Epidemiology of Ophthalmic Diseases

Ophthalmic Res Received: June 21, 2020

Accepted: August 31, 2020
DOI: 10.1159/000511264 Published online: December 4, 2020

Prevalence of Amblyopia and

Associated Risk Factors in Tibetan
Grade One Children
Zhaojun Meng a Jing Fu a Weiwei Chen a, b Lei Li a Han Su a
         

Wei Dai a Yao Yao a
   

aBeijingTongren Eye Center, Beijing Tongren Hospital, Capital Medical University,

Beijing Ophthalmology & Visual Sciences Key Laboratory, Beijing, China; bBeijing Institute of
Ophthalmology, Capital Medical University, Beijing, China

Keywords dents participated in the examinations (97.58% response

Amblyopia · Prevalence · Risk factors · Ophthalmic rate). 1,852 students completed all the related examina-
epidemiology · Clinical trial tions, and 34 of them were diagnosed as amblyopia with the
prevalence of 1.84% (95% confidence interval [CI]: 1.22–
2.45%). Unilateral amblyopia was diagnosed in 23 students
Abstract (1.24%, 95% CI: 0.74–1.75%), including 16 anisometropic, 4
Introduction: Amblyopia is an important public health strabismic, 1 visual deprivational, and 2 mixed. Other 11 stu-
problem and standard screening is quite necessary for early dents were diagnosed as bilateral amblyopia (0.59%, 95%
diagnosis and treatment especially for the remote areas. As CI: 0.24–0.94%), including 9 ametropic, 1 deprivational, and
the place of the largest Tibetan population, the Tibetan Pla- 1 with nystagmus. The mean cylinder refraction and abso-
teau has special geographical characteristics such as high lute value of the spherical equivalent refraction of amblyo-
altitude, time zone, and ethnic composition, where very lit- pic eyes was respectively −2.15 ± 1.52 D and 2.70 ± 2.33 D.
tle information is available about the prevalence of amblyo- Amblyopia was significantly associated with hyperopia
pia and other ocular diseases. The article aims to determine (≥+2.00 D, odds ratio [OR] 8.22, 95% CI 3.42–19.72), astigma-
the prevalence of amblyopia and associated factors in grade tism (≤−2.00 D, OR 6.76, 95% CI 2.56–17.85), and anisome-
1 Tibetan children, living in Lhasa, Tibet Autonomous Re- tropia (≥+0.50 to <+1.00 D, OR 3.95, 95% CI 1.44–10.79;
gion. Methods: A cross-sectional study was conducted. All ≥+1.00 D, OR 21.90, 95% CI 8.24–58.18). Conclusions: The
the participants were scheduled for comprehensive eye ex- prevalence of amblyopia in grade 1 students of Lhasa is rel-
aminations including visual acuity testing, ocular deviation atively higher than that of many other ethnic populations in
and movement evaluation, cycloplegic refraction and ex- China previously reported. Refractive errors including aniso-
aminations of the external eye, anterior segment, media, metropia, hyperopia, and astigmatism are the major risk fac-
and fundus. Amblyopia was assessed in the children accord- tors of amblyopia. © 2020 S. Karger AG, Basel
ing to the standard definition. Results: A total of 1,856 stu-
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

karger@karger.com © 2020 S. Karger AG, Basel Jing Fu

www.karger.com/ore Strabismus and Pediatric Ophthalmology Department
Beijing Tongren Eye Center, Beijing Tongren Hospital
No.1 Dong Jiao Min Xiang St., Dongcheng, Beijing 100730 (China)
Downloaded by:

fu_jing @ 126.com
 Introduction sent was obtained from all of them. The study completed the clinical
trial registration with the number of ChiCTR1900026693.
Lhasa, with an area of 29,518 km2, is one of the main popula-
Amblyopia is an important public health problem be- tion centers of the Tibetan people in Tibet Autonomous Region,
cause its visual impairment can be lifelong and profound China. According to the Chinese census data in 2010, the Tibetan
[1–3]. The latest meta-analysis article estimated 99.2 mil- population of Lhasa is 429,104, accounting for 76.7% of the total
lion people with amblyopia in 2019 worldwide, increas- population of Lhasa. And the 9-year compulsory education is well
ing to 175.2 million by 2030 and 221.9 million by 2040 [4]. implemented in Lhasa, with an enrollment rate of 99.7%. There-
fore, school-based sampling is highly representative for the same-
Understanding the prevalence rate is important for ade- age population.
quate healthcare planning. Substantial studies have been Students who had been living in Lhasa for at least half a year
developed about the prevalence of amblyopia and the re- and would continue to live there for 5 years or more were eligible.
sults vary considerably from 1 to 6% in children [5–9] and Individuals suffering from mental illness or other medical condi-
1.43–5.64% in adults [10, 11]. This range is considered to tions that were unable to cooperate with the baseline survey and
follow-up would be excluded. Stratified cluster sampling was used
be influenced by age, region, ethnicity, diagnosis criteria, in LCES. Based on the evaluation of local government, 27 out of
and other factors. the 28 elementary schools in Lhasa available joining the study were
In China there have been several studies about the stratified into 3 levels. Finally, 1,942 grade 1 students of 7 elemen-
prevalence of amblyopia from discrete regions, such as tary schools were randomly sampled by stratified cluster sampling.
the Anyang Childhood Eye Study (ACES), the Nanjing
Examinations
Pediatric Vision Project (NPVP), and so on which report- The flow diagram of the LCES examination is shown in Figure
ed the rates ranging from 0.82 to 1.47% [12–16]. These 1. Basic systematic examinations were composed of height, weight,
studies mainly focus on the Han, accounting for 91.5% of body circumference, oxygen saturation measured by digital finger-
the whole Chinese population, from the plain area. Nev- tip pulse oximeter (YX301; YUWELL, Jiangsu, China), heart rate
ertheless, very little information is available about the and blood pressure measured by digital automatic blood pressure
monitor (HK-808; HSH, Shenzhen, China).
prevalence of amblyopia in children from Tibetan areas Presenting visual acuity (VA), wearing present correction if
at high altitude. The variations of ethnicity, environment spectacles were carried, was measured monocularly (right fol-
factors such as the climate, light, living habits, and many lowed by left eye) using Lea Symbols 3-m Set charts (250300;
other factors can potentially affect the prevalence. So Goodlite, Elgin, IL, USA) at 3 m. All the subjects were taught to
evaluating the prevalence of amblyopia in Tibetan areas identify the optotypes using the matching card before the exami-
nation. The contralateral eye was occluded with an eye patch for
is of great significance and is also an important supple- monocular measurements. The subjects were asked to read the
ment to other population studies in China and the world. numbers on the right edge of each line (starting from the top line)
The Lhasa Childhood Eye Study (LCES) is conducted until they made a mistake, and then were turned attention to the 2
in the center of Tibetan plateau in Southwest China and lines above the line where they made the initial mistake. The sub-
will last from 2019 to 2024. It aims to research the inci- jects were asked to read each number on successive lines until 3 or
more errors occurred. The last line attempted, combined with the
dence and progression of major ocular diseases affecting number of errors made on it and previous lines, was used to calcu-
Chinese Tibet Plateau children. It is significant to develop late a letter-by-letter logMAR VA score [17].
the study among grade 1 students in view of their sensi- Ocular alignment was assessed using the Hirschberg light reflex
tive period of vision development, the relatively mature test, alternate cover test, and cover-uncover test. Cover tests were
cognitive abilities and the better follow-up. As the base- performed with fixation targets at both distance (6 m) and near (33
cm). The presence of strabismus and its characteristics, type, and
line part of the LCES, the main purpose of current study size were also recorded.
was to determine the prevalence, causes, and associations Objective refraction for every participant was measured before
of amblyopia in grade 1 Tibetan children. and after cycloplegia using an autorefractor (KR-800; Topcon, To-
kyo, Japan). Three readings of sphero-cylindrical autorefraction
were averaged. The cycloplegia was administered with 2 drops of
Methods 1% cyclopentolate (Alcon) and 1 drop of tropicamide/phenyleph-
rine hydrochloride (Santen, Japan) 5 min apart. A third drop of
Study Design and Population cyclopentolate would be administered if pupillary light reflex was
The LCES is a school-based cohort study designed to observe the still present or the pupil size was less than 6.0 mm 30 min later [18].
occurrence and development of different ocular diseases and will Subjective refraction was performed on all participants with VA <
last from 2019 to 2024. The study adhered to the Declaration of Hel- 20/20 in either eye using a trial frame placed and adjusted on the
sinki and obtained Ethics Committee approval from the Institution- participant’s face. Autorefractor readings were used as the starting
al Review Board of Beijing Tongren Hospital, Capital Medical Uni- point and refinements of sphere, cylinder and cylinder axis were
versity (TRECKY2019-146). There was full communication with all performed until the best-corrected distant VA was obtained by
participants’ parents or legal guardians and written informed con- trained optometrists [19].
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

2 Ophthalmic Res Meng/Fu/Chen/Li/Su/Dai/Yao

DOI: 10.1159/000511264
Downloaded by:
 Children registration

Hirschberg light reflex

Vertometer, Heart rate,
test, alternate cover Stereoacuity, lntraocular pressure,
Visual acuity noncycloplegic bloodpressure,
and cover-uncover test, ocular dominance slit-lamp examination
autorefraction Oxygen saturation
ocular movements

Check above records before pupil dilation

Cycloplegia

Anthropometry:height, weight, Digital fundus photography, Cycloplegic autorefraction

body circumference posterior optical coherence tomography and subjective refraction

Quality check of the examination handbooks

Fig. 1. Flow diagram of the LCES examination. VA, visual acuity.

Other eye examination included stereopsis screening (S0001;

STEREO, California, USA), ocular dominance, slit lamp examina- 1,942 sampled
tion (SL-3G; Topcon, Tokyo, Japan), and intraocular pressure 20 lived in Lhasa for
(CT-800; Topcon, Tokyo, Japan). Optical coherence tomography less than one year
and retinal photographs (3D OCT-1; Topcon, Tokyo, Japan) were 15 migrants
taken and assessed to exclude retinal anomalies or pathologies. 4 transferred
The standard operating procedures were produced for staff 1 mental illness
1,902 sampled
training during the study. All examinations were required to be
performed according to procedures. Experienced bilingual staff
were employed to decrease language and cultural barriers. Two 46 refusals and
others
ophthalmologists (FJ and CWW) worked on site to ensure the pro-
cedures were strictly performed. 1,856 attended
Response rate 97.6%
Questionnaire
The questionnaire used in LCES mainly consulted the ACES 4 unable to
[18] and the Sydney Childhood Eye Study (SCES) [20]. In general, complete exam
the questionnaire contained information including birth history,
living and reading habits, history of diagnosis and treatment, and 1,852 analyzed
parents’ information such as education, pregnancy history, and
other potentially required information. The questionnaires were
completed by parents or legal guardians who had been fully in- Fig. 2. Flow chart of the enrollment process in the study.
formed.

Definitions
Unilateral amblyopia was defined as the interocular difference eyes <20/40 (>0.3 logMAR) with bilateral ametropia (≥4.00 D SE
of 2 lines or more with the best-corrected visual acuity (BCVA) hyperopia or ≤−6.00 D SE myopia or ≥2.50 D astigmatism absolute
≤20/32 (≥0.2 logMAR) in the worse eye and met at least one of the value) or with past or present bilateral obstacle of visual axis [21].
following risk factors: strabismus or previous surgery of strabis- Myopia and hyperopia were defined as ≤−0.50 D SE and ≥+2.00
mus, anisometropia (≥1.00 D spherical equivalent (SE) anisohy- D SE, respectively, in 1 or both eyes. Astigmatism was defined as
peropia or ≥3.00 D SE anisomyopia or ≥1.50 D anisoastigmatism), cylindrical refraction absolute value ≥1.00 D. Strabismus was de-
and past or present obstacle of visual axis (e.g., cataract, ptosis, fined as constant or intermittent heterotropia of any magnitude at
corneal opacity). Bilateral amblyopia was defined as BCVA of both distance or near fixation [12].
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

Prevalence and Risk Factors of Amblyopia Ophthalmic Res 3

in Tibetan School Children DOI: 10.1159/000511264
Downloaded by:
 Table 1. Prevalence of amblyopia by different causes in LCES

Amblyopia type Na Prevalencea (95% CI), % Nb Prevalenceb (95% CI), %

Unilateral 23 1.24 (0.74, 1.75) 26 1.40 (0.87, 1.94)

Strabismic 4 0.20 (0.00, 0.43) 4 0.22 (0.00, 0.43)
Anisometropic 16 0.86 (0.44, 1.29) 19 1.03 (0.57, 1.48)
Strabismic + anisometropic 2 0.11 (−0.04, 0.26) 2 0.11 (−0.04, 0.26)
Deprivation 1 0.05 (−0.05, 0.16) 1 0.05 (−0.05, 0.16)
Bilateral 11 0.59 (0.24, 0.94) 17 0.92 (0.48, 1.35)
Ametropic 9 0.49 (0.17, 0.80) 15 0.81 (0.40, 1.22)
Deprivation + ametropic 1 0.05 (−0.05, 0.16) 1 0.05 (−0.05, 0.16)
Other causes (nystagmus) 1 0.05 (−0.05, 0.16) 1 0.05 (−0.05, 0.16)

LCES, Lhasa Childhood Eye Study; CI, confidence interval. a n and prevalence were estimated by the defini-
tion used by MEPEDS (Multiethnic Paediatric Eye Disease Study). b n and prevalence were estimated by the
definition used by PEDIG (Pediatric Eye Disease Investigator Group).

Feedback and Treatment pants was relatively concentrated on 6–8 years. 980 stu-
Every student who completed the examination got a detailed dents (52.92%) were males, and 872 were females
report with diagnostic information and therapic recommenda- (47.08%). 1,758 of the 1,852 students (94.92%) were Ti-
tions in a month, provided by 2 experienced pediatric ophthal-
mologists. betans, while 85 (4.59%) were Han and 9 (0.49%) were of
other ethnicities.
Statistical Analysis
All statistical analyses were performed using SAS statistical Prevalence of Amblyopia
software (version 9.4, SAS Inc., Cary, NC, USA) and two-sided Thirty-four of 1,852 students met the definition of am-
p < 0.05 was considered statistically significant. Values were pre-
sented as mean ± standard deviation for continuous variables and blyopia (1.84%, 95% CI: 1.22–2.45%), and another 3 stu-
percentage for categorical variables. Prevalence was calculated as dents were suspected. All of the 34 definite students with
the percent of participants with amblyopia among all the partici- amblyopia were Tibetans and included 19 girls and 15
pants evaluated. Difference of gender and prematurity between boys. There was no significant difference in age (t = −0.89,
2 groups was compared using χ2 test. And difference in age, BMI, p = 0.38), gender (χ2 = 1.08, p = 0.30), premature (χ2 =
and oxygen saturation between 2 groups was compared using
t test. Multivariate logistic regression models were used to evaluate 1.05, p = 0.31), BMI(t = −0.22, p = 0.83), and oxygen sat-
the effect of each refractive risk factor while adjusting for each uration (t = −1.56, p = 0.13) between students with or
other. All of the confidence intervals (CIs) are given as 95% CIs. without amblyopia. The mean BCVA of amblyopic eyes
was 0.54 ± 0.22 logMAR ranging from 0.2 to 1.0 logMAR.
As shown in Table 1, by the definition used in the ME-
Results PEDS (Multiethnic Pediatric Eye Disease Study) [21],
unilateral amblyopia was diagnosed in 23 children (1.24%,
Subject 95% CI: 0.74–1.75%), and bilateral amblyopia was diag-
Forty of 1,942 students did not meet the inclusion cri- nosed in 11 students (0.59%, 95% CI 0.24–0.94%). Of 11
teria. 1,856 of the eligible 1,902 individuals participated children with bilateral amblyopia, 9 were ametropic, 1
in the examinations (97.58% response rate) and 4 of them had the deprivation factor induced by ptosis in the right
failed to complete the autorefraction examinations. Fi- eye and ametropic factor in the left eye, and 1 had con-
nally 1,852 students completed all the amblyopia-related genital nystagmus without any other eye disease detected.
examinations. The flow chart of enrollment was shown in Twenty-three children with unilateral amblyopia con-
Figure 2. The mean age of participants was 6.83 ± 0.46 tained 16 anisometropic amblyopia, 4 strabismic amblyo-
years (ranging from 5.89 to 10.32 years). Only 2 children pia, 1 stimulus deprivation amblyopia caused by congen-
were <6 years old. There were 1,271 and 555 children aged ital cataract, and 2 mixed amblyopia including strabismus
6–7 and 7–8 respectively, accounting for 98.60% of the and anisometropia.
total. Nineteen children were between the age of 8–9, In addition, 3 children without strabismus and depri-
while 5 above the age of 9. Therefore, the age of partici- vation factor met the VA criteria for unilateral amblyopia,
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

4 Ophthalmic Res Meng/Fu/Chen/Li/Su/Dai/Yao

DOI: 10.1159/000511264
Downloaded by:
 Color version available online
50
46.92 ■ Amblyopic eyes
■ Nonamblyopic eyes

39.93
40

30

Percent, %
20.5 20.59
20

14.7
11.76

10 8.82
5.88 5.88 5.88
5.01 5.39
2.94
0.22 0.5 0.99 0.55 0.28
0.050.06 0.11 0.06
0
≤–6 –5 to –4 to –3 to –2 to –1 to 0 to 1 to 2 to 3 to 4 to 5 to >6
Fig. 3. Distribution of SER in amblyopic
eyes and nonamblyopic right eyes. SER, SER, diopters
spherical equivalent refraction.

Color version available online

■ Amblyopic eyes
53.58 ■ Nonamblyopic eyes

50

40
Percent, %

30
24.48 23.53
20.59 20.59
20 17.65
14.19
11.76
10
5.88
4.51
1.76 0.99 0.5
0
0 to –0.5 to –1.5 to –2.5 to –3.5 to –4.5 <–4.5
Fig. 4. Distribution of cylinder refraction in
Cylinder, diopters
amblyopic eyes and nonamblyopic right
eyes.

but the difference in astigmatism between 2 eyes was 1.25 different criteria, the prevalence based on the PEDIG (Pe-
D, missing the critical value 1.5 D in the diagnosis criteria diatric Eye Disease Investigator Group) criteria and ME-
of amblyopia. If these 3 children were considered as am- PEDS was presented in Table 1. PEDIG defined the uni-
blyopic, the prevalence of amblyopia increased to 1.99%. lateral amblyopia as interocular difference of 2 lines or
The definition of amblyopia varied among different more and bilateal amblyopia as BCVA of both eyes ≤20/40
studies. In order to better explain the prevalence under (≥0.3 logMAR) [22], and according to the above defini-
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

Prevalence and Risk Factors of Amblyopia Ophthalmic Res 5

in Tibetan School Children DOI: 10.1159/000511264
Downloaded by:
 Table 2. Multivariate analysis for refractive factors of amblyopiaa

Refractive errors N Amblyopia, N (%) OR (95% CI) p value

SE refractive error, D
≤−0.50 75 7 (9.33) 2.55 (0.83–7.90) 0.1039
>−0.50 and <+2.00 1,615 11 (0.68) Ref
≥+2.00 162 16 (9.88) 8.22 (3.42–19.72) <0.0001
Astigmatism, D
>−1 and ≤0 1,467 9 (0.61) Ref
>−2 and ≤−1 256 8 (3.13) 2.21 (0.76–6.45) 0.1470
≤−2 129 17 (13.18) 6.76 (2.56–17.85) 0.0001
SE anisometropia, D
≥0 to <+0.50 1,631 10 (0.61) Ref
≥+0.50 to <+1.00 171 9 (5.26) 3.95 (1.44–10.79) 0.0075
≥+1.00 50 15 (30.00) 21.90 (8.24–58.18) <0.0001

OR, odds ratio; CI, confidence interval; SE, spherical equivalent. a Adjusted for all the other refractive risk
factors in a multivariate logistic regression model. The more ametropic eyes in unilateral amblyopia and less am-
etropic eyes in bilateral amblyopia were analyzed.

tion, a total of 43 children were diagnosed as amblyopia The association between refractive errors and preva-
with the prevalence of 2.32% (95% CI: 1.64–3.01%). Uni- lence of amblyopia was shown in Table 2 by multivariate
lateral amblyopia was diagnosed in 26 children (1.40%, logistic regression analysis. Among 34 amblyopic stu-
95% CI: 0.87–1.94%), including 19 anisometropic, 4 stra- dents, 7 were diagnosed with myopia, 16 with hypero-
bismic, 1 stimulus deprivation caused by congenital cata- pia, 25 with astigmatism, and 24 with anisometropia of
ract, and 2 mixed amblyopia containing strabismus and SE equal to or greater than 0.5 D. Amblyopia was associ-
anisometropia. Bilateral amblyopia was diagnosed in 17 ated with hyperopia (≥+2.00 D, OR 8.22, 95% CI 3.42–
children (0.92%, 95% CI 0.48–1.35%) including 15 ame- 19.72), astigmatism (≤−2.00 D, OR 6.76, 95% CI 2.56–
tropic, 1 mixed amblyopia with the deprivation factor in- 17.85), and anisometropia (≥+0.50 to <+1.00 D, OR
duced by ptosis in the right eye and ametropic factor in 3.95, 95% CI 1.44–10.79; ≥+1.00 D, OR 21.90, 95% CI
the left eye, and 1 congenital nystagmus without any oth- 8.24–58.18).
er eye disease detected. According to the American Association of Pediatric
Ophthalmology (AAPOS)–specified threshold levels [23,
Refractive Error and Amblyopia 24], the prevalence of the refractive risk factors was shown
The average of the absolute value of the SE refraction and compared with that in MEPEDS in Table 3. The prev-
of amblyopic eyes was 2.70 ± 2.33 D (ranging from −9.00 alence of anisometropia >1.50 D, hyperopia >3.50 D,
to +6.75 D) compared to 1.17 ± 0.65 D (ranging from astigmatism >1.50 D, and myopia >3.00 D in the LCES
−4.375 to +8.25 D) in nonamblyopic eyes (t = −3.82, p = was respectively 1.13, 1.24, 11.50, and 0.43%, compared
0.0006). The mean cylinder refraction of amblyopic eyes with 2.0, 8.4, 10.4, and 1.0% reported in the MEPEDS
was −2.15 ± 1.52 D (ranging from −5.75 to 0) compared [25].
to −0.63 ± 0.74 D (ranging from −6.5 to 0 D) in nonam-
blyopic eyes (t = −5.80, p < 0.0001). Figures 3 and 4 illus- Strabismus and Amblyopia
trated the SE and cylinder refraction distribution of the There were a total of 68 children diagnosed as manifest
amblyopic and nonamblyopic eyes. Figure 3 showed SE strabismus in the study with the prevalence of 3.7%. Four-
of nonamblyopic eyes mainly concentrated from −1.00 to ty-five of the 68 strabismic children were diagnosed as
+2.00 D, while SE of the amblyopic eyes was relatively intermittent exotropia and other 23 were diagnosed as
dispersed from ≤−6.00 D to ≥+6.00 D. The cylinder re- constant strabismus. Of the 6 strabismic amblyopic chil-
fraction of amblyopic eyes concentrated mainly from the dren including 2 mixed with anisometropia, 4 were diag-
moderate to high degree and nonamblyopic eyes mainly nosed as constant esotropia with the mean deviation an-
concentrated from mild to moderate degree. gle of 42.50 ± 8.29 prism diopters, 1 was diagnosed as
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

6 Ophthalmic Res Meng/Fu/Chen/Li/Su/Dai/Yao

DOI: 10.1159/000511264
Downloaded by:
Table 3. The prevalence of the refractive
risk factors in LCES and MEPEDS by Risk factor AAPOS LCES MEPEDS,
AAPOS levels %
criteria, D estimated prev-
alence, %

Anisometropia >1.50 1.2 1.13% (21) 2.0

Hyperopia >3.50 6 1.24% (23) 8.4
Astigmatism >1.50 9 11.50% (213) 10.4
Myopia >3.00 0.5 0.43% (8) 1.0

AAPOS, the American Association of Pediatric Ophthalmology [23, 24]; MEPEDS,

Multi-ethnic Paediatric Eye Disease Study [25]; LCES, Lhasa Childhood Eye Study.

micro-esotropia with the deviation angle of 10 prism di- ocular diseases for Tibetan children. LCES is also an im-
opters and 1 was diagnosed as constant exotropia with the portant complement to other population studies in Chi-
deviation angle of 40 prism diopters. na, as well as the world because of the distinct environ-
ment, living habits, and cultural characteristics in Tibet.
History and Treatment of Amblyopia In the present study the prevalence of amblyopia was
Six students who met our criteria of amblyopia had a 1.84%, relatively higher compared to most other studies
previous diagnosis and other 28 were newly diagnosed in Asia. Two meta-analysis articles estimated the pooled
during LCES. Four of the six students had been treated global prevalence rate of amblyopia was 1.44 and 1.75%
with spectacles or monocular occlusion therapy reported respectively [4, 26]. Geographically, amblyopia was more
by their parents and other 2 students did not have any prevalent in Europe (2.90%, 95% CI 2.20–3.83%) and
therapy ever. North America (2.41%, 95% CI 1.18–4.88%) than in Asia
In addition, 8 students with related risk factors had (1.09%, 95% CI 0.89–1.33%) and Africa (0.72%, 95% CI
previous diagnosis and treatment of amblyopia reported 0.26–1.98%) [4]. The prevalence is influenced by factors
by their parents. But they were not shown to be amblyopic of age, region, environment, ethnicity, diagnosis criteria,
in this study by our criteria. All of the 8 students had sig- and so on. Different definitions can have significant im-
nificant isoametropia which consisted of 2 astigmatism, pact on amblyopia prevalence estimate. As reported,
2 hyperopia, and 4 astigmatism-mixed hyperopia. If the prevalence was highest when a BCVA ≥ 2 line difference
8 students had not received treatment, they would most criteria was used (2.99%, 95% CI 1.89–4.68%) and lowest
likely been amblyopic. And the prevalence of amblyopia when the BCVA ≤ 20/40 criteria was used (0.76%, 95% CI
would increase to 2.27%. Of all the 45 patients including 0.51–1.14%) [4]. We compared the prevalence of amblyo-
the 34 newly diagnosed, 8 ever diagnosed as amblyopia pia estimated by different definitions used by the ME-
and 3 suspected cases who met the sub-astigmatism PEDS and PEDIG. According to the PEDIG criteria, the
threshold, 12 were wearing spectacles. prevalence of either unilateral or bilateral amblyopia was
higher than that used in MEPEDS. The unilateral ambly-
opic children diagnosed by the PEDIG criteria included
Discussion 3 sub-astigmatism threshold cases who were not diag-
nosed by the MEPEDS criteria. And the bilateral amblyo-
LCES is a school-based cohort study to investigate oc- pic children diagnosed by the PEDIG criteria included 6
ular diseases of school-age children in the Tibetan plateau cases with the VA equalled to 20/40 which did not meet
areas in China where few studies related had been devel- the criteria by MEPEDS (VA < 20/40). Therefore, minor
oped. It is well established and characterized with high difference in the definition would impact on the preva-
response rate (97.58%) and high cycloplegic percentage lence. In order to be comparable we summarized several
(99.84%) in the participants. Lhasa, accounting for more previous studies in Table 4 with the same standard am-
than 1/5 population of Tibet, is the political, cultural, eco- blyopia definition from MEPEDS [21, 27]. Compared
nomic, and educational center of Tibet. It is of represen- with other studies on children close in age and of different
tative significance to observe the incidence and factors nationalities, the prevalence in the Tibetan plateau region
associated with the etiology of amblyopia as well as other was relatively higher [12, 15]. Pi et al. [28] reported a
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

Prevalence and Risk Factors of Amblyopia Ophthalmic Res 7

in Tibetan School Children DOI: 10.1159/000511264
Downloaded by:
Table 4. Prevalence of amblyopia in different studies using the MEPEDS criteria

Study Country Year N Age Ethnicity Prevalence, %

MEPEDS USA 2008 3,350 30–72 months African American (1,663) 1.50
Hispanic/Latino (1,687) 2.60
MEPEDS USA 2013 1,883 30–72 months Asian American (938) 1.81
Non-Hispanic White (945) 1.81
BPEDS USA 2009 1,546 30–71 months Non-Hispanic White (673) 1.80
African American (873) 0.80
STARS Singapore 2010 1,682 30–72 months Singaporean Chinese 1.19
SPEDS Australia 2012 1,422 30–72 months Multiethnic 1.90
(Predominantly White)
ACES China 2014 2,893 6–9 years Chinese Han 1.00
NPVP China 2016 5,667 36–72 months Chinese Han 1.20
YPEDS China 2018 1,695 36–48 months Chinese Han 1.47
MMPS China 2018 3,050 7–8 years (1,656) Chinese Hani 0.82
13–14 years (1,394)
HCES China 2019 7,713 30–83 months Chinese Han 1.09
LCES China 2019 1,852 5–10 years Chinese Tibetan 1.84

MEPEDS, Multi-ethnic Paediatric Eye Disease Study [21, 26]; BPEDS, Baltimore Paediatric Eye Disease Study [7]; STARS, Strabis-
mus, Amblyopia, and Refractive Error in Singaporean Children Study [28]; SPEDS, Sydney Paediatric Eye Disease Study [6]; ACES,
Anyang Childhood Eye Study [12]; NPVP, Nanjing Paediatric Vision Project; YPEDS, Yuhuatai Paediatric Eye Disease Study [13];
MMPS, Mojiang Myopia Progression Study [15]; HCES, Hunan Childhood Eye Study [16]; LCES, Lhasa Childhood Eye Study.

higher prevalence rate of 1.88% in Western China than situation of poor vision. The rate of diagnosis and treat-
our study. But the diagnosis criteria of amblyopia was dif- ment in LCES(17.65, 11.76%) was lower compared to
ferent from ours which defined the BCVA as ≤20/25 or ACES which was also developed during students with
worse. close age in plain area of China (22.22, 14.81%) [12]. Be-
The higher prevalence in our study might be related to sides, 18 students had amblyopia history reported by par-
distinct factors of ethnicity, environment, and econom- ents in questionnaire but did not meet the criteria of di-
ics. Participants of our study were most native Tibetan agnosis in our examination. Among the 18 students 10
(94.92%) who lived on the high altitude area of 3,650 m might be misdiagnosed because they had normal VA
where the diet habit and environment such as climate, without any amblyopic risk factors and treatment history.
light, and pressure were obviously different from plateau. And other 8 students who had related risk factors and
All of above could lead to the physical difference besides previous therapy history did not meet the VA criteria.
the ethnic factor. In economically underdeveloped areas Therefore, standard screening and intervention as well as
the popularity of vision screening is relatively low. The the improvement of parents understanding of amblyopia
risk factors of amblyopia cannot be detected in time at remote areas are necessary for increasing cure rate and
which could lead to delay in the early diagnosis and treat- reducing unnecessary time and economic loss. All of the
ment of amblyopia. Lhasa is relatively remote, and health- students with amblyopia got the feedback of detailed di-
care is not quite universal. Only 6 of 34 students with agnostic information and therapic recommendations af-
amblyopia had a previous diagnosis. Among the 6 stu- ter the examination, provided by 2 experienced pediatric
dents, only 4 accepted treatments with spectacles or mon- ophthalmologists in the group.
ocular occlusion therapy. And another 2 students did not In this study, gender was not significantly associated
get any therapy ever even when their parents knew their with amblyopia prevalence which was in agreement with
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

8 Ophthalmic Res Meng/Fu/Chen/Li/Su/Dai/Yao

DOI: 10.1159/000511264
Downloaded by:
a number of previous studies [29, 30]. Because the age of Second, amblyopia history and treatment in question-
participants in the study was relatively concentrated, it naires were self-reported from the parents, which might
was difficult to identity that age was associated with the be potentially inaccurate. In the follow-up study, we will
prevalence or not. Besides, there was no significant differ- further improve the quality of the questionnaires.
ence in prematurity, BMI, and oxygen saturation between In conclusion the study is representative to observe the
students with or without amblyopia. incidence and factors associated with the etiology of am-
The study showed anisometropia and ametropia were blyopia for Tibetan children. We find the prevalence of
major causes to the prevalence of amblyopia. Eighteen amblyopia is 1.84% in grade 1 students in Lhasa which is
students with anisometropic amblyopia (52.94%) con- higher than that of many other ethnic populations in Chi-
sisted of 11 anisoastigmatism, 3 anisohyperopia, 1 aniso- na previously reported. Anisometropia, hyperopia, and
myopia, and 3 mixed anisometropia (2 anisoastigma- astigmatism are the major risk factors of amblyopia.
tism-mixed anisohyperopia and 1 anisoastigmatism-
mixed anisomyopia) according to the diagnosis criteria.
Some studies reported anisometropia of 1.0 D or more Acknowledgements
was associated significantly with increased odds of am-
We are grateful for the support from the Lhasa government and
blyopia [6, 14]. Our study demonstrated anisometropia
Beijing Hoson Foundation for helping organizing the survey.
of 0.5–1.0 D was also associated significantly with the in-
creased risk of amblyopia which was consistent with oth-
er previous report [31]. Refractive errors which included Statement of Ethics
hyperopia of 2.00 D or more and astigmatism of −2.00 D
or less were also significant risk factors for amblyopia. The study adhered to the Declaration of Helsinki and obtained
Myopia was not identified as significantly associated with Ethics Committee approval from the Institutional Review Board
increased odds of amblyopia in our study. It might be be- of Beijing Tongren Hospital, Capital Medical University
(TRECKY2019-146). The written informed consent was obtained
cause myopia was less common in younger children from the parents or guardians.
(4.05%).
According to the AAPOS-specified threshold levels,
the prevalence of astigmatism >1.50 D in the LCES Conflict of Interest Statement
(11.5%) was higher than that in the MEPEDS (10.4%).
And the prevalence of anisometropia >1.50 D (1.13%), The authors have no conflicts of interest to declare.
myopia >3.00 D (0.43%), and hyperopia >3.50 D (1.24%)
was lower than that in the MEPEDS (2.0, 1.0, 8.4% respec-
tively). The significant difference for the prevalence of hy- Funding Sources
peropia >3.50 D between the 2 studies might be due to the
different age of participants which was mainly concen- The study was supported by High-Level Health Technical Tal-
ent Training Program of Beijing Municipal Health Bureau under
trated on 6–8 years in LCES and 36–60 months in ME- Grant (2015-3-023), Beijing Municipal Science & Technology
PEDS. Commission under Grant (Z171100001017066), and Capital
The proportion of strabismic amblyopia was relatively Health Development Special Fund of China–Major Project under
less in LCES compared with some studies [9], although Grant (SF-2018-1-2051).
the prevalence of manifest strabismus was higher than
that reported in MEPEDS [25]. The reason might be the
strabismus detected in LCES was most intermittent which Author Contributions
usually had minimal impact on VA. And esotropia stra-
J.F. and W.W.C. initiated the study design. Z.J.M., W.W.C.,
bismus was also an important cause for amblyopia in Y.Y., W.D., and H.S. prepared the consent form. All authors are
LCES. involved in the research process and data collation. L.L. assisted in
Our study also had some limitations. First, the study collation and analysis of the data. Z.J.M. and J.F. drafted and final-
was school based rather than population based. This ized the study protocol.
could have biased our findings either towards overesti-
mation or underestimation. But the high enrollment rate
for primary school-age children in Lhasa (99.7%) and the
high response rate in LCES (97.58%) reduced the bias.
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

Prevalence and Risk Factors of Amblyopia Ophthalmic Res 9

in Tibetan School Children DOI: 10.1159/000511264
Downloaded by:
 References
  1 Carlton J, Karnon J, Czoski-Murray C, Smith 12 Fu J, Li SM, Li SY, Li JL, Li H, Zhu BD, et al. 21 Multi-Ethnic Pediatric Eye Disease Study
KJ, Marr J. The clinical effectiveness and cost- Prevalence, causes and associations of ambly- Group. Prevalence of amblyopia and strabis-
effectiveness of screening programmes for opia in year 1 students in Central China: the mus in African American and Hispanic chil-
amblyopia and strabismus in children up to Anyang childhood eye study (ACES). Graefes dren ages 6 to 72 months the multi-ethnic pe-
the age of 4–5 years: a systematic review and Arch Clin Exp Ophthalmol. 2014 Jan;252(1): diatric eye disease study. Ophthalmology.
economic evaluation. Health Technol Assess. 137–43. 2008 Jul;115(7):1229–36.e1.
2008 Jun;12(25):iii, 13 Chen X, Fu Z, Yu J, Ding H, Bai J, Chen J, et 22 Wallace DK, Repka MX, Lee KA, Melia M,
 2 Carlton J, Kaltenthaler E. Amblyopia and al. Prevalence of amblyopia and strabismus in Christiansen SP, Morse CL, et al. Amblyopia
quality of life: a systematic review. Eye. 2011 Eastern China: results from screening of pre- preferred practice pattern®. Ophthalmology.
Apr;25(4):403–13. school children aged 36-72 months. Br J Oph- 2018 Jan;125(1):P105–42.
  3 Felius J, Chandler DL, Holmes JM, Chu RH, thalmol. 2016 Apr;100(4):515–9. 23 Donahue S, Arnold R, Ruben JB. Preschool
Cole SR, Hill M, et al. Evaluating the burden 14 Huang D, Chen X, Zhu H, Ding H, Bai J, Chen vision screening: what should we be detecting
of amblyopia treatment from the parent and J, et al. Prevalence of amblyopia and its asso- and how should we report it? Uniform guide-
child’s perspective. J AAPOS. 2010 Oct;14(5): ciation with refraction in Chinese preschool lines for reporting results from studies of pre-
389–95. children aged 36-48 months. Br J Ophthal- school vision screening. J AAPOS. 2003 Oct;
  4 Fu Z, Hong H, Su Z, Lou B, Pan CW, Liu H. mol. 2018 Jun;102(6):767–71. 7(5):314–6.
Global prevalence of amblyopia and disease 15 Zhu H, Pan C, Sun Q, Huang D, Fu Z, Wang 24 Arnold RW. Amblyopia risk factor preva-
burden projections through 2040: a system- J, et al. Prevalence of amblyopia and strabis- lence. J Pediatr Ophthalmol Strabismus. 2013
atic review and meta-analysis. Br J Ophthal- mus in Hani school children in rural south- Jul–Aug;50(4):213–7.
mol. 2020 Aug;104(8):1164–70. west China: a cross-sectional study. BMJ 25 Varma R. Amblyopia refractive risk factors
  5 Attebo K, Mitchell P, Cumming R, Smith W, Open. 2019 Feb;9(2):e025441. (letter reply). Ophthalmol. 2012 Jun; 119(6):
Jolly N, Sparkes R. Prevalence and causes of 16 Li YP, Zhou MW, Forster SH, Chen SY, Qi X, 1283–4.
amblyopia in an adult population. Ophthal- Zhang HM, et al. Prevalence of amblyopia 26 Hashemi H, MSc RP, Yekta A, Bostamzad P,
mology. 1998 Jan;105(1):154–9. among preschool children in central south Aghamirsalim M, MSc SS, et al. Global and
  6 Pai AS, Rose KA, Leone JF, Sharbini S, Bur- China. Int J Ophthalmol. 2019 May; 12(5): regional estimates of prevalence of amblyo-
lutsky G, Varma R, et al. Amblyopia preva- 820–5. pia: a systematic review and meta-analysis.
lence and risk factors in Australian preschool 17 Chen W, Fu J, Meng Z, Li L, Su H, Dai W, et Strabismus. 2018 Dec;30:1–16.
children. Ophthalmology. 2012 Jan; 119(1): al. Lhasa childhood eye study: the rationale, 27 McKean-Cowdin R, Cotter SA, Tarczy-Hor-
138–44. methodology, and baseline data of a 5 year noch K, Wen G, Kim J, Borchert M, et al.
  7 Friedman DS, Repka MX, Katz J, Giordano L, follow-up of school-based cohort study in the Prevalence of amblyopia or strabismus in
Ibironke J, Hawse P, et al. Prevalence of am- Tibetan plateau region of Southwest China. asian and non-Hispanic white preschool chil-
blyopia and strabismus in white and African BMC Ophthalmol. 2020 Jun;20(1):250. dren: multi-ethnic pediatric eye disease study.
American children aged 6 through 71 months 18 Li SM, Liu LR, Li SY, Ji YZ, Fu J, Wang Y, et Ophthalmology. 2013 Oct;120(10):2117–24.
the Baltimore pediatric eye disease study. al. Design, methodology and baseline data of 28 Pi LH, Chen L, Liu Q, Ke N, Fang J, Zhang S,
Ophthalmology. 2009 Nov;116(11):2128–2. a school-based cohort study in Central China: et al. Prevalence of eye diseases and causes of
  8 Robaei D, Kifley A, Rose KA, Mitchell P. Im- the Anyang childhood eye study. Ophthalmic visual impairment in school-aged children in
pact of amblyopia on vision at age 12 years: Epidemiol. 2013 Dec;20(6):348–59. Western China. J Epidemiol. 2012 Nov;22(1):
findings from a population-based study. Eye. 19 Yuan BL, Friedman DS, Wong TY, Wang FH, 37–44.
2008 Apr;22(4):496–502. Duan XR, Xiao HY, et al. Rationale, design, 29 Chia A, Dirani M, Chan YH, Gazzard G, Au
  9 Robaei D, Rose KA, Ojaimi E, Kifley A, Mar- methodology, and baseline data of a popula- Eong KG, Selvaraj P, et al. Prevalence of am-
tin FJ, Mitchell P. Causes and associations of tion-based study in rural China: the Handan blyopia and strabismus in young singaporean
amblyopia in a population-based sample of eye study. Ophthalmic Epidemiol. 2009 Mar– chinese children. Invest Ophthalmol Vis Sci.
6-year-old Australian children. Arch Oph- Apr;16(2):115–27. 2010 Jul;51(7):3411–7.
thalmol. 2006 Jun;124(6):878–84. 20 Ojaimi E, Rose KA, Smith W, Morgan IG, 30 Mocanu V, Horhat R. Prevalence and risk fac-
10 Elflein HM, Fresenius S, Lamparter J, Pitz S, Martin FJ, Mitchell P. Methods for a popula- tors of amblyopia among refractive errors in
Pfeiffer N, Binder H, et al. The prevalence of tion-based study of myopia and other eye an Eastern European population. Medicina.
amblyopia in Germany: data from the pro- conditions in school children: the Sydney my- 2018 Mar;54(1):6.
spective, population-based Gutenberg health opia study. Ophthalmic Epidemiol. 2005 Feb; 31 Tarczy-Hornoch K, Varma R, Cotter SA,
study. Dtsch Arztebl Int. 2015 May; 112(19): 12(1):59–69. McKean-Cowdin R, Lin JH, Borchert MS, et
338–44. al. Risk factors for decreased visual acuity in
11 Høeg TB, Moldow B, Ellervik C, Klemp K, preschool children: the multi-ethnic pediatric
Erngaard D, la Cour M, et al. Danish rural eye eye disease and Baltimore pediatric eye dis-
study: the association of preschool vision ease studies. Ophthalmology. 2011 Nov;
screening with the prevalence of amblyopia. 118(11):2262–73.
Acta Ophthalmol. 2015 Jun;93(4):322–9.
156.62.3.11 - 12/14/2020 3:35:24 PM
Auckland University of Technology

10 Ophthalmic Res Meng/Fu/Chen/Li/Su/Dai/Yao

DOI: 10.1159/000511264
Downloaded by: