Investigation of myopic open-angle glaucoma (OAG) prevalence in Northeast Asia by systematic review and meta-analysis.
Systematic PubMed, Embase and Cochrane database searches for Northeast Asian population-based studies published up to 30 November 2020 and reporting on myopia and OAG diagnosis. By random-effect models, pooled OAG prevalence in a myopic population and pooled myopic OAG prevalence in a general population were generated, with 95% confidence intervals (CIs).
The meta-analysis encompassed five population-based studies in four countries (12,830 individuals, including 7,723 patients with myopia and 1,112 patients with OAG). In a myopic population, OAG prevalence was 4.10% (95% CI, 3.00–5.70; I2 = 93%); in a general population, myopic OAG prevalence was 1.10% (95% CI, 0.60–1.70; I2 = 94%). A visual examination of funnel plot symmetry raised a suspicion of publication bias. Notwithstanding, Begg and Mazumbar’s adjusted rank correlation test showed no such evidence (
Our systematic review and meta-analysis returned an estimate of OAG prevalence in a myopic Northeast Asian population. Our findings will inform future glaucoma studies as well as public health guidelines for Northeast Asian populations.
Myopia is an increasingly concerning public health issue, especially in East Asia, where it is already extremely widespread [
Glaucoma is the second leading cause of blindness globally, open-angle glaucoma (OAG) being its most common manifestation [
However, whereas the respective prevalence of myopia and glaucoma are well-known, little is confirmed as to the prevalence of their combination. One study reported that 17.2% of Italian glaucoma patients were myopic [
The systematic review and meta-analysis were performed based on a pre-specified protocol, and their methods adhered to both the MOOSE (Meta-analyses Of Observational Studies in Epidemiology) [
We systematically searched for relevant studies in the PubMed, Embase, and Cochrane Library databases. Our strategies were developed with the assistance of an academic librarian with expertise in systematic review and were based on established terminology using MeSH and Embase search terms whenever available. The keywords included ‘glaucoma,’ ‘open-angle glaucoma,’ ‘myopia,’ ‘refractive error,’ and ‘prevalence.’ All of the search details are included in
In the systematic review, the study inclusion criteria were as follows: (1) population-based and (2) evaluation of both OAG and myopia prevalence. The exclusion criteria were as follows: (1) not conducted with humans or adults, (2) narrative and/or systematic review, commentary, case report, (3) involving either secondary glaucoma or angle-closure glaucoma, (4) not conducted in Northeast Asia, (5) published in language other than English, and (6) lacking any detailed definition of OAG.
Two investigators (YJ and AH) extracted data in both an independent and masked fashion using a standardized data extraction method based on those employed by the Cochrane Library’s Database of Systematic Reviews (The Cochrane Collaboration: Review Manager 4.1.1. Nepean, Ontario, Canada). Extracted data following entry into a dedicated database were rechecked by a third investigator (YKK). The following data were extracted: first author’s name, publication year, race/ethnicity of population, study country, subject number, subjects’ age and sex, OAG-diagnostic criteria, myopia definition followed, OAG patient number, myopia patient number, and number of patients with both myopia and OAG.
To assess the methodological quality of the studies, the Newcastle-Ottawa Scale for assessment of comparative nonrandomized study quality was applied (
Because of the anticipated high levels of heterogeneity, we used a random effect model to estimate the pooled OAG prevalence in a myopic population and the pooled myopic OAG prevalence in a general population. Inter-study heterogeneity was quantified by the I2 statistic representing the interstudy variation prevalence that can be attributed to heterogeneity rather than sampling error [
To qualitatively synthesize myopia degree, the prevalence of myopic OAG, and percentage of OAG among myopia, we stratified myopia into low, moderate, moderate-to-high, and high degree categories, as based on spherical equivalent (SE) refractive error up to −3 diopters (D), between −6 D and −3 D (or −4 D), lower than −3 D (or −4 D), and lower than −6 D, respectively.
Publication bias was evaluated in two ways: (1) qualitatively by funnel plot [
The data handling and the statistical analyses were under the purview of a single investigator (YKK) supervised by a statistician with expertise in meta-analysis (SRS). All of the 95% confidence intervals (CIs) and
Our systematic search identified 1,014 articles, 765 from PubMed and 249 from Embase (not in PubMed), among which, 79 were full-text reviewed. Following a thorough review, the final analysis proceeded with five studies (
Five population-based studies involving a total of 26,661 individuals had been conducted in four countries: two in China [
In the South Korean study, the number of OAG patients was 710, the highest among the studies. In the Chinese, Japanese, and Singaporean (Malay population) studies, the OAG patient numbers ranged between 72 and 119. The number of myopic patients, like the number of OAG patients, was highest in South Korea (4,039), followed by Beijing, China (1,837). The myopic patient numbers in the three remaining studies ranged between 239 and 877.
In the Northeast Asian population as a whole, the estimated prevalence of myopia ranged from 9.5% to 42.5%, and that of OAG ranged from 2.8% to 5.1% (
The overall pooled prevalence of OAG among patients with myopia was 4.10% (95% CI, 3.00%–5.70%). There was significant heterogeneity of OAG prevalence in myopia (I2 = 90%,
Four studies (except He et al. [
Myopia has become a major health concern in East Asia [
To the best of our knowledge, this is the first meta-analysis of myopic OAG prevalence and OAG prevalence in myopic Asian populations. This study involved a total of 12,830 subjects, among whom were 7,723 patients with myopia and 1,112 with OAG. The pooled OAG prevalence among those with myopia was 4.1%, and the pooled myopic OAG prevalence in the general population was 1.1%.
Within Northeast Asian populations, there has been significant variability in the prevalence of myopic OAG and the proportion of OAG patients among myopia sufferers. Our current results showed that the prevalence of myopic OAG and OAG among myopic patients were the highest in the South Korean population (1.8% and 6.3%, respectively). As for the other countries, the prevalence of myopic OAG were 1.5% in Japan, 0.5% to 1.2% in China, and 0.7% in Singapore (Malay population); the prevalence of OAG among myopia patients, meanwhile, were 2.7% to 5.0% in China, 4.9% in Japan, and 2.9% in Singapore (Malay population).
The percentage of OAG among patients with low myopia was 2.24% to 4.37%, while that with high myopia was 6.07% to 8.68%. Pan et al. [
There is no single conclusion about the influence of ethnicity on refractive error and glaucoma. The effect of refractive error on the risk of primary open-angle glaucoma (POAG) was modestly but significantly stronger in non-Hispanic whites, and its effect on the risk of normal-tension glaucoma was significantly stronger in Asians and non-Hispanic whites [
There are several limitations to the present meta-analysis that need to be acknowledged. First, there was substantial heterogeneity among the studies, which may have resulted from differences in study designs, populations targeted, analysis strategies, and characteristics of participants. The real effects in particular studies may be significantly different from our estimate of overall effect. Rucker et al. [
In conclusion, we estimated the prevalence of OAG in a myopic Northeast Asian population and the prevalence of myopic OAG in a general Northeast Asian population. Our findings will inform future glaucoma studies as well as public health guidelines for Northeast Asian populations.
No potential conflict of interest relevant to this article was reported.
The authors received no financial support for this article.
None.
Flow diagram showing study selection process for meta-analysis.
Bar graphs showing prevalences of myopia, open-angle glaucoma (OAG), myopic OAG, and OAG among myopic patients.
Forest plot of (A) overall pooled percentage of open angle glaucoma among patients with myopia and (B) overall pooled prevalence of myopic open angle glaucoma. CI = confidence interval.
A funnel plot for meta-analysis of (A) overall pooled percentage of open angle glaucoma among patients with myopia and (B) overall pooled prevalence of myopic open angle glaucoma.
PubMed, Cochrane, and Embase search strategy
Search strategy | ||
---|---|---|
| ||
PubMed and Cochrane | Embase | |
1. Population (disease) | “Glaucoma”[MeSH] OR “Glaucoma”[tiab] OR “OAG”[tiab] OR “POAG”[tiab] | ‘Glaucoma’/exp OR ‘OAG’ OR ‘POAG’ |
2. Exposure | “Myopia”[MeSH] OR “Myopia”[tiab] OR “Myopic”[tiab] OR “Refractive errors”[MeSH] OR “Refractive errors”[tiab] OR “Refractive error”[tiab] | ‘Myopia’/exp OR ‘Myopic’ OR ‘Refraction error’/exp OR ‘Refractive error’ OR ‘Refractive errors’ |
3. Outcome | “Risk factors”[MeSH] OR “Risk factors”[tiab] OR “Risk factor”[tiab] OR “Association”[MeSH] OR “Association”[tiab] OR “Associated”[tiab] OR“Prevalence”[MeSH] OR “Prevalence”[tiab] | ‘Risk factor’/exp OR ‘Risk factors’ OR ‘Determinants’ OR ‘Determinant’ OR ‘Association’/exp OR ‘Associated’ OR ‘Prevalence’/exp OR ‘Prevalence’ |
4. Final search | 1 AND 2 AND 3 | 1 AND 2 AND 3 |
Quality assessment of included studies according to the Newcastle-Ottawa Scale
Study | Selection | Comparability | Outcome/exposure | Total score |
---|---|---|---|---|
Suzuki et al. [ |
4 | 1 | 2 | 7 |
Xu et al. [ |
4 | 1 | 2 | 7 |
Perera et al. [ |
4 | 2 | 2 | 8 |
He et al. [ |
4 | 2 | 2 | 8 |
Kim et al. [ |
4 | 2 | 2 | 8 |
For each of the section (selection, comparability, and outcome/exposure), a maximum of 4, 2, and 3 points could be given, respectively. Thus, the maximum score was 9. A higher score means higher quality.
Characteristics of studies included in the meta-analysis
Study | Country | Total sample size | Age (yr) | Female (%) | Definition of myopia (spherical equivalent in diopters) | Prevalence of myopia (no. of myopic patients) | Prevalence of OAG (no. of OAG patients) | Percentage of OAG among myopic patients (%) | Myopic OAG prevalence (no. of myopic OAG patients) |
---|---|---|---|---|---|---|---|---|---|
Suzuki et al. [ |
Japan | 2,874 | 58.4 ± 11.8 | 56 | <−1.0 | 30.5 (877) | 3.7 (107) | 4.9 | 1.5 (43) |
Xu et al. [ |
China | 4,319 | 55.8 ± 10.3 (40–90) | 56 | <−0.5 | 42.5 (1,837) | 2.8 (119) | 2.7 | 1.2 (50) |
Perera et al. [ |
Singapore |
3,109 | 58.2 ± 10.9 (40–80) | 52 | <−0.5 | 23.5 (731) | 3.3 (104) | 2.9 | 0.7 (21) |
He et al. [ |
China | 2,528 | 63.5 ± 8.8 (50–106) | 58 | All myopia |
9.5 (239) | 2.8 (72) | 5.0 | 0.5 (12) |
Kim et al. [ |
South Korea | 13,831 | 55.1 ± 0.2 | 57 | <−0.5 | 29.2 (4,039) | 5.1 (710) | 6.3 | 1.8 (235) |
Values are presented as number, mean ± standard deviation, mean ± standard deviation (range), or % (number), unless otherwise indicated.
OAG = open-angle glaucoma.
Malay population.
The definition of myopia was not provided.
Glaucoma diagnostic criteria for studies included in the meta-analysis
Study | Diagnosis of glaucoma |
---|---|
Suzuki et al. [ |
CDR ≥0.7 or neuroretinal rim width |
Xu et al. [ |
Optic disc abnormalities (a notch in the neuroretinal rim or abnormally large cup or a localized RNFLD) and GVFD |
Perera et al. [ |
CDR or CDR asymmetry 97.5th percentile or NRRW <0.1 CDR, GVFD, >180° of TM visible on gonioscopy |
He et al. [ |
CDR or CDR asymmetry 97.5th percentile, GVFD, gonioscopy |
Kim et al. [ |
Loss of neuroretinal rim with CDR ≥0.7 or asymmetry ≥0.2, DH, RNFLD, GVFD on FDT, open angle by Van Herick method |
CDR = cup-to-disc ratio; DD = disc diameter; RNFLD = retinal nerve fiber layer defect; GVFD = glaucomatous visual field defect; NRRW = neuro retinal rim width; TM = trabecular meshwork; DH = disc hemorrhage; FDT = frequency doubling technology;
ISNT, in the order of inferior (I) > superior (S) > nasal (N) > temporal (T).
Characteristics of studies included in the meta-analysis according to the degree of myopia
Study | Degree of myopia | Definition of myopia degree (spherical equivalent in diopters) | Total sample size | No. of OAG patients | Myopia-degree-specific OAG prevalence (%) | Percentage of OAG among patients with specific degree of myopia (%) |
---|---|---|---|---|---|---|
Suzuki et al. [ |
Low | <−1.0 to >−3.0 | 389 | 17 | 0.59 | 4.37 |
Moderate-to-high | ≤−3.0 | 488 | 26 | 0.90 | 5.33 | |
Xu et al. [ |
Low | <−0.5 to ≥−3.0 | 1,206 | 27 | 0.32 | 2.24 |
Moderate | <−3.0 to ≥−6.0 | 417 | 10 | 0.12 | 2.40 | |
High | <−6.0 | 214 | 13 | 0.15 | 6.07 | |
Perera et al. [ |
Low | <−0.5 to ≥−4.0 | 583 | 15 | 0.50 | 2.57 |
Moderate-to-high | <−4.0 | 148 | 6 | 0.20 | 4.05 | |
Kim et al. [ |
Low-to-moderate | <−0.5 to >−6.0 | 3,797 | 214 | 1.63 | 5.64 |
High | ≤−6.0 | 242 | 21 | 0.16 | 8.68 |
OAG = open-angle glaucoma.