Session I - Pelvic Trauma

Development of an In-Vivo Model of Acetabular Fractures

Steven A. Olson, MD; Brian K. Bay, PhD; University of California, Davis Medical Center, Sacramento, CA; Scott Smith, MD, University of Tennessee Medical Center, Knoxville, TN; Andrew J. Hamel, BSEE; Neil A. Sharkey, PhD; Pennsylvania State University, University Park, PA

Hypothesis: An in-vivo survival animal model of a displaced articular fracture (dorsal wall of acetabulum) will show significant differences in articular degeneration with and without anatomic reduction.

Methods: Eight adult Nubian goats underwent general anesthetic and had a posterior approach to the right hip. A 1cm x 1cm simulated dorsal (posterior) wall fracture was created by initially predrilling the retroacetabular surface adjacent to the acetabular rim. The animal was then turned to a supine position, and a 10kg mass was dropped 1m onto the ipsilateral knee, in line with the femoral shaft. An accelerometer in the proximal femur recorded the force impulse of this blunt impact. In each case, the fracture was completed with the impact of the femoral shaft creating a blunt chondral injury, in addition to the dorsal wall acetabular fracture.

In four animals, the dorsal wall fracture underwent open reduction internal fixation, and in four animals the dorsal wall fracture was left unreduced and unfixed. In each animal, the contralateral lower extremity was used as a control. The animals were maintained for 90 days and then were killed. Following death of the animal, radiographs were obtained of the hip joints. The pelvis and femoral shafts were harvested and underwent biomechanical analysis of hip joint loading. The pelves were mounted by potting the sacrum in methacrylate to a special jig, to attach to the Instron load cell. The distal femur was similarly potted in methacrylate, and the pelvis and femur were articulated to simulate quadruped locomotion as outlined by Page, et al. Low sensitivity Fuji film was placed on the femoral head and used to measure contact area as well as contact pressure within the hip joint surface. Standardized placement of holes was drilled in the acetabulum to use for referent markers for Fuji film loading. Fuji film was digitized for analysis. Following mechanical testing of both the experimental and control hips, the acetabuli were explanted from the pelvis and placed in a fixative for histologic analysis. The junction of the posterior wall fracture and intact acetabulum at the superior margin of the fracture defect was used for histologic analysis with H&E and Saphranin-O staining. The extent of articular degeneration was assessed histologically using the Mankin rating scale by two independent reviewers.

The biomechanical analysis was performed using repeated measures ANOVA with Bonferoni adjustments. The histologic data was performed with repeated measures ANOVA statistical analysis.

Results: The proximal femoral accelerometer recorded a mean 80G force pulse, reflecting a minimum of 40 MPa blunt impact on the hip joint. All animals survived the 90 day experimental period. There were no wound complications. There were no hip dislocations. All animals healed their fractures at 90 days.

Biomechanical analysis comparing the fractured hip to the contralateral control hip did not reveal a statistically significant alteration in contact area or contact pressure in either the reduced or nonreduced groups.

Histologic analysis demonstrated that there was a statistically significant difference in Mankin grades between the control and fractured sides (p<.01). There was substantially more hypercellularity in cloning in the articular surface of the experimental hip, with or without reduction, as compared to the control hip. There was a nearly significant effect of articular reduction (p=.06) on the extent of articular degeneration. There was no evidence of chondrocyte or osseous necrosis in any of the experimental animals. There was an increased joint space narrowing and subchondral sclerosis in the animals with articular displacement as compared to the reduced animals.

Discussion: Displaced acetabular fractures are associated with the development of posttraumatic arthritis. Biomechanical studies in human cadaveric hips have demonstrated that posterior wall acetabular fractures alter load transmission across the hip joint and result in increased articular contact pressures in the superior acetabulum. This type of biomechanical analysis addresses the alteration of hip joint loading as the result of an acute fracture. Canines have traditionally been used for hip arthroplasty models. Despite quadruped locomotion, these models have been shown to provide an adequate experimental environment to evaluate hip joint mechanics. However, because of the prevalence of hip osteoarthritis in the large canines, we have used the Nubian goat an animal with similar quadruped gait mechanics for our experimental model.

The goal of this project was to develop an in-vivo model of acetabular fractures to investigate the long-term effects of these injuries as healing and articular remodeling occur. In previous work we have demonstrated with a benchtop model that fracture of the posterior wall of the Nubian goat acetabulum results in significant increase in superior (dorsal) wall contact pressures.

Conclusions: This project demonstrates the successful development of an in-vivo survival model of complex articular injuries. Because the fractures were created with impaction of the femoral shaft, the actual articular defect was smaller than anticiapted. This combined with an aggressive healing response resulted in a lack of significant biomechanical alterations and joint loading. However, histologic analysis demonstrated that: (1) The combination of articular fracture and blunt chondral injury results in a significant increased cellularity and cloning of the articular surface as compared to the contralateral control; (2) No chondrocyte or chondral necrosis was seen at 90 days following a substantial blunt impact (40 MPa); (3) Based on this pilot project with four animals per experimental group, a nearly significant difference in histologic appearance of the articular surface was evident as a result of articular reduction. The power analysis demonstrates that future work with this model should require five animals per group to ensure a sufficient statistical sample size.