Abstract
Introduction
Iatrogenic proximal femur hoop-stress fracture is a recognised complication of uncemented hip arthroplasty. It has a reported incidence of two to three percent and increases patient morbidity. We describe a novel technology that predicts fracture in real-time by less than one minute.
Method
Four proximal femora from red deer (Cervus elaphus), similar size to human proximal femora, were prepared to accept an uncemented hip arthroplasty femoral rasp (Finsbury Orthopaedics) and then mounted in a loading machine. The femora were fresh-frozen, defrosted and kept at room temperature in 0.9% saline swabs. The rasp was forced into each femur in repeated loading cycles every 10 seconds, in steps of 100N increasing from 200N to over 2000N until fracture, in a manner to simulate surgery. One sensor was attached to the surface of the proximal femur and one to the femoral rasp. The sensor outputs were recorded, analysed and displayed on a PC using a software algorithm to show signal energy (joules) and amplitude (decibels). The proximal femur was coated with specular marking paint to permit real-time 3-D digital image correlation (DIC) analysis. DIC is an established tool in engineering fracture analysis and utilises two spatially orientated video cameras to measure surface strain and fracture. The femur was observed by the human eye and loaded in cycles until a fracture was seen. The moment of fracture was marked in the recording timeline. DIC was used to confirm fracture.
Results
All femora fractured in the anterior proximal cortex. Signals from both sensors were identical in form and differed by less than five percent in strength during loading. Both signals demonstrated significant increases in energy and amplitude shortly prior to fracture. Early during loading cycles the femoral rasp subsided and became well-fixed within the femur; this was associated with signals of 60-70dB. During later loading cycles the rasp ceased to subside in the femur and was well-fixed in a press-fit; subsequent loading caused fracture and this was preceded by a greater number of stronger signals of over 90dB. The increase occurred 1 to 3 loading cycles prior to fracture, or less than 30 seconds. DIC was used to confirm the presence of a fracture visible to the human eye. At the time of the first significant increase in signal there was no crack visible to the eye or to DIC analysis and the femoral rasp did not subside further into the femur.
Conclusions
During press-fitting of an uncemented femoral rasp in a deer femur a significant change in signal characteristics occurs shortly prior to a fracture being visible to the eye and detectable by DIC analysis. The almost identical signal output from both sensors suggests that one single sensor mounted on a femoral rasp will suffice, thereby removing the need to expose more of the proximal femur during surgery. This technology may be able to predict and therefore prevent femur fracture during uncemented hip arthroplasty. Further research is necessary in animal and human cadavers to explore and validate this research.
