Subsidence of tibial components in knee arthroplasty. A comparison between conventional radiography and roentgen stereophotogrammetry.

RATIONALE AND OBJECTIVES

To assess the ability of conventional radiography to identify subsidence of tibial components.

METHODS

Forty-three cases of knee arthroplasty were followed for 4 years using roentgen stereophotogrammetric analysis as the gold standard.

RESULTS

Roentgen stereophotogrammetric analysis showed subsidence in all uncemented prostheses, whereas two thirds of the cemented prostheses subsided. Medially, the sensitivity of conventional radiography was low, 25%, whereas the specificity was 83%. For the other regions, the sensitivity was almost 100%, although the specificity was as low as 66% frontally and dorsally.

CONCLUSIONS

Medially, where most prostheses were inserted with a slight overhang, conventional radiography identified subsidence correctly. Laterally, frontally, and dorsally, however, conventional radiography overestimated the subsidence by a factor of 2 to 3. This error was caused by bony protrusions arising in the cranial direction from the uncovered bone surfaces, thus changing the points of reference. This problem arose more commonly in uncemented cases, whereas cement seemed to seal the bone.