Shift work and risk of non-cancer mortality in a cohort of German male chemical workers
Objectives Shift work is widely considered to be a health risk. In a previous study, we observed no elevated risk of total mortality in BASF shift workers followed up until the end of 2006. The present study aims to investigate non-cancer mortality, especially mortality caused by ischaemic heart disease (IHD), relative to shift work.
Methods The cohort consisted of 14,038 male shift and 17,105 male day workers from manufacturing plants, who were employed for at least 1 year between 1995 and 2005. Vital status was followed from 2000 to 2009. Cause-specific mortality was obtained from death certificates. Non-cancer mortality as well as mortality specific to diagnoses from I20.0 to I25.9 according to International Classification of Disease version 10 was compared between the two working-time systems. To estimate the impact of shift work on the outcome of interest, Cox proportional hazard model was used to adjust for potential confounders such as age, smoking, alcohol consumption, job level, and disease status at baseline. The effect estimates were then given as hazard ratio (HR) with 95 % confidence interval (CI).
Results Between 2000 and 2009, a total of 1,062 deaths occurred in the cohort: 513 (3.6 %) in shift and 549 (3.2 %) in day workers. Among them were 122 deaths resulting from IHD, 55 (0.39 %) and 67 (0.39 %), respectively. After adjustment for age at entry and job level, no increased risk of non-cancer mortality (HR 0.94; 95 % CI 0.77–1.15) as well as of IHD-caused mortality was found among shift workers (HR 0.77; 95 % CI 0.52–1.14). The risk estimates were robust after further adjustment for more factors in all models and consistently tended to be in favour of shift workers. Considering the duration of exposure to shift, no dose–response relationship was found.
Conclusion The present analysis does not find strong evidence for an increased mortality risk due to non-cancer disease and, more specifically, IHD-caused mortality associated with this shift system. Initial selection based on health criteria as well as ongoing health surveillance and health promotion is likely to have contributed to this result. Shift work over 34 years may lead to a loss of this initial selection advantage over time, but the respective risk estimates lacked statistical precision.