Kaplan-Meier curve showing the probability that the intraocular pressure (IOP) will be controlled at 21 mm Hg or less after Molteno implant insertion. Each tick represents a censored observation, that is, the time of death or the last follow-up for a single eye.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Every SG, Molteno ACB, Bevin TH, Herbison P. Long-term Results of Molteno Implant Insertion in Cases of Neovascular Glaucoma. Arch Ophthalmol. 2006;124(3):355–360. doi:10.1001/archopht.124.3.355
To describe the long-term outcomes of cases of neovascular glaucoma drained by Molteno implants.
A prospective study of 145 eyes (130 patients) followed up for a mean of 3.3 years (range, 0.02 year [5 days] to 18.1 years) in the province of Otago, New Zealand, from 1979 to 2002.
Insertion of a Molteno implant controlled the intraocular pressure at 21 mm Hg or less with a probability (95% confidence interval) of 0.72 (0.64-0.80), 0.60 (0.51-0.69), and 0.40 (0.29-0.50) at 1, 2, and 5 years, respectively. Failure to control intraocular pressure at 1, 2, and 5 years was significantly correlated with persistent iris neovascularization (P<.001, P<.001, and P = .01, respectively). Visual acuity at final follow-up in nonenucleated eyes was maintained or improved in 56 eyes (39%) and deteriorated to light perception or better in 25 (17%) or no light perception in 47 (32%). Seventeen eyes (12%) were enucleated.
The insertion of Molteno implants for neovascular glaucoma maintained or improved vision in 39% of eyes, whereas 12% were eventually enucleated (all of which initially had visual acuity <20/1200). The outcome depended mainly on progression of the underlying vascular disease.
The first cases of neovascular glaucoma drained by long-tube Molteno implants were operated on between 1973 and 1976, before facilities for retinal photocoagulation were available. At that time, implants were used only as a last resort, when a trial of medical treatment had failed and the eyes remained painful with grossly elevated intraocular pressure (IOP). Although insertion of the implants produced a satisfactory reduction of IOP with regression of anterior segment new vessels, the visual results were poor because of the terminal nature of the disease. However, a single case in which the postoperative visual acuity improved to 20/40, falling gradually to 20/200 during the next 4 years, indicated that a good visual result was possible. Subsequently, the argon laser became available and was combined with the use of implants to provide more effective treatment for cases of neovascular glaucoma. Subsequent cases were treated at an earlier stage of the disease by combinations of photocoagulation and drainage via implants in a sequence that was determined by the clinical conditions in individual cases, with special regard for the elevation of IOP and the clarity (or otherwise) of the media. These cases were closely followed and 3 important aspects of their behavior were noted: (1) that relatively moderate elevation of IOP in a short period would cause profound visual loss1,2; (2) that prompt treatment combined with photocoagulation of the underlying retinal lesions did improve the visual outcome3,4; and (3) that patients whose neovascular glaucoma was well controlled by implants, with apparently complete regression of anterior segment new vessels, could relapse in response to general metabolic disease, which included the electrolyte disturbance of renal failure, severe uncontrolled hypertension, and poor diabetic control with ketoacidosis.2,5 This experience indicated quite clearly that the treatment of cases of neovascular glaucoma involved a multidisciplinary approach that should include the use of laser photocoagulation and draining implants for the local ocular disease combined with careful evaluation and management of the patient's general circulatory and metabolic state.1,3,6
We now report the long-term results of a prospective study of 145 consecutive operations for insertion of a Molteno implant in cases of neovascular glaucoma performed at Dunedin Hospital, Dunedin, New Zealand, from January 1, 1979, through October 31, 2002.
Cases of neovascular glaucoma with visual acuity of light perception or better and drained by Molteno implants that were inserted at Dunedin Hospital between 1979 and October 31, 2002, were identified from the computerized database of the Otago Glaucoma Surgery Outcome Study. Consent for data retrieval was obtained from the ethics committee of the Southern Regional Health Authority.
The initial surgical technique (from 1977) involved direct insertion of implants into eyes with uncontrollable IOP and marked rubeosis; less advanced cases were managed by a 2-stage procedure involving placement of the implant on the eye combined with medical control of the IOP and insertion of the tube into the anterior chamber at a later date. Subsequent improvements in surgical technique have included the introduction of dual-chamber Molteno implants that limit the escape of aqueous and reduce hypotonia in the early postoperative period when inserted directly. Dual-chamber implants were used routinely from 1985 onward.4,7 The 2-stage procedure for less acute cases was replaced by a single operation in 1983 in which the drainage tube was occluded by an absorbable polyglycolic acid (Vicryl) suture7 in cases where iris neovascularization developed slowly and the IOP could be temporarily controlled medically. Where this technique was possible, postoperative anterior chamber shallowing did not occur. Immediate drainage of aqueous was used in more acute cases. In the group of cases reported herein, insertion of a standard Molteno implant was used until 1985, after which the pressure ridge implant was used in all cases. There was a statistically significant (P<.01) difference between those, although it did not avoid shallowing of the anterior chamber.
Postoperative management consisted of the administration of topical corticosteroids and cycloplegic agents for several weeks until the eye was white and quiet. Thirty-seven cases with the potential for useful vision were treated as outpatients by means of the systemic administration of oral prednisone, a nonsteroidal anti-inflammatory agent, and colchicine for 5 to 6 weeks to clear the media rapidly and to produce a thinner, more permeable bleb capsule.8-10
Laser photocoagulation was not available for the first 4 years of this study. After 1982, postoperative retinal photocoagulation was carried out as soon as possible after the operation in 44 eyes in which the media were sufficiently clear.
Eyes with potentially useful vision were treated by administration of hypotensive medication if the postoperative IOP was greater than 21 mm Hg. Hypotensive medications included topical β-blockers or epinephrine bitartrate if the β-blockers were contraindicated, and were supplemented if necessary by small doses of acetazolamide or, more recently, topical carbonic anhydrase inhibitors. Miotics (eg, pilocarpine sulfate) were contraindicated because they tended to dilate iris new vessels and sometimes raise the IOP. We have avoided the use of prostaglandin analogues, which can cause vasodilation and intraocular inflammation. Painless blind eyes did not receive hypotensive medication.
Cases in which progression of underlying vascular disease or deterioration in the patient's general state produced recurrence of anterior segment neovascularization and discomfort were treated by means of additional retinal photocoagulation where possible, supplemented if necessary by topical atropine sulfate and corticosteroids.
The preoperative IOP for each eye was taken as the mean IOP in the month before surgery. Postoperatively, IOP was taken as the mean for each postoperative year. The medications used were recorded at each postoperative visit.
Success was defined as an IOP of greater than 6 mm Hg and less than or equal to 21 mm Hg, with or without hypotensive medication, and maintenance of a visual acuity of light perception or better. Failure was defined as a persistent IOP of greater than 21 mm Hg, deterioration of visual acuity to no light perception, phthisis, or enucleation. The time of failure was taken as the first date at which any of these events was recorded.
The preoperative visual acuity for each patient was taken as the best-corrected visual acuity recorded in the month before surgery. The postoperative visual acuity was taken as the best visual acuity for each postoperative year.
Survival methods were used to analyze the probability of implant failure. In this analysis, censoring was at the last follow-up visit for patients who had not experienced an event. The survival curve was calculated using all 145 eyes and estimated the survival probability of a randomly selected eye from 130 patients. Where the time between postoperative visits was more than 1 year (which occurred in 24 eyes), the intervening values were interpolated. The survival curve based only on the first eye treated did not differ from that based on all eyes by more than 4% at any time point. Confidence intervals and P values were based on data for the first eye treated in each patient, which eliminated the effect of the correlation introduced by those patients who underwent bilateral operations. P values were based on the log-rank statistic and confidence intervals were calculated using the Greenwood formula.11
Molteno implants were inserted into 145 eyes (130 patients) with neovascular glaucoma and light perception or better at Dunedin Hospital. There were 15 bilateral cases (12%) and an additional 2 cases with a previous diagnosis of neovascular glaucoma in the fellow eye not treated by insertion of a Molteno implant. The mean age at operation was 68.8 years (range, 19 to 95 years). Postoperative follow-up was for a mean of 3.3 years (range, 0.02 year [5 days] to 18.1 years). Fifteen cases (14 patients) were lost to follow-up after a mean follow-up of 3.0 years (range, 0.2 year [1.5 months] to 9.3 years), and 101 patients (113 eyes) died during follow-up. Patient characteristics, the reason for implant insertion, and previous ocular surgery are given in Table 1. Operative techniques are given in Table 2.
The mean (SD) preoperative IOP was 40.1 (13.0) mm Hg. This fell to 20.0 (9.1) mm Hg at 1 year after surgery and 19.3 (9.9) mm Hg at 2 years, and increased to 21.9 (11.0) mm Hg and 26.8 (10.7) mm Hg at 5 and 10 years, respectively. Insertion of a Molteno implant controlled the IOP at 21 mm Hg or less with a probability (95% confidence interval) of 0.72 (0.64-0.80) at 1 year, 0.60 (0.51-0.69) at 2 years, and 0.40 (0.29-0.50) at 5 years (Figure).
Of the 130 cases where full data were available, 51 (39%) had undergone preoperative panretinal photocoagulation. Eighteen of these 51 cases underwent additional photocoagulation during the early postoperative period. In addition, 26 cases underwent panretinal photocoagulation for the first time after the operation (range, 106-3433 burns; mean, 1189 burns). There was no significant difference in outcome between those who underwent retinal ablation preoperatively or postoperatively and those who did not (P = .40).
Failure to control IOP at 1, 2, and 5 years was significantly correlated with persistent iris neovascularization (P<.001, P<.001, and P = .01, respectively). There was no significant difference in outcome between those who were older or younger than the median age at the time of operation and between those with or without diabetes mellitus or vein occlusion.
The mean number of hypotensive medications being administered preoperatively was 1.27. This number fell to 0.80, 0.73, 0.78, and 0.52 at 1, 2, 5, and 10 years, respectively, after surgery.
The mean preoperative visual acuity was 20/600 (Table 1). Visual acuity at final follow-up was maintained or improved in 56 eyes (39%) and deteriorated to light perception or better in 25 (17%) or no light perception in 47 (32%) (including 14 phthisical eyes). Seventeen eyes (12%) were enucleated, all of which initially had a visual acuity of less than 20/1200. Of the 42 eyes with light perception only, 5 (12%) had improved vision at the final follow-up, with 1 reaching 20/200. The visual acuity at final follow-up (before death or at the most recent follow-up) is given for 128 eyes in Table 3, and visual acuity at 1, 2, 5, and 10 years is given in Table 4.
Visual acuity in the fellow eye was known for 118 eyes (81%), of which 26 (22%) preoperatively and 37 (31%) at final follow-up had visual acuity of hand motions or less.
There were no recorded operative complications in this series. Early postoperative complications and surgical interventions are described in Table 3 and included 2 cases of lenticulocorneal contact requiring surgical reformation of the anterior chamber. Four choroidal detachments (3%) were recorded. Severe iritis (marked anterior chamber inflammation lasting >1 week) occurred in 1 case, and transient hyphema was common (94 eyes [65%]). There were 2 cases of Molteno implant exposure: in one case the single plate implant was removed; in the other case the second plate of a dual-chamber implant was removed, following which adequate drainage has been maintained by the remaining plate. Tube blockage by the iris occurred in 9 cases (6%) and was managed by laser iridoplasty in 8 cases and surgical repositioning of the tube in 1. The most common complication was transient hyphema, which occurred in 94 eyes (65%) and was not affected by surgical technique. Significant shallowing of the anterior chamber occurred in 45 eyes (31%) and did depend on surgical technique. Only 1 (9%) of 11 two-stage and 2 (25%) of the 8 with a tie on the drainage tube developed shallow anterior chambers (3 [16%] of 19 cases) compared with 42 (33%) of the remaining 126 cases. This difference was not statistically significant (P>.05).
However, there was a significant reduction in the incidence of shallowing of the anterior chamber after the introduction of the dual-chamber implant. A comparison between the 126 cases treated with the use of immediate drainage before and after the introduction of dual-chamber implants in 1986 found shallow anterior chambers in 15 (60%) of 25 cases before 1986 compared with 27 (27%) of 101 cases after 1986. This difference was highly significant (P<.01).
Detailed information was available on 127 cases to establish that in 25 cases (20%) the cornea decompensated. At the time of decompensation, 9 eyes had no light perception, whereas the remaining eyes had visual acuities ranging from light perception to 20/1200.
Later intraocular procedures (>3 months postoperatively) in eyes with visual potential included cataract extraction with or without an intraocular lens in 7 eyes, combined lensectomy and vitrectomy in 6, simple vitrectomy in 11, and retinal detachment repair in 1 (Table 3).
This study reports the results of insertion of a Molteno implant in 145 eyes with neovascular glaucoma. All patients were followed up prospectively, which reduced uncertainty resulting from selective attendance and incomplete follow-up. The frail physical and mental state with a significant degree of dementia in many cases resulted in our being unable to maintain contact with 15 patients until death. However, the mean follow-up of these 15 was only slightly less than that of the total group (3.0 vs 3.3 years).
The methods used to measure IOP and visual acuity and the criteria for blindness did not change during the study. Measurement of visual field was unreliable owing to reduced central visual acuity and the inclusion of patients with dementia, and was therefore not reported.
In this series, 27 of the 34 eyes classified as treatment failures after 1 year did not receive hypotensive medication because they had an IOP of less than 30 mm Hg in the operated-on eye and good vision in the fellow eye.
Long-term control of IOP was associated with full regression of anterior segment new vessels, while persistence or reappearance of new vessels was associated with an unstable IOP, which could be raised, normal, or hypotonous.
Preservation of useful vision was more likely in patients who sought care early (<2 weeks from the onset of symptoms) and had relatively good vision (20/160 or better). Favorable signs were prompt regression of iris new vessels, rapid clearing of operative hyphemas, and media that allowed adequate photocoagulation of the underlying retinal lesions. This small group of cases had a relatively good life expectancy, losing only 21% or (2.43 years) of expected life, which suggests that these cases had less severe and less widespread vascular disease than patients with poorer vision, who lost 62% (10.7 years) of expected remaining life.12
Direct comparison of our results with those of other studies is complicated by variable case selection, criteria for success, methods of data analysis, and duration of follow-up.
Cases of neovascular glaucoma have a spectrum of severity ranging from “relatively favorable” cases that are seen early in the course of the disease to cases that are seen late and have severely elevated IOP, opaque media, and minimal vision.12,13 Some published studies14-16 reporting results of treating neovascular glaucoma come from specialized glaucoma centers that receive significant numbers of secondary and tertiary referrals. Such studies may not reflect the full spectrum of disease from a defined geographic area such as those treated at Dunedin Hospital.
Published studies6,13-19 describing the use of artificial draining implants in cases of neovascular glaucoma report success rates varying from 22% to 67%. These results are influenced by differing criteria for success and variable follow-up periods.
The largest series of 60 cases with long-term results of Molteno implant insertion for cases of neovascular glaucoma is that of Mermoud et al13 with a mean follow-up of 24.7 months, compared with 145 cases in the present series with a mean follow-up of 40 months. Because Mermoud et al used the same criteria for success as the present study, it is helpful to compare their results with ours. The mean preoperative and postoperative IOPs were similar in both series. Kaplan-Meier analysis of IOP and preserved vision in the study by Mermoud et al shows the chance of success as 62% at 1 year, 53% at 2 years, and 10.3% at 5 years (1 case), compared with 72% at 1 year, 60% at 2 years, and 40% at 5 years (21 cases) in the present study. Visual acuity in the study by Mermoud et al was the same or improved in 23 eyes (38% of cases) and worse in 37 (62%), with loss of light perception in 29 (48%),13 compared with the same or improved vision in 56 eyes (39%), worse vision (but retaining light perception) in 25 (17%), and loss of light perception in 47 (32%) in the present series at final follow-up. The results of our study are similar to those reported by Mermoud et al13 and others15,16 for the first 2 years, with a steep fall-off in success. However, longer-term follow-up showed a much slower rate of failure, which we believe resulted from selective survival of patients with less severe and less widespread underlying vascular disease.12
Previous published reports2,3,20 emphasize the role of photocoagulation in improving the results when treating cases of neovascular glaucoma. This is clearly the situation in more favorable cases when the media are sufficiently clear. However, both the present study and that of Mermoud et al13 found no statistical difference between eyes that did and did not undergo photocoagulation. We believe that this was probably owing to pressure-induced retinal atrophy that prevented the ongoing formation of vasoformative factors in the more severe cases where the opaque media prevented photocoagulation.
Although Mermoud et al13 found improved outcomes in diabetic patients compared with those with central retinal vein occlusion, we found no significant difference in outcome between these 2 groups (data not shown), which concurs with other studies.14,16
Previous studies1,12-14 have noted the high mortality of cases of neovascular glaucoma. A preliminary study from the Otago Glaucoma Surgery Outcome Study,1 published in 1983, of 39 cases of neovascular glaucoma drained by Molteno implants found a mean survival of only 2.6 years compared with 6.5 years in the recent analysis of neovascular cases in the Otago Glaucoma Surgery Outcome Study.12 This difference is probably the result of major improvements in the general medical management of diabetes mellitus and vascular disease that have occurred in the Otago region in the past 20 years.
Our findings support those of most previous studies and indicate that early diagnosis and prompt treatment of neovascular glaucoma by the use of drainage, photocoagulation, and general medical measures provides useful long-term visual results in 39% of cases. The mean life expectancy of cases of neovascular glaucoma treated by insertion of Molteno implants has improved from 2.6 years in 1983 to 6.5 years in 2004.
Correspondence: Anthony C. B. Molteno, FRCS, Section of Ophthalmology, Department of Medical and Surgical Sciences, University of Otago Dunedin School of Medicine, PO Box 913, Dunedin, New Zealand (email@example.com).
Submitted for Publication: November 29, 2004; final revision received March 23, 2005; accepted April 29, 2005.
Author Contribution: Dr Every had full access to all data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: Dr Molteno has a financial interest in Molteno implants.
Funding/Support: Ms Bevin was supported by the Healthcare Otago Charitable Trust, Dunedin, during this study.
Additional Information: The Healthcare Otago Charitable Trust had no role or influence in any part of this study.