Participants characteristics
In total, 200 participants were included in the study, i.e., 129 patients diagnosed with CSVD. Clinical data are presented in Table 1—no significant group differences in age, sex, ethic, marriage, BMI and shoe size. There was significance difference between CSVD group and controls group in education, drinking, hypertension, type 2 diabetes mellitus (T2DM), history of hypoglycemia and stroke, Other neurological disorders and cardiovascular disorder. Compared to controls, participants with CSVD had shorter stride length, slower stride speed, shorter cadence, longer stance time/phase, longer stride time, shorter swing phase, and smaller average toe-off/heel stride angle either in ST or in DT walking conditions (all P<0.05).
Multivariable logistic regression of potential diagnostic parameters for CSVD.
The potential diagnostic parameters for CSVD were age, T2DM and average heel strike angle both in single-task walking test [age: P = 0.010, OR (odds ratio) = 0.858, 95% CI (confidence interval) = 0.764-0.963; T2DM: P = 0.033, OR = 15.096, 95%CI = 1.242-183.418; average heel strike angle: P<0.001, OR = 0.608, 95%CI = 0.480-0.771]and in dual-task walking test [age: P = 0.020, OR = 0.881, 95% CI = 0.791-0.980; T2DM: P = 0.023, OR = 17.253, 95%CI = 1.476-201.632; average heel strike angle: P = 0.015, OR = 0.731, 95%CI = 0.568-0.941]; drinking in single-task walking test [P = 0.017, OR = 0.031, 95%CI = 0.002-0.532] and stance phase in dual-task walking test[P = 0.041, OR = 1.505, 95% CI = 1.017-2.228] (Table 2).
Correlations analysis for potential diagnostic clinical parameters and gait variables for CSVD.
A correlation analysis was conducted for potential diagnostic clinical parameters and gait variables for CSVD. The results indicated that age, T2DM, stance phase, stride time, stance time were positively associated with CSVD (all p < 0.05) and drinking, stride length, stride speed, cadence, swing phase, average toe-off angle and average heel strike angle negatively correlated with CSVD (all p < 0.05) (Table 3).
Diagnostic accuracy of gait parameters for CSVD
Receiver operating characteristic curves were employed to demonstrate how gait tests differentiated CSVD from healthy controls (Figure 2 and Figure 3). Factors of stride length, stride speed, cadence, stance phase, swing phase, stride time, stance time, toe-off angle and heel strike angle (AUC = 0.858, P < 0.001, 95%CI: 0.807-0.909, sensitivity, 83.1%; specificity, 76.7%;) showed moderate ability to separate CSVD from healthy controls (Table 4) in single-task walking condition and stride length, stride speed, stance phase, swing phase, toe-off angle and heel strike angle (AUC = 0.865, P<0.001, 95%CI: 0.817-0.914, sensitivity: 91.5%, specificity: 70.5%) in dual-task walking condition (Table 5). Among them, heel strike angle in dual-task walking condition was the best one to differentiated CSVD from healthy participants.
Table 1. Characteristics of the study participants.
|
all(n=200)
|
CSVD(n=129)
|
control(n=71)
|
P-value
|
Demographic characteristics
|
|
|
|
age,y
|
70.27±9.48
|
71.06±9.82
|
68.83±8.71
|
0.112
|
Sex (male),%
|
120(60.0%)
|
83(64.3%)
|
37(52.1%)
|
0.091
|
ethic
|
|
|
|
|
the Han nationality
|
148(74.0%)
|
99(76.7%)
|
49(69.0%)
|
0.464
|
the Zhuang nationality
|
48(24.0%)
|
28(21.7%)
|
20(28.2%)
|
others
|
4(2.0%)
|
2(1.6%)
|
2(2.8%)
|
marriage
|
|
|
|
|
married
|
185(92.5%)
|
120(93.0%)
|
65(91.5%)
|
0.705
|
bereave
|
15(7.5%)
|
9(7.0%)
|
6(8.5%)
|
BMI,Kg/m2
|
23.2(2.5)
|
23.3(4.2)
|
22.9(5.2)
|
0.342*
|
shoe size
|
40.0(2.0)
|
40.0(4.0)
|
39.0(5.0)
|
0.230*
|
education
|
|
|
|
|
illiteracy
|
8(4.0%)
|
8(6.2%)
|
0(0%)
|
0.020
|
primary school
|
37(18.5%)
|
28(21.7%)
|
9(12.7%)
|
secondary school and above
|
155(77.5%)
|
93(72.1%)
|
62(87.3%)
|
fall,yes,n%
|
28(14.0%)
|
19(14.7%)
|
9(12.7%)
|
0.689
|
smoke,yes,n%
|
48(24.0%)
|
32(24.8%)
|
16(22.5%)
|
0.719
|
drinking,yes,n%
|
45(22.5%)
|
22(17.1%)
|
23(32.3%)
|
0.013
|
hypertension,yes,n%
|
133(66.5%)
|
99(76.7%)
|
34(47.9%)
|
<0.001
|
T2DM,yes,n%
|
41(20.5%)
|
33(25.6%)
|
8(11.2%)
|
0.016
|
hypoglycemia,yes,n%
|
10(5.0%)
|
0(0%)
|
10(14.1%)
|
<0.001
|
syncope,yes,n%
|
14(7.0%)
|
7(5.4%)
|
7(9.9%)
|
0.240
|
dementia,yes,n%
|
13(6.5%)
|
13(10.1%)
|
0(0%)
|
0.006
|
stroke,yes,n%
|
114(57.0%)
|
108(83.7%)
|
6(8.5%)
|
<0.001
|
Other neurological disorders,yes,n%
|
101(50.5%)
|
100(77.5%)
|
1(1.4%)
|
<0.001
|
cardiovascular disorders,yes,n%
|
137(68.5%)
|
102(79.1%)
|
35(49.3%)
|
<0.001
|
visual system diseases,yes,n%
|
2(1.0%)
|
2(1.6%)
|
0(0%)
|
0.292
|
musculoskeletal system diseases,yes,n%
|
2(1.0%)
|
2(1.6%)
|
0(0%)
|
0.292
|
incontinence,yes,n%
|
5(2.5%)
|
5(3.9%)
|
0(0%)
|
0.093
|
single-task Left (STL)
|
|
|
|
STL-stride length(m)
|
0.99±0.25
|
0.90±0.25
|
1.16±0.13
|
<0.001
|
STL-stride speed(m/s)
|
0.85±0.26
|
0.75±0.26
|
1.03±0.15
|
<0.001
|
STL-cadence (steps/min)
|
101.51±13.39
|
98.31±14.48
|
107.32±8.54
|
<0.001
|
STL-stance phase (%)
|
66.09±3.73
|
67.49±3.65
|
63.54±2.23
|
<0.001
|
STL-swing phase (%)
|
33.91±3.73
|
32.51±3.65
|
36.46±2.23
|
<0.001
|
STL-stride time(s)
|
1.17(0.20)
|
1.23(0.24)
|
1.13(0.13)
|
<0.001*
|
STL-stance time(s)
|
0.76(0.18)
|
0.83(0.21)
|
0.71(0.10)
|
<0.001*
|
STL-swing time(s)
|
0.40(0.03)
|
0.40(0.04)
|
0.41(0.04)
|
0.066*
|
STL-average toe-off angle (°)
|
39.78±9.02
|
36.48±9.17
|
45.77±4.57
|
<0.001
|
STL-Average heel strike angle(°)
|
28.75(13.2)
|
23.9(13.1)
|
34.8(6.9)
|
<0.001*
|
single-task Right (STR)
|
|
|
|
STR-stride length(m)
|
1.00±0.27
|
0.91±0.29
|
1.15±0.14
|
<0.001
|
STR-stride speed(m/s)
|
0.85±0.26
|
0.75±0.26
|
1.03±0.15
|
<0.001
|
STR-cadence (steps/min)
|
101.51±13.39
|
98.31±14.48
|
107.32±8.54
|
<0.001
|
STR-stance phase(%)
|
66.13±4.06
|
67.61±4.10
|
63.45±2.17
|
<0.001
|
STR-swing phase(%)
|
33.87±4.06
|
32.39±4.10
|
36.55±2.17
|
<0.001
|
STR-stride time(s)
|
1.17(0.19)
|
1.23(0.23)
|
1.12(0.13)
|
<0.001*
|
STR-stance time(s)
|
0.76(0.18)
|
0.82(0.21)
|
0.71(0.10)
|
<0.001*
|
STR-swing time(s)
|
0.41(0.05)
|
0.40(0.05)
|
0.41(0.03)
|
0.051*
|
STR-average toe-off angle(°)
|
40.09±8.83
|
36.99±9.17
|
45.72±4.25
|
<0.001
|
STR-Average heel strike angle(°)
|
26.95±9.16
|
23.11±8.62
|
33.93±5.11
|
<0.001
|
dual-task Left (DTL)
|
|
|
|
DTL-stride length (m)
|
0.94±0.26
|
0.84±0.25
|
1.11±0.14
|
<0.001
|
DTL-stride speed (m/s)
|
0.78±0.27
|
0.69±0.26
|
0.95±0.18
|
<0.001
|
DTL-cadence (steps/min)
|
98.99±14.95
|
96.47±16.11
|
103.57±11.31
|
<0.001
|
DTL-stance phase (%)
|
67.00±3.88
|
68.41±3.79
|
64.44±2.49
|
<0.001
|
DTL-swing phase (%)
|
33.00±3.88
|
31.59±3.79
|
35.57±2.49
|
<0.001
|
DTL-stride time (s)
|
1.20(0.24)
|
1.23(0.27)
|
1.16(0.19)
|
<0.001*
|
DTL-stance time (s)
|
0.80(0.21)
|
0.83(0.22)
|
0.74(0.15)
|
<0.001*
|
DTL-swing time (s)
|
0.41(0.05)
|
0.40(0.05)
|
0.41(0.05)
|
0.009*
|
DTL-average toe-off ground angle(°)
|
37.90±9.49
|
34.27±9.56
|
44.50±4.50
|
<0.001
|
DTL-Average heel strike angle(°)
|
25.66±9.10
|
21.82±8.52
|
32.65±5.09
|
<0.001
|
dual-task Right(DTR)
|
|
|
|
DTR-stride length(m)
|
0.93±0.28
|
0.85±0.30
|
1.08±0.14
|
<0.001
|
DTR-stride speed(m/s)
|
0.78±0.27
|
0.69±0.26
|
0.95±0.18
|
<0.001
|
DTR-cadence (steps/min)
|
98.99±14.95
|
96.47±16.11
|
103.57±11.31
|
<0.001
|
DTR-stance phase (%)
|
67.08±4.15
|
68.56±4.15
|
64.40±2.49
|
<0.001
|
DTR-swing phase (%)
|
32.92±4.15
|
31.44±4.15
|
35.60±2.49
|
<0.001
|
DTR-stride time(s)
|
1.18(0.24)
|
1.21(0.27)
|
1.16(0.19)
|
0.002*
|
DTR-stance time(s)
|
0.79(0.22)
|
0.82(0.24)
|
0.73(0.13)
|
<0.001*
|
DTR-swing time(s)
|
0.40(0.05)
|
0.39(0.05)
|
0.40(0.04)
|
0.001*
|
DTR-average toe-off angle (°)
|
40.40(12.50)
|
35.90(14.70)
|
44.30(6.0)
|
<0.001*
|
DTR-Average heel strike angle(°)
|
25.16±9.03
|
21.32±8.33
|
32.15±5.32
|
<0.001
|
STL: single task walking for the left foot; STR: single task walking for the Right foot; DTL: dual-task walking for the left foot; DTR: dual-task walking for the right foot; T2DM: type 2 diabetes mellitus.
*: Wilcoxon rank sum test.
Table 2. Multivariate binary logistic regression analysis for diagnosis of CSVD and controls.
|
Single-task walking test
|
Dual-task walking test
|
Variables
|
P
|
Exp(B)
|
β(95%CI)
|
P
|
Exp(B)
|
β(95%CI)
|
age
|
0.010
|
0.858
|
0.764-0.963
|
0.020
|
0.881
|
0.791-0.980
|
drinking (yes)
|
0.017
|
0.031
|
0.002-0.532
|
0.059
|
0.071
|
0.005-1.106
|
T2DM (yes)
|
0.033
|
15.096
|
1.242-183.418
|
0.023
|
17.253
|
1.476-201.632
|
average heel strike angle
|
<0.001
|
0.608
|
0.480-0.771
|
0.015
|
0.731
|
0.568-0.941
|
stance phase
|
-
|
-
|
-
|
0.041
|
1.505
|
1.017-2.228
|
T2DM: type 2 diabetes mellitus.
Tabel 3. Correlation analysis for variables associated with diagnosis of CSVD and healthy controls.
Variables
|
|
diagnosis (single-task walking test)
|
diagnosis (dual-task walking test)
|
age
|
rho
|
.143*
|
.143*
|
|
P
|
0.044
|
0.044
|
drinking
|
rho
|
-.176*
|
-.176*
|
|
P
|
0.013
|
0.013
|
T2DM
|
rho
|
.170*
|
.170*
|
|
P
|
0.016
|
0.016
|
stride length(m)
|
rho
|
-.483**
|
-.491**
|
|
P
|
0.000
|
0.000
|
stride speed(m/s)
|
rho
|
-.523**
|
-.482**
|
|
P
|
0.000
|
0.000
|
Cadence (steps/min)
|
rho
|
-.335**
|
-.232**
|
|
P
|
0.000
|
0.001
|
stance phase (%)
|
rho
|
.546**
|
.521**
|
|
P
|
0.000
|
0.000
|
swing phase (%)
|
rho
|
-.546**
|
-.521**
|
|
P
|
0.000
|
0.000
|
stride time(s)
|
rho
|
.335**
|
.232**
|
|
P
|
0.000
|
0.001
|
stance time (s)
|
rho
|
.415**
|
.327**
|
|
P
|
0.000
|
0.000
|
average toe-off angle (°)
|
rho
|
-.540**
|
-.530**
|
|
P
|
0.000
|
0.000
|
average heel strike angle (°)
|
rho
|
-.593**
|
-.606**
|
|
P
|
0.000
|
0.000
|
*. Correlation is significant at the 0.05 level (2-tailed).
**. Correlation is significant at the 0.01 level (2-tailed).
Table 4. Variables of ROC analysis for gait parameters distinguishing the individuals with CSVD and healthy controls in single-task walking tests.
Test Result Variable(s)
|
Area
|
Std. Errora
|
Asymptotic Sig.b
|
95%CI
|
sensitivity
|
specificity
|
Youden’s index
|
ST-stride length
|
0.792
|
0.031
|
<0.001
|
0.731-0.852
|
0.887
|
0.620
|
0.507
|
ST-stride speed
|
0.815
|
0.029
|
<0.001
|
0.758-0.873
|
0.845
|
0.690
|
0.535
|
ST-cadence
|
0.702
|
0.036
|
<0.001
|
0.631-0.772
|
0.887
|
0.519
|
0.407
|
ST-stance phase
|
0.171
|
0.029
|
<0.001
|
0.115-0.227
|
0.859
|
0.713
|
-0.572
|
ST-swing phase
|
0.829
|
0.029
|
<0.001
|
0.773-0.885
|
0.859
|
0.713
|
0.572
|
ST-stride time
|
0.298
|
0.036
|
<0.001
|
0.228-0.369
|
0.859
|
0.535
|
-0.394
|
ST-stance time
|
0.250
|
0.034
|
<0.001
|
0.183-0.316
|
0.958
|
0.496
|
-0.454
|
ST-toe-off angle
|
0.826
|
0.029
|
<0.001
|
0.769-0.882
|
0.746
|
0.798
|
0.545
|
ST-heel strike angle
|
0.858
|
0.026
|
<0.001
|
0.807-0.909
|
0.831
|
0.767
|
0.598
|
ST: single-task; 95%CI:95% Confidence Interval.
a.Under the nonparametric assumption; b.Null hypothesis: true area = 0.5.
Table 5. Variables of ROC analysis for gait parameters distinguishing the individuals with CSVD and healthy controls.
Test Result Variable(s)
|
Area
|
Std. Errora
|
Asymptotic Sig.b
|
95%CI
|
sensitivity
|
specificity
|
Youden’s index
|
DT-stride length
|
0.796
|
0.031
|
<0.001
|
0.736-0.856
|
0.901
|
0.651
|
0.553
|
DT-stride speed
|
0.791
|
0.031
|
<0.001
|
0.730-0.851
|
0.901
|
0.574
|
0.475
|
DT-stance phase
|
0.186
|
0.030
|
<0.001
|
0.128-0.244
|
0.831
|
0.682
|
-0.513
|
DT-swing phase
|
0.814
|
0.030
|
<0.001
|
0.756-0.872
|
0.831
|
0.682
|
0.513
|
DT-toe-off angle
|
0.820
|
0.029
|
<0.001
|
0.764-0.876
|
0.915
|
0.620
|
0.536
|
DT-heel strike angle
|
0.865
|
0.025
|
<0.001
|
0.817-0.914
|
0.915
|
0.705
|
0.621
|
DT: dual-task; 95%CI:95% Confidence Interval.
a. Under the nonparametric assumption; b. Null hypothesis: true area = 0.5.