Reference values for gait temporal and loading symmetry of lower-limb amputees can help in refocusing rehabilitation targets

Lower-limb amputees tend to walk asymmetrically when looking at gait temporal and loading parameters, with more time spent and load exerted on the intact limb [1‐9]. Temporal asymmetry is typically measured based on step or stance duration; loading asymmetry based the magnitude of the first peak of the vertical ground reaction force (GRF), and the impulse of GRF [2, 3, 6, 10].

Temporal and loading asymmetries were associated to several comorbidities [5]: increased falls [11], osteoarthritis of the sound limb [10, 12‐15], osteoporosis of the contralateral limb [15, 16], back pain [17‐20]. In addition, walking in public with noticeable asymmetries attracts the general attention [21], which can be very uncomfortable for some prosthesis users. With this background, it is not surprising that a common, almost unquestioned [22], goal for rehabilitation is to regain a symmetric walking [9, 23].

However, the literature does not clearly indicate that striving for perfect symmetry is really and always the best option. Already in 1998, Winter & Sienko [1] stated that “human system with major structural asymmetries in the neuromuscular skeletal system cannot be optimal when gait is symmetrical. Rather, a new non-symmetrical optimal is probably being sought by the amputee within the constraints of his residual system and the mechanics of his prosthesis”. Later in 2005, Schmid and co-workers [3] compared the center of pressure trajectories under the sound and prosthetic foot of transfemoral amputees and concluded that the longer stance on the sound side can be ascribed to the greater ability of the sound leg to advance the step and maintain balance until the prosthetic limb can sustain the body weight. Hof et al. [4] corroborated this explanation in the theoretical framework of the “extrapolated center of mass” [24], concluding that stance time asymmetry is a “sensible adaptation” of experienced transfemoral amputees to improve stability during walking, to overcome the missing lateral ankle strategy of prosthetic feet. More recently, Adamczyk & Kuo [8], with a theoretical and experimental approach involving transtibial amputees, concluded that “some asymmetry may be unavoidable in cases of unilateral limb loss” due to the reduced ankle plantar flexion of the ankle, with direct consequences on stance duration, greater collision work at the sound side, greater work overall, and increased peak force at loading response [25‐27]. Imposing symmetry can actually be detrimental, as also observed by [27, 28].

The evidences from the literature, therefore, indicate that optimal symmetry ratios might exist, to obtain a compromise among stability, forward progression, preservation of body structures and perception of a “normal and symmetric biped locomotion” [21]. Unfortunately, at present not only optimal symmetry ratios are unknown, but we also don’t know typical symmetry ratios or which measures of symmetry are essential and which are redundant.

In our opinion, 3 methodological and 3 clinical questions should be answered to clarify these open issues. The methodological questions are:

The clinical questions are:

Unfortunately, at present it is difficult to answer to these questions based on the available literature, because there are no studies that considered, at the same time 1) both temporal and loading asymmetries, 2) both transfemoral and transtibial amputees treated at the same prosthetic & rehabilitation center, 3) mechanical and electronic knees, 4) energy-storage-and-return feet instead of the SACH (Solid-Ankle Cushion-Heel) foot. Moreover, the number of patients included is typically limited to 8, both for studies on transtibial and transfemoral amputees. Finally, no studies addressed the correlation between temporal and loading parameters.

The aim of this study was to overcome these limitations and answer to questions Q1-Q6 on three groups of well-adapted, traumatic, K3-K4 amputees: transfemoral amputees using a restricted set of mechanical knees (TFM), transfemoral amputees using the C-leg (TFC), transtibial amputees using energy-storage-and-return feet (TT). A additional group of healthy control subjects (“Controls” in short), was also included to highlight general trends.

Gait speed was compared among TFM (1.12 ± 0.13 m/s), TFC (1.17 ± 0.12 m/s), TT (1.23 ± 0.19 m/s) and Controls (1.41 ± 0.21 m/s). ANOVA did not show statistically significant differences among amputees (p = 0.14), but only between Controls and amputees (p = 0.0005).

Further results are reported hereinafter based on their relevance for questions Q1-Q6.

In this study, we addressed three methodological and three clinical questions regarding the temporal and loading symmetry of transfemoral amputees (both mechanical and C-leg users) and transtibial amputees, to support in the development of more targeted rehabilitation goals, that are particularly needed [9, 38].

As a general consideration, the self-selected walking speed was not statistically different among amputees, despite a slight increase in the median from TFM, to TFC, to TT toward Controls. Absolute values compare well with previously reported data [2, 7, 39].

For the sake of clarity, results are discussed below for each question, in comparison with the available literature whenever possible.

In the Introduction, we posed three methodological and three clinical questions regarding the gait temporal and loading symmetry of lower-limb amputees. Based on the results collected on traumatic, K3-K4, transfemoral (mechanical knees and C-leg users) and transtibial patients successfully fit and trained in using their prosthesis, we can answer as follows.

The three methodological questions wanted to establish a minimum set of symmetry indexes to study and if there are limitations in their calculations. First, the first peak of the vertical ground reaction force at loading response cannot be clearly identified in all amputees, and the calculation of its index of symmetry was limited to patients with the typical M-Shaped pattern of the ground reaction force. Second, the analysis of temporal symmetry can be limited to stance, leaving out step symmetry. Third, stance, impulse and first peak symmetries should be separately reported.

The three clinical questions wanted to establish if “typical” levels of temporal and loading symmetry exist and change with the level of amputation and prosthetic components. First, the symmetries of stance, impulse and first peak are all influenced by the level of amputation. In particular, the time spent on the sound side decreases significantly from transfemoral mechanical knee users, to C-leg users, to transtibial patients. The impulse on the sound side decreases significantly from mechanical knee users to C-leg and transtibial patients. Transtibial patients have a higher first peak at loading response on their sound side, while most transfemoral patients do not. Second, advanced prosthetic component seem to positively influence the temporal and loading symmetry. In particular, the C-leg in combination with the Variflex foot improves stance, impulse symmetry and for about 60% of patients smooths the first peak at loading response. About 20% of C-leg users have a stance asymmetry which is below the level of perceived impaired gait, compared to 0% of mechanical knee users. For transtibial patients, comparisons of our results with the literature point toward an improvement of all indexes of symmetry, possibly due to the use of energy-storage-and-return feet instead of SACH feet. Third, it is not always true that amputees overload the sound side. Percentagewise, transfemoral amputees tend to overload the sound side with increased impulse, while TT with increased peak GRF. This might be suggestive of two separate mechanisms for the onset of knee osteoarthritis.

We think that our results can be exploited in the clinical routine. First, clinicians can use our results to set reasonable targets for rehabilitation. Specifically, they can compare the level of symmetry of a new patient with the ranges provided, and put the patient’s performance and advancements during rehabilitation in perspective. Moreover, technical and healthcare professionals might use our findings to compare the effect of different prosthetic components and potentially the effect of different rehabilitation programs. Second, it is often required by payers (e.g. insurances, public healthcare services, or patients), to justify the use of advanced prosthetic components. We think that our results support the use of C-leg and energy-storage-and-return feet on K3-K4 traumatic patients: thanks to the improvement in temporal and loading symmetry compared to mechanical knees and SACH foot, these components can potentially have a positive effect on the asymmetry-related comorbidities analyzed in the Introduction and decrease social stigma. Further research is required to extend these results to other groups of patients, such as K2 and non-traumatic amputees. Finally, our results might suggest possible strategies to mitigate knee osteoarthritis of the sound side. Pending further research, transfemoral amputees might take advantage of prosthetic components with an improved knee-foot coordination to specifically tackle stance time asymmetry. Transtibial patients might benefit from improved socket construction that does not limit knee extension, and prosthetic feet with improved push-off, roll-over shape and range of motion to reduce the first peak at loading response.