Although definitions vary slightly across sources, shift work can be defined as employment in any work schedule that is not a regular daytime schedule (that is, approximately 0900 to 1700). The full spectrum of shift work comprises regular evening or night schedules, rotating shifts, split shifts, on-call or casual shifts, 24-hour shifts, irregular schedules and other non-day schedules. 

Shift work has long been known to disrupt circadian rhythm, sleep, and work-life balance, however, flexible work patterns remain a necessary component for a dynamic, diversified, industrial economy. The association of shift work with vascular disease is controversial. Conflicting data on this association exist, perhaps in part owing to varying methods, populations, and definitions of shift work and vascular or coronary events. 

Furthermore, previous syntheses:

• Are now outdated
• Did not use validated tools for assessing studies
• Did not capture all available data
• Did not apply quantitative techniques to compute summary risk estimates.

Given these uncertainties, the authors of this study from Ontario, Canada, comprehensively analysed the epidemiology of shift work and vascular events as reported in the biomedical literature.¹

The data sources were systematic searches of major bibliographic databases, contact with experts in the field, and review of reference lists of primary articles, review papers, and guidelines. Observational studies that reported risk ratios for vascular morbidity, vascular mortality or all-cause mortality in relation to shift work were included. Control groups could be non-shift ‘day’ workers or the general population. Study quality was assessed with the Downs and Black scale for observational studies. 

The three primary outcomes were myocardial infarction, ischaemic stroke, and any coronary event. Heterogeneity was measured with the I² statistic and computed random effects models.

Some 34 studies in 2,011,935 people were identified. Shift work was associated with myocardial infarction (risk ratio 1.23, 95% confidence interval 1.15 to 1.31; I [2] = 0) and ischaemic stroke (1.05, 1.01 to 1.09; I [2] = 0). 

Coronary events were also increased (risk ratio 1.24, 1.10 to 1.39), albeit with significant heterogeneity across studies (I [2] = 85%). Pooled risk ratios were significant for both unadjusted analyses and analyses adjusted for risk factors. All shift work schedules with the exception of evening shifts were associated with a statistically higher risk of coronary events. 

Shift work was not associated with increased rates of mortality (whether vascular cause-specific or overall). Presence or absence of adjustment for smoking and socioeconomic status was not a source of heterogeneity in the primary studies. There were 6,598 myocardial infarctions (MIs), 17,359 coronary events, and 1,854 ischaemic strokes occurred. On the basis of the Canadian prevalence of shift work of 32.8%, the population-attributable risks related to shift work, were 7.0% for MI, 7.3% for all coronary events and 1.6% for ischaemic stroke.

This is the largest synthesis of shift work and vascular risk reported to date. Previous work has been hampered by a narrow focus on only one type of risk (such as coronary disease), a lack of completeness in identifying all relevant studies, absence of quantitative synthesis through conventional meta-analytic techniques, and failure to use a validated tool to assess the quality of studies. 

On the other hand, several caveats must be noted. Different studies adjusted for different risk factors, although broad consistency between unadjusted and adjusted models was noted. None of the secondary endpoints was statistically significant in relation to shift work, and 95% confidence intervals were wide. 

The large number of distinct cardiovascular maladies represented under the rubric ‘cardiovascular events’ in various iterations of the international classification of diseases coding system could have caused imprecision in risk for this entity. 

In addition, ischaemic stroke was reported by only two studies and cerebrovascular mortality by an additional four studies. Pooled together, this gives 4,592 events (still substantially less than the numbers of coronary events and deaths). 

Finally, the dataset lacked discrete information on the diurnal type of workers ‘morningness’ or ‘eveningness’, so it was not possible to determine whether associations differed across this important characteristic.

Reference

1. Vyas MV, Garg AX, Iansavichus AV et al, BMJ 2012; 345: e4800