- Email: [email protected]
Indicators of dynamic stability in transtibial prosthesis users
Gait & Posture 31 (2010) 375–379
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Indicators of dynamic stability in transtibial prosthesis users C. Kendell a,b,*, E.D. Lemaire a,c, N.L. Dudek a,c, J. Kofman d a
The Ottawa Hospital Rehabilitation Centre, Canada Faculty of Health Sciences, University of Ottawa, Canada c Faculty of Medicine, University of Ottawa, Canada d Department of Systems Design Engineering, University of Waterloo, Canada b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 October 2009 Received in revised form 16 December 2009 Accepted 7 January 2010
An improved understanding of factors related to dynamic stability in lower-limb prosthesis users is important, given the high occurrence of falls in this population. Current methods of assessing stability are unable to adequately characterize dynamic stability over a variety of walking conditions. F-Scan Mobile has been used to collect plantar pressure data and six extracted parameters were useful measures of dynamic stability. The aim of this study was to investigate dynamic stability in individuals with unilateral transtibial amputation based on these six parameters. Twenty community ambulators with a unilateral transtibial amputation walked over level ground, uneven ground, stairs, and a ramp while plantar pressure data were collected. For each limb (intact and prosthetic) and condition, six stability parameters related to plantar center-of-pressure perturbations and gait temporal parameters, were computed from the plantar pressure data. Parameter values were compared between limbs, walking condition, and groups (unilateral transtibial prosthesis users and able-bodied subjects). Differences in parameters were found between limbs and conditions, and between prosthesis users and able-bodied individuals. Further research could investigate optimizing parameter calculations for unilateral transtibial prosthesis users and define relationships between potential for falls and the dynamic stability measures. ß 2010 Elsevier B.V. All rights reserved.
Keywords: Gait Dynamic stability Plantar pressure Biomechanics Prosthetics Center of pressure
1. Introduction Dynamic stability is an important factor for safe mobility. This is especially important for people with lower-limb amputations since they have a higher incidence of falls than the general population. A survey found that 82% of polled individuals with lower-limb amputations had fallen within the previous year; with 49% of those falls occurring while the person wore their prosthesis [1]. For individuals with unilateral amputations, 54% had fallen within the previous year, with 75% of those people reporting more than one fall [2]. Intrinsic factors (i.e., psychological issues, medication, vision loss, age, disease), environmental factors, and prosthetic factors (i.e., fit and alignment) contribute to instability that can result in falling [1,2]. Stability is ‘‘the property of a body that causes it when disturbed from a condition of equilibrium or steady motion to develop forces or moments that restore the original condition’’ [3]. Dynamic stability for lower-limb prosthetic gait has not been thoroughly addressed in the literature, partly due to the lack of an appropriate,
* Corresponding author at: The Ottawa Hospital Rehabilitation Centre, 505 Smyth Road, Ottawa, Ontario, Canada K1H 8M2. Tel.: +1 613 737 8899x75321. E-mail address: [email protected] (C. Kendell). 0966-6362/$ – see front matter ß 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2010.01.003
portable instrument to quantify dynamic stability. Common stability assessment methods used in the clinic and gait laboratory are limited. Clinical tests are designed to be conducted in a short amount of time, with little equipment, and often assess the minimum number of tasks to characterize stability. Motion capture systems restrict assessment to a laboratory setting and require expertise for accurate data collection, analysis, and interpretation. Ideally, a quantitative multi-factorial approach for dynamic stability assessment that provides clinically meaningful output, measures stability over a variety of walking conditions, and is easily applied in the clinic and in the community, is needed. Recently, a method was developed for assessing dynamic stability based on plantar pressure. Using plantar pressure data collected from a multi-celled shoe–insole pressure sensor (F-Scan Mobile, Tekscan Inc., Boston, USA), six dynamic stability parameters relating to plantar center-of-pressure anterior/posterior and medial/lateral motion and gait timing, were identified [4]. The parameters were used to develop a dynamic stability index [5] based on a fuzzy-logic model. The purpose of the current study was to investigate the dynamic stability of individuals with unilateral transtibial amputations using the six parameters. Subjects navigated six conditions: level-ground walking, uneven ground walking, stair ascent/descent, and ramp ascent/descent. Differences between the intact and prosthetic limbs were examined, as
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
376
well as differences between conditions. Stability parameters for individuals with unilateral transtibial amputations were also compared to parameters for able-bodied level-ground walking.
5. Stride time (ST): The time (seconds) from foot strike to foot strike on the same foot. 6. Double support time (DST): The time (seconds) within a single stance phase that both feet are in contact with the ground.
2. Methods 2.1. Subjects Twenty individuals with unilateral transtibial amputations, who were independent community ambulators, volunteered to participate (m = 15, f = 5; age: 61 14 years; weight: 79 16 kg). Individuals with bilateral amputations and those who used a knee–ankle–foot orthosis or ankle–foot orthosis on the intact side were excluded from the study. The project physiatrist and prosthetist screened all subjects to ensure safe participation. Able-bodied subjects consisted of 15 subjects with no impairment that would affect their gait (age: 34 12 years; weight: 69 12 kg). Informed consent was obtained from all participants. 2.2. Data collection F-Scan pressure-sensor insoles were trimmed and fit into each subject’s shoes. The F-Scan Mobile system was used to collect all plantar pressure data at 120 Hz over the following conditions:
The output for several middle strides (three up ramp and three down ramp; five for all other conditions) were extracted for each limb (i.e., intact and prosthetic) of each trial. For each subject and condition, the mean of each stability parameter was calculated over the total number of strides extracted for each limb. 2.4. Data analysis Statistical analysis was conducted using SPSS software (version 15.0 for Windows). Paired t-tests were used to compare stability parameters between the intact and prosthetic limbs for each condition (p < 0.05). A one-way analysis of variance (ANOVA) was used to determine whether stability parameters differed between conditions. If a difference was found, a post hoc Bonferroni’s test was used to find conditions that differed significantly (p < 0.05). For each condition, stability parameter values were compared to values for able-bodied level-ground gait using an independent-samples t-test (p < 0.05).
3. Results Level ground: Data were recorded as each subject walked along a 10 m level walkway. Five successful trials were recorded. Uneven ground: Data were recorded as each subject walked over a row of foam mats (8 m 1 m). Five successful trials were recorded. Ramp: Data were recorded as each subject walked on a ramp (7-degree incline). Ten trials were completed (five ascending and five descending). Stairs: Data were recorded as each subject navigated a 12-step stairwell. Four trials were collected (two ascending and two descending). The order of walking conditions was randomized for each subject. A person walked beside each subject during all walking trials to ensure subject safety. 2.3. Data processing F-Scan plantar pressure data were exported as ASCII files and processed using custom software to calculate the following six stability parameters for each stride, for each foot. 1. Shifts in anterior/posterior center of pressure (AP): Able-bodied gait is characterized by an anterior/posterior center-of-pressure (A/P CoP) trajectory that transitions smoothly from heel to forefoot [6,7]. During stair ascent, the forefoot contacts the ground first, followed by heel contact. To distinguish the forefoot–heel motion during stair ascent from perturbations caused by instability, the first derivative of the raw A/P CoP trajectory was calculated and a dual threshold was implemented. AP is the number of times the first derivative crosses an established dual threshold of 0.5 mm/frame. 2. Shifts in medial/lateral center of pressure (ML): During able-bodied gait, the medial/lateral center-of-pressure (M/L CoP) trajectory begins at the medial heel, moves laterally, and ends at the medial forefoot [7]. To distinguish normal medial/lateral shifts from perturbations caused by instability, a dual threshold was implemented. The first derivative of M/L CoP position was calculated and a dual threshold was implemented. ML is the number of times the first derivative crosses an established dual threshold of 0.5 mm/frame. 3. Cell triggering (CellTrig): When the foot transitions smoothly, each sensor on the insole should only be activated once. CellTrig is the number of times a sensor (i.e., cell) is turned on more than once throughout a stride, normalized by the number of frames in a stride. 4. Maximum lateral force placement (MaxLat): Lateral excursion of the CoP trajectory is associated with increased instability [8,9]. MaxLat is the position of the most laterally activated column of the CoP, expressed as a percentage of the insole width. A higher value indicates increased lateral excursion of the CoP trajectory (i.e., increased instability).
The mean stability parameters for the intact and prosthetic limbs are summarized in Fig. 1, while the differences in means (intact minus prosthetic) are shown in Table 1. Mean values for the intact limb were greater than for the prosthetic limb for 29 of 36 cases (six parameters six conditions). For AP, differences between limbs were significant for all conditions. For MaxLat and ST, there were no significant differences between limbs for any condition. For all other parameters, significant differences occurred for one to three conditions. When significant differences occurred, values were greater on the intact limb (i.e., more unstable). In terms of condition, significant differences between limbs occurred most frequently for downstairs. When comparing conditions, the intact and prosthetic limbs were examined separately (Table 2). Overall, there were significant differences between conditions for every parameter, with the exception of ML, for at least one limb. CellTrig had more significant differences between conditions than any other parameter (four for each limb). When all parameters were considered, the condition that differed most frequently from other conditions was upstairs. Upstairs differed significantly from level ground most frequently (i.e., for four of six parameters for the intact limb and three of six parameters for the prosthetic limb). Data from individuals with transtibial amputations were also compared to data for able-bodied subjects walking on level ground. Table 3 shows that for 42 of the 72 possible cases (two feet six conditions six parameters), prosthesis-user parameter values differed significantly from able-bodied data. In 28 of the 42 cases, prosthesis-user values were greater (i.e. less stable) than able-bodied values. For AP, values for the intact limb were significantly greater than able-bodied values for upstairs and downstairs. AP values for the prosthetic limb were significantly less than able-bodied values for all conditions except upstairs. ML was significantly less for prosthesis users than able-bodied subjects for level ground, upramp, and upstairs on the intact
Table 1 Differences in means between intact and prosthetic limbs (I–P) for the six stability parameters for different conditions. Significant differences (p < 0.05) are marked with an asterisk (*). Variables are explained in the text. Variable
Condition Level
AP (#) ML (#) CellTrig MaxLat (%) ST (s) DST (s)
1.091* 0.785* 4.251 0.798 0.006 0.017
Uneven 1.678* 0.218 3.308 1.607 0.000 0.019
Upramp 1.530* 0.271 2.450 0.926 0.016 0.023*
Downramp 1.515* 0.346 8.152* 2.405 0.001 0.043
Upstairs 1.699* 0.690 3.762 1.521 0.022 0.062*
Downstairs 2.792* 1.492* 8.761* 0.253 0.008 0.032*
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
377
Fig. 1. Mean stability parameter values and standard deviations for the intact and prosthetic limbs for each condition.
Table 2 Stability parameter ANOVA results and conditions that were identified as significantly different by post hoc Bonferonni’s tests. Conditions include: level ground (LG), uneven ground (UG), upramp (UR), downramp (DR), upstairs (US), and downstairs (DS). Stability parameter
Intact
Prosthetic
ANOVA
Bonferroni
ANOVA
Bonferroni
AP (#)
F (5,114) = 4.71, p < 0.05
US > LG DS > LG
F(5,114) = 13.264, p < 0.05
US > LG US > UG US > UR US > DR US > DS
ML (#)
F(5,114) = 0.256, p > 0.05
–
F(5,114) = 0.848, p > 0.05
–
CellTrig
F(5,114) = 5.480, p < 0.05
UG > LG US > LG DS > LG US > UR
F(5,114) = 6.452, p < 0.05
US > LG US > DR UG > LG UG > DR
MaxLat (%)
F(5,114) = 2.783, p < 0.05
–
F(5,114) = 2.478, p < 0.05
DS > UG
ST (s)
F(5,114) = 3.902, p < 0.05
US > LG US > DR
F(5,114) = 3.992, p < 0.05
US > LG US > DR
DST (s)
F(5,114) = 3.356, p < 0.05
US > LG
F(5,114) = 1.855, p > 0.05
–
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
378
Table 3 Mean stability parameter values for each foot, intact limb (I), prosthetic limb (P), and each condition. Values marked with an asterisk (*) are significantly different at the p < 0.05 level from able-bodied data (A). Stability parameter
Condition Level
AP (#) A I P ML (#) A I P
Uneven
Upramp
Downramp
Upstairs
Downstairs
2.496 1.548 0.456*
2.235 0.556*
2.282 0.752*
2.086 0.571*
4.379* 2.680
4.247* 1.455*
4.967 3.622* 2.838*
3.622 2.838
3.618* 3.347*
3.888 3.542*
3.703* 3.013*
4.114 2.622*
CellTrig A I P
29.075 33.960* 29.709
41.824* 38.516*
37.634* 35.184*
37.741* 29.589
45.260* 41.498*
42.867* 34.106*
MaxLat (%) A I P
62.372 63.183 63.981
59.397 57.790*
61.262 62.188
59.536 61.942
64.746 63.225
65.575 65.828
ST (s) A I P
1.065 1.166* 1.160*
1.335* 1.335*
1.304* 1.320*
1.217* 1.216*
1.462* 1.441*
1.342* 1.350*
DST (s) A I P
0.130 0.138 0.155
0.198* 0.179*
0.192* 0.170
0.168 0.125
0.216* 0.154
0.153 0.121
limb, and for all conditions except uneven ground on the prosthetic limb. MaxLat was significantly less for prosthesis users than ablebodied for uneven ground on the prosthetic limb. The parameter that differed from able-bodied most often was ST, which was significantly greater for prosthesis users than able-bodied for both limbs for all conditions (12 cases). MaxLat was only significantly different for 1 out of 12 cases, the least number of cases for all parameters. 4. Discussion A new approach of assessing dynamic stability in unilateral transtibial prosthesis users was presented as a viable multifactorial method for objective assessment within and outside laboratory and clinical environments. This method was based on plantar pressure parameters related to anterior/posterior and medial/lateral plantar center-of-pressure perturbations and gait timing. Each of the six parameters varied in terms of sensitivity, depending on the comparison being made (i.e., between limbs, condition, or group), however, MaxLat was consistently among the parameters with the least number of significant differences (i.e., lowest sensitivity). 4.1. Comparison between limbs When comparing stability parameters between limbs, asymmetry between the intact and prosthetic limbs was evident. Typically, parameter values were higher (i.e., more unstable) on the intact side. For AP, ML, and CellTrig, intact limb values were greater than for the prosthetic limb for all conditions. For MaxLat (level ground, upramp, downramp, downstairs), ST (upramp, downstairs) and DST (level), prosthetic limb values were greater than intact limb values; however, these differences were not significant. It is reasonable that the intact limb is less stable than the prosthetic limb because the intact limb has a greater range of motion and the necessary musculature, proprioception, and motor
control to compensate for disturbances in equilibrium during walking (i.e., the intact foot and ankle can move and adjust to help the person maintain their stability). The ability of the intact limb to adapt to disturbances can contribute to an overall increase in dynamic stability; however, safety issues may occur if an individual approaches their stability threshold on the intact limb. The lack of statistically significant differences between intact and prosthetic limbs is likely due to high inter-subject variability, as expected for a non-targeted sample of prosthesis users. The high sensitivity of the parameters to variability between subjects is positive since this indicates that the parameters are capable of identifying differences in stability between individuals. 4.2. Comparison between conditions With the exception of ML, each variable was affected by condition for at least one limb. When all parameters were considered, values for upstairs differed significantly from other conditions most often. More specifically, upstairs differed from level ground most often. In addition, when significant differences occurred, values for upstairs were always greater and those for level ground always less than the comparison condition. These findings support the assumption that, based on the physical demands for each condition, level ground would be the most stable condition and upstairs would be the least stable. While significant differences between parameters occurred less frequently for the intermediate conditions, most parameters were able to identify significant differences between the least stable and most stable conditions. For ML, values for each condition were similar in magnitude. This suggests that prosthesis users may employ similar mediolateral control strategies for ground, ramps, and stairs. 4.3. Comparison to able-bodied data For AP and ML, there were several conditions for which ablebodied values were greater than those for prosthesis users. It is
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
interesting that these values for able-bodied subjects walking on level ground were greater (i.e., less stable) than for individuals with unilateral transtibial amputations who were navigating more physically demanding conditions. The prosthesis users may have adopted a gait pattern that promotes safe forward progression while minimizing adjustments in anterioposterior and mediolateral motion. The passive prosthetic limb may contribute to such a pattern. The dual threshold used to calculate AP and ML values were based on data from able-bodied individuals. These thresholds may need adjustment for the prosthesis-user population. Individuals with unilateral transtibial amputations may experience a higher number of small perturbations that do not cross the 0.5 mm/s dual threshold. CellTrig was higher for prosthesis users in comparison to able-bodied data for all conditions, and significantly greater for all conditions except level ground and downramp on the prosthetic limb. This supports the theory of a prosthetic gait pattern characterized by frequent small shifts in plantar center of pressure. MaxLat values for able-bodied and prosthesis users were very similar with only one significant difference, MaxLat was significantly less than able-bodied for uneven ground on the prosthetic limb. As with AP and ML, this result could be due to an adapted gait pattern that promoted minimized medial/lateral motion on the prosthetic limb when navigating uneven terrain. ST was significantly greater than able-bodied for all conditions, indicating that individuals with transtibial amputations walk more slowly. DST for prosthesis users was greater than able-bodied for all conditions except downramp and downstairs on the prosthetic limb. A slower walking speed and more time in double support are established strategies for dealing with instability during locomotion. However, for stair and ramp descent, prosthesis users can use the slope of the walking surface to facilitate rollover, and decrease DST, on the prosthetic limb. 4.4. Limitations The prosthesis users and able-subjects in this study were not age matched. Given that the able-bodied subjects were substantially younger than the prosthesis users, some of the differences in stability outcomes may have been due to age-related differences in physical capacity or general health status. MaxLat was based on the total number of columns in an unmodified sensor, and the sensors were cut to fit each subject’s shoes. This approach is efficient for clinical applications, since the number of cells remaining after trimming an insole is unknown without further processing to count the remaining columns. If the number of columns in the trimmed sensor were used to calculate MaxLat, the values would have been larger, indicating increased lateral excursion of the CoP and decreased stability. Since MaxLat was not sensitive to differences between limbs, conditions, and groups, additional research is necessary to determine if changing the calculation method substantially affects the outcomes and whether MaxLat is an appropriate stability parameter for this population. Finally, future research may focus on investigating alternative AP and ML thresholds specific to individuals with unilateral transtibial amputations.
379
5. Conclusion The proposed method of stability assessment examines dynamic stability and provides useful information on how each limb interacts with the ground for several mobility tasks. This information could be used by clinicians to make device prescription decisions and improve prosthetic alignment. The current study highlights differences between the intact and prosthetic limbs, between conditions, and between prosthesis users and ablebodied subjects. The prosthetic limb had consistently lower outcomes, indicating a gait strategy that optimizes dynamic stability on the prosthetic limb and adaptation by the intact limb. The data collected in this study can be used to optimize parameter calculations and determine the appropriate combination of stability parameters required for thorough, multi-factorial, stability assessment in unilateral transtibial amputees. Analysis of subject subgroups, divided by activity or ability level, may reduce variability between subjects and make differences more clear. In future work, critical values could be established that relate each parameter to fall risk. Acknowledgements The authors acknowledge the clinical and technical support of Jocelyn Fawcett, Patrick Lebel, Julie Kim, David Nielen, and Sylvie Maurice-Langis; assistance in the data collection process by Stephen Baskey and Shawn Millar; and facilities support from The Ottawa Hospital Rehabilitation Centre. This project was funded by The Ottawa Hospital Centre for Patient Safety. Conflict of interest statement The authors do not have any financial and personal relationships with other people or organisations that could inappropriately bias their work or be perceived as a conflict of interest. References [1] Kulkarni J, Toole C, Hirons R, Wright S, Morris J. Falls in patients with lower limb amputations: prevalence and contributing factors. Physiotherapy 1996;82(2): 130–6. [2] Miller WC, Speechly M, Deathe B. The prevalence and risk factors of falling and fear among lower extremity amputees. Archives of Physical Medicine and Rehabilitation 2001;82:1031–7. [3] Merriam-Webster Online Dictionary. www.merriamwebster.com/dictionary/ stability [Verified February 4, 2009]. [4] Lemaire ED, Biswas A, Kofman J. Plantar pressure parameters for dynamic gait stability analysis. In: 28th IEEE engineering in medicine and biology society conference; September 2006. [5] Biswas A, Lemaire ED, Kofman J. Dynamic gait stability index based on plantar pressures and fuzzy logic. Journal of Biomechanics 2008;41:1574–81. [6] Rai DV, Aggarwal LM. The study of plantar pressure distribution in normal and pathological foot. Polish Journal of Medical Physics and Engineering 2006;12 (1):25–34. [7] Schmid M, Beltrami G, Zambarbieri D, Verni G. Center of pressure displacements in trans-femoral amputees during gait. Gait and Posture 2005;21:255–62. [8] Kang HG, Dingwell JB. A direct comparison of local dynamic stability during unperturbed standing and walking. Experimental Brain Research 2006;172:35– 48. [9] Patton JL, Pai YC, Lee WA. Evaluation of a model that determines the stability limits of dynamic balance. Gait and Posture 1999;9:38–49.
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Indicators of dynamic stability in transtibial prosthesis users C. Kendell a,b,*, E.D. Lemaire a,c, N.L. Dudek a,c, J. Kofman d a
The Ottawa Hospital Rehabilitation Centre, Canada Faculty of Health Sciences, University of Ottawa, Canada c Faculty of Medicine, University of Ottawa, Canada d Department of Systems Design Engineering, University of Waterloo, Canada b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 October 2009 Received in revised form 16 December 2009 Accepted 7 January 2010
An improved understanding of factors related to dynamic stability in lower-limb prosthesis users is important, given the high occurrence of falls in this population. Current methods of assessing stability are unable to adequately characterize dynamic stability over a variety of walking conditions. F-Scan Mobile has been used to collect plantar pressure data and six extracted parameters were useful measures of dynamic stability. The aim of this study was to investigate dynamic stability in individuals with unilateral transtibial amputation based on these six parameters. Twenty community ambulators with a unilateral transtibial amputation walked over level ground, uneven ground, stairs, and a ramp while plantar pressure data were collected. For each limb (intact and prosthetic) and condition, six stability parameters related to plantar center-of-pressure perturbations and gait temporal parameters, were computed from the plantar pressure data. Parameter values were compared between limbs, walking condition, and groups (unilateral transtibial prosthesis users and able-bodied subjects). Differences in parameters were found between limbs and conditions, and between prosthesis users and able-bodied individuals. Further research could investigate optimizing parameter calculations for unilateral transtibial prosthesis users and define relationships between potential for falls and the dynamic stability measures. ß 2010 Elsevier B.V. All rights reserved.
Keywords: Gait Dynamic stability Plantar pressure Biomechanics Prosthetics Center of pressure
1. Introduction Dynamic stability is an important factor for safe mobility. This is especially important for people with lower-limb amputations since they have a higher incidence of falls than the general population. A survey found that 82% of polled individuals with lower-limb amputations had fallen within the previous year; with 49% of those falls occurring while the person wore their prosthesis [1]. For individuals with unilateral amputations, 54% had fallen within the previous year, with 75% of those people reporting more than one fall [2]. Intrinsic factors (i.e., psychological issues, medication, vision loss, age, disease), environmental factors, and prosthetic factors (i.e., fit and alignment) contribute to instability that can result in falling [1,2]. Stability is ‘‘the property of a body that causes it when disturbed from a condition of equilibrium or steady motion to develop forces or moments that restore the original condition’’ [3]. Dynamic stability for lower-limb prosthetic gait has not been thoroughly addressed in the literature, partly due to the lack of an appropriate,
* Corresponding author at: The Ottawa Hospital Rehabilitation Centre, 505 Smyth Road, Ottawa, Ontario, Canada K1H 8M2. Tel.: +1 613 737 8899x75321. E-mail address: [email protected] (C. Kendell). 0966-6362/$ – see front matter ß 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2010.01.003
portable instrument to quantify dynamic stability. Common stability assessment methods used in the clinic and gait laboratory are limited. Clinical tests are designed to be conducted in a short amount of time, with little equipment, and often assess the minimum number of tasks to characterize stability. Motion capture systems restrict assessment to a laboratory setting and require expertise for accurate data collection, analysis, and interpretation. Ideally, a quantitative multi-factorial approach for dynamic stability assessment that provides clinically meaningful output, measures stability over a variety of walking conditions, and is easily applied in the clinic and in the community, is needed. Recently, a method was developed for assessing dynamic stability based on plantar pressure. Using plantar pressure data collected from a multi-celled shoe–insole pressure sensor (F-Scan Mobile, Tekscan Inc., Boston, USA), six dynamic stability parameters relating to plantar center-of-pressure anterior/posterior and medial/lateral motion and gait timing, were identified [4]. The parameters were used to develop a dynamic stability index [5] based on a fuzzy-logic model. The purpose of the current study was to investigate the dynamic stability of individuals with unilateral transtibial amputations using the six parameters. Subjects navigated six conditions: level-ground walking, uneven ground walking, stair ascent/descent, and ramp ascent/descent. Differences between the intact and prosthetic limbs were examined, as
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
376
well as differences between conditions. Stability parameters for individuals with unilateral transtibial amputations were also compared to parameters for able-bodied level-ground walking.
5. Stride time (ST): The time (seconds) from foot strike to foot strike on the same foot. 6. Double support time (DST): The time (seconds) within a single stance phase that both feet are in contact with the ground.
2. Methods 2.1. Subjects Twenty individuals with unilateral transtibial amputations, who were independent community ambulators, volunteered to participate (m = 15, f = 5; age: 61 14 years; weight: 79 16 kg). Individuals with bilateral amputations and those who used a knee–ankle–foot orthosis or ankle–foot orthosis on the intact side were excluded from the study. The project physiatrist and prosthetist screened all subjects to ensure safe participation. Able-bodied subjects consisted of 15 subjects with no impairment that would affect their gait (age: 34 12 years; weight: 69 12 kg). Informed consent was obtained from all participants. 2.2. Data collection F-Scan pressure-sensor insoles were trimmed and fit into each subject’s shoes. The F-Scan Mobile system was used to collect all plantar pressure data at 120 Hz over the following conditions:
The output for several middle strides (three up ramp and three down ramp; five for all other conditions) were extracted for each limb (i.e., intact and prosthetic) of each trial. For each subject and condition, the mean of each stability parameter was calculated over the total number of strides extracted for each limb. 2.4. Data analysis Statistical analysis was conducted using SPSS software (version 15.0 for Windows). Paired t-tests were used to compare stability parameters between the intact and prosthetic limbs for each condition (p < 0.05). A one-way analysis of variance (ANOVA) was used to determine whether stability parameters differed between conditions. If a difference was found, a post hoc Bonferroni’s test was used to find conditions that differed significantly (p < 0.05). For each condition, stability parameter values were compared to values for able-bodied level-ground gait using an independent-samples t-test (p < 0.05).
3. Results Level ground: Data were recorded as each subject walked along a 10 m level walkway. Five successful trials were recorded. Uneven ground: Data were recorded as each subject walked over a row of foam mats (8 m 1 m). Five successful trials were recorded. Ramp: Data were recorded as each subject walked on a ramp (7-degree incline). Ten trials were completed (five ascending and five descending). Stairs: Data were recorded as each subject navigated a 12-step stairwell. Four trials were collected (two ascending and two descending). The order of walking conditions was randomized for each subject. A person walked beside each subject during all walking trials to ensure subject safety. 2.3. Data processing F-Scan plantar pressure data were exported as ASCII files and processed using custom software to calculate the following six stability parameters for each stride, for each foot. 1. Shifts in anterior/posterior center of pressure (AP): Able-bodied gait is characterized by an anterior/posterior center-of-pressure (A/P CoP) trajectory that transitions smoothly from heel to forefoot [6,7]. During stair ascent, the forefoot contacts the ground first, followed by heel contact. To distinguish the forefoot–heel motion during stair ascent from perturbations caused by instability, the first derivative of the raw A/P CoP trajectory was calculated and a dual threshold was implemented. AP is the number of times the first derivative crosses an established dual threshold of 0.5 mm/frame. 2. Shifts in medial/lateral center of pressure (ML): During able-bodied gait, the medial/lateral center-of-pressure (M/L CoP) trajectory begins at the medial heel, moves laterally, and ends at the medial forefoot [7]. To distinguish normal medial/lateral shifts from perturbations caused by instability, a dual threshold was implemented. The first derivative of M/L CoP position was calculated and a dual threshold was implemented. ML is the number of times the first derivative crosses an established dual threshold of 0.5 mm/frame. 3. Cell triggering (CellTrig): When the foot transitions smoothly, each sensor on the insole should only be activated once. CellTrig is the number of times a sensor (i.e., cell) is turned on more than once throughout a stride, normalized by the number of frames in a stride. 4. Maximum lateral force placement (MaxLat): Lateral excursion of the CoP trajectory is associated with increased instability [8,9]. MaxLat is the position of the most laterally activated column of the CoP, expressed as a percentage of the insole width. A higher value indicates increased lateral excursion of the CoP trajectory (i.e., increased instability).
The mean stability parameters for the intact and prosthetic limbs are summarized in Fig. 1, while the differences in means (intact minus prosthetic) are shown in Table 1. Mean values for the intact limb were greater than for the prosthetic limb for 29 of 36 cases (six parameters six conditions). For AP, differences between limbs were significant for all conditions. For MaxLat and ST, there were no significant differences between limbs for any condition. For all other parameters, significant differences occurred for one to three conditions. When significant differences occurred, values were greater on the intact limb (i.e., more unstable). In terms of condition, significant differences between limbs occurred most frequently for downstairs. When comparing conditions, the intact and prosthetic limbs were examined separately (Table 2). Overall, there were significant differences between conditions for every parameter, with the exception of ML, for at least one limb. CellTrig had more significant differences between conditions than any other parameter (four for each limb). When all parameters were considered, the condition that differed most frequently from other conditions was upstairs. Upstairs differed significantly from level ground most frequently (i.e., for four of six parameters for the intact limb and three of six parameters for the prosthetic limb). Data from individuals with transtibial amputations were also compared to data for able-bodied subjects walking on level ground. Table 3 shows that for 42 of the 72 possible cases (two feet six conditions six parameters), prosthesis-user parameter values differed significantly from able-bodied data. In 28 of the 42 cases, prosthesis-user values were greater (i.e. less stable) than able-bodied values. For AP, values for the intact limb were significantly greater than able-bodied values for upstairs and downstairs. AP values for the prosthetic limb were significantly less than able-bodied values for all conditions except upstairs. ML was significantly less for prosthesis users than able-bodied subjects for level ground, upramp, and upstairs on the intact
Table 1 Differences in means between intact and prosthetic limbs (I–P) for the six stability parameters for different conditions. Significant differences (p < 0.05) are marked with an asterisk (*). Variables are explained in the text. Variable
Condition Level
AP (#) ML (#) CellTrig MaxLat (%) ST (s) DST (s)
1.091* 0.785* 4.251 0.798 0.006 0.017
Uneven 1.678* 0.218 3.308 1.607 0.000 0.019
Upramp 1.530* 0.271 2.450 0.926 0.016 0.023*
Downramp 1.515* 0.346 8.152* 2.405 0.001 0.043
Upstairs 1.699* 0.690 3.762 1.521 0.022 0.062*
Downstairs 2.792* 1.492* 8.761* 0.253 0.008 0.032*
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
377
Fig. 1. Mean stability parameter values and standard deviations for the intact and prosthetic limbs for each condition.
Table 2 Stability parameter ANOVA results and conditions that were identified as significantly different by post hoc Bonferonni’s tests. Conditions include: level ground (LG), uneven ground (UG), upramp (UR), downramp (DR), upstairs (US), and downstairs (DS). Stability parameter
Intact
Prosthetic
ANOVA
Bonferroni
ANOVA
Bonferroni
AP (#)
F (5,114) = 4.71, p < 0.05
US > LG DS > LG
F(5,114) = 13.264, p < 0.05
US > LG US > UG US > UR US > DR US > DS
ML (#)
F(5,114) = 0.256, p > 0.05
–
F(5,114) = 0.848, p > 0.05
–
CellTrig
F(5,114) = 5.480, p < 0.05
UG > LG US > LG DS > LG US > UR
F(5,114) = 6.452, p < 0.05
US > LG US > DR UG > LG UG > DR
MaxLat (%)
F(5,114) = 2.783, p < 0.05
–
F(5,114) = 2.478, p < 0.05
DS > UG
ST (s)
F(5,114) = 3.902, p < 0.05
US > LG US > DR
F(5,114) = 3.992, p < 0.05
US > LG US > DR
DST (s)
F(5,114) = 3.356, p < 0.05
US > LG
F(5,114) = 1.855, p > 0.05
–
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
378
Table 3 Mean stability parameter values for each foot, intact limb (I), prosthetic limb (P), and each condition. Values marked with an asterisk (*) are significantly different at the p < 0.05 level from able-bodied data (A). Stability parameter
Condition Level
AP (#) A I P ML (#) A I P
Uneven
Upramp
Downramp
Upstairs
Downstairs
2.496 1.548 0.456*
2.235 0.556*
2.282 0.752*
2.086 0.571*
4.379* 2.680
4.247* 1.455*
4.967 3.622* 2.838*
3.622 2.838
3.618* 3.347*
3.888 3.542*
3.703* 3.013*
4.114 2.622*
CellTrig A I P
29.075 33.960* 29.709
41.824* 38.516*
37.634* 35.184*
37.741* 29.589
45.260* 41.498*
42.867* 34.106*
MaxLat (%) A I P
62.372 63.183 63.981
59.397 57.790*
61.262 62.188
59.536 61.942
64.746 63.225
65.575 65.828
ST (s) A I P
1.065 1.166* 1.160*
1.335* 1.335*
1.304* 1.320*
1.217* 1.216*
1.462* 1.441*
1.342* 1.350*
DST (s) A I P
0.130 0.138 0.155
0.198* 0.179*
0.192* 0.170
0.168 0.125
0.216* 0.154
0.153 0.121
limb, and for all conditions except uneven ground on the prosthetic limb. MaxLat was significantly less for prosthesis users than ablebodied for uneven ground on the prosthetic limb. The parameter that differed from able-bodied most often was ST, which was significantly greater for prosthesis users than able-bodied for both limbs for all conditions (12 cases). MaxLat was only significantly different for 1 out of 12 cases, the least number of cases for all parameters. 4. Discussion A new approach of assessing dynamic stability in unilateral transtibial prosthesis users was presented as a viable multifactorial method for objective assessment within and outside laboratory and clinical environments. This method was based on plantar pressure parameters related to anterior/posterior and medial/lateral plantar center-of-pressure perturbations and gait timing. Each of the six parameters varied in terms of sensitivity, depending on the comparison being made (i.e., between limbs, condition, or group), however, MaxLat was consistently among the parameters with the least number of significant differences (i.e., lowest sensitivity). 4.1. Comparison between limbs When comparing stability parameters between limbs, asymmetry between the intact and prosthetic limbs was evident. Typically, parameter values were higher (i.e., more unstable) on the intact side. For AP, ML, and CellTrig, intact limb values were greater than for the prosthetic limb for all conditions. For MaxLat (level ground, upramp, downramp, downstairs), ST (upramp, downstairs) and DST (level), prosthetic limb values were greater than intact limb values; however, these differences were not significant. It is reasonable that the intact limb is less stable than the prosthetic limb because the intact limb has a greater range of motion and the necessary musculature, proprioception, and motor
control to compensate for disturbances in equilibrium during walking (i.e., the intact foot and ankle can move and adjust to help the person maintain their stability). The ability of the intact limb to adapt to disturbances can contribute to an overall increase in dynamic stability; however, safety issues may occur if an individual approaches their stability threshold on the intact limb. The lack of statistically significant differences between intact and prosthetic limbs is likely due to high inter-subject variability, as expected for a non-targeted sample of prosthesis users. The high sensitivity of the parameters to variability between subjects is positive since this indicates that the parameters are capable of identifying differences in stability between individuals. 4.2. Comparison between conditions With the exception of ML, each variable was affected by condition for at least one limb. When all parameters were considered, values for upstairs differed significantly from other conditions most often. More specifically, upstairs differed from level ground most often. In addition, when significant differences occurred, values for upstairs were always greater and those for level ground always less than the comparison condition. These findings support the assumption that, based on the physical demands for each condition, level ground would be the most stable condition and upstairs would be the least stable. While significant differences between parameters occurred less frequently for the intermediate conditions, most parameters were able to identify significant differences between the least stable and most stable conditions. For ML, values for each condition were similar in magnitude. This suggests that prosthesis users may employ similar mediolateral control strategies for ground, ramps, and stairs. 4.3. Comparison to able-bodied data For AP and ML, there were several conditions for which ablebodied values were greater than those for prosthesis users. It is
C. Kendell et al. / Gait & Posture 31 (2010) 375–379
interesting that these values for able-bodied subjects walking on level ground were greater (i.e., less stable) than for individuals with unilateral transtibial amputations who were navigating more physically demanding conditions. The prosthesis users may have adopted a gait pattern that promotes safe forward progression while minimizing adjustments in anterioposterior and mediolateral motion. The passive prosthetic limb may contribute to such a pattern. The dual threshold used to calculate AP and ML values were based on data from able-bodied individuals. These thresholds may need adjustment for the prosthesis-user population. Individuals with unilateral transtibial amputations may experience a higher number of small perturbations that do not cross the 0.5 mm/s dual threshold. CellTrig was higher for prosthesis users in comparison to able-bodied data for all conditions, and significantly greater for all conditions except level ground and downramp on the prosthetic limb. This supports the theory of a prosthetic gait pattern characterized by frequent small shifts in plantar center of pressure. MaxLat values for able-bodied and prosthesis users were very similar with only one significant difference, MaxLat was significantly less than able-bodied for uneven ground on the prosthetic limb. As with AP and ML, this result could be due to an adapted gait pattern that promoted minimized medial/lateral motion on the prosthetic limb when navigating uneven terrain. ST was significantly greater than able-bodied for all conditions, indicating that individuals with transtibial amputations walk more slowly. DST for prosthesis users was greater than able-bodied for all conditions except downramp and downstairs on the prosthetic limb. A slower walking speed and more time in double support are established strategies for dealing with instability during locomotion. However, for stair and ramp descent, prosthesis users can use the slope of the walking surface to facilitate rollover, and decrease DST, on the prosthetic limb. 4.4. Limitations The prosthesis users and able-subjects in this study were not age matched. Given that the able-bodied subjects were substantially younger than the prosthesis users, some of the differences in stability outcomes may have been due to age-related differences in physical capacity or general health status. MaxLat was based on the total number of columns in an unmodified sensor, and the sensors were cut to fit each subject’s shoes. This approach is efficient for clinical applications, since the number of cells remaining after trimming an insole is unknown without further processing to count the remaining columns. If the number of columns in the trimmed sensor were used to calculate MaxLat, the values would have been larger, indicating increased lateral excursion of the CoP and decreased stability. Since MaxLat was not sensitive to differences between limbs, conditions, and groups, additional research is necessary to determine if changing the calculation method substantially affects the outcomes and whether MaxLat is an appropriate stability parameter for this population. Finally, future research may focus on investigating alternative AP and ML thresholds specific to individuals with unilateral transtibial amputations.
379
5. Conclusion The proposed method of stability assessment examines dynamic stability and provides useful information on how each limb interacts with the ground for several mobility tasks. This information could be used by clinicians to make device prescription decisions and improve prosthetic alignment. The current study highlights differences between the intact and prosthetic limbs, between conditions, and between prosthesis users and ablebodied subjects. The prosthetic limb had consistently lower outcomes, indicating a gait strategy that optimizes dynamic stability on the prosthetic limb and adaptation by the intact limb. The data collected in this study can be used to optimize parameter calculations and determine the appropriate combination of stability parameters required for thorough, multi-factorial, stability assessment in unilateral transtibial amputees. Analysis of subject subgroups, divided by activity or ability level, may reduce variability between subjects and make differences more clear. In future work, critical values could be established that relate each parameter to fall risk. Acknowledgements The authors acknowledge the clinical and technical support of Jocelyn Fawcett, Patrick Lebel, Julie Kim, David Nielen, and Sylvie Maurice-Langis; assistance in the data collection process by Stephen Baskey and Shawn Millar; and facilities support from The Ottawa Hospital Rehabilitation Centre. This project was funded by The Ottawa Hospital Centre for Patient Safety. Conflict of interest statement The authors do not have any financial and personal relationships with other people or organisations that could inappropriately bias their work or be perceived as a conflict of interest. References [1] Kulkarni J, Toole C, Hirons R, Wright S, Morris J. Falls in patients with lower limb amputations: prevalence and contributing factors. Physiotherapy 1996;82(2): 130–6. [2] Miller WC, Speechly M, Deathe B. The prevalence and risk factors of falling and fear among lower extremity amputees. Archives of Physical Medicine and Rehabilitation 2001;82:1031–7. [3] Merriam-Webster Online Dictionary. www.merriamwebster.com/dictionary/ stability [Verified February 4, 2009]. [4] Lemaire ED, Biswas A, Kofman J. Plantar pressure parameters for dynamic gait stability analysis. In: 28th IEEE engineering in medicine and biology society conference; September 2006. [5] Biswas A, Lemaire ED, Kofman J. Dynamic gait stability index based on plantar pressures and fuzzy logic. Journal of Biomechanics 2008;41:1574–81. [6] Rai DV, Aggarwal LM. The study of plantar pressure distribution in normal and pathological foot. Polish Journal of Medical Physics and Engineering 2006;12 (1):25–34. [7] Schmid M, Beltrami G, Zambarbieri D, Verni G. Center of pressure displacements in trans-femoral amputees during gait. Gait and Posture 2005;21:255–62. [8] Kang HG, Dingwell JB. A direct comparison of local dynamic stability during unperturbed standing and walking. Experimental Brain Research 2006;172:35– 48. [9] Patton JL, Pai YC, Lee WA. Evaluation of a model that determines the stability limits of dynamic balance. Gait and Posture 1999;9:38–49.