To compare the gait of unicompartmental knee arthroplasty (UKA)
and total knee arthroplasty (TKA) patients with healthy controls,
using a machine-learning approach. 145 participants (121 healthy controls, 12 patients with cruciate-retaining
TKA, and 12 with mobile-bearing medial UKA) were recruited. The
TKA and UKA patients were a minimum of 12 months post-operative,
and matched for pattern and severity of arthrosis, age, and body
mass index. Participants walked on an instrumented treadmill until their
maximum walking speed was reached. Temporospatial gait parameters,
and vertical ground reaction force data, were captured at each speed.
Oxford knee scores (OKS) were also collected. An ensemble of trees
algorithm was used to analyse the data: 27 gait variables were used
to train classification trees for each speed, with a binary output
prediction of whether these variables were derived from a UKA or
TKA patient. Healthy control gait data was then tested by the decision
trees at each speed and a final classification (UKA or TKA) reached
for each subject in a majority voting manner over all gait cycles
and speeds. Top walking speed was also recorded.Aims
Patients and Methods
Normal human locomotion entails a rather narrow base of support (BoS), of around 12cm at normal walking speeds. This relatively narrow gait requires good balance, and is beneficial, as it minimises the adduction moment at the knee. Normal knees have a slightly oblique joint line, and slight varus, which allow the normal human to walk rapidly with a narrow BoS. Patients with increased varus and secondary osteoarthritis have a broader BoS, which exacerbates the excessive load, making walking painful and ungainly. We wondered if there would be a difference between the base of support of patients whose knee kinematics had been preserved, by retaining the native jointline obliquity and the acl, in comparison with those whose alignment had been altered to a mechanically correct ‘neutral’ alignment. Of 201 patients measured following knee arthroplasty, 31 unicondylar patients and 35 total knee patients, with a single primary arthroplasty, and no co-morbidities, over 1 year post-operatively were identified. Two control groups of controls, a younger cohort of 112 people and 17 in an age matched older cohort. All operations were performed by the same surgeon. The total knees were cruciate retaining devices, inserted in mechanical alignment, and the unicondylar knees were inserted retaining the native alignment and joint-line obliquity. The gait of all subjects was analysed on an instrumented, calibrated treadmill with underlying force plates. Patients start by walking at a comfortable speed for them for 5 minutes, before the speed of the treadmill is increased at 1/2 km/h increments until maximum walking speed obtained, spending 30 seconds at each. After the flat test, it was then repeated on a downhill slope of 6°. Base of Support is interpreted as the distance between the centre point of heel strike and toe off from one foot to that of the other. The top walking speed in the unicondylar group was significantly greater than that of the total knee group, as we reported in 2013. TKA patients have an average BoS of 14cm, while UKA patients and controls have a 12cm BoS. The BoS did not reduce with speed. This 2cm, or 17% increase in BoS is significant. Shapiro-Wilk tests demonstrate a normal distribution to the results, and ANOVA testing reveals a significant difference (p<0.05) within the groups between the speeds of 4.5 to 9. Post-Hoc Bonferroni testing reveal a significant difference between the TKA group and each of the other three groups. On the downhill test (figure 1), the mean BoS in the TKA group increased to 16cm. This increase is highly significant, with a p value of <0.001, while the increase in the UKA group at higher speeds failed to reach significance, and the controls both stayed at 12cm. 6 Bi-uni knees tested acted just like the UKAs. A narrow base of support minimises excessive loads across the joint line. Maintenance of jointline obliquity and an ACL enables this feature to be returned to normal following uni, or bi-uni, while a well aligned TKA seems to prevent it.Materials and Methods
Discussion
Any arthroplasty that offers superior function needs to be assessed using metrics that are capable of detecting those functions. The Oxford Hip Score (OHS), the Harris Hip Score (HHS) and WOMAC are patient reported outcome measures (PROMs) with well documented ceiling effects: following hip arthroplasty, many patients are clustered close to full marks following surgery. Two recent well conducted randomised clinical trials made exactly this error, by using OHS and WOMAC to detect a differences in outcome between hip resurfacing and hip arthroplasty despite published data already showing in single arm studies that these two procedures score close to full marks using either of these PROMS. We have already reported that patients with hip resurfacing arthroplasty (HRA) were able to walk faster and with more normal stride length than patients with well performing hip replacements. In an attempt to relate this functional superiority to an outcome measure that does not rely upon the use of expensive machinery, we developed a patient centred outcome measure (PCOM) based upon a method developed by Philip Noble's group, and the University of Arizona's Metabolic Equivalent of Task Index (MET). This PCOM allows patients to select the functions that matter to them personally against which the success of their own operation will be measured, with greater sensitivity to intensity than is achieved by the UCLA. Our null hypothesis was that this PCOM would be no more successful than the PROMs in routine use in discriminating between types of hip arthroplasty, and that there would be no difference in gait between patients following these procedures. From our database of over 800 patients whose gait has been assessed in the lab, we identified 22 patients with a well performing conventional THAs, and matched them for age, sex, BMI, height, preop diagnosis with 22 patients with a well performing conventional THA. Both were compared with healthy controls using the novel PCOM and in a gait lab. PROMs for the two groups were almost identical, while HRA scored higher in the PCOM. The 9% difference was significant (p<0.05). At top walking speed, HRA were 10% faster, with a 9% longer stride length, both of these metrics also reached significance. Function following hip replacement is very good, with high satisfaction rates, but the use of a PCOM, and objective measures of function reveal substantial inferiority of THA over THR in two well matched groups. This 9% difference is well over the 5% difference that is considered ‘clinically relevant’. When coupled with the very strong data regarding life expectancy and infection, this functional data makes a compelling case for the use of resurfacing in active adults.Results
Discussion
The Oxford Hip Score (OHS), the Harris Hip Score (HHS) and WOMAC are examples of patient reported outcome measures (PROMs) have well documented ceiling effects, with many patients clustered close to full marks following arthroplasty. Any arthroplasty that offers superior function would therefore fail to be detectable using these metrics. Two recent well conducted randomised clinical trials made exactly this error, by using OHS and WOMAC to detect a differences in outcome between hip resurfacing and hip arthroplasty despite published data already showing in single arm studies that these two procedures score close to full marks using both PROMS. We had observed that patients with hip resurfacing arthroplasty (HRA) were able to walk faster and with more normal stride length than patients with well performing hip replacements, but that these objective differences in gait were not captured by PROMs. In an attempt to capture these differences, we developed a patient centred outcome measure (PCOM) using a method developed by Philip Noble's group. This allows patients to select the functions that matter to them personally against which the success of their own operation will be measured. Our null hypothesis was that this PCOM would be no more successful than the OHS in discriminating between types of hip arthroplasty. 22 patients with a well performing Hip Resurfacing Arthroplasty were identified. These were closely matched by age, sex, BMI, height, preop diagnosis with 22 patients with a well performing conventional THA. Both were compared with healthy controls using the novel PCOM and in a gait lab. PROMs for the two groups were similar, while HRA scored higher in the PCOM. The 9% difference was significant (p<0.05). At top walking speed, HRA were 10% faster, with a 9% longer stride length. Outcome measures should be able to detect differences that are clinically relevant to patients and their surgeons. The currently used hip scores are not capable of delivering this distinction, and assume that most hip replacements are effectively perfect. While the function of hip replacements is indeed very good, with satisfaction rates high, objective measures of function are essential for innovators who are trying to deliver improved functional outcome. The 9% difference in PCOM found in this small study reflects the higher activity levels reported by many, and of similar magnitude to the 10% difference in top walking speed, despite no detectable difference in conventional PROMS. PCOMs may offer further insight into differences in function. For investigators who wish to develop improvements to hip arthroplasty, PCOMs and objective measures of gait may describe differences that matter more to patients than conventional hip scores.Results
Discussion
Operative and non-operative treatment regimens for Achilles tendon ruptures vary greatly but commonly involve rigid casting or functional bracing. The aim of our study was to investigate the extent of tendon apposition following such treatments. Twelve fresh-frozen, adult below knee lower-extremity cadaveric specimens with intact proximal tibiofibular joints were used. Each was prepared by excising a 10cm × 5cm skin and soft tissue window exposing the Achilles tendon. With the ankle in neutral position, the tendon was transfixed with a 2mm k-wire into the tibia, 8cm from its calcaneal insertion. A typical post-rupture gap was created by excising a 2.5cm portion of tendon between 3.5cm and 6cm from its calcaneal insertion. The specimens were then placed into a low profile walker boot (SideKICKTM, Procare) without wedges and a window cut into the back. The distance between the proximal and distal Achilles tendon cut edges was measured and repeated with 1, 2 and 3 (10mm) wedges. Subsequently the specimens were placed into a complete below knee cast in full equinus which was also windowed. The Achilles tendon gap (mean +/− SD) measured: 2.7cm (0.5) with no wedge, 2.3cm (0.4) with 1, 2.0cm (0.4) with 2, 1.5cm (0.4) with 3 wedges and 0.4cm (0.3) in full equinus cast. The choice of treatment had a significant effect on tendon gap (p< 0.0001 – repeated measures ANOVA), and all pairwise comparisons were significantly different (Bonferroni), with all p< 0.001, apart from 0 wedge vs. 1 wedge (p< 0.01) and 1 wedge vs. 2 wedges (p< 0.05). Our results showed that each wedge apposed the tendon edges by approximately 0.5cm with the equinus cast achieving the best apposition. Surgeons should consider this when planning appropriate immobilisation regimes for Achilles tendon ruptures.
During surgical reduction of ankle injuries with syndesmotic instability surgeons often use the anteroposterior (AP) and mortise radiographs to assess reduction. Current literature predicts 50% are malreduced mainly in the sagittal plane. Our aim was to develop a radiographic measure based on the lateral view to assess both the normal and abnormal fibula/tibia relationship after simulated syndesmotic malreduction and to evaluate the effect on commonly used AP and mortise measurements. Nine fresh-frozen cadaveric specimens were dissected to the level of the syndesmosis. AP, mortise and talar dome lateral radiographs were obtained before and following syndesmosis division and posterior fibula displacement. On the lateral radiograph a line was drawn (Orthoview) from the anterior border of the fibula bisecting a line drawn from the anterior to posterior lips of the distal tibia. The ratio of the anterior-posterior segments was calculated. Also a line was drawn from the posterior border of the fibula and the distance was measured to the posterior lip of the tibia. At 0, 2, 4 and 6mm of displacement the ratio measured 1.3±0.2, 1.1±0.2, 0.9±0.2 and 0.7±0.2 respectively with all pairwise comparisons being significantly different. Inter- and intra-observer variability varied from substantial to perfect. The only significant medial clear space (MCS) difference was on the mortise view between 0mm (2.0±0.3mm) and 6mm (2.4±0.4mm) displacement. Our new measure of syndesmotic reduction is reproducible and can detect from 2mm of saggital fibular displacement. At maximum fibular displacement the increase in MCS was less than 1mm. This demonstrates standard mortise radiographs are poor at detecting syndesmotic reduction. An interesting observation was in all specimens prior to any displacement, the posterior fibular line always bisected the posterior lip of the tibia or lay just anterior to it, never posterior. This could serve as a useful adjunct for surgeons when assessing syndesmotic reduction intra-operatively.
The surgery results parallel those reported in other trials. At two years, the treatment costs of the surgery arm were approximately twice those of the rehabilitation arm. The costs of the rehabilitation strategy depend on how many patients opt to have surgery after rehabilitation (22% in this trial). This is a comparison of treatment strategies: there was no clinical or statistical difference in outcome between the strategy of spinal fusion and rehabilitation.
