CORR Insights®: Is Anterior Rotation of the Acetabulum Necessary to Normalize Joint Contact Pressure in Periacetabular Osteotomy? A Finite-element Analysis Study

Where Are We Now?

Developmental dysplasia of the hip (DDH) encompasses a wide spectrum of disorders ranging from a complete fixed dislocation at birth to asymptomatic acetabular dysplasia in the adult [14]. At its core, DDH is a condition of mechanical instability, characterized by inadequate coverage or containment of the femoral head. Associated femoral pathomorphologies, such as coxa valga, excessive femoral anteversion, and cam morphology (aspherical femoral head with reduced head-neck offset) are common in patients with DDH [5, 16].

Altered biomechanics are the primary etiology of noninflammatory joint osteoarthritis (OA). The prevailing theory among researchers is that the damage found in hips with dysplasia results in increased cartilage contact stresses [6, 9, 12]. My own study group’s finite-element models [7] yielded different results, where we found that cartilage contact stresses did not increase in patients with DDH. However, we included the labrum in our finite-element models, which was excluded from previous finite-element models of patients with dysplasia. Inclusion of the labrum provided an additional contact surface through which load could be transferred across the joint interface, thus reducing the necessary support by the cartilage (and hence, the resulting contact stress). Results from our finite-element models support the concept that OA occurs in an “outward-to-in” fashion in patients with DDH. That is, the damage begins at the periphery of the joint, where the labrum is located, and progresses medially to the cartilage over time [4].

Periacetabular osteotomy (PAO) can be used to treat DDH; if it is to be employed, it is best performed before a patient develops clinically important articular cartilage loss. This procedure increases coverage of the femoral head and reduces instability at the joint through surgical reorientation of the acetabulum. While it reduces pain and increases function in most patients, around 40% of patients treated with PAO nonetheless undergo hip arthroplasty [13], which suggests that PAO may not be effective at improving hip joint biomechanics to the level necessary to stave off OA in all treated patients.

In the current study, Kitamura and colleagues [11], using finite-element modeling, found that simulated anterior rotation of the acetabulum brought cartilage contact stresses in the cohort of DDH patients closer to model predictions from a similarly-aged control group with normally-developed hip morphology. The authors suggest that DDH patients with greater posterior pelvic tilt, as indicated by an anterior center-edge angle < 32°, may require additional anterior rotation of the acetabulum to improve cartilage contact mechanics. Collectively, their findings point to the importance of considering physiologic pelvic tilt when planning acetabular reorientation and highlight the need for the attending surgeon to evaluate how coverage is deficient regionally on a per-patient basis.

Where Do We Need To Go?

Biomechanical modeling coupled with surgical navigation may someday be able to improve clinical outcomes after PAO by providing the surgeon with a plan for acetabular reorientation that is customized to the anatomy and expected hip biomechanics of the individual patient. However, there is a risk that the position predicted by the computer model is incorrect, which could have serious, negative clinical consequences. This is one of the primary reasons why surgeons may be hesitant to use and embrace PAO planning/navigation systems. Additional research that verified the predictive value of these preoperative planning systems would help to allay these concerns.

A number of factors—not merely reducing cartilage contact stress—must be considered when we try to determine the target alignment of the acetabulum during PAO. For example, it is possible that the minimization of cartilage contact stress as well as the load to the labrum [7] could produce a surgical plan that facilitates improved clinical outcomes. Additionally, the best position of the acetabulum likely depends on the loading scenario that is applied in the finite-element model. In the current study [11], the authors investigated one loading scenario (single-leg stance) and used the same applied load across all models. However, the magnitude and orientation of the total force acting across the joint is influenced by the mechanics of individual muscles, which are participant- and patient-specific, and depend on the particular activity that is performed. For this reason, future research should consider the role of altered muscle mechanics on cartilage and labrum mechanics. Additionally, future modeling research should evaluate how various predicted acetabular reorientations affect available hip ROM. This is important because excessive anterior rotation of the acetabulum during PAO can lead to iatrogenic femoroacetabular impingement. As with the finite-element models, researchers should include model anatomy for the acetabular labrum in these simulations since the region where the labrum contacts the femur is generally not the same region as where the femur and acetabulum bone surfaces approximate one another (thus, results from bone-to-bone collision models may not be physiologically realistic) [10].

Finally, we cannot underestimate the importance of including accurate, patient-specific anatomy reconstructions in finite-element models used for planning of PAO. Even small undulations, ridges, and hills in the topology of the subchondral bone and cartilage have a major impact on the magnitude and distribution of stresses to cartilage and bone. Researchers should model cartilage with site-specific thickness values when analyzing hip contact mechanics with finite-element analysis since assuming constant thickness will lead to less accurate predictions [2]. Also, femoral cartilage is typically thicker medially than it is laterally, with the opposite holding true for acetabular cartilage. In this way, thinner acetabular cartilage interacts with thicker femoral cartilage at the contact interface (and vice versa). PAO surgery may disrupt this native relationship, which could affect cartilage contact stresses; this effect cannot be ascertained if constant thickness cartilage is assumed.

How Do We Get There?

Including the acetabular labrum, incorporating site-specific cartilage thickness, and considering the dynamic motion of the hip will add complexity to biomechanical models that are used to determine the best alignment of the acetabulum. Including these structures may initially decrease the sample size of studies that evaluate the biomechanical and clinical efficacy of PAO planning and surgical navigation technologies. As such, it is important to evaluate systematically how inclusion of these patient-specific factors influences predictions on the best reorientation of the acetabulum. When the added benefit is not apparent, then investigators can eliminate that particular attribute from the model with confidence.

The ability to reconstruct anatomy of the labrum, cartilage, and bone accurately relies heavily on obtaining high-quality multiplanar images. In my experience, I have found it difficult to distinguish the boundary between femoral and acetabular cartilage throughout the entire articulating surface, even when arthrography was performed. Including traction during the CT or MRI scans substantially improves differentiation of acetabular and femoral cartilage [8], which in turn, makes it more conducive to apply automatic thresholding during segmentation (which saves considerable time). Still, multiplanar images can be time-consuming and laborious to segment. Further, arthrography may not be practical for all patients and clinics that utilize PAO planning systems. One potential solution is to utilize a statistical atlas, which includes information on geometry and its variation across populations, to provide reasonable estimates of site-specific thickness to go along with reconstructions of the bone anatomy [15].

Finally, we should build upon prior research that applied discrete-element analysis to identify the optimal reorientation of the acetabulum for PAO [3]. With discrete-element analysis, bones are reconstructed from multiplanar images and are then modeled as rigid bodies. Cartilage is represented as an array of springs on the bone surface with user-defined spring lengths to represent cartilage thickness. The primary advantage of discrete-element analysis is that stresses can be calculated in a matter of seconds, compared to a finite-element model, which can take several hours or more for a single loading scenario. Given its computational efficiency, discrete-element analysis could be incorporated directly into the navigation system, providing near real-time feedback as to how the distribution of stresses to cartilage and labrum change as a result of various reorientations of the acetabulum. The primary disadvantage of discrete-element analysis is that deformation in the lateral direction of cartilage is not accounted for, which has the effect of overestimating cartilage contact stresses. Nevertheless, when site-specific cartilage thickness is assigned to the springs, discrete-element analysis models predict cartilage contact stress distributions in excellent agreement with more sophisticated, patient-specific finite-element models [1]. Since the major goal of PAO is to redistribute contact stresses away from the acetabular rim, the magnitude of change in stress resulting from acetabular reorientation is likely of less clinical importance than the distribution of the contact stresses. Future research could compare results between discrete-element and finite-element models over a wide range of loading scenarios and hip anatomies to clearly understand the potential implications of using this computationally-efficient approach for PAO planning and navigation.