- Email: [email protected]
Factors Prognostic of Visual Outcome in Patients With Subretinal Hemorrhage
Factors Prognostic of Visual Outcome in Patients With Subretinal Hemorrhage Steven R. Bennett, M.D., James c. Folk, M.D., Christopher F. Blodi, M.D., and Marie Klugman, Ph.D. We reviewed the charts of 29 patients with large subretinal hemorrhages involving the center of the fovea to evaluate factors that might be prognostic of visual outcome. The average final visual acuity was 20/480 with a mean follow-up of three years. Patients with thick hemorrhages (causing an obvious elevation of the fovea) had worse final visual acuity than patients with thin hemorrhages (P = .02). The diameter of the hemorrhage was not a significant predictor of outcome. Patients with aging macular degeneration had poorer final visual acuity (mean, 20/1,700, P .002), and patients with choroidal ruptures had better final visual acuity, (mean 20/35, P < .001) than the remainder of the patients. We found that the presence of aging macular degeneration was a more important predictor of the outcome of legal blindness than the thickness of the hemorrhage (P = .03). Although the prognosis in patients with subfoveal blood is generally poor, some patients have excellent return of vision.
=
RECENT ADVANCES in vitreoretinal surgical techniques have made it possible to remove blood from the subretinal space in the macular area.!" Some patients who have undergone surgery to remove subretinal blood showed im-
Accepted for publication Oct. 19, 1989. From the Department of Ophthalmology, University of Iowa, Iowa City, Iowa (Drs. Bennett, Folk, and Blodi) and the College of Business Administration, Drake University, Des Moines, Iowa (Dr. Klugman). This study was presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Sarasota, Florida, May 2, 1989. This study was supported in part by an unrestricted grant from Research to Prevent Blindness, Inc., New York, New York, and the Retina Research Fund of the University of Iowa. Reprint requests to Christopher F. Blodi. M.D., Department of Ophthalmology, University of Iowa Hospitals, Iowa City, IA 52242.
©AMER:ICAN JOURNAL OF OPHTHALMOLOGY
109:33-37,
proved vision.v" An experimental model of subretinal hemorrhage showed that photoreceptor damage occurs within 24 hours." Perhaps early removal of blood could prevent the visual loss associated with this condition. To determine which patients might benefit from surgical removal of blood from the subretinal space, one must understand the natural history of the condition. Factors that might influence prognosis, such as the cause and size of the hemorrhage, have not been systematically evaluated. A common cause of subretinal hemorrhage is choroidal neovascularization associated with aging macular degeneration. This neovascularization has a poor visual prognosis. 6 . 7 Therefore, an eye that has a subfoveal hemorrhage associated with this type of neovascularization may also fare poorly. A similar hemorrhage caused by another process, such as trauma, might have a different prognosis. We undertook a retrospective study of subretinal hemorrhages of at least one disk diameter in size involving the center of the fovea to evaluate factors that might influence the visual prognosis.
Material
and
Methods
Patient records were identified by a computerized diagnostic retrieval system. Information about approximately 60,000 patients who have had fundus photography at the University of Iowa is contained in the system indexed by coded diagnoses and descriptive terms. The index was searched to identify patients with subretinal hemorrhages. To be included in the study, a patient had to have a subretinal hemorrhage of at least one disk diameter in greatest dimension that involved the center of the fovea. Patients with hemorrhage that was beneath the retinal pigment epithelium or in the vitreous were excluded. Also excluded were patients in whom the hemorrhage did not seem to be the
JANUARY,
1990
33
34
January, 1990
AMERICAN JOURNAL OF OPHTIfALMOLOGY
major factor in the visual loss at the time of diagnosis. For example, if a patient had a large choroidal neovascular membrane with a small rim of hemorrhage, that patient was excluded. At least six months of follow-up information had to be available. A total of 29 patients who met all of the above criteria were identified. Grading of size and thickness of subretinal hemorrhage was based on the fundus photographs. Stereophotographs were available in 28 of the 29 patients. The largest diameter of the hemorrhage was measured with a reticule and divided by the diameter of the disk. Hemorrhages greater than 4 disk diameters in largest diameter were considered large. To be classified as thick, the hemorrhage had to cause an obvious elevation of the retina (Figure). Thickness was assessed at the center of the fovea. Visual acuity was obtained using Snellen charts with the patient's spectacle correction. If a pinhole improved the vision, that vision was used in the study. In some patients, a Snellen equivalent of 20/500 size was used to quantify poor visual acuity. The letter was held progressively closer to the patient until its orientation could be identified, and the Snellen equivalent was calculated from that distance. In charts where counting fingers at a given distance was the only recorded visual acuity, an approximation of the 20/200 Snellen equivalent at the given distance was used to analyze the data. The minimal angle of resolution is the inverse of the Snellen fraction. A visual acuity of hand
Figure (Bennett and associates). Elevation and folding of retina caused by thick subretinal hemorrhage.
motions was assigned a minimal angle of resolution of 5,600 (20/10,000). A visual acuity of light perception was assigned a minimum angle of resolution of 1,000 (20/20,000). A visual acuity of uncentral, unsteady, and unrnaintained was assigned a minimal angle of resolution equal to light perception. All visual acuities were converted to the logarithm of the minimum angle of resolution for statistical analysis." Initial visual acuity was compared to final visual acuity using Student's paired t-test. The mean of each group evaluated was compared using Student's two sample t-test. A loglinear model was used to compare the diagnosis of aging macular degeneration with the thickness of the hemorrhage as a predictor of the outcome of legal blindness (a visual acuity of 20/200 or worse). After statistical manipulation, visual acuities were converted back to the Snellen equivalent for ease of understanding, and these are reported herein.
Results A total of 29 patients met the criteria for entrance in the study. Patients with subretinal hemorrhage seen at our institution were underrepresented because many patients with subretinal hemorrhage had not been photographed at their initial examination. Especially underrepresented were patients with subretinal hemorrhage after retinal detachment surgery. The average follow-up time available on patients in the study was three years. The size of the hemorrhage, cause of the hemorrhage, initial visual acuity, and final visual acuity for each patient are summarized in Table 1. The average visual acuity at the initial visit for the entire study population was 20/860. The visual acuity tended to improve to an average final acuity of 20/480 (P = .07). Initial and final visual acuities for various groups and the P value for the change in acuities are summarized in Table 2. The patients who did not have aging macular degeneration showed significant improvement in their visual acuities, from 20/650 to 20/200, regardless of the size of the hemorrhage (P = .006). Patients with aging macular degeneration tended to worsen slightly from 20/1,300 to 20/1,700 (P = .49). The poorer visual acuity in patients with aging macular degeneration compared to those with other diagnoses was not significant at the initial examination (P = .23), but was highly significant at the final examination (P = .002). Com-
Vol. 109, No. 1
Subretinal Hemorrhage
35
TABLE 1 PATIENT INFORMATION PATIENT NO., AGE (YRS), SEX
1,20, M 2,25, M 3,25, M 4,13, M 5,15, M 6, 78, F 7,82, M 8,70,F 9,85,F 10, 78, F 11,60, M 12,70, M
13, 77, F 14,54, M 15,71, M 16,57, F 17,62, F 18,51, M 19,6 wks, M 20,1, M 21,70, 22,57, 23,77, 24,82, 25,51,
HEMORRHAGE DIAGNOSIS
Trauma
SIZE
Trauma
Large Large
Trauma
Small
Trauma Trauma
Small Small
Aging macular degeneration
Large
Aging macular degeneration Aging macular degeneration
Large Large Large Large
VISUAL ACUITY'
THICKNESS
Thin Thin Thin Thin Thin Thin Thick Thick Thin
INITIAL
FOLLOW·UP FINAL
20/200 20/500 20/70 20/500 20/500 20/500
20/50 20/30 20/40 20/30 20/30 20/2,000
CF (2 ft)
CF (2 ft) CF (3 ft) CF (4 ft)
CF (1 ft)
Small
Thin Thin
Aging macular degeneration Aging macular degeneration
Small Small
Thin Thin
20/500 20/300 20/1000 20/500 20/500
Aging macular degeneration
Large
Aging macular degeneration
Large Large
Thick Thick
20/500
20/3,000 20/8,000 20/320
Aging macular degeneration Aging macular degeneration Aging macular degeneration
LP
20/200 20/300 LP
9 11 53 14 67 12 27 32 9 10 25 108 84 77 36
Thick
CF (3 ft)
CF (3 ft)
Large Large
Thick Thick
HM
Large
Thick
HM CF (4 ft) Uncentral, unsteady, unmaintained
Infant trauma
Large
Thick
Uncentral, unsteady, unmaintained
Uncentral, unsteady,
10 36 16 11 41 62
20/70 CF (6 ft)
Macroaneurysm
Small
Thick
CF (2 ft)
Macroaneurysm
Small Small
Thick Thin
20/200 20/500
Small Large
Thick Thin
CF (3 ft)
20/2,000
20/3,000 20/400 20/1,000 20/200 20/200
Macroaneurysm Macroaneurysm
24
unmaintained
M M M
6 66 6 6
Aging macular degeneration Idiopathic disease Infant trauma
Aging macular degeneration
F
F
(MOB)
26,50, M
Presumed ocular histoplasmosis syndrome Presumed ocular histo-
Small
Thin
20/200
20/30
27,46,F
plasmosis syndrome Presumed ocular histo-
Small
Thin
20/200
CF (4 ft)
144
28,18, M
plasmosis syndrome Pseudoxanthoma
Large
Thin
20/200
20/2,000
29
29, 78, F
elasticum Scleral buckle
Large
Thin
20/200
20/70
29
·LP indicates light perception; CF, counting fingers; HM, hand motions.
parisons of the visual outcome of various groups of patients are summarized in Table 3. The other diagnosis that seemed to affect the prognosis was that of choroidal rupture in adults. These patients showed remarkable improvement to an average final visual acuity of 20/35 (P = .01), and this visual outcome was significantly better than that in the other patients (P < .001). The size of the hemorrhage also had a significant impact on the visual acuity. Both the area covered by the hemorrhage and the thickness of the hemorrhage had a significant effect on the
initial visual acuity. The average initial visual acuity of those patients with a hemorrhage more than 4 disk diameters in greatest diameter was 20/1,380, whereas the average for those with a smaller area of hemorrhage was 20/430 (P = .039). This difference between large and small hemorrhages, however, was not maintained in the final visual acuities (20/600 vs 20/350, P = .51). Patients with thick hemorrhages had significantly worse initial visual acuities (20/2,800) than patients with thin hemorrhages (20/370, P < .001). The worst visual acuity was also present at the final exam-
36
January, 1990
AMERICAN JOURNAL OF OPHTHALMOLOGY
TABLE 2 COMPARISONS OF INITIAL AND FINAL VISUAL ACUITIES
PATIENT GROUP (NO.)
VISUAL ACUITY INITIAL
FINAL
All patients (29)
20/480 20/860 Aging macular degeneration (12) 20/1.300 20/1.700 Non-aging macular degeneration 20/650 20/200 patients (17) Adult choroidal rupture (5) 20/280 20/35 Thick hemorrhage (12) 20/2.800 20/1.300 Non-aging macular degeneration 20/2,400 20/800 thick hemorrhages (6) Aging macular degeneration 20/3.200 20/2,100
P VALUE
.07 .49 .006 .01 .06
.17 .10
thick hemorrhages (6) Thin hemorrhages (17) 20/370 Non-aging macular degeneration 20/320 thin hemorrhages (11)
20/240
.36
20/90
.02
Aging macular degeneration thin hemorrhages (6) Large hemorrhages (17)
20/1,400
.22
Small hemorrhages (12)
20/520
20/1,400 20/600 20/440 20/360
.02 .75
ination (20/1,300 vs 20/240, P = .02). Patients with thick hemorrhages, however, showed an improvement from initial to final visual acuity that approached statistical significance (P = .06). The difference in visual outcome between thick and thin hemorrhages was still significant when patients with aging macular degeneration were excluded from the calculations (20/800 vs 20/90, P = .02). To determine which factor had the most impact on the final visual acuity, we used a loglinear model to compare the predictive value of the diagnosis of aging macular degeneration vs the presence of a thick hemorrhage to the outcome of legal blindness. The diagnosis of aging macular degeneration was the significantly more important predictor (P = .03).
Discussion In this study, patients with aging macular degeneration had poor initial visual acuity and did not change on follow-up, whereas other patients improved significantly. The final visual acuity in patients with aging macular degeneration was significantly worse than that in the rest of the group. A loglinear model also showed that the diagnosis of aging macular degeneration was a more important predictor of poor
final vision than was the thickness of the hemorrhage. Cass" noted that the visual prognosis is better when the hemorrhage is subretinal than when it is below the retinal pigment epithelium. He observed that when patients with subretinal neovascularization from aging macular degeneration bleed, the blood spreads beneath the retinal pigment epithelium and only secondarily breaks through into the subretinal space. 10 In other conditions, such as trauma or the presumed ocular histoplasmosis syndrome, the hemorrhage is present almost exclusively in the subretinal space.":" Our data support the idea that patients with aging macular degeneration fare poorly when choroidal neovascular membranes bleed, even though we excluded patients with hemorrhages below the retinal pigment epithelium. In patients in whom the blood was thick, however, it was impossible to see through the subretinal blood to evaluate for the presence of hemorrhage below the retinal pigment epithelium. Therefore, some or all of our patients with thick hemorrhages may have also had hemorrhages beneath the retinal pigment epithelium. Alternatively, the retina and retinal pigment epithelium may be compromised by the condition of aging macular degeneration so that blood is poorly tolerated. It is also possible that the process of neovascularization in aging
TABLE 3 COMPARISONS BETWEEN VISUAL OUTCOME GROUP
GROUP
(FINAL VISUAL ACUITY)
(FINAL VISUAL ACUITY)
P VALUE
Large hemorrhage
Small hemorrhage
.51
(20/600) Thick hemorrhage (20/1,300) Non-aging macular degeneration with thick hemorrhage (20/800) Aging macular degeneration (20/1,700) Aging macular degeneration (20/1 ,700) Aging macular degeneration with thin hemorrhage (20/1 ,400) Aging macular degeneration with thick hemorrhage (20/2,100) Choroidal rupture
(20/350) Thin hemorrhage
.02
(20/35)
(20/240) Non-aging macular degeneration with thin hemorrhage (20/90) Non-aging macular degeneration (20/200) Choroidal rupture (20{35) Non-aging macular degeneration with thin hemorrhage (20/90) Non-aging macular degeneration with thick hemorrhage (20/800) Nonchoroidal rupture (20/830)
.02
.002 <.001 .003
.33
<.001
Vol. 109, No.1
Subretinal Hemorrhage
macular degeneration may cause more scarring and disruption of the photoreceptors. Regardless of the mechanism, it appears that patients with aging macular degeneration have a worse prognosis from subretinal hemorrhage than do patients with other conditions. The thickness of the hemorrhage appeared to have more impact on the final visual acuity than did the area of hemorrhage, though both had significant effect on the initial vision. Glatt and Macherner" postulated that one mechanism of the damage of subretinal blood is a barrier effect that prevents the photoreceptors from receiving metabolic support from the retinal pigment epithelium and choroid. Our observations support this theory, because a thick hemorrhage would be a more complete barrier and lead to more irreversible damage. The patients who did not have aging macular degeneration and had a thin hemorrhage showed remarkable improvement. The average final visual acuity in this group was 20/90. A thin hemorrhage, therefore, might be a less effective barrier and cause dysfunction that is partially reversible. The natural course of the hemorrhages in patients not affected with aging macular degeneration was improvement in vision. Even in the group with thick hemorrhages, including some patients with aging macular degeneration, there tended to be improvement. Though experimental models have shown irreversible photoreceptor damage from subretinal blood.' perhaps the damage is not complete under the entire hemorrhage. Alternatively, patients may learn to compensate by fixating around a scotoma, improving their measured visual acuity in spite of irreversible damage. The number of patients in this study was small, and the results must be interpreted cautiously because of its retrospective design. It appears, however, that there are substantial differences in the outcome of patients with subretinal hemorrhages depending on cause and size of the hemorrhage. The diagnosis of aging macular degeneration appears to be associated with a poor prognosis. The thickness of the hemorrhage also may be an independent prognostic factor. In patients other than those with aging macular degeneration, however, the condition usually improves. In patients with thin hemorrhages, the return of vision can be excellent. In considering patients for surgery, one would like to select those with the worst prognosis, that is, those with aging macular degeneration. In such patients, however, the blood
37
may be partially below the retinal pigment epithelium. Even if the blood is removed, the remaining neovascular membrane may cause significant scarring and visual loss. Conversely, patients with thin hemorrhages from trauma appear to do well; they have excellent return of vision and would not be good candidates for intervention. Because even the group with thick hemorrhages, including those patients with aging macular degeneration, tended to improve, the inclusion of untreated controls in any study of surgical intervention for subretinal blood is necessary.
References 1. Abrams, G. A.: Retinotomies and retinectomies. In Ryan, S. J. (ed.): Retina, vol. 3. St. Louis, C. V. Mosby, 1989, pp. 321-322. 2. De Juan, E., and Machemer, R.: Vitreous surgery for hemorrhagic and fibrous complications of agerelated macular degeneration. Am. J. Ophthalmol. 105:25,1988. 3. Hanscom, T. A., and Diddie, K. R.: Early surgical drainage of macular subretinal hemorrhage. Arch. Ophthalmo!. 105:1722, 1987. 4. Flynn, H. W., Davis, J. L., PareL J. M., and Lee, W. G.: Applications of a cannulated extrusion needle during vitreoretinal microsurgery. Retina 8:42, 1988. 5. Glatt, H., and Machemer, R.: Experimental subretinal hemorrhage in rabbits. Am. J. Ophthalmo!. 94:762,1982. 6. Bressler, S. B., Bressler, N. M., Fine, S. L., Hillis, A., Murphy, R. P., ou. R. J., and Patz, A.: Natural course of choroidal neovascular membranes within the foveal avascular zone in senile macular degeneration. Am. J. Ophthalmol. 93:157, 1982. 7. Bressler, S. B., Bressler, N. M., Fine, S. L., McCormick, P., and Auer, c.: Subfoveal neovascular membranes in senile macular degeneration. Relationship between membrane size and visual prognosis. Retina 3:7, 1983. 8. Westheimer, G.: Scaling of visual acuity measurements. Arch. Ophthalmo!. 97:327, 1979. 9. Cass, J. D M.: Stereoscopic Atlas of Macular Diseases, ed. 3. St. Louis, C. V. Mosby, 1987, p. 80. 10. - - : Pathogenesis of disciform detachment of the neuroepithelium. III. Senile disciform macular degeneration. Am. J. Ophthalmol. 63:617, 1967. 11. - - : Pathogenesis of disciform detachment of the neuroepithelium. V. Disciform macular degeneration secondary to focal choroiditis. Am. J. Ophthalmo!. 63:661, 1967. 12. - - : Pathogenesis of disciform detachment of the neuroepithelium. VI. Disciform detachment secondary to heredodegenerative, neoplastic and traumatic lesions of the choroid. Am. J. Ophthalmo!. 63:689, 1967.
=
RECENT ADVANCES in vitreoretinal surgical techniques have made it possible to remove blood from the subretinal space in the macular area.!" Some patients who have undergone surgery to remove subretinal blood showed im-
Accepted for publication Oct. 19, 1989. From the Department of Ophthalmology, University of Iowa, Iowa City, Iowa (Drs. Bennett, Folk, and Blodi) and the College of Business Administration, Drake University, Des Moines, Iowa (Dr. Klugman). This study was presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Sarasota, Florida, May 2, 1989. This study was supported in part by an unrestricted grant from Research to Prevent Blindness, Inc., New York, New York, and the Retina Research Fund of the University of Iowa. Reprint requests to Christopher F. Blodi. M.D., Department of Ophthalmology, University of Iowa Hospitals, Iowa City, IA 52242.
©AMER:ICAN JOURNAL OF OPHTHALMOLOGY
109:33-37,
proved vision.v" An experimental model of subretinal hemorrhage showed that photoreceptor damage occurs within 24 hours." Perhaps early removal of blood could prevent the visual loss associated with this condition. To determine which patients might benefit from surgical removal of blood from the subretinal space, one must understand the natural history of the condition. Factors that might influence prognosis, such as the cause and size of the hemorrhage, have not been systematically evaluated. A common cause of subretinal hemorrhage is choroidal neovascularization associated with aging macular degeneration. This neovascularization has a poor visual prognosis. 6 . 7 Therefore, an eye that has a subfoveal hemorrhage associated with this type of neovascularization may also fare poorly. A similar hemorrhage caused by another process, such as trauma, might have a different prognosis. We undertook a retrospective study of subretinal hemorrhages of at least one disk diameter in size involving the center of the fovea to evaluate factors that might influence the visual prognosis.
Material
and
Methods
Patient records were identified by a computerized diagnostic retrieval system. Information about approximately 60,000 patients who have had fundus photography at the University of Iowa is contained in the system indexed by coded diagnoses and descriptive terms. The index was searched to identify patients with subretinal hemorrhages. To be included in the study, a patient had to have a subretinal hemorrhage of at least one disk diameter in greatest dimension that involved the center of the fovea. Patients with hemorrhage that was beneath the retinal pigment epithelium or in the vitreous were excluded. Also excluded were patients in whom the hemorrhage did not seem to be the
JANUARY,
1990
33
34
January, 1990
AMERICAN JOURNAL OF OPHTIfALMOLOGY
major factor in the visual loss at the time of diagnosis. For example, if a patient had a large choroidal neovascular membrane with a small rim of hemorrhage, that patient was excluded. At least six months of follow-up information had to be available. A total of 29 patients who met all of the above criteria were identified. Grading of size and thickness of subretinal hemorrhage was based on the fundus photographs. Stereophotographs were available in 28 of the 29 patients. The largest diameter of the hemorrhage was measured with a reticule and divided by the diameter of the disk. Hemorrhages greater than 4 disk diameters in largest diameter were considered large. To be classified as thick, the hemorrhage had to cause an obvious elevation of the retina (Figure). Thickness was assessed at the center of the fovea. Visual acuity was obtained using Snellen charts with the patient's spectacle correction. If a pinhole improved the vision, that vision was used in the study. In some patients, a Snellen equivalent of 20/500 size was used to quantify poor visual acuity. The letter was held progressively closer to the patient until its orientation could be identified, and the Snellen equivalent was calculated from that distance. In charts where counting fingers at a given distance was the only recorded visual acuity, an approximation of the 20/200 Snellen equivalent at the given distance was used to analyze the data. The minimal angle of resolution is the inverse of the Snellen fraction. A visual acuity of hand
Figure (Bennett and associates). Elevation and folding of retina caused by thick subretinal hemorrhage.
motions was assigned a minimal angle of resolution of 5,600 (20/10,000). A visual acuity of light perception was assigned a minimum angle of resolution of 1,000 (20/20,000). A visual acuity of uncentral, unsteady, and unrnaintained was assigned a minimal angle of resolution equal to light perception. All visual acuities were converted to the logarithm of the minimum angle of resolution for statistical analysis." Initial visual acuity was compared to final visual acuity using Student's paired t-test. The mean of each group evaluated was compared using Student's two sample t-test. A loglinear model was used to compare the diagnosis of aging macular degeneration with the thickness of the hemorrhage as a predictor of the outcome of legal blindness (a visual acuity of 20/200 or worse). After statistical manipulation, visual acuities were converted back to the Snellen equivalent for ease of understanding, and these are reported herein.
Results A total of 29 patients met the criteria for entrance in the study. Patients with subretinal hemorrhage seen at our institution were underrepresented because many patients with subretinal hemorrhage had not been photographed at their initial examination. Especially underrepresented were patients with subretinal hemorrhage after retinal detachment surgery. The average follow-up time available on patients in the study was three years. The size of the hemorrhage, cause of the hemorrhage, initial visual acuity, and final visual acuity for each patient are summarized in Table 1. The average visual acuity at the initial visit for the entire study population was 20/860. The visual acuity tended to improve to an average final acuity of 20/480 (P = .07). Initial and final visual acuities for various groups and the P value for the change in acuities are summarized in Table 2. The patients who did not have aging macular degeneration showed significant improvement in their visual acuities, from 20/650 to 20/200, regardless of the size of the hemorrhage (P = .006). Patients with aging macular degeneration tended to worsen slightly from 20/1,300 to 20/1,700 (P = .49). The poorer visual acuity in patients with aging macular degeneration compared to those with other diagnoses was not significant at the initial examination (P = .23), but was highly significant at the final examination (P = .002). Com-
Vol. 109, No. 1
Subretinal Hemorrhage
35
TABLE 1 PATIENT INFORMATION PATIENT NO., AGE (YRS), SEX
1,20, M 2,25, M 3,25, M 4,13, M 5,15, M 6, 78, F 7,82, M 8,70,F 9,85,F 10, 78, F 11,60, M 12,70, M
13, 77, F 14,54, M 15,71, M 16,57, F 17,62, F 18,51, M 19,6 wks, M 20,1, M 21,70, 22,57, 23,77, 24,82, 25,51,
HEMORRHAGE DIAGNOSIS
Trauma
SIZE
Trauma
Large Large
Trauma
Small
Trauma Trauma
Small Small
Aging macular degeneration
Large
Aging macular degeneration Aging macular degeneration
Large Large Large Large
VISUAL ACUITY'
THICKNESS
Thin Thin Thin Thin Thin Thin Thick Thick Thin
INITIAL
FOLLOW·UP FINAL
20/200 20/500 20/70 20/500 20/500 20/500
20/50 20/30 20/40 20/30 20/30 20/2,000
CF (2 ft)
CF (2 ft) CF (3 ft) CF (4 ft)
CF (1 ft)
Small
Thin Thin
Aging macular degeneration Aging macular degeneration
Small Small
Thin Thin
20/500 20/300 20/1000 20/500 20/500
Aging macular degeneration
Large
Aging macular degeneration
Large Large
Thick Thick
20/500
20/3,000 20/8,000 20/320
Aging macular degeneration Aging macular degeneration Aging macular degeneration
LP
20/200 20/300 LP
9 11 53 14 67 12 27 32 9 10 25 108 84 77 36
Thick
CF (3 ft)
CF (3 ft)
Large Large
Thick Thick
HM
Large
Thick
HM CF (4 ft) Uncentral, unsteady, unmaintained
Infant trauma
Large
Thick
Uncentral, unsteady, unmaintained
Uncentral, unsteady,
10 36 16 11 41 62
20/70 CF (6 ft)
Macroaneurysm
Small
Thick
CF (2 ft)
Macroaneurysm
Small Small
Thick Thin
20/200 20/500
Small Large
Thick Thin
CF (3 ft)
20/2,000
20/3,000 20/400 20/1,000 20/200 20/200
Macroaneurysm Macroaneurysm
24
unmaintained
M M M
6 66 6 6
Aging macular degeneration Idiopathic disease Infant trauma
Aging macular degeneration
F
F
(MOB)
26,50, M
Presumed ocular histoplasmosis syndrome Presumed ocular histo-
Small
Thin
20/200
20/30
27,46,F
plasmosis syndrome Presumed ocular histo-
Small
Thin
20/200
CF (4 ft)
144
28,18, M
plasmosis syndrome Pseudoxanthoma
Large
Thin
20/200
20/2,000
29
29, 78, F
elasticum Scleral buckle
Large
Thin
20/200
20/70
29
·LP indicates light perception; CF, counting fingers; HM, hand motions.
parisons of the visual outcome of various groups of patients are summarized in Table 3. The other diagnosis that seemed to affect the prognosis was that of choroidal rupture in adults. These patients showed remarkable improvement to an average final visual acuity of 20/35 (P = .01), and this visual outcome was significantly better than that in the other patients (P < .001). The size of the hemorrhage also had a significant impact on the visual acuity. Both the area covered by the hemorrhage and the thickness of the hemorrhage had a significant effect on the
initial visual acuity. The average initial visual acuity of those patients with a hemorrhage more than 4 disk diameters in greatest diameter was 20/1,380, whereas the average for those with a smaller area of hemorrhage was 20/430 (P = .039). This difference between large and small hemorrhages, however, was not maintained in the final visual acuities (20/600 vs 20/350, P = .51). Patients with thick hemorrhages had significantly worse initial visual acuities (20/2,800) than patients with thin hemorrhages (20/370, P < .001). The worst visual acuity was also present at the final exam-
36
January, 1990
AMERICAN JOURNAL OF OPHTHALMOLOGY
TABLE 2 COMPARISONS OF INITIAL AND FINAL VISUAL ACUITIES
PATIENT GROUP (NO.)
VISUAL ACUITY INITIAL
FINAL
All patients (29)
20/480 20/860 Aging macular degeneration (12) 20/1.300 20/1.700 Non-aging macular degeneration 20/650 20/200 patients (17) Adult choroidal rupture (5) 20/280 20/35 Thick hemorrhage (12) 20/2.800 20/1.300 Non-aging macular degeneration 20/2,400 20/800 thick hemorrhages (6) Aging macular degeneration 20/3.200 20/2,100
P VALUE
.07 .49 .006 .01 .06
.17 .10
thick hemorrhages (6) Thin hemorrhages (17) 20/370 Non-aging macular degeneration 20/320 thin hemorrhages (11)
20/240
.36
20/90
.02
Aging macular degeneration thin hemorrhages (6) Large hemorrhages (17)
20/1,400
.22
Small hemorrhages (12)
20/520
20/1,400 20/600 20/440 20/360
.02 .75
ination (20/1,300 vs 20/240, P = .02). Patients with thick hemorrhages, however, showed an improvement from initial to final visual acuity that approached statistical significance (P = .06). The difference in visual outcome between thick and thin hemorrhages was still significant when patients with aging macular degeneration were excluded from the calculations (20/800 vs 20/90, P = .02). To determine which factor had the most impact on the final visual acuity, we used a loglinear model to compare the predictive value of the diagnosis of aging macular degeneration vs the presence of a thick hemorrhage to the outcome of legal blindness. The diagnosis of aging macular degeneration was the significantly more important predictor (P = .03).
Discussion In this study, patients with aging macular degeneration had poor initial visual acuity and did not change on follow-up, whereas other patients improved significantly. The final visual acuity in patients with aging macular degeneration was significantly worse than that in the rest of the group. A loglinear model also showed that the diagnosis of aging macular degeneration was a more important predictor of poor
final vision than was the thickness of the hemorrhage. Cass" noted that the visual prognosis is better when the hemorrhage is subretinal than when it is below the retinal pigment epithelium. He observed that when patients with subretinal neovascularization from aging macular degeneration bleed, the blood spreads beneath the retinal pigment epithelium and only secondarily breaks through into the subretinal space. 10 In other conditions, such as trauma or the presumed ocular histoplasmosis syndrome, the hemorrhage is present almost exclusively in the subretinal space.":" Our data support the idea that patients with aging macular degeneration fare poorly when choroidal neovascular membranes bleed, even though we excluded patients with hemorrhages below the retinal pigment epithelium. In patients in whom the blood was thick, however, it was impossible to see through the subretinal blood to evaluate for the presence of hemorrhage below the retinal pigment epithelium. Therefore, some or all of our patients with thick hemorrhages may have also had hemorrhages beneath the retinal pigment epithelium. Alternatively, the retina and retinal pigment epithelium may be compromised by the condition of aging macular degeneration so that blood is poorly tolerated. It is also possible that the process of neovascularization in aging
TABLE 3 COMPARISONS BETWEEN VISUAL OUTCOME GROUP
GROUP
(FINAL VISUAL ACUITY)
(FINAL VISUAL ACUITY)
P VALUE
Large hemorrhage
Small hemorrhage
.51
(20/600) Thick hemorrhage (20/1,300) Non-aging macular degeneration with thick hemorrhage (20/800) Aging macular degeneration (20/1,700) Aging macular degeneration (20/1 ,700) Aging macular degeneration with thin hemorrhage (20/1 ,400) Aging macular degeneration with thick hemorrhage (20/2,100) Choroidal rupture
(20/350) Thin hemorrhage
.02
(20/35)
(20/240) Non-aging macular degeneration with thin hemorrhage (20/90) Non-aging macular degeneration (20/200) Choroidal rupture (20{35) Non-aging macular degeneration with thin hemorrhage (20/90) Non-aging macular degeneration with thick hemorrhage (20/800) Nonchoroidal rupture (20/830)
.02
.002 <.001 .003
.33
<.001
Vol. 109, No.1
Subretinal Hemorrhage
macular degeneration may cause more scarring and disruption of the photoreceptors. Regardless of the mechanism, it appears that patients with aging macular degeneration have a worse prognosis from subretinal hemorrhage than do patients with other conditions. The thickness of the hemorrhage appeared to have more impact on the final visual acuity than did the area of hemorrhage, though both had significant effect on the initial vision. Glatt and Macherner" postulated that one mechanism of the damage of subretinal blood is a barrier effect that prevents the photoreceptors from receiving metabolic support from the retinal pigment epithelium and choroid. Our observations support this theory, because a thick hemorrhage would be a more complete barrier and lead to more irreversible damage. The patients who did not have aging macular degeneration and had a thin hemorrhage showed remarkable improvement. The average final visual acuity in this group was 20/90. A thin hemorrhage, therefore, might be a less effective barrier and cause dysfunction that is partially reversible. The natural course of the hemorrhages in patients not affected with aging macular degeneration was improvement in vision. Even in the group with thick hemorrhages, including some patients with aging macular degeneration, there tended to be improvement. Though experimental models have shown irreversible photoreceptor damage from subretinal blood.' perhaps the damage is not complete under the entire hemorrhage. Alternatively, patients may learn to compensate by fixating around a scotoma, improving their measured visual acuity in spite of irreversible damage. The number of patients in this study was small, and the results must be interpreted cautiously because of its retrospective design. It appears, however, that there are substantial differences in the outcome of patients with subretinal hemorrhages depending on cause and size of the hemorrhage. The diagnosis of aging macular degeneration appears to be associated with a poor prognosis. The thickness of the hemorrhage also may be an independent prognostic factor. In patients other than those with aging macular degeneration, however, the condition usually improves. In patients with thin hemorrhages, the return of vision can be excellent. In considering patients for surgery, one would like to select those with the worst prognosis, that is, those with aging macular degeneration. In such patients, however, the blood
37
may be partially below the retinal pigment epithelium. Even if the blood is removed, the remaining neovascular membrane may cause significant scarring and visual loss. Conversely, patients with thin hemorrhages from trauma appear to do well; they have excellent return of vision and would not be good candidates for intervention. Because even the group with thick hemorrhages, including those patients with aging macular degeneration, tended to improve, the inclusion of untreated controls in any study of surgical intervention for subretinal blood is necessary.
References 1. Abrams, G. A.: Retinotomies and retinectomies. In Ryan, S. J. (ed.): Retina, vol. 3. St. Louis, C. V. Mosby, 1989, pp. 321-322. 2. De Juan, E., and Machemer, R.: Vitreous surgery for hemorrhagic and fibrous complications of agerelated macular degeneration. Am. J. Ophthalmol. 105:25,1988. 3. Hanscom, T. A., and Diddie, K. R.: Early surgical drainage of macular subretinal hemorrhage. Arch. Ophthalmo!. 105:1722, 1987. 4. Flynn, H. W., Davis, J. L., PareL J. M., and Lee, W. G.: Applications of a cannulated extrusion needle during vitreoretinal microsurgery. Retina 8:42, 1988. 5. Glatt, H., and Machemer, R.: Experimental subretinal hemorrhage in rabbits. Am. J. Ophthalmo!. 94:762,1982. 6. Bressler, S. B., Bressler, N. M., Fine, S. L., Hillis, A., Murphy, R. P., ou. R. J., and Patz, A.: Natural course of choroidal neovascular membranes within the foveal avascular zone in senile macular degeneration. Am. J. Ophthalmol. 93:157, 1982. 7. Bressler, S. B., Bressler, N. M., Fine, S. L., McCormick, P., and Auer, c.: Subfoveal neovascular membranes in senile macular degeneration. Relationship between membrane size and visual prognosis. Retina 3:7, 1983. 8. Westheimer, G.: Scaling of visual acuity measurements. Arch. Ophthalmo!. 97:327, 1979. 9. Cass, J. D M.: Stereoscopic Atlas of Macular Diseases, ed. 3. St. Louis, C. V. Mosby, 1987, p. 80. 10. - - : Pathogenesis of disciform detachment of the neuroepithelium. III. Senile disciform macular degeneration. Am. J. Ophthalmol. 63:617, 1967. 11. - - : Pathogenesis of disciform detachment of the neuroepithelium. V. Disciform macular degeneration secondary to focal choroiditis. Am. J. Ophthalmo!. 63:661, 1967. 12. - - : Pathogenesis of disciform detachment of the neuroepithelium. VI. Disciform detachment secondary to heredodegenerative, neoplastic and traumatic lesions of the choroid. Am. J. Ophthalmo!. 63:689, 1967.