In glaucoma drainage implant (GDI) surgery, aqueous humor is drained and excreted from anterior chamber to orbital tissue through a tube. The original method designed by Molteno in 1969 has been modified, and currently, the Molteno implant, Baerveldt tube, and Ahmed valve are mainly used. There have been many reports on long-term (exceeding 3 years) outcomes of treatment using the nonvalved type of GDI in other countries. However, to the best of our knowledge, there has been no report on long-term results of GDI in WAsian eyes exceeding 3 years. In this study, we analyzed survival, prognostic factors contributing to the surgical outcomes, and postoperative complications in patients who underwent nonvalved GDI surgery and GDI combined with trabeculectomy at our hospital, and compared the outcomes in Japan with those reported in other countries.
Subjects and Methods
The treatment course was retrospectively investigated in 124 eyes of 109 consecutive Japanese patients who underwent nonvalved GDI surgery and GDI combined with trabeculectomy in the Ophthalmology Department, Japanese Red Cross Medical Center, between June 1993 and June 2009 and could be followed for 1 year or longer after surgery. Defining states meeting one of the following conditions as death of the eye, survival analysis was performed employing the Kaplan-Meier method and Cox proportional hazards model: (1) The intraocular pressure (IOP) is below 5 or above 22 mmHg, or shows <20% reduction from the preoperative IOP, (2) necessity of oral acetazolamide mediation and/or additional glaucoma surgery, and (3) loss of light perception in visual acuity. Regarding predictors contributing to the surgical outcomes, the age, gender, past medical history of surgery, disease type, gonioscopic findings (the presence or absence of peripheral anterior synechia [PAS]), preoperative IOP, implant type (single/double-plate Molteno, Baerveldt 350), and combined trabeculectomy were subjected to proportional hazards analysis and the hazard ratios (HR) were calculated. Regarding the presence or absence of PAS, PAS indices ≥50% and ≤50% were regarded as presence and absence, respectively, and unclear gonioscopic findings due to a preoperatively poor corneal condition was defined as unclear. Since surgeries of the bilateral eyes and multiple surgeries of the same eye were included in the data, the generalized estimating equation model was employed for analysis. This study was performed after approval by the Clinical Study Ethics Committee of the Japanese Red Cross Medical Center.
- The plate was to be placed at the inferior temporal site as a rule [Fig. 1a and b]. A fornix-based conjunctival flap was prepared for about half of the circumference, and the episcleral tissue was dissected under sub-Tenon anesthesia
- Bridle suture was applied to the two rectus muscles adjacent to the surgical field when Baerveldt 350 and single-plate Molteno were used, and to the three rectus muscles when double-plate Molteno was used
- The plate was fixed to the sclera with 8-0 nylon at a site 8-10 mm from the limbus
- The tip of the tube was ligated with 7-0 nylon by 3-1-1 through [Fig 2] and inserted into anterior chamber. To insert the tube into the anterior chamber, the anterior chamber was perforated about 1.5 mm from the limbs with a 22G needle or 22G V-lance at an angle, and the tube was inserted through this hole using a viscoelastic material to make the tube penetrate at a site posterior to Schwalbe's line. In patients with PAS in the whole circumference due to neovascular glaucoma (NVG) or patients following penetrating keratoplasty, the tube was inserted through the ciliary sulcus [Fig. 3]
- An about 5-6 mm portion of the tube on the limbal side was covered with the donor sclera or an autologous scleral patch
- The conjunctiva was sutured with 8-0 polyglactin, and the viscoelastic material was removed by aspiration irrigation in cases it was used.
In eyes in which the preoperative IOP exceeded 40 mmHg despite being under max treatment and the visual field condition was severer than in the middle stage, trabeculectomy was additionally applied at the side of the tube to prevent intraocular hypertension immediately after surgery [Fig. 1a and b]. Regarding NVG due to central retinal vein occlusion (CRVO) and diabetic retinopathy, GDI surgery was performed after sufficient pan-retinal photocoagulation (PRP). However, in eyes with insufficient PRP in the periphery because of poor mydriasis, retino-cryopexy was applied to 2-3 rows of four circumference in the nearest peripheral retina during GDI surgery in all cases. All surgeries were performed by a single operator (TH).
After surgery, dexamethasone and antibiotic eye drops were administered for about 1 month, and glaucoma eye drops or oral acetazolamide was added when the IOP exceeded 21 mmHg. When the IOP was higher than 21 mmHg after 4 weeks, the suture was released to open the tube using an argon laser in the single-step method [Fig. 3, inset]. When the pressure exceeded 21 mmHg before 4 weeks in cases simultaneously treated with trabeculectomy, laser suture lysis of the trabeculectomy flap was performed. From May 2000, anterior segment optical coherence tomography (AS-OCT) was employed in all eyes to see the tube-inserted site and all eyes with hypotony to detect cilio-choroidal detachment (CCD).
Six patients (6 eyes) who we were unable to follow at our hospital were excluded within 1 year after surgery. The number of the subjects, disease type and GDI type are listed in Table 1. Trabeculectomy + MMC were combined in 5 eyes treated with single-plate Molteno, 43 with double-plate Molteno, and 3 with Baerveldt. The mean duration of follow-up was 3.5 ± 0.7 years (range: 1-16 years), and the mean preoperative IOP was 34.9 ± 19.1 mmHg. On survival rate analysis [Fig. 4], the 1-, 3-, 5-, and 10-year survival rates were 86.1, 71.1, 71.1, and 64.6%, respectively. The cause of eye death [Table 2] was related to the IOP (below 5 mmHg, above 22 mmHg, or < 20% reduction from the preoperative IOP) in 24 eyes, which was the most frequent cause, followed by NVG in 9 (diabetic retinopathy [DMR]: 6, CRVO: 2, familiar exudative vitreo-retinopathy: 1), POAG in 5, SOAG in 8 (lens-induced glaucoma: 2, uveitis, corticosteroid glaucoma, postcorneal transplant surgery, postvitrectomy, others: 1 each), and SACG in 2. Additional glaucoma surgery was necessary in 3 eyes: 2 with NVG (all were DMR) and 1 with POAG. The visual acuity showed lost light perception in 7 eyes: 6 with NVG (DMR: 3, CRVO: 3) and 1 with uveitis-induced SOAG, with a total of 34 eyes 6 (27.4%). Regarding prognostic factors [Table 3] contributing to the surgical outcomes, the HR was high for the age, gender, past medical history of surgery, disease type, PAS finding, preoperative IOP, and combined trabeculectomy, but no significant difference was noted. However, regarding the GDI type, only the HR of Single-plate Molteno was significantly higher than that of double-plate [Table 3].
Postoperative complications were classified into those which occurred within 1 month and more than 1 month after surgery as early and late complications, respectively. Early complications were a shallow anterior chamber in 13 eyes (10.5%), choroidal detachment in 10 (8.1%), anterior chamber hemorrhage in 9 (7.3%), conjunctival separation in 4 (3.2%), and an unfavorable tube position in 1 (0.8%), giving a total of 37 eyes (29.8%). The late complications were bullous keratopathy in 7 eyes (5.6%), an unfavorable tube position in 8 (6.5%), tube erosion in 7 (5.6%), phthisis in 2 (1.6%), persistent hypotony lower than 6 mmHg in 1 (0.8%), and endophthalmitis in 1 (0.8%), giving a total of 26 eyes (21.0%). The tube was re-inserted in all cases of an unfavorable tube position. Excluding 1 eye in which the tube was exposed because of rejection against donor scleral patch, all these were autologous scleral patch-applied cases. The IOP immediately after suture lysis of the tube transiently decreased to 6 mmHg in all cases and exceeded 10 mmHg within 4 weeks in most cases. A 6-9 mmHg IOP after laser suture lysis persisted for more than 6 months in 3 eyes (1 with NVG and 2 with POAG). Spontaneous remission was achieved after 14 months in 1 eye but the 6-9 mmHg pressure persisted for more than 5 years in the other 2 eyes. In one of these, macular edema and the appearance [Fig. 5a and b] and disappearance [Fig. 5c and d] of CCD, which can be observed only by AS-OCT, repeated. The tube has not been opened because of favorable IOP control after surgery in 3 cases (7.5, 2.1, and 2.0 years after surgery, respectively), in which 2 eyes were received combined trabeculectomy.
The corneal endothelial cell count decreased to <1,000/mm2 in 14 eyes, and bullous keratopathy developed in 7 eyes. The period between the GDI surgery and bullous keratopathy ranged between 1 and 9 years. Of the 7 eyes of bullous keratopathy, the preoperative endothelial cell count was <1000/mm2 in 2 eyes and bullous keratopathy developed in both cases. In the other 5 eyes, the preoperative count was between 1000 and 1500/mm2 in 3 and more than 1500/mm2 in 2 [Table 4]. In the former, 2 eyes were with Posner-Schlossman syndrome. One was with primary open angle glaucoma, but underwent 3 trabeculectomy surgeries before GDI surgery, and the corneal endothelial cell count was 1009/mm2. In the latter, re-insertion of IOL was performed for dislocation of the ciliary sulcus-fixed lens in one. In the other, the plate capsule was removed due to an elevated IOP and the tube touched the cornea because of shallow anterior chamber.
Based on the viewpoint of refractory glaucoma, generally, GDI surgery is judged as successful when the IOP decreases to 21 mmHg or lower and a more than 20% reduction from the preoperative IOP is achieved. In the literature, the success rate of nonvalved type GDI surgery was 52-100% at 3-10 years of follow-up. Cataract extraction and Molteno implant insertion controlled the IOP at 21 mmHg or less with a probability of 1.00 (95% CI: 0.93-1.00) at 10 years or more after operation in eyes with cataract and POAG. However, in the eyes, which included not only POAG but also secondary glaucoma, the success rates are 52% to 70.2%. The success rate in the tube versus trabeculectomy study was the highest (70.2%), but this fact may be due to the exclusion of intractable cases, such as active rubeosis, iridocorneal endothelial (ICE) syndrome, aphakia, chronic or recurrent uveitis and eyes which exceeded more than 40 mmHg. Accordingly, the success rate in our study (71.1%) was relatively high compared to those at 5 years after GDI surgery in other studies, and this may have been related to the following two conditions: (1) Ischemia was thoroughly resolved by additionally applying PRP before surgery and retinal cryopexy during surgery in eyes with NVG, which is considered to be the most intractable. Previous studies clarified that the outcome of surgery for NVG was poorer than those of other diseases, but no significant difference was noted in prognostic factors among the diseases in our patients, suggesting that treatment of NVG was successful. (2) For cases with a preoperative IOP exceeding 40 mmHg as well as middle stage or severer visual field impairment, trabeculectomy was concomitantly applied, which may have resolved the early ocular hypertension phase as well as contributed to long-term IOP reduction, because combined trabeculectomy showed high HR in the simple analysis of prognostic factor.
The plate area may also be involved as a factor influencing the success rate of surgery. Heuer reported that the 2-year survival rates after single- and double-plate Molteno implant surgeries were 46% and 71%, respectively, suggesting that this was due to the 2-times wider plate area of the double-plate (268 mm2) than that (143 mm2) of the single-plate Molteno implant. The recently reported ABC study also showed that the IOP reduction rate after 1 year was 2.2 mmHg lower in cases treated with Baerveldt than in those treated with Ahmed, and this was considered to be due to the wider plate area (Ahmed: 184 mm2, Baerveldt: 350 mm2). In our patients, the HR of the single-plate was significantly higher than that of the double-plate in all three types of nonvalved GDI, confirming that the plate area markedly influenced the success rate.
We have employed the nonabsorbable suture method from the beginning of the introduction of GDI. In the original method reported in 1989, the tube tip was ligated with nonabsorbable 7-0 polypropylene and opened later by argon laser suture lysis to prevent postoperative hypotony after surgery using a Molteno implant. A disadvantage of the method involving ligating the tube tip in the anterior chamber is the necessity of widening the hole in the sclera using a 22G needle instead of using 23G or V-lance when the tube is inserted into the anterior chamber because the region of the tube is thickened by ligation. Recently we have improved tube ligation from 3-1 to 1-3-1 through using 7-0 nylon, which enabled much easier insertion of the tube.
The cause of late intraocular hypotension persisting for a prolonged period after suture lysis is overfiltration. Stein et al. closely described the treatment of postoperative hypotony, in which revision surgery was necessary due to overfiltration at 140 days after surgery on average in 1.6% of all cases. In our patients, 6 mmHg or lower late intraocular hypotension was observed in only 1 eye (0.8%). We performed intentional laser suture lysis only in cases with an IOP higher than 21 mmHg after 4 weeks. The tube was left closed in three patients because of favorable postoperative IOP control. The use of absorbable thread for tube ligation in these cases may lead to a high risk of hypotony after suture lysis of the tube. After suture lysis, 6-9 mmHg intraocular hypotension was observed for more than 6 months in 3 eyes and naturally remitted in 1 eye at 14 months. CCD may occur under this physiological intraocular hypotension, but, generally, it is not detectable by funduscopy, and detected only by AS-OCT [Fig. 5]. In these cases, hypotensive maculopathy may occur, and re-ligation of the tube should be considered when it persists for a prolonged period.
In our patients, the incidence of bullous keratopathy was 5.6%, which is higher than those (about 1-2%) in other reports. There may be two reasons for this. The first is that bullous keratopathy occurred in longer observation study. Mills and others reported corneal decompensation occurred in 10% after 44 months (range: 6-107 months). In our study, the median period between the surgery and bullous keratopathy was 4.7 years (range: 1-9 years). The bullous keratopathy occurred more than 5 years after the surgery in 4 eyes out of 7 eyes [Table 4]. In addition to this, corneal endothelial loss is likely to occur as the course is followed for a prolonged period. The second is that there were 4 eyes with ICE syndrome, Posner-Schlossman syndrome and Cytomegalovirus infection [Table 4], which are believed to cause endothelial damage. On the other hand, it is obvious that corneal endothelial loss from normal to <1000/mm2 within a few years may be due to tube-corneal touch during GDI surgery or in the early postop period [Table 4]. Additional surgery after GDI also may have caused tube-corneal touch [Table 4]. In the eyes which had no reason for corneal endothelial loss as mentioned above, the angle of the tube against the cornea or tube inserting position may affect corneal endothelial loss in the long period after the surgery.
The outcomes of nonvalved GDI and GDI combined with trabeculectomy surgery in the long-term (3.5 ± 0.7 years, range: 1-16 years) are comparable or superior to those in other races. Hypotony and corneal endothelial damage are the most serious complications, which threaten visual acuity. Further investigation may be needed as for how to prevent corneal endothelial loss and hypotony after the release of tube ligation.
1. Molteno AC. New implant for drainage in glaucoma. Animal trial Br J Ophthalmol. 1969;53:161–8
2. Mills RP, Reynolds A, Emond MJ, Barlow WE, Leen MM. Long-term survival of Molteno glaucoma drainage devices Ophthalmology. 1996;103:299–305
3. Broadway DC, Iester M, Schulzer M, Douglas GR. Survival analysis for success of Molteno tube implants Br J Ophthalmol. 2001;85:689–95
4. Roy S, Ravinet E, Mermoud A. Baerveldt implant in refractory glaucoma: long-term results
and factors influencing outcome Int Ophthalmol. 2001;24:93–100
5. Gedde SJ, Schiffman JC, Feuer WJ, Herndon LW, Brandt JD, Budenz DL, et al Treatment outcomes in the tube versus trabeculectomy (TVT) study after five years of follow-up Am J Ophthalmol. 2012;153:789–803.e2
6. Molteno AC, Whittaker KW, Bevin TH, Herbison P. Otago Glaucoma Surgery Outcome Study: Long term results of cataract extraction combined with Molteno implant insertion or trabeculectomy in primary glaucoma Br J Ophthalmol. 2004;88:32–5
7. Hill RA, Nguyen QH, Baerveldt G, Forster DJ, Minckler DS, Rao N, et al Trabeculectomy and Molteno implantation for glaucomas associated with uveitis Ophthalmology. 1993;100:903–8
8. Budenz DL, Scott IU, Nguyen QH, Feuer W, Singh K, Nicolela MT, et al Combined Baerveldt glaucoma drainage implant and trabeculectomy with mitomycin C for refractory glaucoma J Glaucoma. 2002;11:439–45
9. Hamard P, Loison-Dayma K, Kopel J, Hamard H, Baudouin C. Molteno implant and refractory glaucoma. Evaluation of postoperative IOP control and complications with a modified surgical procedure J Fr Ophtalmol. 2003;26:15–23
10. Heuer DK, Lloyd MA, Abrams DA, Baerveldt G, Minckler DS, Lee MB, Martone JF. Which is better? One or two? A randomized clinical trial of single-plate versus double-plate Molteno implantation for glaucomas in aphakia and pseudophakia Ophthalmology. 1992;99:1512–9
11. Budenz DL, Barton K, Feuer WJ, Schiffman J, Costa VP, Godfrey DG, et al Treatment outcomes in the Ahmed Baerveldt comparison study after 1 year of follow-up Ophthalmology. 2011;118:443–52
12. Price FW Jr, Whitson WE. Polypropylene ligatures as a means of controlling intraocular pressure with Molteno implants Ophthalmic Surg. 1989;20:781–3
13. Stein JD, McCoy AN, Asrani S, Herndon LW, Lee PP, McKinnon SJ, et al Surgical management of hypotony owing to overfiltration in eyes receiving glaucoma drainage devices J Glaucoma. 2009;18:638–41
14. Nguyen QH, Budenz DL, Parrish RK 2nd. Complications of Baerveldt glaucoma drainage implants Arch Ophthalmol. 1998;116:571–5
15. Lee EK, Yun YJ, Lee JE, Yim JH, Kim CS. Changes in corneal endothelial cells after Ahmed glaucoma valve implantation: 2-year follow-up Am J Ophthalmol. 2009;148:361–7
Source of Support: Nil.
Conflict of Interest: None declared.