Descriptive statistics analysis results. CVI was measured in 164 of 172 eyes; eight eyes were excluded due to poor image quality. The age of the volunteers ranged from five to eighteen years with a mean value ± standard deviation (SD) of 10.05±2.65 years. Mean SE was -1.90 ± 2.48D (ranging from -10.75D to 6.75D). Mean AL is 24.34 ± 1.29mm (ranging from 21.50mm to 27.94mm) (Table 1). The mean CT in the central foveal was 269.87 ± 63.32µm (ranging from 93.00 µm to 443.00µm). The mean Sf-CVI in r=1mm region to the central foveal was 67.65±2.42% (ranging from 61.49 to 77.27%). The mean Sf-CVI in r=3mm region to the central foveal was 67.68±1.98 % (ranging from 62.63 to 74.13%) (Table 2).
Table 1. Demographics, clinical and choroidal characteristics of study subjects (n = 164).
|
Variables
|
Mean ± SD
|
Number (%)
|
Age(years)
|
10.05±2.65
|
85
|
5-9
|
7.88±1.17
|
42 (49.41%)
|
10-13
|
11.49±1.01
|
35 (41.18%)
|
14-18
|
15.25±1.39
|
8 (9.41%)
|
Gender
|
|
85
|
male
|
-
|
45(52.94%)
|
female
|
-
|
40(47.06%)
|
AL (mm)
|
24.34±1.29
|
164
|
≤25
|
23.70±0.75
|
119(72.56%)
|
>25
|
26.11±0.66
|
45(27.43%)
|
SE (diopter)
|
-1.90±2.48
|
164
|
≥-0.25
|
0.84±1.67
|
33(20.12%)
|
-0.25~-3.0
|
-1.49±0.69
|
94(57.32%)
|
-3.0~-6.0
|
-4.02±0.72
|
23(14.02%)
|
≤-6.0
|
-7.66±1.32
|
14(8.54%)
|
Data presented are means ± standard deviations (SD), except for gender, which is number (%).
AL: Axial length; SE: Spherical equivalent; SFCT: sub-foveal choroidal thickness;
CVI: choroidal vascularity index.
Table 2. Choroidal characteristics of study subjects (n = 164).
|
Variables
|
Mean ± SD
|
Range
|
SFCT (µm)
|
269.87±63.32
|
93.00 to 443.00
|
r=1mm
|
|
|
Horizontal CVI (%)
|
67.99±2.69
|
61.49 to 78.72
|
Vertical CVI (%)
|
67.36±2.65
|
60.40 to 75.83
|
Mean CVI (%)
|
67.65±2.42
|
61.49 to 77.27
|
Temporal (%)
|
67.92±3.00
|
58.93 to 77.83
|
Nasal (%)
|
68.05+3.34
|
60.84 to 79.60
|
Superior (%)
|
67.45+2.90
|
61.35 to 75.70
|
Inferior (%)
|
67.27+3.00
|
57.84 to 78.46
|
r=3mm
|
|
|
Horizontal CVI (%)
|
67.97±2.40
|
59.98 to 75.47
|
Vertical CVI (%)
|
67.44±2.08
|
60.65 to 73.05
|
Mean CVI (%)
|
67.68±1.98
|
62.63 to 74.13
|
Temporal (%)
|
67.32+5.74
|
58.51 to 75.82
|
Nasal (%)
|
68.22 + 2.95
|
61.30 to 76.09
|
Superior (%)
|
67.47+2.35
|
61.22 to 73.38
|
Inferior (%)
|
67.42+2.31
|
60.08 to 73.06
|
SFCT: sub-foveal choroidal thickness; CVI: choroidal vascularity index.
Choroidal Remodeling Distribution Pattern. The mean CVI of four regions above was compared in the different myopia groups respectively. In high myopia group, the measurements of CVI can be ordered as N-CVI (70.17, r=1;69.54, r=3), T-CVI (70.09, r=1;69.08, r=3), I-CVI (68.75, r=1;68.62, r=3) and S-CVI (67.95, r=1;68.12, r=3). In general, regardless of r=1mm or r=3mm, the distribution pattern of CVI was similar. Remarkably, irrespective of the group, CVI of the horizontal meridian was consistently greater than that of the vertical meridian, even though the difference was not significant(p>0.05). It was worth mentioning that in all groups the mean N-CVI was always the greatest except the low myopia group in which T-CVI had the greatest value. However, One-way ANOVA analysis showed there were no significant differences in CVI of four regions in either myopia groups(p>0.05).
CVI in the horizontal meridian underwent the largest change as myopia worsened. Temporal and nasal CVI within the r=1mm and r=3mm sub foveal range were positively associated with degree of myopia in young patients (Table 3, Figure 2). Multiple linear regression results revealed significant correlations between SE and T1-CVI (p < 0.05, r2= 0.082, β= 0.194), N1-CVI (p < 0.05, r2= 0.039, β= 0.212) (Table 4). Simple linear regression results revealed that mean Sf1-CVI (p < 0.05, r2 = 0.08) and Sf3-CVI (p < 0.05, r2 = 0.07) were negatively correlated with SE; T1-CVI (p < 0.05, r2 = 0.09) and T3-CVI (p < 0.05, r2 = 0.05) were negatively correlated with SE; N1-CVI (p < 0.05, r2 = 0.05) and N3-CVI (p < 0.05, r2 = 0.04) were negatively correlated with SE (Figure 3).
Table 3. Results of simple linear regression analyses between different SE groups and CVI
CVI
|
Beta
|
p-value
|
R-square
|
T1CVI
N1CVI
S1CVI
|
0.270
0.228
0.132
|
0.000
0.003
0.093
|
0.073
0.052
0.017
|
I1CVI
|
0.130
|
0.097
|
0.017
|
T3CVI
N3CVI
S3CVI
|
0.213
0.184
0.145
|
0.005
0.016
0.063
|
0.045
0.028
0.021
|
I3CVI
|
0.150
|
0.055
|
0.023
|
SE is treated as ordinal categorical variable. The ordinal categorical values are as follows:≥-0.25D=0;-0.25D~-3.0D=1;-3.0D~-6.0D=2; ≤-6.0D=3.
Table 4. Multiple linear regression analyses for age, AL and SE as correlates of CVI(r=1mm)
CVI
|
Variables
|
Beta
|
p-value
|
Adjusted R-square
|
T1CVI
|
Age
AL
SE
|
0.013
0.118
0.194
|
0.878
0.234
0.044
|
0.082
|
N1CVI
|
Age
AL
SE
|
-0.084
0.059
0.212
|
0.308
0.557
0.030
|
0.039
|
S1CVI
|
Age
AL
SE
|
-0.048
0.013
0.351
|
0.571
0.900
0.181
|
0.001
|
I1CVI
|
Age
AL
SE
|
-0.107
0.111
0.085
|
0.204
0.286
0.401
|
0.011
|
All the covariates presented are treated as ordinal categorical variables. The ordinal categorical values
are as follows:(1) Age:5-9yrs=0;10-13yrs=1;14-18yrs=2;(2)AL:≤25mm=0;>25mm=1;(3)≥-0.25D=0;
-0.25D~-3.0D=1;-3.0D~-6.0D=2; ≤-6.0D=3.
CVI was a more stable measure index than CT. For the CVI of different regions, all the coefficients of variation were less than 5%. In contrast, the coefficient of variation of CT was greater than 20%, which was more than 4 times higher compared to the same index of CVI. With adjusted SE and AL, there was no significant difference in mean Sf-CVI across age groups (p > 0.05). There was, however, significant difference in choroidal thickness (CT) across different age groups (p < 0.05).