View Options - Click to expand
Print Options - Click to expand
E-Mail Options - Click to expand

Transition to an Articulating Knee Prosthesis in Pediatric Amputees

Bryan Wilk, BS
Lori Karol, MD
Suzanne Halliday, MS
Don Cummings, CP
Nasreen Haideri, ME
John Stephenson, CP

ABSTRACT

The purpose of this study is to see if gait deviations resulting from the lack of knee flexion in the solid-knee prosthesis are reduced when a young amputee is fitted with an articulating prosthesis at an earlier age than what has been the norm. Three-dimensional kinematic data were collected from seven pediatric amputees (age ranging from 1 year, 5 months to 6 years, 1 month) at three time points: 1) initially, with their nonarticulated prostheses; 2) after gait training with their new, articulated prostheses; and 3) after approximately 1 year of use with the new prostheses. Results show that the gait strategy for advancing the amputated limb changed from initial visit to follow-up. The children no longer used a circumducted gait pattern, and differences in hip range of motion between limbs were decreased. All children successfully transitioned to articulating-knee prostheses and now walk with more normalized gait patterns.

Key Words: pediatric amputee, prosthetic knee, gait, kinematics

Introduction

The current prosthetic prescription for a child with an above-knee or knee-disarticulation requiring amputation is to fit the child with a prosthesis as soon as he or she begins to pull to stand, which generally occurs between the ages of 9 and 16 months.1,2 Common practice is to start gait training with a nonarticulating prosthesis (knee module is usually omitted, but can be included if constructed with a locked knee) during the period when a child first learns to walk.1,3 As the toddler begins to walk independently, the child is usually transitioned to a prosthesis with a functional knee around the age of 3 or 4 years, and sometimes as late as 6 years if the child has a bilateral amputation.1,4

Traditionally, there are two reasons why children are not transitioned earlier to a functional knee. Children can learn to ambulate more easily without having the prosthetic knee buckling at unwanted times. Second, commercially available pediatric knees often cannot be used effectively for children less than 3 years old because the knees are too long to be used in the relatively small space available, or cannot be finished easily with a durable cosmetic cover. Although it is recognized that "outside" metal joints can be used at any age, they have the disadvantage of poor cosmesis, low durability, increased noise, and limited stability.

As the transition age is delayed, the child learns to ambulate with the stiff limb by adopting a gait pattern characterized by an abnormal increase in pelvic motion and increased circumduction of the amputated limb to clear the foot during the swing phase. This learned compensation may be retained when the child later receives a prosthesis with a knee that articulates. The child, therefore, continues to walk in his or her new articulating prosthesis without taking advantage of the available knee flexion. The purpose of this study is to see if gait deviations resulting from the lack of knee flexion in the solid-knee prosthesis are reduced when a young amputee is fit with an articulating prosthesis at an earlier age than what has been the norm.

Methods

After we obtained informed consent from their parents, seven patients, who had already learned to walk using a stiff-knee prosthesis, were referred to the Movement Science Lab (MSL). Average age of the children at time of initial testing with the stiff-knee prosthesis was 2 years, 10 months (ranging from 1 year, 5 months to 6 years, 1 month) (Table 1 ). The age at admission became progressively younger as attempts were made to establish a minimum age as to when a pediatric amputee could be transitioned to an articulating prosthesis. Subjects in the study consisted of two children given diagnoses of proximal femoral focal deficiencies, three with tibial hemimelia, one child with an amputation secondary to amniotic band syndrome, and one child with a knee-disarticulation amputation secondary to a motor vehicle accident.

On the day of admission (test 1), prior to receiving the new articulating prosthesis, each child was tested in his or her existing stiff-knee prosthesis (Table 2 ). All children received inpatient physical therapy, during which time they learned to walk in their new articulated prostheses (Table 3 ). The goals of physical therapy were to transition from sit to stand and pull to stand, have good skin tolerance, weight shifting, single-limb support, step lengths, and knee flexion while walking, and successfully don and doff the prosthesis either independently or with parental assistance. The number of physical therapy sessions varied among patients depending on how well they achieved these goals. When it was determined that the children could successfully ambulate, they were tested in the MSL in their new prostheses (test 2). Children were recalled after an average of 1 year, 3 months (range, 10 months to 1 year, 8 months) to document progress in the articulated prostheses (test 3).

Each patient was followed up through the prosthetics clinic at the Texas Scottish Rite Hospital for Children (TSRHC). Once initially seen, the child was followed up by the same orthopedist and by a team of TSRHC prosthetists and physical therapists. Each patient averaged four follow-up visits between test 2 and test 3 (ranging from two to six visits). Four prosthetists were involved with this study. Once a patient had been casted for his or her nonarticulated system, he or she was fit by the same prosthetist for the articulated system.

Socket design was based on the patient's amputation level (i.e., a knee disarticulation socket would have a "less aggressive" ischial weight-bearing socket instead of an above-knee socket). Knee selection was based on the following criteria: 1) If there was enough space, a pediatric Total KneeR (Century XXII Innovations, Jackson, MI) was selected because of its stance-phase braking mechanism. 2) If there was not enough space for a Total Knee, a DAWR pediatric 4-Bar Knee (DAW Industries, San Diego, CA) was selected for its shorter length and polycentric stability. 3) If there was no room for either of these protheses, an alternative, single-axis knee was used. The prosthetic feet were selected to have a dynamic response (preferably, a Seattle Child's PlayR foot (Seattle Orthopedic Group, Poulsbo, WA.) was used because the height of the keel, recessed into the foot, provided more space for the proximal components). Other feet were selected based on the patient's performance or unique requirements.

Appropriate alignment was determined cooperatively with the prosthetist, physical therapist, and attending physician during the patient's admission for gait training. Alignment was maintained during follow-up visits to the patient's clinic or during follow-up visits to the prosthetics lab. Adjustments were made as needed to assure that the prostheses were fitting well, that the prostheses were the appropriate height, and that the prostheses continued to maintain adequate alignment. In two cases (patient numbers 3 and 5), prostheses were remade when the children outgrew their sockets. Efforts were made to use the same socket and knee design, but in one case the foot had been changed to better match the patient's needs.

At each of the three tests, three-dimensional gait analysis was performed by using six 60 Hz VICON cameras (Oxford Metrics, Oxford, England). Fifteen markers were placed according to the model developed by Davis et al5 at the Newington Children's Hospital. Kinematic data in the sagittal, coronal, and transverse planes were processed by using VICON software (VCM). The dependent variables measured were motion at the hip, knee, and ankle in the sagittal, coronal, and transverse planes. Peak ranges of motion and cadence parameters were calculated. Unbalanced repeated measures analysis of variance and Tukey post hoc testing were used to determine significant interactions and main effects of the measured variables between limbs and between testing times for each subject. The patient with a bilateral amputation contributed two limbs to the sample of the amputated limbs. Neither of his limbs were used in the sample of sound limbs. One-sample t tests were used to assess differences between measured variables compared to a previously published control group of age-matched norms.6 Results with p values less than or equal to .05 were considered statistically significant.

Results

The amputees in this study had walked in their nonarticulating prostheses for an average of 2 years (range, 7 months to 5 years, 2 months). The initial visit, therefore, was used to establish a baseline of comparison for tests 2 and 3. Although baseline cadence was equal to that of age-matched norms, velocity was greater than normal due to greater step lengths for both limbs (p less than or equal to .05). Initially, the pediatric amputees had greater motion at the pelvis than did their age-matched norms (p less than or equal to .05). Motion of the pelvis was analyzed in three planes: sagittal, coronal, and transverse. Pelvic tilt was defined as anterior-posterior rotation in the sagittal plane about a medial-lateral axis. Pelvic obliquity was defined as right-left rotation in the coronal plane about an anterior-posterior axis. Pelvic rotation was defined as rotation in the transverse plane about a vertical axis perpendicular to the floor (Figure 1 ).

Baseline peak hip abduction for the amputated limb was greater than that of age-matched norms (p less than or equal to .05). Asymmetries existed across limbs for both hip rotation and hip flexion (the sound limb had significantly greater motion than that of age-matched norms). It is assumed that no motion occurred at the knee because the prosthesis was constructed without a knee joint. Some motion occurred at the ankle for the amputated limb due to the deformative (dynamic) materials used for both the prosthetic foot and shoe. This motion was, however, minimal when compared to that of a normal ankle joint (p less than or equal to .05).

After the initial course of physical therapy in the articulated prosthesis, step length for both limbs decreased, causing cadence and walking speed to be less than that of age-matched norms (p less than or equal to .05). The pelvis as a whole still moved more than normal in all three planes, but both pelvic obliquity and pelvic rotation decreased slightly from that of the previous testing session. Hip flexion-extension for both limbs decreased, but range of motion for the sound limb still remained significantly greater than that of the amputated limb. Peak hip abduction for both limbs changed so that both were greater than that of age-matched norms (p less than or equal to .05). Hip rotation in the transverse plane for the sound limb was still greater than that of the amputated limb, but to a lesser extent than in the first session. The most significant change from initial visit to second visit occurred at the knee joint, as seen in Figure 2 , with the articulating prosthesis providing 32° of knee flexion. There was no significant change in ankle motion for the sound limb.

After approximately 1 year with the articulated knees, the children significantly increased their walking speed from that of the second testing session to that of age-matched norms by increasing both their cadences and step lengths. Step length for both limbs still remained greater than that of age-matched norms. Both the ranges of pelvic tilt and pelvic obliquity remained greater than normal, but to a lesser degree than that seen at the initial visit. Pelvic rotation, however, significantly decreased from initial test to final follow-up and was now within the normal age-matched range. Difference in hip flexion-extension across limbs remained after approximately 1 year's use with the new prosthesis, but this difference was no longer significant due to the improved symmetry between limbs. Peak hip abduction continued to decrease for the amputated limb, but was not significantly different from the previous two sessions. A decreasing trend, however, was observed. This is clinically significant, because peak hip abduction for both limbs was more symmetrical by the third testing session, and values were within the range for age-matched norms (Table 4 ). No significant changes were seen across testing sessions or between limbs for hip rotation in the transverse plane. Hip rotation for the amputated limb increased from initial visit to follow-up and decreased for the sound limb. Therefore, asymmetry in hip rotation between amputated and sound limb improved (Table 4 ).

The most significant changes were again at the knee, as seen in Figure 2 . Knee flexion for the articulating prosthesis increased to a mean of 49°. More importantly, from a clinical standpoint, there was no longer a significant difference in knee range of motion between limbs. An improved loading response during weight acceptance (denoted by flexion of the knee during early stance) for the sound limb was also observed at follow-up (Figure 1 ). The importance of a loading response during initial stance is to dissipate forces occurring at joints, particularly at the knee.

Discussion

While early fitting with fixed knees greatly simplifies gait training and prosthetic fitting, it forces children to adopt a circumducted gait pattern, as was seen during the initial testing session. The children without an articulating knee advanced their prosthetic limbs by abnormally tilting the pelvis posteriorly and excessively internally rotating and hiking the pelvis on the amputated side during the swing phase.

Posteriorly tilting the pelvis occurs infrequently during the gait cycle, and as such, is usually a deliberate compensation for the advancement of the limb during swing phase.7 Two muscle groups accomplish forward propulsion and initiation of swing: the plantarflexors and the hip flexors. Because the amputees in this study lack plantarflexors on the amputated side, it makes sense that they posteriorly tilt their pelvis to aid in the initiation of swing. The amputated limb also compensated for not having a knee by abnormally abducting at the hip to achieve clearance during swing.

The VCM model used in this study measures hip motion with respect to the pelvis. If the pelvis tilts anteriorly, the hip appears to be in more flexion. If the pelvis tilts posteriorly, the hip appears to be in more extension. In all three testing sessions, there was a greater range of flexion-extension for the sound limb than for the amputated limb. This asymmetry may exist because the patients tilted their pelves posteriorly to advance their amputated limbs.

Remembering that the young amputees had at least 7 months to adapt to their previous stiff-knee prostheses, it is not a surprise that significant changes occurred in their walking patterns when first tested in their new prostheses. Because the new prosthesis allows the knee to bend, the child must adapt to the possibility of the knee buckling during walking. It was observed that the children walked slower, taking fewer shorter steps than in the previous session. This was undoubtedly to make sure that the knee stayed locked during the stance phase of walking. Because hip flexion-extension is dependent upon step length, it also decreased during the second testing session. Pelvic motion decreased slightly, with the exception of pelvic tilt, which increased. Because the children may have been unaccustomed to bending their new knees, they may have used more pelvic tilt to help advance their prosthetic limbs. Peak hip abduction increased for the sound limb, which was surprising. Perhaps the children abducted more to establish a wider base of support for better balance and stability during walking with the new prostheses. The second testing session caused quite a few changes that did not follow any particular trend among the children. It can be speculated that because such a short time was spent walking in the new prostheses prior to testing, the children were still learning how to use the articulating knee prostheses.

After at least 10 months in their new articulated systems, the young amputees appeared to be well adjusted to their prostheses. Walking speeds were back to normal. Although pelvic tilt only decreased slightly from initial session to follow-up, this trend was witnessed in five of seven children. This change, coupled with the significant decrease in pelvic rotation, supports the observation that young amputees are no longer depending on excessive pelvic motion to advance their prosthetic limbs. Asymmetries that were present during the initial testing session were no longer significant at follow-up. Asymmetries in peak hip abduction, hip rotation, and knee flexion no longer remained significant. In fact, the only variable that remained significantly different across limbs was ankle flexion-extension due to the prosthesis.

It is recognized that these changes may have resulted from the maturation process, particularly changes seen in the temporal parameters. As children grow, they reduce their cadences and increase their stride lengths to achieve normal velocities. The data support this trend. Pelvic motion also decreases with age as normal children mature. It is believed, however, that amputees do not necessarily follow this trend, given the compensations involved with amputee walking. The oldest child in the study (6 years old) had more pelvic motion at initial testing than did the much younger amputees (1.5 years old). It therefore appears that the articulated prosthesis may have played a larger role in changing gait than did the maturation process.

The children walked by using a different strategy for advancing their articulated prosthetic limbs. Instead of relying primarily on abducting the limbs and tilting and rotating their pelves to advance their prosthetic limbs during swing, the young amputees have learned to use more knee flexion with the new prostheses to clear their limbs while walking. Parents of the children were satisfied from a functional standpoint; there were no reports of increased falling, and no family requested that their child return to the previous stiff prosthesis.

Conclusion

We were able to fit children as young as 1 year, 5 months of age with mobile knees, which is at a much younger age than is generally accepted. It is of benefit to transition a pediatric amputee to an articulated knee as soon as he or she is developmentally ready for two reasons. Early fitting with an articulated knee reduces the chance that the gait deviations, adopted from the long-term use of a stiff-knee prosthesis, will become ingrained. Secondly, normal childhood activities such as sitting, crawling, squatting, and kneeling may be unrestricted with the use of an articulated prosthesis.

The limitation in prosthetic design for the toddler-aged amputee is in the size of the componentry. While we would prefer to fit all young children with articulating knee joints, some are simply not tall enough to allow for prostheses that can incorporate the knee components. We currently recommend fitting a child with an articulated prosthesis as soon height allows. While the child's gait will not immediately normalize after training with an articulated prosthesis, significant long-term improvements in knee flexion during swing phase and resolution of increased pelvic rotation and hip abduction due to circumduction can be expected. These results prove that a pediatric amputee can achieve a more normalized gait pattern in as little as one year's time.

Acknowledgements

The authors would like to thank Cindy Smith, Sarah Mattes, Scott Colby, and Cecilia Concha of the Movement Science Lab; Erin Berling, Steve Ronde, Wanda McFadden, and Don Virostek of the Prosthetics Department; and Darla Kalb of the Physical Therapy Department for their contributions to the study.


References:

  1. Cummings D, Kapp S. Lower-limb pediatric prosthetics: general considerations and philosophy. JPO. 1992;4(4):196-206.
  2. Hamilton EL. Gait training, part 2: children. In: Kostuik JP, ed. Amputation Surgery and Rehabilitation: The Toronto Experience. New York, NY: Churchill Livingstone; 1981:331-333.
  3. Bochmann D. Prostheses for the limb-deficient child. In: Kostuik JP, ed. Amputation Surgery and Rehabilitation: The Toronto Experience. New York, NY: Churchill Livingstone; 1981:293-305.
  4. Cummings D. The missing link in pediatric prosthetic knees. Biomechanics. 1997;4(2):163-168.
  5. Davis RB. A gait analysis data collection and reduction technique. Human Movement Science. 1991;10:575-587.
  6. Sutherland DH, Olshen RA, Biden EN, Wyatt MP. The Development of Mature Walking. Philadelphia, PA: Mac Kieth Press; 1988.
  7. Perry J. Gait Analysis: Normal and Pathological Function. Thorofare, NJ: SLACK; 1992:265-271.