Early Clinical Experience With the

Baerveldt Implant in Complicated Glaucomas

MICHAEL J. HODKIN, MD., WARREN S. GOLDBLATT, MD.,

CLAUDE F. BURGOYNE, MD., STUART F. BALL, MD.,
AND MICHAEL S. INSLER, M.D.

• PURPOSE: To evaluate our early experience with with hypotony in 13 (26%) patients and corneal
the Baerveldt implant in patients with complicated graft failure in six of 13 (46%) corneal transplant
glaucoma. patients, respectively.
• METHODS: We reviewed the charts of all pa- • CONCLUSION: Within the study follow-up time,
tients with more than six months of follow-up the Baerveldt implant appeared to be safe and
after placement of a Baerveldt implant at LSU Eye effective, with success rates for intraocular pres-
Center. Surgery was considered a success if intra- sure control similar to those reported in a recent
ocular pressure was 21 mm Hg or less (with or retrospective study of the Molteno implant.
without antiglaucoma medications) at the last
postoperative visit, except when further glaucoma MPLANTATION OF A SETON, DEFINED AS A FOREIGN
surgery had been performed or when loss of light implantable material applied to an organ system to
perception occurred. create a fistulous pathway, is an effective surgical
• RESULTS: Fifty eyes (50 patients) were divided option to control recalcitrant glaucoma. Decreased
into six diagnostic groups, with mean follow-up intraocular pressure is attained by the shunting of
times of 16.1 to 19.2 months. Success was aqueous from the anterior chamber to a conjunctival
achieved in 36 of 50 patients (72%): 26 of 35 bleb overlying the implant. Commonly used setons
(74%) patients with aphakia or pseudophakia, include the single- and double-plate Molteno im-
nine of 12 (75%) patients with previously failed plant, anterior chamber tube shunt to encircling band
filtering surgery, three of seven patients with (Schocket procedure), and the valved Krupin-Denver
neovascular glaucoma, all three patients under the shunt.
age of 13 years, nine of 13 (69%) patients who
underwent penetrating keratoplasty, and four of
See also p. 23.
five phakic patients. Overall, visual acuity im-
proved or remained within one line of the preoper-
ative acuity in 32 (64%) patients. The most In February 1990, the Baerveldt implant (Iovision,
frequently observed short- and long-term compli- Inc., Irvine, California), which consists of a non-
cations were serous choroidal effusion associated valved silicone tube (internal and external diameters
of 0.30 and 0.64 mm, respectively) connected to a
Accepted for publication Jan. 12, 1995. large-surface-area (200, 350, or 500 mm2) silicone
From the LSU Eye Center, Louisiana State University Medical plate, was introduced. The implant is placed in a
Center School of Medicine, New Orleans, Louisiana. This study was
supported in part by United States Public Health Service grants single quadrant, usually superotemporally, by tucking
EY02580 and EY02377 from the National Eye Institute, National the ends of the flexible plate beneath adjacent recti
Institutes of Health, Bethesda, Maryland, and by a Career Develop-
ment Award (Dr. Burgoyne) from Research to Prevent Blindness, Inc., muscles (Fig. 1). The large filtration surface area
New York, New York.
Reprint requests to Claude F. Burgoyne, M.D., LSU Eye Center,
combined with ease of insertion is a presumed
2020 Gravier St., Suite B, New Orleans, LA 70112. advantage over other setons.1 Recently, Lloyd and

32 © AMERICAN JOURNAL OF OPHTHALMOLOGY 1995,-120:32-40 JULY 1995

 patient groupings and surgical outcome criteria were
adopted. Patients older than 13 years of age with
nonneovascular glaucomas were grouped into the
following categories: aphakic or pseudophakic,
phakic, failed filtering surgery, and previous penetrat-
ing keratoplasty surgery. The remaining groups in-
cluded all neovascular glaucomas and all glaucomas
in patients younger than 13 years of age.
The categories of surgical outcome are as follows: a
complete success is defined as intraocular pressure of
21 mm Hg or less without medication; qualified
success, intraocular pressure of 21 mm Hg or less with
one or more medications; qualified failure, intraocular
pressure greater than 21 mm Hg with one or more
medications; and complete failure, additional opera-
tions for glaucoma, development of phthisis bulbi, or
loss of light perception attributed to glaucoma. The
Fig. 1 (Hodkin and associates). Diagram of Baerveldt final intraocular pressure levels and visual acuities
implant positioned beneath adjacent rectus muscles in were those of the most recent examination at either
the superotemporal quadrant. A scierai graft is sutured the LSU Eye Center or the referring ophthalmolo-
over the exposed drainage tube. The overlying conjuncti- gist's office. Both conventional methods and life-table
va and Tenon's capsule are not shown. analysis were used to determine surgical outcomes.
All patients underwent Baerveldt implantation in a
associates2 reported their initial results with the similar fashion. A 90-degree, fornix-based peritomy
implant in 13 patients, as well as intermediate-term through the conjunctiva and Tenon's capsule, with
results comparing two sizes of implant in a total of 73 radial relaxing incisions, was performed in the in-
patients.3 tended quadrant of placement (most commonly
On the basis of our early clinical experience with superotemporal). Bare sciera was exposed by dissect-
the Baerveldt implant in 50 patients with complicated ing Tenon's capsule as well as any scar tissue from
glaucomas, we compare the outcomes with those of a previous surgery. Hemostasis was achieved with wet-
previous retrospective study of the Molteno implant.4 field cautery. A muscle hook was used to isolate the
exposed adjacent recti muscles, for example, lateral
and superior recti muscles for supratemporal place-
PATIENTS AND METHODS ment.
The ends of a 350-mm2 Baerveldt implant were
A CHART REVIEW WAS CONDUCTED FOR ALL PATIENTS positioned beneath the recti muscles 2 mm posterior
who received a Baerveldt implant at the LSU Eye to the muscle insertions. Interrupted 6-0 polyglactin
Center, New Orleans, Louisiana. Only patients with sutures placed through the two anterior fixation holes
more than six months of postoperative follow-up were secured the implant to the sciera. For Cases 13
included. From July 1991 through January 1993, through 50, to enable titration of intraocular pressure
Baerveldt implantation was performed in one eye control better in the early postoperative period, the
each of 50 patients with complicated glaucoma, needle-free end of a 5-0 nylon suture was threaded
which is denned as having previously failed conven- through the Baerveldt tube lumen to serve as a partial
tional medical, laser, or (nonseton) surgical treat- occlusion stent. To reduce excessive drainage and
ment, or some combination thereof. All patients (or hypotony in the immediate postoperative period for
their parents) gave informed consent before surgery. all patients, a single 6-0 polyglactin ligature was tied
To allow a more direct comparison of our results around the part of the drainage tube most proximal to
with those of a recent Molteno implant study,4 similar the plate.

VOL. 120, No. i BAERVELDT IMPLANT IN COMPLICATED GLAUCOMAS 33

 A paracentesis was performed 90 degrees away from ointment and combined neomycin, polymyxin b
the intended tube entry site to allow injection of air, sulfate, and dexamethasone ointment. The eye was
balanced saline solution, or sodium hyaluronate dur- loosely patched.
ing the procedure. To create the tube entry site, a Postoperative examinations were performed one
22-gauge needle was passed through sciera 1.5 mm day after surgery, then weekly (or more frequently, if
from the corneoscleral limbus to enter the anterior necessary), and finally at longer intervals, gradually
chamber at an angle directed slightly away from the increasing to several months. Topical medications
iris plane. The Baerveldt tube was cut with enough consisted of antibiotics for one week, 1% atropine
length to extend approximately 3 to 4 mm into the sulfate drops for two weeks, and 1% prednisolone
anterior chamber. A sharp bevel created on the cut acetate. The prednisolone acetate was typically given
end of the tube facilitated entry into the scierai tract. four to six times a day and tapered over a four- to
Approximately six to ten microfenestrations were six-week period or until anterior chamber inflamma-
created in the tube wall with the needle of a 10-0 tion had resolved.
polyglactin suture in approximately one half of the The 5-0 nylon stent suture was removed within two
patient^ to allow some aqueous drainage in the early to five weeks postoperatively, or earlier if intraocular
postoperative period. The microfenestrations were pressure was judged to be unacceptably high. Removal
placed midway between the anterior chamber tube of the suture was usually followed by a prompt
entrance site and the tube connection with the decrease in intraocular pressure and expansion of the
Baerveldt implant. A glycerin-preserved donor sciera bleb overlying the implant. Either topical antiglauco-
graft was thoroughly rinsed and then sutured to host ma medications or a combination of topical and
sciera over the tube with four interrupted 10-0 systemic antiglaucoma medications were added dur-
polyglactin sutures to prevent erosion of the tube ing follow-up as deemed necessary by the surgeon.
through conjunctiva.
To facilitate removal of the 5-0 nylon stent suture
postoperatively, the needle end was externalized RESULTS
through conjunctiva near the inferior fornix and
fixated to bulbar conjunctiva with interrupted 10-0 FIFTY BAERVELDT IMPLANTS WERE PLACED IN 50 EYES BY
polyglactin sutures. The quadrant peritomy was closed a total of six LSU Eye Center faculty. A total of 38
by reapproximating conjunctiva to the corneoscleral (76%) operations were performed by one of us
limbus with interrupted 10-0 polyglactin sutures. (S.F.B.); the remaining five surgeons placed from one
At the conclusion of surgery, a collagen shield to four Baerveldt implants each.
soaked in vancomycin (50 mg/ml) was placed on the Except for the neovascular and under-age-13-years
eye, followed by instillation of 1% atropine sulfate groups, which were mutually exclusive categories,

TABLE 1

PATIENT OVERLAP AMONG DIAGNOSTIC GROUPS*

APHAKIA OR FAILED UNDER

PSEUDOPHAKIA FILTERS NEOVASCULAR AGE 13 YEARS KERATOPLASTY PHAKIC
CATEGORY (N ■ 35) (N = 12) (N = 7) (N-3) (N = 13) {N=5)

Aphakia or pseudophakia — 11 0 0 13 0
Failed filters 11 — 0 0 1 1
Neovascular 0 0 — 0 0 0
Under age 13 years 0 0 0 — 0 0
Keratoplasty 13 1 0 0 — 0
Phakic 0 1 0 0 0 —
"One patient was included in three categories: aphakia or pseudophakia, keratoplasty, and failed filters

34 AMERICAN JOURNAL OF OPHTHALMOLOGY JULY 1995

 TABLE 2

PATIENT DATA AND OUTCOMES

APHAKIA OR FAILED UNDER

PSEUDOPHAKIA FILTERS* NEOVASCULAR AGE 13 YEARS KERATOPLASTV PHAKIC
(N = 35) (N = 12) (N = 7) IN = 3 ) IN = 13) (N = S)

CATEGORY NO. (%) NO. (%) NO. pi) NO. <%) NO. (%) NO. (%)

Age (yrs)
Range 34.9-97.8 34.9-82.2 51.7-90.2 0.6-7.3 57.6-87.8 39.3-57.4
Mean ± S.D. 68.4 ± 12.1 62.6 ± 15.4 70.4 ± 15.5 3.2 ± 3.6 71,1 ± 8.6 47.9 ± 8.5
Race
Black 11 (31) 5(42) 2(29) 1(33) 4(31) —
White 24 (69) 7(58) 5(71) 2(67) 9(69) 5 (100)
Lens status
Phakic — 1(8) 2(29) 3 (100) — 5(100)
Aphakic 8(23) 2(17) 3(43) — 3(23) —
Pseudophakic 27 (77) 9(75) 2(29) — 10 (77) —
Anterior chamber intraocular lens 6(17) 2(17) 1(14) — 2(15) —
Posterior chamber intraocular lens 21 (60) 7(58) 1 (14) — 8(62) —
Diagnosis
Open-angle glaucoma 21 (60) 5(42) — — 10 (77) 2(40)
Secondary angle-closure glaucoma 10 (29) — — — 3(23) —
Combined mechanism 1(3) 3(25) — — — —
Anterior cleavage syndrome 2(6) 2(17) — — — 1(20)
Uveitis-glaucoma-hyphema syndrome 1(3) 1(8) — — — —
Angle recession — 1(8) — — — —
Neovascular glaucoma — — 7 (100)T — — —
Congenital glaucoma — — — 2(67) — —
Glaucoma associated with Sturge-Weber syncIrome — — — 1(33) — —
Iridocorneal endothelial syndrome — — — — — 2(40)
Surgical outcome
Success 26 (74) 9(75) 3(43) 3(100) 9(69) 4(80)
Complete success 13(37) 1(8) 2(29) 2(67) 5(38) 1(20)
Qualified success 13(37) 8(67) 1(14) 1(33) 4(31) 3(60)
Failure 9(26) 3(25) 4(57) — 4(31) 1(20)
Qualified failure 3(9) 1(8) 3(43) — KB) —
Complete failure 6(17) 2(17) 1(14) — 3(23) 1(20)
Further glaucoma surgery 5(14) 2(17) — — 3(23) 1(20)
Follow-up (mos)
Range 6.1-26.1 7.1-26.1 10.6-24.7 17.0-20.4 8.2-24.0 9.3-18.4
Mean ± S.D. 16.3 ± 6 . 1 16.1 ± 6.5 18.3 ± 5.6 19.2 ± 1.9 15.2 ± 5.7 13.7 ± 3.6
Visual outcome
Improved (gain of two or more lines) 6(17) 1(8) — — 4(31) —
No change (± one line) 16 (46) 5(42) 3(43) 1(33) 7(54) 2(40)
Worse (loss of two or more lines) 13 (37)* 7(58) 4(57) 2(67) 2(15) 3 (60)'

*Eight patients (67%) had one failed filtering procedure, and four patients (33%) had two failed filtering procedures.
♦Associated diagnoses were diabetes, two (29%); central retinal vein occlusion, two (29%); siderosis, one (14%); and unknown, two
(29%).
'Attributable to fibrous membrane covering intraocular lens in one patient and uncorrected aphakic visual acuity in one patient.
'Attributable to decompensated cornea in one patient.

VOL.120, No. I BAERVELDT IMPLANT IN COMPLICATED GLAUCOMAS 35

each of the remaining groups shared at least one
patient in common with at least one other group
(Table 1). One patient was included in three catego-
ries: aphakia or pseudophakia, penetrating kerato-
plasty, and failed trabeculectomy.
Demographic, preoperative, and postoperative data
for the six patient groups are summarized in Table 2.
Overall, complete or qualified success was achieved in
36 of the 50 patients (72%): 26 of 35 patients (74%)
with aphakia or pseudophakia, nine of 12 patients
(75%) with previously failed filtering surgery, three of
0 10 20 30 40 50 60 70
the seven patients with neovascular glaucoma, all
Preoperative IOP (mmHg)
three of the patients under the age of 13 years, nine of
the 13 penetrating keratoplasty patients (69%), and Fig. 2 (Hodkin and associates). Successes (open circles)
were defined as patients with intraocular pressure (IOP)
four of the five phakic patients. Best-corrected visual
21 mm Hg or less with or without medication. Failures
acuity improved or remained within one line of (closed circles) were defined as patients with intraocular
preoperative levels in 21 patients (60%) with aphakia pressure greater than 21 mm Hg with medications and
or pseudophakia, six patients (50%) with previously those requiring further glaucoma surgery, as well as
failed filtering surgery, three of seven patients with those who developed phthisis bulbi or lost light percep-
neovascular glaucoma, one of three patients under tion as a result of the glaucoma.
the age of 13 years, 11 penetrating keratoplasty
patients (85%), and two of five phakic patients.
Overall, best-corrected visual acuity improved or To determine the intraocular pressure response
remained within one line of the preoperative acuity in after initiation of maximum aqueous flow, the intra-
32 of 50 patients (64%). One patient (2%) with ocular pressure change was evaluated over time for 19
neovascular glaucoma (final intraocular pressure of 16 successful cases with documented stent suture remov-
mm Hg) lost light perception, presumably because of al for pressure control during the early postoperative
progression of optic nerve damage. period (mean time to suture removal, 19.2 days; range,
For the 36 successful procedures, intraocular pres- five to 59 days) (Fig. 4). Intraocular pressure tended to
sure was reduced by surgery and medications from a increase slowly for up to six months, followed by a
mean ± S.D. preoperative value of 33.7 ± 10.7 mm slow decline.
Hg to 12.6 ± 4.3 mm Hg postoperatively. For the 14
failed procedures, intraocular pressure declined from
33.9 ± 10.6 mm Hg to 21.6 ± 10.6 mm Hg (Fig. 2). DISCUSSION
For the entire study population, life-table analysis
of success after Baerveldt implantation showed a THE BAERVELDT IMPLANT IS COMPOSED OF A LOW-
steady decline over time (Fig. 3), with a median (50th profile, soft (radiopaque) silicone plate that conforms
percentile) time to failure of 721 days. Because of the to the shape of the globe after implantation. A
small number of patients per clinical grouping, group- nonvalved silicone tube drains aqueous from the
specific life-table analysis could not be performed. anterior chamber to the filtering bleb overlying the
Table 3 lists all complications by diagnostic groups. 200-, 350-, or 500-mm2 surface area plate.
Overall, the most frequently observed short-term Although the Molteno implant (Optomat Sup-
complication was serous choroidal effusion, which plies, Dunedin, New Zealand) is currently the most
was seen in 13 patients (26%). The most frequent widely used artificial aqueous drainage device, its
long-term complication was corneal graft failure, surface area (as measured on one side) is limited to
which was seen in six of 13 patients (46%) with 135 mm2 and 270 mm2 for the single- and double-
previous penetrating keratoplasty. plate models, respectively. Furthermore, implantation

36 AMERICAN JOURNAL OF OPHTHALMOLOGY JULY 1995

of the double-plate Molteno implant for patients criteria, data analysis, and patient groupings were
requiring maximum intraocular pressure control re- adopted (Table 4). Although follow-up times were
quires dissection in two quadrants.4,5 In contrast, similar, the numbers of patients in the neovascular
placement of even the largest (500 mm2) Baerveldt and younger-than-age-13 -years groups were substan-
implant is easily accomplished by tucking the seton tially lower in the present study, thus possibly skewing
ends beneath adjacent recti muscles in a single the results. Also, whereas all patients were phakic in
quadrant. Finally, the Baerveldt implant is composed the failed-filtering group in the Molteno study, in this
of silicone, which may be less inflammatory than the study, only one of the 12 patients in our correspond-
polypropylene Molteno plate.6 ing group was phakic.
Comparing the results of shunt device studies is In comparing rates of significant complications, we
difficult, in part because of differences in patient found a higher incidence of serous choroidal effusion
population, success criteria, data analysis, operating and hypotony in our study for patients in the aphakia
surgeons, surgical techniques, postoperative medica- or pseudophakia group (31% vs 4% in theirs) and in
tions, and length of follow-up. To facilitate compari- the failed-filtering group (42% vs 10% in theirs).
son of our study with a large retrospective Molteno However, these differences may in part result from
study by Minckler and associates,4 similar success differing criteria, since the choroidal effusions noted

TABLE 3

COMPLICATIONS BY DIAGNOSTIC GROUP

APHAKIA OR FAILED UNDER

PSEUDOPHAKIA FILTERS NEOVASCULAR AGE 13 YEARS KERATOPLASTY PHAKIC
(N - 35) (N = 121 (N = 7) (N-3) (N = 13) (N = 5)

COMPLICATION NO. (%) NO. (%) NO. (%) NO. (%) NO. (%) NO. (%)

Serous choroidal effusion 11 (31) 5(42) — — 4(31) 1(20)

Corneal graft failure 7(20) 1(8) 1(14) — 6 (46)* —
Epiretinal membrane 5(14) — — — 4(31) —
Corneal edema 2(6) 1(8) — — — —
Hyphema 1(3) 1(8) 1 (14) — 1(8) —
Bacterial keratitis 1(3) — — — 1(8) —
Cystoid macular edema 1(3) — — — — —
Epithelial ingrowth 1 (3) — — — 1(8) —
Exposed Baerveldt implant 1 (3) — — — 1(8) 1(20)
Exposed Baerveldt tube — — — — — 1(20)
Exposed donor sciera 1 (3) — — — 1(8) —
Flat anterior chamber 1(3) 2(17) 1(14) — — 1(20)
Occlusion of tube by 1 (3) — — — 1(8) —
retrocorneal membrane
Occlusion of tube by vitreous 1 (3) — — — 1(8) —
Retinal detachment 1(3) — 1 (14) — — —
Superior oblique tendon 1 (3) 1 (8) 1(14) — — 1(20)
syndrome
Vertical diplopia 1 (3) — — — — —
Vitreous hemorrhage 1(3) 1(8) — — — —
Macular edema — — 1(14) — — —
Cataract — — — 1 (33) — —
Bleb leak — — — — — 1 (20)
Corneal decompensation — — — — — 1(20)

*ln two patients (15%), failure was related to Baerveldt-tube touch, with postoperative times to failure of 2,3 and 2.5 months. In four
patients (31%), failure was not directly Baerveldt related, with postoperative times to failure ranging from 0.9 to 22.4 months.

VOL.120, No. I BAERVELDT IMPLANT IN COMPLICATED GLAUCOMAS 37

 1
^ thelial touch, while the remaining four graft failures

κ
h 1 (mean time to failure, 10.0 months) were thought not
1
0.8
to be directly related to the presence of the Baerveldt

0.6
L= implant. Overall, the 69% success rate that we
obtained for intraocular pressure control with the
I 0.4
Baerveldt implant after penetrating keratoplasty is
comparable to the 7 1 % to 96% range of success rates
reported for various studies of Molteno and Schocket
implants.7'10
Recently, studies of restrictive strabismus after
Baerveldt implantation have been published.1,11'12 Pro-
posed explanations include an increase in the length-
tension curve of the muscle(s) induced by the under-
Fig. 3 (Hodkin and associates). Kaplan-Meier analysis
for implants in all patients in all groups after Baerveldt lying Baerveldt implant plate, its fibrous cocoon, or
implantation. Median (50th percentile) time to failure both; incorporation of the rectus muscle(s) into the
was 721 days. fibrous cocoon surrounding the plate; or possibly a
posterior fixation effect caused by scarring posterior to
the implant. 11112 Smith and associates1 studied 30
by Minckler and associates4 were advanced enough to eyes examined retrospectively for strabismus after
be labeled "kissing choroidals," whereas we recorded placement of a 350-mm 2 Baerveldt implant; 23 (77%)
choroidal effusions of any size as a complication. were found to have a marked heterotropia in primary
In our study, five of the 13 postoperative serous gaze and restriction of gaze into the quadrant of the
choroidal effusions with hypotony (intraocular pres- implant. As a result, 11 (65%) of 17 functionally
sure less than 5 mm Hg) were observed to occur 12.4 binocular patients had diplopia in primary gaze. Most
± 1 1 . 6 (mean ± S.D.) days after stent removal, and commonly, supratemporal Baerveldt placement re-
one was seen five days after lysis of the ligature suture sulted in a hypertropia with exotropia, whereas supra-
with a laser. Three cases occurred spontaneously 4-7
± 2.6 days after Baerveldt implantation, despite
ligature placement during surgery. The remaining 50
four cases were not related closely in time (greater
than one month) either to Baerveldt implantation or 40
to stent removal. After long-term follow-up, four of
the 13 patients with choroidal effusions (31%) were 30
eventually classified as having graft failures. Two of E
E_
the failures were caused by high final intraocular 20
pressure, and the remaining two required choroidal o
drainage surgery and removal of an exposed Baerveldt 10
implant.
In patients with previous penetrating keratoplasty, pre post 1 wk 1 mo 6 mo 1 yr 2 yrs >2 yrs
a graft failure rate of 36% to 60% at a mean follow-up Time Interval After Aqueous Flow Initiated
of about one to two years has been observed in
previous studies of Molteno implantation and Fig. 4 (Hodkin and associates). Mean intraocular pres-
sure (± S.D.) for 19 patients with Baerveldt implants
Schocket procedures.7'10 We observed graft failure in
who were considered to be successes and who also had
six of 13 patients (46%) at a mean follow-up of 15.2
documented stent suture removal for control of intraocu-
months, a rate similar to that of the Molteno implant lar pressure (IOP) in the postoperative period. Pre
studies. Two of the graft failures (mean time to failure, indicates immediately before stent suture removal; post,
2.4 months) were attributed to Baerveldt-tube endo- immediately after stent suture removal.

38 AMERICAN JOURNAL OF OPHTHALMOLOGY JULY 1995

 TABLE 4

COMPARABLE GROUPS IN CURRENT BAERVELDT STUDY AND A RETROSPECTIVE MOLTENO IMPLANT STUDY*

VISUAL OUTCOME
FOLLOW-UP (MOS) SUCCESS SAME OR BETTER

DIAGNOSTIC GROUP NO. OF PATIENTS RANGE MEAN ± S.D. NO. (%) NO. (%(

Aphakia or pseudophakia
Minckler and associates4 48 7-30 16.2 ± 5.9 26 (63) 31 (65)
Current study 35 6-26 16.3 ± 6.1 26 (74) 22 (63)
Failed filters
Minckler and associates' 10 6-25 12.3 ± 6.1 7 (70) 3 (30)
Current study 12 7-26 16.1 ± 6.5 9 (75) 6 (50)
Neovascular glaucomas
Minckler and associates4 18 7-39 20.2 ±12.1 7 (47) 10 (56)
Current study 7 11-25 18.3 ± 5.6 3 (43) 3 (43)
Under age 13 years
Minckler and associates' 14 12-34 22.8 ± 7.8 7 (54) 9 (64)
Current study 3 17-20 19.2 ± 1.9 3 (100). 1 (33)

'Minckler and associates.'

nasal placement resulted in a hypotropia with exotro- based, rather than a limbal-based, conjunctival inci-
pia and restricted elevation in adduction, similar to sion for implantation. A limbal-based flap may pro-
Brown's syndrome. The latter manifestation was at- mote formation of scar tissue beneath the posterior
tributed to a taut superior oblique tendon.1 incision site, possibly leading to muscle scarring with
Our review of 50 patients disclosed only four cases motility restriction (Paul Palmberg, M.D., oral com-
of ocular motility disturbance. Three of them simulat- munication, April 1994). Fourth, since the study by
ed a superior oblique tendon syndrome. Notably, all Smith and associates1 was published, first two and
three were among a subset of seven patients with then four fenestrations were added to the design of
supranasal placement of the Baerveldt implant. The the plate to promote fibrous adhesions between the
fourth case involved vertical diplopia in a patient with sciera and the inner surface of the bleb (oral commu-
supratemporal implantation. Overall, this incidence nication, Iovision, Inc., Irvine, California, January
of strabismus was much lower than that reported by 1994). Theoretically, such adhesions might reduce
Smith and associates.1 At least four factors may bleb height by securing the inner surface of the bleb
account for this difference. First, in the present study, closer to the underlying sciera.112 In our study, 18 of
strabismus complications were not actively investigat- the 50 implants had the two-fenestration modifica-
ed during routine patient examinations. As a result, tion to the plate. None of the ocular motility distur-
strabismus may have been underreported, particularly bances occurred in the patients who received fenes-
in functionally monocular patients, who are much less trated plates.
likely to complain spontaneously of ocular muscle In general, we found that our results, in terms of
imbalance. Second, during implantation we tight- intraocular pressure control and visual outcome, were
ened the tube ligature to prevent any flow to the similar to those of a recent Molteno implant review4
Baerveldt implant plate in the early postoperative and not widely dissimilar to those of the recently
period, whereas Smith and associates1 allowed mini- published early studies by the Baerveldt group.2,3 Our
mal flow. As stated in their discussion, allowing early success rate (74%; 36 of 50 cases) was somewhat
flow before the development of the fibrous cocoon better than the 62% success rate (eight of 13 cases) in
may have been a factor in the development of their initial experience2 in which essentially the same
prominent heterotropia.1 Third, we used a fornix- criteria were used, but was somewhat worse than the

VOL.120, No. I BAERVELDT IMPLANT IN COMPLICATED GLAUCOMAS 39

84% (31 of 37 cases) and 83% (30 of 36 cases) JF. Initial clinical experience with the Baerveldt implant in
success rates in their subsequent prospective compari- complicated glaucomas. Ophthalmology 1994;101:640-50.
3. Lloyd MA, Baerveldt G, Fellenbaum PS, Sidoti PA, Minckler
son of 350-mm2 and 500-mm2 implants, on the basis DS, Martone JF, et al. Intermediate-term results of a random-
of conventional outcome analysis by final intraocular ized clinical trial of the 350- versus the 500-mmz Baerveldt
pressure.3 implant. Ophthalmology 1994;101:1445-64.
4. Minckler DS, Heuer DK, Hasty B, Baerveldt G, Cutting RC,
In conclusion, the Baerveldt implant appears to be Barlow WE. Clinical experience with the single-plate Mol-
an effective means of lowering intraocular pressure in teno implant in complicated glaucomas. Ophthalmology
1988;95:1181-8.
patients with complicated glaucoma. Subjectively (but 5. Lloyd MA, Sedlak T, Heuer DK, Minckler DS, Baerveldt G,
without an objective basis for comparison), we found Lee MB, et al. Clinical experience with the single-plate
the Baerveldt implant easier to place than the Mol- Molteno implant in complicated glaucomas: update of a pilot
study. Ophthalmology 1992;99:679-87.
teno implant, especially for surgeons who do not 6. Wilson RP, Cantor L, Katz LJ, Schmidt CM, Steinmann
routinely implant setons. Except for a high incidence WC, Allee S. Aqueous shunts: Molteno versus Schocket.
of superior oblique tendon syndrome in the subset of Ophthalmology 1992;99:672-8.
7. Kirkness CM. Penetrating keratoplasty, glaucoma and sili-
patients with supranasal placement, we did not ob- cone drainage tubing. Dev Ophthalmol 1987;14:161-5.
serve the high percentage of ocular motility problems 8. McDonnell PJ, Robin JB, Schanzlin DJ, Minckler D,
reported by others after Baerveldt implantation. Fi- Baerveldt G, Smith RE, et al. Molteno implant for control of
glaucoma in eyes after penetrating keratoplasty. Ophthalmol-
nally, our 26% incidence of temporary serous choroi- ogy 1988;95:364-9.
dal effusions with hypotony after stent suture removal 9. Beebe WE, Starita RJ, Fellman RL, Lynn JR, Gelender H.
indicates that a modification of our technique is The use of Molteno implant and anterior chamber tube
shunt to encircling band for the treatment of glaucoma in
required to more finely titrate initial filtration. keratoplasty patients. Ophthalmology 1990;97:1414-22.
10. Sherwood MB, Smith MF, Driebe WT Jr, Stern GA, Beneke
JA, Zam ZS. Drainage tube implants in the treatment of
glaucoma following penetrating keratoplasty. Ophthalmic
REFERENCES Surg 1993;24:185-9.
11. Ball SF, Ellis GS Jr, Herrington RG, Liang K. Brown's
superior oblique tendon syndrome after Baerveldt glaucoma
1. Smith SL, Starita RJ, Fellman RL, Lynn JR. Early clinical
implant [letter]. Arch Ophthalmol 1992;110:1368.
experience with the Baerveldt 350-mm2 glaucoma implant
12. Munoz M, Parrish RK II. Strabismus following implantation
and associated extraocular muscle imbalance. Ophthalmolo-
of Baerveldt drainage devices. Arch Ophthalmol 1993;
gy 1993;100:914-8.
111:1096-9.
2. Lloyd MAE, Baerveldt G, Heuer DK, Minckler DS, Martone

40 AMERICAN JOURNAL OF OPHTHALMOLOGY JULY 1995