M Anne Harris1,2,3, Jill MacLeod3, Joanne Kim3,4, Manisha Pahwa3,5, Michael Tjepkema6, Paul Peters7, Paul A Demers2,3. 1. School of Occupational and Public Health, Ryerson University, Toronto ON, Canada. 2. Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada. 3. Occupational Cancer Research Centre, Cancer Care Ontario, Toronto, ON, Canada. 4. Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, QC, Canada. 5. Centre for Health Economics and Policy Analysis, McMaster University, Hamilton, ON, Canada. 6. Statistics Canada, Health Analysis Division, Ottawa, ON, Canada. 7. Department of Health Sciences, Carleton University, Ottawa, ON, Canada.
Abstract
OBJECTIVES: Shift work with circadian disruption is a suspected human carcinogen. Additional population-representative human studies are needed and large population-based linkage cohorts have been explored as an option for surveillance shift work and cancer risk. This study uses a surveillance linkage cohort and job-exposure matrix to test relationships. METHODS: We estimated associations between shift work and breast, ovarian, and prostate cancer using the population-based Canadian Census Health and Environment Cohort (CanCHEC), linking the 1991 Canadian census to national cancer registry and mortality databases. Prevalence estimates from population labour survey data were used to estimate and assign probability of night, rotating, or evening shifts by occupation and industry. Cohort members were assigned to high (>50%), medium (>25 to 50%), low (>5 to 25%), or no (<5%) probability of exposure categories. Cox proportional hazards modelling was used to estimate associations between shift work exposure and incidence of prostate cancer in men and ovarian and breast cancer in women. RESULTS: The cohort included 1 098 935 men and 939 520 women. Hazard ratios (HRs) indicated null or inverse relationships comparing high probability to no exposure for prostate cancer: HR = 0.96, 95% confidence interval (CI) = 0.91-1.02; breast cancer: HR = 0.94, 95% CI = 0.90-0.99; and ovarian cancer: HR = 0.99, 95% CI = 0.87-1.13. CONCLUSIONS: This study showed inverse and null associations between shift work exposure and incidence of prostate, breast, or ovarian cancer. However, we explore limitations of a surveillance cohort, including a possible healthy worker survivor effect and the possibility that this relationship may require the nuanced exposure detail in primary collection studies to be measurable.
OBJECTIVES: Shift work with circadian disruption is a suspected human carcinogen. Additional population-representative human studies are needed and large population-based linkage cohorts have been explored as an option for surveillance shift work and cancer risk. This study uses a surveillance linkage cohort and job-exposure matrix to test relationships. METHODS: We estimated associations between shift work and breast, ovarian, and prostate cancer using the population-based Canadian Census Health and Environment Cohort (CanCHEC), linking the 1991 Canadian census to national cancer registry and mortality databases. Prevalence estimates from population labour survey data were used to estimate and assign probability of night, rotating, or evening shifts by occupation and industry. Cohort members were assigned to high (>50%), medium (>25 to 50%), low (>5 to 25%), or no (<5%) probability of exposure categories. Cox proportional hazards modelling was used to estimate associations between shift work exposure and incidence of prostate cancer in men and ovarian and breast cancer in women. RESULTS: The cohort included 1 098 935 men and 939 520 women. Hazard ratios (HRs) indicated null or inverse relationships comparing high probability to no exposure for prostate cancer: HR = 0.96, 95% confidence interval (CI) = 0.91-1.02; breast cancer: HR = 0.94, 95% CI = 0.90-0.99; and ovarian cancer: HR = 0.99, 95% CI = 0.87-1.13. CONCLUSIONS: This study showed inverse and null associations between shift work exposure and incidence of prostate, breast, or ovarian cancer. However, we explore limitations of a surveillance cohort, including a possible healthy worker survivor effect and the possibility that this relationship may require the nuanced exposure detail in primary collection studies to be measurable.