Background and purpose — There is limited amount of evidence about optimal revision indications, technique, and implants when performing revision surgery for metal-on-metal (MoM) hip replacements due to adverse reaction to metal debris (ARMD). We assessed which factors are related to re-revisions and complications after a revision of MoM hip arthroplasty because of ARMD. We also aimed to provide information on optimal implants for these revisions.Patients and methods — 420 MoM total hip arthroplasties (THA) and 108 MoM hip resurfacings were implanted and later revised at our institution. We used Cox regression to analyze the factors associated with re-revisions and complications after a revision for ARMD.Results — A re-revision was performed on 27 THAs (6%) and 9 resurfacings (8%). The most common indication for re-revision was recurrent dislocation (20 hips, 4%). Complications not leading to re-revision were seen in 21 THAs (5%) and 6 resurfacings (6%). The most common complication was dislocation treated with closed reduction in 13 hips (2%). Use of revision head size > 36mm was associated with decreased risk for dislocations. Presence of pseudotumor, pseudotumor grade, pseudotumor size, or the choice of bearing couple were not observed to affect the risk for re-revision. Non-linear association was observed between preoperative cobalt and risk for re-revision.Interpretation — As dislocation was the most frequent post-revision complication, large head sizes should be used in revisions. Because size or type of pseudotumor were not associated with risk of re-revision, clinicians may have to reconsider, how much weight is put on the imaging findings when deciding whether or not to revise. In our data blood cobalt was associated with risk for re-revision, but this finding needs further assessment.