Posterior approach and uncemented stems increases the risk of reoperation after hemiarthroplasties in elderly hip fracture patients

Abstract

Background Hemiarthroplasties are performed in great numbers worldwide but are seldom registered on a national basis. Our aim was to identify risk factors for reoperation after fracture-related hemiarthroplasty in Norway and Sweden.

Material and methods A common dataset was created based on the Norwegian Hip Fracture Register and the Swedish Hip Arthroplasty Register. 33,205 hip fractures in individuals > 60 years of age treated with modular hemiarthroplasties were reported for the period 2005–2010. Cox regression analyses based on reoperations were performed (covariates: age group, sex, type of stem and implant head, surgical approach, and hospital volume).

Results 1,164 patients (3.5%) were reoperated during a mean follow-up of 2.7 (SD 1.7) years. In patients over 85 years, an increased risk of reoperation was found for uncemented stems (HR = 2.2, 95% CI: 1.7–2.8), bipolar heads (HR = 1.4, CI: 1.2–1.8), posterior approach (HR = 1.4, CI: 1.2–1.8) and male sex (HR = 1.3, CI: 1.0–1.6). For patients aged 75–85 years, uncemented stems (HR = 1.6, 95% CI: 1.2–2.0) and men (HR = 1.3, CI: 1.1–1.6) carried an increased risk. Increased risk of reoperation due to infection was found for patients aged < 75 years (HR = 1.5, CI: 1.1–2.0) and for uncemented stems. For open surgery due to dislocation, the strongest risk factor was a posterior approach (HR = 2.2, CI: 1.8–2.6). Uncemented stems in particular (HR = 3.6, CI: 2.4–5.3) and male sex increased the risk of periprosthetic fracture surgery.

Interpretation Cemented stems and a direct lateral transgluteal approach reduced the risk of reoperation after hip fractures treated with hemiarthroplasty in patients over 75 years. Men and younger patients had a higher risk of reoperation. For the age group 60–74 years, there were no such differences in risk in this material.

CR and JEG planned the study. AMF was responsible for preparing the common dataset. JEG was responsible for the Norwegian dataset and CR was responsible for the Swedish dataset. CR, JEG, AMF, and JK performed the statistical analyses. CR wrote the manuscript with substantial contributions from all authors, who also participated in data collection and in interpretation of the results.

We thank Anna Sandelin at the Center of Registers in Region Västra Götaland for preparing the Swedish dataset. We also thank the staff of the Norwegian and Swedish registries for their work. Finally, we are grateful to all the surgeons and coordinators who loyally report to the national registries.

No competing interests declared.