Our findings suggest that patients who underwent RA-TKA had a faster return to function compared to conventional TKA using IMU-based outcomes within the first six weeks after surgery. In particular, RA-TKA was associated with a quicker recovery as measured by a range of motion, impact load, bone stimulus, and earlier return of symmetrical limb function compared to conventional TKA. The differences between RA-TKA and conventional TKA were not noted using PROMs at any time point.
IMUs enable quantitative assessment of post-operative knee function using specific biomechanical metrics that have previously been limited to gait lab analysis7–11. These non-invasive wearable devices contain various sensors which combine together to allow estimation of the orientation of the sensor in space. as well as the impact load with each step (as assessed by the peak acceleration during foot contact). By measuring both operative and non-opeerative limbs, a patient's contralateral limb can be used as an internal control by providing an assessment of limb load symmetry. IMUs have been used previously to capture motion data to determine how a patient's surgical limb progresses after anterior cruciate ligament (ACL) surgery. Thomson et al.37 reported that among patients that underwent ACL reconstruction, IMU metrics were able to detect differences between the affected and unaffected limb despite patients meeting functional criteria to return to sport that could potentially place them at risk for re-injury.
These data present novel findings evaluating outcomes using wearable IMUs among patients undergoing conventional TKA and RA-TKA. A previous feasibility study by Bolam et al.22 reported that IMUs could be used in post-operative rehabilitation programs to track the impact load, bone stimulus, and impact asymmetry improvement over the early post-operative period in patients following knee arthroplasty. In this study, patients that underwent RA-TKA had improved ROM in the first six weeks after surgery compared to conventional TKA using IMU-based metrics. These results are consistent with previous studies that reported improved maximal knee flexion after RA-TKA compared to conventional TKA using traditional measures of range of motion38. In addition, we found patients undergoing RA-TKA progressed faster towards a symmetrical gait as measured by impact load IMU measurements. The increased bone impact loads detected with the IMUs among patients that underwent RA-TKA also resulted in improvements in the total load exposure, or bone stimulus score, which might encourage early osseointegration and reduced subsidence. However, longer-term follow-up is needed to determine when a symmetrical gait is achieved, as persistent limb loading asymmetry was noted at our final 6-week follow-up in this study. Anderson et al.39 reported using smartphone-based sensor data that symmetrical gait patterns were achieved at 13 weeks post-TKA. However, the study could not quantify IMU-specific metrics such as impact load and bone stimulus with smartphone-based sensors. Our findings demonstrate a large discrepancy in percent improvement in IMU-based metrics between the groups, suggesting that patients are able to tolerate more impact intensity after a robotic TKA than conventional instrumentation. One important explanation, and introduction of potential error, for this could be attributed to the differences in measured-resection or gap-balancing techniques and alignment philosophies among the three surgeons when performing conventional TKA. Despite the robotic technique being similar for all three surgeons, the conventional jig-based technique may have varied based on surgeon preference as described earlier.
Despite differences between IMU-based metrics, PROMs failed to detect any differences within six weeks of TKA when RA-TKA were compared to conventional TKA. These results are consistent with previous studies that have failed to report a difference when comparing PROMs between the patients that underwent conventional or RA-TKA40,41. Two recent prospective randomised controlled trials from Kim et al.40 and Song et al.41 reported no clinical difference between RA-TKA and conventional TKA when measured using traditional PROMS, including HSS, KSS, UCLA Activity Scale and WOMAC scores. Our findings suggest that IMU-based metrics are more sensitive than traditional PROMs in detecting clinically important post-operative recovery milestones, such as return to normal gait and impact loads through the surgical limb. While both measures can capture physical function, PROMs may miss important improvements in function in the early post-operative period due to the rudimentary questions. These IMU-based metrics have the potential to enable a more detailed understanding of patients' early post-operative progress to understand how various platforms, implants, or surgical techniques (i.e. alignment philosophies) influence patient outcomes. However, future research with longer follow-up is warranted to address whether subtle differences detected with wearable sensors are clinically meaningful.
This study has several limitations that must be considered when interpreting the findings. First, this study had an average follow-up of one year. Future studies with a longer follow-up interval of two or more years will allow us to monitor and report post-operative complications, including rates of revision, in the longer term and provide a more accurate assessment of the benefits or non-benefits of robotic TKA using wearable sensors. Second, while the large variability in the IMU dataset may be seen as a limitation, it reflects the real variability in patient mobility and their day-to-day activities, including the type and amount of activity performed by an individual on any given day. IMUs provide a measure of the variability in the activity levels of each patient, which traditional PROMs are not able to capture quantitatively. Third, this was a pragmatic study that did not blind patients to their surgery. However, IMUs provide objective functional measurements that should be less subject to measurement bias and over- or under-reporting of function.