BACKGROUND AND PURPOSE: There is no clear evidence on whether migration or wear is the best predictor for later acetabular cup loosening. We aimed to investigate whether early wear or migration, measured via radiostereometric analysis (RSA), predicts later cup loosening. We also compared long-term aseptic loosening rates between conventional (CPE) and highly crosslinked polyethylene (XLPE) cups. METHODS: Data was drawn from a randomized controlled trial (RCT) (ClinicalTrials.Gov NCT00698672) of 150 patients receiving cemented total hip arthroplasties (THAs), with 10-year RSA follow-up. 5 groups were assessed based on implant combinations (Charnley or Spectron EF stems with CPE or XLPE cups and CoCr or Oxinium heads). Migration and wear up to 2 years were evaluated against 18-year cup survival using receiver operating characteristic (ROC) curves. RESULTS: 19 cups (17 CPE, 2 XLPE) were loose at final follow-up. The area under the ROC curve (AUC) was 0.56 (95% confidence interval [CI] 0.40-0.73) for early migration and 0.85 (CI 0.77-0.94) for early polyethylene (PE) wear, with a difference of 0.29 (CI 0.09-0.49). Hazard ratio for loosening was 0.88 (CI 0.20-3.89) for early migration > 0.2 mm and 19.4 (CI 2.55-147) for early wear > 0.2 mm. At 18 years, survival free of aseptic loosening was 65% (CI 48-77) for CPE and 96% (CI 85-99) for XLPE cups, with a 9-fold higher risk of loosening for CPE. CONCLUSION: Early polyethylene wear, not migration, predicted long-term cup loosening. XLPE showed superior long-term performance over CPE with less wear, cup loosening, and revision.