It is uncertain whether Thunderbeat has a place in harvesting the left internal mammary artery and whether skeletonisation is superior to pedicle harvested LIMA. Some investigations have shown improved flowrates in the skeletonised graft. The aim of this study was to compare three groups of harvesting techniques: Pedicled, surgical skeletonised and skeletonised with Thunderbeat in terms of flow rates in the left internal mammary artery and postoperative in-hospital outcomes. Patients undergoing coronary artery bypass grafting with the left internal mammary artery to the anterior descending artery were randomized to pedicled (n = 56), surgical skeletonised (n = 55), and skeletonised with Thunderbeat (n = 54). Main outcomes were blood flow and pulsatility index in the graft. No statistical difference between groups regarding flow in LIMA or pulsatility index. Similarly, no difference in postoperative bleeding or days of hospitalisation. The duration of harvesting was faster for the pedicled technique compared with surgical skeletonised and skeletonised with Thunderbeat (mean total min: pedicled 20.2 min SD ± 5.4; surgical skeletonised 28.6 min SD ± 8.7; skeletonised with Thunderbeat 28.3 min SD ± 9.11, p < 0.001). No grafts discarded due to faulty harvesting and there was no graft failure within hospital stay. We found no difference between the harvesting methods except for a significantly faster harvesting time with the pedicled technique. However, non-touch skeletonised left internal mammary artery harvesting with Thunderbeat seems to be an effective alternative to traditional surgical skeletonised LIMA. The future will reveal whether patency is harvesting dependent.