May 2019
Volume 60, Issue 6
Open Access
Glaucoma  |   May 2019
Optic Disc Microhemorrhage in Primary Open-Angle Glaucoma: Clinical Implications for Visual Field Progression
Author Affiliations & Notes
  • Ahnul Ha
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
  • Young Kook Kim
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
  • Sung Uk Baek
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
  • Ki Ho Park
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
  • Jin Wook Jeoung
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
  • Correspondence: Jin Wook Jeoung, Department of Ophthalmology, Seoul National University Hospital, Seoul National University College of Medicine, 101 Daehak-ro, Jongno-gu, Seoul 03080, Korea; neuroprotect@gmail.com
  • Footnotes
     AH and YKK contributed equally to the work presented here and should therefore be regarded as equivalent authors.
Investigative Ophthalmology & Visual Science May 2019, Vol.60, 1824-1832. doi:10.1167/iovs.19-26673
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Ahnul Ha, Young Kook Kim, Sung Uk Baek, Ki Ho Park, Jin Wook Jeoung; Optic Disc Microhemorrhage in Primary Open-Angle Glaucoma: Clinical Implications for Visual Field Progression. Invest. Ophthalmol. Vis. Sci. 2019;60(6):1824-1832. doi: 10.1167/iovs.19-26673.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: To establish the existence of minute-sized optic disc hemorrhage (DH; i.e., optic disc microhemorrhage [micro-DH]) in primary open-angle glaucoma (POAG) and to evaluate its clinical implications for glaucoma progression.

Design: Retrospective analysis of prospectively collected data.

Methods: POAG patients with macro-DH who had met the following conditions were included: follow-up period 7 years or longer (at least 3 years before and at least 4 years after the date of first macro-DH), and more than nine reliable VF results. Micro-DH was defined as a less than 0.01-mm2 area DH that is undetectable on conventional stereo disc photography (SDP) but can be discriminated by enhanced SDP. SDPs were enhanced by customized image-compensation software. Each enhanced image was evaluated to determine the presence of micro-DH. VF progression was confirmed by standard automated perimetry's guided progression analysis.

Results: Among the 107 POAG eyes with macro-DH, micro-DH was detected prior to macro-DH in 39 (36.4%), the median time lag being 13.6 months. Over the course of the mean 7.1 ± 0.8-year follow-up period, 40 of 107 eyes showed VF progression: 21 (53.8%) of the 39 eyes of the micro-DH positive group and 19 (27.9%) of the 68 eyes of the micro-DH–negative group (P = 0.008). In the micro-DH–positive group, the cumulative VF-progression probability was significantly greater (P = 0.001), and the overall VF-deterioration rate was much faster (−1.01 ± 0.58 vs. −0.78 ± 0.49 dB/year, P = 0.029).

Conclusions: Micro-DH was found prior to macro-DH detection in a significant proportion of POAG patients; micro-DH, moreover, was associated with earlier and faster VF progression.

Optic disc hemorrhage (DH) is among the most important glaucoma-progression risk factors.1,2 In many studies, DH has been strongly associated with glaucomatous deterioration both structurally and functionally.37 Given the close association of DH with progressive glaucomatous damage, accurate assessment of DH presence or absence in glaucomatous eyes is essential to proper management. 
According to evidence uncovered in the Ocular Hypertension Treatment Study, a careful review of stereo disc photography (SDP) can detect DH more sensitively than can clinical examination.8 DH detection on SDP is, compared with clinical examination, less affected by eye movement, thereby allowing clinicians more time for review. Nevertheless, on SDP, some DHs can be missed, either accidentally or inevitably. The reason is that, notwithstanding the continuous improvements in imaging modalities that have enabled better and higher-resolution visualization of the optic nerve head (ONH) surface, red-colored blood vessels and the red-orange-colored retina interfere with DH recognition, especially small-sized DH. 
In the field of neurology, with the improvement of magnetic resonance imaging techniques, interest in small-sized hemorrhage (i.e., cerebral microhemorrhage) has grown.9 Cerebral microhemorrhage is evidence of underlying small-vessel pathology and has been established as being significantly associated with recurrent cerebrovascular diseases, among which are stroke, intracerebral hemorrhage (ICH), and cerebral amyloid angiopathy.10 In cases of ICH, hematoma location has been found to have a positive correlation with baseline-microhemorrhage distribution.11 A prospective study on survivors of ICH or infraction, furthermore, reported that history of microhemorrhage is related to disease recurrence as well as poor prognosis.12 Likewise, in the field of ophthalmology, it might be clinically significant for glaucomatous eyes if optic disc microhemorrhage is detected as a preceding change, prior to detection of macro-DH; as such, it would have a crucial role to play as an early indicator of vulnerability to glaucomatous change. 
We have devised an SDP-compensation technique for enhancement of DH visibility. Applying this method in the present study, we determined the existence of minute-sized DH (i.e., optic disc microhemorrhage [micro-DH]) in primary open-angle glaucoma (POAG) eyes. Additionally, we evaluated the clinical implications of micro-DH detection for subsequent disease progression. 
Methods
This study was approved by the Seoul National University Hospital (SNUH) Institutional Review Board and faithfully adhered to the tenets of the Declaration of Helsinki. All the participants provided their written informed consent prior to their inclusion in the POAG cohort of an ongoing prospective study at SNUH. The methodology of the register has been described in a previous paper.13 
Study Subjects
POAG patients examined between January 2004 and December 2016 and included in the POAG cohort at SNUH's Glaucoma Clinic were considered as study subjects. Patients followed up for a minimum of 7 years at 6-month intervals were consecutively included in this study after a retrospective medical record review in June 2016. All underwent a complete ophthalmic examination including best-corrected visual acuity assessment, refraction, slit-lamp biomicroscopy, gonioscopy, Goldmann applanation tonometry (GAT; Haag-Streit, Koniz, Switzerland), and dilated-funduscopic examination. Additionally, all of the subjects underwent central corneal thickness measurement (Orbscan 73II; Bausch & Lomb Surgical, Rochester, NY, USA), digital color SDP (TRC-50IX; Topcon Corp., Tokyo, Japan), red-free retinal nerve fiber layer (RNFL) photography, spectral-domain optical coherence tomography (SD-OCT, Cirrus HD; Carl Zeiss Meditec, Dublin, CA, USA), and Humphrey VF (HVF) using central 30-2 Swedish interactive threshold algorithm (SITA) standard tests (HFA II; Humphrey Instruments Inc., Dublin, CA, USA). During the follow-up period, serial SDP was acquired at intervals of 6 months or shorter, and HVF examinations were performed at intervals of 6 to 12 months. 
The present study enrolled POAG patients meeting all of the following inclusion criteria: one or more macro-DH on serial SDP; follow-up duration longer than 7 years (at least 3 years before and at least 4 years after the initial macro-DH); more than nine reliable VF results; no systemic diseases such as diabetes mellitus, systemic hypertension, or cardiovascular disease; nonusage of antiplatelet and/or anticoagulation medication. The enrolled subjects all received one or more topical glaucoma medication; in all cases, average IOP was lowered relative to the baseline by at least 20% and maintained without additional laser or glaucoma-surgical intervention. 
The exclusion criteria were as follows: best-corrected visual acuity worse than 20/40 in Snellen equivalent; spherical equivalent less than −6 diopters (D) and more than 3 D, and/or history of optic-neuropathic (other than glaucoma) or retinal disease possibly affecting VF results. 
Discrimination of Macro-DH
SDPs were evaluated for macro-DH retrospectively and independently by three glaucoma specialists (AH, SUB, JWJ) masked to patients' clinical information; any discrepancies were resolved by majority voting. DH was deemed unrelated to glaucoma for one or more of the following reasons: swollen or otherwise evidently abnormal due to nonglaucomatous optic neuropathy; multiple proximitous retinal hemorrhages suggestive either of diabetic retinopathy or retinal vascular abnormality; potentially DH-causative acute posterior vitreous detachment. Patients were enrolled only if they had developed macro-DH during interval testing; in cases of bilateral macro-DH meeting every inclusion criterion, one eye was randomly selected. 
SDP Enhancement
In the present study, SDP was enhanced by means of a postprocessing compensation algorithm customized for that purpose. All image postprocessing was performed using a commercial image processing tool (Adobe Photoshop CS3, version 10.0.1; Adobe, Inc., San Jose, CA, USA) by a single glaucoma image processing specialist (YKK). The algorithm proceeds as follows: first, it reduces media-opacity-related noise (thereby increasing the sharpness of the ONH surface structure and peripapillary area on SDP) by using the Curves tool (specifically by clicking on the image Levels curve and dragging on it); next, it improves the visibility of the DH boundary by improving the contrast between the DH and background fundus, specifically by changing the blood vessel or hemorrhage color from red to bright red and, conversely, the background retina color from red-orange to light brown. See the flowchart for the detailed process of SDP enhancement (Fig. 1) and see also a comparison of an original SDP image with its enhanced version (Fig. 2). 
Figure 1
 
Flow chart of postprocessing compensation algorithm. All image postprocessing was performed using a commercial image processing tool (Adobe, Inc.). First, it reduces media opacity–related noise by using the Curves tool; next, it improves the visibility of the optic disc hemorrhage boundary by improving the contrast between the hemorrhage and background fundus, specifically by changing the blood vessel or hemorrhage color from red to bright red and, conversely, the background retina color from red-orange to light brown.
Figure 1
 
Flow chart of postprocessing compensation algorithm. All image postprocessing was performed using a commercial image processing tool (Adobe, Inc.). First, it reduces media opacity–related noise by using the Curves tool; next, it improves the visibility of the optic disc hemorrhage boundary by improving the contrast between the hemorrhage and background fundus, specifically by changing the blood vessel or hemorrhage color from red to bright red and, conversely, the background retina color from red-orange to light brown.
Figure 2
 
A comparison of an original SDP image with its enhanced version. Original SDP image (left), and SDP image enhanced by customized post-processing compensation algorithm (right). Note that in the enhanced image, the reduced noise and increased contrast improve the visibility of the disc margin and DH.
Figure 2
 
A comparison of an original SDP image with its enhanced version. Original SDP image (left), and SDP image enhanced by customized post-processing compensation algorithm (right). Note that in the enhanced image, the reduced noise and increased contrast improve the visibility of the disc margin and DH.
Definition of Micro-DH
Micro-DH was defined as a minute-sized DH that cannot be detected by original SDP but can be discriminated by enhanced SDP. For distinguishing micro-DH from macro-DH, an optimal cutoff value was determined, as follows: first, 42 cases of small-sized DH detected by enhanced SDP were compiled; then, by a previously reported methodology,14 DH area was measured. In brief, pixel numbers were compared between the disc and DH areas on SD-OCT RNFL deviation map/SDP overlay images using ImageJ software (version 1.45s; Wayne Rasband, http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA), and the formula DH area = (DH area / disc area) ratio × computed SD-OCT disc area was applied. Three glaucoma specialists (AH, SUB, JWJ) independently re-evaluated the original SDP of the 42 cases, as masked to the patient's clinical information. We then determined, as the cutoff size for micro-DH, the area of DH that cannot be discriminated by original SDP but can be discriminated by enhanced SDP. Thus, micro-DH was defined as a minute-sized DH smaller than 0.01-mm2
Discrimination of Micro-DH
For micro-DH discrimination, enhanced SDP images were retrospectively reevaluated independently by three glaucoma specialists (AH, SUB, JWJ) all blinded to the patient's clinical information (including macro-DH location). The enhanced SDP images reviewed had been taken 3 years prior and up to 4 years after the date of first macro-DH detection (Fig. 3). In 78 randomly selected enhanced SDPs, the three observers evaluated images once again in order to calculate interobserver and intraobserver agreement. 
Figure 3
 
Schematic representation of study protocol. The SDP images that were enhanced had been taken 3 years prior and up to 4 years after the date of first optic disc hemorrhage (macro-DH) detection. VF progression was evaluated over the course of the entire 7-year follow up.
Figure 3
 
Schematic representation of study protocol. The SDP images that were enhanced had been taken 3 years prior and up to 4 years after the date of first optic disc hemorrhage (macro-DH) detection. VF progression was evaluated over the course of the entire 7-year follow up.
In cases where enhanced-SDP-detected, small-sized DH had a greater than 0.01 mm2 area, it was considered to be a missed macro-DH. There were three cases of small-sized DH that two glaucoma specialists detected only with enhanced SDPs but with an area greater than 0.01 mm2 area. All three glaucoma specialists agreed that the missed macro-DHs actually were observable in the original SDP, albeit with great caution. 
Spatial Relationship Between Micro-DH and Macro-DH
In order to assess the spatial relationship between macro- and micro-DH, micro-DH–detected, enhanced SDP images were superimposed on macro-DH–detected enhanced SDP images and aligned by imaging software (Adobe, Inc.) according to vascular landmarks. In cases where a micro-DH was within the 0.5 clock-hour (15°) angular extent of each macro-DH margin, it was considered to be in the same location as the macro-DH.15 
Assessment of Progression
Two experienced graders (AH, SUB) masked to the patients' clinical and VF information reviewed each patient's SDP/RNFL photographs. They then determined, based on the baseline photographs, structural progression's presence or absence. Optic disc progression was considered to be the extent of neuroretinal rim thinning. RNFL progression was defined as widening, deepening, or a newly apparent RNFL defect. If the opinions of the two examiners on structural progression differed, a third independent grader evaluated the case. 
For assessment of functional glaucoma progression, the subjects' VF test results were evaluated at the latest visit of the patient's follow-up. VF progression—defined as two or three consecutive VF test results showing at three or more identical test points, a significant decrease from the baseline pattern deviation—was determined by means of event-based analysis using the Humphrey field analyzer with guided progression analysis software. With this software, VF progression is classified as either “possible progression” or “likely progression.” In the present study, “likely progression” only was deemed to be VF progression. For minimization of learning effects, the first VF result was excluded for all patients, as were any results deemed unreliable (fixation loss rate >20% or false-positive and/or false-negative error rates >25%).1618 
Data Analysis
The normally distributed data were compared by independent Student's t-test. The categorical data were analyzed by χ2 or Fisher's exact test. Interobserver and intraobserver agreements were assessed in terms of the proportion of agreement and by the kappa statistic. The kappa statistic measures the agreement of categorical data and corrects for chance agreement. The value of kappa can range from −1 (total disagreement) to +1 (complete agreement). A value of 0 indicates chance agreement. In the present study, we used the interpretation of kappa values proposed by Landis and Koch: a kappa of 0.00 or less is considered poor agreement; 0.01 to 0.20 is slight; 0.21 to 0.40 is fair; 0.41 to 0.60 is moderate; 0.61 to 0.80 is substantial; 0.81 to 1.00 is almost perfect agreement.19 Kaplan-Meier survival analysis and the log-rank test were used to compare, between the groups, structural or functional progression's cumulative risk ratio. The first time at which progression was identified was regarded as the survival analysis endpoint. The end of follow up was the point at which patients without progression were censored. The hazard ratios (HRs) of the associations between the potential risk factors and glaucoma progression were determined by Cox proportional hazards modeling. For each factor, a univariate analysis was performed, and those having a P-value less than 0.1 were included in the multivariate model. The final multivariate model was developed by means of backward elimination, and the adjusted HRs with 95% confidence intervals (CIs) were calculated. Statistical analyses were performed using a commercial statistical package (SPSS 22.0; SPSS, Inc., Chicago, IL, USA). A two-sided P value less than 0.05 was considered to be statistically significant. 
Results
Among our POAG cohort, a total of 311 patients who met all other inclusion criteria but never had macro-DH during 7 years of follow-up were excluded from further analysis, and 107 eyes of 107 POAG patients meeting the inclusion criteria with at least 1 macro-DH were consecutively enrolled. That is, the proportion of experiencing macro-DH in our cohort was 25.6% (107/418) during the follow-up duration of 7 years. 
Among those 107 eyes, micro-DH presented before macro-DH in 39 (36.4%), with a median time lag of 13.6 ± 4.2 months (range: 6.0–24.3 months). Figure 4 provides serial enhanced SDP images of a case with micro-DH. 
Figure 4
 
Representative case of POAG eye with recurrent optic disc hemorrhage (macro-DH) and microhemorrhage (micro-DH). The first row shows original SDP images during the 7-year follow up; the second and third rows show the enhancements of the first-row images. Note that the first macro-DH was detected in 2013 and that on the enhanced SDP images, micro-DH was detected in 2010 and 2012, respectively, at the macro-DH-correspondent location.
Figure 4
 
Representative case of POAG eye with recurrent optic disc hemorrhage (macro-DH) and microhemorrhage (micro-DH). The first row shows original SDP images during the 7-year follow up; the second and third rows show the enhancements of the first-row images. Note that the first macro-DH was detected in 2013 and that on the enhanced SDP images, micro-DH was detected in 2010 and 2012, respectively, at the macro-DH-correspondent location.
The percentage intraobserver agreement on the presence and location of micro-DH ranged from 96.1% to 98.7%, with kappa values ranging from 0.911 to 0.971 (all P < 0.001). Complete agreement between the three observers was obtained in 73 of the 78 cases (93.6%; kappa of 0.788; P < 0.001). Agreement of at least two of three observers was obtained in 76 cases (97.4%). The percentage agreement between pairs of observers ranged from 93.6% to 97.4%, with corresponding kappa values ranging from 0.851 to 0.971 (all P < 0.001). 
Demographic and Clinical Characteristics of Preceding-Micro-DH-Positive and Preceding-Micro-DH-Negative Groups
The demographic and clinical characteristics of the preceding-micro-DH–positive and –negative groups are provided in Table 1. The preceding-micro-DH–positive group had a significantly lower baseline IOP than the preceding-micro-DH–negative group (14.5 ± 3.4 vs. 16.3 ± 3.8 mm Hg, P = 0.016). 
Table 1
 
Comparison of Demographic and Clinical Characteristics Between DH Eyes With and Without Preceding-Micro-DH
Table 1
 
Comparison of Demographic and Clinical Characteristics Between DH Eyes With and Without Preceding-Micro-DH
Distribution of Macro- and Micro-DH According to DH Locations
The DH position was described in terms of the proximal and octant locations (Table 2). The average number of macro-DHs did not show statistically significant differences between the micro-DH–positive and –negative groups (P = 0.357). The frequencies of the proximal and octant locations of macro-DH were similar in the micro-DH–positive and –negative groups (P = 0.911, χ2 test and P = 0.932, Fisher's exact test). The locations of macro-DH did not show significant differences when compared with those of micro-DH in the micro-DH–positive group (P = 0.939, χ2 test and P = 0.848, Fisher's exact test, Fig. 5). 
Table 2
 
Characteristics of DH in Eyes With and Without Preceding-Micro-DH
Table 2
 
Characteristics of DH in Eyes With and Without Preceding-Micro-DH
Figure 5
 
Optic DH diagrams for preceding-micro-DH–positive eyes. Diagram representing macro-DH (left), and diagram of micro-DH (right). Cold colors (blue, green) are employed to represent relatively lower frequency, and warm colors (yellow, red) to represent the higher frequency of hemorrhage.
Figure 5
 
Optic DH diagrams for preceding-micro-DH–positive eyes. Diagram representing macro-DH (left), and diagram of micro-DH (right). Cold colors (blue, green) are employed to represent relatively lower frequency, and warm colors (yellow, red) to represent the higher frequency of hemorrhage.
Comparison of Glaucoma Progression Between Preceding-Micro-DH–Positive and –Negative Groups
Structural progression was observed in 33 eyes (84.6%) of the preceding-micro-DH–positive group (n = 39), and 49 eyes (72.1%) of the preceding-micro-DH–negative group (n = 68; P = 0.140). The cumulative probability of structural nonprogression by Kaplan-Meier analysis was significantly greater in the preceding-micro-DH–negative group (P = 0.023, log-rank test, Fig. 6); the 5-year survival rates for VF progression were 0.21 ± 0.07 in the preceding-micro-DH–positive group and 0.40 ± 0.06 in the –negative group. 
Figure 6
 
Kaplan-Meier curves comparing cumulative nonprogression probability in optic disc hemorrhage (macro-DH) patients with and without preceding microhemorrhage (micro-DH). Patients with detected micro-DH showed, relative to those without micro-DH, a greater cumulative structural and VF progression probability (P = 0.023 and P = 0.001, log-rank test). The tables below the Kaplan-Meier curve indicate the numbers at risk at specific time points.
Figure 6
 
Kaplan-Meier curves comparing cumulative nonprogression probability in optic disc hemorrhage (macro-DH) patients with and without preceding microhemorrhage (micro-DH). Patients with detected micro-DH showed, relative to those without micro-DH, a greater cumulative structural and VF progression probability (P = 0.023 and P = 0.001, log-rank test). The tables below the Kaplan-Meier curve indicate the numbers at risk at specific time points.
Among the preceding-micro-DH–positive group eyes (n = 39), VF progression was observed in 21 (53.8%), and among the preceding-micro-DH–negative eyes (n = 68), in 19 (27.9%; P = 0.008). The cumulative probability of VF progression by Kaplan-Meier analysis was significantly greater in the preceding-micro-DH–positive group (P = 0.001, log-rank test, Fig. 6); the 5-year survival rates for VF progression were 0.54 ± 0.08 in the preceding-micro-DH–positive group and 0.73 ± 0.06 in the negative group. See Supplementary Figure S1 and S2 for Humphrey field analyzer's mean deviation slope comparison and results of the subgroup analysis. 
Factors Associated With VF Progression in POAG Patients With Macro-DH
According to the univariate Cox proportional hazards model, four factors were associated with VF progression, namely: a greater number of macro-DH and micro-DH, lower baseline IOP, and a smaller-percentage reduction of IOP. The multivariate Cox proportional hazards model revealed that a greater number of preceding-micro-DH (P = 0.023) and a smaller-percentage reduction in IOP (P = 0.037) were significantly associated with VF progression. The full statistical results including HRs and CIs are summarized in Table 3
Table 3
 
Univariate and Multivariate HRs and 95% CIs of Significant Risk Factors for VF Deterioration in Glaucoma Eyes With DH
Table 3
 
Univariate and Multivariate HRs and 95% CIs of Significant Risk Factors for VF Deterioration in Glaucoma Eyes With DH
Micro-DH Evaluation After Initial Macro-DH
To compare the difference between the occurrence of micro-DH before and after macro-DH, enhanced SDPs, which had been taken up to 4 years after the date of initial macro-DH, were evaluated. Among the 107 study eyes, during the 4 years of follow up from the date of initial macro-DH, a total of 5 (4.7%) showed micro-DH in the corresponding location of initial macro-DH. 
Discussion
In this study, micro-DH, as detected prior to macro-DH, was confirmed for more than one-third of POAG eyes; also, the presence of preceding-micro-DH was associated with subsequent VF progression that was both earlier and faster. To our knowledge, this is the first report demonstrating, based on longitudinal observational study results, a precedence effect of micro-DH on glaucomatous DH. 
The underlying mechanisms of DH remain unclear. However, given its strong association with glaucoma progression as well as high spatial correlation with structural alteration, the DH location on the ONH and peripapillary RNFL is considered to be the site of progressive microvascular change that is related to accelerated glaucomatous damage.20,21 Further, Radcliffe et al.22 showed that structural glaucomatous progression was not significantly different before and after the event of a DH, suggesting ongoing structural damage prior to the DH. In terms of functional progression, De Moraes et al.23 found that even before DH detection, localized VF damage had begun in corresponding regions. These findings suggest that fast disease progression predisposes eyes to DH occurrence. 
On enhanced SDP images, micro-DH existed prior to conventional-SDP-detected-DH (i.e., macro-DH) in a significant percentage of POAG eyes. The median time lag between micro-DH and macro-DH was 13.6 ± 4.24 months. When comparing macro-DH eyes with and without preceding micro-DH respectively, those with micro-DH showed, on average, detection of VF progression 26 months earlier. Based on these findings, we tentatively conclude that significant ONH and peripapillary RNFL change had initiated several months or about 1 year before macro-DH detection and micro-DH should be regarded as a disease-progression predictor that is clinically meaningful. 
From another perspective, however, micro-DH can be interpreted as a residue of macro-DH re-absorption in the course of macro-DH recurrence. That is, macro-DH that has almost disappeared can be detected, and identified, as “micro-DH.” Previous studies have found that glaucomatous eyes with recurrent DH show much faster structural progression and more significant VF deterioration than cases of solitary (non-recurrent) DH.24,25 In our study, a total of 5 eyes of 107 (4.7%) showed micro-DH after the initial macro-DH. It is difficult to determine, based on our data, if micro-DH is a residual hemorrhage or a preceding change of macro-DH. However, considering the much lower incidence of micro-DH after macro-DH, micro-DH seems to be a preceding change of macro-DH. The difference in the duration of hemorrhage would be one possible explanation for this result. Micro-DH might be caused by persistent leaking in vulnerable points, which can be smaller in size but longer in duration. Meanwhile, if the residual hemorrhage after macro-DH takes only a few days from 0.01 mm2 size to complete absorption, it is less likely to be found as a micro-DH. Thus, although we cannot establish whether micro-DH should be regarded as an event separate from recurrent DH or a premonitory sign of it, it seems clear that in either case, micro-DH can be considered to be a sign that both appears earlier than macro-DH and is associated with subsequent glaucoma progression. 
Interestingly, preceding-micro-DH–positive eyes had a baseline IOP significantly lower than that of the preceding-micro-DH–negative group. In a prospective study comparing low- and high-tension glaucoma groups, Kitazawa et al.26 found that DH was four times more prevalent in the former than in the latter; other studies have reported similarly high DH prevalence in low-tension glaucoma.27,28 Furthermore, DH recurrence has been shown to be more common in low-tension glaucoma.26,29 All of the findings noted above tend to support the assertion that, in lower baseline IOP POAG eyes, microvascular abnormality might have a significant role to play as a common pathology in glaucoma progression.30 
Glaucoma progression is a complex process affected by multiple factors; however, because IOP is the only treatable risk factor, its sufficient lowering is essential for patients who are likely to progress. Akagi et al.5 demonstrated a beneficial effect of treatment intensification in reducing the post-DH rate of RNFL thinning. In terms of post-DH functional progression, Medeiros et al.31 showed additional IOP lowering to be associated with VF progression rate reduction. Considering the earlier VF deterioration in patients who were preceding-micro-DH positive, identification of micro-DH might indeed be of clinical significance, especially for timely management of glaucoma. 
Study Limitations
First, because we included POAG eyes with confirmed macro-DH, it was not possible, in this study, to determine if micro-DH can occur in either glaucomatous eyes with no history of macro-DH or normal eyes. If micro-DH exists without macro-DH, it would also be important to evaluate its association with glaucoma progression. Second, it is possible that some cases of micro-DH were either missed or, due to the limited resolution of the enhanced SDP images, undetectable. It is also possible that patients whose micro-DH was missed were included in the micro-DH–negative group, since SDPs were recorded at 6-month intervals. However, misclassification of micro-DH patients as “non-micro-DH” would in fact bias the results toward the null. Thus, this study's finding of a statistically significant intergroup difference might indicate an even stronger association between micro-DH and glaucoma progression. Nonetheless, it would be important to evaluate the clinical implications of micro-DH in further research with more frequent SDPs. Third, glaucoma medication type might also affect the frequency of micro-DH. The Low-Pressure Glaucoma Treatment Study reported that there were more DH events in eyes with timolol compared with eyes with brimonidine, though the difference was not statistically significant.30 Among our study subjects, some also used medications other than timolol and brimonidine, and in many cases, medications were changed over the course of the follow up. Thus, it was difficult to assess the effect of glaucoma medication on the frequency of micro-DH. Further research with larger patient cohorts would be expected to obtain fuller information on the effects of glaucoma medications on micro-DH development. Finally, whereas we studied a group of mostly low-baseline-IOP POAG eyes (93.2% of the subjects had a baseline IOP ≤ 21 mm Hg), DH's topographic characteristics reportedly differ between normal- and high-baseline IOP POAG eyes.14 As such, our results actually might not be directly applicable to other populations of glaucoma patients. 
Conclusions
In a significant percentage of POAG eyes with macro-DH, micro-DH was detected, and the presence of preceding-micro-DH was associated with both earlier and faster progression. Therefore, micro-DH can be considered to be a predictive sign of ongoing, or accelerated, glaucomatous damage. Further studies investigating the possibility of a treatment-reinforcement benefit for POAG eyes with micro-DH are warranted. 
Acknowledgments
Disclosure: A. Ha, None; Y.K. Kim, None; S.U. Baek, None; K.H. Park, None; J.W. Jeoung, None 
References
Bengtsson B, Leske MC, Yang Z, Heijl A; Early Manifest Glaucoma Trial Group. Disc hemorrhages and treatment in the early manifest glaucoma trial. Ophthalmology. 2008; 115: 2044–2048.
Coleman AL, Miglior S. Risk factors for glaucoma onset and progression. Surv Ophthalmol. 2008; 53: S3–S10.
Nitta K, Sugiyama K, Higashide T, Ohkubo S, Tanahashi T, Kitazawa Y. Does the enlargement of retinal nerve fiber layer defects relate to disc hemorrhage or progressive visual field loss in normal-tension glaucoma? J Glaucoma. 2011; 20: 189–195.
Ahn JK, Park KH. Morphometric change analysis of the optic nerve head in unilateral disk hemorrhage cases. Am J Ophthalmol. 2002; 134: 920–922.
Akagi T, Zangwill LM, Saunders LJ, et al. Rates of local retinal nerve fiber layer thinning before and after disc hemorrhage in glaucoma. Ophthalmology. 2017; 124: 1403–1411.
Kim JM, Kyung H, Azarbod P, Lee JM, Caprioli J. Disc haemorrhage is associated with the fast component, but not the slow component, of visual field decay rate in glaucoma. Br J Ophthalmol. 2014; 98: 1555–1559.
Drance S, Anderson DR, Schulzer M; Collaborative Normal-Tension Glaucoma Study Group. Risk factors for progression of visual field abnormalities in normal-tension glaucoma. Am J Ophthalmol. 2001; 131: 699–708.
Budenz DL, Anderson DR, Feuer WJ, et al. Detection and prognostic significance of optic disc hemorrhages during the ocular hypertension treatment study. Ophthalmology. 2006; 113: 2137–2143.
Greenberg SM, Vernooij MW, Cordonnier C, et al. Cerebral microbleeds: a guide to detection and interpretation. Lancet Neurol. 2009; 8: 165–174.
Viswanathan A, Chabriat H. Cerebral microhemorrhage. Stroke. 2006; 37: 550–505.
Lee S, Bae H, Kwon S, et al. Cerebral microbleeds are regionally associated with intracerebral hemorrhage. Neurology. 2004; 62: 72–76.
Yates PA, Villemagne VL, Ellis KA, et al. Cerebral microbleeds: a review of clinical, genetic, and neuroimaging associations. Front Neurol. 2014; 4: 205.
Ha A, Kim YK, Jeoung JW, et al. Association of angle width with progression of normal-tension glaucoma: a minimum 7-year follow-up study. JAMA Ophthalmol. 2019; 137: 13–20.
Kim YK, Park KH, Yoo BW, Kim HC. Topographic characteristics of optic disc hemorrhage in primary open-angle glaucoma. Invest Ophthalmol Vis Sci. 2014; 55: 169–176.
Kim YK, Park KH. Lamina cribrosa defects in eyes with glaucomatous disc haemorrhage. Acta Ophthalmol. 2016; 94: e468–e473.
Birt CM, Shin DH, Samudrala V, et al. Analysis of reliability indices from Humphrey visual field tests in an urban glaucoma population. Ophthalmology. 1997; 104: 1126–1130.
Lee WJ, Baek SU, Kim YK, et al. Rates of ganglion cell-inner plexiform layer thinning in normal, open-angle glaucoma and pseudoexfoliation glaucoma eyes: a trend-based analysis. Invest Ophthalmol Vis Sci. 2019; 60: 599–604.
Lee EJ, Kim T-W, Kim J-A, Kim J-A. Central visual field damage and parapapillary choroidal microvasculature dropout in primary open-angle glaucoma. Ophthalmology. 2018; 125: 588–596.
Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977; 33: 159–174.
Quigley HA, Addicks EM, Green WR, Maumenee AE. Optic nerve damage in human glaucoma II. The site of injury and susceptibility to damage. Arch Ophthalmol. 1981; 99: 635–649.
Lee EJ, Kim T-W, Kim M, Weinreb RN. Recent structural alteration of the peripheral lamina cribrosa near the location of disc hemorrhage in glaucoma recent alteration of LC and disc hemorrhage. Invest Ophthalmol Vis Sci. 2014; 55: 2805–2815.
Radcliffe NM, Chung E, Demetriades AM, Christos PJ. Structural glaucomatous progression before and after occurrence of an optic disc haemorrhage. Br J Ophthalmol. 2015; 99: 21–25.
De Moraes CG, Prata TS, Liebmann CA, Tello C, Ritch R, Liebmann JM. Spatially consistent, localized visual field loss before and after disc hemorrhage. Invest Ophthalmol Vis Sci. 2009; 50: 4727–4733.
Ishida K, Yamamoto T, Sugiyama K, Kitazawa Y. Disk hemorrhage is a significantly negative prognostic factor in normal-tension glaucoma. Am J Ophthalmol. 2000; 129: 707–714.
Kim SH, Park KH. The relationship between recurrent optic disc hemorrhage and glaucoma progression. Ophthalmology. 2006; 113: 598–602.
Kitazawa Y, Shirato S, Yamamoto T. Optic disc hemorrhage in low-tension glaucoma. Ophthalmology. 1986; 93: 853–857.
Gloster J. Incidence of optic disc haemorrhages in chronic simple glaucoma and ocular hypertension. Br J Ophthalmol. 1981; 65: 452–456.
Susanna R, Basseto FL. Hemorrhage of the optic disc and neurosensorial dysacousia. J Glaucoma. 1992; 1: 248–253.
Hendrickx KH, van den Enden A, Rasker MT, Hoyng PF. Cumulative incidence of patients with disc hemorrhages in glaucoma and the effect of therapy. Ophthalmology. 1994; 101: 1165–1172.
Furlanetto RL, De Moraes CG, Teng CC, et al. Risk factors for optic disc hemorrhage in the low-pressure glaucoma treatment study. Am J Ophthalmol. 2014; 157: 945–952.
Medeiros FA, Alencar LM, Sample PA, et al. The relationship between intraocular pressure reduction and rates of progressive visual field loss in eyes with optic disc hemorrhage. Ophthalmology. 2010; 117: 2061–2066.
Figure 1
 
Flow chart of postprocessing compensation algorithm. All image postprocessing was performed using a commercial image processing tool (Adobe, Inc.). First, it reduces media opacity–related noise by using the Curves tool; next, it improves the visibility of the optic disc hemorrhage boundary by improving the contrast between the hemorrhage and background fundus, specifically by changing the blood vessel or hemorrhage color from red to bright red and, conversely, the background retina color from red-orange to light brown.
Figure 1
 
Flow chart of postprocessing compensation algorithm. All image postprocessing was performed using a commercial image processing tool (Adobe, Inc.). First, it reduces media opacity–related noise by using the Curves tool; next, it improves the visibility of the optic disc hemorrhage boundary by improving the contrast between the hemorrhage and background fundus, specifically by changing the blood vessel or hemorrhage color from red to bright red and, conversely, the background retina color from red-orange to light brown.
Figure 2
 
A comparison of an original SDP image with its enhanced version. Original SDP image (left), and SDP image enhanced by customized post-processing compensation algorithm (right). Note that in the enhanced image, the reduced noise and increased contrast improve the visibility of the disc margin and DH.
Figure 2
 
A comparison of an original SDP image with its enhanced version. Original SDP image (left), and SDP image enhanced by customized post-processing compensation algorithm (right). Note that in the enhanced image, the reduced noise and increased contrast improve the visibility of the disc margin and DH.
Figure 3
 
Schematic representation of study protocol. The SDP images that were enhanced had been taken 3 years prior and up to 4 years after the date of first optic disc hemorrhage (macro-DH) detection. VF progression was evaluated over the course of the entire 7-year follow up.
Figure 3
 
Schematic representation of study protocol. The SDP images that were enhanced had been taken 3 years prior and up to 4 years after the date of first optic disc hemorrhage (macro-DH) detection. VF progression was evaluated over the course of the entire 7-year follow up.
Figure 4
 
Representative case of POAG eye with recurrent optic disc hemorrhage (macro-DH) and microhemorrhage (micro-DH). The first row shows original SDP images during the 7-year follow up; the second and third rows show the enhancements of the first-row images. Note that the first macro-DH was detected in 2013 and that on the enhanced SDP images, micro-DH was detected in 2010 and 2012, respectively, at the macro-DH-correspondent location.
Figure 4
 
Representative case of POAG eye with recurrent optic disc hemorrhage (macro-DH) and microhemorrhage (micro-DH). The first row shows original SDP images during the 7-year follow up; the second and third rows show the enhancements of the first-row images. Note that the first macro-DH was detected in 2013 and that on the enhanced SDP images, micro-DH was detected in 2010 and 2012, respectively, at the macro-DH-correspondent location.
Figure 5
 
Optic DH diagrams for preceding-micro-DH–positive eyes. Diagram representing macro-DH (left), and diagram of micro-DH (right). Cold colors (blue, green) are employed to represent relatively lower frequency, and warm colors (yellow, red) to represent the higher frequency of hemorrhage.
Figure 5
 
Optic DH diagrams for preceding-micro-DH–positive eyes. Diagram representing macro-DH (left), and diagram of micro-DH (right). Cold colors (blue, green) are employed to represent relatively lower frequency, and warm colors (yellow, red) to represent the higher frequency of hemorrhage.
Figure 6
 
Kaplan-Meier curves comparing cumulative nonprogression probability in optic disc hemorrhage (macro-DH) patients with and without preceding microhemorrhage (micro-DH). Patients with detected micro-DH showed, relative to those without micro-DH, a greater cumulative structural and VF progression probability (P = 0.023 and P = 0.001, log-rank test). The tables below the Kaplan-Meier curve indicate the numbers at risk at specific time points.
Figure 6
 
Kaplan-Meier curves comparing cumulative nonprogression probability in optic disc hemorrhage (macro-DH) patients with and without preceding microhemorrhage (micro-DH). Patients with detected micro-DH showed, relative to those without micro-DH, a greater cumulative structural and VF progression probability (P = 0.023 and P = 0.001, log-rank test). The tables below the Kaplan-Meier curve indicate the numbers at risk at specific time points.
Table 1
 
Comparison of Demographic and Clinical Characteristics Between DH Eyes With and Without Preceding-Micro-DH
Table 1
 
Comparison of Demographic and Clinical Characteristics Between DH Eyes With and Without Preceding-Micro-DH
Table 2
 
Characteristics of DH in Eyes With and Without Preceding-Micro-DH
Table 2
 
Characteristics of DH in Eyes With and Without Preceding-Micro-DH
Table 3
 
Univariate and Multivariate HRs and 95% CIs of Significant Risk Factors for VF Deterioration in Glaucoma Eyes With DH
Table 3
 
Univariate and Multivariate HRs and 95% CIs of Significant Risk Factors for VF Deterioration in Glaucoma Eyes With DH
Supplement 1
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×