Introduction: Periprosthetic fractures have long been recognized as one of the major complications after hip resurfacing arthroplasty. Both biomechanical factors and pathological changes of bone tissue might hypothetically influence its occurrence. We analyzed retrieved femoral remnants to identify possibly different fracture modes.

Material and Methods: 83 hips revised for periprosthetic fracture (134.5 days in situ±159.2) were analyzed macroscopically, contact radiographically and histologically. Most cases (80.7%) were treated for advanced stages of osteoarthritis. Hips with preoperative femoral head necrosis were not included. 49 (59.0%) patients were men (57.8 years old±8.5) and 34 (41.0%) women (55.1 years old±10.0; p=.3445). Occurrence of reactive changes and of avascular necrosis in addition to amount of osteonecrosis were used as the major histological criteria for classification of the fracture as acute biomechanical, acute postnecrotic or chronic.

Statistical analyses were performed using statistical software. Probability of Type I error was set to 5% (alpha=0.05).

Results: 37 (44.6%) femoral neck fractures (83.9 days±87.7) occurred earlier than the remaining 46 (55.4%) head fractures (174.1 days±89.7; p=.0129). 50(60.2%) remnants revealed complete osteonecrosis and were thus classified as acute postnecrotic fractures, 29 (34.9%) chronic fractures were characterized by finding of pseudoarthrosis or preformed callus and the remaining 4 (4.8%) were classified as acute mechanic. Acute mechanic fractures (17.5 days±8.0) failed earlier than both acute postnecrotic (146.3 days±181.7; p=.0049) and chronic (130.8 days±120.6; p=.0017) fractures.

Osteonecrosis was found in 81 (97.6%) hips revised after fracture (p< .0001). The vertical size of avascular necrosis in hips after acute postnecrotic fracture (21.1mm±8.5) was bigger (p< .0001) than in both chronic (7.3mm±7.3) and acute mechanic (0.9 mm±1.2) fractures.

Even though 33 (66.0%) of 50 patients with acute postnecrotic fracture were men (p=.0237), no significant differences between males and females were found with respect to age of patients (p=.3445) or duration of prosthesis implantation (p=.1232).

Conclusion: We analyzed hips revised for periprosthetic fracture after the resurfacing arthroplasty. Three distinct fracture modes of this complication could have been identified morphologically. Osteonecrosis secondary to the hip resurfacing arthroplasty appeared to be causative for more than a half of all fractures in present cohort. Mechanical and biomechanical factors related to the procedure might have possibly influenced the occurrence of both postnecrotic fractures and cases with vital reactive changes of tissues neighbouring the fracture line.

The proposed classification may help to understand causes of periprosthetic fractures after hip resurfacing arthroplasty.

Background: A general increase in total number of primary total hip arthroplasty (THA) has been observed in Denmark from 3.828 in 1995 to 7.645 in 2006. During the same period the number of pa-tients treated at private clinics has also increased. To our knowledge no studies, comparing patient characteristics and treatment quality between public and private hospitals, have been published.

We compared patients’ characteristics and outcome following THA in private and public hos-pitals.

Materials and Methods: We used data from the Danish Hip Arthroplasty Registry to identify 69 249 primary THA’ies performed between 1 January 1995 to 31 December 2006.

To detect eventual difference in patient characteristics- age, gender, diagnosis leading to THA, Carlson’s comorbidity score and Charnley category were evaluated.

We matched 3 658 cases operated in private with 3 658 controllers operated in public hospitals on propensity score. Scoring parameters were age, gender, diagnosis leading to THA, Carlson’s comorbidity score, Charnley category, operating time, type of anesthesia and type of prosthesis.

We used multivariate logistic regression on propensity score matched data to assess association between type of hospital and outcome by computing relative risks and 95% Confidence Interval (CI). Outcomes were perioperative complications, readmission within 3 months, re-operation within 2 years, implant failure after 5 years, and mortality within 3 months of surgery.

Results: Private hospitals operated on older females, patients with primary osteoarthritis and low comorbidity and Charnley category 1.

Patients in private and propensity matched controls from public hospitals showed no differences in age, gender, diagnosis leading to THA, Carlson’s comorbidity score, Charnley category, operating time, type of anesthesia and type of prosthesis (p-value < 0,0001).

Based on matched data, private hospitals had lower relative risk for perioperative complications (0.39, 0.26–0.60), reoperations (0.59, 0.41–0.83) and readmissions (0.57, 0.42–0.77) compared with public. There was no difference in mortality or implant failure.

Discussion and Conclusions: We had no data on surgeon, general health and socioeconomic status of the patients. In addition, reported data from private clinics have not been validated in contrast to public hospitals

We found significant difference between patient characteristics operated at public versus private hospitals. No difference was evident regarding mortality and implant failure but for complications, reoperations and readmissions between private and public hospitals.