The prediction of collapse progression has been the key question in treating ONFH patients. Evolvement of classification systems and techniques in interpreting imaging studies were directed to elucidate such inquiry. Thus, indications of each treatment options including conservative care, joint preserving techniques, or arthroplasties have been widely studied, with a goal of basic consensus to preserve one’s natural hip joint as long as possible. In a recent study based on U.S. nationwide database, Sodhi et al. reported rates of arthroplasty (94.03%) were far greater than those of other procedures including osteotomy, partial arthroplasty, core decompression, and bone graft.[14] However, in young population, joint preserving techniques should be considered in prior to arthroplasties due to the latter’s higher complication rate, invasiveness, and implant life span which could lead to revision surgery.[4, 14, 15]
In particular, biomechanical significance of the lateral pillar had been suggested in preventing collapse,[9, 10, 16] but to date, cohort studies of which were scarcely reported.[11] Thus we designed a study of propensity-score matched cohorts to further minimize selection bias prone in retrospective data analysis. Additionally, FEM analysis was selected to investigate the fact that not all osteonecrosis placed in the lateral trisection undergo collapse, presumably due to dynamic basis rather than its sole mechanical structures. Hence, emphasis of this study is put on the phenomenon that not the absolute force on the lateral pillar but the presence of stress concentration on the critical areas determines collapse progression. Also, in this setting, it looks impracticable to set an absolute cutoff value in predicting collapse, in that it may vary under various local and global conditions among different individuals’ hip joints.
For demographics, age (p=0.958), sex (p=1.000), and as well as BMI (p=0.944) between the two groups had no significant difference as propensity scoring was estimated for age, sex, and BMI to minimize confounders. For pre-analytic comparison of characteristics of necrotic areas, location was classified as ‘lateral’ or ‘non-lateral’, where the latter includes centrally or medially located lesion based on the significance of the lateral pillar. All necrotic lesions in the collapsed group had pre-collapse lesions in the lateral trisection (n=16), while lesions in the non-collapsed group were located half in the lateral (n=8) and the other half in the non-lateral (n=8) trisections. As a result, location (p=0.015) and size (p=0.015) had significant difference, which follows the current understandings of intrinsic risk factors of necrosis progression.
Upon the hypothesis of stress concentration on the lateral pillar might accelerate collapse, FEM analysis was conducted. To reflect extreme forces that can be applied to the hip joint, arbitrary stress load of 4,500N was set considering that peak 870% of body weight can be applied when stumbling in a 53kg weighing individual. [17] As a result, stress concentration was focused on the lateral pillar in 87.5% of hips in the collapsed group while stress dispersion through the pillars were observed in 81.2% of the hips in the non-collapsed group (p=0.001). Additionally, when stress concentration to the lateral pillar was present, distal force transmission through the primary compression trabeculae were always coupled. That is, stress concentration converging to the vertical axis would predict near-future collapse with high probability, whereas stress dispersion through medial (inferior) and lateral (superior) cortices of proximal femur is crucial in maintaining support of the anatomical structure. Therefore, the importance of structural support is emphasized not only at the lateral pillar, but inevitably also the primary compression trabeculae owing to its extended transmission of the yield stress distally.
Limitations of this study includes small sample size of 16 patients per cohort, hence propensity-score matching was used to reinforce clinical significance given the small size in this retrospective case cohort study. Second, this study lacks quantitative analysis via such as von Mises stress[9, 11], stress index[7], or value of equivalent stress.[8] But as stated above, emphasis of this study is put on the qualitative cognition on the stress distribution rather than suggesting quantified cutoff values. In turn, proving effects of reinforced support to the lateral pillars using buttresses such as with fibular strut grafts or tantalum rod implantation would be analyzed through FEM in the future studies.
In conclusion, FEM analyses of followed-up ONFH suggest stress concentration to the lateral pillar and the primary compression trabeculae predicts collapse with high probability. Graphical output as an end result in pre-collapse ONFH provides a simple and intuitive, but yet valuable information to aid surgeons in treatment selection. Moreover, FEM generation is achieved easily using ordinary CT data of a patient, depicting stress distribution which can be recognized at a glance. Therefore, especially for young patients, holding out the lateral pillar and the primary compression trabeculae through joint preserving procedures might be the key in preventing further collapse of the femoral head.