Population characteristics
The mean positive KAM-time integral, negative KAM-time integral, and maximum KAM were 0.13±0.05 Nms/kg, 0.0039±0.010 Nms/kg, and 0.36±0.12 Nm/kg, respectively. The characteristics of the study population are summarized in Table 1. Male lower extremities had greater values of the positive KAM-time integral (male vs. female, 0.14±0.06 vs. 0.13±0.05 Nms/kg), the maximum KAM (0.37±0.12 vs. 0.35±0.12 Nm/kg), mTFA (1.70±2.43° vs. 0.61±2.46°), knee joint line orientation (KJLO) (1.88±2.16° vs. 1.03±2.18°), AJLO (-2.00±3.86° vs. -2.89±4.00°), and joint line convergence angle (JLCA) (0.64±1.27° vs. 0.41±1.52°). They had lower values of medial proximal tibial angle (MPTA) (85.8±2.16° vs. 86.6±2.16°) (Table 2).
Table 1 Characteristics of the lower extremities included in the analysis
Data summary of lower extremities
|
|
No.
|
937
|
Age (years)
|
36.8 (SD, ±16.5)
|
Sex (men:women)
|
471:466 (50.3%:49.7%)
|
Side (right:left)
|
470:467 (49.8%:50.2%)
|
Height (cm)
|
165.8 (SD, ±8.9)
|
Weight (kg)
|
64.0 (SD, ±12.8)
|
BMI (kg/m2)
|
23.1 (SD, ±3.4)
|
Kinetic data
|
|
Maximum KAM (Nm/kg)
|
0.36 (SD, ±0.12)
|
Positive KAM-time integral (Nms/kg)
|
0.13 (SD, ±0.05)
|
Negative KAM-time integral (Nms/kg)
|
0.0039 (SD, ±0.010)
|
Radiographic measurements
|
|
Medial K-L grade (0/I/II/III/IV)
|
780/139/17/1/0 (83.2%/14.8%/1.8%/0.1%/0%)
|
Lateral K-L grade (0/I/II/III/IV)
|
856/77/4/0/0 (91.4%/8.2%/0.4%/0%/0%)
|
mTFA (°)
|
1.16 (SD, ±2.50)
|
MPTA (°)
|
86.2 (SD, ±2.19)
|
LDFA (°)
|
86.8 (SD, ±2.05)
|
KJLO (°)
|
1.46 (SD, ±2.21)
|
AJLO (°)
|
-2.44 (SD, ±3.96)
|
JLCA (°)
|
0.53 (SD, ±1.40)
|
Tibiofemoral translation
|
3.24 (SD, ±2.50)
|
SD, standard deviation; BMI, body mass index; KAM, knee adduction moment; K-L, Kellgren-Lawrence; mTFA, mechanical tibiofemoral angle; MPTA, medial proximal tibial angle; LDFA, lateral distal femoral angle; KJLO, knee joint line orientation; AJLO, ankle joint line orientation; JLCA, joint line convergence angle
Table 2 Characteristics of the lower extremities according to sex
Characteristics of lower extremities
|
Male (n=471)
|
Female (n=466)
|
P value
|
Age (years)
|
36.4 (SD, ±16.7)
|
37.3 (SD, ±16.3)
|
0.389
|
Side (right:left)
|
237:234 (50.3%:49.7%)
|
233:233 (50%:50%)
|
0.922
|
Height (cm)
|
172.3 (SD, ±6.4)
|
159.3 (SD, ±5.8)
|
<0.001
|
Weight (kg)
|
71.2 (SD, ±11.5)
|
56.6 (SD, ±9.3)
|
<0.001
|
BMI (kg/m2)
|
23.9 (SD, ±3.2)
|
22.3 (SD, ±3.5)
|
<0.001
|
Kinetic data
|
|
|
|
Maximum KAM (Nm/kg)
|
0.37 (SD, ±0.12)
|
0.35 (SD, ±0.12)
|
0.008
|
Positive KAM-time integral (Nms/kg)
|
0.14 (SD, ±0.06)
|
0.13 (SD, ±0.05)
|
<0.001
|
Negative KAM-time integral (Nms/kg)
|
0.0033 (SD, ±0.0083)
|
0.0046 (SD, ±0.0119)
|
0.642
|
Radiographic measurements
|
|
|
|
Medial K-L grade (0/I/II/III/IV)
|
403/64/4/0/0 (85.6%/13.6%/0.8%/0%/0%)
|
377/75/13/1/0 (80.9%/16.1%/2.8%/0.2%/0%)
|
0.039
|
Lateral K-L grade (0/I/II/III/IV)
|
432/37/2/0/0 (91.7%/7.9%/0.4%/0%/0%)
|
424/40/2/0/0 (91.0%/8.6%/0.4%/0%/0%)
|
0.896
|
mTFA (°)
|
1.70 (SD, ±2.43)
|
0.61 (SD, ±2.46)
|
<0.001
|
MPTA (°)
|
85.8 (SD, ±2.16)
|
86.6 (SD, ±2.16)
|
<0.001
|
LDFA (°)
|
86.8 (SD, ±1.95)
|
86.7 (SD, ±2.15)
|
0.448
|
KJLO (°)
|
1.88 (SD, ±2.16)
|
1.03 (SD, ±2.18)
|
<0.001
|
AJLO (°)
|
-2.00 (SD, ±3.86)
|
-2.89 (SD, ±4.00)
|
<0.001
|
JLCA (°)
|
0.64 (SD, ±1.27)
|
0.41 (SD, ±1.52)
|
0.012
|
Tibiofemoral translation
|
3.36 (SD, ±2.40)
|
3.11 (SD, ±2.60)
|
0.125
|
SD, standard deviation; BMI, body mass index; KAM, knee adduction moment; K-L, Kellgren-Lawrence; mTFA, mechanical tibiofemoral angle; MPTA, medial proximal tibial angle; LDFA, lateral distal femoral angle; KJLO, knee joint line orientation; AJLO, ankle joint line orientation; JLCA, joint line convergence angle
Correlation study between the KAM and other variables
The results of the correlation analyses between the KAM and other variables are shown in Table 3. mTFA had the strongest correlation with both the positive KAM-time integral and the maximum KAM (r=0.376 and 0.383, respectively) followed by AJLO (r=-0.258 and -0.269). Both the positive KAM-time integral and the maximum KAM had positive correlation with age (r=0.145 and 0.103), medial Kellgren-Lawrence (K-L) grade (r=0.140 and 0.106), and lateral distal femoral angle (LDFA) (r=0.184 and 0.199) but negative correlation with MPTA (r=-0.240 and -0.250). Male lower extremities had positive correlation with the positive KAM-time integral and the maximum KAM (r=0.125 and 0.087, respectively). Lateral K-L grade had significant positive correlation with only the positive KAM-time integral (r=0.088), but not with the maximum KAM or the negative KAM-time integral. The negative KAM-time integral had negative and positive correlation with mTFA (r=-0.095) and MPTA (r=0.076), respectively. Among the KAM variables, the positive KAM had strong positive correlation with the maximum KAM (r=0.924) and negative correlation with the negative KAM (r=-0.374). The maximum KAM also had negative correlation with the negative KAM (r=-0.442).
Multiple regression analysis for the KAM variables
Multiple regression analyses for the positive and negative KAM-time integral and the maximum KAM that include variables which were significantly correlated in the correlation study except the KAM variables themselves were performed. Final models were obtained after backward elimination steps. Sex, medial K-L grade, mTFA, MPTA, and AJLO comprised the final model for the positive KAM-time integral while only mTFA and AJLO were included in the model for the maximum KAM. When more backward eliminations from the final model for the positive KAM-time integral were performed to compare the adjusted R2 values, both mTFA and AJLO independently explained the variance of the positive KAM-time integral the most while mTFA was the main explanatory parameter of the maximum KAM. The model for negative KAM-time integral included only mTFA (Table 4).
Positive KAM-time integral = 0.295 + (-0.009) x Sex + 0.012 x medial K-L grade + 0.006 x mTFA + (-0.002) x MPTA + (-0.003) x AJLO (R2 = 0.1917, adjusted R2 = 0.1874)
(1 and 2 for male and female, respectively)
Maximum KAM = 0.330 + 0.017 x mTFA + (-0.006) x AJLO (R2 = 0.1843, adjusted R2 = 0.1825)
Negative KAM-time integral = 0.0044 + (-0.0004) x mTFA (R2 = 0.0091, adjusted R2 = 0.0081)
Table 4 Factors associated with the KAM variables in the multiple regression analysis
(a)
|
Non-standardized
|
Standardized β
|
t
|
P-value
|
VIF
|
β
|
Standard error
|
Sex
Male: 1
Female: 2
|
-0.009
|
0.003
|
-0.087
|
-2.806
|
0.005
|
1.097
|
Medial K-L grade
|
0.012
|
0.004
|
0.096
|
3.123
|
0.002
|
1.084
|
mTFA
|
0.006
|
0.001
|
0.256
|
6.736
|
<0.001
|
1.663
|
MPTA
|
-0.002
|
0.001
|
-0.076
|
-2.099
|
0.036
|
1.498
|
AJLO
|
-0.003
|
0.000
|
-0.206
|
-6.736
|
<0.001
|
1.077
|
Intercept
|
0.295
|
0.078
|
|
3.794
|
<0.001
|
|
R2=0.1917, adjusted R2=0.1874
(b)
|
Non-standardized
|
Standardized β
|
t
|
P-value
|
VIF
|
β
|
Standard error
|
mTFA
|
0.017
|
0.001
|
0.342
|
11.329
|
<0.001
|
1.686
|
AJLO
|
-0.006
|
0.001
|
-0.198
|
-6.558
|
<0.001
|
1.064
|
Intercept
|
0.330
|
0.004
|
|
75.473
|
<0.001
|
|
R2 = 0.1843, adjusted R2=0.1825
(c)
|
Non-standardized
|
Standardized β
|
t
|
P-value
|
β
|
Standard error
|
mTFA
|
-0.0004
|
0.0001
|
-0.0955
|
-2.9330
|
0.0034
|
Intercept
|
0.0044
|
0.0004
|
|
11.9840
|
<0.001
|
R2 = 0.0091, adjusted R2 = 0.0081
(d)
Adjusted R2 for Positive KAM-time integral
|
0.1874
|
Sex + medial K-L grade + mTFA + MPTA + AJLO
|
0.1814
|
Medial K-L grade + mTFA + MPTA + AJLO
|
0.1844
|
Sex + medial K-L grade + mTFA + AJLO
|
0.1797
|
Sex + mTFA + + MPTA + AJLO
|
0.1487
|
Sex + medial K-L grade + mTFA + MPTA
|
0.1487
|
Sex + medial K-L grade + MPTA + AJLO
|
Adjusted R2 for Maximum KAM
|
0.1825
|
mTFA + AJLO
|
0.1458
|
mTFA
|
0.0712
|
AJLO
|
(a) Positive KAM-time integral (b) maximum KAM (c) Negative KAM-time integral (d) adjusted R2 for predictors of the positive KAM-time integral and maximum KAM; VIF, variance inflation factor; K-L, Kellgren-Lawrence; mTFA, mechanical tibiofemoral angle; MPTA, medial proximal tibial angle; KJLO, knee joint line orientation; AJLO, ankle joint line orientation
Analysis according to mTFA by the interval of 1°
To investigate the relationship between the KAM values and limb alignment, the lower extremities were grouped according to mTFA by the interval of 1°. There was a tendency that the positive KAM-time integral and the maximum KAM increased as mTFA increased (P <0.001, Table 5). The positive KAM-time integral and the maximum KAM of alignment groups which showed significant difference in the post-hoc analysis of ANOVA are shown in Table 6. The positive KAM-time integral values of lower extremities of varus alignment groups with 3°≤mTFA<8° were significantly greater than those of neutral or valgus alignment groups with -5°≤mTFA<0°. Similar relationship was observed in the post-hoc analysis of the maximum KAM among those with 3°≤mTFA<8° and -5°≤mTFA<0°, except that the P-value of the post-hoc analysis of the maximum KAM between the alignment groups -1°≤mTFA<0° and 5°≤mTFA<6° was slightly greater than 0.05 (P=0.088). Although the positive KAM-time integral and the maximum KAM showed decreasing tendency in alignment groups of -5°≤mTFA<0°, there was no significant difference of the positive KAM-time integral and the maximum KAM within these groups (Table 6). There was no increasing tendency of the negative KAM as the amount of valgus increased (Table 5). Every alignment group with 3° to 7° of mTFA included more male than female extremities whereas female lower extremities were more prevalent in alignment groups with -4° to 0° of mTFA (P<0.001).
Table 5 The positive and negative KAM-time integral, and the maximum KAM values according to the alignment groups of one-degree interval
|
No.
|
Male
No.
|
Female
No.
|
Age (years)
|
Positive KAM-time integral (Nms/kg)
|
Negative KAM-time integral (Nms/kg)
|
Maximum KAM
(Nm/kg)
|
mTFA≥10
|
1 (0.1%)
|
0
|
1
|
59.0
|
0.266
|
0.000
|
0.656
|
9≤mTFA<10
|
1 (0.1%)
|
1
|
0
|
61.0
|
0.194
|
0.000
|
0.492
|
8≤mTFA<9
|
1 (0.1%)
|
1
|
0
|
61.0
|
0.254
|
0.000
|
0.658
|
7≤mTFA<8
|
9 (1.0%)
|
3
|
6
|
49.6
|
0.193 (SD, ±0.071)
|
0.001 (SD, ±0.002)
|
0.478 (SD, ±0.145)
|
6≤mTFA<7
|
18 (1.9%)
|
14
|
4
|
48.7
|
0.181 (SD, ±0.062)
|
0.002 (SD, ±0.003)
|
0.479 (SD, ±0.139)
|
5≤mTFA<6
|
15 (1.6%)
|
11
|
4
|
43.7
|
0.174 (SD, ±0.062)
|
0.007 (SD, ±0.019)
|
0.441 (SD, ±0.115)
|
4≤mTFA<5
|
73 (7.8%)
|
50
|
23
|
42.2
|
0.164 (SD, ±0.051)
|
0.003 (SD, ±0.006)
|
0.429 (SD, ±0.101)
|
3≤mTFA<4
|
103 (11.0%)
|
65
|
38
|
39.8
|
0.153 (SD, ±0.054)
|
0.003 (SD, ±0.006)
|
0.407 (SD, ±0.120)
|
2≤mTFA<3
|
131 (14.0%)
|
67
|
64
|
36.7
|
0.142 (SD, ±0.055)
|
0.003 (SD, ±0.007)
|
0.383 (SD, ±0.115)
|
1≤mTFA<2
|
138 (14.7%)
|
84
|
54
|
39.9
|
0.136 (SD, ±0.051)
|
0.003 (SD, ±0.013)
|
0.373 (SD, ±0.121)
|
0≤mTFA<1
|
141 (15.0%)
|
67
|
74
|
36.0
|
0.128 (SD, ±0.047)
|
0.005 (SD, ±0.011)
|
0.350 (SD, ±0.103)
|
-1≤mTFA<0
|
121 (12.9%)
|
50
|
71
|
33.2
|
0.119 (SD, ±0.043)
|
0.005 (SD, ±0.012)
|
0.334 (SD, ±0.105)
|
-2≤mTFA<-1
|
86 (9.2%)
|
29
|
57
|
32.8
|
0.108 (SD, ±0.042)
|
0.004 (SD, ±0.009)
|
0.308 (SD, ±0.099)
|
-3≤mTFA<-2
|
59 (6.3%)
|
12
|
47
|
30.4
|
0.104 (SD, ±0.045)
|
0.006 (SD, ±0.014)
|
0.293 (SD, ±0.121)
|
-4≤mTFA<-3
|
22 (2.3%)
|
8
|
14
|
26.8
|
0.097 (SD, ±0.037)
|
0.007 (SD, ±0.010)
|
0.289 (SD, ±0.089)
|
-5≤mTFA<-4
|
13 (1.4%)
|
7
|
6
|
29.8
|
0.096 (SD, ±0.074)
|
0.006 (SD, ±0.011)
|
0.265 (SD, ±0.121)
|
-6≤mTFA<-5
|
2 (0.2%)
|
1
|
2
|
41.0
|
0.156 (SD, ±0.066)
|
0.001 (SD, ±0.001)
|
0.397 (SD, ±0.146)
|
-7≤mTFA<-6
|
1 (0.1%)
|
1
|
1
|
23.0
|
0.328
|
0.000
|
0.600
|
-8≤mTFA<-7
|
2 (0.2%)
|
1
|
1
|
23.5
|
0.078 (SD, ±0.046)
|
0.004 (SD, ±0.005)
|
0.221 (SD, ±0.100)
|
Total
|
937
|
471
|
466
|
|
|
|
|
|
|
|
SD, standard deviation; KAM, knee adduction moment; mTFA, mechanical tibiofemoral angle
Interobserver reliability test
Table 7 shows ICCs for each variable. Every coefficient value was greater than 0.8, implying good to excellent interobserver reliability 25.
Table 7 Interobserver reliability of radiographic measurements
Interobserver reliability of radiographic measurements
|
ICC
|
Medial K-L grade (0/I/II/III/IV)
|
0.823 (95% CI, 0.671-0.909)
|
Lateral K-L grade (0/I/II/III/IV)
|
1
|
mTFA (°)
|
0.996 (95% CI, 0.991-0.998)
|
MPTA (°)
|
0.949 (95% CI, 0.898-0.974)
|
LDFA (°)
|
0.98 (95% CI, 0.96-0.99)
|
KJLO (°)
|
0.966 (95% CI, 0.93-0.984)
|
AJLO (°)
|
0.987 (95% CI, 0.974-0.994)
|
JLCA (°)
|
0.879 (95% CI, 0.763-0.94)
|
Tibiofemoral translation
|
0.986 (95% CI, 0.97-0.993)
|
ICC, intraclass correlation coefficient; CI, confidence interval; K-L, Kellgren-Lawrence; mTFA, mechanical tibiofemoral angle; MPTA, medial proximal tibial angle; LDFA, lateral distal femoral angle; KJLO, knee joint line orientation; AJLO, ankle joint line orientation; JLCA, joint line convergence angle