale cohort study, the incidence of prostate cancer among shift workers did not differ from that among daytime workers. In the authors’ opinion, further follow-up of this relatively young cohort is required.
In 2007 the International Agency for Research on Cancer (IARC) classified shift work involving circadian disruption as a probable human carcinogen (1). Most of the evidence for this is based on animal experiment models and only a few reliable epidemiological studies. The cancer best described to date is breast cancer; this has mainly been investigated in female nurses and flight attendants (2, 3).
Shift work can also increase the incidence of prostate cancer, as plausibly described in a review article published in 2012 (4). Mechanistically, it is possible that nocturnal light exposure and disrupted circadian rhythms cause cancer, or that tumor growth is favored by hormonal mediation. Nocturnal light exposure suppresses melatonin secretion in the pineal gland. In vitro and in vivo experiments have shown that melatonin prevents or inhibits tumor growth, particularly that of prostate cancer, and that there is a dose–effect relationship (1, 5).
The epidemiological evidence on prostate cancer comes from only a few studies involving individual data on shift work (6–8) (Table 1). In studies of flight personnel, duration of employment or number of flight hours was used as a proxy measure (9–11). In one large Scandinavian study, the probability of exposure was estimated using a job exposure matrix (12).
This article investigates the incidence of prostate cancer in this cohort and examines earlier findings on overall tumor incidence in more detail (15). The complete scientific report is available online (16).
Study design, study population
This study is based on a cohort already described in detail by Oberlinner et al. (17) of all male production workers who were employed for at least one year at BASF SE, Ludwigshafen between 1995 and 2005. Information on shift work and potential confounding factors such as age, smoking, professional status, and duration of employment were recorded using personnel files and medical records. BASF SE knows whether these individuals are now alive or deceased because former employees receive a company pension.
For this study the cohort was limited to the 12 609 shift workers and 15 219 day workers resident in Rhineland-Palatinate. Incident cancer cases between 2000 and 2009 were identified using the Rhineland-Palatinate Cancer Registry by matching pseudonymized data (18–20). The registry's degree of completeness since 2000 is at over 90% (21). Cancer Registry data specialists checked all potential matches and classified them as good, medium, or poor.
In order to investigate the extent to which health-related behavior depends on shift status, evaluation was broken down by tumor stage T (22). T stage during the observation period is known for 70 to 75% of all registered prostate cancers. Where it is not known, this is usually when a biopsy was performed due to an abnormal prostate-specific antigen (PSA) test at stage T1.
In line with the data protection regulations for comparing cohort studies, which have been approved by the Rhineland-Palatinate State Official for Data Protection (www.datenschutz.rlp.de), pseudonymized epidemiological information on exposure and health was sent to the Institute for Medical Biostatistics, Epidemiology and Informatics (IMBEI, Institut für Medizinische Biometrie, Epidemiologie und Informatik) at Mainz University Medical Center for evaluation. All dates and ages were rounded off to the nearest year before the data was sent.
Two different 12-hour shift patterns are authorized at the Ludwigshafen location. The older pattern (3 × 12 hours) was used for fifty years; the 4 × 12 hours patterns was introduced in 1992 and is now used almost exclusively. In the 3 × 12 hours pattern, a 12-hour day shift from 6 a.m. to 6 p.m. is followed by 24 hours off, a 12-hour night shift, and another 24 hours off. In the 4 × 12 hours pattern, the second free period lasts 48 hours. Both patterns are forward-rotating and have the same annual working time—1709 hours—and the same number of night shifts. This is achieved using different numbers of free shifts and leave days.
Whether individuals worked shifts or daytime hours on the start date of January 1, 1995, or on inclusion in the cohort, was known for all employees. This status was taken to be constant in this study. Switches between daytime and shift work are rare. An initial validation study investigated the resulting misclassification (15).
For the period from 1995 onwards, in a sample of 921 production employees, the status of four of 324 daytime workers changed to shift work, and that of 24 of 597 shift workers to daytime work. There was no change for 893 (97%) of the employees (14). Information on shift status before 1995 is available only on paper.
The authors investigated a sample of 300 employees, 200 of whom worked shifts and 100 worked daytime hours on January 1, 1995. Their shift status was recorded retrospectively for their whole working lives. The mean duration of misclassified shift status was estimated at 3.3% for daytime workers and 15.6% for shift workers. Both these validation studies suggest that misclassification has not substantially biased the study findings.
The number of cancer cases observed and the number expected on the basis of incidence rates in the population of Rhineland-Palatinate were compared using standardized incidence rates (SIRs) and 95% confidence intervals (95% CIs).
Differences between cancer incidence among shift workers and daytime workers were investigated using Cox regression and reported as hazard ratios (HRs) with 95% confidence intervals. Adjustment was performed for the potential confounding factors of age and professional status. Lifestyle factors associated with high social status are known risk factors for prostate cancer. As used in earlier publications, professional status—categorized into skilled work/supervision and manual labor—was used as a proxy for social status. Smoking is not a known risk factor for prostate cancer and was included only in sensitivity analysis (23).
A dose–effect relationship between duration of employment in shift work and prostate cancer was investigated in two ways:
- Time-dependent dichotomous indicator variables: 20 years of employment or less versus more than 20 years of employment
- Further subdivision: less than 30 years, 30 to <35 years, 35 to <40 years, 40 to <45 years, 45 to <50 years, and 50 years or more of employment
To investigate a potential latency in the development of prostate cancer, a latency period of 20 years was used for sensitivity analyses; these were based on the study findings on breast cancer. All evaluations were performed using SAS 9.4 (24).
The cohort included 27 828 employees. Two individuals with the same recorded beginning and end of employment dates were excluded. During the 10-year study period (270 068 person-years), 555 and 518 cases of cancer were observed in daytime and shift workers respectively, of which 191 and 146 respectively were prostate cancers (Table 2).
Data was successfully matched with that of the Rhineland-Palatinate Cancer Registry; this yielded only six uncertain cases, of which three were among the 337 cases of prostate cancer.
The majority of prostate cancer cases (n = 322; 95.5%) were observed in employees who had been employed for 20 years or longer. Stages T2 (n = 216; 64.1%) and T3 (n = 71; 21.1%) were the most frequently diagnosed. In the chi-square test there was no difference between the T stage distribution for daytime and shift workers in either manual (p = 0.60) or other (p = 0.21) workers (p = 0.34).
Both daytime and shift workers had a higher incidence of prostate cancer than the general population (SIR = 1.44, 95% CI: 1.22 to 1.70; SIR = 1.51, 95% CI: 1.30 to 1.74). In both groups, the SIR for stage TI prostate cancer or prostate cancer with no recorded stage was less than 1; for all other stages it was more than 1. The highest values recorded were for stage T2 (SIR = 2.05; 95% CI: 1.71 to 2.44 for shift workers; SIR = 1.73; 95% CI: 1.38 to 2.14 for daytime workers) (Table 3). The 20-year latency period used in analysis had little effect on the findings, as prostate cancer was diagnosed during the first 20 years of employment in only 15 employees.
Shift workers’ risk of prostate cancer was no higher than that of daytime workers (HR = 0.93; 95% CI: 0.73 to 1.18) (Table 4). There was a non-significant increase in the risk of a diagnosis of stage TI or T4 prostate cancer (HR = 1.26; 95% CI: 0.44 to 3.86 versus HR = 1.36; 95% CI: 0.25 to 6.18) and a decreased risk for stages T2 and T3 (HR = 0.84; 95% CI: 0.62 to 1.15 versus HR = 0.90; 95% CI: 0.53 to 1.52). Adjustment for smoking had little effect on the findings (HR = 0.88; 95% CI: 0.69 to 1.13). The 20-year latency period used in analysis had only a weak effect on the overall HR (HR = 0.89; 95% CI: 0.69 to 1.14) and stage-specific risk estimators.
In dose–effect analysis, the dichotomous indicator for duration of employment (less than 20 years versus 20 years or more) did not qualitatively change the findings. However, the risk of prostate cancer in those with longer employment histories was significantly higher than in those with shorter employment histories. Evaluation using further subdivision for duration of employment (Figure) indicates that the risk of stage T2 prostate cancer increases with duration of employment.
This research covered 337 cases in 27 828 employees and is therefore the largest cohort study on the risk of prostate cancer in shift and daytime workers to date. Tumor stage-specific comparisons were being presented for the first time. The cohort was put together in 1995 on the basis of personnel files and medical records (17) and since then has been followed actively. The health of the study participants was already well documented (13, 14).
While the literature contains evidence that shift work that involves circadian disruption increases the risk of breast cancer, and some studies have reported an increased risk of prostate cancer, this study found no difference between shift and daytime workers (HR = 0.93; 95% CI: 0.73 to 1.18) (6–8, 25).
In this cohort there was a non-significantly increased risk of stage T1 and T4 prostate cancer in shift workers versus daytime workers, and a non-significantly reduced risk for stages T2 and T3. This pattern remained after additional adjustment for the confounding factors of smoking or duration of employment and when a latency period was assumed. Although the number of early stage T1 cancers was low and the SIR for shift and daytime workers was less than 1, this pattern indicated a group of shift workers who lived more health-consciously than their colleagues who worked during the day. The number of prostate cancers with a recorded T stage was 90.8% in this cohort and thus higher than the mean for Rhineland-Palatinate. Similarly, SIRs for prostate cancer of 1.44 and 1.56 in shift and daytime workers respectively versus the general population have already been reported (15). Both indicate detection bias as a result of the good medical care offered equally to all BASF employees. Prostate cancer screening is not actively offered to employees but can be performed on request (26). While increased SIRs for some cancers, such as cancer of the bladder, were due to occupational exposure in this cohort, no factors that might explain an increased risk of prostate cancer were recorded (27).
In addition to the size of the cohort, the strengths of this study included its documentation, based on personnel files and medical records. This means that recall bias can be ruled out. One feature of the cohort is the balanced numbers of daytime and shift workers from the same parts of the company and with the same working conditions, making them comparable in terms of risk profile, age distribution, and socioeconomic status. This maximized the power of internal comparisons and at the same time minimized the risk of bias caused by other work-related factors. In order to improve comparability further still, figures were adjusted for duration of employment, a proxy for work-related health risks resulting from chemical exposure, for example. Cancer cases were identified by matching data with the Rhineland-Palatinate Cancer Registry, which is complete and contains a high proportion of histologically confirmed cases. The tumor stage T was known in more than 90% of cases. This made stage-specific evaluation possible, so employees’ health-related behavior and the effect of company medical prevention measures could be taken into account (28).
Data matching with the Rhineland-Palatinate Cancer Registry meant that only employees in the area covered by the registry could be included, and that cancer incidence could only be observed from the year 2000 onwards. A further weakness of the study results from shift status misclassification. As described in the Methods section, validation studies assessed shift status misclassification as low, so it could not have had a major effect on the study findings. More recently, a cohort of trainees with no adverse occupational factors has been put together, and the distribution of future shift and daytime workers has been investigated with regard to smoking and other cardiovascular risk factors (29).
In addition to the obvious sleep rhythm disruptions and associated effects on melatonin levels, shift work may have further effects on the risk of prostate cancer. It is suspected that vitamin D, cutaneous production of which requires ultraviolet radiation, inhibits prostate cancer (23). In addition, nutrition, particularly antioxidant intake, affects carcinogenesis and differs between daytime and shift workers (30, 31).
A working group of the International Agency for Research on Cancer (IARC) has debated the idea that duration of employment is a crude proxy for the cumulative effects of chronodisruption (32) and that shift work could be seen as a deviation from the natural day. Differences in chronoregulating genes may also play a role in carcinogenesis. Other authors argue that the individual chronotype or “social jetlag” is a more suitable measure of risk (3, 33). An initial cross-sectional study on this subject has recently been published (34).
The shift workers in this cohort worked according to two very similar forward-rotating shift patterns. The findings of this study cannot therefore simply be extrapolated to the variety of different shift patterns followed in Europe (32, 35).
Recording the individual chronotype is an approach for further investigation of this cohort. In the authors’ opinion, further follow-up of incidence rates in this still relatively young study population is important. Comparison of these findings with other shift patterns regarding prostate and other cancers may also be informative.
The authors would like to thank the scientific advisory board established by the DGUV for its many useful discussions and comments.
Ethics Committee vote
This study received a positive vote from the Ethics Committee of the Rhineland-Palatinate State Medical Association.
This study was funded by German Social Accident Insurance
(DGUV, Deutsche Gesetzliche Unfallversicherung), funding code FP-0332.
Conflict of interest statement
Dr. Yong is employed in the Medical Department of BASF SE.
Dr. Nasterlack was previously employed in the Medical Department of BASF SE. He holds shares in BASF SE.
Prof. Blettner has received fees for statistical processing of another study by BASF SE.
Dr. Hammer and Dr. Emrich declare that no conflict of interest exists.
Manuscript received on 5 November 2014, revised version accepted on
7 April 2015.
Translated from the original German by Caroline Shimakawa-Devitt, M.A.
Dr. Gaël P. Hammer
Morphologic Tumour Registry
Laboratoire National de Santé E.P.
1, rue Louis Rech
Health 2015; DOI: 10.5271/sjweh.3487. Epub ahead of print CrossRef
Prof. Dr. rer. nat. Blettner
Registre Morphologique des Tumeurs, Laboratoire National de Santé,
Luxembourg: Dr. P. H. Hammer
Cancer Registry Rhineland-Palatinate, Institute of Medical Biostatistics, Epidemiology and Informatics (IMBEI) at the University Medical Center of the Johannes Gutenberg University Mainz : Dr. rer. nat. Emrich
Occupational Medicine and Health Protection, BASF SE, Ludwigshafen:
Dr. med. Nasterlack, Dr. med. Yong
|1.||Erren TC, Morfeld P, Stork J, et al.: Shift work, chronodisruption and cancer?—The IARC 2007 challenge for research and prevention and 10 theses from the Cologne Colloquium 2008. Scand J Work Environ Health 2009; 35: 74–9 CrossRef MEDLINE|
|2.||Straif K, Baan R, Grosse Y, et al.: Carcinogenicity of shift-work, painting, and fire-fighting. Lancet Oncol 2007; 8: 1065–6 CrossRef|
|3.||Erren TC, Falaturi P, Morfeld P, et al.: Shift work and cancer: the evidence and the challenge. Dtsch Arztebl Int 2010; 107: 657–62 VOLLTEXT|
|4.||Sigurdardottir LG, Valdimarsdottir UA, Fall K, et al.: Circadian disruption, sleep loss, and prostate cancer risk: a systematic review of epidemiologic studies. Cancer Epidemiol Biomarkers Prev 2012; 21: 1002–11 CrossRef MEDLINE PubMed Central|
|5.||Reiter RJ, Tan DX, Erren TC, et al.: Light-mediated perturbations of circadian timing and cancer risk: a mechanistic analysis. Integr Cancer Ther 2009; 8: 354–60 CrossRef MEDLINE|
|6.||Conlon M, Lightfoot N, Kreiger N: Rotating shift work and risk of prostate cancer. Epidemiology 2007; 18: 182–3 CrossRef MEDLINE|
|7.||Kubo T, Ozasa K, Mikami K, et al.: Prospective cohort study of the risk of prostate cancer among rotating-shift workers: findings from the Japan collaborative cohort study. Am J Epidemiol 2006; 164: 549–55 CrossRef MEDLINE|
|8.||Kubo T, Oyama I, Nakamura T, et al.: Industry-based retrospective cohort study of the risk of prostate cancer among rotating-shift workers. Int J Urol 2011; 18: 206–11 CrossRef MEDLINE|
|9.||Hammer GP, Auvinen A, De Stavola B, et al.: Mortality from cancer and other causes in commercial airline crew: a joint analysis of cohorts from ten countries. Occup Environ Med 2014; 71: 313–22 CrossRef MEDLINE|
|10.||Pukkala E, Aspholm R, Auvinen A, et al.: Incidence of cancer among Nordic airline pilots over five decades: occupational cohort study. BMJ 2002; 325: 567 CrossRef MEDLINE PubMed Central|
|11.||Pukkala E, Helminen M, Haldorsen T, et al.: Cancer incidence among Nordic airline cabin crew. Int J Cancer 2012; 131: 2886–97 CrossRef MEDLINE|
|12.||Pukkala E, Martinsen JI, Lynge E, et al.: Occupation and cancer—follow-up of 15 million people in five Nordic countries. Acta Oncol 2009; 48: 646–790 CrossRef MEDLINE|
|13.||Yong M, Nasterlack M, Germann C, et al.: Shift work and risk of non-cancer mortality in a cohort of German male chemical workers. Int Arch Occup Environ Health 2014; 87: 763–73 CrossRef CrossRef|
|14.||Yong M, Nasterlack M, Messerer P, et al.: A retrospective cohort study of shift work and risk of cancer-specific mortality in German male chemical workers. Int Arch Occup Environ Health 2014; 87: 175–83 CrossRef CrossRef MEDLINE|
|15.||Yong M, Blettner M, Emrich K, et al.: A retrospective cohort study of shift work and risk of incident cancer in German male chemical workers. Scand J Work Occup Health 2014; 40: 502–10 CrossRef MEDLINE|
|16.||Hammer GP, Yong M, Blettner M: Kohortenstudie zur Untersuchung der Krebsinzidenz und Mortalität bei Schichtarbeitern in der BASF. Report no. FF-FP0332. Sankt Augustin: Deutsche Gesetzliche Unfallversicherung e. V. (DGUV) 2014.|
|17.||Oberlinner C, Ott MG, Nasterlack M, et al.: Medical program for shift workers—impacts on chronic disease and mortality outcomes. Scand J Work Environ Health 2009; 35: 309–18 CrossRef MEDLINE|
|18.||Brenner H, Schmidtmann I: Effects of record linkage errors on disease registration. Methods Inf Med 1998; 37: 69–74 MEDLINE|
|19.||Michaelis J, Miller M, Pommerening K, et al.: A new concept to ensure data privacy and data security in cancer registries. Medinfo 1995; 8 Pt 1: 661–5 MEDLINE|
|20.||Pommerening K, Miller M, Schmidtmann I, et al.: Pseudonyms for cancer registries. Methods Inf Med 1996; 35: 112–21 MEDLINE|
|21.||Emrich K, Ressing M, Zeißig S, Seebauer G, Blettner M: Krebs in Rheinland-Pfalz 2011. Mainz: Krebsregister Rheinland-Pfalz, Registerstelle 2014.|
|22.||International Union Against Cancer: The TNM classification of malignant tumours. 7th edition. Chichester, West Sussex: John Wiley & Sons 2009.|
|23.||Platz EA, Giovannucci E. Prostate cancer. In: Schottenfeld D, Fraumeni JFJ (eds.): Cancer epidemiology and prevention. 3rd edition. New York: Oxford University Press 2006; 1128–50 CrossRef|
|24.||SAS 9.4. Cary, SAS Institute Inc. NC. USA: 2014.|
|25.||Wang F, Yeung KL, Chan WC, et al.: A meta-analysis on dose-response relationship between night shift work and the risk of breast cancer. Ann Oncol 2013; 24: 2724–32 CrossRef MEDLINE|
|26.||Ott MG, Oberlinner C, Lang S, et al.: Health and safety protection for chemical industry employees in a rotating shift system: program design and acute injury and illness experience at work. J Occup Environ Med 2009; 51: 221–31 CrossRef MEDLINE|
|27.|| Parent ME, Siemiatycki J: Occupation and prostate cancer. |
Epidemiol Rev 2001; 23: 138–43 CrossRef MEDLINE
|28.||Ott MG, Yong M, Zober A, et al.: Impact of an occupational health promotion program on subsequent illness and mortality experience. Int Arch Occup Environ Health 2010; 83: 887–94 CrossRef MEDLINE|
|29.|| Yong M, Germann C, Lang S, et al.: Primary selection into shift work and change of cardiovascular risk profile. Scand J Work Environ |
Health 2015; DOI: 10.5271/sjweh.3487. Epub ahead of print CrossRef
|30.||Pan A, Schernhammer ES, Sun Q, et al.: Rotating night shift work and risk of type 2 diabetes: two prospective cohort studies in women. PLoS Med 2011; 8: e1001141 CrossRef MEDLINE PubMed Central|
|31.||Lowden A, Moreno C, Holmback U, et al.: Eating and shift work—effects on habits, metabolism and performance. Scand J Work Environ Health 2010; 36: 150–62 CrossRef MEDLINE|
|32.||Stevens RG, Hansen J, Costa G, et al.: Considerations of circadian impact for defining ’shift work’ in cancer studies: IARC Working Group Report. Occup Environ Med 2011; 68: 154–62 CrossRef MEDLINE|
|33.||Erren TC: Shift work and cancer research: can chronotype predict susceptibility in night-shift and rotating-shift workers? Occup Environ Med 2013; 70: 283–4 CrossRef MEDLINE|
|34.||Yong M, Fischer D, Germann C, Vetter C, Lang S, Oberlinner C: Haben Früh- oder Spättyp einen Einfluss auf die Gesundheit, gemessen anhand des Work Ability Index? München: Deutsche Gesellschaft für Arbeitsmedizin und Umweltmedizin e. V. 2015.|
|35.||Harrington JM: Health effects of shift work and extended hours of work. Occup Environ Med 2001; 58: 68–72 CrossRef PubMed Central|