Demographic characteristics
4500 participants were recruited in the initial research. A total of 4411 participants were available for the statistical analysis after excluding 88 subjects with missing data, 1 subject diagnosed with Down syndrome. Table 1 presents the demographic characteristics of the analysis cohort by myopic or non-myopic. The response rate of the questionnaire was 749 out of 4411. One hundred and thirty-two subjects chose “unknown” in the questionnaire and were excluded during analysis.
Table 1 Demographic characteristics
Variables
|
Myopic
|
Non-myopic
|
Total
|
P value
|
Age (years)
|
10 (8, 11)
|
8 (7, 10)
|
9 (7, 10)
|
<0.001a
|
Height (cm)
|
141.0
(133.0, 149.0)
|
132.0
(125.0, 141.0)
|
135.0
(127.0, 144.5)
|
<0.001a
|
Weight (kg)
|
36.1
(29.7, 44.4)
|
29.9
(25.3, 37.0)
|
31.8
(26.4, 40.0)
|
<0.001a
|
BMI (kg/m2)
|
17.95
(16.43, 20.36)
|
17.07
(15.81, 18.94)
|
17.36
(15.99, 19.46)
|
<0.001a
|
Gender
|
|
|
|
<0.001b
|
Male
|
684(30.0%)
|
1598(70.0%)
|
2282
|
|
female
|
800(37.6%)
|
1329(62.4%)
|
2129
|
|
Wearing spectacles
|
|
|
|
<0.001b
|
Yes
|
377(80.9%)
|
89(19.1%)
|
466
|
|
No
|
1107(28.1%)
|
2838(71.9%)
|
3945
|
|
Total
|
1484 (33.6%)
|
2927(66.4%)
|
4411
|
|
a. Mann-Whitney U test b. Chi-square test
Data are presented as medians and quartiles (p25, p75) or as n (%).
Prevalence of myopia
The overall prevalence of myopia was 33.6% (1484 of 4411). The prevalence of high myopia was 0.6% (27 of 4411). The prevalence rate of myopia in different grades and sexes was shown in Table 2. The incidence of myopia increased significantly as the grade levels rose (x²=522.169, p <0.001), from 8.6% in first grade to 55.6% in sixth grade. After adjusted by Bonferroni method, significant difference was found between every two grades from 1st to 4th grade, and between 4th and 6th grade. There was no significant difference in the incidence of myopia between 4th and 5th grade, 5th and 6th grade. The prevalence of myopia in girls was significantly higher than that in boys(37.6% vs. 30.0%, x²=28.517, p<0.001).
Table 2. Prevalence rate of myopia in different grades and sexes
Grade
|
Male
|
Female
|
Difference between sexes
|
Total (%)
|
Increase rate
%
|
P value
|
Myopic
|
%
|
Myopic
|
%
|
1st
|
42
|
5.1
|
28
|
3.4
|
-1.7%
|
817
|
|
|
2nd
|
78
|
9.6
|
96
|
11.8
|
2.2%
|
812
|
12.9
|
x²=52.891, p<0.001a
|
3rd
|
141
|
17.2
|
127
|
15.5
|
-1.7%
|
819
|
11.3
|
x²=26.327, p<0.001a
|
4th
|
137
|
19.4
|
180
|
25.5
|
6.1%
|
705
|
12.2
|
x²=24.006, p<0.001a
|
5th
|
148
|
21.2
|
196
|
28.0
|
6.9%
|
699
|
4.2
|
x²=2.543, p=0.111a
|
6th
|
138
|
24.7
|
173
|
30.9
|
6.3%
|
559
|
6.4
|
x²=5.133, p=0.023a
|
Total
|
684
|
15.5
|
800
|
18.1
|
|
4411
|
|
|
a. Chi-square test, difference between grades
Refraction and ocular biometric parameters
The mean SERs were 0.31 D, -0.27 D, -0.58 D, -0.63D, -0.83D and -1.15 for participants from 1st to 6th grade respectively (Figure 1). Significant differences (p<0.01) were found between each two grades, except between 4th and 5th grade (p = 0.117). Spearman’s correlation test showed that the AL and AL/CRC were closely associated with SER(AL: r=-0.531, p<0.001; AL/CRC: r=-0.639, p<0.001).
The mean ALs were 22.60, 23.11, 23.54, 23.56, 23.63 and 23.92 mm from 1st to 6th grade respectively. Significant differences (p<0.01) were found between each two grades, except between 3rd and 4th (p = 0.060), 5th and 6th grade (p = 0.268). The mean ratio of AL to CRC were 2.89, 2.96, 2.99, 3.01, 3.03, 3.05 from 1st to 6th grade respectively. Significant differences (p<0.01) were found between each two grades, except between 5th and 6th grade (p = 0.931). The refractive characteristics of the myopic and non-myopic groups were shown in Table 3.
Table 3. Refractive characteristics
Variables
|
Myopic
|
Non-myopic
|
Total
|
P value
(Myopic vs. Non-myopic)
|
AL (mm)
|
23.92
(23.35, 24.61)
|
23.01
(22.50, 23.49)
|
23.27
(22.69, 23.91)
|
<0.001a
|
CRC (mm)
|
7.79 (7.62, 7.96)
|
7.84 (7.67, 8.00)
|
7.82 (7.65, 7.99)
|
0.03 a
|
SER (D)
|
-1.50
(-2.50, -0.88)
|
0.13
(-0.25, 0.50)
|
-0.13
(-1.13, -0.13)
|
<0.001 a
|
ACD (mm)
|
3.77 (3.63, 3.93)
|
3.56 (3.39, 3.74)
|
3.64 (3.45, 3.82)
|
<0.001 a
|
CCT(μm)
|
548.97±32.67
|
550.75±31.22
|
550.14±31.72
|
0.357 b
|
AL/CRC
|
3.07(3.01, 3.14)
|
2.94(2.89, 2.98)
|
2.97(2.91, 3.04)
|
<0.001 a
|
Data are presented as medians and quartiles (p25, p75) or as the means with SD
a. Mann-Whitney U test
b. Independent t-test
There were 384 participants (8.7%) out of 4411 with adequate hyperopia reserve in total. The prevalence of moderate and hyperopia was 0.8% (36/4411). As is shown in Figure 2, the percentage of subjects with adequate hyperopia reserve was low, especially in younger participants. Subjects with age over 12 years old were not included in this figure.
Relationship between body stature and refractive characteristics
The height, weight and BMI were found to be significantly associated with AL (r=0.471, p <0.001; r=0.440, p <0.001; r=0.276, p <0.001, respectively), SER (r=-0.376, p <0.001; r=-0.334, p <0.001; r=-0.189, p <0.001, respectively) and AL/CRC (r=0.445, p <0.001; r=0.406, p <0.001; r=0.240, p <0.001, respectively).
Multiple logistic regression analysis
Univariate analysis showed that educational attainment of parents, time of homework, near visual work, reading and writing posture, time of outdoor activities were not related with incidence of myopia (results shown in Supplementary material 3). Relevant risk factors validated by univariate analysis were listed in Table 4. Multiple logistic regression analysis was then performed. Results showed that parental myopia was associated with a greater prevalence of myopia (Table 4).
Table 4. Analysis of risk factors of myopia by multiple logistic regression
|
N (%)
|
Myopia
|
OR (95%CI)
|
p
|
Age*
|
617
|
1.66 (1.48-1.86)
|
<0.001*
|
BMI
|
617
|
|
0.920
|
Female
|
294 (47.6%)
|
|
0.395
|
Parent myopia*
|
|
|
|
Non
|
302 (48.9%)
|
ref
|
|
One myopic
|
220 (35.7%)
|
1.98 (1.32-2.97)
|
0.001*
|
both
|
95 (15.4%)
|
2.28 (1.36-3.83)
|
0.002*
|
Take a break outdoors*
|
|
|
|
No
|
414 (67.1%)
|
ref
|
|
Yes
|
203 (32.9%)
|
0.67 (0.45-1.00)
|
0.049*
|
PE class
|
|
|
0.675
|
<3
|
66 (10.7%)
|
|
0.725
|
3
|
306 (49.6%)
|
|
0.974
|
4
|
195 (31.6%)
|
|
0.621
|
5
|
53 (8.6%)
|
|
0.585
|
>5
|
26 (4.2%)
|
|
0.939
|
Cram school
|
|
|
0.384
|
No
|
253 (41.0%)
|
|
0.481
|
<1h
|
56 (9.1%)
|
|
0.734
|
1-2h
|
135 (21.9%)
|
|
0.273
|
2-3h
|
73 (11.8%)
|
|
0.961
|
>3h
|
129 (20.9%)
|
|
0.434
|
Electronic screen use
|
|
|
0.445
|
No
|
40 (6.5%)
|
|
0.363
|
<30min
|
203 (32.9%)
|
|
0.638
|
30-60min
|
189 (30.6%)
|
|
0.668
|
1-2h
|
118 (19.1%)
|
|
0.390
|
2-3h
|
38 (6.2%)
|
|
0.094
|
>3h
|
58 (9.4%)
|
|
0.492
|
Data are presented as n (%).
* statistically significant
Other independently associated factors for myopia prevalence were age (p <0.001) and lack of outdoor activity during break between classes (p=0.049). Parental myopia was also associated with longer axial length (regression coefficient 0.100 [95% CI, 0.026-0.174], p=0.008) and greater AL/CRC (regression coefficient 0.012[95% CI, 0.004-0.020], p=0.005) (Table 5).
Multiple logistic regression analysis showed that older age, the female sex, higher BMI, more cram school classes were independently associated with longer axial length. Factors which independently associated with greater AL/CRC were similar, except BMI.
Table 5. Analysis of association between risk factors and AL and AL/CRC by multiple linear regression
|
AL
|
AL/CRC
|
regression coefficient, 95% CI
|
p value
|
regression coefficient, 95% CI
|
p value
|
Age*
|
0.266 [0.224-0.308]
|
<0.001*
|
0.029[0.025-0.034]
|
<0.001*
|
BMI*
|
0.024 [0.001-0.047]
|
0.044*
|
0.002[0.000-0.005]
|
0.102
|
Female*
|
0.582 [0.453-0.710]
|
<0.001*
|
0.021[0.007-0.036]
|
0.004*
|
Parental myopia*
|
0.100 [0.026-0.174]
|
0.008*
|
0.012[0.004-0.020]
|
0.005*
|
Take a break outdoors
|
-0.090 [-0.225-0.046]
|
0.194
|
-0.008[-0.023-0.008]
|
0.318
|
PE class
|
0.016 [-0.054-0.085]
|
0.662
|
0.005[-0.003-0.013]
|
0.229
|
Cram school*
|
0.045 [0.002-0.089]
|
0.042*
|
0.005[0.000-0.010]
|
0.047*
|
Electronic screen use
|
-0.010 [-0.060-0.039]
|
0.682
|
0.000[-0.006-0.006]
|
0.985
|
* statistically significant