Abstract
Monitoring the performance of hip replacements post-operatively is tedious and costly, necessitating radiological examinations as well as other specialized examinations such as whole blood metal ion levels. In an effort to control escalating costs, we conducted an ethically approved clinical trial to assess the efficacy of basic acoustic monitoring equipment to asses these implants.
Method.
An electronic stethoscope was successfully used to record sounds from the hips of participants with different bearing surfaces. The sounds were recorded while conducting a standardized movement sequence. A 5th order Savitzky-Golay filter with a window width of 21 points was used to remove background noise. The recordings were also listened to by ear and three primary classes of sounds were identified. Frequency components contained in the classes were identified using spectrograms and Welch power density spectra. The sounds were correlated with different patient factors including component positioning, BMI and length of time that the implant was in situ. The skewness and kurtosis of the power spectra were calculated and found to be different for each class. Further frequency analysis was conducted with the aid of the discrete wavelet transform. This met with some success as different frequency levels were found in each sound class.
Results.
All bearing surfaces produced some noise. The most sounds were produced by the ceramic-on-metal group, even though not in the audible range, and those participants with a body mass index in the obese range. Sounds were also detected in the ceramic-on-polyethylene implants. However, no consistent links between these factors and the sounds produced could be identified. Specifically, the lack of correlation between sound occurrence and length of implantation indicates that this technique is not useful in predicting possible failures or future complications in real time.
Conclusions.
The sounds themselves did not immediately reveal any information about the implants. This method was deemed impractical as a real-time diagnostic technique. Our study though has demonstrated that inexpensive acoustic monitoring devices can monitor noise emissions. Our data needs to be refined to make these investigations reliable and clinically relevant.
