Background and purpose - Straight collarless polished tapered stems have been linked to an increased risk for periprosthetic femur fractures in comparison with anatomically shaped stems, especially in elderly patients. Therefore, we evaluated the effect of an orthopedic department's full transition from the use of a cemented collarless, polished, tapered stem to a cemented anatomic stem on the cumulative incidence of postoperative periprosthetic fracture (PPF). Patients and methods - This prospective single-center cohort study comprises a consecutive series of 1,077 patients who underwent a cemented hip arthroplasty using either a collarless polished tapered stem (PTS group, n = 543) or an anatomic stem (AS group, n = 534). We assessed the incidence of PPF 2 years postoperatively and used a Cox regression model adjusted for age, sex, ASA class, cognitive impairment, BMI, diagnosis, and surgical approach for outcome analysis. Results - Mean age at primary surgery was 82 years (49-102), 73% of the patients were female, and 75% underwent surgery for a femoral neck fracture. The PPF rate was lowered from 3.3% (n = 18) in the PTS group to 0.4% (n = 2) in the AS group. The overall complication rate was also lowered from 8.8% in the PTS group to 4.5% in the AS group. In the regression model only cognitive dysfunction (HR 3.8, 95% CI 1.4-10) and the type of stem (PTS vs AS, HR 0.1, CI 0.0-0.5) were correlated with outcome. Interpretation - For elderly patients with poor bone quality use of cemented anatomic stems leads to a substantial reduction in periprosthetic fracture rate without increasing other complications.