Failure mode classification for tumor endoprostheses: retrospective review of five institutions and a literature review.

BACKGROUND: Massive endoprostheses provide orthopaedic oncologists with many reconstructive options after tumor resection, although failure rates are high. Because the number of these procedures is limited, failure of these devices has not been studied or classified adequately. This investigation is a multicenter review of the use of segmental endoprostheses with a focus on the modes, frequency, and timing of failure.
METHODS: Retrospective reviews of the operative databases of five institutions identified 2174 skeletally mature patients who received a large endoprosthesis for tumor resection. Patients who had failure of the endoprosthesis were identified, and the etiology and timing of failure were noted. Similar failures were tabulated and classified on the basis of the risk of amputation and urgency of treatment. Statistical analysis was performed to identify dependent relationships among mode of failure, anatomic location, and failure timing. A literature review was performed, and similar analyses were done for these data.
RESULTS: Five hundred and thirty-four failures were identified. Five modes of failure were identified and classified: soft-tissue failures (Type 1), aseptic loosening (Type 2), structural failures (Type 3), infection (Type 4), and tumor progression (Type 5). The most common mode of failure in this series was infection; in the literature, it was aseptic loosening. Statistical dependence was found between anatomic location and mode of failure and between mode of failure and time to failure. Significant differences were found in the incidence of failure mode Types 1, 2, 3, and 4 when polyaxial and uniaxial joints were compared. Significant dependence was also found between failure mode and anatomic location in the literature data.
CONCLUSIONS: There are five primary modes of endoprosthetic failure, and their relative incidences are significantly different and dependent on anatomic location. Mode of failure and time to failure also show a significant dependence. Because of these relationships, cumulative reporting of segmental failures should be avoided because anatomy-specific trends will be missed. Endoprosthetic design improvements should address failure modes specific to the anatomic location.