Professional Documents
Culture Documents
● PURPOSE: The Tube Versus Trabeculectomy (TVT) ions among glaucoma surgeons regarding the preferred
Study will compare the safety and efficacy of nonvalved surgical approach in patients who have undergone previ-
tube shunt surgery to trabeculectomy with mitomycin C ous cataract extraction and/or failed filtering surgery.
in patients with previous intraocular surgery. Forthcoming data from the TVT Study should provide
● DESIGN: Multicenter randomized clinical trial. valuable information comparing two surgical procedures
● METHODS: SETTING: 17 Clinical Centers. STUDY POPU- commonly used in this patient population. (Am J Oph-
LATION: Patients 18 to 85 years of age who have under- thalmol 2005;140:275–287. © 2005 by Elsevier Inc. All
gone previous trabeculectomy, cataract extraction with rights reserved.)
intraocular lens implantation, or both and have inade-
T
quately controlled glaucoma with intraocular pressure HE SURGICAL MANAGEMENT OF GLAUCOMA HAS
(IOP) >18 mm Hg and <40 mm Hg on tolerated medical evolved over the past several decades. Trabeculectomy
therapy. INTERVENTIONS: Study patients were randomized (or guarded filtration surgery) replaced full thickness
to undergo placement of a 350-mm2 Baerveldt glaucoma procedures when it became clear that similar success rates
implant or trabeculectomy with mitomycin C (0.4 mg/ml could be achieved with a lower rate of complications. Eyes
for 4 minutes). MAIN OUTCOME MEASURES: IOP, complica- that had undergone previous ocular surgery, including cata-
tion rates, visual acuity, visual field, quality of life, ract extraction and/or failed glaucoma filtering surgery, were
reoperations for glaucoma, and need for supplemental found to be at higher risk for subsequent trabeculectomy
medical therapy. failure. Wound-healing modulation with antifibrotic agents,
● RESULTS: A total of 212 patients were enrolled be- such as 5-fluorouracil (5-FU) and mitomycin C (MMC), was
tween October 1999 and April 2004. The age of the subsequently introduced and improved the success rate of
study population was 71.0 ⴞ 10.4 years (mean ⴞ SD), trabeculectomy.1–3 The use of 5-FU and MMC as adjuncts
and 53% were women. The baseline IOP was 25.3 ⴞ 5.3 with glaucoma filtering surgery has now become widespread
mm Hg (mean ⴞ SD). There were no significant differ- in clinical practice.4
ences in the demographic and ocular characteristics Although wound-healing modulation has increased the
between the 2 treatment groups. likelihood of intraocular pressure control (IOP) following
● CONCLUSIONS: Practice patterns vary in the surgical filtering surgery, there has been a growing concern within
management of glaucoma, and there are differing opin- the ophthalmic community about the complications asso-
ciated with the use of 5-FU and MMC. The prevalence of
Accepted for publication March 7, 2005.
From the Bascom Palmer Eye Institute, Miller School of Medicine, bleb-related infections, bleb leaks, and bleb dysesthesia
University of Miami, Miami, Florida (S.J.G., J.C.S., W.J.F., R.K.P. II); associated with a perilimbal filtering bleb suggests the need
Department of Ophthalmology, Medical College of Wisconsin, Milwau- to consider alternatives. Glaucoma drainage implants (or
kee, Wisconsin (D.K.H.); and Department of Ophthalmology, University
of California, Davis, Sacramento, California (J.D.B.). tube shunts) have been growing in popularity in recent
Supported by a research grant from Pfizer, Inc, New York, New York, years, and these devices offer an alternative surgical ap-
by grant EY014801 from the National Eye Institute, National Institutes of proach to trabeculectomy to reduce IOP in glaucoma
Health, Bethesda, Maryland, and by Research to Prevent Blindness, Inc,
New York, New York. patients. Favorable results have been reported with glau-
*A listing of additional investigators who participated in this study coma drainage implants (GDIs) in the treatment of
appears in the Appendix. refractory glaucomas.5–15 Comparable rates of severe com-
Inquiries to Steven J. Gedde, MD, Bascom Palmer Eye Institute, 900
N.W. 17th Street, Miami, Florida 33136; fax: 305-326-6478; e-mail: plications have been observed with GDI surgery and
sgedde@med.miami.edu trabeculectomy with antifibrotic agents.16
Purpose To compare the safety and efficacy of nonvalved tube shunt surgery to trabeculectomy with mitomycin C
The Tube Versus Trabeculectomy (TVT) Study is a ating recruitment. An effort was made to recruit every
multicenter randomized clinical trial comparing the safety eligible patient into the study. Following a discussion of
and efficacy of placement of a Baerveldt glaucoma implant the study, informed consent was obtained in accord with
(Advanced Medical Optics, Irvine, California, USA) to Health Insurance Portability and Accountability Act
trabeculectomy with the adjunctive use of MMC in eyes (HIPAA) regulations.
that have undergone previous filtering surgery, cataract
surgery with intraocular lens implantation, or both. ● STUDY ORGANIZATION: Multicenter clinical trials re-
quire an organizational structure that provides efficient
operations and facilitates communication. The Appendix
METHODS describes the study organization and lists the participating
centers and committees. Investigators at 17 Clinical Cen-
A SYNOPSIS OF THE TVT STUDY DESIGN FEATURES IS PRO- ters have been responsible for screening potential study
vided in Table 1. The Institutional Review Board at each patients, enrolling eligible patients, and following the
Clinical Center approved the study protocol before initi- patients according to the study protocol set forth in detail
Baseline 1 Day 1 Week 1 Month 3 Months 6 Months 1 Year 18 Months 2 Years 3 Years 4 Years 5 Years
Trabeculectomy With Mitomycin C. All patients ran- low-up visits 1 day, 1 week, 1 month, 3 months, 6 months,
domized to the trabeculectomy group underwent a 1 year, 18 months, 2 years, 3 years, 4 years, and 5 years
trabeculectomy superiorly. A limbus-based or fornix- postoperatively.
based conjunctival flap was dissected, depending on the
surgeon’s preference. A fluid retaining sponge soaked in Visual Acuity. Distance visual acuity is measured using 2
MMC (0.4 mg/ml) was applied to the superior sclera for techniques, Snellen visual acuity and the standard proce-
4 minutes before or after scleral flap dissection, depend- dure developed for the Early Treatment Diabetic Retinop-
ing on the surgeon’s preference. This area was then athy Study (ETDRS).17 Refraction is performed before
copiously irrigated with balanced saline solution (BSS). measurement of visual acuity by both techniques at the
A paracentesis tract was created in the peripheral baseline examination and annual follow-up visits. Snellen
cornea, and a corneal-based scleral flap was dissected visual acuity is measured at the baseline examination and
approximately half scleral thickness. A block of limbal every follow-up visit. ETDRS visual acuity is tested at the
tissue was excised underneath the trabeculectomy flap. baseline examination and annual follow-up visits. The
The scleral flap was reapproximated to the scleral bed examining clinician provides a reason for Snellen visual
with interrupted or releasable 10-0 nylon suture. Fol- acuity less than 20/30 at the baseline examination, or
lowing injection of BSS into the anterior chamber
reduction of visual acuity by 2 or more lines from baseline
through the paracentesis tract, the anterior chamber
at the 6-month follow-up visit or after.
remained formed with a visible leak present around the
scleral flap at equilibrium. The dimensions of the scleral
Slit-lamp Biomicroscopy. Examination of the anterior
flap, the size of the inner block, and the number and
segment using slit-lamp biomicroscopy is performed at the
tension of the scleral flap sutures were at the surgeon’s
baseline examination to document the preoperative status
discretion. The type of suture material and method of
conjunctival closure (single-layered or double-layered of the eye, and at all follow-up visits to detect any changes
closure) were of the surgeon’s choice. BSS was injected in ocular status during the course of the study that may be
through the paracentesis tract to elevate the bleb at the attributable to the disease or treatment.
conclusion of the case. A moistened fluorescein strip
was used to check for a conjunctival leak or leakage Seidel Testing. Seidel testing is performed at each fol-
from the paracentesis tract, and additional sutures were low-up visit. If the Seidel test is positive, the leak is graded
placed as needed. The use of intraoperative and postop- as an ooze, frank leak, or brisk leak.
erative medications was determined by the surgeon.
Tonometry. Goldmann applanation tonometry is used
● STUDY MEASUREMENTS: A list of study measurements to measure IOP, except when irregular corneal astigma-
for scheduled follow-up visits is presented in Table 2. An tism, corneal scarring, or corneal edema precludes accurate
appointment schedule was generated for each patient at readings. In these cases, the Tono-Pen (Mentor) is used.
enrollment by the SCC and sent to the patient’s Clinical IOP readings are repeated until 2 consecutive or noncon-
Center. The date of surgery was the study entry date for secutive measurements are obtained differing by 1 mm Hg
patients in the TVT Study, and all follow-up visits were or less, and the average of the 2 readings serves as the IOP
computed from this date. Enrolled patients complete fol- measurement for the visit.
Age (years)
Mean ⫾ SD 71.0 ⫾ 10.4 70.9 ⫾ 11.0 71.1 ⫾ 9.9 .89*
Range 32–85 40–85 32–85
Gender, n (%)
Male 100 (47) 43 (40) 57 (54) .055†
Female 112 (53) 64 (60) 48 (46)
Race, n (%)
White 95 (45) 52 (49) 43 (41) .53‡
Black 82 (39) 40 (37) 42 (40)
Hispanic 30 (14) 12 (11) 18 (17)
Other 5 (2) 3 (3) 2 (2)
Hypertension, n (%) 124 (59) 61 (57) 63 (60) .76†
Diabetes mellitus, n (%) 67 (32) 31 (29) 36 (34) .49†
SD ⫽ standard deviation.
*Student t-test.
†
Chi-square test.
‡
Exact permutation chi-square test.
Motility Evaluation. Diplopia is a significant com- is directed toward identifying signs of proliferative
plication that may occur following Baerveldt implantation. retinopathy or a scleral buckling element. The presence
The incidence of permanent restrictive strabismus associ- of either excludes the eye from the study. At all
ated with the Baerveldt glaucoma implant is not precisely follow-up visits, ophthalmoscopy is performed to evalu-
known, as this complication has not been studied prospec- ate for posterior segment complications such as serous
tively. To address this issue, a formal motility evaluation is choroidal effusions, suprachoroidal hemorrhage, or hy-
performed for all patients at baseline and at the 1-year and potonous maculopathy.
5-year follow-up visits. Additionally, motility evaluation is
performed in those patients with diplopia at the 6-month Perimetry. Quantitative automated perimetry is per-
follow-up visit or after. Transient diplopia following formed within 1 month of enrollment and at annual
Baerveldt implantation is not uncommon. This study will follow-up visits. The Humphrey Field Analyzer is used to
focus on the incidence and nature of permanent restrictive conduct a 24-2 threshold test with a size III white stimulus
strabismus associated with the Baerveldt glaucoma im- in all patients.
plant. The cover-uncover and alternate cover tests are
performed with the patient looking in primary gaze, as well Quality of Life Assessment. There has been a growing
as in upgaze, downgaze, left gaze, and right gaze. Motility interest within the medical community to determine the
evaluation is performed with the patient looking in the impact of different medical interventions on a patient’s
distance and fixating at a near target. Any heterophorias or functional status and quality of life. The 25-item National
heterotropias are identified, and the deviation is measured Eye Institute-Visual Function Questionnaire (NEI VFQ-
with a hand-held prism. In patients who are unable to 25) is a quality of life instrument that was designed to
fixate for cover testing, the deviation is measured by evaluate visual disability and its impact on daily function-
centering the corneal light reflexes with a prism using the ing in ophthalmic patients.18 The SCC administers the
modified Krimsky method. NEI VFQ-25 through telephone interview at the baseline
examination and follow-up visits. Patients at the Bascom
Gonioscopy. Gonioscopy is performed at the baseline Palmer Eye Institute also have the option of completing
examination. The anterior chamber angle is examined for the questionnaire in the clinic at the time of their
neovascularization under high magnification for the pur- follow-up visit.
pose of excluding eyes with neovascularization from the
study. ● OUTCOME MEASURES: Because glaucoma surgeries are
intended to reduce IOP, IOP is a primary outcome measure
Ophthalmoscopy. A dilated fundus examination is per- in the TVT Study. Concern about complications associ-
formed at the baseline examination and all follow-up ated with both surgical procedures was a major stimulus for
visits. At the baseline examination, particular attention the study, so the rate of complications is another primary
Overall Group (n ⫽ 212) Tube Group (n ⫽ 107) Trabeculectomy Group (n ⫽ 105) P-value
ACIOL ⫽ anterior chamber intraocular lens; CACG ⫽ chronic angle-closure glaucoma; ETDRS ⫽ Early Treatment Diabetic Retinopathy
Study; IOP ⫽ intraocular pressure; MD ⫽ mean deviation; NEI-VFQ ⫽ National Eye Institute-Visual Function Questionnaire; PCIOL ⫽ posterior
chamber intraocular lens; PG ⫽ pigmentary glaucoma; POAG ⫽ primary open-angle glaucoma; PSD ⫽ pattern standard deviation; PXFG ⫽
pseudoexfoliative glaucoma; SD ⫽ standard deviation.
*Chi-square test.
†
Student t-test.
‡
Exact permutation chi-square test.
§
Mann-Whitney U test.
#
Interval between last intraocular surgery and surgical treatment in study.
储
34 patients did not receive a preoperative 24-2 Humphrey visual field.
¶
10 patients did not receive a preoperative NEI-VFQ.
TABLE 5. Glaucoma Medications at Baseline in the Tube Versus Trabeculectomy Study
*Chi-square test.
†
Exact permutation chi-square test.
outcome measure. A secondary outcome measure will be recruited to allow for losses to follow-up, and recruitment
failure, defined by the following criteria: was terminated with 212 patients.
1. IOP ⬎21 mm Hg or not reduced by 20% below
baseline on two consecutive follow-up visits after 3 ● QUALITY ASSURANCE: Data collection is standardized
months through description of all study measures in a comprehensive
2. IOP ⱕ5 mm Hg on two consecutive follow-up visits Manual of Procedures and the use of uniform data collection
after 3 months forms. All personnel were certified before beginning data
3. Additional glaucoma surgery collection for the study. Patient eligibility was independently
4. Loss of light perception vision reviewed by the SCC before enrollment. All data forms are
Eyes that have not failed and are not on supplemental received and secured at the SCC. Each form is data entered
medical therapy are considered complete successes. Eyes by a data entry clerk and then verified by double entry by the
that have not failed but require supplemental medical SCC Research Coordinator. Edit checks, including missing
therapy are defined as qualified successes. Other outcome and questionable data, are clarified with Clinical Centers.
measures include visual acuity, visual field, and quality of The study is regularly reviewed by the SDMC. This
life. committee monitors adherence to the study protocol at
each Clinical Center, and it reviews treatment reports
● SAMPLE SIZE AND POWER CALCULATIONS: Sample prepared by the SCC for evidence of adverse or beneficial
size calculations were performed based on projected differ- treatment effects. The SDMC would have terminated the
ences in complication rates and IOP levels between study if treatment benefits or treatment risks were so great
treatment groups. In particular, the rate of late-onset for 1 treatment group that continuation of the trial was
bleb-related infection was estimated to be 5.0%19 in the deemed unethical.
trabeculectomy group and 0.05% in the tube group. The
incidence of persistent postoperative diplopia was assumed ● STATISTICAL ANALYSIS: The comparability of treat-
to be 5.0% in the tube group and 0.05% in the trabecu- ment groups at baseline was evaluated using 2-sided
lectomy group. The study was originally designed with a Student t test, Mann-Whitney U test, and 2 tests. Interim
sample size of 195 in each group to detect a difference of analyses are regularly prepared by the SCC and reviewed
5.0% versus 0.05% with a 2-sided significance level of 0.05 by the SDMC to assess treatment benefits and risks for
and a power of 0.80. The SDMC determined that a sample each group.
size of 95 patients in each treatment group would provide
sufficiently narrow 95% confidence limits of 2% to 11%
around the expected 5.0% complication rates. This sample RESULTS
size would detect a difference between 10.0% and 0.05%
with the original significance level and power. A TOTAL OF 212 PATIENTS WERE ENROLLED IN THE TVT
With a 2-sided significance level of 0.05, a sample size of Study at 17 Clinical Centers between October 1999 and
95 patients in each group would be expected to detect a April 2004. All patients received their assigned treatment.
difference in mean IOP of 2.5 mm Hg with 0.80 power. For Baerveldt implantation was performed in 107 patients, and
the secondary outcome of successful surgery, this sample 105 patients underwent trabeculectomy with MMC.
size would have a power of 0.80 to detect a treatment group The demographic characteristics of the overall study group
hazard ratio of ⬍0.63 and ⬎1.7. Additional patients were and both treatment groups are presented in Table 3. The
5-FU ⫽ 5-fluorouracil; ECCE ⫽ extracapsular cataract extraction; ICCE ⫽ intracapsular cataract extraction; MMC ⫽ mitomycin C;
PECE ⫽ phacoemulsification cataract extraction.
*P-value ⫽ .79 comparing the difference in proportions of the four strata between the two treatment groups using the chi-square test.
†
Exact permutation chi-square test.
mean age of the overall group at enrollment was 71.0 ⫾ 10.4 groups in the number of previous surgeries and the time
years (mean ⫾ SD), and 112 patients (53%) were women. interval between the last intraocular surgery and surgical
There was a trend toward a difference in gender distribution treatment in the study.
among treatment groups with 64 women (60%) in the tube The most common diagnosis was primary open-angle
group and 57 men (54%) in the trabeculectomy group, but glaucoma in 172 eyes (81%). Chronic angle-closure glau-
this did not reach statistical significance (P ⫽ .055). By coma was present in 18 eyes (8%), pseudoexfoliative
self-attribution, 95 patients (45%) were white, 82 (39%) were glaucoma in 8 eyes (4%), pigmentary in 1 eye (1%), and
black, 30 (14%) were Hispanic, and 5 (2%) were other races. other glaucoma types were diagnosed in 13 eyes (6%).
Hypertension was present in 124 patients (59%), and 67 Most eyes were pseudophakic with 160 (75%) containing
(32%) had diabetes mellitus. There were no significant a posterior chamber intraocular lens and 7 eyes (3%) had
differences in the demographic characteristics between treat- an anterior chamber intraocular lens (aphakic eyes were
ment groups. excluded from the study); the remaining 45 eyes (21%)
The ocular characteristics of the entire study group and were phakic. Eyes were stratified based on type of previous
each treatment group are summarized in Table 4. No signif- intraocular surgery; Table 6 lists the enrollment numbers
icant differences were detected in the ocular characteristics by stratum.
between treatment groups. Of the 212 study eyes, 112 (53%) The mean ETDRS visual acuity was 63.6 ⫾ 21.9 (mean ⫾
were left eyes. The IOP at baseline for the overall study group SD). The median Snellen visual acuity was 20/35. Reduced
was 25.3 ⫾ 5.3 mm Hg (mean ⫾ SD). Corneal abnormalities visual acuity less than 20/30 was attributed to glaucoma in 78
precluded reliable Goldmann applanation tonometry in two eyes (37%), macular disease in 17 eyes (8%), and cataract
patients, and the Tono-Pen (Mentor) was used to measure formation in 10 eyes (5%). The average mean deviation with
IOP in these cases. The average number of glaucoma medi- Humphrey visual field testing was ⫺15.9 ⫾ 9.9 dB (mean ⫾
cations at the time of enrollment was 3.1 ⫾ 1.2 (mean ⫾ SD), and the average pattern standard deviation was 7.0 ⫾
SD), and a majority of study eyes were receiving three or more 3.5 (mean ⫾ SD). Several patients had poor vision that
medications. Table 5 provides information on the types of precluded meaningful automated perimetry, and 34 patients
glaucoma medications that were used at baseline. Approxi- did not receive a preoperative visual field. The mean score on
mately half of the eyes (52%) had ocular laser therapy before the NEI-VFQ was 71.2 ⫾ 16.9 (mean ⫾ SD). There were 10
enrollment. Laser trabeculoplasty had been performed in 73 patients who could not be contacted by telephone before
eyes (34%), and 20 (9%) had laser iridotomy. The mean surgical treatment to administer the baseline quality of life
number of previous incisional surgeries was 1.3 for the overall questionnaire. Diplopia was reported by seven patients (3%)
group. There was no significant difference between treatment during the baseline examination.
Follow-up (months)
Authors Procedure Eyes Success Rate IOP Success Criteria (mm Hg) Mean Range
Heuer et al.21 Trab with 5-FU A 79% (27/34) ⱕ21 with meds 7.0 3.0–15.6
ⱕ25 without meds
Heuer et al.22 Trab with 5-FU A 69% (33/48) ⱕ21 with meds 18.5 6–34
ⱕ25 without meds
Weinreb23 Trab with 5-FU A 86% (13/15) ⱕ21 with meds 12
ⱕ25 without meds
FFSS1 Trab with 5-FU A/P 72% (58/81) ⱕ21 12
FFSS2 48% (37/77) 36
Palmer24 Trab with MMC A/P 75% (6/8) ⱕpredetermined level 12.3 6–30
Prata et al.25 Trab with 5-FU P 75% (12/16) ⱕ21 13.4
Trab with MMC 70% (21/30) 10.4
Minckler et al.5 SP Molteno A/P 63% (26/41) ⱕ21 16.2 7–30
Freedman & Rubin6 SP Molteno A/P 83% (20/24) ⱕ21 22 8.1–53.3
Lloyd et al.7 SP/DP Molteno A/P 56% (28/50) ⱕ21 and ⬎5 48.6 7–78
Heuer et al.8 SP Molteno A/P 50% (25/50) ⱕ21 and ⬎6 14.9 6 –29
DP Molteno 75% (38/51) 16.4 7–30
Hodkin et al.9 Baerveldt A/P 74% (26/35) ⱕ21 16.3 6.1–26.1
Mills et al.10 SP/DP Molteno A/P 58% (14/24) ⱕ22 45 6–107
Huang et al.11 Ahmed A 88% (28/32) ⬍22 and ⬎5 13.4 4–44
P 88% (84/96)
Broadway et al.12 SP/DP Molteno A 70% (21/30) ⬍22 43
P 66% (23/35)
Roy et al.13 Baerveldt A 75% (6/8) ⱕ21 and ⬎6 37.6 12–68
A ⫽ aphakia; DP ⫽ double-plate; 5-FU ⫽ 5-fluorouracil; IOP ⫽ intraocular pressure; MMC ⫽ mitomycin C; P ⫽ pseudophakia; SP ⫽
single-plate.
Follow-up (months)
Authors Procedure Success Rate IOP Success Criteria (mm Hg) Mean Range
Heuer et al.21 Trab with 5-FU 89% (8/9) ⱕ21 with meds 7.8 4.5–13
ⱕ25 without meds
Heuer et al.22 Trab with 5-FU 81% (13/16) ⱕ21 with meds 18.5 9–27
ⱕ25 without meds
Weinreb23 Trab with 5-FU 86% (13/15) ⱕ21 with meds 12
ⱕ25 without meds
FFSS1 Trab with 5-FU 79% (19/24) ⱕ21 12
FFSS2 61% (14/23) 36
Chen et al.26 Trab with MMC 78% (35/45) ⬍21 36 12–96
Palmer24 Trab with MMC 100% (7/7) ⱕpredetermined level 21.7 6–42
Singh et al.25 Trab with MMC 83% (10/12) ⱕ21 12.8 4–24
Andreanos et al.28 Trab with MMC 83% (20/24) ⱕ20 18 11–34
You et al.29 Trab with MMC 89% (39/44) ⱕ21 38.2 6–53
Minckler et al.5 SP Molteno 70% (7/10) ⱕ21 12.3 6–25
Lloyd et al.7 SP/DP Molteno 75% (9/12) ⱕ21 and ⬎5 41.4 15–64
Hodkin et al.9 Baerveldt 75% (9/12) ⱕ21 16.1 7.1–26.1
Mills et al.10 SP/DP Molteno 44% (4/9) ⱕ22 42 8–78
Broadway et al.12 SP/DP Molteno 58% (34/59) ⬍22 43
Roy et al.13 Baerveldt 88% (15/17) ⱕ21 and ⬎6 37.6 12–68
Chihara and associates compared the effect of tube shunt the two treatment groups, the Ahmed implant group had a
surgery using a White pump shunt in 16 eyes with that of greater adjunctive medication requirement. This study was
a trabeculectomy with 5-FU in 31 eyes with neovascular performed in Saudi Arabia and Sri Lanka, and it included
glaucoma.14 The probability of success (IOP ⱕ26 mm Hg patients with all glaucoma types and some eyes that had
and ⱖ5 mm Hg) was 53.0% in the implant group and undergone previous intraocular surgery. A follow-up study
45.4% in the trabeculectomy group at 3 years with Kaplan- continued enrollment in Sri Lanka of 123 patients with
Meier survival analysis. Bluestein and Stewart compared primary open-angle and angle-closure glaucoma without
the outcomes of 22 patients who underwent single-plate prior intraocular surgery.31 Lower IOPs were noted for the
Molteno implantation with those of a matched group who trabeculectomy group during the first year. However, the
had trabeculectomy with 5-FU.15 Similar mean IOP levels IOPs and cumulative probabilities of success were compa-
of 15.4 ⫾ 8.0 mm Hg in the Molteno group and 16.1 ⫾ 5.6 rable between the two study groups with longer follow-up.
mm Hg in the trabeculectomy group were observed after 6 The TVT Study has restricted its study population to
months of follow-up. patients who have undergone previous intraocular surgery.
In nonrandomized prospective studies and retrospective A standard dosage of MMC was used as an adjunct in all
case series, selection biases may produce treatment groups patients randomized to the trabeculectomy group. It has
with different underlying risk factors for failure. Results become common practice for glaucoma specialists to use an
reported in such studies must be interpreted with caution. A adjunctive antifibrotic agent when performing filtering
randomized clinical trial aims to produce comparison groups surgery in previously operated eyes.4 MMC is a more
that differ only by the treatment that their subjects receive potent inhibitor of fibroblast proliferation than 5-FU, and
and offers the best means for comparing treatments. studies have shown that trabeculectomy with the adjunc-
Wilson and associates performed a randomized clinical tive use of MMC yields lower IOPs than 5-FU.32,33
trial comparing the Ahmed glaucoma valve implant to We decided to exclude several secondary glaucomas
trabeculectomy in 117 patients.30 The use of an adjunctive from the TVT Study, including neovascular glaucoma,
antifibrotic agent in patients randomized to trabeculec- uveitic glaucoma, and glaucoma associated with the
tomy was not standardized and left to the discretion of the iridocorneal endothelial syndrome and epithelial or
surgeon. The cumulative probabilities of success (IOP ⬍21 fibrous downgrowth. It was the consensus of the inves-
mm Hg and at least 15% reduction in IOP from preoper- tigators that use of a GDI is the preferred surgical
ative levels) was 83.6% for trabeculectomy and 88.1% for approach with these types of refractory glaucomas.
Ahmed implant. While there was no statistically signifi- Severe posterior blepharitis and patient unwillingness to
cant difference in success and complication rates between discontinue contact lens use after surgery were exclusion