- Email: [email protected]
Three-point gait crutch walking: Variability in ground reaction force during weight bearing
86
Three-Point Gait Crutch Walking: Variability in Ground Reaction Force During Weight Bearing Sheng Li, MD, Charles W. Armstrong, PhD, Daniel Cipriani, MEd, PT ABSTRACT. Li S, Armstrong CW, Cipriani D. Three-point gait crutch walking: variability in ground reaction force during weight bearing. Arch Phys Med Rehabil 2001;82:86-92. Objective: To investigate variability in ground reaction force (GRF) and kinematics on both sides during 3-point partial weight-bearing (PWB) crutch walking. Design: Within-subject comparisons of kinematic and kinetic data collected at different levels of 3-point crutch walking: 10%, 50%, and 90% PWB at comfortable speeds. Setting: An applied biomechanics lab in a university setting. Participants: Twelve healthy college students (9 women, 3 men). Main Outcome Measures: Spatial and temporal variables, major peak kinematic data, and peak GRFs from force platforms during the gait cycle. Results: Large variations were found in replicating the target levels of PWB, particularly at 10% and 90% PWB. Subjects had a shorter stance phase and longer swing phase during the crutch walking gait cycle. Velocity significantly decreased (p ⫽ .006) because of decreased cadence (p ⫽ .002). Slightly greater hip abduction and external rotation on the noninvolved side and slightly less hip adduction and internal rotation on the involved side indicated that the center of gravity shifted slightly from the involved side toward the noninvolved side. There was no increase in vertical GRF, and there was a relatively constant loading pattern on the noninvolved side. Conclusions: Subjects have difficulty replicating a prescribed weight-bearing restriction. A shift of the center of gravity toward the noninvolved side may reduce the weight distribution on the involved side. Key Words: Crutches; Gait; Rehabilitation; Walking; Weight bearing. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation RAUMA TO A LOWER EXTREMITY may impair its ability to bear weight. Ankle sprains, fractures, and knee T injuries typically force an individual to adopt a protective weight-bearing status until there is sufficient healing.1,2 Weight-bearing restrictions depend on treatment needs, which range from non–weight bearing to partial weight bearing (PWB) to full weight bearing. PWB helps patients in different
From the Department of Kinesiology, University of Toledo (Li, Armstrong); and the Department of Physical Therapy, Medical College of Ohio (Cipriani), Toledo, OH. Accepted in revised form March 28, 2000. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Sheng Li, MD, Dept of Kinesiology, Room 20-Recreation Hall, Pennsylvania State University, University Park, Pennsylvania, PA 16801, e-mail: sx1282.psu.edu. 0003-9993/01/8201-5794$35.00/0 doi:10.1053/apmr.2001.16347
Arch Phys Med Rehabil Vol 82, January 2001
stages of rehabilitation, with the amount of weight bearing based on the severity of the injury and healing status. Axillary crutches are designed to help individuals walk with reduced weight bearing on the involved lower extremity. They also help an individual maintain balance3 and can help relieve pain in the hip, knee, ankle, and foot.1 Different crutch walking gait patterns include 2-point, 3-point, 4-point, swing-to, and swing-through patterns. Three-point gait crutch walking is commonly used because it provides for varied levels of weight bearing, from non–weight bearing to full weight bearing.2 In this pattern, both crutches are advanced simultaneously with the involved side, and the noninvolved side is then advanced. Winstein et al4 and Bhambhami et al5 found that individuals were not reliable in following weight-bearing restrictions. After 80 trials with feedback from a floor scale during swing-to crutch walking, Winstein et al4 found that individuals still had an error rate of approximately 10% in replicating a target PWB level. Patients usually use crutches without having training in PWB. This likely results in high error rates in replicating prescribed levels of weight bearing. Although research on crutch walking has generally focused on the involved limb, adaptive compensatory changes in the gait pattern of the noninvolved limb may also be of concern. Stallard et al6 investigated ground reactive force (GRF) during swing-through gait crutch walking and found that individuals tended to land with 25% greater vertical force in a single-leg landing and nearly 33% greater force when landing on both legs than when walking normally. These findings are significant in terms of potential stress on the ipsilateral noninvolved leg because the use of crutches may exacerbate such existing conditions as osteoarthritis or rheumatoid arthritis. Stress on the noninvolved leg emphasizes the importance of measuring the forces on both legs during crutch walking. The potential of crutch walking for inducing injuries is not limited to the lower extremity.7-10 Multiple forms of upper extremity injuries, including injuries to the brachial plexus, hand neuropathies, and brachial artery occlusions, have been identified in the literature. Collectively, the issues associated with the pathomechanics of crutch walking are important clinically, considering the common use of axillary crutches and the potential for indirect injury. There have been many studies of crutch walking. Unfortunately, there are few data about PWB. Thus, this study examines the biomechanical characteristics of a 3-point crutch walking gait pattern in 3 different weight-bearing conditions: 10%, 50%, and 90% weight-bearing status. By examining the associated kinetic, kinematic, and temporal and distance characteristics of gait under prescribed levels of weight bearing, we hoped to gain insight into the pathomechanics of crutch walking gait. METHODS Subjects Nine women and 3 men (mean age ⫾ standard deviation, 25 ⫾ 7.6yr; height, 164 ⫾ 4cm; weight, 54 ⫾ 3.1kg) were recruited for this study. To minimize variability in response to
87
THREE-POINT GAIT CRUTCH WALKING, Li
experimental conditions, only subjects with no known history of musculoskeletal, neurologic, cardiovascular, or pulmonary disorders were included. Individuals with muscle weakness or length discrepancy in lower limbs were excluded. Each subject underwent a brief physical examination. Before testing, a written consent form explaining the study, which was approved by the university’s human subject research review committee, was given to each subject. Instrumentation GRF data during walking was recorded by 2 standard strain gauge force platformsa that were embedded in the center of a 10-meter walkway. The boundary of each force platform was identified to the subjects by the tester. Six raw voltage signals from the force platforms were transmitted to the computer through an analog-to-digital converterb at a sampling rate of 480Hz. Kinematic data were collected by a computerized video 3-dimensional analysis system (Hires 3D).c Six cameras (Falcon Hires)c arranged around the walkway at approximately 60° intervals captured the subject’s motion at 60Hz using passive retroflective markers placed on the subject. Markers were applied to the subject according to the Helen Hayes Static marker set, bilaterally on the wrists, elbows, shoulders, anterosuperior iliac spine, midthighs, medial and lateral knee centers, midshanks, medial and lateral ankle centers, heels and forefeet, and sacrum. Video data were digitized and processed using EVA 4.01 software.c All kinematic data were smoothed with a low pass filter using a cutoff at 6Hz. A static test identified the location of each marker on the body for each subject. Video and GRF data were output to OrthoTrak software version 4.1,c which calculated the kinetic, kinematic, and temporal and distance measures. Procedures Two days before they were tested, the subjects were shown how to use adjustable axillary crutches appropriately for a 3-point gait pattern. First, crutch length was adjusted to match each subject’s height, with approximately 1 hand width between the axilla and the top of the crutches.2 Second, based on each subject’s weight, 3 levels of target weight for PWB were determined on a standard scale, 10% (low), 50% (medium), and 90% (high) body weight. Subjects were allowed a few practice trials on the scale with concurrent feedback to feel the requisite amount of weight. They were encouraged to walk using a 3-point gait pattern. During PWB crutch walking, 1 leg was randomly assumed to be injured and to take PWB only. That leg was identified as the involved leg (or side), and the other side was the noninvolved side.
After being instructed on 3-point gait crutch walking, subjects were asked to practice it at various levels of PWB for 30 minutes each day at home for 2 days. They then returned for the crutch walking gait assessment, at which time they practiced 3-point crutch walking for approximately 5 minutes before measurements were taken. The subjects reported having completed a mean amount of 45.4 ⫾ 16.8 minutes of practice time and were now comfortable and familiar with crutch walking. On the test day, each subject performed 8 normal walking trials, followed by 8 trials at each level of the PWB crutch walking tests. After the system was calibrated, hemispheric, reflective markers were placed on the subject’s skin bilaterally. A static test was performed before the trial and was used as a calibration and standard. Additionally, each subject’s weight and height were measured and recorded. To ensure a gait typical of their usual walking pattern, subjects were instructed to walk barefoot at their own selected walking speeds along the walkway. They were given multiple practice trials before testing to ensure that they were walking at a normal rhythm. Only those trials in which each foot contacted a force platform separately and that were free of contact between the crutches and the force plates were retained for analysis. Subjects were tested in the same order: normal walking first, followed by 10%, 50%, and 90% PWB crutch walking. After the normal walking test, subjects walked repeatedly across the walkway for approximately 5 minutes with a 3-point pattern of crutch walking at different levels of PWB. Eight trials at each level were performed in the same way as the normal walking trials. Statistical Analysis Group means for each variable were based on the average of 8 trails for each subject for each condition. Standard descriptive statistics followed by repeated-measures analysis of variance were used in the statistical analysis. The main factors were SIDE (2 levels: left, right) and PWB (4 levels: normal, 10%, 50%, 90% PWB).d Tukey post hoc comparisons were used to compare individual cell means. RESULTS Temporal and spatial changes, including cadence, stride length, velocity, step width, double support time, support time, and nonsupport time in the 3-point gait crutch walking are listed in table 1. Compared with normal walking, the cadence (F3,33 ⫽ 14.76; p ⫽ .002) and velocity (F3,33 ⫽ 10.33; p ⫽ .005) of both sides were decreased significantly in PWB crutch walking. Statistical analysis also showed no significant differ-
Table 1: Summary of Spatial Parameters Normal Involved
Noninvolved
10% PWB Involved
Noninvolved
Cadence 111.9 ⫾ 11.35 110.4 ⫾ 10.76 74.5 ⫾ 11.92 70.9 ⫾ 11.05 (steps/s) Stride length 126.1 ⫾ 9.47 124.7 ⫾ 9.54 116.7 ⫾ 19.55 111.8 ⫾ 7.49 (cm) Velocity (cm/s) 118.3 ⫾ 18.24 115.2 ⫾ 17.14 70.8 ⫾ 15.8 66.5 ⫾ 12.52 Step width (cm) 10.6 ⫾ 2.26 9.7 ⫾ 2.26 Double support 9.7 ⫾ 2.19 10.9 ⫾ 1.84 24 ⫾ 27.84 14.3 ⫾ 11.99 time (%) Support time (%) 60.3 ⫾ 1.5 60.6 ⫾ 1.69 40.6 ⫾ 14.58 67.8 ⫾ 4
50% PWB Involved
75.6 ⫾ 15.22 118 ⫾ 14.32
Noninvolved
70.1 ⫾ 10.99
90% PWB Involved
71 ⫾ 13.59
Noninvolved
70.7 ⫾ 13.05
113.7 ⫾ 11.07 118.8 ⫾ 13.74 115.7 ⫾ 10.9
71.2 ⫾ 15.62 66.3 ⫾ 13.47 9.6 ⫾ 1.93 25.3 ⫾ 30.96 12.2 ⫾ 7.06
70.4 ⫾ 15.41 68.2 ⫾ 14.49 9 ⫾ 1.81 28.6 ⫾ 28.52 10.1 ⫾ 1.86
43.8 ⫾ 13.05
45.9 ⫾ 16.71
66.2 ⫾ 3.6
65 ⫾ 2.52
Values expressed as mean ⫾ standard deviation (SD), n ⫽ 12.
Arch Phys Med Rehabil Vol 82, January 2001
88
THREE-POINT GAIT CRUTCH WALKING, Li
ence between the involved and noninvolved sides. Reductions in velocity in both sides resulted from decreases in cadence. Step width significantly decreased in PWB crutch walking compared with normal walking (F3,33 ⫽ 15.53; p ⫽ .001). There was no significant difference in double support time between the 2 sides in the different walking conditions. Support time for the involved side and noninvolved side were significantly different (F3,33 ⫽ 19.35; p ⫽ .002), but crutch use did not have a significant influence on the different levels of PWB crutch walking. Support time (stance phase) for the involved side significantly decreased, ie, the percentage of the stance phase in the gait cycle decreased for the involved side, and there was no significant difference among different levels of PWB crutch walking. However, double support time did not have significant changes. As indicated, there were significant changes in support time for the involved side. Furthermore, the range-of-motion (ROM) patterns of the pelvis, hip, knee, ankle, and foot for both sides changed during crutch walking. The results of these kinematic changes appeared to be linked to shifts of the center of gravity from the involved side toward the noninvolved side. The following descriptions of the 3-dimensional movement pattern changes of each part of the body were based on detailed analysis. Trunk. The trunk was maintained in the neutral position in the mediolateral (M/L) and transverse planes during PWB crutch walking, whereas it had rotation and up and down movement in normal walking. In the sagittal plane, the trunk maintained a forward lean position of approximately 8° during PWB, but it was maintained at a position of a 2° backward lean during normal walking. Pelvis. In the sagittal plane, the pelvis tilted forward approximately 22° on both sides during PWB crutch walking throughout the gait cycle, approximately 7° more than the normal position. The movement pattern of the pelvis in the M/L and transverse planes for PWB crutch walking resembled that of normal walking. Hip. The main changes were in the transverse plane; the involved side was internally rotated to 2° to 6° (fig 1A), whereas the noninvolved side was rotated in the opposite direction, with external rotation at approximately 2° to 6° (fig 1B). In the sagittal plane, the patterns of hip flexion and extension did not change, but hip joints of both sides could not reach to the full extension position through the gait cycle because the pelvis was kept in a forward position with limited hip extension. Hip extension increased as the amount of weight bearing increased for both sides. The involved side was maintained in slight adduction (fig 2A) and the noninvolved side in slight abduction (fig 2B) at foot strike in the M/L plane. Knee. In general, compared with the noninvolved side, the main changes in ROM were on the involved side during the different walking conditions. For the sagittal plane, the involved side had a ROM of approximately 52° in normal walking, but it had only approximately 40° of ROM for crutch walking, with some variations among different weight-bearing conditions (table 2). The patterns for the involved side also changed from normal walking to crutch walking; the typical knee flexion in the midstance phase disappeared for crutch walking. The initial knee flexion angle also decreased as the percentage of weight bearing increased for the involved side. Ankle and foot. The main changes were around the toe-off phase for the ankle and foot on the involved side in both dorsiflexion-plantarflexion and supination-pronation. The ankle was maintained in the dorsiflexed position, and the foot in a pronated position. There were small variations among PWB Arch Phys Med Rehabil Vol 82, January 2001
status with crutch walking, but the patterns were similar. However, the noninvolved side was nearly normal. GRF changes. Vertical GRFs for both sides are shown in figure 3, and peak values for GRF in each plane are listed in table 2. Real vertical GRFs on the involved side from the different levels of PWB crutch walking varied and were inconsistent with expected levels of PWB. For example, an expected 10% PWB had a real vertical GRF of 36% ⫾ 16% of body weight (BW), 90% PWB had a real vertical GRF of 62% ⫾ 20% of BW, and 50% PWB had an expected value of 50% ⫾ 25% of BW. However, GRF also changed on the noninvolved side. The noticeable change was a relatively constant loading phase during stance. Another interesting observation is that the value of the vertical GRF on the noninvolved side did not appear to be influenced to the same degree as the anteroposterior (AP) GRF. There were also wide variations among subjects. Changes of GRF in the AP and M/L planes were mainly in values, not in patterns. There was a large decrease in AP GRF on the involved side, from 16% of BW during normal walking to 5% of BW during crutch walking. Decreases in M/L GRF were close to 50% in value, but without substantial changes in the patterns (table 2). DISCUSSION Support time (stance phase) significantly decreased on the involved side and significantly increased on the noninvolved side; consequently, nonsupport time (swing phase) decreased. During 3-point gait pattern crutch walking, double support time for the involved side was shared by the noninvolved leg, and the remainder of the support time (single support phase) for the involved side was shared by the crutches and involved side. Therefore, the shared loading period for the involved side increased in duration, and the noninvolved side shared some of the loading of the involved side. This may help explain the GRF patterns presented in later discussions. Kinematic changes resulted from the adaptation to crutch walking and were obvious on both sides. The trunk and pelvis were maintained in a position that resulted in a forward lean with the pelvis tilted forward. Although the trunk and pelvis were relatively fixed in the sagittal plane, the pelvis had relatively normal movement in the other 2 planes. The use of crutches limited the ROM of the trunk and pelvis, diminishing the lateral movement and rotation. Compared with normal walking, the involved side during crutch walking had diminished hip flexion and adduction, less knee flexion, and decreased ankle plantarflexion at the toe-off, and the foot remained in pronation throughout the gait cycle. The noninvolved side showed slightly greater hip abduction, external rotation, and knee flexion. The pelvis showed a relatively normal movement pattern in rotation and M/L movement. There was slightly greater hip abduction and external rotation on the noninvolved side and slightly less hip adduction and internal rotation on the involved side. This indicated a shift of the center of gravity from the involved side slightly toward the other side. This shift would help diminish weight bearing on the involved side. In 3-point gait, the crutches moved forward together with the involved side. Therefore, the crutches helped propel the body forward at the toe-off, reducing the need for the ankle plantarflexion that is necessary during normal walking. This change was also related to GRF patterns on the involved side, discussed later. GRF indicated a tremendous change between normal walking and PWB crutch walking. The subjects could not perform PWB at the target levels set for them. This result matched those of other studies6,11 indicating an inability of subjects to produce
THREE-POINT GAIT CRUTCH WALKING, Li
89
Fig 1. Hip joint rotation (deg) in the transverse plane (N, normal walking; L, 10%PWB; M, 50%PWB; H, 90%PWB). (A) The involved side has internal rotation of approximately 2° to 6°, whereas (B) the noninvolved side has external rotation of approximately 2° to 6° during walking.
PWB accurately. Apparently, it was easier for the subjects to perform 50% PWB than either of the other 2 levels. A possible explanation for the subjects’ inability to match the requisite levels of weight bearing may be that without pathologic states, they were not receiving sensory (pain) feedback from the involved limb. This could be the reason subjects had large variations in controlling loading during PWB. In crutch walking, both the shoulders and hands are involved in reducing loading on the involved limb. At the extremes (10% and 90% PWB), sensory feedback from the shoulders and hands may have been more difficult to interpret than feedback from the intermediate level (50% PWB). This may account for the variability in accuracy of weight-bearing reproduction across these levels. For the involved side during PWB, there was no obvious difference in GRF except for the vertical component. Vertical GRF changed in both value and patterns for both sides. But AP and M/L GRFs changed only in value, not in terms of patterns. For both sides, the vertical GRF peak values diminished; there
also was a slight reduction in value for the noninvolved side during crutch walking. One explanation is simply that part of the BW is transmitted to the crutches.3 Based on our findings, we could add another factor: the shift of the center of gravity toward the noninvolved side. This means that loading is more dependent on the noninvolved than involved side. In addition, in terms of variation in the patterns of the GRFs, 2 other factors may explain these changes. First, a larger initial knee angle (IKA) occurred during crutch walking, which may have had a role in reducing impact force.12 LaFortune et al12 found that IKA correlated with the impact force: the larger the IKA, the lower the impact force. Second, because the crutches continuously supported the BW during the stance phase for the involved side, particularly at the toe-off phase, vertical GRF showed no second loading peak on the involved side during crutch walking. It is difficult to explain why there was no increase in vertical GRF and no peak in the loading phase on the noninvolved side. According to Nilsson and Thorstensson,13 increased velocity Arch Phys Med Rehabil Vol 82, January 2001
90
THREE-POINT GAIT CRUTCH WALKING, Li
Fig 2. Hip joint abduction and adduction (deg) in the M/L plane (N, normal walking; L, 10%PWB; M, 50%PWB; H, 90%PWB). (A) The involved side is maintained in slight adduction, and (B) the noninvolved side is maintained in slightly abducted position at foot strike in the M/L plane.
Table 2: Peak Values of GRF in 3 Directions Normal
V GRF max V GRF min M/L GRF Max M/L GRF min AP GRF max AP GRF min
10% PWB
Involved
Noninvolved
Involved
.98 ⫾ .22 0⫾0 .05 ⫾ .02 0 ⫾ .02 .16 ⫾ .04 0 ⫾ .05
1.02 ⫾ .23 0⫾0 .05 ⫾ .02 0 ⫾ .01 .17 ⫾ .04 0 ⫾ .03
.36 ⫾ .16 0⫾0 .02 ⫾ .01 0⫾0 .03 ⫾ .02 0 ⫾ .01
Values expressed as mean ⫾ SD of BW. Abbreviations: V, vertical; max, maximum; min, minimum.
Arch Phys Med Rehabil Vol 82, January 2001
Noninvolved
.75 ⫾ .42 0⫾0 .02 ⫾ .02 0 ⫾ .01 0.1 ⫾ .07 0 ⫾ .04
50% PWB Involved
0.5 ⫾ .25 0⫾0 .03 ⫾ .01 0 ⫾ .01 .05 ⫾ .04 0 ⫾ .04
Noninvolved
.85 ⫾ .36 0⫾0 .02 ⫾ .01 0 ⫾ .01 .11 ⫾ .06 0 ⫾ .04
90% PWB Involved
.62 ⫾ 0.2 0⫾0 .03 ⫾ .02 0 ⫾ .01 .07 ⫾ .03 0 ⫾ .04
Noninvolved
.96 ⫾ .21 0⫾0 .03 ⫾ .02 0 ⫾ .01 .13 ⫾ .05 0 ⫾ .03
THREE-POINT GAIT CRUTCH WALKING, Li
91
Fig 3. Vertical GRF (BW) on both sides (N, normal walking; L, 10%PWB; M, 50%PWB; H, 90%PWB). (A) Vertical GRF on the involved side is not consistent with expected levels. (B) No increases in vertical GRF occur on the noninvolved side.
could cause an increase in all components of GRF. Compared with normal walking, decreased velocity was found during crutch walking. Also, AP and M/L GRFs on the noninvolved side decreased in this study. Furthermore, based on our findings, another factor could contribute to these changes. Some of the double support time for the noninvolved side was shared by the crutches, the noninvolved leg, and the involved leg. Thus, the crutches were supporting part of the BW, leading to a reduction in vertical GRF on the noninvolved side. During double support time for the noninvolved side, the first part of this duration was from foot strike of the noninvolved side to the push-off of the involved side, and the second part was from foot strike of the involved side to push-off of the noninvolved side. These 2 phases represent the period when the loading peaks normally occur. If BW is shared by the crutches during these periods, vertical GRF on the noninvolved side would then
be reduced. Additionally, there would be diminished peaks and a resultant relatively constant loading phase. Therefore, when double support time for the involved side increased, the time shared by the crutches and the noninvolved side increased, resulting in a larger reduction in GRF on the noninvolved side and leading into a relative constant loading plateau. Decreases in AP and M/L GRF components on the involved side were probably caused by less weight bearing and decreased velocity. Lack of increases in GRF on the noninvolved side also contrasted with findings reported in the literature. Based on the study of Stallard et al,6 there would be one quarter to one third increases in vertical GRF in the landing leg (the noninvolved side) during swing-through crutch walking. In the swingthrough gait pattern, 2 crutches are placed in front of the body and the body is propelled through them, with 1 or 2 legs landing ahead of the crutches. However, the crutches were Arch Phys Med Rehabil Vol 82, January 2001
92
THREE-POINT GAIT CRUTCH WALKING, Li
placed in front of the body simultaneously with the involved leg in a 3-point gait pattern. This difference in gait pattern results in the crutches sharing BW with the landing leg during 3-point crutch gait, in contrast to what occurs during swingthrough crutch walking. Because the body is pulled through the crutches and has a short duration of double support, the body has a relative large landing speed during the swing-through gait pattern. This may explain the difference in the vertical GRF observed in this study compared with that of swing-through gait crutch walking. CONCLUSION Our subjects could not accurately reproduce PWB level of 3-point crutch walking, particularly during high-level (90%) or low-level (10%) PWB crutch walking. During 3-point crutch walking, the involved side has a shorter stance phase and a longer swing phase than that of the noninvolved side. There are large individual variations in GRF at each level of PWB during 3-point crutch walking. Decreases in GRF on the involved side result from a shift of the center of gravity toward the noninvolved side, transmission of part of BW by the crutches, and decreased velocity. Decreased GRF on the noninvolved side results from the crutches supporting part of BW, as well as decreased velocity. This study has clinical implications. Because patients cannot control the load on the involved side as prescribed, therapists and physicians must consider this situation when they prescribe crutches. A shift of the center of gravity toward the noninvolved side could help reduce the percentage of weight bearing of the involved side. It might be helpful if clinicians instructed patients to lean on the noninvolved side to a certain degree in attempting to control the target level of PWB. References 1. Duesterhaus Minor MA, Duesterhaus Minor S. Patient care skills. 3rd ed. Norwalk (CT): Appleton & Lange; 1995.
Arch Phys Med Rehabil Vol 82, January 2001
2. Harkess JW, Ramsey WC, Harkess JW. Principles of fracture and dislocation. In: Rockwood CA, Green DP, Bucholz RW, editors. Rockwood and Green’s fractures in adults. 4th ed. Philadelphia: Lippincott-Raven; 1996. p. 102-7. 3. Whittle MW. Gait analysis: an introduction. 2nd ed. Oxford: Butterworth-Heinemann; 1996. 4. Winstein CJ, Pohl PS, Cardinale C, Green A, Schlotz L, Waters CS. Learning a partial-weight-bearing skill: effectiveness of two forms of feedback. Phys Ther 1996;76:985-93. 5. Bhambhami YN, Clarkson HM, Gomes PS. Axillary crutch walking: effects of three training programs. Arch Phys Med Rehabil 1990;71:484-9. 6. Stallard J, Sankarankutty M, Rose GK. Lower-limb vertical ground-reaction forces during crutch walking. J Med Eng Technol 1978;2:201-3. 7. Feldman DR, Vujic I, Mckay D, Callcott F, Uflacker R. Crutchinduced axillary artery injury. Cardiovasc Intervent Radiol 1995; 18:296-9. 8. Raikin S, Froimson MI. Bilateral brachial plexus compressive neuropathy (crutch palsy). J Orthop Trauma 1997;11:136-8. 9. Shabas D, Scheiber M. Suprascapular neuropathy related to the use of crutches. Am J Phys Med 1986;65:298-300. 10. Tripp HF, Cook JW. Axillary artery aneurysms. Mil Med 1998; 163:653-5. 11. Baxter ML, Allington RO, Koepke GH. Weight-distribution variables in the use of crutches and canes. Phys Ther 1969;49:360-5. 12. LaFortune MA, Hennig EM, Lake MJ. Dominant role of interface over knee angle for cushioning impact loading and regulating initial leg stiffness. J Biomech 1996;29:1523-9. 13. Nilsson J, Thorstensson A. Ground reaction forces at different speeds of human walking and running. Acta Physiol Scand 1989; 136: 217-27. Suppliers a. dOR-5; Advanced Mechanical Technology, Inc, 176 Waltham St, Watertown, MA 02472. b. National Instruments, Corp, 11500 N Mopac Expwy, Austin, TX 78759. c. Motion Analysis, Corp, 3617 Westwind Blvd, Santa Rosa, CA 95403. d. SPSS, version 9.0; SPSS, Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
Three-Point Gait Crutch Walking: Variability in Ground Reaction Force During Weight Bearing Sheng Li, MD, Charles W. Armstrong, PhD, Daniel Cipriani, MEd, PT ABSTRACT. Li S, Armstrong CW, Cipriani D. Three-point gait crutch walking: variability in ground reaction force during weight bearing. Arch Phys Med Rehabil 2001;82:86-92. Objective: To investigate variability in ground reaction force (GRF) and kinematics on both sides during 3-point partial weight-bearing (PWB) crutch walking. Design: Within-subject comparisons of kinematic and kinetic data collected at different levels of 3-point crutch walking: 10%, 50%, and 90% PWB at comfortable speeds. Setting: An applied biomechanics lab in a university setting. Participants: Twelve healthy college students (9 women, 3 men). Main Outcome Measures: Spatial and temporal variables, major peak kinematic data, and peak GRFs from force platforms during the gait cycle. Results: Large variations were found in replicating the target levels of PWB, particularly at 10% and 90% PWB. Subjects had a shorter stance phase and longer swing phase during the crutch walking gait cycle. Velocity significantly decreased (p ⫽ .006) because of decreased cadence (p ⫽ .002). Slightly greater hip abduction and external rotation on the noninvolved side and slightly less hip adduction and internal rotation on the involved side indicated that the center of gravity shifted slightly from the involved side toward the noninvolved side. There was no increase in vertical GRF, and there was a relatively constant loading pattern on the noninvolved side. Conclusions: Subjects have difficulty replicating a prescribed weight-bearing restriction. A shift of the center of gravity toward the noninvolved side may reduce the weight distribution on the involved side. Key Words: Crutches; Gait; Rehabilitation; Walking; Weight bearing. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation RAUMA TO A LOWER EXTREMITY may impair its ability to bear weight. Ankle sprains, fractures, and knee T injuries typically force an individual to adopt a protective weight-bearing status until there is sufficient healing.1,2 Weight-bearing restrictions depend on treatment needs, which range from non–weight bearing to partial weight bearing (PWB) to full weight bearing. PWB helps patients in different
From the Department of Kinesiology, University of Toledo (Li, Armstrong); and the Department of Physical Therapy, Medical College of Ohio (Cipriani), Toledo, OH. Accepted in revised form March 28, 2000. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Sheng Li, MD, Dept of Kinesiology, Room 20-Recreation Hall, Pennsylvania State University, University Park, Pennsylvania, PA 16801, e-mail: sx1282.psu.edu. 0003-9993/01/8201-5794$35.00/0 doi:10.1053/apmr.2001.16347
Arch Phys Med Rehabil Vol 82, January 2001
stages of rehabilitation, with the amount of weight bearing based on the severity of the injury and healing status. Axillary crutches are designed to help individuals walk with reduced weight bearing on the involved lower extremity. They also help an individual maintain balance3 and can help relieve pain in the hip, knee, ankle, and foot.1 Different crutch walking gait patterns include 2-point, 3-point, 4-point, swing-to, and swing-through patterns. Three-point gait crutch walking is commonly used because it provides for varied levels of weight bearing, from non–weight bearing to full weight bearing.2 In this pattern, both crutches are advanced simultaneously with the involved side, and the noninvolved side is then advanced. Winstein et al4 and Bhambhami et al5 found that individuals were not reliable in following weight-bearing restrictions. After 80 trials with feedback from a floor scale during swing-to crutch walking, Winstein et al4 found that individuals still had an error rate of approximately 10% in replicating a target PWB level. Patients usually use crutches without having training in PWB. This likely results in high error rates in replicating prescribed levels of weight bearing. Although research on crutch walking has generally focused on the involved limb, adaptive compensatory changes in the gait pattern of the noninvolved limb may also be of concern. Stallard et al6 investigated ground reactive force (GRF) during swing-through gait crutch walking and found that individuals tended to land with 25% greater vertical force in a single-leg landing and nearly 33% greater force when landing on both legs than when walking normally. These findings are significant in terms of potential stress on the ipsilateral noninvolved leg because the use of crutches may exacerbate such existing conditions as osteoarthritis or rheumatoid arthritis. Stress on the noninvolved leg emphasizes the importance of measuring the forces on both legs during crutch walking. The potential of crutch walking for inducing injuries is not limited to the lower extremity.7-10 Multiple forms of upper extremity injuries, including injuries to the brachial plexus, hand neuropathies, and brachial artery occlusions, have been identified in the literature. Collectively, the issues associated with the pathomechanics of crutch walking are important clinically, considering the common use of axillary crutches and the potential for indirect injury. There have been many studies of crutch walking. Unfortunately, there are few data about PWB. Thus, this study examines the biomechanical characteristics of a 3-point crutch walking gait pattern in 3 different weight-bearing conditions: 10%, 50%, and 90% weight-bearing status. By examining the associated kinetic, kinematic, and temporal and distance characteristics of gait under prescribed levels of weight bearing, we hoped to gain insight into the pathomechanics of crutch walking gait. METHODS Subjects Nine women and 3 men (mean age ⫾ standard deviation, 25 ⫾ 7.6yr; height, 164 ⫾ 4cm; weight, 54 ⫾ 3.1kg) were recruited for this study. To minimize variability in response to
87
THREE-POINT GAIT CRUTCH WALKING, Li
experimental conditions, only subjects with no known history of musculoskeletal, neurologic, cardiovascular, or pulmonary disorders were included. Individuals with muscle weakness or length discrepancy in lower limbs were excluded. Each subject underwent a brief physical examination. Before testing, a written consent form explaining the study, which was approved by the university’s human subject research review committee, was given to each subject. Instrumentation GRF data during walking was recorded by 2 standard strain gauge force platformsa that were embedded in the center of a 10-meter walkway. The boundary of each force platform was identified to the subjects by the tester. Six raw voltage signals from the force platforms were transmitted to the computer through an analog-to-digital converterb at a sampling rate of 480Hz. Kinematic data were collected by a computerized video 3-dimensional analysis system (Hires 3D).c Six cameras (Falcon Hires)c arranged around the walkway at approximately 60° intervals captured the subject’s motion at 60Hz using passive retroflective markers placed on the subject. Markers were applied to the subject according to the Helen Hayes Static marker set, bilaterally on the wrists, elbows, shoulders, anterosuperior iliac spine, midthighs, medial and lateral knee centers, midshanks, medial and lateral ankle centers, heels and forefeet, and sacrum. Video data were digitized and processed using EVA 4.01 software.c All kinematic data were smoothed with a low pass filter using a cutoff at 6Hz. A static test identified the location of each marker on the body for each subject. Video and GRF data were output to OrthoTrak software version 4.1,c which calculated the kinetic, kinematic, and temporal and distance measures. Procedures Two days before they were tested, the subjects were shown how to use adjustable axillary crutches appropriately for a 3-point gait pattern. First, crutch length was adjusted to match each subject’s height, with approximately 1 hand width between the axilla and the top of the crutches.2 Second, based on each subject’s weight, 3 levels of target weight for PWB were determined on a standard scale, 10% (low), 50% (medium), and 90% (high) body weight. Subjects were allowed a few practice trials on the scale with concurrent feedback to feel the requisite amount of weight. They were encouraged to walk using a 3-point gait pattern. During PWB crutch walking, 1 leg was randomly assumed to be injured and to take PWB only. That leg was identified as the involved leg (or side), and the other side was the noninvolved side.
After being instructed on 3-point gait crutch walking, subjects were asked to practice it at various levels of PWB for 30 minutes each day at home for 2 days. They then returned for the crutch walking gait assessment, at which time they practiced 3-point crutch walking for approximately 5 minutes before measurements were taken. The subjects reported having completed a mean amount of 45.4 ⫾ 16.8 minutes of practice time and were now comfortable and familiar with crutch walking. On the test day, each subject performed 8 normal walking trials, followed by 8 trials at each level of the PWB crutch walking tests. After the system was calibrated, hemispheric, reflective markers were placed on the subject’s skin bilaterally. A static test was performed before the trial and was used as a calibration and standard. Additionally, each subject’s weight and height were measured and recorded. To ensure a gait typical of their usual walking pattern, subjects were instructed to walk barefoot at their own selected walking speeds along the walkway. They were given multiple practice trials before testing to ensure that they were walking at a normal rhythm. Only those trials in which each foot contacted a force platform separately and that were free of contact between the crutches and the force plates were retained for analysis. Subjects were tested in the same order: normal walking first, followed by 10%, 50%, and 90% PWB crutch walking. After the normal walking test, subjects walked repeatedly across the walkway for approximately 5 minutes with a 3-point pattern of crutch walking at different levels of PWB. Eight trials at each level were performed in the same way as the normal walking trials. Statistical Analysis Group means for each variable were based on the average of 8 trails for each subject for each condition. Standard descriptive statistics followed by repeated-measures analysis of variance were used in the statistical analysis. The main factors were SIDE (2 levels: left, right) and PWB (4 levels: normal, 10%, 50%, 90% PWB).d Tukey post hoc comparisons were used to compare individual cell means. RESULTS Temporal and spatial changes, including cadence, stride length, velocity, step width, double support time, support time, and nonsupport time in the 3-point gait crutch walking are listed in table 1. Compared with normal walking, the cadence (F3,33 ⫽ 14.76; p ⫽ .002) and velocity (F3,33 ⫽ 10.33; p ⫽ .005) of both sides were decreased significantly in PWB crutch walking. Statistical analysis also showed no significant differ-
Table 1: Summary of Spatial Parameters Normal Involved
Noninvolved
10% PWB Involved
Noninvolved
Cadence 111.9 ⫾ 11.35 110.4 ⫾ 10.76 74.5 ⫾ 11.92 70.9 ⫾ 11.05 (steps/s) Stride length 126.1 ⫾ 9.47 124.7 ⫾ 9.54 116.7 ⫾ 19.55 111.8 ⫾ 7.49 (cm) Velocity (cm/s) 118.3 ⫾ 18.24 115.2 ⫾ 17.14 70.8 ⫾ 15.8 66.5 ⫾ 12.52 Step width (cm) 10.6 ⫾ 2.26 9.7 ⫾ 2.26 Double support 9.7 ⫾ 2.19 10.9 ⫾ 1.84 24 ⫾ 27.84 14.3 ⫾ 11.99 time (%) Support time (%) 60.3 ⫾ 1.5 60.6 ⫾ 1.69 40.6 ⫾ 14.58 67.8 ⫾ 4
50% PWB Involved
75.6 ⫾ 15.22 118 ⫾ 14.32
Noninvolved
70.1 ⫾ 10.99
90% PWB Involved
71 ⫾ 13.59
Noninvolved
70.7 ⫾ 13.05
113.7 ⫾ 11.07 118.8 ⫾ 13.74 115.7 ⫾ 10.9
71.2 ⫾ 15.62 66.3 ⫾ 13.47 9.6 ⫾ 1.93 25.3 ⫾ 30.96 12.2 ⫾ 7.06
70.4 ⫾ 15.41 68.2 ⫾ 14.49 9 ⫾ 1.81 28.6 ⫾ 28.52 10.1 ⫾ 1.86
43.8 ⫾ 13.05
45.9 ⫾ 16.71
66.2 ⫾ 3.6
65 ⫾ 2.52
Values expressed as mean ⫾ standard deviation (SD), n ⫽ 12.
Arch Phys Med Rehabil Vol 82, January 2001
88
THREE-POINT GAIT CRUTCH WALKING, Li
ence between the involved and noninvolved sides. Reductions in velocity in both sides resulted from decreases in cadence. Step width significantly decreased in PWB crutch walking compared with normal walking (F3,33 ⫽ 15.53; p ⫽ .001). There was no significant difference in double support time between the 2 sides in the different walking conditions. Support time for the involved side and noninvolved side were significantly different (F3,33 ⫽ 19.35; p ⫽ .002), but crutch use did not have a significant influence on the different levels of PWB crutch walking. Support time (stance phase) for the involved side significantly decreased, ie, the percentage of the stance phase in the gait cycle decreased for the involved side, and there was no significant difference among different levels of PWB crutch walking. However, double support time did not have significant changes. As indicated, there were significant changes in support time for the involved side. Furthermore, the range-of-motion (ROM) patterns of the pelvis, hip, knee, ankle, and foot for both sides changed during crutch walking. The results of these kinematic changes appeared to be linked to shifts of the center of gravity from the involved side toward the noninvolved side. The following descriptions of the 3-dimensional movement pattern changes of each part of the body were based on detailed analysis. Trunk. The trunk was maintained in the neutral position in the mediolateral (M/L) and transverse planes during PWB crutch walking, whereas it had rotation and up and down movement in normal walking. In the sagittal plane, the trunk maintained a forward lean position of approximately 8° during PWB, but it was maintained at a position of a 2° backward lean during normal walking. Pelvis. In the sagittal plane, the pelvis tilted forward approximately 22° on both sides during PWB crutch walking throughout the gait cycle, approximately 7° more than the normal position. The movement pattern of the pelvis in the M/L and transverse planes for PWB crutch walking resembled that of normal walking. Hip. The main changes were in the transverse plane; the involved side was internally rotated to 2° to 6° (fig 1A), whereas the noninvolved side was rotated in the opposite direction, with external rotation at approximately 2° to 6° (fig 1B). In the sagittal plane, the patterns of hip flexion and extension did not change, but hip joints of both sides could not reach to the full extension position through the gait cycle because the pelvis was kept in a forward position with limited hip extension. Hip extension increased as the amount of weight bearing increased for both sides. The involved side was maintained in slight adduction (fig 2A) and the noninvolved side in slight abduction (fig 2B) at foot strike in the M/L plane. Knee. In general, compared with the noninvolved side, the main changes in ROM were on the involved side during the different walking conditions. For the sagittal plane, the involved side had a ROM of approximately 52° in normal walking, but it had only approximately 40° of ROM for crutch walking, with some variations among different weight-bearing conditions (table 2). The patterns for the involved side also changed from normal walking to crutch walking; the typical knee flexion in the midstance phase disappeared for crutch walking. The initial knee flexion angle also decreased as the percentage of weight bearing increased for the involved side. Ankle and foot. The main changes were around the toe-off phase for the ankle and foot on the involved side in both dorsiflexion-plantarflexion and supination-pronation. The ankle was maintained in the dorsiflexed position, and the foot in a pronated position. There were small variations among PWB Arch Phys Med Rehabil Vol 82, January 2001
status with crutch walking, but the patterns were similar. However, the noninvolved side was nearly normal. GRF changes. Vertical GRFs for both sides are shown in figure 3, and peak values for GRF in each plane are listed in table 2. Real vertical GRFs on the involved side from the different levels of PWB crutch walking varied and were inconsistent with expected levels of PWB. For example, an expected 10% PWB had a real vertical GRF of 36% ⫾ 16% of body weight (BW), 90% PWB had a real vertical GRF of 62% ⫾ 20% of BW, and 50% PWB had an expected value of 50% ⫾ 25% of BW. However, GRF also changed on the noninvolved side. The noticeable change was a relatively constant loading phase during stance. Another interesting observation is that the value of the vertical GRF on the noninvolved side did not appear to be influenced to the same degree as the anteroposterior (AP) GRF. There were also wide variations among subjects. Changes of GRF in the AP and M/L planes were mainly in values, not in patterns. There was a large decrease in AP GRF on the involved side, from 16% of BW during normal walking to 5% of BW during crutch walking. Decreases in M/L GRF were close to 50% in value, but without substantial changes in the patterns (table 2). DISCUSSION Support time (stance phase) significantly decreased on the involved side and significantly increased on the noninvolved side; consequently, nonsupport time (swing phase) decreased. During 3-point gait pattern crutch walking, double support time for the involved side was shared by the noninvolved leg, and the remainder of the support time (single support phase) for the involved side was shared by the crutches and involved side. Therefore, the shared loading period for the involved side increased in duration, and the noninvolved side shared some of the loading of the involved side. This may help explain the GRF patterns presented in later discussions. Kinematic changes resulted from the adaptation to crutch walking and were obvious on both sides. The trunk and pelvis were maintained in a position that resulted in a forward lean with the pelvis tilted forward. Although the trunk and pelvis were relatively fixed in the sagittal plane, the pelvis had relatively normal movement in the other 2 planes. The use of crutches limited the ROM of the trunk and pelvis, diminishing the lateral movement and rotation. Compared with normal walking, the involved side during crutch walking had diminished hip flexion and adduction, less knee flexion, and decreased ankle plantarflexion at the toe-off, and the foot remained in pronation throughout the gait cycle. The noninvolved side showed slightly greater hip abduction, external rotation, and knee flexion. The pelvis showed a relatively normal movement pattern in rotation and M/L movement. There was slightly greater hip abduction and external rotation on the noninvolved side and slightly less hip adduction and internal rotation on the involved side. This indicated a shift of the center of gravity from the involved side slightly toward the other side. This shift would help diminish weight bearing on the involved side. In 3-point gait, the crutches moved forward together with the involved side. Therefore, the crutches helped propel the body forward at the toe-off, reducing the need for the ankle plantarflexion that is necessary during normal walking. This change was also related to GRF patterns on the involved side, discussed later. GRF indicated a tremendous change between normal walking and PWB crutch walking. The subjects could not perform PWB at the target levels set for them. This result matched those of other studies6,11 indicating an inability of subjects to produce
THREE-POINT GAIT CRUTCH WALKING, Li
89
Fig 1. Hip joint rotation (deg) in the transverse plane (N, normal walking; L, 10%PWB; M, 50%PWB; H, 90%PWB). (A) The involved side has internal rotation of approximately 2° to 6°, whereas (B) the noninvolved side has external rotation of approximately 2° to 6° during walking.
PWB accurately. Apparently, it was easier for the subjects to perform 50% PWB than either of the other 2 levels. A possible explanation for the subjects’ inability to match the requisite levels of weight bearing may be that without pathologic states, they were not receiving sensory (pain) feedback from the involved limb. This could be the reason subjects had large variations in controlling loading during PWB. In crutch walking, both the shoulders and hands are involved in reducing loading on the involved limb. At the extremes (10% and 90% PWB), sensory feedback from the shoulders and hands may have been more difficult to interpret than feedback from the intermediate level (50% PWB). This may account for the variability in accuracy of weight-bearing reproduction across these levels. For the involved side during PWB, there was no obvious difference in GRF except for the vertical component. Vertical GRF changed in both value and patterns for both sides. But AP and M/L GRFs changed only in value, not in terms of patterns. For both sides, the vertical GRF peak values diminished; there
also was a slight reduction in value for the noninvolved side during crutch walking. One explanation is simply that part of the BW is transmitted to the crutches.3 Based on our findings, we could add another factor: the shift of the center of gravity toward the noninvolved side. This means that loading is more dependent on the noninvolved than involved side. In addition, in terms of variation in the patterns of the GRFs, 2 other factors may explain these changes. First, a larger initial knee angle (IKA) occurred during crutch walking, which may have had a role in reducing impact force.12 LaFortune et al12 found that IKA correlated with the impact force: the larger the IKA, the lower the impact force. Second, because the crutches continuously supported the BW during the stance phase for the involved side, particularly at the toe-off phase, vertical GRF showed no second loading peak on the involved side during crutch walking. It is difficult to explain why there was no increase in vertical GRF and no peak in the loading phase on the noninvolved side. According to Nilsson and Thorstensson,13 increased velocity Arch Phys Med Rehabil Vol 82, January 2001
90
THREE-POINT GAIT CRUTCH WALKING, Li
Fig 2. Hip joint abduction and adduction (deg) in the M/L plane (N, normal walking; L, 10%PWB; M, 50%PWB; H, 90%PWB). (A) The involved side is maintained in slight adduction, and (B) the noninvolved side is maintained in slightly abducted position at foot strike in the M/L plane.
Table 2: Peak Values of GRF in 3 Directions Normal
V GRF max V GRF min M/L GRF Max M/L GRF min AP GRF max AP GRF min
10% PWB
Involved
Noninvolved
Involved
.98 ⫾ .22 0⫾0 .05 ⫾ .02 0 ⫾ .02 .16 ⫾ .04 0 ⫾ .05
1.02 ⫾ .23 0⫾0 .05 ⫾ .02 0 ⫾ .01 .17 ⫾ .04 0 ⫾ .03
.36 ⫾ .16 0⫾0 .02 ⫾ .01 0⫾0 .03 ⫾ .02 0 ⫾ .01
Values expressed as mean ⫾ SD of BW. Abbreviations: V, vertical; max, maximum; min, minimum.
Arch Phys Med Rehabil Vol 82, January 2001
Noninvolved
.75 ⫾ .42 0⫾0 .02 ⫾ .02 0 ⫾ .01 0.1 ⫾ .07 0 ⫾ .04
50% PWB Involved
0.5 ⫾ .25 0⫾0 .03 ⫾ .01 0 ⫾ .01 .05 ⫾ .04 0 ⫾ .04
Noninvolved
.85 ⫾ .36 0⫾0 .02 ⫾ .01 0 ⫾ .01 .11 ⫾ .06 0 ⫾ .04
90% PWB Involved
.62 ⫾ 0.2 0⫾0 .03 ⫾ .02 0 ⫾ .01 .07 ⫾ .03 0 ⫾ .04
Noninvolved
.96 ⫾ .21 0⫾0 .03 ⫾ .02 0 ⫾ .01 .13 ⫾ .05 0 ⫾ .03
THREE-POINT GAIT CRUTCH WALKING, Li
91
Fig 3. Vertical GRF (BW) on both sides (N, normal walking; L, 10%PWB; M, 50%PWB; H, 90%PWB). (A) Vertical GRF on the involved side is not consistent with expected levels. (B) No increases in vertical GRF occur on the noninvolved side.
could cause an increase in all components of GRF. Compared with normal walking, decreased velocity was found during crutch walking. Also, AP and M/L GRFs on the noninvolved side decreased in this study. Furthermore, based on our findings, another factor could contribute to these changes. Some of the double support time for the noninvolved side was shared by the crutches, the noninvolved leg, and the involved leg. Thus, the crutches were supporting part of the BW, leading to a reduction in vertical GRF on the noninvolved side. During double support time for the noninvolved side, the first part of this duration was from foot strike of the noninvolved side to the push-off of the involved side, and the second part was from foot strike of the involved side to push-off of the noninvolved side. These 2 phases represent the period when the loading peaks normally occur. If BW is shared by the crutches during these periods, vertical GRF on the noninvolved side would then
be reduced. Additionally, there would be diminished peaks and a resultant relatively constant loading phase. Therefore, when double support time for the involved side increased, the time shared by the crutches and the noninvolved side increased, resulting in a larger reduction in GRF on the noninvolved side and leading into a relative constant loading plateau. Decreases in AP and M/L GRF components on the involved side were probably caused by less weight bearing and decreased velocity. Lack of increases in GRF on the noninvolved side also contrasted with findings reported in the literature. Based on the study of Stallard et al,6 there would be one quarter to one third increases in vertical GRF in the landing leg (the noninvolved side) during swing-through crutch walking. In the swingthrough gait pattern, 2 crutches are placed in front of the body and the body is propelled through them, with 1 or 2 legs landing ahead of the crutches. However, the crutches were Arch Phys Med Rehabil Vol 82, January 2001
92
THREE-POINT GAIT CRUTCH WALKING, Li
placed in front of the body simultaneously with the involved leg in a 3-point gait pattern. This difference in gait pattern results in the crutches sharing BW with the landing leg during 3-point crutch gait, in contrast to what occurs during swingthrough crutch walking. Because the body is pulled through the crutches and has a short duration of double support, the body has a relative large landing speed during the swing-through gait pattern. This may explain the difference in the vertical GRF observed in this study compared with that of swing-through gait crutch walking. CONCLUSION Our subjects could not accurately reproduce PWB level of 3-point crutch walking, particularly during high-level (90%) or low-level (10%) PWB crutch walking. During 3-point crutch walking, the involved side has a shorter stance phase and a longer swing phase than that of the noninvolved side. There are large individual variations in GRF at each level of PWB during 3-point crutch walking. Decreases in GRF on the involved side result from a shift of the center of gravity toward the noninvolved side, transmission of part of BW by the crutches, and decreased velocity. Decreased GRF on the noninvolved side results from the crutches supporting part of BW, as well as decreased velocity. This study has clinical implications. Because patients cannot control the load on the involved side as prescribed, therapists and physicians must consider this situation when they prescribe crutches. A shift of the center of gravity toward the noninvolved side could help reduce the percentage of weight bearing of the involved side. It might be helpful if clinicians instructed patients to lean on the noninvolved side to a certain degree in attempting to control the target level of PWB. References 1. Duesterhaus Minor MA, Duesterhaus Minor S. Patient care skills. 3rd ed. Norwalk (CT): Appleton & Lange; 1995.
Arch Phys Med Rehabil Vol 82, January 2001
2. Harkess JW, Ramsey WC, Harkess JW. Principles of fracture and dislocation. In: Rockwood CA, Green DP, Bucholz RW, editors. Rockwood and Green’s fractures in adults. 4th ed. Philadelphia: Lippincott-Raven; 1996. p. 102-7. 3. Whittle MW. Gait analysis: an introduction. 2nd ed. Oxford: Butterworth-Heinemann; 1996. 4. Winstein CJ, Pohl PS, Cardinale C, Green A, Schlotz L, Waters CS. Learning a partial-weight-bearing skill: effectiveness of two forms of feedback. Phys Ther 1996;76:985-93. 5. Bhambhami YN, Clarkson HM, Gomes PS. Axillary crutch walking: effects of three training programs. Arch Phys Med Rehabil 1990;71:484-9. 6. Stallard J, Sankarankutty M, Rose GK. Lower-limb vertical ground-reaction forces during crutch walking. J Med Eng Technol 1978;2:201-3. 7. Feldman DR, Vujic I, Mckay D, Callcott F, Uflacker R. Crutchinduced axillary artery injury. Cardiovasc Intervent Radiol 1995; 18:296-9. 8. Raikin S, Froimson MI. Bilateral brachial plexus compressive neuropathy (crutch palsy). J Orthop Trauma 1997;11:136-8. 9. Shabas D, Scheiber M. Suprascapular neuropathy related to the use of crutches. Am J Phys Med 1986;65:298-300. 10. Tripp HF, Cook JW. Axillary artery aneurysms. Mil Med 1998; 163:653-5. 11. Baxter ML, Allington RO, Koepke GH. Weight-distribution variables in the use of crutches and canes. Phys Ther 1969;49:360-5. 12. LaFortune MA, Hennig EM, Lake MJ. Dominant role of interface over knee angle for cushioning impact loading and regulating initial leg stiffness. J Biomech 1996;29:1523-9. 13. Nilsson J, Thorstensson A. Ground reaction forces at different speeds of human walking and running. Acta Physiol Scand 1989; 136: 217-27. Suppliers a. dOR-5; Advanced Mechanical Technology, Inc, 176 Waltham St, Watertown, MA 02472. b. National Instruments, Corp, 11500 N Mopac Expwy, Austin, TX 78759. c. Motion Analysis, Corp, 3617 Westwind Blvd, Santa Rosa, CA 95403. d. SPSS, version 9.0; SPSS, Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.