The knee adduction moment during gait in subjects with knee osteoarthritis is more closely correlated with static alignment than radiographic disease severity, toe out angle and pain

The knee adduction moment during gait in subjects with knee osteoarthritis is more closely correlated with static alignment than radiographic disease severity, toe out angle and pain

Journal of Orthopaedic Research ELSEVIER Journal of Orthopaedic Research 20 (2002) 101-107 ~ www.elsevier.com/locate/orthres The knee adduction m...

770KB Sizes 10 Downloads 16 Views

Journal of Orthopaedic Research

ELSEVIER

Journal of Orthopaedic Research 20 (2002) 101-107

~

www.elsevier.com/locate/orthres

The knee adduction moment during gait in subjects with knee osteoarthritis is more closely correlated with static alignment than radiographic disease severity, toe out angle and pain D.E. Hurwitz a

Depurtmen t

a,b3*,

A.B. Ryals

J.P. Case b,e, J.A. Block

b,c,

T.P. Andriacchi

a,f

of Orthopedic Surgery, Rush- Preshyteriun-St. Luke's Medical Center, Rush Medicul College of Rush Uniwrsity,

1645 West Congress Purkrwy, Chicago. I L 60612, U S A Departnzent of Internul Medicine, Seciion qf' Rheuniutology, Rush-Preshqiteriun-St. Luke's Medical Center, Rush Medical College of' Rush Uniuersity, 1645 West Congress Parkwuy, Chicago, I L 60612, USA Dquurtnient of' Biochemistry. Rush-Preshj~terian-St.L u k ~ ' sMedicul Center, Rush Meclical Collcge uf' Rush Unizwsitj, 1645 West Congress Purkway, Chicugo, I L 60612, U S A " Department of' Bioengineering, Unicersity of' Illinois at Chicugo, Chicugo. IL, U S A Section of Rheumatology, Cook County Hospiiul, Chicago, f L , USA Depurttnent oJ Mechunicul Engineering/Functional Restoration, Sanford Uniuersity, Stanford, CA, USA

'

Accepted 29 May 2001

Abstract

This study tested whether the peak external knee adduction moments during walking in subjects with knee osteoarthritis (OA) were correlated with the mechanical axis of the leg, radiographic measures of OA severity, toe out angle or clinical assessments of pain, stiffness or function. Gait analysis was performed on 62 subjects with knee OA and 49 asymptomatic control subjects (normal subjects). The subjects with OA walked with a greater than normal peak adduction moment during early stance 0, = 0.027). In the OA group, the mechanical axis was the best single predictor of the peak adduction moment during both early and late stance ( R = 0.74, p < 0,001). The radiographic measures of OA severity in the medial compartment were also predictive of both peak adduction moments (R = 0.43 to 0.48, p < 0.001) along with the sum of the WOMAC subscales ( R = -0.33 to -0.31, p < 0.017). The toe out angle was predictive of the peak adduction moment only during late stance ( R = -0.45, p < 0.001). Once mechanical axis was accounted for, other factors only increased the ability to predict the peak knee adduction moments by 10-18%~While the mechanical axis was indicative of the peak adduction moments, it only accounted for about 50% of its variation, emphasizing the need for a dynamic evaluation of the knee joint loading environment. Understanding which clinical measures of OA are most closely associated with the dynamic knee joint loads may ultimately result in a better understanding of the disease process and the development of therapeutic interventions. 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved.

Introduction Increased dynamic loads on the medial compartment of the knee may contribute to the development or progression of knee osteoarthritis (OA), as increased load on the articular cartilage is one factor associated with the pathogenesis of OA [8,17]. Given that knee OA often affects the medial compartment, studies have characterized the dynamic knee joint loads using the external knee adduction moment [6,12,16,19-211. A greater ad-

*Corresponding author. Tel.: + 1-312-942-5791; fax: +I-312-9422101. E-muil address; [email protected] (D.E. Hurwitz).

duction moment corresponds to increased load on the medial compartment relative to that of the lateral compartment [181. The external knee adduction moment has also been correlated with the bone distribution between the medial and lateral compartments of the proximal tibia [13]. Subjects with knee OA with medial joint space narrowing have been shown to have greater than normal peak knee adduction moments [6,16]. Moreover, studies of subjects with knee OA have shown that surgical outcome [ 16,211, pain relief [ 12,191 and radiographic disease severity [20] are related to the peak external knee adduction moment. One mechanism that has been suggested to decrease the external knee adduction moment is to increase the toe out angle [16,21]. The external knee adduction moment has also

0736-0266/02/$ - see front matter 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved. PII: S 0 7 3 6 - 0 2 6 6 ( 0 1 ) 0 0 0 8I - X

been correlated with the mechanical axis in young normal subjects [ 3 ] , preoperative total knee replacement subjects [ l l ] and some subjects with knee OA [9,22]. In contrast, no such correlation was found among subjects with primary osteoarthritis or rheumatoid arthritis with either flexion, varus or valgus deformities [lo] or patients with varus gonarthrosis scheduled for a high tibial osteotoniy [7,16]. Differences between the study populations with respect to the extent of varus deformity, radiographic disease severity, pain, toe out angle or the development of other compensatory gait adaptations may account for the variable findings. The relationships between the external knee adduction moment and mechanical axis, radiographic disease severity, toe out angle, pain and functional assessment have not yet been concurrently examined in one population. Instead most studies have focused on the effect of only one or two of these factors on the external knee adduction moment and often times restricted their attention to either presurgical candidates with more advance symptomatic OA or to asymptomatic normal subjects. Understanding how these factors relate to dynamic loads during gait in subjects with milder symptomatic knee OA (i.e. those in whom surgical management is not being considered due to adequate clinical response with medical therapy) may ultiinately help identify subjects who are more susceptible to progressive knee OA and who may benefit from early therapeutic interventions. The present study assesses what percentage of the variation in the peak external knee adduction moment is accounted for by mechanical axis, radiographic disease severity, toe out angle, pain and functional assessment in a single group of subjects with milder symptomatic knee OA. The study tests the following hypotheses: The external knee adduction moment is correlated with the mechanical axis, radiographic disease severity, toe out angle, and pain and functional assessment. In the presence of each other, each of these factors explains part of the variance in the external knee adduction moment. Methods The subjects with knee OA were part of a double blinded study examining the effect of pain medication on dynamic loads at the knee

joint during gait. Subjects underwent gait analysis and a radiographic and clinical evaluation after discontinuing their arthritis medication (nonsteroidal anti-inflammatory drugs, analgesics) for a two-week washout period. The study was Institutional Review Board approved, and all subjects granted informed consent. Subjects were included in the double blinded study if they had knee pain and were classified as functional class I I or I11 by ARA criteria [I]. The present study was restricted to those 62 subjects (30 females, 32 males, age 62 k 10 years) enrolled in the double blinded study who had a Kellgren and Lawrence grade [14] of a t least 1 and who had a weightbearing anteroposterior radiograph of the knee in full extension so that static alignment of the lower limb could be measured. Nine (14%) of the subjects had a Kellgren and Lawrence grade of 1. indicating possible osteophytic lipping; 24 (39%) had a grade of 2, indicating definite osteophyte, possible joint space narrowing; 22 (36'%1)had a grade of 3 , indicating moderate multiple osteophytes, definite joint space narrowing, some sclerosis, and possible bone contour deformity; and 7 (11%) had a grade of 4, indicating large osteophytes, marked joint space narrowing, severe sclerosis, definite bone contour deformity. The niechanical axis was measured by one investigator from a single radiograph that included the hip, knee and ankle [IS]. Neutral alignment was defined as zero, varus alignment as greater than zero and valgus alignment as less than Lero. The average mechanical axis was 5 5 5" (range -9" to 15"). 50 subjects had varus knees (80'%1);four, neutral (7%); and eight, valgus (13%1). Weight-bearing anteroposterior radiographs of the knee were interpreted by a rheumatologist who was blinded to the clinical features of the study subjects and who was experienced in the assessment and grading of OA by radiographic criteria. Radiographs were evaluated according to the criteria disseminated by the Osteoarthritis Research Society and employed the "Atlas of individual radiographic features in osteoarthritis" [2], which has previously been validated for inter- and intra-observer variability. Specifically, qualitative assessments of each of the radiographic parameters were performed, and each parameter was assigned a grade between 0 and 3, with 0 being normal and 3, severe involvement [2]. Joint space narrowing and subchondral sclerosis in the medial and lateral tibiofemoral compartments were each assessed separately, and osteophytes at the margins of the femoral condyles o r tibial plateaus were graded (Table 1). Pain and function were assessed for the subjects with OA using the visual analog version 3.0 of the Western Ontario and McMaster University Rating (WOMAC-VA3.0). The pain (maximum 500 points), stitrness (maximum 200 points), and physical function (maximum 1700 points) subscales and the sum of these three subscales (maximum 2400 points) were assessed. The average pain, stiffness and function WOMAC subscales were 187 f 94, 95 i 51, and 654 ?C 341, respectively, and their sum averaged 936 & 464. A group of 49 asymptomatic control subjects (25 females, 24 males. age 59 i 10 years) were also tested in the gait laboratory. All asymptomatic control subjects (normal subjects) had n o clinical diagnosis of OA or rheumatoid O A or a history of knee trauma or pain. They had no significant involvement in their lower extremity or back, meaning they were not being treated for any clinical conditions and none of their activities were restricted due to any medical o r musculoskeletal conditions. The number of right and left sides analyzed in the normal subjects (26 or 53% right, 23 or 47% left) match the distribution of right and left affected sides in the OA group (32 or 52% right, 30 or 48% left). The mass and height of the normal group (76 i 16 kg. I .70 0.10 m, respectively) were not significantly different from those of the OA group (79 =t12 kg, 1.71 i0.1 1 m, respectively) @ > 0.298). The OA subjects were o n average 4 years older than the normal sub-

*

Table 1 Radiographic measures of OA severity for the 62 subjects with knee OA (hde

0: none I: mild 2: moderate 3: severe

Medial joint space narrowing 1 (2%)

26 (425'0) 27 (43%) 8 (1 3%)

Lateral joint space narrowing

Medial sclerosis

Lateral sclerosis

Osteoph yte

4s (73%) 14 (23%) 2 (3%) I (1%)

3 (5%) 38 (61%) 19 (31'%)) 2 (3%)

18 (29%) 37 (60'%,) 7 ( 1 11%)) 0 (OY")

9 (14%) 25 (40'%1) 19 (31%) 9 ( 1S'YO)

D.E. Hurwitz ct ul. / Journal of Orthopcreclic Reseurch 20 (2002) 101-107

jects 0,= 0.049). While this difference was almost statistically significant, it is highly unlikely that this small age difference affected the gait data (95% confidence interval, -7 years to -0.02 years). The instrumentation for gait analysis included a n optoelectronic system with a passive retroreflective marker system (CFTC - Computerized Functional Testing Corporation, Chicago I L) and a multicomponent force plate (Bertec, Columbus, OH). The average error of the three-dimensional coordinates was 7 mm, and the average angular accuracy was 1" with a maximum variation of 3" across the Calibrdtion volume. The three-dimensional spatial positions of markers placed on the lateral aspect of the superiormost iliac crest, at the center of the greater trochanter, over the mid-point of the lateral joint line of the knee, on the lateral aspect of the malleolus, and at the base of the fifth metatarsal were calculated to determine the limb motion [4,5]. The three-dimensional external moments and intersegmental joint forces were calculated using inverse dynamics and by modeling the leg as a collection of rigid links or segments (slender rods) [4,5].The link model included the assumption that no axial rotation occurred about the long axis of each segment. The inertial properties for each rigid segment were lumped at its mass center (lumped mass approximation) and were included in the calculation of the external moments and intersegmental forces. The external moments were transformed into the local coordinate system of each joint and were expressed as a percentage of each subject's body weight multiplied by height. The peak external adduction moment during gait is reproducible with a mean difference between tests on separate days of O.I'%i body weight * height [3]. In this paper, the knee adduction moment is the external knee adduction moment after it has been normalized to the subject's body weight and height. During the gait evaluation subjects were asked to walk multiple times at three self-selected speeds of slow, normal and fast. A representative walking trial at approximately 1 m/s was chosen for analysis, and the average speed of the analyzed trials was 1.01 f 0 . 1 5 m/s for the OA group and 1.03 0.09 m/s for the normal group 0, = 0.504). Significant differences in the peak knee adduction moments between the asymptomatic normal and OA groups were identified with ttests. Pearson correlations and multiple linear regression models were used to test for significant relationships between the peak adduction moments and the mechanical axis, radiographic disease severity, toe out angle, pain and functional assessment among the subjects with OA. Multiple regression models and partial correlation coefficients were used to identify which variables contributed to the adduction moment in the presence of each other. The final multiple regression model was identified using stepwise forward regression (p,,,= 0.05.p,,, = 0.1). A significance level of ( x < 0.05) was used throughout. The appropriateness of the stepwise forward regression models was evaluated by examining histograms of the residuals as well as plots of the residuals against the independent and predicted values.

*

Results The external knee adduction moment normally consists of two peaks one during the first half or early stance, and one during the second half or late stance (Fig. I). However, 29% of the asymptomatic normal subjects (14 of 49) and 52% of the subjects with OA (32 of 62) did not have a definitive second peak adduction moment during late stance. Among the subjects with a distinct second peak adduction moment, the second peak adduction moment and the peak internal rotation moment occurred at about the same time. If the second peak adduction moment and the peak internal rotation moment were always coincident, then the regression line between the second peak adduction moment and the adduction moment during the peak internal rotation moment would have a slope of one and an intercept of zero. Based on the data of the 65 subjects with a distinct

I03

External Frcmtal and Transverse Plane Moments During Stance bbdudion

bkrnal Rddion c

c w .-

a,E

3 .n

8' 8

Intmal Rddion

Fig. I . (Left) Among subjects with two distinct peak adduction moments, the second peak adduction moment is coincident with the peak internal rotation moment. (Right) Among subjects without a distinct second peak adduction moment, the adduction moment that occurred during the peak internal rotation moment was considered to be the second peak adduction moment.

second peak adduction moment, the slope of the regression line was 1.01 (95Yn confidence interval, 0.97 to 1.04), which was not significantly different from one (p > 0.5), and the intercept was -0.01 (95% confidence interval, -0.17 to 0.02), which was not significantly different from zero (p = 0.137). Thus, for subjects without a distinct second peak adduction moment late in stance, the adduction moment that occurred at the time of the peak internal rotation moment was used. The first peak adduction moment during early stance of the subjects with knee OA was significantly greater than that of the asymptomatic normal subjects 0, = 0.027) (Fig. 2). In contrast, the second peak adduction moment during late stance was not significantly different between the two groups (p = 0.391). In both groups the first peak adduction moment was significantly greater than the second peak adduction moment (p = 0.001). Among the subjects with OA, the mechanical axis was the best single predictor of both the first and second peak adduction moments ( R = 0 . 7 4 , ~< 0.001; R = 0.75, p < 0.001, respectively) (Table 2, Fig. 3). The next best single predictors of the first peak adduction moment were the two radiographic measures of disease severity in the medial compartment (i.e. medial joint space narrowing, R = 0.48, p < 0.001 and medial sclerosis, R = 0.43, p < 0.OOl). The WOMAC stiffness, function and

D.E.Hurwitz et ul. I Journul of Orrhopaedic Reseurch 20 (2002) I01 107

I 04

= 0

OASubJects Asymptomatic Normal Subjects

p = 0.027

.

R = 0.751 p < 0.001

I

E

R=-0.452 p =< 0.001

.a, o)

I

E 5 rn ._

I

First Peak Adduction Moment

..:.

2

p = 0.391

U x

T

4 -

O 4 1 m

:

0.

3-

2-

1,

Second Peak Adduction Moment

Fig. 2. The subjects with knee OA had a significantly greater than normal peak adduction moment early in stance, while the peak adduction moment later in stance was not significantly different between the two groups.

sum of the subscales were negatively correlated with the first peak adduction moment ( R = -0.34 to -0.26; p =: 0.007 to p = 0.047) indicating that subjects with OA with worse clinical symptoms had lower first peak adduction moments. The three next best single predictors for the second peak adduction moment also included the radiographic measures of disease severity in the medial compartment ( R = 0.44, p < 0.001 medial joint space narrowing grade; R = 0.47, p < 0.001 medial sclerosis grade) as well as the toe out angle ( R = -0.45, p < 0.001, Fig. 3). The negative correlation between toe out angle and the second peak adduction moment indicated that subjects with OA with an increased toe out angle had a lower second peak adduction moment. Similar to the first peak

-10 Valgus

-5

0 5 10 Mechanical Axis

15

0

10

varus

(Degrees)

20 (Degrees)

Fig. 3. (Left) The best predictor of the second peak adduction moment was the mechanical axis. (Right) The toe out angle was negatively correlated with the second peak adduction moment, indicating that an increased toe out angle decreases the adduction moment.

adduction moment, the WOMAC function and sum of the subscales were negatively correlated with the second peak adduction moment ( R = - 0 . 3 3 , ~= 0.010;R = - 0 . 3 1 , ~= 0.016, respectively). After accounting for the effects of mechanical axis, only the osteophyte grade ( R = 0.42, p < 0.001, partial correlation coefficient) and the WOMAC function subscale ( R = -0.27, p = 0.042, partial correlation coefficient) were significantly correlated with the first peak adduction moment. The osteophyte grade and WOMAC function subscale only accounted for an additional 8% and 3% of the variation in the first peak adduction moment, respectively, when used in con-

Table 2 Correlation coefficients and the significance levels" First peak adduction moment

Second peak adduction moment

Mechanical axis Toe otil angle

R = 0.735, p < 0.001 R = -.206, p = 0.108

R=0.751,p<0.001 R = -0.452, p C 0.001

Radiogrupliic meusures of diseuse seuerity K-L grade Medial joint space narrowing Lateral joint space narrowing Medial sclerosis Lateral sclerosis Osteophytes

R = 0.301, p = 0.01 7 R = 0.481, p 0.001 R = -0.178, p = 0.165 R = 0.429, p <: 0.001 R = -0.170, p = 0.186 R=0.357,p=0.004

R = 0.266, p = 0.036 R = 0.436, p < 0.001 R = -0.113, p = 0.382 R=0.467,p<0.001 R = -0.071, p 0.518 R = 0.205, p = 0.109

WOMAC scores Paill Stiffness Function S uni

R = -0.250, p z= 0.052 R = -0.343. p = 0.047 R = -0.343, p = 0.007 R = -0.330, p 10.009

R = -0.210, p = 0.104 R = -0.224, p = 0.083 R = -0.326, p = 0.010 R=-0.306, p=0.016

'I

The values in italic are statistically significant.

30

Toe OutAngle

7

D.E. Hurwii; et (11. I Journul

of Orthopueilic Reseurch 20 (2002) 101-107

I05

Table 3 The adjusted R2 for. predictors of the first and second peak adduction moments as identified by the forward stepwise regression model First peak ndtluctron mon?ent 0.53 0.61 0.63

Mechanical axis Mechanical axis osteophyte Mechanical axis + osteophyte + WOMAC function

Second peak udduction moment 0.56 0.61 0.7 I 0.73 0.75

1:Mechanical axis 2:Mechanical axis toe 3:Mechanical axis + toe 4:Mechanical axis + toe 5:Mechanical axis + toe

+

+

junction with the mechanical axis. It was unsurprising that many of the other variables no longer explained a significant portion of the variation in the first peak adduction moment after accounting for the mechanical axis due to the fact that the mechanical axis was significantly correlated with the Kellgren and Lawrence grade ( R = 0.31, p = 0.014), the medial joint space narrowing grade ( R = 0.45, p < 0.001), the lateral joint space narrowing grade ( R = -0.41. p = 0.001) and the medial sclerosis grade ( R = 0.43, p = 0.001). The mechanical axis was not correlated with the remaining radiographic measures of disease severity, WOMAC subscales or toe out angle (p > 0.067). As identified by the forward stepwise regression algorithm, the mechanical axis, maximum osteophyte grade and WOMAC function score were the final set of predictors of the first peak adduction moment (Table 3) and these variables together explained 65% of the variance in the first peak adduction moment (R2 = 0.65, adjusted R2 = 0.63). However, the inclusion of the WOMAC function score only explained an additional 2-3% of the variation in the first peak adduction moment when compared to that explained by mechanical axis and maximum osteophyte grade. For the second peak adduction moment, once the effects of mechanical axis were accounted for, only the toe-out angle ( R = -0.45, p < 0.001, partial correlation coefficient) and the lateral joint space narrowing (R = 0 . 3 2 , = ~ 0.012, partial correlation coefficient) were significantly correlated with the second peak adduction moment. The toe out angle and lateral joint space narrowing only accounted for an additional 9% and 5% of the variation in the second peak adduction moment, respectively, when used in conjunction with the mechanical axis. The final model for predicting the second peak adduction moment as identified using the forward stepwise regression was mechanical axis, toeout angle, lateral joint space narrowing grade and WOMAC function and pain subscales (Table 3). These five variables accounted for 77% of the variation in the second peak adduction moment (R2 = 0.77, adjusted R2 = 0.75). However, mechanical axis, toeout and lateral joint space narrowing

out out Lateral joint space out + Lateral joint space WOMAC function out angle + Lateral joint space WOMAC function + WOMAC pain

+

+

+

grade together accounted for 72% of the variation in the second peak adduction moment (R' = 0.72, adjusted R2 = 0.71) with the WOMAC function and pain subscales each accounting for only an additional 2-3% the variation. Residual analysis of the forward stepwise regression models indicated this linear model was an appropriate choice with the residuals being normally distributed and the residual plots indicating there was constant variance and no evidence of nonlinearity.

Discussion Among all the variables examined, the static alignment of the knee as assessed by the mechanical axis was the best single predictor of the peak external knee adduction moment in these subjects with milder symptomatic knee OA. Even when other measures of radiographic disease severity, toe out angle, pain and functional assessment were accounted for, the majority of the variance in the peak adduction moments remained accounted for by the mechanical axis. Thus, subjects with varus knees had larger peak knee adduction moments than subjects with neutral or valgus knees. However, the mechanical axis only accounted for about 50% of the variation in the peak adduction moments. Therefore it is important to dynamically evaluate the knee joint loading environment in order to account for individual gait adaptations which further affect the knee joint loads. The correlation in the present study between the mechanical axis and the adduction moment in the subjects with knee OA was stronger than that of Goh et al. [9] who studied preoperative high tibial osteotomy candidates or Weidenhielm et al. [22] who studied subjects with medial knee OA. This relationship between the mechanical axis and peak knee adduction moment was not previously found among preoperative candidates for high tibial osteotomy [16]. It is presumed that the preoperative high tibial osteotomy subjects had greater clinical symptoms or disease severity than the

106

D.E. Hurnlitz et 01. I Jourml of Orthopardic Reseurcli 20 (2002) I0I -107

current study population. These factors may have had a greater effect on the gait adaptations than that of niechanical axis among the preoperative high tibial osteotomy candidates. In addition, the average varus deformity in the preoperative high tibial osteotomy group was greater than in the present study. Possibly at greater varus angles, the relationship between the mechanical axis and adduction moment is no longer linear, and the two variables are no longer correlated. In addition there may have been insufficient intersubject variation in the mechanical axis with which to detect a correlation in the smaller sample size of this study. Among the radiographic measures of disease severity those most specific to the medial compartment were more related to the mechanical axis and the external knee adduction moment. Previously it has been shown that disease severity as assessed by Kellgren and Lawrence grade or the quantification of the medial joint space width was predictive of the overall peak adduction moment among those with evidence of medial compartment OA [20]. In the present study the Kellgren and Lawrence grade and the qualitative evaluation of the medial joint space width were also correlated with the first and second peak adduction moments. However, the mechanical axis was a better predictor of both the first and second peak adduction moments. Once mechanical axis was accounted for, neither the Kellgren and Lawrence grade nor the medial joint space narrowing grade contributed significantly to explaining the variance in the peak adduction moments. This was due to the fact that the mechanical axis was significantly correlated to these two measures of radiographic disease severity. It has been suggested that increasing the toe out angle during gait is a mechanism for decreasing the knee adduction moment [21]. Presumably an increased toe out angle would result in the ground reaction force being located closer to the knee joint center during the later part of stance and would thus result in a lower second peak adduction moment. The data of Andrews et al. [3] and the data from the current study support this concept. The first peak adduction moment in the OA group, which typically is the larger peak moment, was not significantly correlated with the toe out angle. The toe out angle was not correlated with the mechanical axis. Therefore, subjects with OA with large varus deformities did not automatically adopt an increased toe out angle as a compensatory gait adaptation. Possibly among subjects with greater symptoms this adaptation would be automatically adopted. If subjects were to walk with an increased toe out angle, it is expected that the external knee adduction throughout the entire later part of stance would be reduced and not just the second peak adduction moment. However, the first peak adduction moment would most likely remain unchanged. Consistent with the previous finding that the change in pain was inversely related to a change in adduction

moment [ 121, there was a negative correlation between the WOMAC subscores and the peak adduction moments. In general these correlations were even weaker than those based on radiographic disease severity in the medial compartment for both peak adduction moments or toe out angle for the second peak adduction moment. High variability between subjects in how pain or functional abilities are perceived as most likely contributes to the lower correlations between the peak adduction moments and the WOMAC subscores. Another common automated algorithm in addition to Forward Stepwise for developing regression models is Backward Stepwise. The results from this study are hardly different if the Backward Stepwise Algorithm were used instead of the Forward Stepwise Algorithm. In fact for the first peak, the results are identical. For the second peak the only difference is the inclusion of the medial joint space narrowing, in addition to the variables already included in the Forward Stepwise Model (mechanical axis, toe-out, lateral joint space, WOMAC Function, WOMAC Pain). The inclusion of this extra variable only results in a 1%) increase in the adjusted R2 value (0.76, Backward Stepwise; 0.75 Forward Stepwise). In summary, this study indicated that static alignment as assessed by the mechanical axis was indicative of the dynamic loads on the medial compartment as assessed by the external knee adduction moment. This study was cross-sectional, thus it cannot be determined if higher external knee adduction moments result in greater disease severity and consequently a more varus knee. Conversely, it cannot be determined if the varus deformity itself initiates the increased dynamic load. Once mechanical axis was accounted for, other radiographic measures of disease severity, toe out angle or functional assessment only increased the ability to predict the peak external knee adduction moment by an additional 10-18%. While the mechanical axis was indicative of the peak adduction moments, it only accounted for about 50% of the variation in the peak adduction moments, emphasizing the need for a dynamic evaluation of the loading environment of the knee joint. Understanding which clinical measures of OA or OA progression are most closely associated with the dynamic knee joint loads may ultimately result in a better understanding of the disease process and the development of therapeutic interventions.

Acknowledgements

This work was supported with NIH SCOR grant AR39239. The authors would like to thank Dr. Ali Karrar for making the radiographic measurements of disease severity. The authors would also like to thank Caryn Clares and Jeffrey Sum for their help with the gait

D.E. Hurbcitr

rt

01. I Joumol of Orthopnediir Rcscurch 20 (2002) 101-107

data collection and processing and Dr. Susan Shotte for her assistance with the statistical analysis.

References [I] Altman RD, Asch E, Bloch D, Bole G , Borenstein D, Brandt K, et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986;29:103949, [2] Altman RD, Hochberg M, Murphy Jr. WA, Wolfe F, Lequesne M. Atlas of individual radiographic features in osteoarthritis. Osteoarthritis Cartilage 1995;3:3-70. [3] Andrews M, Noyes FR, Hewett TE, Andriacchi TP. Lower limb alignment and foot angIe are related to stance phase knee adduction in normal subjects: a critical analysis of the reliability of gait analysis data. J Orthop Res 1996:14289-95. [4] Andriacchi TP, Natarajan RN, Hurwitz DE. Musculo-skeletal dynamic locomotion and clinical applications. New York: Raven Press; 1997. p. 37-69. [S] Andriacchi TP, Strickland AB. Lower limb kinetics applied to the study of normal and abnormal walking. Dordrecht: Martiunus Nijhoff; 1985. p. 83-102. [6] Baliunas A, Ryals AR, Hurwitz DE, Kerrar A, Andriacchi TP. Gait adaptations associated with early radiographic tibiofemoral knee osteoarthritis. In: 46th Annual Meeting Orthopaedic Research Society; 2000. p. 260. [7] Bryan JM, Hurwitz DE, Bach BR, Bittar T, Andriacchi TP. A predictive model of outcome in high tibia1 osteotomy. Trans ORS 1997;718. [S] Frost H. Perspectives: a biomechanical model of the pathogenesis of arthroses. Anat Rec 1994;240:19-31. [9] Goh JC, Bose K, Khoo BC. Gait analysis study on patients with varus osteoarthrosis of the knee. Clin Orthop 1993;223-31. [lo] Harrington IJ. Static and dynamic loading patterns in knee joints with deformities. J Bone Joint Surg [Am] 1983;65:247 -59.

I07

1111 Hilding MB, Lanshammar H, Ryd L. A relationship between dynamic and static assessments of knee joint load. Gait analysis and radiography before and after knee replacement in 45 patients. Acta Orthop Scand 1995;66:317-20. [I21 Hurwitz DE, Ryals AR, Block JA, Sharma L, Schnitzer TJ, Andriacchi TP. Knee pain arid joint loading in subjects with knee osteoarthritis. J Orthop Res 2000;18:572-9. [I31 Hurwitz DE. Sumner DR, Andriacchi TP, Sugar DA. Dynamic knee loads during gait predict proximal tibial bone distribution. J Biomech 1998:31:423-30. [I41 Kellgren JH, Lawrence JS. Atlas of standard radiographs. Department of Rheumatology and Medical Illustrations. University of Manchester, Blackwell, Oxford; 1963. [I51 Moreland JR, Bassett LW, Hanker GJ. Radiographic analysis of the axial alignment of the lower extremity. J Bone Joint Surg [Am] 1987;69:745-9. [I61 Prodromos CC, Andriacchi TP, Galante JO. A relationship between gait and clinical changes following high tibial osteotomy. J Bone Joint Surg [Am] 1985;67:1 188-94. [17] Radin EL, Burr DB, Caterson B, Fyhrie D. Brown TD, Boyd RD. Mechanical determinants of osteoarthrosis. Semin Arthritis Rheum 1991;21: 12--21. [ I 81 Schipplein OD, Andriacchi TP. Interaction between active and passive knee stabilizers during level walking. J Orthop Res 1991;9:113-9. [I91 Schnitzer TJ, Popovich JM, Anderson GB, Andriacchi TP. Effect of piroxicam on gait in patients with osteoarthritis of the knee. Arthritis Rheum 1993;36:1207-13. [20] Sharma L, Hurwitz DE, Thonar EJ, Sum JA, Lenz ME, Dunlop DD, et al. Knee adduction moment, serum hyaluronan level. and disease severity in medial tibiofemoral osteoarthritis. Arthritis Rheum 1998;41:123340. [21] Wang JW, Kuo KN, Andriacchi TP, Galante JO. The influence of walking mechanics and time on the results of proximal tibial osteotorny. J Bone Joint Surg [Am] 1990;72:905 -9. [22] Weidenhielm L, Svensson OK, Brostrom LA, Mattsson E. Adduction moment of the knee compared to radiological and clinical parameters in moderate medical osteoarthrosis of the knee. Ann Chir Gynaecol 1994:83:23642.