Optic Disc Hemorrhage
Disc hemorrhages are characteristic linear hemorrhages perpendicular to the optic disc, most commonly on the superotemporal or inferotemporal margin. The cause of disc hemorrhages has not been fully characterized, and mechanical and vascular hypotheses with evidence supporting both have been proposed. Although disc hemorrhages can occur in eyes with no signs of glaucoma, in glaucomatous eyes, disc hemorrhages are a risk factor for progression. Disc hemorrhages themselves cannot be treated but their presence may warrant intensification of intraocular pressure (IOP) lowering therapy in glaucomatous eyes. The purpose of this article is to provide a review of disc hemorrhage for the trainee and practicing ophthalmologist.
Figure 1. A selection of typical appearing disc hemorrhages.
Alternate names: optic disc hemorrhage, Drance hemorrhage, splinter hemorrhage, nerve fiber layer hemorrhage, optic nerve head hemorrhage.
Disease: Disc Hemorrhages
Disc hemorrhages are splinter or flame-shaped hemorrhages oriented perpendicular to the optic disc margin (Figure 1). Classically, these hemorrhages are located in the prelaminar optic disc, cross the peripapillary zone, and extend into the adjacent superficial retinal nerve fiber layer, although they may not occupy the entire length from disc to retina. Alternately, deeper disc hemorrhages may appear round and blotchy. Less commonly, a disc hemorrhage may be noted in the peripapillary retinal nerve fiber layer reaching within one disc diameter of the optic disc margin.
Etiology and Pathophysiology
Mechanical and Vascular Hypotheses
Although disc hemorrhages have been the subject of tremendous research, the exact mechanism by which they appear remains elusive. Two primary hypotheses exist: mechanical and vascular.
Proponents of the mechanical theory hypothesize that disc hemorrhages result from mechanical shearing at the lamina cribrosa5 or because of damage to the capillary network at the border of retinal nerve fiber layer defect enlargement.6 Essentially, this theory suggests that the primary insult is neurodegenerative, and the hemorrhage is a secondary event resulting from tissue damage. Other authors suggest various vascular etiologies,7–11 for example, ischemic microinfarction in the optic nerve head7 or perturbation of the blood-retinal barrier.10,11 In this theory, a yet unknown primary vascular problem increases tissue susceptibility to damage.
Given the presence of disc hemorrhages in glaucomatous and non-glaucomatous eyes,4,12 it is probable that multiple factors may contribute to their development. The etiology of optic disc hemorrhages remains an area of active research.
The Relationship to Rim Notches
A neural rim notch is a localized divot in the neuroretinal rim. In their study of 33 eyes for which optic disc images prior to development of a disc hemorrhage were available for review, Law and colleagues noted that all eyes with preexisting neural rim notches had a subsequent disc hemorrhage at or adjacent to the notches.13 Their observation that a rim notch may precede disc hemorrhage (on average, by 21.5 months) and that disc hemorrhages occurred at or adjacent to rim notches13 differed from older work3,14 that suggested disc hemorrhages precede notches. That disc hemorrhages occur at or near preexisting rim notches lends evidence to the theory that glaucomatous damage begins prior to the appearance of a hemorrhage.
Several population-based studies12,15–20 have reported disc hemorrhage prevalence estimates ranging from 0.6%15,16 to 1.4%.12 The Beaver Dam Eye Study and Blue Mountains Eye Study, both of which had similar methodologies, reported overall disc hemorrhage prevalence of 0.9%17 and 1.4%,12 respectively. Disc hemorrhage prevalence in the first survey of residents of Dalby, Sweden was 0.8%,19 and in repeated surveys of Dalby residents, prevalence was 0.7%.18 Two large scale eye disease screening projects on healthy Japanese adults each revealed a prevalence of 0.6%.15,16 The Beijing Eye Study, a population-based survey in Northern China, reported a prevalence of 1.2%.20 Although the observation of a disc hemorrhage should prompt a thorough investigation for glaucoma, the overall low prevalence of disc hemorrhages in population-based studies limits their usefulness as an effective screening tool for glaucoma.2,21
As with the overall prevalence of disc hemorrhage, the reported prevalence of disc hemorrhage in different types of glaucoma also varies. Using the Blue Mountain Eye Study as an example, disc hemorrhages were present in 13.8% of participants with open-angle glaucoma (OAG; 8% of patients with high-pressure glaucoma and 25% of patients with low-pressure glaucoma), 1.5% of patients with ocular hypertension, and 1% of normals.12
Most other reports of disc hemorrhage prevalence are subject to selection bias because they have reported on subjects with ocular hypertension, suspected glaucoma, or diagnosed glaucoma often selected from clinical or hospital populations. In studies of subjects with open angle glaucoma,14,19,22–24 for example, disc hemorrhage prevalence ranges from 4.7%22 to 58.3%.19
Overall, disc hemorrhages have been observed more frequently in early compared with advanced glaucoma25,26 and in patients with normal-tension rather than high-pressure glaucoma.12,23,27
Multiple studies have identified both local and systemic associations with disc hemorrhages.
Related to the eye, increased vertical cup to disc ratio,12,22,28 relatively low IOP,3,23,29 nerve fiber layer atrophy,30 worsening visual fields,31–34 and multiple other ocular variables are associated with disc hemorrhages.
An association between peripapillary atrophy, specifically beta zone, and disc hemorrhage was demonstrated in a cohort of mostly normal-tension glaucoma patients with unilateral disc hemorrhage,35 and a large scale study of apparently healthy subjects revealed the prevalence of peripapillary atrophy (both alpha and beta zone) was significantly greater in eyes with disc hemorrhage compared to those without disc hemorrhage.15 However, in a study of bilateral OAG eyes, the eye with a disc hemorrhage and the contralateral eye did not vary with respect to the size of the alpha or beta zone peripapillary atrophy.36
A selection of systemic associations and authors who noted these associations are shown in Table 1.
Physical Examination and Clinical Diagnosis
Clinical examination of the optic disc and peripapillary retina is performed with a high magnification convex lens and slit lamp biomicroscope. The findings should be described and the location of a disc hemorrhage should be drawn and photographed.2 Interestingly, in the Ocular Hypertensive Treatment Study (OHTS), significantly more disc hemorrhages (84%) were detected on annual photographic review at a dedicated reading center than were detected on clinical exam and photographic review (16%),4 a finding that emphasizes the challenge of identifying disc hemorrhages on exam alone.
At present, optic disc imaging techniques are unable to identify disc hemorrhages, emphasizing the importance of imaging as an adjunct to, rather than a replacement for, clinical examination.
Classically, disc hemorrhages associated with glaucoma are found mostly in the inferotemporal and superotemporal regions of the optic disc (Figure 2).16,29,37,38 The observation that disc hemorrhages occur in the area of the optic disc where glaucoma damage occurs supports a relationship between disc hemorrhages and glaucomatous optic neuropathy.16,29
Duration and Recurrence
Weekly photography has demonstrated that optic disc hemorrhages persist for 2 to 35 weeks and on average are present for about 11 weeks.23 A recurrent disc hemorrhage was observed in 64% of eyes, 92% of which occurred within 28 weeks of the prior hemorrhage.23
As noted above (link), documentation of a disc hemorrhage is accomplished both by drawing the location of the hemorrhage with respect to the optic nerve and taking disc photographs. Because current optic disc imaging techniques do not capture disc hemorrhages, clinical examination remains an essential component of diagnosis.
Although disc hemorrhages are strongly associated with glaucoma, there are other causes. Textbooks,2,39 reviews,1 articles,4 and letters21 cite diabetes mellitus, optic disc drusen, ischemic optic neuropathies, vascular diseases of the retina, systemic hypertension, leukemia, and systemic lupus erythematosus among the possible causes of linear hemorrhages at the optic disc, although little primary literature addresses these associations. Posterior vitreous detachment can also cause optic disc hemorrhage.40,41 A thorough history and close evaluation for evidence of nonglaucomatous optic neuropathies, disc edema, retinal abnormalities, or retinal vasculature changes may help to distinguish these entities as a potential cause of disc hemorrhage.
Whether disc hemorrhages can occur in normal eyes is discussed below (link).
The observation of a disc hemorrhage should prompt a thorough investigation for glaucoma, and individuals with disc hemorrhages should be considered glaucoma suspects.3 Early identification of glaucoma and subsequent significant IOP reduction is the only treatment at present to prevent vision loss, perimetric progression, and changes to the disc and retinal nerve fiber layer.
In individuals with known glaucoma, a disc hemorrhage may be a sign of active disease and progression (link to prognosis). Thus, while a disc hemorrhage does not need to be treated, per se, its presence may signal the need to initiate or intensify IOP-lowering therapy.
Whether IOP reduction lowers the frequency of disc hemorrhages has not been fully established.4,42–44 Some studies suggest a decrease in disc hemorrhage frequency with significant IOP reduction,43,44 which may be more readily achieved in primary open angle (POAG) and glaucoma suspects compared with normal tension glaucoma (NTG) eyes.44 In contrast, assignment to the treatment arm of the OHTS study reduced the incidence of developing POAG but did not significantly reduce the incidence of disc hemorrhages, although the authors suggest this finding may have been due to the relatively small number of disc hemorrhages observed.4
Numerous studies show disc hemorrhages to be an important risk factor in the development and progression of glaucoma.
Development of Glaucoma
The eight year cumulative incidence of POAG in the OHTS was 13.6% in eyes with disc hemorrhages, compared with 5.2% in eyes without disc hemorrhage.4 In disc hemorrhage eyes in which the end point of POAG developed, the median time between disc hemorrhage and POAG was 13 months.4 Because the presence of a disc hemorrhage at baseline excluded potential subjects from inclusion in OHTS, the risk of POAG associated with disc hemorrhages may have been underestimated.4 In the Collaborative Normal-Tension Glaucoma Study (CNTGS), a patient with a disc hemorrhage at enrollment had 2.72 times the change of developing a perimetric endpoint.33
Do all individuals with disc hemorrhage develop glaucoma?
Although many authors and glaucoma specialists have developed opinions as to whether nonglaucomatous eyes can have disc hemorrhages, there is no clear consensus in the literature. In the Blue Mountain Eye Study, 70% of disc hemorrhages were observed in eyes without glaucoma (diagnosed by perimetric criteria in the presence of optic disc criteria).12 Similar diagnostic criteria were used in the Beaver Dam Eye Study, and 96% of disc hemorrhages were observed in eyes without glaucoma.17 In the OHTS, 86.7% of subjects with a disc hemorrhage did not develop a POAG endpoint in a median follow up 30.7 months after disc hemorrhage occurrence.4
In contrast, two studies found no disc hemorrhages in healthy eyes; they were observed only in suspect22 and glaucomatous eyes.22,25 Kitazawa and colleagues found that only 0.4% of normal patients observed consecutively at the University of Tokyo had a disc hemorrhage; a much higher prevalence of hemorrhage was observed in patients with low tension glaucoma and POAG (20.5% and 4.2%, respectively).23
Some of the variability between these studies is likely due to their definitions of glaucoma and length of follow up. Although diagnostic criteria for glaucoma may not be met at the time of disc hemorrhage observation, many eyes may eventually progress to glaucoma. As emphasized elsewhere (link), the presence of a disc hemorrhage should prompt a thorough investigation for glaucoma.
Progression of Glaucoma
In patients with glaucoma, studies support associations between disc hemorrhages and perimetric progression,31–34 optic disc changes,34 retinal ganglion cell loss,45 and retinal nerve fiber layer thinning.30
Disc hemorrhage at the time of study enrollment was a risk factor for perimetric progression in the CNTGS.33 In fact, a recent review of 85 articles investigating prognostic factors for visual field progression in patients with OAG found a clear association between visual field progression and disc hemorrhage in patients with NTG.31 The association between disc hemorrhage and perimetric progression is likely similar in the various glaucomas. For example, Rasker and colleagues observed an association between visual field deterioration and disc hemorrhage in eyes with normal pressure glaucoma, POAG, and ocular hypertension.32 In the Early Manifest Glaucoma Trial, as the percentage of patient visits with disc hemorrhages increased, as noted by clinical assessment rather than photographic review, the risk of perimetric and photographic optic disc criteria progression increased.34
A longitudinal cohort of eyes with glaucoma recruited from the Diagnostic Innovations Glaucoma Study demonstrated significantly faster rates of estimated retinal ganglion cell loss in eyes with disc hemorrhages compared to glaucomatous eyes without disc hemorrhages over an average study period of nearly four years.45 Similarly, recent spectral-domain optical coherence tomography research revealed an association between disc hemorrhage and progressive retinal nerve fiber layer thinning in glaucomatous eyes.30 Figure 3 demonstrates inferotemporal focal retinal nerve fiber layer thinning at 18 months following a disc hemorrhage in the same location.
Effect of Treatment
Interesting, in the OHTS, assignment to the treatment arm of the study reduced the incidence of developing POAG but did not significantly reduce the incidence of disc hemorrhages, although the authors suggest that this finding may have been due to the relatively small number of disc hemorrhages observed.4
Exfoliation glaucoma (XFG), which is the most common type of secondary open angle glaucoma,46 tends to progress more quickly and become more severe compared to POAG.46–49 Disc hemorrhage has been shown to be an independent risk factor for progression of XFG.50
Recurrent Disc Hemorrhages: The two-population hypothesis
As noted above (link), a weekly photography study demonstrated recurrent disc hemorrhage in 64% of eyes, and 72% of recurrent hemorrhages occurred in the same optic disc quadrant.23 Similarly, Ishida and colleagues observed recurrent disc hemorrhages in 71.9% of subjects in their cohort of untreated NTG patients, and 78.2% of recurrences were in the same quadrant.38 Interestingly, they observed that eyes with recurrent disc hemorrhages were significantly more likely to demonstrate visual field progression compared to eyes with only a single observed disc hemorrhage (mean follow up, 5.6 years).38 The CNTGS also found a higher frequency of disc hemorrhages in patients whose glaucoma progressed even when adjusted for follow up time.33
Observations like these have led some authors to propose two populations of glaucoma patients: bleeders and nonbleeders.23,38,43,44,51 Some eyes are suggested to be particularly predisposed to disc hemorrhage even with adequate IOP-lowering therapy, and this may be especially true of eyes with NTG.43 The significance of the two population hypothesis with respect to progression and treatment has not been fully established.
Additional research is needed to determine exactly why disc hemorrhages occur and to clarify why only some glaucoma patients develop disc hemorrhages so we may better understand the associated implications for prognosis and management.
- Uhler TA, Piltz-Seymour J. Optic disc hemorrhages in glaucoma and ocular hypertension: implications and recommendations. Curr Opin Ophthalmol. 2008;19(2):89–94.
- Schacknow PN, Samples JR eds. The glaucoma book: a practical, evidence-based approach to patient care. New York: Springer; 2010.
- Drance SM. Disc hemorrhages in the glaucomas. Surv Ophthalmol. 1989;33(5):331–337.
- 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(12):2137–2143.