Reappraisal of Biomicroscopic Classification of

Stages of Development of a Macular Hole

J. DONALD M. GASS, M.D.

• PURPOSE: To update the biomicroscopic classifi­ contain no retinal receptors (pseudo-opercula).

cation and anatomic interpretations of the stages Surgical reattachment of the retina surrounding the
of development of age-related macular hole and hole and centripetal movement of the foveolar
provide explanations for the remarkable recovery retina induced by gliosis may restore foveal anatomy
of visual acuity that occurs in some patients after and function to near normal.
vitreous surgery.

K
• METHODS: Recent biomicroscopic observations ELLY AND WENDELL1 AND OTHERS 24 HAVE REPORT-
of various stages of macular holes are used to ed excellent recovery of visual acuity after
postulate new anatomic explanations for these vitrectomy in some patients with age-related
stages. macular hole. As many as 48% of patients undergoing
• RESULTS: Biomicroscopic observations include surgery may regain visual acuity of 20/40 or better.4
the following: (1) the change from a yellow spot The rationale for the surgery and the presumed
(stage 1-A) to a yellow ring (stage 1-B) during the explanation for its success was reattachment of the
early stages of foveal detachment is unique to retina surrounding the macular hole. It seems unlike­
patients at risk of macular hole; (2) the prehole ly, however, that reattachment of the 200- to 500-
opacity with a small stage 2 hole may be larger μπι-wide rim of retinal detachment surrounding a
than the hole diameter; and (3) the opacity resem­ 400- to 500-μπι defect in the foveolar retina can be
bling an operculum that accompanies macular the only explanation for the remarkable visual im­
holes is indistinguishable from a pseudo-opercu- provement that occurs in some patients after surgery.
lum found in otherwise normal fellow eyes. The present classification of stages of macular hole
• CONCLUSIONS: The change from a yellow spot development is based largely on biomicroscopic obser­
(stage 1-A) to a yellow ring (stage 1-B) is caused vations and presumed anatomic changes, not clinico-
primarily by centrifugal displacement of retinal pathologic correlations.5 Previous anatomic interpreta­
receptors after a dehiscence at the umbo. The hole tions of these stages have suggested that most macular
may be hidden by semiopaque contracted prefoveo- holes develop as the result of a 360-degree tear in the
lar vitreous cortex bridging the yellow ring (stage peripheral foveola, with loss of the central retina in
1-B occult hole). Stage 1-B occult holes become formation of an operculum, which is demonstrable in
manifest (stage 2 holes) either after early separa­ approximately 75% of macular holes.5
tion of the contracted prefoveolar vitreous cortex The purposes of this report are to update the
from the retina surrounding a small hole or as an biomicroscopic classification and anatomic interpreta­
eccentric can-opener-like tear in the contracted tion of the stages of development of age-related macu­
prefoveolar vitreous cortex, at the edge of larger lar holes and to provide possible explanations for the
stage 2 holes. Most prehole opacities probably unexpected good visual results after surgical treatment
of patients with this disorder. This reinterpretation of
Accepted for publication Nov. 11, 1994. biomicroscopic observations suggests the following: (1)
From the Bascom Palmer Eye Institute, Department of Ophthalmol­ that most macular holes begin as an occult central
ogy, University of Miami Medical Center, Miami, Florida.
No reprints are available. neurosensory retinal dehiscence at the umbo followed

752 © AMERICAN JOURNAL OF OPHTHALMOLOGY 119:752-759 JUNE 1995

by centrifugal retraction and concentration of the in patients at risk for developing a macular hole.
nearly full complement of retinal receptors, (2) that Biomicroscopic evidence of a full-thickness macular
this mechanism of hole formation is usually accompa­ hole may appear initially as (1) a small round central
nied by the appearance of a prehole opacity consist­ hole in the center of the yellow ring in the absence of
ing of the contracted and condensed vitreous cortex a prehole vitreous opacity, (2) as a round central hole
but no retinal receptor cells (pseudo-operculum), and associated with a prehole vitreous opacity that may be
(3) that a peripheral foveolar tear in the neurosensory larger in diameter than the hole, and (3) most often
retina with operculum formation probably occurs as an eccentric hole beginning at the inner margin of
infrequently. the yellow ring and later extending 360 degrees, to
The following key concepts redefine the anatomic form a prehole vitreous opacity smaller in diameter
interpretation of the biomicroscopic classification: (1) than the hole. Approximately 75% of fully developed
while spontaneous contraction and condensation of macular holes are associated with a prehole opacity
the vitreous cortex probably involve a broad area of resembling an operculum. These opacities are indis­
the posterior vitreoretinal interface, they are focally tinguishable biomicroscopically from pseudo-opercula
intense in the region of the foveolar and perifoveolar that occur after spontaneous vitreofoveal separation
area; (2) this focal condensation of the prefoveolar in eyes with normal function and biomicroscopic
vitreous cortex, which is normally transparent, causes appearance.
it, in most cases, to become semitransparent and Figure 1 illustrates diagrammatically the presumed
biomicroscopically visible as an interface, when it anatomy of the various stages of macular hole devel­
bridges a macular hole, or as a prefoveolar opacity opment. The Table summarizes the biomicroscopic
indistinguishable from an operculum, after vit- classification of stages of development of a macular
reofoveal separation; (3) the retinal xanthophyll, hole, as well as the old and new anatomic interpreta­
which is highly concentrated within the retinal tion of these stages. Spontaneous tangential contrac­
receptors comprising the 400- to 500^m-diameter tion of the external part of the prefoveolar cortical
foveolar retina, is most likely responsible for the vitreous detaches the foveolar retina, and a yellow
yellow spot seen in stage 1-A lesions; and (4) the spot appears (stage 1-A hole; Fig IB). As the foveal
yellow ring of stage 1-B lesions, particularly in its later retina elevates to the level of the surrounding thick
stages of development, is probably primarily caused by perifoveal retina, the retinal receptor layer is stretched
centrifugal displacement of the foveolar receptor cells or elongated, and thinning of the foveolar retina
with xanthophyll and not by displacement of the around the umbo causes a change from yellow spot to
xanthophyll alone. a small donut-shaped yellow ring lesion (stage 1-B,
impending macular hole; Fig. 1C). This change from
a yellow spot to a ring is followed by a break in the
continuity of the receptor cell layer at the umbo,
MATERIAL AND METHODS
structurally the weakest point in the retina. The
SINCE PUBLICATION OF THE BIOMICROSCOPIC CLASSIFY retinal receptors, their radiating nerve fibers, and the
cation of age-related macular holes in 1987, I have xanthophyll retract centrifugally beneath the con­
examined more than 100 patients with various stages tracted vitreous cortex, and the yellow ring enlarges
of macular hole formation. and develops a more well defined central semitranslu-
cent zone (stage 1-B lesion, occult macular hole; Figs.
ID and 2, top left). Initially, the internal limiting
membrane of the foveolar retina and the thin layer of
RESULTS horizontally oriented Mueller cell processes separating
ALTHOUGH A YELLOW SPOT IN THE CENTER OF THE
it from the retinal receptor cells may not be involved
fovea may occur biomicroscopically in any disorder in the central retinal break. Regardless, as long as the
that causes serous foveal retinal detachment, the contracted prefoveolar vitreous cortex bridges the
change from a yellow spot (stage 1-A) to a yellow ring hole, it may be visible biomicroscopically as a semi-
(stage 1-B) is peculiar to foveal detachment occurring translucent interface. Thus the change from a stage

VOL.119, No. 6 REAPPRAISAL OF MACULAR HOLE CLASSIFICATION 753

 A. NORMAL FOVEA

S
ï t t * . - » »-*. ..
' · : — ' ■y-·-'--

—!■ iliniril-....' * - ' - ■ ■ · - "

— ■ ■ - —

B. STAGE 1-A IMPENDING HOLE C. STAGE 1-B IMPENDING HOLE

S
t w o w v u w. .t. - ' ^ , ι τ ι ^ . · . . - ' ·■··- ■
^ - —
:«Η.ί.·>ιι ■■ I M ■■ ■■Trfi'if'iTl
D. STAGE 1-B OCCULT HOLE E. STAGE 1-B OCCULT HOLE

EHÊÊÎ

F. STAGE 2 HOLE G. STAGE 2 HOLE

J c IfliB^B
H. STAGE 3 HOLE I. STAGE 4 HOLE

Fig. 1 (Gass). Stages of development of a senile macular hole. A, Normal fovea. Layer of vitreous cortex (vc) lying on
internal limiting membrane of retina. B, Stage 1-A impending hole. Early contraction of outer part of vitreous cortex
with foveolar detachment. C, Stage 1-B impending hole. Further vitreous contraction and condensation of the
prefoveolar vitreous cortex with foveal detachment. D and E, Stage 1-B occult hole. Dehiscence of the retinal receptor
layer at the umbo with centrifugal retraction of the retinal receptors. F, Stage 2 hole with early separation of condensed
prefoveolar vitreous cortex with formation of pseudo-operculum that is larger than the hole. G, Stage 2 hole with tear in
vitreous cortex at junction of the prefoveolar vitreous cortex and edge of macular hole. H, Stage 3 hole with
pseudo-operculum. I, Stage 4 hole after posterior vitreous separation.

1-B impending hole to a stage 1-B occult hole cannot vitreous cortex becomes visible as a semitranslucent
be detected biomicroscopically. Reactive proliferation prehole opacity lying anterior to a small foveolar hole
of Mueller cells and retinal astrocytes occurring (stage 2 hole; Fig. IF). Initially the diameter of this
within the area of the receptor cell dehiscence may opacity is often larger than that of the foveolar hole.
contribute to the opacification of the tissue bridging In a few patients with early stage 1-B lesions, separa­
the defect and, in some cases, may cause ruffled edges tion of the prefoveolar vitreous cortex may be accom­
of the retinal dehiscence surrounded by fine radiating panied by avulsion of part of the foveolar retina and
retinal folds (Fig. 2, top left). Spontaneous vitreofo- by true operculum formation. Biomicroscopy, howev­
veal separation may occur soon after the central er, cannot determine the presence or absence of
retinal dehiscence, and the contracted prefoveolar retinal tissue in the prehole opacity. In some patients,

754 AMERICAN JOURNAL OF OPHTHALMOLOGY JUNE 1995

 TABLE

BIOMICROSCOPIC CLASSIFICATION OF AGE-RELATED MACULAR HOLE

ANATOMIC INTERPRETATION

STAGE BIOMICROSCOPIC FINDINGS OLD NEW

1 -A (impending Central yellow spot, loss of foveolar Early serous detachment of Same
hole) depression, no vitreofoveolar foveolar retina
separation
1-B (impending Yellow ring with bridging interface, Serous foveolar detachment with Same for small ring. For larger
or occult hole) loss of foveolar depression, lateral displacement of xanthophyll ring, central occult foveolar hole
no vitreofoveolar with centrifugal displacement of
separation foveolar retina and xanthophyll,
with bridging contracted
prefoveolar vitreous cortex.
Cannot detect transition from
impending to occult hole.
2 Eccentric oval, crescent, or horse­ Hole (tear) in peripheral foveolar Hole (tear) in contracted
shoe retinal defect inside edge retina prefoveolar vitreous bridging
of yellow ring round retinal hole, no loss of
foveolar retina
Central round retinal defect with
rim of elevated retina
With prefoveolar opacity Hole with operculum,* rim of Hole with pseudo operculum,f
retinal detachment rim of retinal detachment
Without prefoveolar opacity Hole, no posterior vitreous Same
detachment from optic disk
and macula
3 Central round 2 400 μηι diameter
retinal defect, no Weiss's ring,
rim of elevated retina
With prefoveolar opacity Hole with operculum, no posterior Hole with pseudo-operculum,
vitreous detachment from no posterior vitreous
optic disk and macula detachment
Without prefoveolar opacity Hole, no posterior vitreous Same
detachment from optic disk
and macula
4 Central round retinal defect, rim
of elevated retina, Weiss's ring
With prefoveolar opacity* Hole with operculum and posterior Hole with pseudo-operculum
vitreous detachment from and posterior vitreous
optic disk and macula detachment from optic
disk and macula
Without prefoveolar opacity Hole and posterior vitreous Same
detachment from optic
disk and macula

'Operculum contains foveolar retina

tPseudo-operculum contains no retinal receptors.
'Usually found near temporal border of Weiss's ring.

the contracted prefoveolar vitreous cortex, either scopically. In such cases very small stage 2 holes
while it remains attached to and bridges the macular without a prehole opacity may be evident. In most
hole or after it separates from the perifoveolar retina, patients, however, the contracted vitreous cortex is
may be transparent and undetectable biomicro- semitransparent and remains attached to the inner

VOL.119, N o . 6 REAPPRAISAL OF MACULAR HOLE CLASSIFICATION 755

 Fig. 2 (Gass). Top left, Right eye examined May 7, 1984.
Stage 1-B lesion, probably occult hole. Note ruffled
inner edge of yellow ring and radiating retinal stria. Top
right, Right eye examined Aug. 6, 1984. Stage 2 retinal
hole (arrowheads) with oval dehiscence (arrow) in
contracted vitreous cortex at the superior margin of the
retinal hole. Note that the yellow ring at the edge the
macular hole is serrated and most prominent along its
inferior half. Bottom left, Left eye of same patient Oct.
15, 1984. Large stage 2 hole (arrowheads) with reniform
dehiscence (arrow) in the contracted vitreous cortex at
the inferotemporal edge of the macular hole. Note the
serration and prominence of the yellow ring along the
nasal edge of the macular hole.

retinal surface surrounding the retinal hole as the hole, and the yellow pigmentation fades, possibly as a
foveolar retina continues to retract centrifugally result of diffusion of xanthophyll out of the retina in
(stage 1-B occult hole; Fig. IE). Biomicroscopically this area. Over a period of days or weeks, further
there is progressive enlargement of the yellow ring, enlargement of the macular hole and additional
which may become serrated along the inner margin contraction of the prefoveolar vitreous cortex cause a
that corresponds to the edge of the occult round 360-degree tear in the contracted prefoveolar vitreous
retinal hole (Fig. 2). Eventually the first biomicro- cortex at the edge of the retinal hole, separating it
scopic evidence of a dehiscence may occur in the from the less condensed vitreous cortex surrounding
semitransparent vitreous cortex at the inner edge of the hole. The prefoveal vitreous cortex is visible
the yellow ring (stage 2 hole; Figs. IG and 2, top right biomicroscopically as an opacity (pseudo-operculum)
and bottom left). In the area of the dehiscence, the suspended anterior to the hole on the posterior
serration of the yellow ring disappears, presumably surface of the thin layer of transparent vitreous gel
because of relief of traction on the edge of the retinal that bridges the hole and lies along the inner retinal

756 AMERICAN JOURNAL OPHTHALMOLOGY JUNE 1995

surface in the macula. Centrifugal retraction of the retinal opercula have been observed histopathologi-
foveolar retinal receptors continues until the diameter cally in two eyes, one with posttraumatic macular
of the hole, in all but a few cases, reaches 400 to 600 hole and the other with an idiopathic macular hole,
μπι (stage 3 hole; Fig. 1H). All stages of progressive they have not been observed in most idiopathic
enlargement of the hole are considered to be stage 2 macular holes studied histopathologically.6·7
holes. Since the ultimate diameter of the hole is The yellow spot seen in a stage 1-A impending
variable, for purposes of classification, I suggest that hole is seen in serous detachment of the foveal retina
all holes less than 400 μπι in diameter be considered caused by other disorders, for example, idiopathic
as stage 2 holes. After separation of the vitreous from central serous retinopathy and epiretinal membranes.8
the entire macular surface and optic disk, the hole is The change from the yellow spot to a yellow ring,
designated as stage 4, irrespective of its diameter (Fig. however, occurs only with foveolar detachment
II). caused by the anatomic changes and forces generated
at the vitreoretinal interface occurring during macu­
lar hole formation. Approximately 50% of the pa­
tients who initially have stage 1 lesions experience
DISCUSSION spontaneous vitreofoveal detachment and return of
visual acuity to near normal levels.5,9 Most of these
IF THE ANATOMIC INTERPRETATIONS SUMMARIZED IN patients have stage 1-B lesions at the time of initial
Figure 1 are correct, then the implications include the examination.5 The macula in these patients typically
following: (1) most macular holes develop as the returns to a normal appearance or, in approximately
result of a central retinal dehiscence at the umbo, one third of patients, shows evidence of one or more
followed by centrifugal displacement of the relatively inner lamellar holes. Most, but not all, patients whose
normal complement of retinal receptors; (2) this macula returns to normal have a pseudo-operculum
dehiscence occurs soon after the change from a suspended immediately in front of the foveola. These
yellow spot (stage 1-A impending hole) to a yellow findings provide further evidence that some, or per­
ring lesion (stage 1-B impending hole), but in most haps most, of the prefoveolar opacities in stage 2 and
cases it is not detectable with a thin slit beam as a 3 holes are pseudo-opercula. Failure to find pseudo-
defect in the center of the ring because of the opercula in all patients after spontaneous vitreofoveal
presence of the semitranslucent condensed cortical separation suggests that the contracted prefoveolar
vitreous that bridges the hole (stage 1-B occult hole); vitreous cortex in some patients is transparent and
(3) most of the prehole opacities overlying stage 2 and not detectable. Inner lamellar holes presumably devel­
stage 3 holes are condensed prefoveal vitreous cortex op as the result of a superficial dehiscence of the
(pseudo-opercula), not opercula; and (4) after suc­ internal limiting membrane during stage 1-A or early
cessful vitreous surgery, which includes tamponade of stage 1-B macular hole formation.
the hole with an intravitreal gas bubble and which is Although no data are available concerning the
done within one year after commencement of hole relative size of yellow ring lesions that progress to a
formation, the anatomy of the central retina and its macular hole compared with those that do not, it is
visual function may be restored to near normal levels probable that there is a direct relationship between
in some patients as a result of retinal reattachment the ring diameter, visual acuity, and the chance for
and centripetal repositioning of the retinal receptors. hole development. Some patients with relatively large
If this description of the anatomic changes occur­ stage 1-B yellow rings, however, may abort the
ring in macular hole development is correct, histo- process, with return of the macula to normal appear­
pathologic examination of the prehole opacities ance and function. In such cases the retinal internal
should show that most of them are composed of limiting membrane, Mueller cell processes, and prolif­
vitreous cortical collagen and are accompanied in erating retinal glial tissue, which bridge the defect
some cases by internal limiting membrane of the caused by centrifugal displacement of the neurosen-
retina, Mueller cell processes, and reactive glial prolif­ sory retina, may act as a scaffold and stimulus to
eration, but not by retinal receptor cells. Although centripetal replacement of the foveolar retina to its

VOL.119, No. 6 REAPPRAISAL OF MACULAR HOLE CLASSIFICATION 757

normal position after spontaneous vitreofoveal sepa­ centripetal movement of the paracentral retinal re­
ration and retinal reattachment. The pseudo- ceptors and their xanthophyll may occur after retinal
operculum that develops in some patients after spon­ reattachment.11 This hypothesis is supported also by
taneous retinal reattachment may cast a yellow shad­ the histopathologic studies of Funata and associates12
ow on the retina during slit-lamp biomicroscopy.10 and Madreperla and associates.13 Gross anatomic and
This observation is further evidence suggesting a histopathologic examination of both eyes of a patient
central retinal receptor dehiscence that allows diffu­ whose visual acuity improved from 20/400 to 20/40 in
sion of the retinal xanthophyll into the condensed both eyes after macular hole surgery showed closure of
vitreous cortex before its separation from the retinal the macular holes, reattachment of the retina, and
surface. After reattachment of the retina and closure return of xanthophyll to the center of the macula. In
of the receptor cell defect, which occur either sponta­ the left eye, closure of the hole was associated with
neously or after vitreous surgery, the mild hyperfluo- glial proliferation and probable centripetal displace­
rescence usually evident in stage 1-B lesions and the ment of the retinal receptors. In the right eye, retinal
more intense fluorescence present in most stage 2, 3, reattachment and merging of the edges of the hole
and 4 holes are no longer demonstrable in most were not associated with gliosis. In another patient
patients. This disappearance of the hyperfluorescence preoperative visual acuity was 20/80 and the postop­
suggests that there is a return of xanthophyll to the erative visual acuity was 20/40.13 One month postop-
retinal receptors as well as glial tissue in the foveolar eratively the patient died, and histopathologic exami­
area. nation showed closure of the hole and close
Before the appearance of a slit-beam defect within approximation of the retinal receptors centrally by
the yellow ring, I have been unable to determine proliferating Mueller cells.13
which stage 1-B lesions will progress and which will In summary, I postulate that one of the important
abort. The size of the ring and the level of visual reasons for the reported surgical success in the
acuity will probably be helpful but not completely treatment of macular holes is that, contrary to
reliable in predicting progression. In a previous study, previous belief, most macular holes begin as a central
I found that seven of ten patients with stage 1 retinal dehiscence and not as the result of a foveolar
impending holes that progressed to a stage 3 hole had tear and operculum formation. Circumstantial evi­
visual acuity of 20/70 or worse at the initial examina­ dence supports the following hypotheses: (1) the
tion, whereas seven of nine patients who aborted the progressively enlarging yellow ring seen during the
process had visual acuity of 20/50 or better at the early stages of development of a macular hole is
initial examination.5 Until more data concerning the caused by centrifugal displacement of the foveolar
relationship between ring diameter and risk of hole retinal receptor cells, as well as the xanthophyll, after
formation and concerning the success of vitreous a central retinal dehiscence; (2) many of the lesions
surgery are available, it is advisable to observe patients that we now classify as stage 1-B, because of a
for biomicroscopic evidence of a full-thickness dehis­ biomicroscopically discernible interface within the
cence within the yellow ring and a visual acuity of yellow ring, are occult holes that began centrally soon
20/70 or worse before surgery is considered. Even the after the change from a yellow spot to a yellow ring
presence of a biomicroscopic defect within the ring lesion; (3) stage 1-B occult holes in some patients may
does not absolutely guarantee that a full-thickness close spontaneously after vitreoretinal separation; (4)
hole is present. It is probable that in some patients the the eccentric holes that develop near the inner
contracted prefoveolar vitreous cortex and the attenu­ margin of the yellow ring are tears in the condensed
ated foveolar retina are sufficiently transparent that vitreous cortex and not the retina; and (5) most of
they cannot be detected biomicroscopically. the prehole vitreous opacities previously referred to as
The recovery of 20/20 visual acuity, the disappear­ retinal opercula are instead contracted prefoveolar
ance of the focal hyperfluorescence corresponding vitreous cortex or pseudo-opercula. Until these hy­
with the hole angiographically, and the disappear­ potheses are either verified or refuted by light and
ance of absolute central scotomas observed in some electron microscopic examination of prehole opaci­
patients after macular hole surgery indicate that ties, I suggest caution in making a diagnosis of a

758 AMERICAN JOURNAL OF OPHTHALMOLOGY JUNE 1995

macular hole before the patient develops biomicro- 6. Frangieh GT, Green WR, Engel HM. A histopathologic
study of macular cysts and holes. Retina 1981;1:311-36.
scopic evidence of either a central or paracentral 7. Guyer DR, Green WR, de Bustros S, Fine SL. Histopatholog­
full-thickness hole in the foveolar area. ic features of idiopathic macular holes and cysts. Ophthal­
mology 1990;97:1045-51.
8. Gass JDM, Joondeph BC. Observations concerning patients
with suspected impending macular holes. Am J Ophthalmol
REFERENCES 1990;109:638-46.
9. de Bustros S. Vitrectomy for prevention of macular holes:
results of a randomized multicenter clinical trial. Ophthal­
1. Kelly NE, Wendel RT. Vitreous surgery for idiopathic macular mology 1994;101:1055-9.
holes: results of a pilot study. Arch Ophthalmol 1991;
10. Gass JDM, Van Newkirk M. Xanthic scotoma and yellow
109:654-9.
foveolar shadow caused by a pseudo-operculum after vitreofo-
2. Glaser BM, Michels RG, Kuppermann BD, Sjaarda RN, Pena
veal separation. Retina 1992;12:242-4.
RA. Transforming growth factor-beta 2 for the treatment of
11. Sjaarda RN, Frank DA, Glaser BM, Thompson JT, Murphy
full-thickness macular holes: a prospective randomized study.
RP. Resolution of an absolute scotoma and improvement of
Ophthalmology 1992;99:1162-72.
relative scotoma after successful macular hole surgery. Am J
3. Poliner LS, Tornambe PE. Retinal pigment epitheliopathy Ophthalmol 1993;116:129-39.
after macular hole surgery. Ophthalmology 1992;99:1671-7.
12. Funata M, Wendel RT, de la Cruz Z, Green WR. Clinico-
4. Lansing MB, Glaser BM, Liss H, Hanham A, Thompson JT,
pathologic study of bilateral macular holes treated with pars
Sjaarda RN, et al. The effect of pars plana vitrectomy and
plana vitrectomy and gas tamponade. Retina 1992; 12:
transforming growth factor-beta 2 without epiretinal mem­
289-98.
brane peeling on full-thickness macular holes. Ophthalmolo­
13. Madreperla SA, Geiger GL, Funata M, de la Cruz Z, Green
gy 1993;100:868-71.
WR. Clinicopathologic correlation of a macular hole treated
5. Gass JDM. Idiopathic senile macular hole: its early stages and
by cortical vitreous peeling and gas tamponade. Ophthalmol­
pathogenesis. Arch Ophthalmol 1988;106:629-39.
ogy 1994;101:682-6.

VOL.119, No. 6 REAPPRAISAL OF MACULAR HOLE CLASSIFICATION 759