This study of a population-based, cross-sectional survey of Korean adults 40 years and older found a significant association between myopia and glaucoma. The odds of having a glaucoma diagnosis were increased 2.2-fold and 4.6-fold for subjects with moderate and severe myopia, respectively, compared to those with emmetropia. In addition, the odds of a visual field defect on FDT were increased 2.3-fold and 20.9-fold in mild and severe myopia, respectively. The odds of having structural damage to the optic disc were increased 1.8-fold and 2.3-fold in moderate and severe myopia, respectively. The odds of having a VCDR greater than or equal to 0.7 were increased 1.7-fold in severe myopia.
The relationship found in this study between myopia and glaucoma is similar to other studies in Asian populations. In the Beijing Eye Study,
15 glaucoma susceptibility was higher in those with worse than −6 D of myopia. Though the frequency of glaucoma increased with more severe myopia, the relationship was not significant for refractive errors less severe than −6 D. In the Handan Eye Study,
22 groups with −3 D to −6 D of myopia had a 4.7 (1.6–13.5, 95% CI) greater odds of having POAG, though not statistically significant at other levels of refractive error. In the Tajimi Study,
14 individuals with −1 to −3 D of myopia had a 1.85 (1.03–3.31) times greater odds of glaucoma and individuals with less than −3 D had a 2.60 (1.56–4.35) times greater odds of glaucoma. In the Singapore Malay Eye Study,
16 refractive error of less than −4 D was associated with a 2.80 (1.07–7.37) greater odds of glaucoma. Similar findings have been shown in American studies of a predominantly Caucasian population in the Beaver Dam Eye Study,
6 in Latinos in the Los Angeles Latino Eye Study (LALES),
8 and in a nationally representative sample of the United States.
7 Similar findings have also been reported in other countries including The Netherlands,
9 Sweden,
10 Australia,
11 Barbados,
12 and India.
13 The present study is the first to investigate the association between refractive status and glaucoma in a cross-sectional survey of the Korean population.
While the findings from prior studies, specifically those in East Asian countries, could be extrapolated to the Korean population, the prevalence of POAG, primary angle-closure glaucoma (PACG), and NTG varies significantly between differing ethnicities and nationalities. In one analysis of a United States managed care network database,
17 there were notable differences in the prevalence of POAG, PACG, and NTG in different racial groups. The prevalence of POAG was highest in Blacks (12.2%), then Asians (6.5%), Latinos (6.4%), and lowest in Whites (5.6%). Primary angle-closure glaucoma prevalence was highest in Asians (2.2%), then Latinos (1.5%), Blacks (1.2%) and Whites (1.0%). Normal-tension glaucoma was highest in Asians (2.1%), Blacks (1.6%), Latinos (1.1%), and Whites (1.0%). Further analysis separating Asian Americans by country of origin demonstrated additional distinctions in glaucoma subtypes. In general, NTG is more common among Japanese Americans, whereas PACG is more common in Chinese, Vietnamese, and Pakistani Americans.
The predominant form of glaucoma in Korea is NTG. In a previous analysis of native Koreans in the KNHANES, POAG, PACG, and NTG (a subset of POAG) were present in 2.0 ± 0.2%, 0.1 ± 0.1%, and 1.9 ± 0.2%, respectively.
18 In a different survey of a rural Korean community, prevalence of POAG, NTG, and PACG was 3.5%, 2.7%,
20 and 0.7%,
23 translating to 77% of OAG being normotensive (≤21 mm Hg). The proportions of POAG and PACG in vary widely in different Asian countries, with POAG accounting for between 65% and 95% of glaucoma.
24 –29 The varying ratios of POAG to PACG may be attributed to true epidemiologic differences by country, but also potentially to differing examination techniques, definitions of POAG and PACG, and regional factors such as differing proportions of rural and urban populations. Thus, it is apparent that there are clear differences in subtypes of glaucoma by race and nationality, and it may be inappropriate to assume homogeneity by race or ethnicity alone.
The high prevalence of NTG in Korea, and other countries, is of particular importance as NTG may be more difficult to recognize and diagnose. In Sweden,
30 a comparison was made of individuals with glaucoma who were previously undiagnosed (screened) and patients with a diagnosis and had sought ophthalmologic care on their own (self-selected). The study showed that NTG was four times as common in the screened group (52.9%) compared with the self-selected group (13.5%). The authors hypothesized that this may represent patients with NTG being more commonly undiagnosed and less likely to seek eye health services. The high levels of undiagnosed glaucoma were true despite 62% of the screened patients having seen an ophthalmologist prior to the study, and 17% within the two years before screening. In large population studies, previously undiagnosed glaucoma was found in 49% to 53% of participants in the Thessaloniki Eye Study,
31 Baltimore Eye Survey,
32 Blue Mountain Eye Study,
33 Barbados Eye Study,
34 and Rotterdam Study.
35 Identifying risk factors for glaucoma, particularly NTG, could improve disease detection and treatment initiation at earlier stages of disease. The results of this study suggest that increased vigilance in individuals with more severe myopia may be needed to detect undiagnosed ocular pathologies.
Myopia has been increasing at a rapid rate in the younger population, and there could be dramatic increases in the prevalence of myopia as this population ages. An analysis of the KNHANES showed that the prevalence of myopia (<−0.75 D) in Korea is 53.7% for all ages and as high as 78.8% in the 19 to 29 age group.
18 Jung et al.
36 studied a group of 19-year-old Korean males, and found that 96.54% had myopia (<−0.5 D), 31.00% mild myopia (−0.5 to −2.99 D), 43.92% had moderate myopia (−3.0 D to −5.99 D), and 21.62% had high myopia (<−5.99 D). Some theories for higher myopia levels include more time studying and performing near-work,
37 and appears to be less prevalent among children spending more time outdoors.
38
The prevalence of myopia worse than −6 D was 2.3% in the current study. The prevalence of refractive error less than −6 D was 2.8%, 3.6%, and 5.5% in China,
15 Singapore,
39 and Japan,
40 respectively. It is possible that the 40 years and older South Korean population has a lower rate of high myopia compared with other Asian populations. The reported prevalence of high myopia could also be affected by differences in the study populations with respect to environmental factors, such as education level, time spent on near-work
37 or outdoor activities,
38 and urbanization.
41,42 It is difficult to determine if the prevalence of high myopia is truly lower than other Asian countries, or is due to a combination of reasons stated above. While our study showed that more severe myopia was associated with higher odds of glaucoma, it is unclear whether individuals with differing severities of myopia have different rates of glaucoma progression. In a retrospective study of 143 patients receiving medical treatment for NTG, no significant association was found between refractive status and the rate of progression of NTG on visual field testing.
43 Conversely, there are studies that have shown more visual field progression in glaucoma patients with more severe myopia.
44,45
A potential confounding factor is the optic disc appearance with high myopia. Myopic individuals tend to have abnormal optic disc findings, such as tilted optic nerve heads
46 and parapapillary atrophy.
47 In addition, individuals with severe myopia may have physiologically larger optic discs. In one study, highly myopic eyes with less than −8 D of refractive error had a larger disc area and a larger VCDR, 0.77 in high myopia and 0.71 in others.
47 However, for refractive errors between −8 and +4 D, there was no correlation between refractive status and optic disc size. In the Singapore Malay Eye Study, highly myopic eyes (<−6 D) were associated with a larger optic disc area, but a smaller cup-to-disc ratio, compared with those with emmetropia.
48 In the Rotterdam Study, while more myopia was associated with an increased optic disc area, there was no association between refractive error and cup-to-disc ratios.
49 At severe levels of myopia, optic disc area appears to be greater but the relationship between refractive error and cup-to-disc ratio is unclear.
In our analysis, only 1.0% of participants had a refractive error less than −8 D of myopia. The unadjusted and adjusted odds of having a VCDR greater than or equal to 0.7 were statistically similar in our study between individuals with emmetropia and mild, moderate, or severe myopia. On re-analysis, had our severe myopia group (<−6 D) been split into two groups, (−6 to −8 D and <−8 D), the odds of a glaucoma diagnosis would have changed from 4.6 (<−6 D), to 3.7 (−6 D to −8 D), and 5.7 (<−8 D). Refractive error may only alter optic disc parameters at extreme levels of myopia, which comprised a minor portion of our population. Even by excluding those individuals with less than −8D of myopia, the relationship between glaucoma and myopia persists.
Another potential limitation is the use of FDT instead of standard automated perimetry to assess visual field defects. The ISGEO category 1 criteria for glaucoma consist of a combination of optic disc and visual field results consistent with glaucoma. Currently, automated static threshold perimetry is the preferred tool for testing visual fields.
50 While FDT may not be the first choice for visual field testing in glaucoma diagnosis, it offers the benefit of a fast and reliable test that is suited for mass screenings and may predict glaucomatous functional damage earlier than standard perimetry.
51
In highly myopic individuals, it can be unclear if visual field defects found on perimetry are due to glaucoma or myopia. Individuals with a tilted disc can appear to have visual field defects that later improve with myopic correction.
52 While myopic individuals without a diagnosis of glaucoma may have visual field defects, those defects can present differently from defects due to glaucomatous damage.
53 However, atypical optic disc shape, RNFL defects, and visual field defects not of the classical glaucoma pattern were found in highly myopic patients with a diagnosis of OAG.
54 For those with severe myopia, the odds of an abnormal visual field were higher than the odds of glaucoma by ISGEO criteria, which may be due to some combination of glaucoma and/or myopia rather than glaucoma alone.
While Humphrey Field analysis is typically the test of choice for perimetry, a previous study
55 comparing FDT with Humphrey Field Analyzer found 100% sensitivity and specificity for advanced glaucoma, 96% sensitivity and specificity for moderate glaucoma, and 85% sensitivity and 90% specificity for early glaucoma. Reliable algorithms for detecting glaucoma using FDT in screening modes have previously been investigated. Trible et al.
56 found similar sensitivity and specificity for detecting glaucoma among algorithms using (1) the single most severe point, (2) two points, or (3) a cluster of three or more points. The one point algorithm had a specificity of 95%, and sensitivities of 39%, 86%, and 100% for early, moderate, and severe glaucoma. In another study, the best algorithm for detecting glaucoma was a visual field defect of two or more abnormal test locations, with a sensitivity of 91% and a specificity of 94%.
57 However, it should be emphasized that the detection rates for glaucoma defects in these studies may not apply to cases of high myopia, in which field defects may be due to myopic retinopathy. We excluded patients with AMD to reduce the number of visual field defects due to non-glaucomatous damage. However, other retinal abnormalities could have affected the results from visual fields.
The ISGEO criteria was intended to provide one method for assessing glaucoma in cross-sectional surveys.
21 It combines functional and structural findings to generate estimates of glaucoma prevalence, which may differ if based on a complete eye assessment in a standard clinic setting. A notable benefit is that the ISGEO criteria provide a consistent definition for glaucoma that allows for comparisons with multiple studies that have used the ISGEO criteria. For this study, it allows for comparisons to other East Asian populations such as in China,
24,25 Singapore,
29 and Japan.
26
However, the ISGEO criteria do not include adjustment for the optic disc size or qualitative factors, such as observable RNFL defects or adherence to the ISNT rule.
58 The addition of qualitative factors would increase individual screening and diagnosis time, which is multiplied in large population studies. Also, the validity of a new definition for glaucoma with qualitative factors would need to be confirmed.
59 The ISGEO criteria utilize a VCDR cutoff, with the 97.5th percentile being 0.7 in this current study. Analysis by glaucoma specialists in previous studies have shown that the VCDR was less than 0.7 in 32.3% of POAG in a study in the Handan Eye Study
22 and 50% in the LALES.
60 The ISGEO criteria do not take into account the progression of disease in diagnosis of glaucoma. Given the cross-sectional nature of this study, it is possible that those classified as having glaucoma are actually glaucoma suspects. While the ISGEO criteria may be of use in cross-sectional studies and screening events to classify glaucoma, a preferred method would be a clinical diagnosis, based on a comprehensive adult eye examination by trained ophthalmologists and longitudinal observation for progression of disease.
61
If FDT had been administered to all eye exam participants who met criteria for visual field examination, a more representative estimate of glaucoma prevalence could have been obtained. However, the KNHANES did not provide any information regarding the reason why some subjects who met criteria for FDT examination did not receive this exam. To investigate the possibility of selection bias, we compared the proportion of FDT testing across eligible subjects in each refractive status group and found no significant difference. A more accurate measure of glaucoma prevalence could have been obtained if all individuals meeting ISGEO criteria 1 for structural damage had received the FDT exam. For comparison, the prevalence of glaucoma diagnosis was 3.0% in our study, which is similar to but slightly lower than the 3.5% prevalence of POAG reported in a different survey of a Korean population.
20
The KNHANES had no information for certain ocular parameters, including central corneal thickness and axial length measurements. Axial length would have been useful in determining whether or not refractive error due to axial myopia, corneal curvature, or lens changes are associated with glaucoma,
16 but this information was not available in the KNHANES. Myopia has been associated with higher rates of cataracts and cataract surgery.
62 In one study, highly myopic individuals received cataract surgery earlier (59.6 years) than controls (67.5 years).
63 We excluded pseudophakic and aphakic eyes from the current study because axial length measurements were not available and autorefraction was used to classify myopia. Though subject age is available, there was no information on subject age at the time of their cataract surgery. Because high myopes are at higher risk for getting cataracts and getting earlier cataract surgery, it is possible that the excluded aphakic and pseudophakic group may have proportionally more cases of high myopia and therefore more cases of glaucoma. In this situation, we could have underestimated the prevalence of myopia and glaucoma in the current study.
Intraocular pressure was found to be statistically higher in individuals with all levels of myopia compared with emmetropia. However, the difference in IOP was less than 1 mm Hg, which may be clinically insignificant given both the small difference in IOP and potential error in measurement.
In summary, in a population-based sample of Korean adults aged 40 years and older, we found an association between myopia and glaucoma diagnosis, visual field defects on FDT exam, optic disc structural damage, VCDR greater than or equal to 0.7 and higher IOP. However, these associations from a cross-sectional survey do not demonstrate causation, and future prospective studies are needed to compare the progression of glaucoma in subjects with varying severity of myopia. Since myopia prevalence has been rising at a rapid rate, identifying whether or not myopia is a risk factor for glaucoma may become increasingly useful in helping patients and physicians screen for and diagnose glaucoma.