The possibilities of uncemented glenoid component--a finite element study.

OBJECTIVE

(1) Determine the initial stress distributions within an uncemented implanted glenoid during elevation of the arm and to investigate whether failure is caused by stresses generated within this implant-bone structure. (2) Compare stress patterns between the uncemented design and two basic models of cemented prostheses.

DESIGN

The uncemented component consists of a polyethylene cup with a metal-backing. All material interfaces were assumed to be fully bonded.

BACKGROUND

Cemented glenoid components have been frequently vulnerable to failure within itself and at the cement-bone interface. A 3-D finite element analysis of an uncemented design is required to investigate whether clinical observations on failure can be better explained with a stress analysis.

METHODS

A 3-D finite element submodel an uncemented prosthesis was generated using CT-scan data and realistic loading conditions (humeral abduction, 30-180 degrees ). The submodelling approach was based on an overall solution of a complete scapula acted upon by all muscles, ligaments and joint reaction forces.

RESULTS

High Von Mises stresses (20-70 MPa) were generated in the metal-backing during abduction. Stresses were reduced in the polyethylene cup by 17-20% as compared to the cemented designs. Stresses in the underlying bone were substantially lower than to the natural glenoid. Stress-shielding can be observed in the trabecular bone underlying the prosthesis. The implant-bone interface is secure against interface failure at moderate loads, although the implant-bone (metal-bone) interface around the superior edge of the prosthesis is subject to high stresses (normal: 11.85 MPa, shear: 6.67 MPa) as compared to the cemented prosthesis. Whereas, the cement-bone interface, appears more likely to fail either at locations adjacent to the keel or at locations around the superior edge of the cemented design. The uncemented design therefore appeared to be a reasonable alternative to fixation with cement.

RELEVANCE

Results of this study strongly agree with clinical and radiographic findings. Although indications of stress-shielding and separation of modular parts of the prosthesis were apparent, the implant-bone interface seems less likely to fail as compared to cemented designs. Once initial fixation of the implant is achieved, the uncemented design appears to have better prospects than cemented ones.