Effects of Disease Severity on Response to Lateral Wedged Shoe Insole for Medial Compartment Knee Osteoarthritis

Methods 

Patients and Healthy Volunteers 

The recruited subjects were 23 patients with bilateral medial compartment knee OA (ie, 46 knees), representing all consecutive patients examined at the Gait Analysis Laboratory, Division of Physical Therapy and Rehabilitation, Fukui University Hospital, between December 1998 and February 1999, who fulfilled the following exclusion criteria. The exclusion criteria were knee pain due to pathologic conditions other than OA, such as rheumatoid arthritis and pseudogout, and knee with flexion contracture greater than 20°. In addition, patients with hip problems, symptomatic lumbar spine disease, spinal cord disease, or those showing ataxic gait related to neurologic disorders were excluded from the study. We also studied 19 age-matched, healthy volunteers (38 knees) as the control group. The demographic data of the patients and controls are shown in table 1.

Table 1. Demographic Data of Patients and Controls

	Characteristics	Knee OA	Controls	P
	No. of subjects	23	19
	Age (y)	67.0±8.7	66.4±9.3	NS†
	Sex (male/female)	6/17	5/14	NS‡
	Height (cm)	150.2±7.2	148.3±9.0	NS†
	Body weight (kg)	60.3±7.1	51.5±8.5	<.001†
	Hip-knee-ankle angle (deg)⁎	−6.2±4.4	NA

NOTE. Values are mean ± standard deviation (SD) or as otherwise indicated. Abbreviations: NA, not applicable; NS, not significant.

⁎ The medial deviation angle from 180° at the intersection between mechanical axes of the femur and tibia. † Nonpaired t test. ‡ Chi-square test.

We measured the hip-knee-ankle angle on a full-length anteroposterior radiograph taken for each OA patient in standing position. The angle represented the medial deviation angle from 180° at the intersection between the mechanical axes of the femur and tibia, and was a plus value for valgus alignment and minus value for varus alignment of the knee. The severity of OA was evaluated by radiography using the Kellgren and Lawrence grading system,17 in which grade 0 is normal; grade I is possible osteophytic lipping; grade II is definite osteophytes and possible joint space narrowing; grade III is moderate and/or multiple osteophytes, definite joint space narrowing, some sclerosis, and possible bony attrition; and grade IV is large osteophytes, marked joint space narrowing, severe sclerosis, and definite bony attrition. Using this system, 11 knees were classified as grade I, 11 as grade II, 13 as grade III, and 11 as grade IV.

The study protocol was approved by the ethics reviewing committee guidelines of our university, and a written informed consent was obtained from each subject.

Gait Analysis 

The gait analysis system and its performance as well as data processing have been described previously by our group (fig 1).18, 19 The computerized gait analysis system (model G1812A)a used in this study included 2 forceplates (each 250×40cm) and a light source spot measuring device (model G2250M).a The latter consisted of light-emitting diodes attached to the body and 4 optoelectronic cameras that capture the emitted light. The 2 forceplates are placed on the floor on either side of the centerline of the 8-m long walkway. Three-dimensional ground reaction forces obtained by the forceplates were synchronized with goniometric measurements using an analog-to-digital converter (model G1812CS)a and a microcomputer (model G3852CS).a The errors of the light source spot measuring device were less than 1% of the field of view, and errors of the forceplates were less than 2% of full scale deflection.

View Large Image Download to PowerPoint

Fig 1. Outline of the gait analysis system. Abbreviations: A, acceleration of the marker motion to the lateral side; M, external adduction moment at the center of the knee.

We attached markers for the light source spot measuring device on the left and right anterior superior iliac spine of the ileum, iliac crest, greater trochanter, lateral side of the femoral condyles, lateral malleolus of the fibula, lateral side of the calcaneus, and the head of the fifth metatarsal bone. The markers of iliac crest and greater trochanter were 5cm apart and attached to the skin using jigs (see fig 1). After 3 walking trials to get used to the lateral wedged insole, the subject was asked to ambulate on the forceplates 3 times without lateral wedged insoles at his/her natural speed with both arms folded in front of the chest so as not to interfere with the lights from the light source (baseline measurements). Then, wearing lateral silicon rubber-made wedged insoles with 10-mm lateral elevation extending from the posterior aspect of the heel to approximately 6cm distal to the center of the heel,b OA patients were asked to ambulate on the forceplates 3 times at a speed similar to that at baseline and gait was recorded and analyzed. The sampling frequency was set at 50Hz.

Results 

Table 2 lists the values of step length, stride width, and walking velocity. These parameters were significantly lower in OA patients than the control. The use of lateral wedged insoles in OA patients did not change the step length and walking velocity, but significantly increased the stride width. Peak adduction moment during the stance phase of the gait cycle was significantly higher in knee OA patients than the control (table 3, fig 2). The use of lateral wedged insoles significantly decreased peak adduction moment values of OA patients with Kellgren-Lawrence grades I and II, whereas the reductions in these values were insignificant in Kellgren-Lawrence grades III and IV. The magnitude of reduction of peak adduction moment, when lateral wedged insoles were applied, was 5.1% (range, −3.7% to 14.3%) in Kellgren-Lawrence grade I, 6.6% (range, 0%−9.0%) in grade II, 3.3% (range, −4.7% to 7.1%) in grade III, and 5.0% (range, −3.1% to 11.1%) in grade IV. There was no gender effect in the observed changes in peak adduction moment values.

Table 2. Parameters Used in Gait Analysis and Results

	Parameters	Knee OA				Healthy Controls	P†	Cohen d
		No Insole	With Insole	P⁎	Cohen d
	Step length (cm)	38.8±9.0	39.8±8.0	.266	NA	48.0±5.1	<.001	1.26
	Stride width (cm)	12.3±2.7	13.9±2.8	.001	0.58	14.6±2.6	<.001	0.87
	Walking velocity (m/s)	36.7±9.0	37.0±8.3	.315	NA	50.7±8.4	<.001	1.61

NOTE. Values are mean ± SD.

⁎ Paired t test between knee OA patients with and without insoles. † Nonpaired t test between knee OA patients without insoles and healthy controls.

Table 3. Peak Adduction Moment and the First Acceleration Peak

	Parameters	Peak Adduction Moment (Nm/kg)				First Acceleration Peak (m/s2)
		No Insoles	With Insoles	P⁎	Cohen d	No Insoles	With Insoles	P⁎	Cohen d
	Knee OA	0.90±0.20	0.86±0.19	<.001	.21	1.65±0.69	1.36±0.76	.018	.40
	Kellgren-Lawrence grade
	I	0.79±0.21†	0.75±0.18	<.043	.21	1.52±0.52	1.02±0.51	.024	.97
	II	0.91±0.11‡	0.85±0.12	<.001	.52	1.57±0.56	1.12±0.67	.043	.73
	III	0.91±0.26‡	0.88±0.24	<.058		1.56±0.45	1.42±0.50	.400
	IV	1.01±0.19‡	0.96±0.17	<.086		1.95±1.10§	1.85±1.07	.800
	Healthy controls	0.59±0.12†‡	NA			1.19±0.56§	NA

NOTE. Values are mean ± SD.

⁎ Paired t test between knee OA patients with and without insoles. † P=.047, d=1.17, between Kellgren-Lawrence grade I and healthy subjects (Scheffé post hoc test). ‡ P<.001, among Kellgren-Lawrence grades II (d=2.78), III (d=1.58), IV (d=2.64), and healthy subjects (Scheffé post hoc test). § P=.020, d=.86, between Kellgren-Lawrence grade IV and healthy subjects (Scheffé post hoc test).

View Large Image Download to PowerPoint

Fig 2. Joint moments of the knee in sagittal plane. Each datum was averaged during an averaged gait cycle.

The mean value of first acceleration peak after heel strike was higher in patients with OA than the control (see table 3). Application of lateral wedged insoles resulted in significant reductions of first acceleration peak values in OA patients with Kellgren-Lawrence grades I and II, but the reductions were insignificant in Kellgren-Lawrence grades III and IV. The magnitude of first acceleration peak reduction in patients wearing lateral wedged insoles was 32.9% (range, −18.2% to 78.1%) in Kellgren-Lawrence grade I, 28.7% (range, −36.3% to 98.2%) in grade II, 9.0% (range, −64.3% to 50.3%) in grade III, and 5.1% (range, −96.0% to 64.8%) in grade IV. There was no sex effect in the observed changes in first acceleration peak values. The correlation between the peak adduction moment and first acceleration peak values under the condition of not wearing lateral wedged insoles (R2=.006, P=.447), and the correlation between the magnitude of reduction of peak adduction moment and that of first acceleration peak during wearing the lateral wedged insoles (R2=.046, P=.151), were not significant in OA patients.

Discussion 

A positive correlation between the magnitude of the knee moment and walking velocity has been already reported.23 Walking velocity may also influence the magnitude of first acceleration peak. However, in the present study, application of lateral wedged insoles did not result in significant changes in these parameters in knee OA patients. Our results of the intrinsic effect of the lateral wedged insoles in knee OA patients excluded the effect of walking velocity. On the other hand, increase of stride width may affect the magnitude of peak adduction moment or first acceleration peak, because a wide-based gait may lead to changes in the moments at the knee joint via the more laterally shifted location of the center of pressure. Increase of stride width might be part of the biomechanic effect of lateral wedged insoles.

The adduction moment of the knee is a major determinant of medial-to-lateral load distribution,24 and is considered the most influential factor, producing medial joint force in joints with varus deformity.25, 26 Therefore, it seems responsible for the biomechanical abnormality of the medial compartment OA of the knee.27 Sharma et al28 reported that the dynamic load during walking correlates with the severity of the disease in tibiofemoral knee OA. They suggested that the magnitude of knee adduction moment possibly influences the structural outcome in medial compartment knee OA. Moreover, Miyazaki et al21 reported that the baseline adduction moment at the knee could predict radiographic progression of OA at the 6-year follow-up in medial compartment knee OA patients.

In our study, reduction of peak adduction moment (average, 4.4%) was seen in OA patients after the use of insoles and the effect was significant in patients with Kellgren-Lawrence grades I and II but not III and IV. Although they did not show data of static alignment, Kerrigan et al10 reported that the use of 5° lateral wedged insoles reduced the peak adduction moment value by about 6% and the 10-mm height lateral wedged insoles reduced the peak by about 8% in patients with Kellgren-Lawrence grades III and IV. Kakihana et al12 also described a 5.6% reduction of the knee joint varus moment during walking by the application of 6° lateral wedged insoles. They did not report the radiographic severity of OA or static alignment. Maly et al13 showed no differences in peak adduction moment values when subjects wore their routine footwear and the valgus heel wedge which promoted valgus alignment of the foot. However, they also did not examine the correlation between peak adduction moment and radiographic severity of OA and did not show the data on static alignment. Our results of reduction in peak adduction moment value (average, 4.4%), on application of the lateral wedged insoles, were smaller than the data reported by Kerrigan10 and Kakihana.12 In addition, the effects of reduction of peak adduction moment were not significant in Kellgren-Lawrence grades III and IV knees in disagreement with the results of Kerrigan.10

The relation between the adduction moment of the knee and static alignment has been reported by several studies.28, 29, 30, 31 Moreover, the adduction moment of the knee correlates with the severity of OA28, 31, 32 and our results provide support to these studies. We suggest that the static alignment of OA patients differs between our patients and those of previous studies. We propose that the mechanism of reduction of adduction moment of the knee with the usage of insoles is a lateral shift in the center of pressure on the foot, thereby reducing the moment arm of the adduction moment, which was hypothesized by Maly.13 Patients with Kellgren-Lawrence grades III and IV had severe varus deformity compared with those of Kellgren-Lawrence grades I and II. Changes in the center of pressure and changes in the moment arm induced by the insoles are probably trivial in patients with Kellgren-Lawrence grades III and IV.

Our results suggest obvious kinetic effect of lateral wedged insoles on peak adduction moment in mild OA patients, and these effects are relatively small in moderate to severe knee OA. Toda et al6 developed a new type of lateral wedged insoles with strapping of the subtalar joint, using a belt support that can correct femorotibial valgus. In another study,7 they also reported that the use of subtalar-strapped insole not only improved visual analog scale of pain and Lequense index scores consisting of pain and functional scores, but also significantly reduced the femorotibial angle in patients with Kellgren-Lawrence grades II, III, and IV at 6-month follow-up assessment. The use of lateral wedged insoles without subtalar strapping, which was used in the present study, might have certain limitations; for example, talus movement might prevent calcaneal valgus correction.6 More effective insoles that can transmit the corrective force to the femorotibial joint via the calcaneus and talus are needed for patients with severe knee OA.

Lateral thrust is the visualized dynamic bowing out of the knee laterally, that is, the abrupt first appearance of genu varus (or the abrupt worsening of existing genu varus) while the limb is bearing weight during ambulation.16 Chang et al16 reported that lateral thrust increased the odds of progression among varus aligned knees when considered separately, suggesting that knees with thrust were a subset of varus-aligned knees at particularly high risk for progression of OA. In our study, 17.6% reduction in first acceleration peak, which reflects the magnitude of lateral thrust motion, was noted after the use of insoles in OA patients and the change was significant in patients with Kellgren-Lawrence grades I and II but not III and IV. Ogata et al14 reported that the amplitude of first acceleration peak decreased with the usage of valgus insoles in all 50 knees with medial OA, and the mean reduction in first acceleration peak amplitude with the usage of insoles was 23.7%, a value similar to that found in our study. They did not show the relation between radiographic severity of OA and reduction of the amplitude of first acceleration peak, but showed the relation between radiographic severity of OA and pain relief when shoe insoles were applied. Thus, they recommended the use of insoles for patients with painful early medial compartment OA. We do agree with their recommendation and our results demonstrated that the use of insoles reduce the magnitude of lateral thrust in patients with mild OA.

There was no relation between the magnitude of reduction of peak adduction moment and that of first acceleration peak when wearing lateral wedged insoles in OA patients. In addition, the peak adduction moment values did not correlate with first acceleration peak before wearing lateral wedged insoles, suggesting different pathogenic processes. Based on our data, we recommend the use of insoles for mild to moderate OA patients, however, other biomechanic factors (muscular strength, joint laxity, proprioception) must be considered also as factors that could alter the effects of lateral wedged insoles.

Our work described the kinetic and kinematic effects of lateral wedged insoles during ambulation and the relation between the radiographic severity of knee OA and biomechanic effects of lateral wedged insoles. In the present study, we did not attempt to examine the relation between radiographic severity of knee OA and clinical results of wearing lateral wedged insoles; thus, further study may be required to assess such effects. However, our results support the proposal made by several groups that the use of lateral wedged insoles is recommended for mild knee OA but not for severe OA based on pain relief2, 8, 14 and improvement of motional function.2

Conclusions 

We confirmed in the present study the kinetic and kinematic benefits of wearing lateral wedged insoles in grades I and II medial compartment knee OA patients. However, there was no correlation between kinetic and kinematic effects, suggesting different pathogenic mechanisms. However, we support the clinical recommendation of the use of lateral wedged insoles for patients only with early and mild knee OA.