Commentary & Perspective

Commentary & Perspective on
"Osteoporosis and Anterior Femoral Notching in Periprosthetic Supracondylar Femoral Fractures: A Biomechanical Analysis"
by Major Scott B. Shawen, MD, et al.

Commentary & Perspective by
Thomas D. Brown, PhD*,
Department of Orthopaedic Biomechanics,
University of Iowa Hospitals and Clinics, Iowa City, IA

It is widely recognized that no good can come of anterior femoral notching in total knee arthroplasty and that despite best efforts to leave the anterior cortex inviolate, notching unfortunately occurs in a minority of cases. Shawen et al. quite properly addressed the question of which variable(s) best predict supracondylar fracture when anterior femoral notching occurs.

Biomechanically, fractures occur when local bone stresses exceed local bone strength. For a given global load (torsion, in this particular experimental set-up), two factors influence the stress-to-strength ratio: first, the intrinsic mechanical competence of bone substance, and second, bone geometry. The authors have adopted essentially a logistic regression approach that incorporates several plausible measures of both of these factors. They reached the clear conclusion that the most discriminating combination of these measures are distal femoral bone-mineral density assessed with dual-energy x-ray absorpiometry (DEXA) and an estimate of the polar moment of inertia. From a biomechanical perspective, that combination is certainly intuitively correct.

Nevertheless, readers of The Journal should be cautious in relying on the seemingly high predictive power (r² = 0.79, p < 0.001) of this combination of measures. The authors understandably adopted a laboratory model that minimized confounding variables. A consequence of this experimental design, however, is that the investigators held constant what may in fact be the surgical variable to which fracture risk is most sensitive—the magnitude of mechanical compromise in the immediate vicinity of the notch. In the field of fracture mechanics, local details of geometry and/or discontinuities of material properties (i.e., "stress risers") determine the degree to which bone strength is compromised. Extrapolating that observation to the present surgical problem, one can see that all notches are not created equal. I believe that variability in the extent of notching probably would have overwhelmed the predictive capability of both distal femoral bone-mineral density as measured by DEXA and the polar moment of inertia, if the authors had chosen to include notch depth as an independent variable.

When one examines the authors' data on the polar moment of inertia, some questions arise. (This parameter drew my attention because, from a biomechanical perspective, these data seemed at least indirectly related to the relative severity of notching.) The values that Shawen et al. reported for the polar moment of inertia in their study were somewhat simplified according to a formula mentioned by Culp et al.¹. That formula assumes the negligible contribution of cancellous bone, an elliptical cortical cross-sectional geometry preoperatively, and the removal of anterior material by creation of a notch of specified depth t. Culp et al.¹ noted that their formula, which approximated the periosteal and endosteal surfaces as two concentric elliptical surfaces, became inapplicable when the notch depth exceeded the anterior cortical thickness, presumably because they realized that the notch then became an "open section," which is well recognized in elementary mechanics as resulting in a discontinuously abrupt reduction of torsional strength. Shawen et al. made a point of creating "full-thickness" anterior cortical defects, which calls into question the validity of their application of the formula used by Culp et al. (Even that formula itself is an extremely simplified method of calculating the polar moment of inertia, and, to my knowledge, it has not been compared with the gold-standard method of calculating the polar moment of inertia from the cross-sectional geometry of bone as measured on computed tomography scans. The former probably would not fare well by comparison.)

One other biomechanical issue bears mention so that readers can more fully interpret the findings of this study. The authors duly noted that the application of torsional loads in their study did not replicate realistic physiology in that their cadaver preparations did not include ligaments. Even that important factor aside, however, collecting valid data on the torsional strength of defects so close to the end of a long bone is very problematic because of a phenomenon that is known in mechanics as Saint Venant's principle. One frequent statement of St. Venant's principle is that stresses in the near neighborhood of a grip or a support are strongly determined by the geometry of that grip or support. As a "rule of thumb," measurements should be made at least one characteristic diameter away from the grip or support before this local effect due to clamping of the specimen diminishes. The polyester casting resin in which the specimens were embedded in the study by Shawen et al. was extremely close, according to Saint-Venant's principle, to the notch on the distal side. Therefore, in this particular experiment, variability from specimen to specimen in the nuances of distal casting resin would be expected to have exerted a very strong influence on the stresses that developed nearby at the all-important corner of the notch.

Given these methodological limitations, even if the application of functional loads had been a well-controlled variable (which they were not), it seems something of a stretch to suggest that fracture risk can be reliably predicted in the clinical setting by additional DEXA and computed tomographic studies analogous to those utilized in this cadaveric study. To be sure, it certainly makes very good sense to regard patients with anterior femoral notching as having an elevated risk of periprosthetic fracture. However, an equal or perhaps even greater prognostic purpose probably would be served—at considerably less expense—by the use of carefully oriented plain radiographs that clearly demonstrate the precise geometry of the notch, including the depth, to allow stratification of patients with anterior femoral notching according to their relative risk of fracture.

*The author did not receive grants or outside funding in support of the research or preparation of this manuscript. He did not receive payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the author is affiliated or associated.

References

1. Culp RW, Schmidt RG, Hanks G, Mak A, Esterhai JL Jr, Heppenstall RB. Supracondylar fracture of the femur following prosthetic knee arthroplasty. Clin Orthop. 1987;222:212-22.