Abstract
Introduction
In cementless THA the incidence of intraoperative fracture has been reported to be as high 28% [1]. To mitigate these surgical complications, investigators have explored vibro-acoustic techniques for identifying fracture [2–5]. These methods, however, must be simple, efficient, and robust as well as integrate with workflow and sterility. Early work suggests an energy-based method using inexpensive sensors can detect fracture and appears robust to variability in striking conditions [4–5]. The orthopaedic community is also considering powered impaction as another way to minimize the risk of fracture [6– 8], yet the authors are unaware of attempts to provide sensor feedback perhaps due to challenges from the noise and vibrations generated during powered impaction. Therefore, this study tests the hypothesis that vibration frequency analysis from an accelerometer mounted on a powered impactor coupled to a seated femoral broach can be used to distinguish between intact and fractured bone states.
Methods
Two femoral Sawbones (Sawbones AB Europe, SKU 1121) were prepared using standard surgical technique up to a size 4 broach (Summit, Depuy Synthes). One sawbone remained intact, while a calcar fracture approximately 40mm in length was introduced into the other sawbone. Broaching was performed with a commercially available pneumatic broaching system (Woodpecker) for approximately 4 secs per test (40 impactions/sec) with hand-held support. Tests were repeated 3 times for fractured and intact groups as well as a ‘control’ condition with the broach handle in mid-air (ie not inserted into the sawbone).
Two accelerometers (PCB M353B18) positioned on the femoral condyle and the Woodpecker impactor captured vibration data from bone-broach-impactor system (Fig1).
Frequency analysis from impaction strikes were postprocessed (Labview). A spectrogram and area under FFT (AUFFT) [4] were analysed for comparisons between fractured and intact bone groups using a nested ANOVA.
Results
Vibration frequency patterns between respective groups were best observed using an accelerometer positioned on the impaction device rather than on a sawbone (fig1). Qualitative assessment revealed that spectrograms showed no obvious difference for characteristic vibration frequencies between intact and fractured bone groups. A frequency signal at approximately 10kHz was absent for control impactions but present with bone impactions (Fig2). Quantitative assessment revealed AU-FFT was noticeably higher for intact bone groups than fractured bone groups for sampled impactions using a nested experimental design for statistics (p=0.11).
Discussion
Our pilot study demonstrates that application of powered impaction combined with vibration frequency analysis has the potential to distinguish between an intact and fractured sawbone in a way that minimises instrumentation footprint and complexity of workflow in OR with a new generation of impaction device targeted at reducing and detecting bone fractures. Further investigation should validate these methods by evaluating the variation with sawbones and simulated bone fractures.
