Participants

Patients who were newly diagnosed with NTG were enrolled prospectively in the study by the following inclusion and exclusion criteria. Inclusion criteria were presence of typical optic nerve head changes, including increased cupping, focal or diffuse loss of the neuroretinal rim, or both; glaucomatous VF loss on at least 2 consecutive tests; open angle on gonioscopy; IOP of 21 mm Hg or less (without topical treatment) in repeated measurements on different days; and follow-up at our clinic of at least 2 years with visits at 4- to 6-month intervals. Exclusion criteria were the presence of diabetes mellitus (defined as fasting blood glucose of more than 126 mg/dL or postprandial 2-hour blood glucose of more than 200 mg/dL; guidelines for diabetes mellitus, the Korean Society of Diabetes 2008) or systemic cardiovascular disease, smoking, or taking vasoactive drugs or drugs known to affect the autonomic nervous system. Those having abnormally high blood pressure (BP) on measurement, namely, more than 140 mm Hg systolic and more than 90 mm Hg diastolic (guidelines for the management of hypertension, the Korean Society of Hypertension 2008) were excluded. Patients taking drugs for treatment of systemic hypertension or heart disease, such as calcium channel blockers, angiotensin-converting enzyme inhibitors, β-adrenergic blockers, and α-1 adrenergic blockers, also were excluded. Subjects with known connective tissue disease that might have affected the results of nailfold capillaroscopy or subjects with any symptoms suggesting autonomic nervous system involvement also were excluded. Subjects with any intraocular or neurologic disease that could lead to VF loss and consistently unreliable VFs (defined as false negative ≥ 15%, false positive ≥ 15%, and fixation losses ≥ 20%) were excluded from the study. The subjects who visited our glaucoma clinic between 2007 and 2010 who fulfilled the inclusion and exclusion criteria and agreed to undergo heart rate variability assessment and nailfold capillaroscopy were enrolled. For subjects with unilateral disease, the eye with the disease was recruited. If a subject had bilateral disease, the worse eye was recruited.

Each subject underwent a complete ophthalmologic examination, including visual acuity, refraction, slit-lamp biomicroscopy, gonioscopy, Goldmann applanation tonometry, dilated stereoscopic examination of the optic disc, red-free fundus photography, optical coherence tomography (Stratus OCT system version 3; Zeiss-Humphrey Ophthalmic Systems, Dublin, California, USA), Heidelberg retinal tomography (Heidelberg Engineering GmbH, Heidelberg, Germany), and Humphrey VF examination using the Swedish interactive threshold algorithm standard 24-2 test (Carl Zeiss Meditec, Dublin, California, USA). BP was measured with a digital automatic blood pressure monitor (Dinamap model Pro 18 Series DP110X-RW, 100V2; GE Medical Systems Information Technologies, Inc, Milwaukee, Wisconsin, USA), with the participant seated after 5 minutes of rest. BP was measured twice, 5 minutes apart. Mean arterial BP was defined as diastolic BP + 1/3 (systolic BP − diastolic BP). Mean ocular perfusion pressure (MOPP) was defined as 2/3 (mean arterial BP − IOP).

An abnormal glaucomatous VF was defined as the consistent presence of a cluster of 3 or more nonedge points on the pattern deviation plot with a probability of occurring in less than 5% of the normal population, with 1 of these points having the probability of occurring in less than 1% of the normal population; a pattern standard deviation at P < .05; or glaucoma hemifield test results outside normal limits. VF defects had to be repeatable on at least 2 consecutive tests.

All patients were referred to the rheumatology outpatient clinic of the Department of Internal Medicine at Seoul St. Mary’s Hospital, Seoul, Korea. One of the authors (S.-H.P.) performed the physical examination and the recording of medical status. Hematologic status, clinical chemistry, hepatitis, HIV serologic analysis, and urine analysis were performed to determine the health status of the subject. Heart rate variability assessment and nailfold capillaroscopy were performed on the same day by 2 independent technicians in the rheumatology department. All patients were given an information leaflet advising them to refrain from drinking caffeine or alcohol for 1 day before the tests. Patients also were requested to avoid activities that would affect blood pressure, including running and jumping, for at least 2 hours before the tests.

Heart Rate Variability Assessment

Heart rate variability assessment was performed before the nailfold capillaroscopy. The method has been described in detail elsewhere. Briefly, noninvasive BP was obtained from the radial artery at the wrist using an automated oscillometric device. The echocardiography was monitored for 5 minutes after 30 minutes of rest in the supine position. Echocardiography signals were transferred to a Medicore Heart Rate Analyzer (model SA-3000P; Medicore, Seoul, Korea). The intervals between each successive normal QRS complex were determined initially. The standard deviation of qualified normal-to-normal intervals (SDNN) is the standard deviation of all the normal rate-to-rate intervals in a 24-hour echocardiography recording in milliseconds. In the present study, testing was performed over a 5-minute period under controlled conditions. A reduction of the SDNN reflects low heart rate variability, which means a high tone of sympathetic activity of the heart. Power spectral density analysis provided information on how power was distributed (the variance) as a function of frequency. The typical power spectrum of a 5-minute heart rate tachogram is divided into 3 bands or regions consisting of a 0.15- to 0.4-Hz high-frequency band, a 0.04- to 0.15-Hz low-frequency band, and a 0.0033- to 0.04-Hz very low-frequency band. The high-frequency band reflects fast changes in beat-to-beat variability resulting from parasympathetic activity. The low-frequency band is considered a fair approximation of sympathetic activity. The very low-frequency band reflects mostly sympathetic stimulation. Total power is the average of the total power (power in the band less than 0.04 Hz) of each 5-minute segment, including the very low-frequency, low-frequency, and high-frequency power spectrum bands. The low-frequency to high-frequency is the ratio of low-frequency to high-frequency power. A higher ratio indicates increased sympathetic activity or reduced parasympathetic activity.

Patients were divided into 3 groups according to SDNN. Our previous study showed that this parameter differed significantly in patients with glaucoma versus normal healthy controls. The lowest heart rate variability group consisted of patients within the lowest tertile in terms of SDNN. The middle heart rate variability group consisted of patients within the middle tertile, and the highest heart rate variability group consisted of patients within the highest tertile. The lowest heart rate variability group has a reduced value of SDNN, which means a higher tone of sympathetic activity with higher autonomic dysfunction of the heart than other groups.

Nailfold Capillaroscopy

Nailfold capillaroscopy was performed in a separate room after the heart rate variability assessment. The method has been described in detail elsewhere. Briefly, each patient was seated with the dorsum of the hand facing upward, with halogen lights illuminating the nails coated with immersion oil, under a nailfold microscope. Nailfolds of the second, third, and fourth digits of both hands were observed using a light microscope (Olympus SZ-PT, Tokyo, Japan) under ×100 and ×400 magnification. All microphotographs were transmitted to a computer via a digital camera (Polaroid, Minnetonka, Minnesota, USA).

The capillaroscopy results were analyzed by a single observer (S.-H.P.) who did not know the ophthalmologic diagnosis and the results of heart rate variability testing. The capillary vessel architecture, morphologic features, and distribution; the number of capillary vessels; and the presence of splinter hemorrhages were evaluated. Dilated loops were defined as detection of a microvascular loop with a homogeneous increase in diameter of more than 50 μm. Avascular areas were defined as the absence of capillaries in a field of at least 500 μm. Nailfold hemorrhages were considered present when more than 2 punctuated hemorrhages per finger or confluent areas of hemorrhage were observed.

Patients were categorized by nailfold capillaroscopy into 2 groups according to 3 signs: dilated loops, avascular areas, and nailfold hemorrhages. To be classified as abnormal, only 1 positive finding among the 3 signs was required. Patients with normal findings on all 3 signs were classified into the normal capillaroscopy group.

Dividing the Visual Field Defects by Location

We selected consecutive eyes with VF defect clusters confined to the central 10 degrees or outside the central 10 degrees in 1 hemifield. The central 10 degrees and outside the central 10 degrees of the VF were divided into superior and inferior sectors ( Figure 1 ). Patients with overlapping VF defect clusters in different sectors were excluded from the analysis. The mean of the deviations of the pattern standard deviation (PSD) plot at 20 points for superior and inferior periphery sectors and at 6 points for superior and inferior center sectors were calculated.

Diagram showing the pattern standard deviation plot of the visual field examination. Four sectors and 52 test points of the right eye are shown.

Statistical Analysis

Statistical analysis was performed using the SPSS statistical package (SPSS, Inc, Chicago, Illinois, USA). The Student t test or a 1-way analysis of variance was used to compare differences between or among groups. The post hoc test with the Scheffe method was used for multiple comparison. The chi-square test was used where appropriate to compare frequencies. The frequency of VF defect locations in each sector was compared between groups by the chi-square test. The means of the deviations in each sector of the PSD plot were analyzed to compare the depth of VF defects with the Student t test. The relationship between mean deviation (MD) and PSD was evaluated by linear regression analysis according to the defect locations in each group. To find out factors related to the location of the VF defects, univariate and multivariate logistic regression analyses were carried out. The dependent variables were either central or peripheral VF defects. The independent variables were age, sex, spherical equivalent, MD, mean IOP, systolic BP, diastolic BP, MOPP, group of the heart rate variability assessment, and presence of abnormalities of nailfold capillaroscopy. Because the sex, group of the heart rate variability assessment, and presence of abnormalities of nailfold capillaroscopy are nominal in scale, these independent variables were investigated as independent factors using a regression model, and dummy variables were performed with male, heart rate variability group, and no abnormalities in nailfold capillaroscopy as a standard. The variables that retained significance at P < .10 in univariate analysis were included in the multivariate model. P < .05 was considered to represent statistical significance.