RESUMO
BACKGROUND: The World Health Organization (WHO) and the International Labour Organization (ILO) are developing joint estimates of the work-related burden of disease and injury (WHO/ILO Joint Estimates), with contributions from a large network of individual experts. Evidence from mechanistic data and prior studies suggests that exposure to long working hours may cause stroke. In this paper, we present a systematic review and meta-analysis of parameters for estimating the number of deaths and disability-adjusted life years from stroke that are attributable to exposure to long working hours, for the development of the WHO/ILO Joint Estimates. OBJECTIVES: We aimed to systematically review and meta-analyse estimates of the effect of exposure to long working hours (three categories: 41-48, 49-54 and ≥55 h/week), compared with exposure to standard working hours (35-40 h/week), on stroke (three outcomes: prevalence, incidence, and mortality). DATA SOURCES: A protocol was developed and published, applying the Navigation Guide to systematic reviews as an organizing framework where feasible. We searched electronic databases for potentially relevant records from published and unpublished studies, including Ovid MEDLINE, PubMed, EMBASE, Scopus, Web of Science, CISDOC, PsycINFO, and WHO ICTRP. We also searched grey literature databases, Internet search engines, and organizational websites; hand-searched reference lists of previous systematic reviews; and consulted additional experts. STUDY ELIGIBILITY AND CRITERIA: We included working-age (≥15 years) individuals in the formal and informal economy in any WHO and/or ILO Member State but excluded children (aged < 15 years) and unpaid domestic workers. We included randomized controlled trials, cohort studies, case-control studies and other non-randomized intervention studies with an estimate of the effect of exposure to long working hours (41-48, 49-54 and ≥55 h/week), compared with exposure to standard working hours (35-40 h/week), on stroke (prevalence, incidence or mortality). STUDY APPRAISAL AND SYNTHESIS METHODS: At least two review authors independently screened titles and abstracts against the eligibility criteria at a first review stage and full texts of potentially eligible records at a second stage, followed by extraction of data from qualifying studies. Missing data were requested from principal study authors. We combined relative risks using random-effects meta-analysis. Two or more review authors assessed the risk of bias, quality of evidence and strength of evidence, using the Navigation Guide and GRADE tools and approaches adapted to this project. RESULTS: Twenty-two studies (20 cohort studies, 2 case-control studies) met the inclusion criteria, comprising a total of 839,680 participants (364,616 females) in eight countries from three WHO regions (Americas, Europe, and Western Pacific). The exposure was measured using self-reports in all studies, and the outcome was assessed with administrative health records (13 studies), self-reported physician diagnosis (7 studies), direct diagnosis by a physician (1 study) or during a medical interview (1 study). The outcome was defined as an incident non-fatal stroke event in nine studies (7 cohort studies, 2 case-control studies), incident fatal stroke event in one cohort study and incident non-fatal or fatal ("mixed") event in 12 studies (all cohort studies). Cohort studies were judged to have a relatively low risk of bias; therefore, we prioritized evidence from these studies, but synthesised evidence from case-control studies as supporting evidence. For the bodies of evidence for both outcomes with any eligible studies (i.e. stroke incidence and mortality), we did not have serious concerns for risk of bias (at least for the cohort studies). Eligible studies were found on the effects of long working hours on stroke incidence and mortality, but not prevalence. Compared with working 35-40 h/week, we were uncertain about the effect on incidence of stroke due to working 41-48 h/week (relative risk (RR) 1.04, 95% confidence interval (CI) 0.94-1.14, 18 studies, 277,202 participants, I2 0%, low quality of evidence). There may have been an increased risk for acquiring stroke when working 49-54 h/week compared with 35-40 h/week (RR 1.13, 95% CI 1.00-1.28, 17 studies, 275,181participants, I2 0%, p 0.04, moderate quality of evidence). Compared with working 35-40 h/week, working ≥55 h/week may have led to a moderate, clinically meaningful increase in the risk of acquiring stroke, when followed up between one year and 20 years (RR 1.35, 95% CI 1.13 to 1.61, 7 studies, 162,644 participants, I2 3%, moderate quality of evidence). Compared with working 35-40 h/week, we were very uncertain about the effect on dying (mortality) of stroke due to working 41-48 h/week (RR 1.01, 95% CI 0.91-1.12, 12 studies, 265,937 participants, I2 0%, low quality of evidence), 49-54 h/week (RR 1.13, 95% CI 0.99-1.29, 11 studies, 256,129 participants, I2 0%, low quality of evidence) and 55 h/week (RR 1.08, 95% CI 0.89-1.31, 10 studies, 664,647 participants, I2 20%, low quality of evidence). Subgroup analyses found no evidence for differences by WHO region, age, sex, socioeconomic status and type of stroke. Sensitivity analyses found no differences by outcome definition (exclusively non-fatal or fatal versus "mixed") except for the comparison working ≥55 h/week versus 35-40 h/week for stroke incidence (p for subgroup differences: 0.05), risk of bias ("high"/"probably high" ratings in any domain versus "low"/"probably low" in all domains), effect estimate measures (risk versus hazard versus odds ratios) and comparator (exact versus approximate definition). CONCLUSIONS: We judged the existing bodies of evidence for human evidence as "inadequate evidence for harmfulness" for all exposure categories for stroke prevalence and mortality and for exposure to 41-48 h/week for stroke incidence. Evidence on exposure to 48-54 h/week and ≥55 h/week was judged as "limited evidence for harmfulness" and "sufficient evidence for harmfulness" for stroke incidence, respectively. Producing estimates for the burden of stroke attributable to exposures to working 48-54 and ≥55 h/week appears evidence-based, and the pooled effect estimates presented in this systematic review could be used as input data for the WHO/ILO Joint Estimates. PROTOCOL IDENTIFIER: https://doi.org/10.1016/j.envint.2018.06.016. PROSPERO REGISTRATION NUMBER: CRD42017060124.