LXII Edward Jackson Lecture: Open Angle
Glaucoma After Vitrectomy
STANLEY CHANG, MD
● PURPOSE: To present data and an hypothesis for the
late development of open angle glaucoma (OAG) after
vitrectomy.
● DESIGN: A retrospective observational case series.
● METHODS: The records of 453 eyes that had under-
gone vitrectomy were reviewed for postoperative OAG.
Eyes with confounding factors were excluded. Sixty-eight
eyes of 65 patients that underwent routine vitrectomy
were followed for a mean of 56.9 months (range, seven to
192 months). For the main outcome measures, patients
were classified into three groups: patients with suspected
glaucoma, patients in whom glaucoma developed after the
operation, and patients with pre-existing glaucoma.
● RESULTS: In glaucoma suspects, the mean intraocular
pressure was significantly higher in the operated eye
compared with the fellow eye (P ⴝ .0001). In eyes with
new onset glaucoma, 23 of 34 eyes (67.6%) had it in the
vitrectomized eye only. In phakic eyes, the time interval
between vitrectomy and the development of glaucoma
(mean, 45.95 months) was significantly longer than eyes
that were nonphakic at the time of vitrectomy (mean,
18.39 months; P ⴝ .0115). When the interval between
cataract surgery in phakic eyes to the development of
glaucoma was compared with the interval from vitrec-
tomy to glaucoma diagnosis in the nonphakic group, the
difference was not statistically significant. In eyes with
glaucoma before the operation, the mean number of
antiglaucoma medications that were required to control
the intraocular pressure was significantly higher in the
vitrectomized eye, compared with the fellow eye (2.9
Accepted for publication Feb 6, 2006.
From the Edward S. Harkness Eye Institute, Department of Ophthal-
mology, Columbia University, New York, New York.
Supported, in part, by an unrestricted grant to the Department of
Ophthalmology from Research to Prevent Blindness, Inc, New York, New
York, the K.K. Tse and Ku Teh Ying Endowed Professorship, and the
Louis V. Gerstner, Jr, Clinical Research Center, Columbia University
Medical Center, New York, New York.
I would like to thank Norihiko Yoshida, MD, and Paulo Escario, MD,
Columbia University, for their diligence and precision in the review and
statistical analysis of massive amounts of data and Michael Chiang, MD,
and James Tsai, MD, Columbia University, for their constructive review
of the study design and their statistical support.
Inquiries to Stanley Chang, MD, Edward S. Harkness Eye Institute,
635 West 165th St, New York, NY 10032; e-mail: sc434@columbia.edu
0002-9394/06/$32.00 © 2006 BY ELSEVIER INC. ALL
doi:10.1016/j.ajo.2006.02.014
medications ⴞ 1.2 vs 2.0 medications ⴞ 1.4; P ⴝ .0215;
n ⴝ 14).
● CONCLUSION: There is an increased risk of OAG after
vitrectomy. The presence of the lens may be protective.
In established OAG before the operation, the number
of antiglaucoma medications may increase after sur-
gery. Oxidative stress is hypothesized to have a role in
the pathogenesis. (Am J Ophthalmol 2006;141:
1033–1043. © 2006 by Elsevier Inc. All rights re-
served.)
I
AM MOST GRATEFUL TO THE AMERICAN ACADEMY OF
Ophthalmology and the American Journal of Ophthal-
mology for their invitation to give this lecture honoring
the achievements of one of the “founding fathers” of Amer-
ican Ophthalmology, Dr Edward Jackson. Joining ranks with
61 of the most prominent leaders in ophthalmology who have
given this lecture previously is a distinction that is both
absolutely incredible but also humbling.1 It has been exactly
40 years since the last Jackson Memorial Lecture was given
by an ophthalmologist from New York, Dr Irving Leopold.
At that time, Dr Leopold had a faculty appointment at
Columbia University before taking the chair at Mount
Sinai School of Medicine the following year. Four other
Columbia faculty members have given this distinguished
lecture: John Dunnington, Algernon Reese, Ludwig Von
Sallman, and George Smelser. I am very pleased to share
this honor.
Dr Jackson exhibited leadership and excellence in many
facets of ophthalmology (Figure 1).2,3 He was the quintes-
sential clinician who was analytical and precise in applying
his broad experience and knowledge to solving clinical
problems. He was a prolific author who wrote ⬎600
articles and many books on ophthalmology and refraction.
He led the movement to elevate the standards for training
ophthalmologists and medical students in eye disorders. He
was greatly respected and admired by his peers and was a
founder and first leader of professional organizations such
as the Western Ophthalmologic, Otological, Laryngologi-
cal, and Rhinological Association (later renamed the
American Academy of Ophthalmology and Otolaryngol-
ogy), the American Board of Ophthalmology, and first
Editor-in-Chief of the American Journal of Ophthalmology.
RIGHTS RESERVED. 1033

FIGURE 1. Edward Jackson.
Above all, Edward Jackson was the ultimate teacher. He
organized the first postgraduate course in ophthalmology in
the United States and was always sought out by his
residents and associates with their interesting patients. He
was dedicated, kind, and generous with his time to his
students. He is responsible for introducing the concept of
instruction courses at the annual American Academy of
Ophthalmology meeting, a program that started in 1921.
Education is vital to the progress of ophthalmology. It is
with this spirit that I would like to dedicate this lecture to
the many teachers who have shared their knowledge,
transferred their skills, and given guidance professionally
and personally throughout my career. Their interest and
mentorship have made me successful.
Since the revolutionary introduction of vitrectomy in
1971 by Machemer and associates,4,5 this procedure has
become the third most frequently performed ophthalmic
surgical operation, after cataract and excimer laser refrac-
tive surgery. It is estimated that approximately 225,000
vitrectomies are done in the United States each year. The
known long-term complications of vitrectomy are rela-
1034 AMERICAN JOURNAL
FIGURE 2. Preoperative fundus drawing of the right eye
(Patient 1) with large posterior retinal tears and retinal detach-
ment. The intraocular pressures were OD 17 mm Hg, OS 18
mm Hg.
tively few. In the phakic eye, the lens develops a progres-
sive increase in nuclear sclerosis, eventually requiring
cataract surgery.6 The pseudophakic eye that undergoes
vitrectomy is more likely to need a posterior capsulotomy
with YAG laser.7 In the posterior segment, there is a
potential for late retinal detachment, often in the areas of
the sclerotomy incisions, when tears develop in the vitre-
ous base from vitreous incarceration at the sclerotomy
sites.
After vitrectomy, a rise in intraocular pressure (IOP) is
not an uncommon event in the immediate postoperative
period.8 –12 Often the pressure rise is related to the use of
viscoelastics, an expanding gas bubble,12 silicone oil tam-
ponade, hemorrhage, inflammation, or corticosteroid re-
sponse. Frequently, medical treatment is effective until the
surgical healing response subsides. Infrequently, angle clo-
sure glaucoma occurs, sometimes resulting in a permanent
glaucoma. In eyes with silicone oil, late glaucoma may
develop from emulsification of the oil or from the sudden
passage of oil into the anterior chamber that obstructs the
passage of fluid into the trabecular meshwork.13
The development of new open angle glaucoma (OAG)
as a late event after vitrectomy has had little attention.8,14
In most case series studies of vitrectomy outcomes, there
are little data that indicate that OAG is a late complica-
tion. The clinical courses of the following three patients
illustrate the possibility that vitrectomy may predispose the
eye to OAG.
OF OPHTHALMOLOGY JUNE 2006
 FIGURE 4. Preoperative fundus drawing (Patient 2) with large
posterior horseshoe tear that caused a macular-sparing retinal
detachment. Vitrectomy, perfluorocarbon liquid, endolaser, and
gas tamponade were done without scleral buckling. The in-
traocular pressures were normal throughout the perioperative
period.

FIGURE 3. Pre- and postoperative photographs (Patient 1).

(Top) A large posterior tear at 12 o’clock was treated with
vitrectomy, perfluorocarbon liquid, endolaser, and gas tampon-
ade. No scleral buckle was used. (Bottom) After the operation,
the visual acuity was 20/25.

CASE STUDIES
● PATIENT 1: A 57-year-old Caucasian man was referred
in September 2000 for large retinal tears in the right eye.
Four years earlier, he had had a 120-degree giant retinal
tear with three inferior horseshoe tears in the left eye that
FIGURE 5. Postoperative fundus montage of attached retina
were treated with laser photocoagulation. Three days
(Patient 2); laser photocoagulation was placed only around the
before my examination, he noted floaters in the right eye, retinal tears. The corrected visual acuity was 20/30.
and an inferior shadow developed two days later. The
visual acuity in the right eye was 20/20 with correction.
There was mild vitreous debris and hemorrhage. A bullous Because of the large posterior tears, the retinal detach-
retinal detachment was present from 10:30 to 2 o’clock, with ment in the right eye was managed by primary vitrectomy,
two large horseshoe tears at 12 and 1:30 o’clock (Figures 2 perfluorocarbon liquid tamponade, endophotocoagulation,
and 3, Top). Two additional horseshoe tears were located and perfluorocarbon gas injection. A scleral buckle was not
equatorially at 6:00 and at 7:30 o’clock. In the left eye the done. Postoperatively, one month later, the retina was
retina was attached, and the tears were surrounded by completely attached with reabsorption of the gas bubble
laser photocoagulation scars. The IOPs were OD 17 mm and corrected visual acuity of 20/25. One year later and
Hg, OS 18 mm Hg. receiving no ocular medications, his corrected visual acuity

VOL. 141, NO. 6 OPEN ANGLE GLAUCOMA AFTER VITRECTOMY 1035

FIGURE 6. Patient 2. Four years after vitrectomy, the intraoc-
ular pressure was OD 23 mm Hg, OS 17 mm Hg. The patient
was followed as a glaucoma suspect for 12 months until
asymmetry of the optic cups was noted with enlargement in the
right eye. Timolol 0.5% was started in the right eye only.
was 20/40 in the right eye, and the IOP was normal. A
nuclear cataract was noted in the right eye. The patient
elected to proceed with uneventful phacoemulsification
and intraocular lens implantation in the right eye in October
2001, thirteen months after vitrectomy. His IOP remained
normal initially after the procedure, but eight months after
cataract surgery IOP had increased to 24 mm Hg. Gonioscopy
showed an open angle. Three months later, the IOP was
29 mm Hg, and timolol 0.5% was started. In January 2003
with timolol therapy, the IOP was 28 mm Hg, and Cosopt
was started twice daily. One year later, the IOP increased to
30 mm Hg while he was receiving Cosopt; twice daily
latanoprost 0.005% was added daily. The IOP in the left eye
was normal throughout this period without any medications.
At the last visit 55 months after vitrectomy, the visual acuity
remains OD 20/25 ⫹ 2, OS 20/70 (Figure 3, Bottom). The
IOPs were OD 19 mm Hg, OS 15 mm Hg. The medica-
tions are Cosopt twice daily, latanoprost hs (at bedtime),
in the right eye only.
● PATIENT 2: A 43-year-old African-American woman
was first examined in January 1991 with symptoms of
floaters in the right eye and was found to have visual acuity
of 20/20. An operculated retinal hole was found at 8
o’clock at the equator with paravascular lattice degenera-
tion from 11 to 12 o’clock. Both areas were surrounded
with laser photocoagulation. In July 1997, the patient
suddenly noted a visual field defect and dimming of vision
in the right eye; a retinal detachment was found (Figure 4).
The detachment was bullous from 10 o’clock clockwise to
5 o’clock and extended posteriorly close to the macula.
The tears superiorly had coalesced into one large poste-
quatorial retinal tear at 12 o’clock. A primary vitrectomy
with perfluorocarbon liquid, endophotocoagulation, and
1036 AMERICAN JOURNAL
FIGURE 7. Graph of intraocular pressure (IOP) over the
clinical course of 17 years (Patient 3). IOP was normal and
symmetric for many years before vitrectomy for macular
pucker. The IOP increased in the vitrectomized left eye just
before cataract surgery but did not require treatment until six
years later. The IOP in the fellow eye was normal. BID ⴝ twice
daily.
perfluorocarbon gas injection were done. The retina was
reattached with a 20/30 corrected visual acuity (Figure 5).
In March 1998, nine months after vitrectomy, visual acuity
decreased to 20/50, and the patient underwent uneventful
phacoemulsification with intraocular lens implantation.
The IOP remained normal in both eyes until March 2002,
when the IOP was elevated in the right eye (OD 23 mm
Hg, OS 17 mm Hg). She was followed as a glaucoma
suspect in the right eye for twelve months, when the cup
in the right optic disk appeared enlarged compared with
the left and IOPs were OD 23 mm Hg, OS 17 mm Hg.
Timolol 0.5% was started in the right eye only (Figure 6).
On timolol therapy, five months later, the IOP increased
to 24 mm Hg, and Cosopt twice daily was started. With
Cosopt twice daily at last follow-up examination, 77
months after vitrectomy, the IOPs ranged between OU 13
to 14 mm Hg, with no medications in the left eye.
● PATIENT 3: A 61-year-old Caucasian woman was seen in
June 1988 with metamorphopsia in the left eye of one month
duration. Her corrected visual acuity was OD 20/25, OS
20/20. Cellophane maculopathy was noted in the left eye. In
September 1994, the visual acuity in the left eye was reduced
to 20/60-2, and the macular pucker was diagnosed. The IOPs
were 19 mm Hg OU. Vitrectomy with membrane peeling was
recommended, but the patient decided to wait until July 1995
to undergo surgery when the vision worsened slightly. Vit-
rectomy was uncomplicated, and the recuperation was unre-
markable. The IOPs were OD 18 mm Hg and OS 20 mm Hg.
The patient was followed regularly, and a brunescent nuclear
cataract developed in the operated left eye. In July 1998, the
IOPs were measured at OD 20 mm Hg and OS 22 mm Hg at
OF OPHTHALMOLOGY JUNE 2006
 TABLE 1. Clinical Characteristics Potentially Confounding
the Diagnosis of Open Angle Glaucoma After Vitrectomy
and Excluded From the Study
Patients with more than two vitrectomy operations
Previous history of anterior inflammation (eg, uveitis,
endophthalmitis)
Rubeosis irides or proliferative diabetic retinopathy with
severe ischemia or other retinal vascular disease associated
with iris neovascularization
Use of silicone oil, with or without emulsification
Severe penetrating trauma with corneal opacification or ciliary
body loss
Angle closure glaucoma
Patients receiving intravitreal triamcinolone
Patients chronically using topical or periocular steroids
Vitrectomy combined with a glaucoma surgical procedure (eg,
endocyclophotocoagulation or pars plana tube shunt)
one month before cataract surgery. After phacoemulsification
and intraocular lens implantation, the visual acuity improved
from 20/60 to 20/30. The patient was followed regularly, and
macular drusen were found in both eyes in September 2000.
From the years 2000 to 2005, the IOPs were frequently higher
by 3 to 4 mm Hg in the vitrectomized left eye. In March
2005, the IOP in that eye was found to be 26 mm Hg and
later 34 mm Hg on repeat testing by her local ophthalmolo-
gist. Gonioscopy revealed an open angle. Latanoprost was
started in the left eye only, and two weeks later the IOPs were
OD 14 mm Hg, OS 22 mm Hg.
This patient had unilateral OAG in the vitrectomized
eye 116 months (nine years, eight months) after vitrec-
tomy and 80 months (six years, eight months) after
phacoemulsification and intraocular lens implantation. A
graph of the IOPs over the clinical course of this patient is
shown in Figure 7.
In summary, all three patients underwent an uncompli-
cated vitrectomy with improvement of visual acuity, later
followed by cataract surgery. After vitrectomy, the IOP
returned to normal and was similar to that in the fellow eye
after the operation. As expected, a nuclear cataract developed
in each of the patients, and they underwent an uneventful
phacoemulsification procedure with intraocular lens implan-
tation (nine to 36 months after vitrectomy). At some period
after cataract surgery (in one patient, ⬎6 years), each patient
had OAG in the operated eye only, with the unoperated
fellow eye maintaining a normal IOP. Thus, I grew suspicious
that the vitrectomy, and later the cataract surgery, contrib-
uted to the development of glaucoma.
METHODS
WITH THE ASSISTANCE OF MY COLLEAGUES AND FELLOWS
at Columbia, I embarked on a retrospective chart review of
all patients who had undergone vitrectomy at the Edward
VOL. 141, NO. 6 OPEN ANGLE GLAUCOMA AFTER VITRECTOMY
TABLE 2. Mean Age of the Patients with Postoperative
Open Angle Glaucoma at the Time of Vitrectomy
Group Years N
1 Suspect 57.7 ⫾ 22.35 (range, 6–78) 10
2 New onset 62.3 ⫾ 11.96 (range, 26–88) 40
3 Preoperative open
angle glaucoma 67.3 ⫾ 5.44 (range, 59–77) 15
Harkness Eye Institute since 1995. Approval for this
review was obtained from the Columbia University Med-
ical Center Institutional Review Board, with exemption of
Informed Consent because of its retrospective nature. The
study was compliant with Health Insurance Portability and
Accountability Act (HIPAA) regulations.
Nine eyes (13.2%) had one previous vitrectomy. All
patients underwent conventional pars plana vitrectomy
with a 20-gauge, three-port system in conjunction with a
retrobulbar anesthetic block with a lidocaine 2% and
bupivacaine 0.75% mixture with 1/200,000 epinephrine.
Intraoperatively, perfluoro-n-octane (Perfluoron) was used
as an adjunct in 12 eyes (17.6%), in 11 eyes for the
management of retinal detachment, and in one eye to float
a dislocated intraocular lens anteriorly. At the end of the
surgical procedure, 35 eyes (51.5%) received an air tam-
ponade, and 26 eyes (38.2%) received a nonexpanding
concentration of fluorinated intraocular gas:air mixture
(sulfur hexafluoride, perfluorohexane, perfluoropropane).
Seven eyes (10.3%) received no internal tamponade. Ten
eyes (14.7%) had previous or concomitant scleral buckling
at the time of vitrectomy. After the operation, follow-up
visits routinely included visual acuity and slit-lamp and
fundus examinations. All IOPs were measured with Gold-
man applanation tonometry.
A long follow-up period was required because of long
latency for the development of OAG, and the charts of all
vitrectomy patients with a follow-up period of ⬎six
months were examined. Of 1156 charts that were re-
viewed, 703 patients were excluded because of a follow-up
period of ⱕsix months. Of the remaining 453 patients,
patients were identified to include those with a higher IOP in
the operated eye, OAG suspects, or OAG that required
treatment. Some patients were excluded because of the
possibility that the elevated IOP may have resulted from the
ocular disease process, after multiple surgeries, or from other
confounding factors such as intravitreal corticosteroids or
silicone oil. These exclusion criteria are listed in Table 1.
In all instances documented gonioscopic findings con-
firmed open angle status for inclusion in the study. In all,
we identified 68 eyes in 65 patients and separated them
into three categories as described below.
● GROUP 1: GLAUCOMA SUSPECTS: In this group of
patients (n ⫽ 10), the IOP in the vitrectomized eye was
1037
 TABLE 3. Indications for Vitrectomy in Patients With
Postoperative Open Angle Glaucoma
Indication Group 1 Group 2 Group 3 Total
Macular pucker 4 17 11 32
Macular hole 4 6 0 10
Retinal detachment 1 11 1 13
Vitreous hemorrhage 0 6 2 8
Retained lens fragments 0 2 1 3
Dislocated intraocular lens 0 1 0 1
Serous detachment 2
degrees to optic pit 1 0 0 1
higher than the fellow eye by ⱖ4 mm Hg on three or more
postoperative visits. The IOP could be normal or as high as
25 mm Hg in the operated eye. No visual field loss in the
operated eye or asymmetry of optic cup size (⬎0.2) was
present. These patients were not treated with antiglau-
coma therapy. If both eyes had vitrectomy or the fellow eye
had other significant eye disease that could affect IOP, the
patients were not eligible for inclusion.
● GROUP 2: NEW ONSET GLAUCOMA: These patients
(n ⫽ 43 eyes, 40 patients) were believed to have glaucoma
and underwent treatment because of elevated IOP, optic
disk changes, and visual field change that was compatible
with the optic nerve appearance. The eyes that were
included had IOP of ⱖ30 mm Hg. Also included were
patients with elevated IOP and disk or visual field changes
that were consistent with glaucoma. Another group that
also was included were patients with normal tension
glaucoma in which the optic disk cupping was asymmetric
and discovered after the operation with high normal or
slightly elevated IOP. Most of these patients were sent to
a glaucoma subspecialist for an opinion. After July 2002,
central corneal thickness was often measured with ultra-
sound pachymetry (Sonomed 300P PacScan, Lake Success,
New York, USA) to assist in the assessment for normal
tension glaucoma.
● GROUP 3: PRE-EXISTING GLAUCOMA: This group (n ⫽ 15)
included patients with pre-existing OAG that was treated
with medication before vitrectomy. In this group of pa-
tients, the goal was to study the effect of vitrectomy on the
progression of glaucoma in the operated eye.
For all patients, the following parameters were recorded:
age, gender, high myopia (ⱖ7 diopters), lens status, indi-
cations for vitrectomy, length of time after vitrectomy to
phacoemulsification, length of time after vitrectomy and
cataract surgery to the development of OAG, glaucoma
treatment before and after vitrectomy, the course of
glaucoma progression (medications, progression of visual
field loss, surgical therapy), and the length of follow-up time.
Student t tests were done to compare variables among the
groups.
1038 AMERICAN JOURNAL
TABLE 4. Lens Status Before and After Vitrectomy in Eyes
With Postoperative Open Angle Glaucoma
Before Vitrectomy After Vitrectomy
Group Phakic Pseudophakic/Aphakic Phakic Pseudophakic/Aphakic
1 6 4/0 1 9/0
2 25 14/4 1 38/4
3 10 5/0 0 15/0
TABLE 5. Mean Follow-up Interval of Eyes With Open
Angle Glaucoma After Vitrectomy
Group Months
1 55.2 ⫾ 20.6 (range, 15–74)
2 62.2 ⫾ 47.1 (range, 7–192)
3 42.8 ⫾ 37.3 (range, 9–156)
All patients 56.9 ⫾ 42.4
RESULTS
THE MEAN AGE OF THE PATIENTS AT THE TIME OF VITREC-
tomy for each group is shown in Table 2. There were no
statistically significant differences between the groups.
The indications for vitrectomy for each of the groups are
listed in Table 3. Macular pucker or macular hole ac-
counted for 61.8% of all eyes, and the remaining eyes did
not have complex problems.
The preoperative and postoperative lens status in all
groups is noted in Table 4. Forty-one eyes were phakic
(60.3%) before vitrectomy, and only two eyes (2.9%) were
phakic at the last follow-up examination.
The mean follow-up interval between the vitrectomy
and last follow-up visit is given in Table 5 for all groups.
There were no statistically significant differences between
groups.
In group 1, the glaucoma suspect group, the mean IOP
of three consecutive follow-up visits was compared in the
vitrectomized eye with the fellow unoperated eye. In this
group, the mean IOP was 19.5 ⫾ 2.7 mm Hg in the
vitrectomized eye, compared with 14.3 ⫾ 3.0 mm Hg in
the fellow eye (P ⫽ .0001, paired t test).
In group 2, eyes in which new onset glaucoma developed
after vitrectomy, the time interval between vitrectomy and
the diagnosis of glaucoma was significantly longer in
phakic eyes compared with the interval between vitrec-
tomy and glaucoma diagnosis in nonphakic eyes. In ini-
tially phakic eyes, the mean time interval was 45.95 ⫾
44.79 months (range, one to 174 months; n ⫽ 20), compared
with 18.39 ⫾ 13.76 months (range, three to 59 months; n ⫽
23) in eyes that were nonphakic at the time of vitrectomy
(P ⫽ .0115). When comparing the interval between
cataract extraction to glaucoma diagnosis in initially pha-
kic eyes (mean, 32.17 ⫾ 30.86 months; n ⫽ 18) and the
OF OPHTHALMOLOGY JUNE 2006

FIGURE 8. In general, postvitrectomy open angle glaucoma
was controlled successfully with medical therapy. However, a
monocular patient who was treated for macular pucker with
laser applied to inferior retinal tears required surgery. The
cupping increased in the right optic disk, and a tube shunt
implant was placed.
interval to glaucoma in nonphakic eyes (at the time of
vitrectomy; mean, 18.39 ⫾ 13.76 months; n ⫽ 23), the
difference was no longer statistically significant (P ⫽
.0627). In a subset of group 2 patients, 34 patients had
normal fellow eyes that did not have significant disease
(except for cataract surgery) and could be used for com-
parison. The number of antiglaucoma medications that
were taken in the vitrectomized eye (mean, 1.79 ⫾ 0.95)
was greater than used in the fellow eye (mean, 0.65 ⫾ 1.0),
and this was highly statistically significant (n ⫽ 34; P ⫽
.0001). Of these, 23 patients (67.6%) received treatment
for glaucoma only in the vitrectomized eye.
Group 3 eyes included patients who were treated for OAG
before vitrectomy in one eye. In 14 patients, the fellow eye
did not undergo vitrectomy and could be used for comparison.
At the last follow-up visit, the mean number of glaucoma
medications that were used in eyes that had vitrectomy was
2.9 ⫾ 1.2 and, in the fellow eye, was 2.00 ⫾ 1.4 (P ⫽ .0215).
At the last follow-up visit, the mean IOP was 14.43 ⫾ 2.0
mm Hg in vitrectomized eyes, compared with 14.36 ⫾ 3.8
mm Hg in the fellow eye (P ⫽ .9265).
In most patients, the glaucoma was well controlled and did
not cause much change in visual function. One patient
(group 2) underwent a glaucoma shunt procedure because of
the progression of cupping on maximal medical therapy
(Figure 8). In two other patients (both group 3), the visual
field loss progressed centrally, but no further surgery was done
because of macular disease (macular hole, macular pucker).
DISCUSSION
WE REVIEWED THE LITERATURE THAT REPORTED THE CLINI-
cal outcomes after vitrectomy for various retinal conditions to
VOL. 141, NO. 6 OPEN ANGLE GLAUCOMA AFTER VITRECTOMY
see whether OAG was reported as a late complication of
vitrectomy. In a review of several reports of large surgical
series for macular holes,15–17 epimacular membranes,6,18 –20
and vitrectomy for retinal detachment,21–26 there was little
information about the rates of glaucoma. In one series of
71 eyes with pseudophakic retinal detachment that was
managed with vitrectomy or vitrectomy and scleral buck-
ling, 14 eyes (19.7%) were noted to be on antiglaucoma
therapy before the operation. At the nine month follow-up
examination, 27 eyes (38%) were treated with topical
medication for glaucoma.27 The authors did not comment
on the significance of this finding.
The data from this study strongly suggest that vitrec-
tomy increases the risk of the development of OAG in the
operated eye. Most commonly, glaucoma was unilateral
and developed in the eye that underwent vitrectomy and
subsequent cataract surgery. The development of glaucoma
may take ⱖ10 or more years. The specific indications for
vitrectomy did not appear to be a risk factor. Because this
is a retrospective chart review, it is difficult to determine
the true incidence of OAG after vitrectomy because of the
variability in follow-up periods and the large number of
patients who were excluded because of short follow-up
periods. We estimate that up to 15% to 20% of eyes may
be at risk for the development of OAG after vitrectomy.
This suggests that up to 30,000 new cases of glaucoma may
develop annually in the United States after vitrectomy.
The time from vitrectomy in phakic eyes to the devel-
opment of OAG (mean, 45.9 months) compared with
same interval in pseudophakic eyes (mean, 18.4 months)
was statistically significant. In contrast, the time from
cataract surgery to the development of OAG in phakic
eyes was not statistically significant from eyes that were
pseudophakic at the time of vitrectomy. This suggests that
the presence of the lens may have a protective role and
may delay the development of OAG. The progression of
nuclear sclerosis in the lens of the vitrectomized eye is
well-known. In a series of 75 patients, the cumulative
incidence of progressive nuclear sclerosis was 72% in the
operated eye and 15% in the fellow eye after twenty-four
months.28,29 The progression of nuclear sclerosis was more
likely in older patients. After removal of the vitreous, the lens
undergoes a myopic shift in the refractive state of the eye and
develops a slow, but relentless, clouding of the nucleus in
almost all eyes. When the vitreous is not removed, the lens
does not develop cataract for up to five years.30
Postvitrectomy cataract was managed with phacoemul-
sification and intraocular lens implantation in all eyes. The
possibility that the cataract surgery increases the risk of
OAG after the operation was considered. However, several
case series that examined the effect of phacoemulsification
on IOP in nonglaucomatous eyes reported that the IOP
decreases after the operation.31–33 In several series that
reported the outcomes of patients with OAG who under-
went phacoemulsification, the glaucoma was easier to
manage. The number of medications that were required to
1039
FIGURE 9. In the nonvitrectomized phakic eye, the oxygen
tension (pO2) is lower (arrow) in the center of the vitreous
than in the vitreous near the retinal surface.
control IOP was significantly reduced, and the IOP was
lower than preoperative levels.34 –37 These effects persisted
in eyes that were followed for a mean of 2.8 years.36 Thus,
phacoemulsification does not appear to affect IOP in norm-
al and glaucomatous eyes and may even reduce it after
operation.
In this study, we found that, in eyes with pre-existing
glaucoma, the glaucoma frequently worsened in the vitrec-
tomized eye, because the number of medications that were
used to control the IOP increased after surgery. The factors
that contributed to these findings are not clear, but there
are several possible explanations. First, the surgical inflam-
mation and debris could have reduced the aqueous outflow
facility further. Another consideration is that the optic
disk in the vitrectomized eye becomes more susceptible to
damage after the operation. Last, it is possible that the
removal of the vitreous body alters the biochemical environ-
ment in such a way that aqueous outflow becomes reduced.
The delayed onset of glaucoma in the vitrectomized eye
suggests that there is a diffusible factor in the vitreous that
appears after the cortical vitreous is removed. This factor
alters the trabecular meshwork over time and reduces
aqueous outflow that results in elevated IOP and glaucoma.
After the immediate postoperative period, there is usually
no visible inflammation in the eye after uncomplicated
surgery. It is my hypothesis that oxygen is the main factor
and that oxidative stress affects the cells of the trabecular
meshwork over time, causing a rise in IOP in some eyes. In
the phakic eye, the lens is protective by metabolizing most
of the oxygen; but after cataract surgery, the trabecular
meshwork is subjected to oxidative damage by an increased
oxygen level from the vitreous.
1040 AMERICAN JOURNAL
FIGURE 10. After vitrectomy, the gradient disappears, and
oxygen tension (pO2) increases (arrow) in the central vitreous,
surrounding the lens, eventually causing cataract.
Oxygen tension has been measured before and after
vitrectomy in the normal rabbit eye.38 Before vitrectomy,
the oxygen tension is highest near the retinal surface and
decreases in the anterior vitreous just posterior to the
center of the lens (Figure 9). There is a further gradient
within the lens, with the lowest oxygen tension in the
center of the lens. After vitrectomy, the oxygen tension
near the lens rises to two to three times the preoperative
level, and the concentration gradient in the vitreous
disappears. In humans, Holekamp and associates39 mea-
sured the oxygen tension in the vitreous in patients who
underwent vitrectomy. They reported the elevation of
oxygen tension after vitrectomy and found in eight pa-
tients who underwent reoperation that a statistically sig-
nificant elevation of oxygen tension persisted for many
months after the initial surgery (three to twenty months;
mean, ten months).
In the phakic eye after vitrectomy, the elevated oxygen
levels are modulated by antioxidant mechanisms in the
vitreous and lens.40,41 For instance, in the presence of
transition metals such as copper and iron, ascorbic acid
may be oxidized with the result that hydrogen peroxide is
produced. Glutathione and catalase then detoxify the
hydrogen peroxide. Oxygen crosses cell membranes readily
in the lens. Most of the oxygen is consumed by oxidative
phosphorylation that takes place in the outer layers of the
lens where mitochondria are most populous.41 The lens
nucleus has very low levels of oxygen; however, when
ambient levels increase, protein oxidation begins in the
center of the lens, and nuclear cataract develops (Figure 10).
The lens also contains high levels of ascorbate that diffuse
from the fluid that surrounds it. The ascorbate and gluta-
thione within the lens also consume oxygen. The high
OF OPHTHALMOLOGY JUNE 2006

FIGURE 11. After cataract surgery, the increased level of
oxygen in the vitreous no longer is metabolized by the lens and
mixes with aqueous humor that passes through the trabecular
meshwork (arrows). pO2 ⴝ oxygen tension.
rate of progression of nuclear sclerosis after vitrectomy is
a sign that the lens does not have sufficient antioxidant
capacity to deal completely with the prolonged elevation
of oxygen levels. After cataract surgery, the increased
levels of vitreous oxygen mix with the aqueous fluid and
subsequently pass through the aqueous outflow channels
(Figure 11).
The pathologic mechanisms that cause glaucoma are
multiple and complex.42 However, there is good evidence
that oxidative damage to cells in the trabecular meshwork
alters its outflow capacity and may contribute to the
pathogenesis of OAG. There are several pathways by which
elevated levels of hydrogen peroxide and free radicals may
cause damage to cells in the trabecular meshwork.43,44
Antioxidant mechanisms such as superoxide dismutase
decrease in human trabecular meshwork cells with increas-
ing age.45 Oxidative DNA damage was found to be increased
in the trabecular meshwork of patients with OAG.46 – 48
Deletion of GSTM1 (encoding gene for glutathione
S-transferases) gene predisposes individuals to more oxidative
DNA damage and an increased risk of glaucoma.47 Oxidized
proteins are degraded by the 20S proteasome, and this
activity is reduced in the trabecular meshwork of patients
with glaucoma.49,50 The high levels of ascorbate in the
aqueous convert oxygen, which provides a source for
hydrogen peroxide. Human trabecular meshwork cells that
are exposed briefly to H2O2 showed impaired adhesion to
the extracellular matrix.51 The loss of adhesiveness was
believed to be related to the rearrangement of cytoskeletal
structures. Continued exposure to H2O2 might lead to
trabecular endothelial cell loss and compromised outflow.
VOL. 141, NO. 6 OPEN ANGLE GLAUCOMA AFTER VITRECTOMY
FIGURE 12. Flow diagram of the role of oxidative stress that is
hypothesized to cause an increase in nuclear cataract and open
angle glaucoma. The presence of a human lens delays the
development of open angle glaucoma.
Figure 12 provides a diagram that gives a summary of the
sequences of oxidative stress and its relationship to postvit-
rectomy cataract and glaucoma that are hypothesized.
The results found in this study have several important
clinical implications. Most of these patients were followed
for a number of years after vitrectomy. The glaucoma was
usually diagnosed as part of routine follow-up examina-
tions. In general, the disease was found at a relatively early
stage and treated promptly with good clinical outcomes.
Only one patient required a glaucoma surgery because of
progression under maximal medical therapy. Because of
the retrospective nature of this study, it is difficult to know
the exact incidence of OAG after vitrectomy. Most pa-
tients were lost to follow up or returned to their referring
physician. In the patients who did not receive follow up
with an ophthalmologist, the consequences might be more
serious because most likely the early changes of glaucoma
would be relatively asymptomatic. Thus, it is extremely
important to stress the importance of long-term follow-up
examination to patients who undergo vitrectomy, espe-
cially in those eyes that have had the lens removed.
This study excluded patients with OAG who might
have had other confounding factors in the development of
the glaucoma. Such factors included the use of silicone oil,
severe proliferative diabetic retinopathy, and periocular or
intravitreal triamcinolone injection. It is not clear whether
the mechanisms of oxidative stress that are proposed here
might also be a contributing factor in the development of
these other forms of glaucoma. Only prospective studies
might provide further insight into this problem.
Finally, there has been a progressive expansion in the
role of vitrectomy in retinal disorders. For instance, a
1041
recent survey of retinal surgeons found that 39% of
pseudophakic retinal detachments were managed with
vitrectomy with or without scleral buckling.52 In previous
years, this number was approximately 10%. In addition,
the trends have also promoted the use of vitrectomy and
gas injection even for the management of retinal detach-
ment with inferior retinal breaks.53–55 Increasingly, we are
beginning to see the use of vitrectomy for debris such as
floaters. Because these patients tend to be younger in age,
the lifetime risk of the development of glaucoma may be
higher, and an alternative vitreous-sparing procedure for
the repair of retinal detachment might be preferred.
In summary, the risk of OAG is increased after vitrec-
tomy, and this reports the outcomes of a series of patients
in whom glaucoma developed after the operation. The
glaucoma is more likely to appear years after cataract
surgery. In patients with pre-existing glaucoma, the eyes
required more medication after vitrectomy. The hypothesis
of the pathogenetic mechanisms presented seems plausible.
In most cases, the conditions were well-controlled and
rarely required surgery. Further work is necessary to deter-
mine the incidence of and to investigate and prove the
causes of this finding.
REFERENCES
1. Liesegang TJ, Hoskins HD, Jensen AD. The significance of
the Edward Jackson Lecture. Am J Ophthalmol 2005;139:
530 –532.
2. Crisp WH. Edward Jackson, student and teacher. Am J
Ophthalmol 1943;26:1–12.
3. Albert DM. Edward Jackson in 1896: a man and his specialty
at the crossroads: LXII Edward Jackson Memorial Lecture,
part I. Am J Ophthalmol 1996;122:469 – 475.
4. Machemer R, Buettner HM, Norton EW, Parel JM. Vitrec-
tomy: a pars plana approach. Trans Am Acad Ophthalmol
Otolaryngol 1971;75:813– 820.
5. Machemer R, Parel JM, Norton EW. Vitrectomy: a pars
plana approach: technical improvements and further results.
Trans Am Acad Ophthalmol Otolaryngol 1972;76:462– 466.
6. Margherio RR, Cox MS Jr, Trese MT, Murphy PL, Johnson
J, Minor LA. Removal of epimacular membranes. Ophthal-
mology 1985;92:1075–1083.
7. Chang MA, Parides MK, Chang S, Braunstein RE. Outcome
of phacoemulsification after pars plana vitrectomy. Ophthal-
mology 2002;109:948 –954.
8. Weinberg RS, Peyman GA, Huamonte FU. Elevation of
intraocular pressure after pars plana vitrectomy. Albrecht
Von Graefes Arch Klin Exp Ophthalmol 1976;200:157–161.
9. Han DP, Lewis H, Lambrou FH Jr, Mieler WF, Hartz A.
Mechanisms for intraocular pressure elevation after pars
plana vitrectomy. Ophthalmology 1989;96:1357–1362.
10. Thompson JT, Sjaarda RN, Glaser BM, Murphy RP. In-
creased intraocular pressure after macular hole surgery. Am J
Ophthalmol 1996;121:615– 622.
11. Chen CJ. Glaucoma after macular hole surgery. Ophthalmol-
ogy 1998;105:94 –100.
1042 AMERICAN JOURNAL
12. Chen PP, Thompson JT. Risk factors for elevated intraocular
pressure after the use of intraocular gases in vitreoretinal
surgery. Ophthalmic Surg Lasers 1997;28:37– 42.
13. Costarides AP, Alabata P, Bergstrom C. Elevated intraocular
pressure following vitreoretinal surgery. Ophthalmol Clin
North Am 2004;17:507–512.
14. Tranos P, Asaria R, Aylward W, Sullivan P, Franks W. Long
term outcome of secondary glaucoma following vitreoretinal
surgery. Br J Ophthalmol 2004;88:341–343.
15. Ezra E, Gregor ZJ. Surgery for idiopathic full-thickness
macular hole: two-year results of a randomized clinical trial
comparing natural history, vitrectomy, and vitrectomy plus
autologous serum: Morfields Macular Hole Study Group
report no. 1. Arch Ophthalmol 2004;122:224 –236.
16. Leonard RE 2nd, Smiddy WE, Flynn HW Jr, Feuer WJ.
Long-term visual outcomes in patients with successful mac-
ular hole surgery. Ophthalmology 1997;104:1648 –1652.
17. Scott IU, Moraczewski AL, Smiddy WE, Flynn HW Jr, Feuer
WJ. Long-term anatomic and visual acuity outcomes after
initial anatomic success with macular hole surgery. Am J
Ophthalmol 2003;135:633– 640.
18. Pesin SR, Olk RJ, Grand MG, et al. Vitrectomy for prem-
acular fibroplasia: prognostic factors, long-term follow-up, and
time course of visual improvement. Ophthalmology 1991;
98:1109 –1114.
19. Michels RG. Vitrectomy for macular pucker. Ophthalmology
1984;91:1384 –1388.
20. Margherio RR, Trese MT, Margherio AR, Cartright K.
Surgical management of vitreomacular traction syndrome.
Ophthalmology 1989;96:1437–1445.
21. Sharma YR, Karunanithi S, Azad RV, et al. Functional and
anatomic outcome of scleral buckling versus primary vitrec-
tomy in pseudophakic retinal detachment. Acta Ophthalmol
Scand 2005;83:293–297.
22. Ahmadieh H, Moradian S, Faghihi H, et al. Pseudophakic
and Aphakic Retinal Detachment (PARD) Study Group.
Anatomic and visual outcomes of scleral buckling versus
primary vitrectomy in pseudophakic and aphakic retinal
detachment: six-month follow-up results of a single opera-
tion: report no. 1. Ophthalmology 2005;112:1421–1429.
23. Miki D, Hida T, Hotta K, Shinoda K, Hirakata A. Compar-
ison of scleral buckling and vitrectomy for retinal detach-
ment resulting from flap tears in superior quadrants. Jpn J
Ophthalmol 2001;45:187–191.
24. Halberstadt M, Chatterjee-Sanz N, Brandenberg L, Koerner-
Stiefbold U, Koerner F, Garweg JG. Primary retinal reattach-
ment surgery: anatomical and functional outcome in phakic
and pseudophakic eyes. Eye 2005;19:891– 898.
25. Campo RV, Sipperley JO, Sneed SR, et al. Pars plana
vitrectomy without scleral buckle for pseudophakic retinal
detachments. Ophthalmology 1999;106:1811–1816.
26. Devenyi RG, de Carvalho Nakamura H. Combined scleral
buckle and pars plana vitrectomy as a primary procedure for
pseudophakic retinal detachments. Ophthalmic Surg Lasers
1999;30:615– 618.
27. Stangos AN, Petropoulos IK, Brozou CG, Kapetanios AD,
Whatham A, Pournaras CJ. Pars-plana vitrectomy alone vs.
vitrectomy with scleral buckling for primary rhegmatogenous
pseudophakic retinal detachment. Am J Ophthalmol 2004;
138:952–958.
OF OPHTHALMOLOGY JUNE 2006
28. de Bustros S, Thompson JT, Michels RG, Enger C, Rice TA,
Glaser BM. Nuclear sclerosis after vitrectomy for idiopathic
epiretinal membranes. Am J Ophthalmol 1988;105:160 –164.
29. Melberg NS, Thomas MA. Nuclear sclerotic cataract after
vitrectomy in patients younger than 50 years of age. Oph-
thalmology 1995;102:1466 –1471.
30. Sawa M, Ohji M, Kusaka S, et al. Nonvitrectomizing vitreous
surgery for epiretinal membrane: long-term follow up. Oph-
thalmology 2005;112:1402–1408.
31. Shingleton BJ, Gamell LS, O’Donoghue MW, Baylus SL,
King R. Long-term changes in intraocular pressure after clear
corneal phacoemulsification: normal patients versus glau-
coma suspect and glaucoma patients. J Cataract Refract Surg
1999;25:885– 890.
32. Suzuki R, Kuroki S, Fujiwara N. Ten year follow up of
intraocular pressure after phacoemulsification and aspiration
with intraocular lens implantation performed by the same
surgeon. Ophthalmologica 1997;211:79 – 83.
33. Pohjalainen T, Vesti E, Uusitalo RJ, Laatikainen L. Intraoc-
ular pressure after phacoemulsification and intraocular lens
implantation in nonglaucomatous eyes with and without
exfoliation. J Cataract Refract Surg 2001;27:426 – 431.
34. Mathalone N, Hyams M, Nieman S, Buckman G, Hod Y,
Geyer O. Long-term intraocular pressure control after clear
corneal phacoemulsification in glaucoma patients. J Cataract
Refract Surg 2005;31:479 – 483.
35. Hayashi K, Hayashi H, Nakao F, Hayashi F. Effect of cataract
surgery on intraocular pressure control in glaucoma patients.
J Cataract Refract Surg 2001;27:1779 –1786.
36. Pohjalainen T, Vesti E, Uusitalo RJ, Laatikainen L. Phaco-
emulsification and intraocular lens implantation in eyes with
open-angle glaucoma. Acta Ophthalmol Scan 2001;79:313–
316.
37. Hudovernik M, Pahor D. Intraocular pressure after phaco-
emulsification with posterior chamber lens implantation in
open-angle glaucoma. Klin Monatsbl Augenheilkd 2003;220:
835– 839.
38. Barbazetto IA, Liang J, Chang S, Zheng L, Spector A, Dillon
JP. Oxygen tension in the rabbit lens and vitreous before and
after vitrectomy. Exp Eye Res 2004;78:917–924.
39. Holekamp NM, Shui YB, Beebe DC. Vitrectomy surgery
increases oxygen exposure to the lens: a possible mechanism
for nuclear cataract formation. Am J Ophthalmol 2005;139:
302–310.
40. McNulty R, Wang H, Mathias RT, Ortwerth BJ, Truscott RJ,
Bassnett S. Regulation of tissue oxygen levels in the mam-
malian lens. J Physiol 2004;559.3;883– 898.
VOL. 141, NO. 6 OPEN ANGLE GLAUCOMA AFTER VITRECTOMY
41. Truscott RJ. Age-related nuclear cataract-oxidation is the
key. Exp Eye Res 2005;80:709 –725.
42. Gabelt BT, Kaufman PL. Changes in aqueous humor dynam-
ics with age and glaucoma. Prog Retin Eye Res 2005;24:612–
637.
43. Nguyen KP, Chung ML, Anderson PJ, Johnson M, Epstein
DL. Hydrogen peroxide removal by the calf aqueous outflow
pathway. Invest Ophthalmol Vis Sci 1988;29:976 –981.
44. Kahn MG, Giblin FJ, Epstein DL. Glutathione in calf trabecular
meshwork and its relation to aqueous humor outflow facility.
Invest Ophthalmol Vis Sci 1983;24:1283–1287.
45. Freedman SF, Anderson PJ, Epstein DL. Superoxide dis-
mutase and catalase of calf trabecular meshwork. Invest
Ophthalmol Vis Sci 1985;26:1330 –1335.
46. Chen JZ, Kadlubar FF. A new clue to glaucoma pathogenesis.
Am J Med 2003;114:697– 698.
47. Izzotti A, Sacca SC, Cartiglia C, De Flora S. Oxidative
deoxyribonucleic acid damage in the eye of glaucoma pa-
tients. Am J Med 2003;114:638 – 646.
48. Sacca SC, Pascotto A, Camicione P, Capris P, Izzotti A.
Oxidative DNA damage in the human trabecular meshwork:
clinical correlation in patients with primary open-angle
glaucoma. Arch Ophthalmol 2005;123:458 – 463.
49. Caballero M, Liton PB, Epstein DL, Gonzalez P. Proteasome
inhibition by chronic oxidative stress in human trabecular
meshwork cells. Biochem Biophys Res Commun 2003;308:
346 –352.
50. Wright G, Terada K, Yano M, Sergeev I, Mori M. Oxidative
stress inhibits the mitochondrial import of preproteins and
leads to their degradation 2001;263:107–117.
51. Zhou L, Li Y, Yue BY. Oxidative stress affects cytoskeletal
structure and cell-matrix interactions in cells from an ocular
tissue: the trabecular meshwork. J Cell Physiol 1999;180:
182–189.
52. American Society of Retinal Specialists. Preferences and
trends survey 2004. Last accessed: July 1, 2005. Available at:
http://www.retinaspecialists.org.
53. Sharma A, Grigoropoulos V, Williamson TH. Management
of primary rhegmatogenous retinal detachment with inferior
breaks. Br J Ophthalmol 2004;88:1372–1375.
54. Martinez-Castillo V, Boixadera A, Verdugo A, Garcia-
Arumi J. Pars plana vitrectomy alone for the management of
inferior breaks in pseudophakic retinal detachment without
facedown position. Ophthalmology 2005;112:1222–1226.
55. Wickham L, Connor M, Aylward GW. Vitrectomy and gas
for inferior break retinal detachments: Are the results com-
parable to vitrectomy, gas, and scleral buckle? Br J Ophthal-
mol 2004;88:1376 –1379.
1043
 Biosketch
Stanley Chang, MD, is currently a Edward Harkness Professor and Ku Teh Ying and K. K. Tse Professor and Chairman,
Department of Ophthalmology, Columbia University, New York, New York. Dr Chang graduated from the College of
Physicians and Surgeons of Columbia University and completed his Ophthalmology residency training at Massachusetts
Eye and Ear Infirmary and a vitreoretinal fellowship at Bascom Palmer Eye Institute. Dr Chang pioneered the development
of perfluorocarbon liquids and panoramic viewing, and surgical techniques for the management of complex retinal
detachments.

1043.e1 AMERICAN JOURNAL OF OPHTHALMOLOGY JUNE 2006