Journal of NeuroEngineering and Rehabilitation

The SNS median for TFC (1.11) compares well with the median SNS that can be calculated from the results reported in [ 3 ] (1.09). Furthermore, our results can be compared with the study of Nolan et al. [ 2 ] , that involved 4 transfemoral and 4 transtibial amputees using a single hinge knee and a SACH foot. Once appropriately converted to our indexes, Nolan ’ s results are reported in Table 4 . Results, can also be compared with Bateni et al. [ 41 ] , which reported a mean stance asymmetry for TT of about 7% (calculated as the ratio of the mean between sides). Compared to these studies, our SNS values are lower. In particular, 63% of TT and 20% of TFC have a SNS lower than ±5%, which makes them unperceived by others as “ impaired ” walkers with regard to temporal symmetry [ 21 ] . This is not surprising given the different prosthetic components used and the fact that our patients followed a specialized rehabilitation training. Our SNS results for TT are also in very good agree- ment with results reported by Jarvis et al. [ 38 ] for young veterans (median 1.04, IQR = 0.03). For TFC, our SNS is higher (1.11 compared to 0.98) but the IQR is much smaller (0.09 compared to 0.20). This remarks that the training for transfemoral amputees is more challenging. When looking at IMS, the TFM median was statisti- cally different from TFC and TT: TFM asymmetry is twice that of TFC and 16 times TT ’ s. The comparison with Nolan et al. [ 2 ] is striking: our TFM had an impulse asymmetry which is half Nolan ’ s; for TT it is 10 times less. This result points, again, in the direction of the benefits of energy-storage-and-return feet and more advanced knees. Improvement in loading asymmetry with energy-storage-and-return feet and feet with improved roll-over shape has been previously reported in [ 25 , 27 , 42 ] , and match well with simulation studies [ 8 ] . Finally, P1S results show statistically significant differ- ences between TF and TT (Fig. 7 ). About 59% of TF have a higher peak on the prosthetic side . Our results agree with Castro et al., which did not report an increased peak GRF on the sound side, but rather an increase in the GRF im- pulse. TT clearly show an asymmetric loading with higher values for the sound side (70% of patients), but 3 times less than that reported by Nolan and co-workers. As previously reported, it is reasonable to ascribe this improvement to the use of energy-storage-and-return feet compared to SACH [ 27 , 43 ] . Question Q5 Question Q5 asked if advanced prosthetic components improve temporal and loading symmetry, and if C-leg users have better results than mechanical knee users of the same mobility level. Results support a positive answer. Results have been partially discussed while addressing Q1 and Q4 and can be summarized stating that TFC were statistically different from TFM for SNS and IMS. Results for IMS bring TFC to undistinguishable results to TT. Also, the C-leg in combination with Variflex triggers a new GRF pattern that possibly ensures an increased comfort during walking (Question Q1). This requires fur- ther experimental confirmations. Petersen et al. [ 44 ] have previously reported about SNS in C-leg users compared to TFM. However, that study was not able to prove a statistically significant improvement but just a trend, probably due to the small number of subjects included (5) with different amputation etiologies. Our results confirm that trend, with statistically significant differences. More generally, a considerable body of knowledge is available about the positive effects of the C-leg on amputees ’ mobility [ 31 , 45 – 47 ] , gait kinematic [ 32 – 40 ] , kinetic [ 39 ] and step-length symmetry [ 32 ] . Our findings match well with this general trend toward improved symmetries. As discussed in Q4, the comparison of the literature with our results for TT suggests a possible positive effect of energy-storage-and-return feet in comparison with SACH, for all the indexes of symmetry. Question Q6 Question Q6 asked if it is always true that amputees overload the sound side both in terms of first peak and impulse of GRF, thus contributing to the development of osteoarthritis. Results support a negative answer. As previously discussed about Q4, if we focus on IMS, 100% of TFM overload the sound side . This percentage decreases to 75% of TFC and 57% of TT. If we look at P1S, 41% TF load more the sound side . However, this percentage rises to 70% for TT. Based on these different percentages of TT and TF for IMS and P1S, it could be argued that two different mechanisms might be related to knee osteoarthritis for the two groups: peak overload for TT (measured by P1S), and extended duration of force action (impulse) for TF (measured by IMS). Given the higher prevalence of knee osteoarthritis in TF compared to TT [ 5 , 10 ] , it might be speculated that the second mechanism is more detrimental than the first. Table 4 Results from Nolan et al. [ 2 ] , converted to the indexes of symmetry used in this study. SNS (stance), IMS (impulse) and P1S (first peak) SNS IMS P1S TFM 1,27 1,69 1,22 TT 1,05 1,36 1,25 CONTROLS 1,03 1,08 1,08 Having named N the indexes in [ 2 ] , the new values follow from this equation: New = (2 + N )/(2- N ) TFM transfemoral mechanical knee users, TT transtibial amputees Cutti et al. Journal of NeuroEngineering and Rehabilitation 2018, 15 (Suppl 1):61 Page 38 of 72