97
ferent regions of the retina. Thus, this technique has been
used not only for retinal diseases with localized lesions
but also for macular diseases such as macular holes and
diabetic macular edema [9].
Recently, successful outcomes of pars plana vitrecto-
my for diabetic macular edema have been reported [3,
13]. It would be ideal if cases could be selected with the
greatest possibility for improvement of vision. We real-
ize that the equipment to perform OCT and multifocal
ERGs is not available in the standard ophthalmological
office. We have therefore determined how the data ob-
tained with them correlate with the visual acuity and
fundoscopic appearances.
Methods
Forty-five eyes of 25 patients with clinically significant diabetic
macular edema as defined by the Early Treatment Diabetic Reti-
nopathy Study [1] or cystoid macular edema were studied. The pa-
tients’ ages ranged from 44 to 77 years, and visual acuity mea-
sured with log MAR charts ranged from –0.1 to 1.5. All patients
underwent a complete ophthalmic examination, including indirect
ophthalmoscopy, fluorescein angiography, and best-corrected vi-
sual acuity. Fluorescein angiography did not show ischemic mac-
ulopathy in any of the diabetic eyes in this study. Forty-three eyes
of 24 patients had received photocoagulation for non-perfusion ar-
eas outside the vascular arcade, and none had received macular
photocoagulation prior to this study. The duration of the macular
edema could not be determined in most of the patients, because
the macular edema was detected at their first visit to our clinic.
Both eyes were studied in 20 patients; only one eye was studied in
5 patients because the other eye did not show macular edema by
either ophthalmoscopy or OCT. Twenty-one subjects without oph-
thalmic abnormalities, ranging in age from 43 to 84 years, com-
prised the normal control group.
Fig. 1 Trace array of 61 local responses of multifocal electroret-
inograms obtained from A a 59-year-old normal subject and B a
57-year-old patient with diabetic macular edema. Ovals indicate
seven central hexagons. The calibration markers represents
200 nV vertically and 80 ms horizontally.
This study followed the tenets of the Declaration of Helsinki. In-
formed consent was obtained from all subjects after the nature and
possible consequences of the study had been explained to them.
Optical coherence tomography (OCT) was performed in all pa-
tients using the Humphrey model 2000 (Humphrey Instruments,
San Leandro, Calif.). The patient’s pupil was fully dilated with
topical 0.5% tropicamide, and the fundus was scanned with a
probe beam positioned so that the horizontal and vertical planes
crossed the central fovea, as determined from fundus photographs.
The scan length was usually 2.8 mm. The retinal thickness at the
fovea was measured automatically as the distance between the in-
ner retinal surface and the retinal pigment epithelium.
Macular edema was divided into two types: cystoid macular
edema and diffuse macular edema. Cystoid macular edema was
characterized by intraretinal cystoid spaces in the macular area.
Diffuse macular edema was characterized by increased retinal
thickness with reduced intraretinal reflectivity and expanded areas
of lower reflectivity. Serous retinal detachment in the macular area
was not detected by OCT in any of the patients.
Multifocal ERGs were recorded with the Visual Evoked Re-
sponse Imaging System (VERIS Science, Mayo, Nagoya, Japan).
The stimulus matrix consisted of 61 hexagons that were arranged
concentrically and covered approximately 50 deg of the central vi-
sual field. The size of each hexagon was scaled to give approxi-
mately equal ERG responses. Each hexagon was independently al-
ternated between black and white according to a pseudorandom
binary m-sequence at a rate of 75 Hz.
patient’s pupil was fully dilated with topical 0.5% tropicamide,
and the opposite eye was occluded. After the patients had under-
gone optical correction for a viewing distance of 30 cm, they were
instructed to maintain their fixation on the fixation spot, or the cu-
neiform indicator for patients who could not see the spot. The sig-
nals were amplified 100,000 times with a bandpass filter of
10–300 Hz. The recording period was comprised of eight seg-
ments of 30 s, providing a total recording time of 4 min. Because
the responses from the central hexagon were very small and not
measurable in many diabetic cases, the responses from the seven
central hexagons, covering approximately 10 deg, were summed
and used as the macular response (Fig. 1).
Cone ERGs to full-field flash stimuli were recorded from 32
of the 46 diabetic eyes and in 16 of the 21 normal control eyes
in order to evaluate the cone function over the entire retina.
A Ganzfeld stimulator provided white flash stimuli and white
background illumination (50 cd/m2). The stimulus frequency was
5 Hz, and 20 responses were averaged using a Neuropack 2 avera-
ger (Nihon Kohden, Tokyo).
Statistical comparisons between the diabetic patients with cys-
toid macular edema, those with diffuse macular edema, and nor-
mal subjects were performed by means of ANOVA and Scheffé’s
F tests.
A bipolar Burian-Allen contact lens electrode was used for the
recording, and a ground electrode was placed on the earlobe. The