Background and purpose — Almost all prosthetic implant brands include several variations. Most studies on implant performance investigate an implant system without sub-analysis of implant attributes. We studied the influence of design variations during the last 2 decades on implant survival of the most frequently used cemented femoral stem, the Lubinus SPII, reported to the Swedish Arthroplasty Register (SHAR). Patients and methods — Between 2000 and 2018, 100,032 cemented Lubinus SP II stems had been reported to SHAR. Patients with primary osteoarthritis operated on with stem length 150 mm together with a cemented cup from the same manufacturer (n = 76,530) were included in this analysis. Primary study outcome was non-infectious stem revision. Cox regression with adjustment for age, sex, surgical approach, and year of surgery was used. Hazard rates (HR) are presented with 95% confidence intervals (CI). Results — Average follow-up was 7.6 years (SD 4.2). At 18 years the overall stem survival was 99.1 (CI 98.9–99.3). Increased revision rate was observed for stems with extra offset, when a long or an extra-long head length had been used. Smaller stem sizes, in particular the smallest stem size (01), substantially increased the rate of mechanical failure as reflected by an almost 10 times increased rate of revision compared with the standard size (2). Interpretation — In this study with larger sample size and longer follow-up than previously presented from the same register, we observed more pronounced effects of previously documented design variations. Based on our results, surgeons using the Lubinus SPII stems are advised to consider an alternative solution if a Lubinus stem size 01, Lubinus extra offset, or an extra-long head seems to be the most suitable choice at the preoperative planning.