Based on the maintained database (source database) on effects in nuclear workers (NW), a selection of major studies of the relationship between mortality from diseases of the circulatory system (CVD; codes 390–459 according to ICD-9 and I00–I99 according to ICD-10) and external radiation dose. The sample included 30 papers and covered cohorts from 6 countries plus an NW cohort from 15 countries. For the sample, in most cases based on published standardized mortality rates (SMR), the relative risks (RR) of mortality from CVD were calculated for the selected dose groups with subsequent processing of the material for outliers. Initial: n = 207; final sample: n = 199; covers very low (0–10 mSv; 15.8 % of the sample), low (>10–100 mSv; 45.8 %) and moderate (>100–1000 mSv; 36.4 %) doses; data for high doses (>1000 mSv; n = 4; 2 % of the sample), due to dubiousness, were excluded. A systematic review and pooled analysis of the RR for mortality from CVD depending on the dose on an ordinal scale was performed on the final sample. For the entire dose range (0–1000 mSv) and for moderate doses, statistically significant trends in increasing RR were found when expressed in five types of regressions (except for the logarithmic one for the entire range). Although the r values were small (0.230–0.293), the effect was clear. The ERR per 1 Gy (Sv) calculated for moderate doses using linear regression was 0.54. This value is higher than those obtained previously in meta-analyses, but should be considered as the most adequate. No dose relationship was found for the very low + low dose range (0–100 mGy); the r coefficients for the regressions were either negligible or negative at statistical insignificance. For the subthreshold dose range for CVD mortality after exposure (according to UNSCEAR and ICRP: 500 mSv), only a weak trend towards an increase in RR was found, statistically insignificant, despite the large sample size (n = 191), while for the dose range 500–1000 mSv, the highest tendency among the pooled analyzes was revealed to increase the risk depending on the level of exposure (r = 0.297–0.423; statistically insignificant due to the small sample size: n = 8). It is concluded that for mortality from CVD after irradiation, the threshold value of 0.5 Gy established by UNSCEAR and ICRP and confirmed in the present pooled analysis should be strictly adhered to. Due to the lack of effects of low doses, it is inappropriate to raise the issue of low dose effects in the context of these pathologies.