The influence of artefacts due to metal implants on the range of ion beams is investigated, using a geometrically well-defined head and pelvic phantom together with inserts from steel, titanium and tungsten. The ranges along various beam paths including artefacts were calculated from the TPS and compared to known calculations for phantoms without any insert. In the head phantom, beams intersecting the streak artefacts lead to errors in the range of around or below 1%, which is mainly due to a cancellation of various effects. Beams through the metal or close to it show an underestimation of 3.5% of the range for tungsten. For the pelvic phantom, a large underestimation of the range is observed for a lateral path through the metal insert. In the case of tungsten and steel, range errors of −5% and −18% are observed, respectively. Such beam paths are typically used for pelvic tumours in radiotherapy with ion beams. For beams in the anterior–posterior direction through the inserts, an overestimation of ion ranges of up to 3% for titanium and 8% for steel is expected, respectively. Beam paths outside the metal insert show a large cancellation for the lateral beams (leading to errors of around 1% only) and somewhat higher errors for anterior–posterior beams (around 3% for titanium and 6% for steel). The analysis of CT data of patients with dental implants of gold as compared to patients with healthy teeth also showed a significant effect of the artefacts on the distribution of HU in the data, namely a redistribution of HU to higher and lower values as compared to patients with healthy teeth. The corresponding mean range variation was a 2.5% reduction in the data with artefacts as compared to the data without artefacts. It is concluded that beam paths through metal implants should generally be avoided in proton and ion therapy. In this case, the underestimation of ion range due to artefacts alone may amount to 3% for dental fillings and up to 5% and 18% for hip prosthesis made of titanium and steel, respectively. It is important to note that the size of the metal inserts cannot be determined correctly from the images, so that a correction of the ranges in metal also leads to large uncertainties. Finally, it should be stressed that the stated relative deviations are strictly valid only for the investigated phantoms and can only give a rough estimate on the size of range uncertainties that may appear in real patients.