The purpose of this study was to determine the effect of contrast material and retrospective ECG-gating on radiation dose in an adult cardiac dual-energy computed tomography (DECT). Sixty-two patients underwent CT cardiac examination with a Somatom Definition Flash DECT using tube voltages of 100 kV without filter and 140 kV with a tin filter (mean effective mA s: arterial 72.39 and 62.94, venous 93.21 and 78.45, and late phase 134.5 and 118.2). The arterial and late phases were examined with retrospective ECG-gating, but gating was not used for the venous phase. Seventy milliliters (70 ml) iodinated contrast material (CM) was injected into the patient during examination. The effective doses (ED) were calculated from dose-length-product (DLP) and computed tomographic dose index volume (CTDIvol) using the latest k-factor (0.028). Pearson’s correlation coefficient was used for statistical tests on continuous variables. Mean CTDIvol and DLP were lower in the late phase (10.15 ± 1.5 mGy and 202.9 ± 23 mGy cm) compared to the arterial phase (19.69 ± 3 mGy and 394 ± 90 mGy cm). Differences between the arterial and late phase were statistically significant (p = 0.005), and mean values for the late phase were 48.5% lower than mean values for the arterial phase. Mean CTDIvol and DLP were lower in venous (7.72 ± 1 mGy and 154.3 ± 17 mGy cm) compared to late phase (10.15 ± 1.5 mGy and 202.9 ± 23 mGy cm). The difference between venous and late phase was statistically significant (p < 0.001). The mean results for the venous phase were 24% lower than those for the late phase. This study shows that contrast material (CM) absorbs radiation significantly and increases dose by 48.5% in an adult cardiac dual-energy CT with retrospective ECG-gating. Care must be taken to determine the type, concentration, and volume of CM used for the scan. The dual-energy non-ECG-gated technique decreased radiation dose by 24% compared to the ECG-gated technique. ECG-gated cardiac examination should be limited to patients with strong clinical indications. SNR and HU increased with decreasing energy. The image noise values showed a negligible difference in the arterial and late phase datasets, and this did not affect the diagnostic quality of the image evaluation.