Shift work has been associated with a higher risk of heart disease. The epidemiological literature is heterogeneous with some studies finding risks for shift workers twice as high as for day workers, while other studies do not find shift workers to be at a higher risk of heart disease. We have examined the literature, using a meta analytic approach to give both a more precise estimate of the risk and to explore reasons for the apparent heterogeneity. The analysis suggests that among the longitudinal studies using individual exposure classification at least two different study populations existed based on statistical testing, one finding an increased risk of heart disease with a meta risk estimate of 1.31 (95% confidence interval (CI): 1.17-1.45), the other finding no excess risk with an estimate of 0.96 (95% CI: 0.85-1.08). Study characteristics including publication year, geographic area, exposure estimation, outcome, follow up period, confounding control or overall quality of study did not explain the heterogeneity. In conclusion shift work seems in some instances, not explained by study characteristics, to raise the risk of heart disease.