PURPOSE: We aimed to compare the long-term outcomes of small-head (28 mm) metal-on-metal (MoM) total hip arthroplasty (THA) to ceramic-on-polyethylene (CoP) THA using the same cup. METHODS: All primary elective MoM and CoP THAs performed 1998-2011 were prospectively included in a local registry. Patients were followed until 31 December 2022. Outcomes were all-cause revision, complications and mortality. The uncemented Morscher 28 mm monobloc press-fit cup was used in all THAs. RESULTS: Overall, 3257 THAs were included, 864 MoM (mean age 63) and 2393 CoP THAs (mean age 72). Mean follow-up of the cohort was 12.9 years (maximum 26.8 years). Revision for any cause was performed in 85 MoM and 79 CoP THAs. Cumulative incidence of all-cause revision at 20 years was 13.2% (95% CI 10.6 to 16.3) in MoM and 6.3% (95% CI 4.8 to 8.3) in CoP group. Adjusted hazard ratio for all-cause revision was 1.88 (95% CI 1.34 to 2.65) comparing MoM vs. CoP. Diagnoses at revision were mainly aseptic loosening (33%) and adverse local tissue reactions (33%) in MoM and aseptic loosening in CoP group (44%). The smoothed hazard function revealed the largest difference in instantaneous revision rate between three and 14 years postoperative. After that period no difference was observed. CONCLUSION: Overall, the cumulative risk of all-cause revision was almost twice as high in patients with a small head MoM as compared to a CoP THA over the 20-year period. However, most of the excess in revisions among MoM patients occurred between three and 14 years postoperative.