Background and purpose — Little is known about the long-term migration pattern of cementless stems in total hip arthroplasty (THA). Furthermore, the role of bioactive coatings in fixation, and thus migration, remains uncertain. Hydroxyapatite (HA) is the most commonly used bioactive coating. However, delamination of the coating might induce loosening. Alternatively, fluorapatite (FA) has proved to be more thermostable than HA, thereby potentially increasing longevity. We assessed the long-term migration of cementless stems with different coatings using radiostereometric analysis (RSA), thereby establishing a reference for acceptable migration. Patients and methods — 61 THAs in 53 patients were randomized to receive either a HA, FA, or uncoated Mallory-Head Porous stem during the years 1992 to 1994. Primary outcome was stem migration measured using RSA and secondary outcome was the Harris Hip Score (HHS). Evaluation took place preoperatively and postoperatively on the second day, at 6, 12, 25 and 52 weeks, and annually thereafter. At the 25-year follow-up, 12 patients (17 THAs) had died and 1 patient (1 THA) was lost to follow-up. Due to the high number of missing second-day postoperative RSA radiographs, the 1-year postoperative RSA radiograph was used as baseline for the comparative analyses. Results — Mean follow-up was 17 years (SD 6.6). All stems showed initial rapid migration with median subsidence of 0.2 mm (–0.1 to 0.6) and median retroversion of 0.9° (–3.2 to 2) at 12 months, followed by stable migration reaching a plateau phase. No stem was revised, albeit 1 stem showed continuous subsidence up to 1.5 mm. Comparing the different coatings, we could not find a statistically significant difference in overall 25-year migration (p-values > 0.05). Median subsidence at 15-year follow-up was for HA –0.1 mm (–0.4 to 0.2), for FA 0 mm (–0.1 to 0.2), and for uncoated stems 0.2 mm (–0.1 to 0.5). Median internal rotation at 15-year follow-up was for HA not available, for FA 1.1° (–0.5 to 2.6), and for uncoated stems 0° (–0.5 to 0.4). HHS were also comparable (p-values > 0.05), with at 15-year follow-up for HA 85 points (41–99), for FA 76 points (61–90), and for uncoated stems 79 points (74–90). Interpretation — The long-term migration pattern of cementless stems using different bioactive coatings has not previously been described. No beneficial effect, or side effect at long-term follow-up of bioactive coatings, was found. The provided migration data can be used in future research to establish thresholds for acceptable migration patterns cementless stem designs.