The excess relative risk (ERR) of mortality for circulatory disease among nuclear workers was reanalyzed by taking into consideration the annual dose as the dose rate using publicly available epidemiological data of the Hanford site dedicated to the cohort study of nuclear workers in the US, the UK, and Canada. Values of the dose rate (cut-points) were chosen at 2 mSv y−1 intervals from 2 to 40 mSv y−1, and risk estimates were made for 32,988 workers, considering the doses accumulated below and above each cut-point to have different effects. Similarly to that in the previous study for cancer by Sasaki et al., examinations of sensitivity analysis were also carried out for different risk models, lag periods, and impacts of adjusting the monitoring period to find the effect of the dose rate. As a result, emergence of a statistically significant difference between βL1, which is the ERR for the doses accumulated below the specified cut-point, and βH1, which is the ERR for that above the specified cut-point, was observed for cut-point of the dose rate of 4, 6, 34, 36, and 38 mSv y−1. While statistically negative values were estimated for βL1 (4 and 6 mSv y−1) and for βH1 (34, 36, and 38 mSv y−1), the overall relationship between the ERR and the cut-point of the dose rate was found to be similar to that obtained by the analysis of the mortality for all cancers excluding leukemia.