Using data from the Norwegian Hip Fracture Register, 8639 cemented and 2477 uncemented primary hemiarthroplasties for displaced fractures of the femoral neck in patients aged > 70 years were included in a prospective observational study. A total of 218 re-operations were performed after cemented and 128 after uncemented procedures. Survival of the hemiarthroplasties was calculated using the Kaplan-Meier method and hazard rate ratios (HRR) for revision were calculated using Cox regression analyses. At five years the implant survival was 97% (95% confidence interval (CI) 97 to 97) for cemented and 91% (95% CI 87 to 94) for uncemented hemiarthroplasties. Uncemented hemiarthroplasties had a 2.1 times increased risk of revision compared with cemented prostheses (95% confidence interval 1.7 to 2.6, p < 0.001). The increased risk was mainly caused by revisions for peri-prosthetic fracture (HRR = 17), aseptic loosening (HRR = 17), haematoma formation (HRR = 5.3), superficial infection (HRR = 4.6) and dislocation (HRR = 1.8). More intra-operative complications, including intra-operative death, were reported for the cemented hemiarthroplasties. However, in a time-dependent analysis, the HRR for re-operation in both groups increased as follow-up increased.

This study showed that the risk for revision was higher for uncemented than for cemented hemiarthroplasties.

We undertook a prospective randomised controlled trial involving 400 patients with a displaced intracapsular fracture of the hip to determine whether there was any difference in outcome between treatment with a cemented Thompson hemiarthroplasty and an uncemented Austin-Moore prosthesis. The surviving patients were followed up for between two and five years by a nurse blinded to the type of prosthesis used.

The mean age of the patients was 83 years (61 to 104) and 308 (77%) were women. The degree of residual pain was less in those treated with a cemented prosthesis (p < 0.0001) three months after surgery. Regaining mobility was better in those treated with a cemented implant (p = 0.005) at six months after operation. No statistically significant difference was found between the two groups with regard to mortality, implant-related complications, re-operations or post-operative medical complications.

The use of a cemented Thompson hemiarthroplasty resulted in less pain and less deterioration in mobility than an uncemented Austin-Moore prosthesis with no increase in complications.