Ischemic Macular Edema Recognition and Favorable Natural History in Branch Vein Occlusion Daniel Finkelstein, MD
\s=b\ Eyes with macular edema caused by retinal branch vein occlusion underwent masked evaluation of fluorescein angiography to determine complete macular perfusion vs incomplete macular perfusion (capillary dropout, ischemia). Cases evaluated as incomplete macular perfusion showed a greater frequency of improvement (91%) in visual acuity than did perfused cases (29%) (P=.003) after a mean follow-up of 39 months. Ischemic edema is often transient, and is associated with a good outcome in visual acuity (median final visual acuity was 20/30). Perfused macular edema has a poorer prognosis for visual acuity (median final visual acuity was 20/80). Macular ischemia is usually associated with a broken foveal capillary ring. Previous animal research on ischemic brain edema has shown that following brain ischemia, an intracellular and an extracellular hypertonic environment lead to intracellular and extracellular edema (cytotoxic edema), which is often followed by vascular protein leakage (vasogenic edema). A similar occurrence in ischemic retina could explain the transient edema reported herein, with good outcome in visual acuity following the spontaneous resolution of edema.
(Arch Ophthalmol. 1992;110:1427-1434)
Accepted
for publication December 13, 1991. From the Department of Ophthalmology, The Wilmer Institute, The Johns Hopkins University School of Medicine, Baltimore, Md. Reprint requests to Department of Ophthalmology, The Wilmer Institute, The Johns Hopkins Hospital, Baltimore, MD 21205 (Dr Finkelstein).
edema is a common cause of ]\/ -"-'-acular loss of visual in dif¬
acuity many ferent diseases1; visual prognosis and
mechanisms for the formation of macu¬ lar edema and for its resolution, either spontaneous or with laser photocoagu¬ lation, are poorly understood.2 The data reported herein attempt to clarify one aspect of macular edema, that which is associated, at least in part, with isch¬ emia, as documented by an absence of capillary perfusion in the macula in high-quality fluorescein angiography following retinal branch vein occlusion. The term edema refers to a condition in which the tissue's water content is increased.3 In the macula, edema is recognized clinically by a thickening of the macular retina that is often accom¬ panied by cystoid spaces. Objective doc¬ umentation of macular edema is most readily obtained by fluorescein angiog¬ raphy since most (but not all4) occur¬ rences of macular edema are associated with a disruption in the blood-retinal barrier at the level of the retinal capil¬ laries. The entity of branch vein occlusion provides a unique model for the study of retinal vascular abnormalities, such as macular edema and nonperfusion (isch¬ emia), because the retinal vascular ab¬ normalities often remain stable, and the stability often can be documented by flu¬ orescein angiography. As in other retinal vascular disorders, exact mechanisms for the production of macular edema caused by branch vein occlusion, and for its spontaneous resolution or for resolu¬ tion of the edema with laser photocoag-
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
ulation,
are
ever, two
poorly understood.·'11 How¬ previous studies on branch
vein occlusion have indicated that when there is a broken foveal capillary ring, the visual prognosis is worsened.12·13 Such observations are of importance in consid¬ ering hypotheses regarding causes of macular edema, mechanisms of sponta¬ neous resolution, and mechanisms of treatment effect with laser photocoagu¬ lation; a reexamination of the correlation between the visual acuity prognosis and a broken foveal capillary ring forms one aspect of the basis for the data reported herein. Also confirmed is an association of the broken foveal capillary ring with macular ischemia, as documented by dis¬ tinct areas of macular capillary nonper¬ fusion on high-quality fluorescein an¬
giography.
In the neurologic literature, it is widely reported and generally clinically accepted that there are two major types of brain edema, termed vasogenic and cytotoxic.1416 Vasogenic brain edema refers to edema caused by an escape of water with plasma protein constituents from cerebral blood vessels; cytotoxic brain edema refers to cellular change in the brain parenchyma producing edema because of intracellular hypertonicity followed by extracellular hypertonicity. The entity of "ischemie brain edema" consists of cytotoxic brain edema fol¬ lowed by vasogenic brain edema. These
studies in brain edema appear remark¬
ably analogous and relevant to the tran¬ sient macular edema, reported herein for the first time in the retina, that is associated with capillary nonperfusion.
Comparison is provided for the visual acuity outcome in two groups of pa¬
tients with macular edema and loss of visual acuity after branch vein occlu¬ sion, comparing one group with macular ischemia (usually associated with a bro¬ ken foveal capillary ring) with another group with good macular perfusion (usually associated with an intact foveal capillary ring); it is demonstrated that the group with macular ischemia has a statistically significant improved visual prognosis, without laser therapy. The visual outcome for this group with mac¬ ular ischemia, then, has important clin¬ ical implications regarding consider¬ ation of laser therapy in the treatment of macular edema after branch vein oc¬ clusion. Discussed also is the correlation between macular capillary nonperfu¬ sion and ischemia with possible mecha¬ nisms of macular edema and its resolu¬ tion. The macular edema associated with good capillary perfusion is always associated with fluorescein leakage on the fluorescein angiogram since the edema is of the vasogenic type, with leakage of the fluorescein molecule through a break in the blood-retinal barrier. The macular edema associated with macular ischemia may or may not be associated with fluorescein leakage since the majority of the edema may be secondary to intracellular and extracel¬ lular hypertonicity with consequent wa¬ ter movement that may or may not be associated with a significant break in the blood-retinal barrier. MATERIALS AND METHODS All of the cases coded as branch vein occlusion and followed up by the author from 1975 to 1989 were reviewed. The diagnosis of branch vein occlusion was made according to the characteristic, segmental presence of in¬ traretinal hemorrhage with corresponding segmental intraretinal vascular abnormali¬ ties. Patients with macular edema after branch vein occlusion were included in this study if 1. They presented for examination within 1 year of onset of symptoms. 2. Best corrected visual acuity was 20/40 or worse.
3. They had a diagnosis of macular edema based on one or both of the following criteria: retinal thickening involving the fovea on slit-lamp contact lens examination and/or cystoid maculopathy involving the fovea on slit-lamp contact lens examination (and/or stereofundus photography and/or fluores¬ cein angiography). 4. Color photography and fluorescein an¬
giography were performed. 5. They could be available for follow-up at
least 6 months after the first examination. Patients with pseudophakia or aphakia were not eligible for the study. Patients with diabetic retinopathy or any other disease that could cause loss of visual acuity were also ineligible.
Using these eligibility and exclusion crite¬ ria, 67 cases were identified. The transit phase of the fluorescein angiogram for each case was prepared as a strip of six negative
frames to be read in masked fashion by two senior readers at the Central Vein Occlusion Study Reading Center (Miami, Fla). The two readers evaluated each fluorescein negative independently and in masked fashion, never having seen these fluorescein angiograms previously. The readers did not see the late phase of the fluorescein angiogram and had no access to the visual acuity in these cases. The readers were asked to (1) evaluate whether the perifoveal capillary ring at the border of the foveal avascular zone was intact or broken, (2) determine whether any distinct area of capillary nonperfusion could be identified within 1 disc diameter of the center of the fovea, and (3) identify fluores¬ cein angiograms whose quality or definition was insufficient to make these determina¬ tions with certainty. Of the 67 fluorescein angiograms presented to the readers, 30 an¬ giograms were judged of high quality for the determinations requested; the readers inde¬ pendently provided the same interpretation for each of these 30 angiograms. Statistical significance was evaluated by Fisher's Exact Test. For many cases, several fluorescein angio¬ grams were obtained during the course of follow-up examination; if the initial angio¬ gram at time of presentation did not demon¬ strate sufficient capillary detail of the foveal region, a later fluorescein angiogram, if of sufficient quality, was chosen for analysis. In all cases in which a later angiogram was cho¬ sen, it did not appear that there was any change in the basic characteristic of the oc¬ clusion; that is, the occlusion had not "pro¬
gressed."1'
For purposes of illustration in this report, the best single fluorescein frame was chosen, although for certain cases, the availability of multiple fluorescein angiograms or stereofluorescein angiograms helped to clarify fur¬ ther the status of the foveal capillary ring18 and/or macular capillary nonperfusion. REPORT OF CASES
Thirty eyes fulfilled the eligibility and ex¬ clusion criteria and were judged by two masked readers, independently agreeing, to have a sufficiently high-quality fluorescein angiogram to ascertain perifoveal capillary filling or lack of filling with a high degree of certainty. Of these 30 eyes, 23 had macular capillary nonperfusion and seven had intact macular capillary perfusion. The fluorescein angiogram frame that best demonstrates the foveal capillary integrity or lack of integrity in the transit phase is illustrated for each case in Figs 1 and 2. Summaries of key fea¬ tures of these 30 cases are presented in Ta¬ bles 1 and 2. Three representative cases are presented in more detail. Patient IR.—A 49-year-old hypertensive white man with a 3-month history of de¬ creased vision in his right eye was first exam¬ ined in March 1983. Visual acuity was 20/70 OD and 20/15 OS. Fundus examination of the right eye demonstrated an inferotemporal oc¬ clusion (Fig 3, top left) with cystoid macular edema. There was no foveal hemorrhage.
Fluorescein angiography demonstrated a break in the foveal capillary ring along with extensive adjacent capillary nonperfusion (Fig 3, top right) with late fluorescein stain¬ ing in the perifoveal region, but without flu¬ orescein staining in a cystoid pattern (Fig 3, bottom left). With no laser therapy, visual acuity in the right eye improved to 20/40 in 2 months and to 20/15 in 6 months. At 6 months, the late phase of the fluorescein an¬ giogram showed less macular staining (Fig 3, bottom right); on ophthalmoscopic examina¬ tion, there was no longer cystoid macular edema. The patient was followed up for 6 years with maintenance of 20/15 visual acu¬ ity in the right eye. Patient 4.—A 71-year-old hypertensive white woman with a 4-month history of de¬ creased vision in her left eye was first exam¬ ined in September 1987. The patient had been diabetic for 5 years, and had been tak¬ ing insulin to control the diabetes. Vision was noted to have been decreased in the right eye for an unknown duration. Visual acuity was 20/100 OD and 20/60 OS. Fundus examina¬ tion showed no diabetic retinopathy in either eye. Fundus examination of the right eye demonstrated an old branch vein occlusion with macular involvement, which was the cause of visual loss. On fundus examination of the left eye, an inferotemporal macular branch vein occlusion was noted (Fig 4, top left) with cystoid edema involving the fovea and no foveal hemorrhage. Fluorescein an¬ giography demonstrated a break in the foveal capillary ring and adjacent capillary nonperfusion in the transit phase (Fig 4, top center) and prominent cystoid fluorescein accumulation in the late phase (Fig 4, top right). Without laser therapy, visual acuity in the left eye improved to 20/40 in 3 months and to 20/25 in 13 months. Fluorescein an¬ giography at 13 months demonstrated less staining and no cystoid staining (Fig 4, bot¬ tom left and bottom right). The duration of cystoid edema on ophthalmoscopy was 3 months. PATIENT 24.—A 57-year-old hypertensive black woman with a 4-month history of decreased vision in her right eye was first examined in August 1981. Visual acuity was 20/40 OD and 20/15 OS. Fundus examination of the right eye demonstrated a superotemporal macular branch vein occlusion (Fig 5, left) with cystoid macular edema and no foveal hemorrhage. The results of the fundus examination of the left eye were normal. Fluorescein angiography demonstrated an intact foveal capillary ring and good adjacent macular capillary perfusion (Fig 5, center) with cystoid leakage in the late phase (Fig 5, right). It is assumed that this case repre¬ sents pure vasogenic edema without a cyto¬ toxic component. Without laser therapy, vi¬ sual acuity decreased to 20/50 in 8 months, 20/70 in 15 months, 20/100 in 17 months, and 20/200 in 24 months, then remained stable at 20/200 for the remainder of follow-up, which extended to 88 months. Cystoid macular edema was present for 35 months.
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
RESULTS
Sixty-seven eyes met the eligibility and exclusion criteria as stated in the "Materials and Methods" section. Of
Fig 1.—Composite fluorescein angiogram
of all eyes with macular ischemia.
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
Fig 2.—Composite fluorescein angiogram
show eyes with macular ischemia
(A to H) and those with complete macular perfusion (I
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
to
O).
Table 1.—Clinical Characteristics of Patients With Ischemie Macular Edema*
Patient No.
Age, y 49 49 52 71
10 12 13 14 15 16 17
19 20 21 22 23
56 59 73 71 57 50 45 62 78 67 56 71 60 78 57 57 68 65 76
Hyper¬
tension Yes Yes Yes Yes Yes No Yes Yes No Yes Yes No Yes No Yes Yes Yes No Yes No No Yes No
Eye
Duration,
Type
mo
l/Q/T Q/S/T
Ring Intact No No No No No No
I
l/M/T Q/S/T H/S Q/S/T Q/S/T l/Q/T l/Q/T
No No No No No No No No Yes
l/M/T
l/Q/T Q/S/T Q/S/T Q/S/T l/Q/T Q/S/T M/S/T M/S/T Q/S/T M/S/T l/Q/T Q/S/T
No No No No No No Yes
Yes
Foveal
Hemorrhage No No No No Yes No No Yes No No No No No Yes Yes No Yes No Yes Yes No No Yes
Cystoid
Macular Edema Yes Yes No Yes No No No Yes Yes No No No No No Yes Yes Yes No Yes No No Yes Yes
Duration of Cystoid Macular
Initial Visual
Final Visual
Edema,
Acuity
Acuity
20/70 20/80 10/200 20/60
20/15 20/15 20/20 20/25
20/50 20/70 20/100 20/100 20/60 20/60 20/40 20/100 20/200
20/30 20/30 20/40 20/30 20/20 20/30 20/15 20/40 20/200
28 12 17
20/100 20/50 20/50 20/100
20/30 20/30 20/20 20/80
64 37 28 20
86 76 28 32
20/125 20/80 20/50 20/100
20/125 20/30 20/15 20/50
12
20/100 20/50
20/20 20/30
17 72 25 60 13
mo
24
37 12
17
Duration to Best Vision,
Duration of
mo
mo
35 14 13
20
Follow-up, 72 35 23 13 47 30 40 14 23 103 12
14 12
60 13
*Data include the patient's age at time of presentation; hypertension (defined as having taken antlhypertensive medication before presentation); right (R) or left (L) eye; duration of symptoms before the first examination; type of occlusion fundus segment Involved, where I indicates Inferior; S, superior; T, temporal; Q, quad¬ rant; and M, macular; whether the Innermost foveal capillary ring was Intact; whether foveal hemorrhage was present at the time of presentation; whether cystoid macular edema was present at the time of presentation; follow-up duration of cystoid macular edema; Initial visual acuity; final visual acuity at last follow-up; follow-up duration until the best vision was obtained; and total duration of follow-up.
Table 2.—Clinical Characteristics of Patients With Perfused Macular Edema*
Patient No.
Age,
24 25
57 72
26 27 28 29
65 65 56 47 71
30
y
Hyper¬
tension Yes No No No Yes No No
Eye
"For clarification of data and
Duration,
Type
mo
M/S/T M/S/T Q/S/T l/M/T
Ring Intact Yes Yes Yes
I
M/S/T
expansion
of abbreviations,
the 67 cases, 30 (45%) could be graded with certainty for presence or absence of macular ischemia. Of these, 23 were graded independently by the masked observers as including defini¬ tive macular capillary nonperfusion adjacent to the foveal avascular zone and seven were judged as completely perfused. Of the 67 eyes presented to the masked readers at the Reading Center, the 37 eyes that could not be evaluated for perfusion characteristics had inadequate quality fluorescein an¬ giography (24 cases) or involved dis¬ agreement between the readers on perfusion status (13 cases). As judged by the readers, of the 23 cases with
Yes Yes Yes No see
Cystoid Foveal
Hemorrhage No No No No Yes No No
Macular Edema Yes Yes No No Yes Yes No
Duration of Cystoid Macular
Initial Visual
Final Visual
Edema,
Acuity
Acuity
20/40 20/140
20/400 20/140 20/30 20/40 20/400
mo
35
44
20/50 20/40 20/100 20/400 20/70
Duration to Best Vision,
Duration of
mo
mo
Follow-up, 88
17
116 47 48
20/25 20/80
footnote In Table 1.
clearly identifiable macular capillary nonperfusion, 20 (87%) cases also could be identified as having a broken foveal ring at the border of the foveal
avascular zone. Of the seven cases with good macular capillary perfusion, six (86%) had an intact foveal capil¬ lary ring at the border of the foveal avascular zone. Consequently, for the cases reported herein, there is good correlation between an intact foveal capillary ring with good capillary per¬ fusion and a broken capillary ring with macular ischemia. For the 23 eyes with macular capillary ischemia, the median initial visual acuity was 20/80 and the median final visual acu-
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
20/30. By comparison, for the with good macular perfu¬ sion, the median initial visual acuity was 20/70 and the median final visual acuity was 20/80. The median follow-up time for the 23 cases with macular ischemia was 28 months; the median follow-up time for the seven cases with good macular per¬ fusion was 47 months. For the 23 cases with macular capillary nonperfusion, 11 (48%) demonstrated cystoid macular edema, which ranged from 3 to 37 months in duration (median, 6 months). Ten of the 11 patients with cystoid edema had a final visual acuity of 20/30 or better; the three patients with the longest duration
ity
was
seven cases
of cystoid edema, 17 months, 24 months, and 37 months, achieved final visual acu¬ ity of 20/30 or better. Of the 23 patients with macular cap¬ illary nonperfusion, 21 (91%) improved in visual acuity. The duration in months to best visual acuity varied from 3 to 64 months (median, 14 months). For the seven eyes with good macu¬ lar capillary perfusion, two (29%) eyes improved in visual acuity. The statisti¬ cally significant difference in proportion improved between the two groups (Fisher's Exact Test, P=.003), demon¬ strating the improved visual acuity prognosis for the group of eyes with macular capillary nonperfusion. COMMENT
Macular edema, which can occur with diabetic retinopathy, branch vein occlu¬ sion, central vein occlusion, and follow¬ ing cataract surgery, is a frequent cause of loss of vision for which laser therapy is often inadequate. The prognosis for visual acuity is often so variable that one can provide a patient only with a rough estimate of future outcome. In an attempt to learn more about the nature of macular edema, this study was de¬ signed to evaluate a discrete and care¬ fully chosen subset of macular edema cases, those cases associated with mac¬ ular ischemia in an otherwise stable disease entity. Retinal branch vein oc¬ clusion provides a unique opportunity to study such a circumstance because it is a disease often associated with mac¬ ular ischemia and is a well-defined en¬ tity with an acute onset in an otherwise normal eye with vascular changes that only rarely progress. The highest-quality fluorescein an¬
giogram was required to study capillary details that would, with certainty, de¬ fine perfusion or nonperfusion (isch¬ emia). All of the 67 eyes that met the
eligibility criteria
for macular edema, loss of vision, and no other confounding ocular disease, with sufficient follow-up, were presented to two fluorescein read¬ ers at the Central Vein Occlusion Study Reading Center. Their masked evalua¬ tion ofthe fluorescein transit only, with¬ out knowledge of visual acuity or follow-up visual acuity, led to the selec¬ tion of 30 cases in which there was independent agreement regarding the status of macular perfusion or macular nonperfusion (ischemia). The analysis of these 30 eyes demonstrated, with sta¬ tistical significance, that macular isch¬ emia is associated with a much im¬ proved visual outcome (91%) compared with eyes with good macular perfusion (29%). In addition, a strong association was noted between eyes with macular ischemia and a broken foveal capillary
Fig 3.—Top left, Red-free photography demonstrates inferotemporal branch vein occlusion of right eye. Top right, The transit phase of the fluorescein angiogram demonstrates capillary nonperfusion and a broken foveal capillary ring (arrow). Bottom left, The late phase of the flu¬ orescein angiogram demonstrates fluorescein staining of the macula, but without a cystoid pattern. Bottom right, The late phase of the fluorescein angiogram demonstrates less macular staining at 6-month follow-up visit.
the
ring (Figs 3 and 4), as well as a strong association between eyes with good macular perfusion and an intact foveal capillary ring (Fig 5).
regarding good
macular perfusion vs macular ischemia because the fluores¬ cein angiograms generally were not of high enough quality to provide the dif¬ ferentiation (Branch Vein Occlusion Study Reading Center, oral communi¬ cation, August 1989). However, pa¬ tients in the Collaborative Branch Vein Occlusion Study with distinct areas of capillary nonperfusion in the macula were excluded from that study. Conse¬ quently, it would be expected that whenever they could be so identified,
For the cases with macular ischemia, when the edema that accompanies the
capillary nonperfusion spontaneously resolved, visual acuity improved in 91%. It is suggested, therefore, that macular edema and cystoid macular edema may be produced by capillary nonperfusion as an
acute and transient event. On
oc¬
casion, that edema may be documented by the appearance of cystoid leakage on the fluorescein angiogram (Fig 4), but
often documentation is not possible in this fashion (exemplifying the condition of "nonleaking cystoid") and is only vis¬ ible on ophthalmoscopy and/or slit-lamp contact lens examination (Fig 3). The Collaborative Branch Vein Oc¬ clusion Study Group recently reported the visual acuity prognosis for 35 un¬ treated patients with macular edema after branch vein occlusion and loss of visual acuity to 20/40 or worse.19 That study reported that even without laser therapy, 37% of eyes showed a sponta¬ neous visual improvement. This im¬ provement is similar to the 29% im¬ provement demonstrated for the eyes with good macular perfusion reported herein (Table 2). The cases reported in the Collaborative Branch Vein Occlu¬ sion Study were not divided into types
with significant macular capillary nonperfusion were excluded from the cases
Collaborative Branch Vein Occlusion Study, so that it is reasonable that the percentage improvement for patients in the Collaborative Branch Vein Occlu¬ sion Study is similar to the good macu¬ lar perfusion cases reported herein. There are no previous studies that have examined the prognosis for visual acuity of patients with branch vein occlu¬ sion accompanied by documented macu¬ lar ischemia; however, two previous studies have examined the prognosis for visual acuity in patients who might have similar branch vein occlusions with bro¬ ken foveal capillary rings. 12·1: Contrary to the cases reported herein, both of these studies report that branch vein occlusion with a broken foveal capillary ring had a worsened prognosis for visual acuity. These previous reports, however, are
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
Fig 4.—Top left, Red-free photography demonstrates acute branch vein occlusion of the left eye. Top center, The transit phase of the fluores¬ cein angiogram demonstrates macular capillary nonperfusion and a break in the foveal capillary ring (arrow). Top right, The late phase of the fluorescein angiogram demonstrates central cyst of cystoid macular edema (arrow). Bottom left, The transit phase of the fluorescein angiogram at 13-month follow-up demonstrates a break in the foveal capillary ring (arrow). Bottom right, The late phase of the fluorescein angiogram at 13-month follow-up demonstrates less staining and no cystoid staining.
Fig 5.—Left, Red-free photography demonstrates fundus appearance of a superotemporal macular branch vein occlusion. Center, Fluorescein angiography demonstrates good macular capillary perfusion and an intact foveal capillary ring. Right, The late phase of the fluorescein angio¬ gram demonstrates cystoid macular edema with foveal involvement. difficult to compare with the series re¬ ported herein since details of each case and high-quality angiography are not
presented.
Ischemie macular edema appears to a transient edema with visual acuity improvement as it spontaneously re¬ solves. Presumably, the macula is suffi¬ ciently oversupplied with vasculature for its metabolic needs that small seg¬ ments of ischemia do not of themselves produce acuity loss after edema has re¬ solved. In contrast, perfused edema frequently persists, with persistence of visual acuity decrease. One plausible mechanism might be that death and disruption of neural tissue from isch¬ emia produces an intracellular and extracellular hypertonic environment, be
forcing water movement from any ocu¬ lar water source and subsequent
edema; a consequent disturbance of the retinal vasculature leading to protein leakage from disruption of the bloodretinal barrier may or may not occur. There will be fluorescein leakage only if this secondary damage to the bloodretinal barrier occurs; water that moves toward a hypertonic environment does not carry fluorescein with it. Such a phenomenon is compatible with the ob¬ servation of cystoid edema occurring without fluorescein leakage in a cystoid pattern in retinitis pigmentosa.'1 In cases of ischemia, after the tissue debris (and consequent hypertonic environ¬ ment) is removed, the edema would be expected to resolve spontaneously. If
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
the ischemia is insufficient to cause widespread macular neural tissue loss, as hypothesized to be the case in the branch vein occlusion cases presented herein, remarkable improvement in vi¬ sual acuity could then occur. This concept of two separate types of neural edema, hyperpermeable (vaso¬ genic) edema and hypertonic (cytotoxic) edema, has never received prior men¬ tion in the ophthalmic literature, but it has received considerable attention in the neurologic literature.14"16 The entity of branch vein occlusion provides a unique model to examine hypertonic and hyperpermeable edema since both situations can be seen to occur as acute events accompanying the occlusion. In his presidential address to the
American Association of Neuropatholo¬ gists in 1966, Klatzo14 drew the sharp dis¬ tinction between vasogenic brain edema and cytotoxic brain edema. By 1977, the study group on brain edema in stroke provided a review article on brain edema emphasizing that focal ischemia begins as cytotoxic edema, followed by vaso¬ genic edema, as supported by animal ex¬ perimentation and information from ionic and tracer studies.15 For this partic¬ ular pattern of development of ischemie edema, the Study Group on Brain Edema established a separate classification known
as
"ischemie brain edema." This
study group review stated the following: "Necrosis, with the breaking up of struc¬ tural components of brain parenchyma, creates an enlargement of extracellular spaces, which may be distended further
by a marked increase in water movement due to lysosomal enzymatic digestion of
structural cellular elements and release of osmotically active substances." Ischemie cerebral edema was again reviewed in 1979 by O'Brien,16 who stated: "Complete ischemia rapidly causes a sequence of events with cessa¬ tion of electrical activity, failure of the sodium pump, depletion of glucose and increase in lactate. Throughout this time the blood-brain barrier remains intact and there is virtually no leakage of protein." It is clear from experimen¬ tal brain studies that the initial edema is intracellular followed by extracellular caused by osmotic changes in the cell from failure of the sodium pump, as well as extracellular from tissue damage, with both of these circumstances lead¬ ing on occasion to an opening of vascu¬ lar permeability. These brain events that are so nicely documented in the neurologic literature might well be expected to occur in the
retina.
Fortunately, the acute edema accompanying ischemie vein occlusion does not produce the problems that oc¬ cur
in the central
nervous
system of
dangerously increasing intracranial pressure. Consequently, as the osmotic abnormalities, both intracellular and extracellular, are resolved over time, the edema spontaneously involutes
with return of foveal function when the edema is resorbed. Further explorations and awareness of these two separate types of edema may be important in evaluating other retinal vascular abnormalities as well, particularly for background diabetic retinopathy. For example, the recently reported Early Treatment Diabetic Retinopathy Study Research Group20 recommended grid laser photocoagulation to areas of retinal ischemia in the context of grid treatment for clinically significant macular edema. However, it might be that grid laser treatment is not necessary in these areas of ischemia (but only needed in fluorescein leaking areas), as suggested by the spontaneous resolu¬ tion of ischemie edema in the branch vein occlusion cases reported herein. Interestingly, basic laboratory ani¬ mal work on the mechanism of effect of grid laser photocoagulation lends sup¬ port to the hypothesis presented herein, suggesting that decreased reti¬ nal blood flow may decrease macular edema.21 Wilson et al21 showed that grid laser photocoagulation of the normal monkey macula, placed in a pattern identical to the grid pattern that was successful in the Collaborative Branch Vein Occlusion Study, produced a nar¬ rowing and absence of the retinal vas¬ culature (presumably as an autoregulatory effect when the retina was thinned by the photocoagulation) that would be
expected to decrease retinal vascular flow. Similarly, the decrease of retinal capillary flow in segmentai macular ischemia of branch vein occlusion might be expected to result in the same type
of decrease in edema. There are important clinical observa¬ tions and recommendations that follow from the data reported herein. For cases that meet the eligibility criteria, when a distinct area of macular capil¬ lary ischemia can be identified with certainty on fluorescein angiography, it is likely that visual acuity will improve significantly without therapy. Conse¬ quently, the data reported herein un¬ derscore the recommendation of the Collaborative Branch Vein Occlusion Study that stated, "We have no basis for recommending early treatment."la Pa¬ tients with branch vein macular edema, then, should be followed up until highquality fluorescein angiography can be obtained. Laser photocoagulation should not be considered if macular ischemia is present with macular edema, unless it is clear that improve¬ ment in visual acuity is not spontane¬ ously occurring. Laser photocoagula¬ tion should be considered if good macular perfusion is present with per¬ sistent macular edema, reducing visual acuity without spontaneous improve¬ ment.
The entity of branch vein occlusion has proven to be a unique model for the study of retinal neovascularization,22 as previously reported. The series of cases reported herein suggest that branch vein occlusion may also provide a unique model for the study of certain aspects of macular edema. This study was supported by research grants EY06142 and EY-01765 from the National Eye Institute, Bethesda, Md.
References 1. Goldberg MF. Diseases affecting the inner blood-retinal barrier. In: Cunha-Vaz JG, ed. The Blood-Retinal Barriers. New York, NY: Plenum Press; 1979:309-363. 2. Jampol LM. Macular edema. In: Ryan SJ, ed. Retina. St Louis, Mo: Mosby-Year Book; 1989;2: 81-88. 3. Henkind P, Bellhorn RW, Schall B. Retinal edema: postulated mechanism(s). In: Cunha-Vaz JG, ed. The Blood-Retinal Barriers. New York, NY: Plenum Press; 1979:251-268. 4. Gass JDM. Stereoscopic Atlas Macular Diseases. St Louis, Mo: Mosby-Year Book; 1970:154. 5. Finkelstein D, Clarkson JG, Branch Vein Occlusion Study Group. Branch and central retinal vein occlusions: focal points. Clin Modules Ophthalmol. 1987;5:1-10. 6. Sedney SC. Photocoagulation in Retinal Vein Occlusion. The Hague, the Netherlands: Dr W Junk; 1976. 7. Coscas G, Dhermy P. Occlusions veineuses retinienne. New York, NY: Masson Publishing USA Inc; 1978. 8. Michels RD, Gass JDM. The natural course of
retinal branch vein occlusion. Trans Am Acad
Ophthalmol Otolaryngol. 1974;78:166-177. 9. Gutman FA, Zegarra H. Macular edema secondary to occlusion of the retinal veins. Surv Oph-
thalmol. 1984;28:464-470. 10. Coscas G, Gaudric A. Natural course of nonaphakic cystoid macular edema. Surv Ophthalmol.
1984;28:471-484.
11. Orth D, Patz A. Retinal branch vein occlusion. Surv Ophthalmol. 1978;22:357-376. 12. Clemett RS, Kohner EM, Hamilton AM. The visual prognosis in retinal branch vein occlusion. Trans Ophthalmol Soc U K. 1973;93:523-535. 13. Shilling JS, Jones CA. Retinal branch vein occlusion: a study of argon laser photocoagulation in the treatment of macular oedema. Br J Ophthalmol. 1984;68:196-198. 14. Klatzo I. Neuropathological aspects of brain edema. J Neuropathol Exp Neurol. 1967;26:1-14. 15. Katzman R, Clasen R, Klatzo I, Meyer JS, Pappius HM, Waltz AG. Report of Joint Committee for Stroke Resources, IV: brain edema in stroke. Stroke. 1977;8:512-540. 16. O'Brien MD. Ischemic cerebral edema: a re-
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
view. Stroke. 1979;10:623-628. 17. Branch Vein Occlusion Study Group. Progressing' branch vein occlusion. Invest Ophthalmol Vis Sci. 1983;24(suppl):172. 18. Bresnick GH, Condit R, Syrjala S, Palta M, Groo A, Korth K. Abnormalities of the foveal avascular zone in diabetic retinopathy. Arch Ophthalmol. 1984;102:1286-1293. 19. Branch Vein Occlusion Study Group. Argon laser photocoagulation for macular edema in branch vein occlusion. Am J Ophthalmol. 1984;98: 271-282. 20. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema: Early Treatment Diabetic Retinopathy Study Report Number 1. Arch Ophthalmol. 1985;103:1796-1806. 21. Wilson DJ, Finkelstein D, Quigley HA, Green WR. Macular grid photocoagulation: an experimental study on the primate retina. Arch Ophthalmol. 1988;106:100-105. 22. Wise GN. Factors influencing retinal new vessel formation. Am J Ophthalmol. 1961;52:637\x=req-\ 650.
\s=b\ Eyes with macular edema caused by retinal branch vein occlusion underwent masked evaluation of fluorescein angiography to determine complete macular perfusion vs incomplete macular perfusion (capillary dropout, ischemia). Cases evaluated as incomplete macular perfusion showed a greater frequency of improvement (91%) in visual acuity than did perfused cases (29%) (P=.003) after a mean follow-up of 39 months. Ischemic edema is often transient, and is associated with a good outcome in visual acuity (median final visual acuity was 20/30). Perfused macular edema has a poorer prognosis for visual acuity (median final visual acuity was 20/80). Macular ischemia is usually associated with a broken foveal capillary ring. Previous animal research on ischemic brain edema has shown that following brain ischemia, an intracellular and an extracellular hypertonic environment lead to intracellular and extracellular edema (cytotoxic edema), which is often followed by vascular protein leakage (vasogenic edema). A similar occurrence in ischemic retina could explain the transient edema reported herein, with good outcome in visual acuity following the spontaneous resolution of edema.
(Arch Ophthalmol. 1992;110:1427-1434)
Accepted
for publication December 13, 1991. From the Department of Ophthalmology, The Wilmer Institute, The Johns Hopkins University School of Medicine, Baltimore, Md. Reprint requests to Department of Ophthalmology, The Wilmer Institute, The Johns Hopkins Hospital, Baltimore, MD 21205 (Dr Finkelstein).
edema is a common cause of ]\/ -"-'-acular loss of visual in dif¬
acuity many ferent diseases1; visual prognosis and
mechanisms for the formation of macu¬ lar edema and for its resolution, either spontaneous or with laser photocoagu¬ lation, are poorly understood.2 The data reported herein attempt to clarify one aspect of macular edema, that which is associated, at least in part, with isch¬ emia, as documented by an absence of capillary perfusion in the macula in high-quality fluorescein angiography following retinal branch vein occlusion. The term edema refers to a condition in which the tissue's water content is increased.3 In the macula, edema is recognized clinically by a thickening of the macular retina that is often accom¬ panied by cystoid spaces. Objective doc¬ umentation of macular edema is most readily obtained by fluorescein angiog¬ raphy since most (but not all4) occur¬ rences of macular edema are associated with a disruption in the blood-retinal barrier at the level of the retinal capil¬ laries. The entity of branch vein occlusion provides a unique model for the study of retinal vascular abnormalities, such as macular edema and nonperfusion (isch¬ emia), because the retinal vascular ab¬ normalities often remain stable, and the stability often can be documented by flu¬ orescein angiography. As in other retinal vascular disorders, exact mechanisms for the production of macular edema caused by branch vein occlusion, and for its spontaneous resolution or for resolu¬ tion of the edema with laser photocoag-
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
ulation,
are
ever, two
poorly understood.·'11 How¬ previous studies on branch
vein occlusion have indicated that when there is a broken foveal capillary ring, the visual prognosis is worsened.12·13 Such observations are of importance in consid¬ ering hypotheses regarding causes of macular edema, mechanisms of sponta¬ neous resolution, and mechanisms of treatment effect with laser photocoagu¬ lation; a reexamination of the correlation between the visual acuity prognosis and a broken foveal capillary ring forms one aspect of the basis for the data reported herein. Also confirmed is an association of the broken foveal capillary ring with macular ischemia, as documented by dis¬ tinct areas of macular capillary nonper¬ fusion on high-quality fluorescein an¬
giography.
In the neurologic literature, it is widely reported and generally clinically accepted that there are two major types of brain edema, termed vasogenic and cytotoxic.1416 Vasogenic brain edema refers to edema caused by an escape of water with plasma protein constituents from cerebral blood vessels; cytotoxic brain edema refers to cellular change in the brain parenchyma producing edema because of intracellular hypertonicity followed by extracellular hypertonicity. The entity of "ischemie brain edema" consists of cytotoxic brain edema fol¬ lowed by vasogenic brain edema. These
studies in brain edema appear remark¬
ably analogous and relevant to the tran¬ sient macular edema, reported herein for the first time in the retina, that is associated with capillary nonperfusion.
Comparison is provided for the visual acuity outcome in two groups of pa¬
tients with macular edema and loss of visual acuity after branch vein occlu¬ sion, comparing one group with macular ischemia (usually associated with a bro¬ ken foveal capillary ring) with another group with good macular perfusion (usually associated with an intact foveal capillary ring); it is demonstrated that the group with macular ischemia has a statistically significant improved visual prognosis, without laser therapy. The visual outcome for this group with mac¬ ular ischemia, then, has important clin¬ ical implications regarding consider¬ ation of laser therapy in the treatment of macular edema after branch vein oc¬ clusion. Discussed also is the correlation between macular capillary nonperfu¬ sion and ischemia with possible mecha¬ nisms of macular edema and its resolu¬ tion. The macular edema associated with good capillary perfusion is always associated with fluorescein leakage on the fluorescein angiogram since the edema is of the vasogenic type, with leakage of the fluorescein molecule through a break in the blood-retinal barrier. The macular edema associated with macular ischemia may or may not be associated with fluorescein leakage since the majority of the edema may be secondary to intracellular and extracel¬ lular hypertonicity with consequent wa¬ ter movement that may or may not be associated with a significant break in the blood-retinal barrier. MATERIALS AND METHODS All of the cases coded as branch vein occlusion and followed up by the author from 1975 to 1989 were reviewed. The diagnosis of branch vein occlusion was made according to the characteristic, segmental presence of in¬ traretinal hemorrhage with corresponding segmental intraretinal vascular abnormali¬ ties. Patients with macular edema after branch vein occlusion were included in this study if 1. They presented for examination within 1 year of onset of symptoms. 2. Best corrected visual acuity was 20/40 or worse.
3. They had a diagnosis of macular edema based on one or both of the following criteria: retinal thickening involving the fovea on slit-lamp contact lens examination and/or cystoid maculopathy involving the fovea on slit-lamp contact lens examination (and/or stereofundus photography and/or fluores¬ cein angiography). 4. Color photography and fluorescein an¬
giography were performed. 5. They could be available for follow-up at
least 6 months after the first examination. Patients with pseudophakia or aphakia were not eligible for the study. Patients with diabetic retinopathy or any other disease that could cause loss of visual acuity were also ineligible.
Using these eligibility and exclusion crite¬ ria, 67 cases were identified. The transit phase of the fluorescein angiogram for each case was prepared as a strip of six negative
frames to be read in masked fashion by two senior readers at the Central Vein Occlusion Study Reading Center (Miami, Fla). The two readers evaluated each fluorescein negative independently and in masked fashion, never having seen these fluorescein angiograms previously. The readers did not see the late phase of the fluorescein angiogram and had no access to the visual acuity in these cases. The readers were asked to (1) evaluate whether the perifoveal capillary ring at the border of the foveal avascular zone was intact or broken, (2) determine whether any distinct area of capillary nonperfusion could be identified within 1 disc diameter of the center of the fovea, and (3) identify fluores¬ cein angiograms whose quality or definition was insufficient to make these determina¬ tions with certainty. Of the 67 fluorescein angiograms presented to the readers, 30 an¬ giograms were judged of high quality for the determinations requested; the readers inde¬ pendently provided the same interpretation for each of these 30 angiograms. Statistical significance was evaluated by Fisher's Exact Test. For many cases, several fluorescein angio¬ grams were obtained during the course of follow-up examination; if the initial angio¬ gram at time of presentation did not demon¬ strate sufficient capillary detail of the foveal region, a later fluorescein angiogram, if of sufficient quality, was chosen for analysis. In all cases in which a later angiogram was cho¬ sen, it did not appear that there was any change in the basic characteristic of the oc¬ clusion; that is, the occlusion had not "pro¬
gressed."1'
For purposes of illustration in this report, the best single fluorescein frame was chosen, although for certain cases, the availability of multiple fluorescein angiograms or stereofluorescein angiograms helped to clarify fur¬ ther the status of the foveal capillary ring18 and/or macular capillary nonperfusion. REPORT OF CASES
Thirty eyes fulfilled the eligibility and ex¬ clusion criteria and were judged by two masked readers, independently agreeing, to have a sufficiently high-quality fluorescein angiogram to ascertain perifoveal capillary filling or lack of filling with a high degree of certainty. Of these 30 eyes, 23 had macular capillary nonperfusion and seven had intact macular capillary perfusion. The fluorescein angiogram frame that best demonstrates the foveal capillary integrity or lack of integrity in the transit phase is illustrated for each case in Figs 1 and 2. Summaries of key fea¬ tures of these 30 cases are presented in Ta¬ bles 1 and 2. Three representative cases are presented in more detail. Patient IR.—A 49-year-old hypertensive white man with a 3-month history of de¬ creased vision in his right eye was first exam¬ ined in March 1983. Visual acuity was 20/70 OD and 20/15 OS. Fundus examination of the right eye demonstrated an inferotemporal oc¬ clusion (Fig 3, top left) with cystoid macular edema. There was no foveal hemorrhage.
Fluorescein angiography demonstrated a break in the foveal capillary ring along with extensive adjacent capillary nonperfusion (Fig 3, top right) with late fluorescein stain¬ ing in the perifoveal region, but without flu¬ orescein staining in a cystoid pattern (Fig 3, bottom left). With no laser therapy, visual acuity in the right eye improved to 20/40 in 2 months and to 20/15 in 6 months. At 6 months, the late phase of the fluorescein an¬ giogram showed less macular staining (Fig 3, bottom right); on ophthalmoscopic examina¬ tion, there was no longer cystoid macular edema. The patient was followed up for 6 years with maintenance of 20/15 visual acu¬ ity in the right eye. Patient 4.—A 71-year-old hypertensive white woman with a 4-month history of de¬ creased vision in her left eye was first exam¬ ined in September 1987. The patient had been diabetic for 5 years, and had been tak¬ ing insulin to control the diabetes. Vision was noted to have been decreased in the right eye for an unknown duration. Visual acuity was 20/100 OD and 20/60 OS. Fundus examina¬ tion showed no diabetic retinopathy in either eye. Fundus examination of the right eye demonstrated an old branch vein occlusion with macular involvement, which was the cause of visual loss. On fundus examination of the left eye, an inferotemporal macular branch vein occlusion was noted (Fig 4, top left) with cystoid edema involving the fovea and no foveal hemorrhage. Fluorescein an¬ giography demonstrated a break in the foveal capillary ring and adjacent capillary nonperfusion in the transit phase (Fig 4, top center) and prominent cystoid fluorescein accumulation in the late phase (Fig 4, top right). Without laser therapy, visual acuity in the left eye improved to 20/40 in 3 months and to 20/25 in 13 months. Fluorescein an¬ giography at 13 months demonstrated less staining and no cystoid staining (Fig 4, bot¬ tom left and bottom right). The duration of cystoid edema on ophthalmoscopy was 3 months. PATIENT 24.—A 57-year-old hypertensive black woman with a 4-month history of decreased vision in her right eye was first examined in August 1981. Visual acuity was 20/40 OD and 20/15 OS. Fundus examination of the right eye demonstrated a superotemporal macular branch vein occlusion (Fig 5, left) with cystoid macular edema and no foveal hemorrhage. The results of the fundus examination of the left eye were normal. Fluorescein angiography demonstrated an intact foveal capillary ring and good adjacent macular capillary perfusion (Fig 5, center) with cystoid leakage in the late phase (Fig 5, right). It is assumed that this case repre¬ sents pure vasogenic edema without a cyto¬ toxic component. Without laser therapy, vi¬ sual acuity decreased to 20/50 in 8 months, 20/70 in 15 months, 20/100 in 17 months, and 20/200 in 24 months, then remained stable at 20/200 for the remainder of follow-up, which extended to 88 months. Cystoid macular edema was present for 35 months.
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
RESULTS
Sixty-seven eyes met the eligibility and exclusion criteria as stated in the "Materials and Methods" section. Of
Fig 1.—Composite fluorescein angiogram
of all eyes with macular ischemia.
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
Fig 2.—Composite fluorescein angiogram
show eyes with macular ischemia
(A to H) and those with complete macular perfusion (I
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
to
O).
Table 1.—Clinical Characteristics of Patients With Ischemie Macular Edema*
Patient No.
Age, y 49 49 52 71
10 12 13 14 15 16 17
19 20 21 22 23
56 59 73 71 57 50 45 62 78 67 56 71 60 78 57 57 68 65 76
Hyper¬
tension Yes Yes Yes Yes Yes No Yes Yes No Yes Yes No Yes No Yes Yes Yes No Yes No No Yes No
Eye
Duration,
Type
mo
l/Q/T Q/S/T
Ring Intact No No No No No No
I
l/M/T Q/S/T H/S Q/S/T Q/S/T l/Q/T l/Q/T
No No No No No No No No Yes
l/M/T
l/Q/T Q/S/T Q/S/T Q/S/T l/Q/T Q/S/T M/S/T M/S/T Q/S/T M/S/T l/Q/T Q/S/T
No No No No No No Yes
Yes
Foveal
Hemorrhage No No No No Yes No No Yes No No No No No Yes Yes No Yes No Yes Yes No No Yes
Cystoid
Macular Edema Yes Yes No Yes No No No Yes Yes No No No No No Yes Yes Yes No Yes No No Yes Yes
Duration of Cystoid Macular
Initial Visual
Final Visual
Edema,
Acuity
Acuity
20/70 20/80 10/200 20/60
20/15 20/15 20/20 20/25
20/50 20/70 20/100 20/100 20/60 20/60 20/40 20/100 20/200
20/30 20/30 20/40 20/30 20/20 20/30 20/15 20/40 20/200
28 12 17
20/100 20/50 20/50 20/100
20/30 20/30 20/20 20/80
64 37 28 20
86 76 28 32
20/125 20/80 20/50 20/100
20/125 20/30 20/15 20/50
12
20/100 20/50
20/20 20/30
17 72 25 60 13
mo
24
37 12
17
Duration to Best Vision,
Duration of
mo
mo
35 14 13
20
Follow-up, 72 35 23 13 47 30 40 14 23 103 12
14 12
60 13
*Data include the patient's age at time of presentation; hypertension (defined as having taken antlhypertensive medication before presentation); right (R) or left (L) eye; duration of symptoms before the first examination; type of occlusion fundus segment Involved, where I indicates Inferior; S, superior; T, temporal; Q, quad¬ rant; and M, macular; whether the Innermost foveal capillary ring was Intact; whether foveal hemorrhage was present at the time of presentation; whether cystoid macular edema was present at the time of presentation; follow-up duration of cystoid macular edema; Initial visual acuity; final visual acuity at last follow-up; follow-up duration until the best vision was obtained; and total duration of follow-up.
Table 2.—Clinical Characteristics of Patients With Perfused Macular Edema*
Patient No.
Age,
24 25
57 72
26 27 28 29
65 65 56 47 71
30
y
Hyper¬
tension Yes No No No Yes No No
Eye
"For clarification of data and
Duration,
Type
mo
M/S/T M/S/T Q/S/T l/M/T
Ring Intact Yes Yes Yes
I
M/S/T
expansion
of abbreviations,
the 67 cases, 30 (45%) could be graded with certainty for presence or absence of macular ischemia. Of these, 23 were graded independently by the masked observers as including defini¬ tive macular capillary nonperfusion adjacent to the foveal avascular zone and seven were judged as completely perfused. Of the 67 eyes presented to the masked readers at the Reading Center, the 37 eyes that could not be evaluated for perfusion characteristics had inadequate quality fluorescein an¬ giography (24 cases) or involved dis¬ agreement between the readers on perfusion status (13 cases). As judged by the readers, of the 23 cases with
Yes Yes Yes No see
Cystoid Foveal
Hemorrhage No No No No Yes No No
Macular Edema Yes Yes No No Yes Yes No
Duration of Cystoid Macular
Initial Visual
Final Visual
Edema,
Acuity
Acuity
20/40 20/140
20/400 20/140 20/30 20/40 20/400
mo
35
44
20/50 20/40 20/100 20/400 20/70
Duration to Best Vision,
Duration of
mo
mo
Follow-up, 88
17
116 47 48
20/25 20/80
footnote In Table 1.
clearly identifiable macular capillary nonperfusion, 20 (87%) cases also could be identified as having a broken foveal ring at the border of the foveal
avascular zone. Of the seven cases with good macular capillary perfusion, six (86%) had an intact foveal capil¬ lary ring at the border of the foveal avascular zone. Consequently, for the cases reported herein, there is good correlation between an intact foveal capillary ring with good capillary per¬ fusion and a broken capillary ring with macular ischemia. For the 23 eyes with macular capillary ischemia, the median initial visual acuity was 20/80 and the median final visual acu-
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
20/30. By comparison, for the with good macular perfu¬ sion, the median initial visual acuity was 20/70 and the median final visual acuity was 20/80. The median follow-up time for the 23 cases with macular ischemia was 28 months; the median follow-up time for the seven cases with good macular per¬ fusion was 47 months. For the 23 cases with macular capillary nonperfusion, 11 (48%) demonstrated cystoid macular edema, which ranged from 3 to 37 months in duration (median, 6 months). Ten of the 11 patients with cystoid edema had a final visual acuity of 20/30 or better; the three patients with the longest duration
ity
was
seven cases
of cystoid edema, 17 months, 24 months, and 37 months, achieved final visual acu¬ ity of 20/30 or better. Of the 23 patients with macular cap¬ illary nonperfusion, 21 (91%) improved in visual acuity. The duration in months to best visual acuity varied from 3 to 64 months (median, 14 months). For the seven eyes with good macu¬ lar capillary perfusion, two (29%) eyes improved in visual acuity. The statisti¬ cally significant difference in proportion improved between the two groups (Fisher's Exact Test, P=.003), demon¬ strating the improved visual acuity prognosis for the group of eyes with macular capillary nonperfusion. COMMENT
Macular edema, which can occur with diabetic retinopathy, branch vein occlu¬ sion, central vein occlusion, and follow¬ ing cataract surgery, is a frequent cause of loss of vision for which laser therapy is often inadequate. The prognosis for visual acuity is often so variable that one can provide a patient only with a rough estimate of future outcome. In an attempt to learn more about the nature of macular edema, this study was de¬ signed to evaluate a discrete and care¬ fully chosen subset of macular edema cases, those cases associated with mac¬ ular ischemia in an otherwise stable disease entity. Retinal branch vein oc¬ clusion provides a unique opportunity to study such a circumstance because it is a disease often associated with mac¬ ular ischemia and is a well-defined en¬ tity with an acute onset in an otherwise normal eye with vascular changes that only rarely progress. The highest-quality fluorescein an¬
giogram was required to study capillary details that would, with certainty, de¬ fine perfusion or nonperfusion (isch¬ emia). All of the 67 eyes that met the
eligibility criteria
for macular edema, loss of vision, and no other confounding ocular disease, with sufficient follow-up, were presented to two fluorescein read¬ ers at the Central Vein Occlusion Study Reading Center. Their masked evalua¬ tion ofthe fluorescein transit only, with¬ out knowledge of visual acuity or follow-up visual acuity, led to the selec¬ tion of 30 cases in which there was independent agreement regarding the status of macular perfusion or macular nonperfusion (ischemia). The analysis of these 30 eyes demonstrated, with sta¬ tistical significance, that macular isch¬ emia is associated with a much im¬ proved visual outcome (91%) compared with eyes with good macular perfusion (29%). In addition, a strong association was noted between eyes with macular ischemia and a broken foveal capillary
Fig 3.—Top left, Red-free photography demonstrates inferotemporal branch vein occlusion of right eye. Top right, The transit phase of the fluorescein angiogram demonstrates capillary nonperfusion and a broken foveal capillary ring (arrow). Bottom left, The late phase of the flu¬ orescein angiogram demonstrates fluorescein staining of the macula, but without a cystoid pattern. Bottom right, The late phase of the fluorescein angiogram demonstrates less macular staining at 6-month follow-up visit.
the
ring (Figs 3 and 4), as well as a strong association between eyes with good macular perfusion and an intact foveal capillary ring (Fig 5).
regarding good
macular perfusion vs macular ischemia because the fluores¬ cein angiograms generally were not of high enough quality to provide the dif¬ ferentiation (Branch Vein Occlusion Study Reading Center, oral communi¬ cation, August 1989). However, pa¬ tients in the Collaborative Branch Vein Occlusion Study with distinct areas of capillary nonperfusion in the macula were excluded from that study. Conse¬ quently, it would be expected that whenever they could be so identified,
For the cases with macular ischemia, when the edema that accompanies the
capillary nonperfusion spontaneously resolved, visual acuity improved in 91%. It is suggested, therefore, that macular edema and cystoid macular edema may be produced by capillary nonperfusion as an
acute and transient event. On
oc¬
casion, that edema may be documented by the appearance of cystoid leakage on the fluorescein angiogram (Fig 4), but
often documentation is not possible in this fashion (exemplifying the condition of "nonleaking cystoid") and is only vis¬ ible on ophthalmoscopy and/or slit-lamp contact lens examination (Fig 3). The Collaborative Branch Vein Oc¬ clusion Study Group recently reported the visual acuity prognosis for 35 un¬ treated patients with macular edema after branch vein occlusion and loss of visual acuity to 20/40 or worse.19 That study reported that even without laser therapy, 37% of eyes showed a sponta¬ neous visual improvement. This im¬ provement is similar to the 29% im¬ provement demonstrated for the eyes with good macular perfusion reported herein (Table 2). The cases reported in the Collaborative Branch Vein Occlu¬ sion Study were not divided into types
with significant macular capillary nonperfusion were excluded from the cases
Collaborative Branch Vein Occlusion Study, so that it is reasonable that the percentage improvement for patients in the Collaborative Branch Vein Occlu¬ sion Study is similar to the good macu¬ lar perfusion cases reported herein. There are no previous studies that have examined the prognosis for visual acuity of patients with branch vein occlu¬ sion accompanied by documented macu¬ lar ischemia; however, two previous studies have examined the prognosis for visual acuity in patients who might have similar branch vein occlusions with bro¬ ken foveal capillary rings. 12·1: Contrary to the cases reported herein, both of these studies report that branch vein occlusion with a broken foveal capillary ring had a worsened prognosis for visual acuity. These previous reports, however, are
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
Fig 4.—Top left, Red-free photography demonstrates acute branch vein occlusion of the left eye. Top center, The transit phase of the fluores¬ cein angiogram demonstrates macular capillary nonperfusion and a break in the foveal capillary ring (arrow). Top right, The late phase of the fluorescein angiogram demonstrates central cyst of cystoid macular edema (arrow). Bottom left, The transit phase of the fluorescein angiogram at 13-month follow-up demonstrates a break in the foveal capillary ring (arrow). Bottom right, The late phase of the fluorescein angiogram at 13-month follow-up demonstrates less staining and no cystoid staining.
Fig 5.—Left, Red-free photography demonstrates fundus appearance of a superotemporal macular branch vein occlusion. Center, Fluorescein angiography demonstrates good macular capillary perfusion and an intact foveal capillary ring. Right, The late phase of the fluorescein angio¬ gram demonstrates cystoid macular edema with foveal involvement. difficult to compare with the series re¬ ported herein since details of each case and high-quality angiography are not
presented.
Ischemie macular edema appears to a transient edema with visual acuity improvement as it spontaneously re¬ solves. Presumably, the macula is suffi¬ ciently oversupplied with vasculature for its metabolic needs that small seg¬ ments of ischemia do not of themselves produce acuity loss after edema has re¬ solved. In contrast, perfused edema frequently persists, with persistence of visual acuity decrease. One plausible mechanism might be that death and disruption of neural tissue from isch¬ emia produces an intracellular and extracellular hypertonic environment, be
forcing water movement from any ocu¬ lar water source and subsequent
edema; a consequent disturbance of the retinal vasculature leading to protein leakage from disruption of the bloodretinal barrier may or may not occur. There will be fluorescein leakage only if this secondary damage to the bloodretinal barrier occurs; water that moves toward a hypertonic environment does not carry fluorescein with it. Such a phenomenon is compatible with the ob¬ servation of cystoid edema occurring without fluorescein leakage in a cystoid pattern in retinitis pigmentosa.'1 In cases of ischemia, after the tissue debris (and consequent hypertonic environ¬ ment) is removed, the edema would be expected to resolve spontaneously. If
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
the ischemia is insufficient to cause widespread macular neural tissue loss, as hypothesized to be the case in the branch vein occlusion cases presented herein, remarkable improvement in vi¬ sual acuity could then occur. This concept of two separate types of neural edema, hyperpermeable (vaso¬ genic) edema and hypertonic (cytotoxic) edema, has never received prior men¬ tion in the ophthalmic literature, but it has received considerable attention in the neurologic literature.14"16 The entity of branch vein occlusion provides a unique model to examine hypertonic and hyperpermeable edema since both situations can be seen to occur as acute events accompanying the occlusion. In his presidential address to the
American Association of Neuropatholo¬ gists in 1966, Klatzo14 drew the sharp dis¬ tinction between vasogenic brain edema and cytotoxic brain edema. By 1977, the study group on brain edema in stroke provided a review article on brain edema emphasizing that focal ischemia begins as cytotoxic edema, followed by vaso¬ genic edema, as supported by animal ex¬ perimentation and information from ionic and tracer studies.15 For this partic¬ ular pattern of development of ischemie edema, the Study Group on Brain Edema established a separate classification known
as
"ischemie brain edema." This
study group review stated the following: "Necrosis, with the breaking up of struc¬ tural components of brain parenchyma, creates an enlargement of extracellular spaces, which may be distended further
by a marked increase in water movement due to lysosomal enzymatic digestion of
structural cellular elements and release of osmotically active substances." Ischemie cerebral edema was again reviewed in 1979 by O'Brien,16 who stated: "Complete ischemia rapidly causes a sequence of events with cessa¬ tion of electrical activity, failure of the sodium pump, depletion of glucose and increase in lactate. Throughout this time the blood-brain barrier remains intact and there is virtually no leakage of protein." It is clear from experimen¬ tal brain studies that the initial edema is intracellular followed by extracellular caused by osmotic changes in the cell from failure of the sodium pump, as well as extracellular from tissue damage, with both of these circumstances lead¬ ing on occasion to an opening of vascu¬ lar permeability. These brain events that are so nicely documented in the neurologic literature might well be expected to occur in the
retina.
Fortunately, the acute edema accompanying ischemie vein occlusion does not produce the problems that oc¬ cur
in the central
nervous
system of
dangerously increasing intracranial pressure. Consequently, as the osmotic abnormalities, both intracellular and extracellular, are resolved over time, the edema spontaneously involutes
with return of foveal function when the edema is resorbed. Further explorations and awareness of these two separate types of edema may be important in evaluating other retinal vascular abnormalities as well, particularly for background diabetic retinopathy. For example, the recently reported Early Treatment Diabetic Retinopathy Study Research Group20 recommended grid laser photocoagulation to areas of retinal ischemia in the context of grid treatment for clinically significant macular edema. However, it might be that grid laser treatment is not necessary in these areas of ischemia (but only needed in fluorescein leaking areas), as suggested by the spontaneous resolu¬ tion of ischemie edema in the branch vein occlusion cases reported herein. Interestingly, basic laboratory ani¬ mal work on the mechanism of effect of grid laser photocoagulation lends sup¬ port to the hypothesis presented herein, suggesting that decreased reti¬ nal blood flow may decrease macular edema.21 Wilson et al21 showed that grid laser photocoagulation of the normal monkey macula, placed in a pattern identical to the grid pattern that was successful in the Collaborative Branch Vein Occlusion Study, produced a nar¬ rowing and absence of the retinal vas¬ culature (presumably as an autoregulatory effect when the retina was thinned by the photocoagulation) that would be
expected to decrease retinal vascular flow. Similarly, the decrease of retinal capillary flow in segmentai macular ischemia of branch vein occlusion might be expected to result in the same type
of decrease in edema. There are important clinical observa¬ tions and recommendations that follow from the data reported herein. For cases that meet the eligibility criteria, when a distinct area of macular capil¬ lary ischemia can be identified with certainty on fluorescein angiography, it is likely that visual acuity will improve significantly without therapy. Conse¬ quently, the data reported herein un¬ derscore the recommendation of the Collaborative Branch Vein Occlusion Study that stated, "We have no basis for recommending early treatment."la Pa¬ tients with branch vein macular edema, then, should be followed up until highquality fluorescein angiography can be obtained. Laser photocoagulation should not be considered if macular ischemia is present with macular edema, unless it is clear that improve¬ ment in visual acuity is not spontane¬ ously occurring. Laser photocoagula¬ tion should be considered if good macular perfusion is present with per¬ sistent macular edema, reducing visual acuity without spontaneous improve¬ ment.
The entity of branch vein occlusion has proven to be a unique model for the study of retinal neovascularization,22 as previously reported. The series of cases reported herein suggest that branch vein occlusion may also provide a unique model for the study of certain aspects of macular edema. This study was supported by research grants EY06142 and EY-01765 from the National Eye Institute, Bethesda, Md.
References 1. Goldberg MF. Diseases affecting the inner blood-retinal barrier. In: Cunha-Vaz JG, ed. The Blood-Retinal Barriers. New York, NY: Plenum Press; 1979:309-363. 2. Jampol LM. Macular edema. In: Ryan SJ, ed. Retina. St Louis, Mo: Mosby-Year Book; 1989;2: 81-88. 3. Henkind P, Bellhorn RW, Schall B. Retinal edema: postulated mechanism(s). In: Cunha-Vaz JG, ed. The Blood-Retinal Barriers. New York, NY: Plenum Press; 1979:251-268. 4. Gass JDM. Stereoscopic Atlas Macular Diseases. St Louis, Mo: Mosby-Year Book; 1970:154. 5. Finkelstein D, Clarkson JG, Branch Vein Occlusion Study Group. Branch and central retinal vein occlusions: focal points. Clin Modules Ophthalmol. 1987;5:1-10. 6. Sedney SC. Photocoagulation in Retinal Vein Occlusion. The Hague, the Netherlands: Dr W Junk; 1976. 7. Coscas G, Dhermy P. Occlusions veineuses retinienne. New York, NY: Masson Publishing USA Inc; 1978. 8. Michels RD, Gass JDM. The natural course of
retinal branch vein occlusion. Trans Am Acad
Ophthalmol Otolaryngol. 1974;78:166-177. 9. Gutman FA, Zegarra H. Macular edema secondary to occlusion of the retinal veins. Surv Oph-
thalmol. 1984;28:464-470. 10. Coscas G, Gaudric A. Natural course of nonaphakic cystoid macular edema. Surv Ophthalmol.
1984;28:471-484.
11. Orth D, Patz A. Retinal branch vein occlusion. Surv Ophthalmol. 1978;22:357-376. 12. Clemett RS, Kohner EM, Hamilton AM. The visual prognosis in retinal branch vein occlusion. Trans Ophthalmol Soc U K. 1973;93:523-535. 13. Shilling JS, Jones CA. Retinal branch vein occlusion: a study of argon laser photocoagulation in the treatment of macular oedema. Br J Ophthalmol. 1984;68:196-198. 14. Klatzo I. Neuropathological aspects of brain edema. J Neuropathol Exp Neurol. 1967;26:1-14. 15. Katzman R, Clasen R, Klatzo I, Meyer JS, Pappius HM, Waltz AG. Report of Joint Committee for Stroke Resources, IV: brain edema in stroke. Stroke. 1977;8:512-540. 16. O'Brien MD. Ischemic cerebral edema: a re-
Downloaded From: http://archopht.jamanetwork.com/ by a University of Pennsylvania User on 06/20/2015
view. Stroke. 1979;10:623-628. 17. Branch Vein Occlusion Study Group. Progressing' branch vein occlusion. Invest Ophthalmol Vis Sci. 1983;24(suppl):172. 18. Bresnick GH, Condit R, Syrjala S, Palta M, Groo A, Korth K. Abnormalities of the foveal avascular zone in diabetic retinopathy. Arch Ophthalmol. 1984;102:1286-1293. 19. Branch Vein Occlusion Study Group. Argon laser photocoagulation for macular edema in branch vein occlusion. Am J Ophthalmol. 1984;98: 271-282. 20. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema: Early Treatment Diabetic Retinopathy Study Report Number 1. Arch Ophthalmol. 1985;103:1796-1806. 21. Wilson DJ, Finkelstein D, Quigley HA, Green WR. Macular grid photocoagulation: an experimental study on the primate retina. Arch Ophthalmol. 1988;106:100-105. 22. Wise GN. Factors influencing retinal new vessel formation. Am J Ophthalmol. 1961;52:637\x=req-\ 650.
