ORIGINAL STUDY
Posterior Surgical Revision of Failed Fornix-based
Trabeculectomy
Doaa H. Sobeih, MD*† and Paul R. Cotran, MD*
Purpose: To evaluate outcomes after revision of failed fornix-based
trabeculectomy using a posterior conjunctival incision and
mitomycin C.
Methods: Cases were identified using Current Procedural Termi-
nology codes. Information from clinical records was analyzed ret-
rospectively. Complete success was defined as intraocular pressure
(IOP) ≤ 16 mm Hg with no glaucoma medications and IOP reduc-
tion of > 20% from preoperative levels. Qualified success was
defined as IOP ≤ 16 with or without medications and IOP reduction
of > 20%. Secondary outcomes included IOP, number of glaucoma
medications used, visual acuity, and complications.
Results: Sixty eyes of 56 patients underwent surgical revision of
failed trabeculectomy. The complete success rate at 1 year was
43.7%, declining to 41.7% at 2 years and all time points up to
5 years; qualified success rates were 68.9% at 1 year and at all
subsequent time points. Mean IOP declined from 21.2 ± 7.7 (SD)
mm Hg preoperatively to 10.9 ± 4.3 mm Hg at 1 year, to
10.5 ± 3.8 mm Hg at 2 years, and to 9.9 ± 4.6 mm Hg at 3 years. The
mean number of glaucoma medications used fell from 2.5 ± 0.5
preoperatively to 1.0 ± 0.4 at 1 year, 0.9 ± 0.4 at 2 years, and
0.9 ± 0.4 at 3 years. Complications included bleb leak in 3 eyes
(5.0%), hypotony with choroidal detachment in 3 eyes (5.0%) and
hyphema in 1 eye (1.7%). Eight eyes required additional glaucoma
surgery.
Conclusions: Surgical bleb revision with mitomycin C using a pos-
terior incision in cases of failed fornix-based trabeculectomy can
provide effective control of IOP. This conjunctiva-sparing proce-
dure should be considered as a viable alternative to a tube shunt or
repeat trabeculectomy.
Key Words: glaucoma surgery, trabeculectomy, bleb failure,
mitomycin C
(J Glaucoma 2017;26:947–953)
T rabeculectomy continues to be the most common glau-
coma surgical procedure and is routinely performed with
the adjunctive use of antifibrotic agents such as mitomycin
C (MMC).1,2 The fornix-based conjunctival flap has largely
supplanted the earlier preference for a limbus-based flap;
this and other refinements in technique have been widely
adopted as in the Moorfields Safer Surgery System.3 When
Received for publication March 6, 2017; accepted August 5, 2017.
From the *Department of Ophthalmology, Lahey Hospital and Medical
Center, Burlington, MA; and †Matrouh Hospital, Marsa Matrouh,
Egypt.
Presented in part at the American Glaucoma Society Annual Meeting,
March 2016, Ft. Lauderdale, FL.
Disclosure: The authors declare no conflict of interest.
Reprints: Paul R. Cotran, MD, Department of Ophthalmology, Lahey
Hospital and Medical Center, 1 Essex Center Drive, Peabody, MA
01960 (e-mail: paul.r.cotran@lahey.org).
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/IJG.0000000000000770
J Glaucoma  Volume 26, Number 10, October 2017
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
trabeculectomy is successful, the low intraocular pressure
(IOP) obtained can prevent visual field loss over the long-
term, particularly in the eyes of patients with advanced
glaucoma.4,5
Although cataract formation, hypotony, and endoph-
thalmitis are potential complications of trabeculectomy, the
most common adverse outcome of trabeculectomy is pri-
mary bleb failure with resulting IOP elevation and con-
tinued progression of the disease. The failure rate after
primary trabeculectomy with MMC has been reported to be
between 10% and 56%.6–8 When the bleb loses function,
another drainage procedure is often required, and the choice
is usually between a repeat trabeculectomy adjacent to the
first, a tube shunt procedure, or a same-site bleb revision.
Same-site revision has an advantage over repeat trabecu-
lectomy and tube shunt surgery which both require the
dissection of another quadrant of conjunctiva, compromis-
ing future surgical options. Trabeculectomy revision may be
performed via a conjunctival puncture using a needle or fine
blade, usually referred to as a bleb needling procedure.
Studies suggest that needling is effective in encapsulated
blebs (ie, Tenon cyst) but has a higher failure rate when the
scleral flap is scarred down and the conjunctiva is flattened,
or when more than a year has elapsed since the primary
trabeculectomy.9–11 Incisional trabeculectomy revision with
MMC offers another option for patients with non-
functioning blebs. The larger conjunctival opening, which
can be initiated via an anterior or posterior approach,
allows the surgeon to directly visualize and remove the scar
tissues over the scleral flap that cause the bleb to lose
function. In the few previous reports on trabeculectomy
revision, the authors have used an anterior conjunctival
incision or have operated on eyes with a limbus-based pri-
mary trabeculectomy. In contrast, all of our bleb revisions in
recent years have been performed upon fornix-based tra-
beculectomies and we strongly prefer a posterior con-
junctival incision to initiate the surgery. To our knowledge,
this is the first study to look specifically at outcomes of
surgical revision with MMC using a posterior approach in
patients with a failed fornix-based trabeculectomy.
METHODS
This is a retrospective chart review of adult patients
who have undergone surgical bleb revision for non-
functioning fornix-based trabeculectomy. The sample con-
sists of patients operated upon for this indication by one
glaucoma surgeon (P.R.C.) at Lahey Hospital and Medical
Center. The study protocol was approved by the Institu-
tional Review Board/Ethics Committee at Lahey Hospital
Foundation and conformed with HIPAA policy and with
the recommendations of the Declaration of Helsinki. Charts
were reviewed for all 103 surgical procedures between
January 2007 and March 2015 which used Current Procedural
www.glaucomajournal.com | 947
Sobeih and Cotran
Terminology code 66250 (revision or repair of operative
wound of anterior segment). Ninety-seven of these were
trabeculectomy revision procedures. We included for analysis
eyes in which revisions were performed for the purpose of
restoring aqueous drainage, but excluded eyes if the clinical
record specifically indicated that the operation was to excise
a Tenon cyst (3 eyes). We also excluded those eyes that had
bleb revision for other indications such as bleb leakage (12),
dysthesesia (6), or hypotony (14). We also excluded eyes
with a history of more than one previous trabeculectomy
(1 eye) or a previous tube-shunt (1 eye). Prior bleb needling
with antimetabolites was not a reason for exclusion. Seven
eyes of 7 patients who had cataract surgery concurrent with
bleb revision were included in the sample. In total, 37 eyes
were excluded. The remaining 60 eyes (of 56 patients)
analyzed for this study all had an initial MMC-augmented
trabeculectomy using a fornix-based conjunctival incision
with a peripheral iridectomy performed in all cases. All 60 of
the revisions used a posterior (limbus-based) incision and were
augmented with MMC.
Indications for surgical bleb revision were often mul-
tifactorial but the most common reason was inadequate
control of IOP with medical therapy. In most eyes there was
concurrent evidence of increasing visual field loss, and some
eyes also had progressive optic nerve fiber layer thinning by
optical coherence tomography. In a few cases the decision to
reoperate was influenced by a patient’s desire to reduce
medication usage. Eyes were also evaluated and selected as
appropriate candidates for the procedure. This assessment
included testing for adequate mobility of the conjunctiva
overlying and posterior to the original scleral flap. Most
eyes with a failed trabeculectomy had little or no elevation
of the conjunctiva over the sclera, and these were referred to
in clinic notes as “flattened” blebs. Eyes in which the con-
junctiva was extensively scarred to the underlying sclera or
very thin and avascular were not chosen for this technique.
Gonioscopy was performed to confirm that the internal
sclerostomy was free of obstruction.
Surgical Technique
All procedures were performed using a peribulbar
anesthetic. A 6-0 silk traction suture was placed through the
superior cornea and used to rotate the globe downwards. A
6 to 8 mm long conjunctival incision was made parallel to
the limbus and 10 to 12 mm posterior to it, depending on the
adequacy of exposure. The incision location was centered
behind the original scleral flap, which was at 12.00. Using
nontoothed forceps and mini-Westcott scissors, the con-
junctiva was gently mobilized and elevated as the dissection
progressed anteriorly toward the scleral flap, taking care to
avoid the superior rectus muscle and its vasculature. As all
revisions were done in patients who had undergone fornix-
based trabeculectomy, there was generally no scar tissue
noted until the dissection reached the area of the flap.
Although blunt dissection was preferred to elevate the
conjunctiva over the scleral flap, it was necessary in some
cases to use a sharper instrument such as a 57 blade (Beaver-
Visitec Inc.) to incise scar tissue and create a cleavage plane.
Once the conjunctiva was freed from adhesions to the sclera,
the usual finding was that the scleral flap itself was closed,
with a thin sheet of scar tissue overlying the entire rec-
tangular flap and preventing aqueous flow. Two cut halves
of a polyvinyl alcohol corneal shield were saturated with
0.4 mg/mL MMC. A 5×5 mm thin sponge was cut from a
cellulose spear and saturated with MMC. The conjunctival
948 | www.glaucomajournal.com Copyright
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma  Volume 26, Number 10, October 2017
flap was lifted with forceps, allowing the 2 semicircular
sponges to be applied to the scleral surface posterior and
lateral to the flap. The square sponge was applied directly
over the fibrosed scleral flap. After 3 minutes the sponges
were removed and the ocular surface rinsed. A paracentesis
was made. The edges of the fibrosed flap were incised with
Vannas scissors or a sharp blade, and at that point clear
aqueous was usually noted to egress. If not, a crescent blade
was used to separate the scleral flap from its bed until
unobstructed aqueous flow occurred. In almost all cases the
scleral flap could be completely mobilized and lifted to
inspect the sclerostomy. A cyclodialysis spatula was inserted
under the posterior edge of the scleral flap and advanced
through the sclerostomy and into the anterior chamber.
Next, one or two 10-0 nylon sutures were placed to reappose
the scleral flap to its bed and reduce flow. Balanced salt
solution was instilled through the paracentesis to assess the
amount of leakage under the flap. A small amount of flow
was considered acceptable, but the goal was to have mini-
mal leakage through the scleral flap after suturing so that
the anterior chamber remained formed. The conjunctival
flap was then pulled posteriorly to its original position and
the incision was closed with a running 9-0 Vicryl suture on a
BV100 needle (V402G, Ethicon Inc.) using the technique of
locking every other throw. The bleb was elevated by
injecting balanced salt solution through the paracentesis or
by pressing on the posterior edge of the scleral flap through
the conjunctiva. Fluorescein was then placed on the con-
junctival wound closure to ensure that it was not leaking;
any leaking areas were resutured. A subconjunctival injec-
tion of 4 mg of dexamethasone was given in the inferior
fornix. All patients received a bead of combination steroid-
antibiotic ointment before patching the eye. In those
patients receiving concurrent cataract surgery, a temporal
clear cornea phacoemulsification technique was used with
insertion of an acrylic PCIOL before the trabeculectomy
revision.
Postoperative Management
Follow-up care was similar to that for standard tra-
beculectomy. Patients were examined weekly over the first
postoperative month, then at least monthly for the next
3 months. Laser suturelysis was performed as needed to
lower IOP. For the first month topical steroids were applied
8 times daily, these were gradually tapered over 4 to
6 months. Topical antibiotics were applied 4 times daily for
the first week. Supplemental glaucoma drops were added to
the regimen at any time in the postoperative course as
deemed necessary to maintain IOP at or below the target
IOP. If any eye had a shallow anterior chamber or choroidal
detachment postoperatively, cyclopentolate 1% drops were
given 3 times daily until the condition had resolved. Patients
were instructed to begin digital ocular pressure for 5 to 10
seconds 3 times daily if laser suturelysis failed to decrease
IOP and form a bleb; this was continued for up to 1 month.
Data Collection
Patient medical records were reviewed to obtain the
following information: demographic data, ocular history,
type of glaucoma, mean of the last 2 preoperative IOP
measurements, IOP at all postoperative visits, number of
glaucoma medications used at all visits, best-corrected visual
acuity (BCVA) at all visits, postoperative interventions
and complications, and any subsequent ocular surgeries.
IOP was measured by calibrated Goldmann Applanation
r 2017 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma  Volume 26, Number 10, October 2017 Revision of Failed Fornix-based Trabeculectomy

Tonometry (Haag-Streit USA). BCVA was measured using

the Snellen notation and converted to decimal form for TABLE 1. Characteristics of Patients Undergoing
Bleb Revision With Mitomycin C
analysis. The dates and details of the original trabeculec-
tomy and of the subsequent surgical revision were obtained Age ± SD (y)
from the operative notes. Mean 75 ± 10
Range 36-92
Outcome Analysis Sex [n (%)]
The main outcome measures were the success rates. Male 30 (54)
Female 26 (46)
The secondary outcome measures included postoperative
Race [n (%)]
IOP, the number of glaucoma medications needed after White 49 (88)
surgery, visual acuity, and rates of complications. Criteria Black 3 (5)
for success used for Kaplan-Meier survival analysis were as Hispanic 3 (5)
follows: eyes which had IOP ≤ 16 mm Hg with no glaucoma Asian 1 (2)
medications and a reduction in IOP by at least 20% from Type of glaucoma [n (%)]
preoperative levels were considered as a complete success. Primary open-angle 44 (72)
Qualified success was defined as an eye with IOP ≤ 16 mm Pseudoexfoliation 7 (12)
Hg with or without supplemental medication, and an IOP Uveitic 6 (10)
Neovascular 1 (2)
reduction of at least 20% from preoperative levels. Failure
Pigmentary 1 (2)
was defined as IOP above 16 mm Hg, or IOP not reduced by Angle closure 1 (2)
at least 20% from preoperative levels, or need for further Time to revision (y)
surgical interventions, or visually significant hypotony Mean 3.6
which was defined as IOP < 6 mm Hg with loss of 2 or more Range 0.25-9.6
lines of BCVA and lasting > 2 months. For survival analysis
the date of revision failure was the first visit 3 or more
Seven of the study eyes underwent combined phacoemulsi-
months after surgery when any of the above criteria
fication/PCIOL and trabeculectomy revision. Of the other
were met.
53 eyes, 37 (70%) were already pseudophakic. All of the
Statistical Methods study eyes were on topical antiglaucoma therapy at the time
Five-year survival curves using the Kaplan-Meier life- of bleb revision. Five eyes (8.3%) had received an office-
table analysis were used to estimate the probability of ach- based bleb needling with MMC before the surgical revision.
ieving complete or qualified success based on the previously Fifty (83.3%) of eyes completed at least 12 months of fol-
described criteria. For the Kaplan-Meier curves the unit of low-up, 41 (68.3%) of eyes completed 24 months of follow-
analysis was the patient rather than the eye, thus in bilat- up, 34 (57%) of eyes completed 36 months of follow-up and
erally operated patients 1 eye was randomly selected for 23 (38.4%) of eyes completed 60 months of follow-up.
analysis. Eyes were followed until failure, or last evaluation
verified as nonfailure. For calculation of mean IOP, mean Success Rates
number of medications used, and visual acuity, the unit of By Kaplan-Meier survival analysis (Fig. 1), the quali-
analysis was the eye. For the comparison of preoperative fied success rate following surgical trabeculectomy revision
versus postoperative IOP a random effect mixed model was
used to compare the IOP value at each time point with the
presurgery IOP. This model controlled for multiple eyes
within a single person by including a random effect for
person and a random effect of eye within each person. A
random effect generalized mixed Poisson model was used to
analyze the medication count at each time point. Least
mean square mean IOP and number of medications were
calculated at each time point and compared with the pre-
surgery IOP using a Dunnett test to control for multiple
testing. A random effect mixed model was also used to
compare the visual acuity values at each postoperative time
point with the presurgery visual acuity controlling for
multiple eyes within a single person, and multiple meas-
urements on each eye. For the analyses of mean IOP, mean
medication count, visual acuity, and complications, eyes
were not censored. All statistical analyses were performed
with SAS software version 9.4 TS Level 1M2 (SAS Institute,
Cary, NC). P-values <0.05 were considered as significant.
FIGURE 1. Kaplan-Meier plots for eyes undergoing trabeculectomy
RESULTS revision. Qualified success is indicated by the dashed line; complete
A total of 60 eyes of 56 patients who underwent same- success is indicated by the solid line. The number of eyes at risk at
each time point is indicated in the 2 rows above the x-axis; row 1 is
site surgical bleb revision between January 1, 2007 and April for complete success and row 2 is for qualified success. Censored
1, 2015 and met study criteria for inclusion were analyzed. eyes were those eyes that did not fail, but were not available for
Four patients had bilateral revisions. Patient age, sex, race, analysis either because they were lost to follow-up or had not yet
type of glaucoma, and time between initial trabeculectomy reached the indicated time point at the end of the study. Figure 1
and subsequent surgical revision are summarized in Table 1. can be viewed in color online at www.glaucomajournal.com.

Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved. www.glaucomajournal.com | 949
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
Sobeih and Cotran J Glaucoma  Volume 26, Number 10, October 2017

was 80.4% at 3 months, 72.9% at 12 months, and 68.9% at 60

the 18, 24, 36, 48, and 60 months postoperative time points.
Using the more stringent complete success standard, the
success rates were 62.5% at 3 months, 49.6% at 6 months, 50
43.7% at 12 months, and 41.7%, at all subsequent time
points. In both the qualified and the complete success groups
most of the failures occurred within the first 6 months after

IOP at 1 year (mmHg)

surgery. The Kaplan-Meier curve is based on the data from
56 eyes of 56 patients. Thirty-two of the 56 patients failed
the criteria for complete success at least once during follow-
up; the other 24 patients were censored for complete success.
Of these, 11 were followed from between 3 and 48 months, 2
died, 2 were lost to follow-up, and 9 were still being followed
when the study ended at 60 months. Fifteen of these 24
patients were also censored for qualified success; of these 5
were followed from between 3 and 48 months, 2 died, 2 were
lost to follow-up, and 6 were still being followed at
60 months.
IOP and Medications
Figure 2 shows the mean preoperative and postoperative
IOPs over time in the study eyes. The SD is reported with the
means. The preoperative mean IOP of 21.2 ± 7.7 mm Hg
(SD) was reduced to a mean of 9.72 ± 7.9 ± mm Hg on the
first postoperative day but with a wide range (0 to 34 mm Hg,
median 7.0). Mean IOP at the 3 month postoperative
time point was 11.93 mm Hg, decreasing at 6 months
to 11.41 ± 5.7 mm Hg, to 10.9 ± 4.3 mm Hg at 12 months,
9.90 ± 3.6 mm Hg at 18 months, 10.4 ± 4.7 mm Hg at
24 months, 9.9 ± 4.3 mm Hg at 36 months, 11.02 ± 5.1 mm Hg
at 48 months and 10.7 ± 5.0 mm Hg at 60 months. Reductions
of mean IOP compared with preoperative values were highly
significant at all postoperative time points (P < 0.001). Figure 3
is a scatter graph showing, the distribution of preoperative
versus postoperative IOPs of all eyes remaining in the sample
at the 12 month time point, and also shows, for each eye,
whether glaucoma medications were being used, or not. In this
analysis, 36/50 eyes (72%) had IOP lowered by 30% or more
with or without supplemental medications at the 12 month
time point, and 42/50 eyes (84%) had IOP lowered 20% or
more, with or without medications. Thirty-one of 50 eyes
(62%) had IOP ≤ 12, with or without medications.
Figure 4 displays the preoperative and postoperative
number of glaucoma medications needed by patients to
meet their individual target IOPs. The mean number of
preoperative medications was 2.5 ± 0.5; this decreased to
40
30
20
10
0
0 10 20 30 40 50 60
Preoperative IOP (mmHg)
FIGURE 3. Scatter graph showing intraocular pressure (IOP) at 1 year
after revision surgery versus preoperative IOP with reference lines for
0% (solid black), 20% (gray dot), or 30% (gray dash) improvement
relative to preoperative IOP. Each box represents 1 eye (n = 50). Eyes
treated with medications at 1 year are shown as solid squares, those
without medications at 1 year are shown as open circles. Figure 3 can
be viewed in color online at www.glaucomajournal.com.
0.8 ± 0.4 at both the 3 and 6 month postoperative time
points. There was a rise in medication use to 1.0 ± 0.4 at
12 months , to 0.9 ± 0.4 at the 18, 24, and 36 month time
points, to 1.0+0.5 at 48 months, and to 1.2 ± 0.5 at
60 months The reductions in medication use compared with
preoperative counts were highly significant at all time points
(P < 0.0001). Eleven of the 23 eyes (48%) completing
60 months of follow-up used no supplemental medications.
Visual Acuity
There was no statistically significant change in the
mean BCVA after trabeculectomy revision surgery at any
time point (P > 0.05) in patients not undergoing concurrent
cataract surgery. As expected, there was a statistically
significant improvement in the visual acuity in 7 eyes
receiving combined phacoemulsification/PCIOL and revi-
sion (P < 0.001). One patient had persistent loss of more

35
4
Intraocular pressure (mmHg)

30 60
3.5 60
Number of Medications

25 3
23
2.5 50 26
20 60 53 44 41 34
50 26 23 2 60 53
15 44 41 34
1.5
10
1
5 0.5

0 0

Time after surgery (months) Time after surgery (months)

FIGURE 2. Bar graph showing the mean intraocular pressure FIGURE 4. Graph showing the mean number of glaucoma
before and after trabeculectomy revision. The sample size (num- medications used before and following trabeculectomy revision.
ber of eyes) at each interval is indicated above the bar. Vertical The sample size (number of eyes) at each interval is indicated
lines, SD. above the bar. Vertical lines, SD.

950 | www.glaucomajournal.com Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.

J Glaucoma  Volume 26, Number 10, October 2017
than 2 lines of visual acuity postoperatively associated with
hypotony maculopathy.
Complications
Complications of bleb revision surgery developed in 7
(11.7%) of studied eyes (Table 2). Early conjunctival wound
edge leaks were noted in 3 eyes (5.0%), clinically significant
hypotony with choroidal detachments occurred in 3 eyes
(5.0%), and a transient hyphema occurred in 1 patient
(1.5%). One of the wound leaks resolved after treatment
with a large bandage contact lens, and 2 required resuturing
in the office. Two of these leaking blebs ultimately failed due
to uncontrolled IOP. Of the 3 hypotonous eyes with cho-
roidal detachments, 2 resolved with conservative manage-
ment. The third persisted with the patient unable to have
surgery. Because of the decrease of BCVA of at least 2 lines
(Snellen notation) for > 2 months, this eye was classified as
a failure. Eight eyes (13.3%) needed additional glaucoma
surgery due to elevated IOP: 2 received a second trabecu-
lectomy, 1 had a second surgical revision, 4 had tube shunt
procedures, and 1 had a laser cycloablation treatment.
The average time between the trabeculectomy revision and
the subsequent glaucoma surgery was 6.3 months (range:
0.25 to 2.0 y).
DISCUSSION
In this paper we retrospectively analyzed a series of
same-site trabeculectomy revisions which were performed
with a posterior conjunctival incision. A strength of this
study is its specificity; all procedures were done on eyes
which had undergone a primary trabeculectomy using a
fornix-based approach, and all operations were performed
by one surgeon with a standardized technique including a
consistent concentration and duration of exposure to MMC.
Determination of glaucoma surgical success rates are
dependent upon the definition of success or failure and thus
vary from study to study; in this report we used 2 sets of
criteria that are clinically meaningful in eyes with moderate
to advanced glaucoma and are consistent with guidelines
from the World Glaucoma Association.12 For most of our
patients the goal of therapy was an IOP below 16 mm Hg, in
many cases even lower IOPs were desired. IOP targets were
established for every patient and medications were routinely
added, even in cases of successful bleb reformation, if the
IOP remained above the target. Complete success was
defined as IOP ≤ 16 mm Hg on no supplemental medi-
cations and an IOP reduction of at least 20% below pre-
operative levels. A total of 41.7% of the analyzed cases met
these criteria at all time points between 18 and 60 months
following revision. Although bleb appearance was not an
outcome measure reported in this work, all patients cate-
gorized as complete successes had visibly elevated filtering
blebs as documented in the clinical record. In total, 68.9% of
patients were qualified successes, that is: were able to
maintain IOP ≤ 16 mm Hg and > 20% below preoperative
TABLE 2. Complications of Surgical Bleb Revision
Complication Type n (%)
Hypotony 3 (5)
Wound leak (early) 3 (5)
Hyphema 1 (1.6)
Choroidal detachment 3 (5)
Further glaucoma surgery 8 (13)
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
Revision of Failed Fornix-based Trabeculectomy
levels with the use of supplemental medications at the 12
through 60 months follow-ups. Reductions in both mean
IOP and mean number of medications used were highly
significant at all postoperative time points, but we believe
that the success rates are more accurate indications of the
efficacy of trabeculectomy revision, given the wide spread in
observed postoperative IOPs. For example, several eyes had
IOPs <6 mm Hg at multiple postoperative visits. These were
only classified as failures if the visual acuity declined by 2
lines or more of Snellen acuity, but those very low IOPs
contributed to lower mean IOPs for the group.
There are few previous studies of surgical revision with
MMC after trabeculectomy failure. Hirunpatravong et al13
analyzed a large series of same-site trabeculectomy revisions
with MMC, using 3 increasingly rigorous criteria for suc-
cess. Their criteria B (IOP < 16 mm Hg and IOP reduction
of > 25%), allowing use of supplemental medications, is
similar to our qualified success criteria. At postoperative
year 3 they report success rates of 47.8%, compared with
68.9% in our study at the same time point. In contrast with
our study of exclusively limbus-based (posterior incision)
revisions, the series of Hirunpatravong and colleagues
includes 115/130 cases performed as fornix-based (anterior
incision) revisions, and only 15/130 as limbus-based. It also
cannot be determined from their report whether the primary
trabeculectomies had been fornix-based or limbus-based.
Mean postoperative IOP at 3 years in their study was
11.8 mm Hg with a mean use of 1.28 medications, compared
with 9.9 mm Hg and 1.0 medications in the current study. It
is possible that our moderately better outcomes are due to
the posterior location of the conjunctival incision or to our
protocol of longer and more intensive postoperative use of
topical steroids. Coote et al14 reported a series of 57 bleb
revisions in 52 patients; the great majority (48/57) of sur-
geries were performed on limbus-based blebs using a pos-
terior incision. The authors, using success criteria of
IOP ≤ 16 mm Hg with no supplemental medications,
reported 1 year and 3 year success rates of 95% and 84%,
respectively, higher than those found in most contemporary
reports of primary trabeculectomy with MMC. In their
series MMC was applied after dissection of the scleral flap
and the scleral flap was not resutured. Viscoelastic was used
to fill the anterior chamber and was not removed. These
methods may have enhanced early filtration postoperatively
and contributed to high success rates; remarkably, they
noted only 4 eyes with choroidal effusions and reported no
shallow anterior chambers. Coote and colleagues did not
include hypotony in their definition of failure and it is
possible that some patients with very low IOPs are included
in their successful cases. It also appears from the description
of the surgical technique that many of the included eyes had
encapsulated rather than flattened bleb morphology. Anand
and Arora15 reviewed the results of surgical trabeculectomy
revision with MMC in 54 eyes and defined complete success
as IOP < 18 mm Hg and IOP reduction of 20% from pre-
operative levels with no supplemental medications. They
found that 48% of eyes at 1 year and 38% at 3 years were
complete successes. Qualified success rates were 72% at
1 year and 64% at 3 years, and allowed use of antiglaucoma
medications. Unlike the present series, the eyes in their study
had originally received limbus-based trabeculectomy and
their technique used an anterior (fornix-based) dissection of
the conjunctiva. Limbus-based and fornix-based blebs are
morphologically distinct, with limbus-based blebs often
having a dense posterior scar at the site of the incision,
www.glaucomajournal.com | 951
Sobeih and Cotran
commonly referred to as the “ring of steel.” This makes the
dissection of these 2 types of failed blebs markedly different
from the surgeon’s point of view. Coote et al14, in the pre-
viously discussed study, specifically included dissection of
this posterior cicatrix in their description of the surgical
technique. In a limbus-based failed bleb one might reason-
ably use an anterior incision to avoid the “ring of steel” scar.
After years of performing bleb revisions to restore aqueous
flow in fornix-based failed blebs, we strongly prefer the
posterior incisional approach. We observed that the limbal
conjunctiva in the fornix-based trabeculectomies is often
scarred or partially recessed, making the anterior dissection
difficult and, more importantly, contributing to problems
with the subsequent closure. Wound leaks at the limbus
were common in our experience using this technique. In
contrast, we found that the posterior incision, made in virgin
conjunctiva, was easy to initiate and the subsequent running
suture closure was reliable and rarely leaked. The posterior
approach also allows the surgeon to perform a rapid dis-
section to the scleral flap, as there is no posterior “ring of
steel” in the blebs that were originally created with the
fornix-based incision. Although scarring was always
encountered at the scleral flap, and there was a risk of
perforating the conjunctiva anterior to this point, we found
that there was no need to dissect and elevate the conjunctiva
all the way to the limbus in order to open the scleral flap and
establish flow. The dissection over the scleral flap is the most
challenging part of this procedure, yet there were no con-
junctival buttonholes observed in this series of cases. Careful
dissection using blunt-tipped mini-Westcott scissors and
pressing downwards while using the scissors, was key to
avoiding conjunctival trauma.
A posterior incision for bleb revision is by definition
limbus-based, and there is appropriate concern that this
technique may result in a revised and functioning bleb with
the classic limbus-based morphology—a localized, high bleb
with a long-term risk of a thin, avascular conjunctival sur-
face. Although we observed the formation of a posterior
cicatrix in a number of our cases, we did not note any bleb
leaks or cases of blebitis during the study, nor any cases of
dysethesia due to a high bleb or a corneal bleb.
In contrast to the limited data available on the results
of surgical trabeculectomy revision, there is extensive
information on outcomes of same-site revision using a nee-
dle or fine blade and augmented with MMC. Published
studies included eyes with a range of bleb morphologies
including flattened, cystic and encapsulated blebs and
reported success rates range between 27% and 80%.11,16–19
Tulidowicz-Bielak et al,18 using a MMC augmented nee-
dling technique in 40 eyes with nonfunctioning, flattened
filtering blebs, and using complete success criteria of
IOP ≤ 18 mm Hg without medications, achieved a 1 year
success rate of 42.5% and qualified success, allowing the use
of medications, of 77.5%. Panarelli and colleagues published
a case series of 27 trabeculectomy revisions performed using
an MVR blade through a posterior puncture site in blebs
that were specifically described as having a flattened mor-
phology. Using success criteria of IOP < 14 mm Hg with or
without medications, they found success rates at 1, 2, and
3 years postoperatively of 62%, 58%, and 53%,
respectively.19 In this small series, 8/27 eyes had transient
choroidal effusions and 5/27 required anterior chamber
reformation. Several eyes required more than one procedure
to achieve success. In this and other publications on nee-
dling of a failed bleb, hypotony, often with choroidal
952 | www.glaucomajournal.com Copyright
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma  Volume 26, Number 10, October 2017
effusions, was seen more commonly than in our study,
presumably because there was no resuturing of the scleral
flap to control the level of aqueous flow.18,19 Needling
procedures intended to open a scleral flap are particularly
difficult if the scleral flap and fibrotic scar tissue cannot be
seen through the conjunctiva and Tenon’s capsule. In con-
trast, revision through a larger incision allows for more
complete release of subconjunctival scar tissue, better visu-
alization and dissection of the scleral flap, adequate cau-
terization of bleeding vessels, and titration of aqueous flow
with flap sutures. These are significant advantages that can
improve outcomes and reduce complications.
In addition to the option of same-site revision in the
patient with a failed filtration procedure, a repeat trabecu-
lectomy at an adjacent superonasal or superotemporal site
may be attempted. Law et al20 reviewed the results of repeat
trabeculectomy with MMC in 75 eyes and found that at
3 years 44% of eyes were successes using criteria of
IOP < 16 mm Hg and a 25% reduction from preoperative
levels with or without medications. Meyer et al21 reported a
2 year complete (with no supplemental medication) success
rate for repeat trabeculectomy with MMC of 32.1% and a
qualified (with medication) success rate of 46.7%, using IOP
criteria of <16 mm Hg and IOP reduced > 40% from
preoperative levels. The results of these reports can be
compared with our 3 year success rates of 68.9% (with
supplemental medication) and 41.7% (without medication)
for same-site revision. Allowing for some differences in
success criteria, our trabeculectomy revision group achieved
better tonometric outcomes than patients in these 2 studies
of repeat trabeculectomy. In addition, trabeculectomy
revision has an advantage over repeat trabeculectomy in
not requiring dissection and disruption of additional con-
junctival tissue, nor an additional sclerostomy or iridec-
tomy. The sparing of conjunctiva is a very important benefit
of same-site bleb revision, as many glaucoma patients need
multiple external drainage procedures over time.
Another possibility for an eye with a nonfunctioning bleb
is a tube shunt procedure. We did not find a publication
specifically reporting on efficacy of tube shunt implantation
following failure of trabeculectomy, but information can be
gleaned from the prospective Tube Versus Trabeculectomy
study in which 63 of 107 eyes randomized to receive Baerveldt
350 implants had previously undergone trabeculectomy.22
This group at 3 years had a complete success rate of 26.5%
and a qualified success rate (allowing use of medications) of
79.6%.23 It should be noted that the success criteria in the
Tube Versus Trabeculectomy study only required that IOP be
<21 mm Hg. A tube shunt is a reasonable alternative to
surgical trabeculectomy revision but, in addition to requiring
the dissection of an additional quadrant of conjunctiva,
introduces the risk of device-related problems such as tube or
plate exposure, corneal decompensation, and restriction of
ocular movement.24
The current study has several limitations including the
retrospective nature of the work and loss of patients to
follow-up, the potential for inaccuracy in our inclusion and
exclusion criteria, and the lack of a standardized post-
operative protocol for management, visit intervals, and data
collection. Failed trabeculectomy patients could receive a
recommendation to undergo other procedures including
tube shunt, cycloablation, and repeat trabeculectomy
instead of surgical revision. This choice was made at the
surgeon’s discretion, and in consultation with the patient
and family. In some cases trabeculectomy revision was not
r 2017 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma  Volume 26, Number 10, October 2017
chosen due to lack of conjunctival mobility or excessive
vascularity or scarring of the failed bleb. This may have
biased our results towards increased success, but in our
opinion selection of the most appropriate procedure for a
given patient is a key part of the surgical art. As mentioned
earlier, this technique for bleb revision involves a posterior,
or limbus-based incision. It is possible that complications
that have been associated with limbus-based bleb mor-
phology, including dysethesia, corneal blebs, bleb leaks, and
blebitis, may develop with longer follow-up than the 5 years
of this study. Not all of the qualified successes were asso-
ciated with clearly elevated blebs, and it is possible that in
these eyes observed reduction in IOP after revision may
have been due to other factors such as improved adherence
to medical therapy. A further limitation of this study is that
most of our patients were older white adults, and the out-
comes presented may not apply to other populations. A
randomized prospective comparison of incisional trabecu-
lectomy revision with repeat trabeculectomy or tube shunt
using standardized treatment protocols would be helpful to
compare the outcomes of surgical options for patients with
failed blebs, as would a prospective comparison of needling
revision with incisional revision. Such studies would ideally
include comparison of the economic costs of the procedures.
In summary, this study demonstrated the effectiveness
of same-site trabeculectomy revision using a posterior
approach to lower IOP and reduce glaucoma medication
requirement. Over two-thirds of patients analyzed in this
study were qualified successes with IOPs maintained
≤ 16 mm Hg during a follow-up period of up to 5 years.
Complications were few and there was only one case of
permanently decreased visual acuity attributed to the sur-
gical procedure. Variations in individual wound healing
responses remain poorly understood, but management of
patients after trabeculectomy revision should include close
observation, high frequency topical steroid applications,
timely lysis of scleral flap sutures, and digital pressure when
appropriate. The surgical technique is straightforward, but
there is a learning curve in avoiding conjunctival trauma,
creating the proper planes of dissection, and titrating flow
through the scleral flap. It should not be assumed that the
procedure will be successful if only undertaken rarely; as
with all glaucoma drainage surgeries there is a benefit to
experience in performance of the technique.
ACKNOWLEDGMENT
The authors thank Robin Ruthazer, Tufts Medical Center,
Boston, MA for assistance with the statistical analysis.
REFERENCES
1. Arora KS, Robin AL, Corcoran KJ, et al. Use of various
glaucoma surgeries and procedures in Medicare beneficiaries
from 1994 to 2012. Ophthalmology. 2015;122:1615–1624.
2. Beatty S, Potamitis T, Kheterpal S, et al. Trabeculectomy
augmented with mitomycin C application under the scleral flap.
Br J Ophthalmol. 1998;82:397–403.
3. Dhingra S, Khaw PT. The Moorfields Safer Surgery System.
Middle East Afr J Ophthalmol. 2009;16:112–115.
4. Jayaram H, Strouthidis NG, Kamal DS. Trabeculectomy for
normal tension glaucoma: outcomes using the Moorfields Safer
Surgery technique. Br J Ophthalmol. 2016;100:332–338.
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
Revision of Failed Fornix-based Trabeculectomy
5. The Advanced Glaucoma Intervention Study (AGIS): 7. The
relationship between control of intraocular pressure and visual
field deterioration. Am J Ophthalmol. 2000;130:429–440.
6. Fontana H, Nouri-Mahdavi K, Lumba J, et al. Trabeculectomy
with mitomycin C: outcomes and risk factors for failure
in phakic open-angle glaucoma. Ophthalmology. 2006;113:
930–936.
7. Scott IU, Greenfield DS, Schiffman J, et al. Outcomes of
primary trabeculectomy with the use of adjunctive mitomycin.
Arch Ophthalmol. 1998;116:286–291.
8. Landers J, Martin K, Sarkies N, et al. A twenty-year follow-up
study of trabeculectomy: risk factors and outcomes. Ophthal-
mology. 2012;119:694–702.
9. Gutiérrez-Ortiz C, Cabarga C, Teus MA. Prospective evalua-
tion of preoperative factors associated with successful mitomy-
cin C needling of failed filtration blebs. J Glaucoma. 2006;15:
98–102.
10. Rotchford AP, King AJ. Needling revision of trabeculectomies:
bleb morphology and long-term survival. Ophthalmology. 2008;
115:1148–1153.
11. Kaushik S, Tiwari A, Pandav SS, et al. Use of ultrasound
biomicroscopy to predict long-term outcome of sub-Tenon
needle revision of failed trabeculectomy blebs: a pilot study. Eur
J Ophthalmol. 2011;21:700–707.
12. Shaarawy TM, Sherwood MB, Grehn F. eds. Guidelines on
Design and Reporting of Glaucoma Surgical Trials. Amsterdam,
Netherlands: World Glaucoma Association. Amsterdam Kugler
Publications; 2009:15–24.
13. Hirunpatravong P, Reza A, Romero P, et al. Same-site
trabeculectomy revision for failed trabeculectomy: outcomes
and risk factors for failure. Am J Ophthalmol. 2016;170:
110–118.
14. Coote MA, Gupta V, Vasudevan S, et al. Posterior revision
for failed blebs: long-term outcomes. J Glaucoma. 2011;20:
377–382.
15. Anand N, Arora S. Surgical revision of failed filtration
surgery with mitomycin C augmentation. J Glaucoma. 2007;16:
456–461.
16. Anand N, Khan A. Long-term outcomes of needle revision of
trabeculectomy blebs with mitomycin C and 5-fluorouracil: a
comparative safety and efficacy report. J Glaucoma. 2009;18:
513–520.
17. Shetty RK, Wartluft L, Moster MR. Slit-lamp needle revision
of failed filtering blebs using high-dose mitomycin C.
J Glaucoma. 2005;14:52–56.
18. Tulidowicz-Bielak M, Tomaszewska J, Kosior-Jarecka E, et al.
Efficacy and safety of mitomycin C-augmented revisions of
non-functioning filtering blebs after trabeculectomy-7 year
results. Klin Oczna. 2013;115:177–183.
19. Panarelli JF, Vinod K, Huang G, et al. Transconjunctival
revision with mitomycin-C following failed trabeculectomy.
J Glaucoma. 2016;25:618–622.
20. Law SK, Shih K, Tran DH, et al. Long-term outcomes of
repeat vs initial trabeculectomy in open-angle glaucoma. Am J
Ophthalmol. 2009;148:5.
21. Meyer LM, Graf NE, Philipp S, et al. Two-year outcome of
repeat trabeculectomy with mitomycin C in primary open-angle
and PEX glaucoma. Eur J Ophthalmol. 2015;25:185–191.
22. Gedde SJ, Schiffman JC, Feuer WJ, et al. The tube versus
trabeculectomy study: design and baseline characteristics of
study patients. Am J Ophthalmol. 2005;140:275–287.
23. Gedde SJ, Schiffman JC, Feuer WJ, et al. Three-year follow-up
of the tube versus trabeculectomy study. Am J Ophthalmol.
2009;148:670–684.
24. Budenz DL, Feuer WJ, Barton K, et al. Postoperative
complications in the Ahmed Baerveldt Comparison Study
During Five Years of Follow-up. Ahmed Baerveldt Compar-
ison Study Group. Am J Ophthalmol. 2016;163:75–82.
www.glaucomajournal.com | 953