Shunt Revision Versus Additional Tube ShuntImplantation After Failed Tube Shunt Surgery

in Refractory Glaucoma

AVANEE A. SHAH, MD, DARRELL WUDUNN, MD, PHD, AND LOUIS B. CANTOR, MD

● PURPOSE: To compare tube shunt revision with addi-tional tube shunt after failed tube shunt surgery.● METHODS: We identified 281 patients who underwenta primary tube shunt procedure from 1985 to 1998 atIndiana University and reviewed 33 eyes of 33 patientsthat had failed and required further surgery. Shuntrevision was performed in 12, whereas an additionalshunt was placed in 21 eyes. Intraocular pressure, anti-glaucoma medications, visual acuity, and complicationswere noted. Success was defined as at least a 25%reduction in intraocular pressure that was deemed clini-cally adequate. Qualified success was defined as a 25%intraocular pressure reduction but with additional medi-cations or a significant reduction in medications withstable intraocular pressure for preoperative intraocularpressure less than 21 mm Hg.● RESULTS: Preoperative intraocular pressures (mean 695% confidence interval) for the revision and additionaltube groups were 28.8 6 5.8 mm Hg and 29.8 6 2.7 mmHg (P 5 .73), with an average follow-up period of 25.2months (range, 3 to 108 months) and 34.8 months(range, 6 to 84 months), respectively. Final mean in-traocular pressure was 25.3 6 6.7 mm Hg for therevision group and 17.7 6 3.4 mm Hg for the additionaltube group (P 5 .037). Forty-two percent in the revisiongroup versus 62% in the additional tube group achievedat least a qualified success (P 5 .30, Fisher exact test).Corneal edema was a common complication, especially inthe additional tube group. Limitations of this studyinclude the small sample sizes and the uneven distribu-tion of neovascular glaucoma between the two groups(six of 12 in the revision group vs two of 21 in theadditional tube group; P 5 .015, Fisher exact test).

● CONCLUSIONS: Our series showed that after failedtube shunt surgery, an additional tube shunt offers betterintraocular pressure control than revision by excision ofan encapsulated bleb. (Am J Ophthalmol 2000;129:455–460. © 2000 by Elsevier Science Inc. All rightsreserved.)

T UBE SHUNT IMPLANTS HAVE BEEN USED IN THE MAN-

agement of complicated glaucomas. Indications forthe use of tube shunts include those eyes with

refractory glaucoma, such as neovascular glaucoma, uveiticglaucoma, glaucoma in aphakia and pseudophakia, con-genital and developmental glaucoma, and eyes that havefailed previous filtration surgery.1,2

A tube shunt implant consists of a silicone tube placedin the anterior chamber that drains the aqueous to aposteriorly placed plate, disk, or encircling element reser-voir in the subconjunctival space.1 The Molteno,3Ahmed,4 Schocket,5 Baerveldt,6 and Krupin7 implants areexamples of such devices. Tube shunts have been provento be successful in the treatment of refractory glaucoma,8but failure to control intraocular pressure adequately doesoccur in some cases.

The major cause of failure is excessive fibrosis.2,9,10 Thethick fibrous capsule around the reservoir leads to in-creased intraocular pressure. Etiologic factors for this fi-brotic response depend on the individual patient, butseverity of glaucoma and multiple previous surgeries seemto be risk factors.9,11

When failure occurs in such eyes, excision of theencapsulated bleb may be attempted or an additional tubeshunt may be placed. Some may consider revision bycapsule excision to be less extensive than implantation ofan additional tube shunt.10 One study reported a 52%success rate with capsule excision alone.10 Based on ourexperience, the revision procedure can be technicallydifficult because of excessive bleeding and scar tissue,whereas placing another tube shunt in a new location mayavoid such vascularized scar tissue. The purpose of thisretrospective study was to determine whether revision byexcision of the encapsulated bleb or implantation of an

Accepted for publication Oct 14, 1999.From the Glaucoma Service, Department of Ophthalmology, Indiana

University School of Medicine, Indianapolis, Indiana.This study was supported in part by an unrestricted grant from Research

to Prevent Blindness.Correspondence to Louis B. Cantor, MD, Department of Ophthalmol-

ogy, Indiana University School of Medicine, 702 Rotary Circle, Room145, Indianapolis, IN 46202; fax: (317) 278-1007; e-mail: lcantor@iupui.edu

© 2000 BY ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED.0002-9394/00/$20.00 455PII S0002-9394(99)00410-9

additional tube shunt affords better intraocular pressurecontrol.

METHODS

BY A CHART REVIEW OF 281 PATIENTS WITH A HISTORY OF

a tube shunt implant (from 1985 to 1998), we identified33 eyes of 33 patients who experienced increasingintraocular pressures requiring further surgical interven-tion. Revision by excision of an encapsulated bleb wasperformed in 12 eyes, whereas an additional tube shuntwas placed in 21 eyes. In general, the decision thatdetermined which of the procedures would be done wasbased on preoperative clinical examinations. If thepatient appeared to have an elevated functioning bleboverlying one or both plates of the implant, additionaltube shunt implantation was usually preferred overrevision. If the bleb was low or absent over the implant,revision was performed. The types of primary andadditional implants used included the Molteno (IOP,Inc, Costa Mesa, California), Baerveldt (Pharmacia &Upjohn, Kalamazoo, Michigan), and Ahmed (NewWorld Medical, Rancho Cucamonga, California) de-vices (Table 1).

The revision and additional tube shunt implantationprocedures were performed at the Indiana UniversityGlaucoma Service between 1988 and 1998. Demographicdata, including age, sex, and race, were noted. For eachpatient, glaucoma diagnosis, history of intraocular surgeriesand laser procedures, and types of tube shunt implants wererecorded. The time intervals from the first tube shunt torevision or additional tube shunt and from the revision oradditional tube shunt to final follow-up were determined.All charts were reviewed until the most current follow-upvisit, with a minimal follow-up period of 3 months.Evaluation measures included intraocular pressure, numberof antiglaucoma medications, and best-corrected Snellenvisual acuity. These data were recorded immediately beforethe first tube shunt, revision or additional tube shunt, andat the final follow-up. Major postoperative complicationswere noted.

Criteria for success were defined as at least a 25%reduction in intraocular pressure, on equal or fewer med-ications, and the final pressures were deemed clinicallyadequate. Qualified success was defined as at least a 25%reduction in intraocular pressure with additional medica-tions. For patients with preoperative intraocular pressureless than 21 mm Hg, qualified success was also defined as asignificant reduction of medications (at least three) withan intraocular pressure within 2 mm Hg of preoperativepressure. Failures did not meet any of the above criteria,required further surgical intervention for intraocular pres-sure control, experienced phthisis, or progressed to no lightperception.

Failure of the revision procedure was considered the end

point. Six patients in the revision group went on to havean additional tube shunt placed, but these patients werenot included in the additional tube group.

Statistical analysis was accomplished using GraphPadPrism 3.0. Data are presented as the mean 6 95%confidence interval of the mean. The Student t test wasused to compare the two groups regarding baseline andoutcome measures and to determine statistical significance.A P value less than .05, two-tailed test, was consideredstatistically significant. Kaplan-Meier product limit sur-vival analysis was also performed.

RESULTS

COMPREHENSIVE INDIVIDUAL DATA FOR PATIENTS WITH

revision and patients with additional tube shunt implan-tation are shown in Table 1.

The revision group included 12 eyes, whereas theadditional tube group included 21 eyes. Both groups weresimilar in terms of age, sex, race, number of prior surgeries,preoperative intraocular pressure, and number of antiglau-coma medications.

The average age at the time of surgery was 60.4 6 8.7years for the revision group and 58.7 6 8.5 years for theadditional tube group (P 5 .79). Previous surgeries andlaser procedures before revision or additional tube shunt,excluding the first tube shunt implantation, includedtrabeculectomy, goniotomy, cataract extraction, penetrat-ing keratoplasty, vitrectomy, cyclocryotherapy, retinal de-tachment repair, argon laser trabeculoplasty, Nd:YAGperipheral iridectomy, Nd:YAG posterior capsulotomy,and Nd:YAG lysis of a fibrin membrane in the tube. Thetotal number of such procedures was similar for both groups(revision group, 3.3 6 1.2 procedures; additional tubegroup, 3.4 6 1.3 procedures, P 5 .92).

There were six patients (50%) with neovascular glau-coma in the revision group, whereas only two such patients(10%) received an additional tube shunt (P 5 .015, Fisherexact test).

Both groups were similar regarding preoperative intraoc-ular pressures and medications before the first tube shuntand before the revision or additional tube shunt proce-dures. The mean intraocular pressure before the first tubeshunt surgery was 37.9 6 7.3 mm Hg for the revision groupand 37.9 6 4.8 mm Hg for the additional tube group (P 5.99). Mean preoperative medications before the first tubeimplant were 2.4 6 0.7 for the revision group and 2.3 60.5 for the additional tube group (P 5 .90). The averagetime to failure after the first tube shunt was 24.3 months(range, 2 to 58 months) for the revision group and 17.9months (range, 3 to 84 months) for the additional tubegroup (P 5 .37). The mean preoperative intraocularpressure was 28.8 6 5.8 mm Hg with 1.9 6 0.7 medica-tions before revision and 29.8 6 2.7 mm Hg with 2.1 6 0.4medications before additional tube shunt implantation

AMERICAN JOURNAL OF OPHTHALMOLOGY456 APRIL 2000

(P 5 .73 and P 5 .64, respectively). The average time tofinal follow-up was 25.2 months (range, 3 to 108 months)for the revision group and 34.8 months (range, 6 to 84months) for the additional tube group (P 5 .36).

The intraocular pressure at final follow-up was signifi-

cantly lower in the additional tube group (25.3 6 6.7 mmHg for the revision group; 17.7 6 3.4 mm Hg for theadditional tube group, P 5 .037). The mean percentreduction in intraocular pressure at final follow-up was alsosignificantly greater in the additional tube group (7% for

TABLE 1. Individual Patient Data

Age

(yrs) Sex Race

Type of

Glaucoma

No. of

Previous

Surgeries

Primary

Tube

Type

Interval*

(mos)

Additional

Tube

Type

Preoperative Postoperative

Follow-up†

(mos) Complications

IOP

(mm Hg) VA Meds

IOP

(mm Hg) VA Meds

Revision group

80 M W UGH, POAG 5 DPM 50 — 19 20/20 4 30 CF 1 1 CE (graft)

68 M W NVG 2 DPM 4 — 50 20/200 1 45 20/100 1 1 —

54 F W NVG 0 DPM 22 — 29 20/200 2 26 CF 3 49 VH

68 M W NVG 0 DPM 3 — 27 CF 2 44 CF 1 1 —

51 F W NVG 4 DPM 57 — 25 HM 0 11 HM 0 32 CE

59 F W Aph ACG 2 DPM 58 — 13 CF 4 14 CF 1 4 CE; PKP

61 F W NVG 4 Ahmed 7 — 47 20/200 0 18 20/80 0 4 —

74 M W POAG 4 ACTSEB 41 — 26 20/30 2 15 20/60 2 66 —

32 F W Aph ACG 4 ACTSEB 10 — 27 20/60 2 13 20/80 1 108 CE

46 M W NVG 5 ACTSEB 28 — 26 20/100 2 27 20/100 1 28 CE; VH

47 F B Uveitic ACG 7 ACTSEB 10 — 28 HM 2 37 20/30 2 1 —

85 M W POAG 3 SPM 2 — 28 20/100 2 23 20/400 3 7 —

Additional tube group

48 F W Aph OAG 5 DPM 14 Ahmed 36 20/40 3 17 20/30 0 13 VH

1 F W Cong glc 1 SPM 5 SPM 33 No fix 1 29 20/100 2 55 —

67 F W POAG 3 SPM 19 DPM 30 20/400 3 22 HM 1 60 CE

38 F W Aph OAG 2 SPM 25 DPM 20 20/100 0 18 20/100 0 42 —

68 M B POAG 4 DPM 7 DPM 32 20/30 2 23 20/50 3 69 —

67 F W POAG 3 SPM 26 DPM 30 20/100 2 16 HM 1 9 CE; VH

85 F W UGH 2 SPM 3 SPM 37 20/200 3 31 CF 1 34 Choroidal; CE

78 F W POAG 2 DPM 4 DPM 34 20/30 2 11 20/25 1 64 —

69 M W NVG 0 SPM 7 SPM 28 20/60 2 24 20/100 1 6 —

65 M W POAG 3 DPM 12 DPM 19 20/60 3 17 20/40 0 16 —

46 F B NVG, Ghost Cell 2 DPM 27 DPM 36 CF 2 24 LP 2 20 —

65 F W POAG, SACG 4 DPM 7 DPM 28 20/30 1 14 20/30 2 70 —

55 M W Aph OAG 4 SPM 4 SPM 23 20/25 2 2 NLP 0 84 CE; PKP; IRC;

CME; phthisis

49 F B Uveitic ACG 1 DPM 23 Baerveldt 27 20/60 2 11 20/20 0 18 —

68 F W CACG 15 DPM 9 DPM 29 CF 4 0 LP 0 14 CE; phthisis

20 M W Aph OAG 2 SPM 84 DPM 37 20/25 1 18 20/20 0 20 —

60 M W UGH 3 SPM 25 DPM 19 20/60 1 12 20/80 1 68 CE; PKP; VH

63 M W CACG 5 DPM 13 Ahmed 42 20/60 3 25 20/60 1 22 CE; PKP

72 M W SACG, POAG 4 DPM 44 Ahmed 23 20/40 2 26 CF 3 26 CE

79 M W POAG 2 Ahmed 11 Ahmed 29 20/30 2 14 20/50 1 10 —

69 M W SACG 5 Ahmed 6 Baerveldt 33 20/50 3 18 NLP 1 10 CE; PKP

ACTSEB 5 anterior chamber tube shunt to an encircling band; Aph ACG 5 aphakic angle-closure glaucoma; Aph OAG 5 aphakic

open-angle glaucoma; B 5 black; CACG 5 chronic angle-closure glaucoma; CE 5 corneal edema; CF 5 counting fingers; Choroidal 5

choroidal detachment requiring surgical drainage; CME 5 cystoid macular edema; Cong glc 5 congenital glaucoma; DPM 5 double-plate

Molteno; Ghost cell 5 ghost cell glaucoma; HM 5 hand motion; IOP 5 intraocular pressure; IRC 5 iridocyclitis; LP 5 light perception; Meds 5

number of antiglaucoma medications; NLP 5 no light perception; NVG 5 neovascular glaucoma; No fix 5 no fixation; PKP 5 penetrating

keratoplasty; POAG 5 primary open-angle glaucoma; SACG 5 secondary angle-closure glaucoma; SPM 5 single-plate Molteno; UGH 5

uveitis-glaucoma-hyphema syndrome; Uveitic ACG 5 uveitic angle-closure glaucoma; VA 5 visual acuity; VH 5 vitreous hemorrhage; W5

white.

*Interval between primary tube shunt and revision or additional tube shunt implantation.†Last follow-up visit or time of failure.

SHUNT REVISION VS ADDITIONAL TUBE SHUNT IMPLANTATIONVOL. 129, NO. 4 457

the revision group; 40% for the additional tube group, P 5.012). This level of intraocular pressure control wasachieved with 1.3 6 0.6 antiglaucoma medications in therevision group and 1.0 6 0.4 in the additional tube group(P 5 .35).

As defined by our criteria, only five (42%) of 12 patientsin the revision group versus 13 (62%) of 21 patients in theadditional tube group achieved at least a qualified success(P 5 .30, Fisher exact test). In the revision group therewere four successes (34%), one qualified success (8%), andseven failures (58%). In the additional tube group therewere 10 successes (48%), three qualified successes (14%),and eight failures (38%). Kaplan-Meier survival analysis isshown in Figure 1. At 2 years, the probability of adequateintraocular pressure control was significantly higher in theadditional tube group (84% in the additional tube group;56% in the revision group, P 5 .034, Mantel-Haenszel logrank test). Thereafter, however, the differences betweenthe two survival curves did not reach statistical signifi-cance (overall, P 5 .167, Mantel-Haenszel log rank test).Median survival (intraocular pressure control) for therevision group was 28 months, whereas median survival forthe additional tube group was 84 months.

The single-plate Molteno implant as the additionalimplant had a lower success rate (zero of four) comparedwith the double-plate Molteno implant (nine of 11), theAhmed implant (three of four), and the Baerveldt implant(one of two).

At final follow-up, best-corrected Snellen visual acuityhad improved by at least 2 lines from preoperative visualacuity in two patients (17%) with revision and in threepatients (15%) with the additional tube. Six patients(50%) in the revision group and seven (33%) in theadditional tube group maintained visual acuity within 1

line of preoperative acuity. Four (33%) in the revisiongroup and 11 (52%) in the additional tube group experi-enced worsening of vision by at least 2 lines.

The rate of complications was high in both groups.Corneal edema was the most significant complication inboth groups. Five (42%) in the revision group and nine(43%) in the additional tube group had corneal edema(P 5 1.0), although several of the patients had edemabefore the surgery. In the revision group, four of the fivewith corneal edema had a prior history of edema, whereasonly one of the nine in the additional tube group had ahistory of corneal edema. The incidence of new cornealedema was higher in the additional tube group, but thedifference was not statistically significant (eight of 21 inthe additional tube group; one of 12 in the revision group,P 5 .107, Fisher exact test). Penetrating keratoplasty wasrequired in one patient after revision and in four patientsafter the additional tube shunt surgery. In the revisiongroup, two patients (17%) had vitreous hemorrhage thateventually cleared. In the additional tube group, vitreoushemorrhage that resolved occurred in three patients(14%), choroidal detachment requiring surgical drainagein one (5%), iridocyclitis in one (5%), cystoid macularedema in two (10%), and phthisis in two (10%).

DISCUSSION

FEW STUDIES HAVE EVALUATED TREATMENT OPTIONS FOR

patients who have failed tube shunt surgery. In our study,we retrospectively compared intraocular pressure controlin patients who underwent revision by excision of anencapsulated bleb over the seton implant and those whohad implantation of an additional tube shunt.

Failure of tube shunts is most commonly related to anexcessive fibrotic response.2,9,10 This seems to be related tothe severity of glaucoma and number of prior surgeries.9,11

We have found revision by excision of the thick fibrouscapsule to be technically difficult, and the fibrous capsuleoften reforms. Additional tube shunts are usually placed inareas without prior surgical procedures. This may make ittechnically easier, because there is usually less bleedingand scar tissue. Also, the tendency for excessive fibrosis inthe new surgical site may be somewhat decreased.

Capsule excision after failed Molteno surgery was exam-ined by Valimaki and associates.10 They reported a 52%success rate with revision alone (intraocular pressure lessthan 22 mm Hg), whereas 42% required a second tubeshunt for intraocular pressure control.

Our study showed that placing an additional tube shuntwas more effective than revision in controlling intraocularpressure. The additional tube shunt provided a lower finalintraocular pressure (P 5 .037) and a greater percentreduction in intraocular pressure (P 5 .012) than revision.The Kaplan-Meier survival analysis showed better intraoc-ular pressure control in the additional tube group at 2 years

FIGURE 1. Kaplan-Meier survival analysis of intraocular pres-sure control after revision and additional tube shunt implanta-tion. Criteria for success and failure are described in theMethods section. Error bars indicate standard error.

AMERICAN JOURNAL OF OPHTHALMOLOGY458 APRIL 2000

(P 5 .034). After 2 years, the survival difference betweenthe two groups was no longer statistically significant. Thismay be in part because of the fewer number of subjectsfollowed beyond 2 years.

Various drainage tube implants were used in our study.The single-plate Molteno implant, which has the smallestsurface area, was associated with more failures to controlintraocular pressure than the other larger implants whenused as the additional implant. However, the sample sizeswere too small to make definitive conclusions.

Visual acuity remained the same or improved frompreoperative visual acuity in eight (67%) of the revisiongroup patients and in 10 (48%) of the additional tubegroup patients. After primary Molteno tube shunt implan-tation, Mills and associates12 noted that 68% of patientsmaintained at least the same level of preoperative visualacuity. In some of our patients, worsening of vision may beattributed to corneal edema rather than solely to glau-coma.

The complication rate was high in both groups, withcorneal edema being the most significant and seriouscomplication. This is not unexpected given the history ofprior surgical procedures. Mills and associates12 reported an18% combined incidence of corneal decompensation andgraft failure after primary Molteno tube shunt implanta-tion. In our study, five patients (42%) in the revision groupand nine (43%) in the additional tube group experiencedcorneal edema. But the incidence of new corneal edemawas much greater after the additional tube shunt proce-dure. Only one patient underwent penetrating keratoplastyafter revision, whereas four patients required the procedureafter additional tube shunt implantation.

This study has several limitations The sample size ofthe two groups was small and unequal. Because the studywas retrospective, there may be selection bias and thusa nonrandom sample. Although the primary implantappeared to have an elevated bleb in the additional tubegroup, the preoperative intraocular pressure control wasno different from that of the revision group, in whichthe primary implant bleb appeared to be low or absent.In either group, the primary implant was not adequatelycontrolling pressure. Both groups were similar in othercriteria, but the distribution of diagnoses differed. Therevision group had more patients with neovascularglaucoma (six of 12) compared with the additional tubegroup (two of 21). Neovascular glaucoma tends to bevery difficult to control and has the poorest progno-sis.11–13 This may have contributed to the relativesuccess in the additional tube group. In the study byValimaki and associates,10 only one of 12 subjectsundergoing revision had neovascular glaucoma. Thestrength of our study is that it evaluated and comparedthe two treatment options.

In this study we considered revision and additional tubeshunt implantation. Other options include cyclodestruc-tive procedures, which are usually considered for patients

with end-stage glaucoma.14–18 Cyclocryotherapy as ameans of ciliary body ablation to lower intraocular pressurehas had variable success rates ranging from 29% to 76%and is associated with a high incidence of vision loss andphthisis bulbi.14,15 Noncontact transscleral Nd:YAG lasercyclophotocoagulation has success rates ranging from 45%to 72% but is also associated with a significant risk ofvision loss.16 Contact diode laser cyclophotocoagulationhas success rates ranging from 52% to 66%.17,18 Futurestudies comparing cyclodestructive procedures with shuntrevision or additional tube shunt implantation in eyes inwhich the initial shunt has failed may further define thebest treatment option in these difficult cases.

The patients in our study had already failed primary tubeshunt implants and hence had relatively poor prognoses forsubsequent surgical success. We compared the results ofrevision and additional tube shunt placement in theseeyes. Although there are limitations to our study, thegeneral trend showed better intraocular pressure controlafter additional tube shunt implantation. The most fre-quent complication was corneal edema. For most patientswith primary tube shunt failure, we think that an addi-tional tube shunt is a better choice than a revisionprocedure.

REFERENCES1. Shields MB. Drainage implant surgery. In: Shields MB,

editor. Textbook of glaucoma (4th ed). Baltimore: Williamsand Wilkins, 1998:538–546.

2. Melamed S, Fiore PM. Molteno implant surgery in refractoryglaucoma. Surv Ophthalmol 1990;34:441–448.

3. Molteno ACB. New implant for drainage in glaucoma:clinical trial. Br J Ophthalmol 1969;53:606–615.

4. Coleman AL, Hill R, Wilson MR, et al. Initial clinicalexperience with the Ahmed glaucoma valve implant. Am JOphthalmol 1995;120:23–31.

5. Schocket SS, Lakhanpal V, Richards RD. Anterior chambertube shunt to an encircling band in the treatment ofneovascular glaucoma. Ophthalmology 1982;89:1188–1194.

6. Lloyd MAE, Baerveldt G, Heuer DK, Minckler DS,Martone JF. Initial clinical experience with the Baerveldtimplant in complicated glaucomas. Ophthalmology 1994;110:640 – 650.

7. Krupin T, Ritch R, Camras CB, et al. A long Krupin-Denvervalve implant attached to a 180° scleral explant for glaucomasurgery. Ophthalmology 1988;95:1174–1180.

8. Minckler DS, Heuer DK, Hasty B, Baerveldt G, Cutting RC,Barlow WE. Clinical experience with the single-plate Mol-teno implant in complicated glaucomas. Ophthalmology1988;95:1181–1188.

9. Lavin MJ, Franks WA, Wormald RPL, Hitchings RA.Clinical risk factors for failure in glaucoma tube surgery.Arch Ophthalmol 1992;110:480–485.

10. Valimaki J, Tuulonen A, Airaksinen PJ. Capsule excisionafter failed Molteno surgery. Ophthalmic Surg Lasers 1997;28:382–386.

11. Spiegel D, Shrader RR, Wilson RP. Anterior chamber tubeshunt to an encircling band (Schocket procedure) in thetreatment of refractory glaucoma. Ophthalmic Surg 1992;23:804–807.

SHUNT REVISION VS ADDITIONAL TUBE SHUNT IMPLANTATIONVOL. 129, NO. 4 459

12. Mills RP, Reynolds A, Emond MJ, Barlow WE, Leen MM.Long-term survival of Molteno glaucoma drainage devices.Ophthalmology 1996;103:299–305.

13. Mermoud A, Salmon JF, Alexander P, Straker C, MurrayADN. Molteno tube implantation for neovascular glaucoma.Ophthalmology 1993;100:897–902.

14. Benson MT, Nelson ME. Cyclocryotherapy: a review of casesover a 10 year period. Br J Ophthalmol 1990;74:103–105.

15. Caprioli J, Strang S, Spaeth G, Poryzees E. Cyclocryotherapyin the treatment of advanced glaucoma. Ophthalmology1985;92:947–954.

16. Dickens CJ, Nguyen N, Mora JS, et al. Long-term results ofnoncontact transscleral neodymium:YAG cyclophotocoagu-lation. Ophthalmology 1995;102:1777–1781.

17. Kosoko O, Gaasterland DE, Pollack IP, Enger CL, the DiodeLaser Ciliary Ablation Study Group. Long-term outcome ofinitial ciliary ablation with contact diode laser transscleralcyclophotocoagulation for severe glaucoma. Ophthalmology1996;103:1294–1302.

18. Bloom PA, Tsai JC, Sharma K, et al. “Cyclodiode” trans-scleral diode laser cyclophotocoagulation in the treatment ofadvanced refractory glaucoma. Ophthalmology 1997;104:1508–1520.

The full-text of AJO is now available online at www.ajo.com. AuthorsInteractivet, currently available in limited form, is undergoing an upgrade.

AMERICAN JOURNAL OF OPHTHALMOLOGY460 APRIL 2000