Metabolic syndrome as an independent risk factor for glaucoma: a nationally representative study

In this retrospective, cross-section study, we analyzed all data from the 2019‒2021 Korean National Health and Nutrition Examination Survey (KNHANES). The KNHANES is a nationwide, representative, and population-based annual survey conducted by the Korea Disease Control and Prevention Agency (KDCA) that monitors the health and nutritional status of the Korean population [21]. The KNHANES adopts a three-stage sample design [22]. First, primary sample units (PSUs) are selected from a list of census blocks or resident registration addresses, with each PSU containing approximately 50–60 households. Second, 20 households are chosen through field surveys for household screening. Third, all members aged 1 year and above in the selected households participate in the survey. Each year, a total of approximately 10,000 individuals from 4800 households are sampled across all 192 PSUs. Over the 2019–2021 period, a total of 576 PSUs with 14,400 households are sampled for the KNHANES. The stratification variables are province, city/township/district, and housing type, while the implicit stratification variables were sex, age, residential area, and head of household's education level. Finally, sampling weights are assigned to each participant to generalize the units for representing the Korean population based on the estimated number of households and population for each year, so that the sum of household weights was equal to the total number of households in South Korea, and the sum of individual weights by survey category was equal to the population of the corresponding age group in South Korea [22]. Detailed information is available from the KNHANES website (https://​knhanes.​kdca.​go.​kr/​knhanes/​eng, accessed on May 8, 2023).

Figure 1 shows the flowchart of the study population selection. Among 22,559 patients listed in KNHANES 2019–2021, we excluded those aged < 19 y (n = 3868), insufficient data to evaluate MetS (n = 788), post-glaucoma surgery status (n = 16); currently under treatment for glaucoma (n = 192), and unknown glaucoma status (n = 6196). After excluding these, we analyzed data from 11,499 participants.

This study conformed to the ethical guidelines of the 1964 Declaration of Helsinki and its amendments. Written, informed consent was obtained from all eligible KNHANES participants. The Institutional Review Board (IRB) at Eulji University College of Medicine approved the study protocol (IRB number: 2022-07-022).

Height (cm) and weight (kg) were measured, then body mass index (BMI, kg/m²) was calculated. Waist circumference (cm) was measured at the midline between the lowest margin of the rib and highest margin of the iliac crest. Abdominal obesity was defined as waist circumference ≥ 90 (men) and ≥ 85 cm (women). Systolic (SBP) and diastolic (DBP) blood pressure (mmHg) was measured three times in seated participants at least 5 min after resting. The mean values of the last two measurements were defined as SBP and DBP. The mean blood pressure (MBP, mmHg) was calculated as DBP + 1/3 × (SBP − DBP). Cigarette use was classified as follows: nonsmoker (never smoked or had smoked < 100 cigarettes throughout life), ex-smoker (smoked ≥ 100 cigarettes throughout life but quit smoking by the time of the survey), or current smoker (smoked ≥ 100 cigarettes throughout life and had not quit by the time of the survey). Alcohol use was classified as currently, or not currently consumed. Physical activity was assessed using the Global Physical Activity Questionnaire [23], as the metabolic equivalent of task (MET) hours per day (MET-h/day). Participants were categorized into groups with low, moderate, or high physical activity (< 7.5, 7.5–30, and > 30 METs/h/day, respectively). Total energy intake was calculated using the 24-h recall method. Educational levels were categorized as elementary, middle, and high school, and college or university, and monthly household income was classified into quartiles. Fasting plasma glucose (FPG), serum total cholesterol, triglyceride, and high-density lipoprotein (HDL) cholesterol were measured in blood samples collected after fasting for at least 8 h.

To ensure quality control of ophthalmological exams and the survey, the KDCA and the Korean Ophthalmological Society (KOS) implemented team education and training programs twice a year. Four well-trained nurses were responsible for conducting visual field tests, IOP measurements, and optical coherence tomography (OCT) on the participants. All participants who aged 19 years or older underwent automated visual field examination employing frequency doubling technology (FDT), N30-5 screening, and Humphrey Matrix (Carl Zeiss Meditec) following standard protocols. Only reliable visual fields were included, characterized as having ≤ 1 fixation losses and false-positive responses. An abnormal visual field was denoted by the presence of at least one location exhibiting reduced sensitivity [24]. IOP was measured using the probe of an Icare PRO rebound tonometer (Icare Finland Oy, Helsinki, Finland) placed within 3–7 mm of the cornea, perpendicular to the central cornea in seated participants looking straight ahead. The IOP was measured by taking six consecutive measurements for each eye. The interpretation of the readings involved discarding the highest and lowest values among the six measurements and calculating the average of the remaining four readings [24]. To capture images of the fundus of the eye, non-mydriatic fundus photography (VISUCAM 224, Carl Zeiss Meditec) with a field angle of 45° was employed, which was reduced to 30° in cases of small pupils. Digital images were taken with physiological mydriasis in all participants. For each participant, one fundus image, covering both the macula and optic disc, was obtained for each eye. Additionally, red-free photography was acquired by digitally transforming the original color fundus photography [24]. Every participant in the study underwent a thorough examination of both the posterior and anterior segments using the Cirrus high-definition OCT (Cirrus HD-OCT 500, Carl Zeiss Meditec). Posterior segment imaging was obtained using the built-in program, which included the Macular Cube 512 × 128 images, two high-definition 5 Line Raster scans, and the Optic Disc Cube 200 × 200 protocol. Additionally, anterior segment imaging was conducted using the built-in 15.5 mm Wide Chamber View protocol [24].

Using the diagnostic criteria for MetS based on the National Cholesterol Education Program Adult Treatment Panel III [25], we defined MetS when at least three of the following criteria were met: abdominal obesity, FPG ≥ 100 mg/dL or use of oral hypoglycemic agents or insulin therapy; serum triglycerides ≥ 150 mg/dL or under control with lipid-lowering agents; serum HDL cholesterol < 40 (men) and < 50 (women) mg/dL in women; SBP ≥ 130 mmHg, DBP ≥ 85 mmHg, or under control with anti-hypertensive agents. The participants were then assigned to groups with (n = 4384) and without (n = 7115) MetS.

Using data from visual field test and OCT, POAG was defined according to the International Society for Geographic and Epidemiological Ophthalmology criteria [26]. A patient was diagnosed with POAG if they fell under Category I, II, or III.

Category I required both structural damage and corresponding visual field defects to be present together. The criteria included: (1) glaucomatous structural damage, such as thinning or notching of the optic disc, vertical cup-to-disc ratio of ≥ 0.7 (OCT), asymmetry of ≥ 0.2 between the two eyes, presence of retinal nerve fiber layer defect, or optic disc hemorrhage, with the location of the damage noted as superior, inferior, or none; and (2) glaucomatous visual field defects, defined as reduced sensitivity corresponding to retinal nerve fiber layer or optic disc abnormalities on the FDT perimetry test, with a fixation error and false positive rate ≤ 1, and the location of the defect noted as superior, inferior, or none.

Category II required clear structural damage without proven visual field defects, and the criteria included: (1) glaucomatous structural damage, such as thinning or notching of the optic disc with a vertical cup-to-disc ratio of ≥ 0.9, asymmetry of ≥ 0.3 between the two eyes, or presence of optic disc and corresponding retinal nerve fiber layer defects, with the location of the damage noted as superior, inferior, or none, and (2) absence of or fixation error and false positive rate ≥ 2 on FDT perimetry test.

Category III included patients who could not undergo visual field testing due to inability to observe the optic disc. Criteria included: (1) corrected visual acuity of less than 0.05 and IOP measured by rebound tonometry of ≥ 23 mmHg.

The KOS established a glaucoma reading committee, which consisted of 24 primary readers and 5 secondary readers in the first round of evaluations in 2019. In the second round, there were 21 primary readers and 6 secondary readers in 2020, and 36 primary readers and 6 secondary readers in the first round of evaluations in 2021. These readers reviewed the results of the visual field tests as well as OCT, and made diagnoses regarding the presence of glaucoma. Each patient's data was assessed by two primary readers, and they jointly made the initial assessment. The data was distributed among the readers based on regions, allowing them to review the relevant data for each specific week.

Finally, we defined ocular HTN as IOP > 21 mmHg and classified participants into POAG with ocular HTN group, and NTG (POAG without ocular HTN) group [27].

Data are presented as means ± standard error (SE) for continuous variables and as ratios (% with SE) for categorical variables. Differences in continuous and categorical variables between groups with and without MetS were compared using weighted t-tests and chi-square tests, respectively.

We estimated ORs with 95% CI using univariate and multivariate weighted logistic regression analyses for POAG according to MetS status. Age, sex, and BMI were adjusted in Model 1. These adjusted variables plus socioeconomic factors and personal habits of total energy intake, smoking status, alcohol consumption, physical activity, education, and monthly household income were included in Model 2. The variables adjusted in model 2 plus metabolic factors, including MBP, FPG, and serum total cholesterol were included in Model 3. We further adjusted for IOP in Model 4 for sensitivity analysis.

We determined dose–response relationships between the number of MetS components and POAG. We created a cubic spline curve, then applied univariate and multivariate weighted logistic regression analyses for POAG per MetS incremental component.

Table 1 shows the clinical characteristics of the study population with and without MetS. The proportions of men, current smokers, and individuals with low levels of physical activity were significantly higher in the group with, than without MetS. They also had higher mean values for age, BMI, waist circumference, MBP, and a higher proportion of individuals in the lowest quartile of monthly household income and individuals with an elementary school education than those without MetS. Less participants consumed alcohol in the group with, than without MetS. Total energy intake between the groups did not significantly differ.

Figure 2 shows the prevalence of POAG according to MetS status. The prevalence of POAG was 5.7% (252 of 4384) and 3.5% (250 of 7115) patients with and without MetS, respectively.

Table 2 shows the results of weighted logistic regression analysis of POAG risk according to MetS. The OR for POAG was higher in all models of the group with, than without MetS. The unadjusted OR was 1.85 (95% CI 1.52–2.25). The adjusted ORs (95% CI) for POAG in models 1, 2, 3 and 4 in the MetS group were 1.37 (1.04–1.80), 1.51 (1.10–2.08), 1.49 (1.04–2.13), and 1.47 (1.04–2.09), respectively. Risk for NTG was higher in patients with, than without MetS, whereas risk of POAG with ocular HTN did not significantly differ between the groups (Additional file 2: Table S1).

The cubic spline curve in Fig. 3 shows that risk of POAG increased along with more MetS components. Table 3 shows the results of the weighted logistic regression analysis of POAG risk per increase in MetS components. The unadjusted OR was 1.28 (95% CI 1.19–1.38). The adjusted ORs (95% CI) for POAG per increment in the number of MetS components in models 1, 2, 3, and 4 were 1.17 (1.05–1.31), 1.21 (1.07–1.37), 1.19 (1.03–1.37), and 1.20 (1.04–1.39), respectively.

Additional file 2: Table S2 shows the results of weighted logistic regression analysis of ocular HTN in patients with and without metabolic syndrome. The findings indicate that there is no significant relationship between MetS and ocular HTN. A significant dose–response relationship between the number of MetS components and risk of ocular HTN was not evident (Additional file 1: Fig. S1), however, the relationships were significant in univariable model (OR: 1.17; 95% CI 1.01–1.35) and multivariable models 1, 2, and 3, with adjusted ORs (95% CI) of 1.20 (1.00–1.44), 1.24 (1.03–1.51), and 1.21 (1.00–1.47), respectively for ocular HTN per increment in MetS components (Additional file 2: Table S3).