You are here

Article Text

Article menu

Download PDFPDF

Clinical science
Failed glaucoma drainage implant: long-term outcomes of a second glaucoma drainage device versus cyclophotocoagulation
  1. Jamie Lea Schaefer1,2,
  2. Monica A Levine1,
  3. Gina Martorana1,
  4. Helen Koenigsman1,
  5. M Fran Smith1,
  6. Mark B Sherwood1
  1. 1Department of Ophthalmology, University of Florida, Gainesville, Florida, USA
  2. 2Department of Ophthalmology, University at Buffalo, Buffalo, New York, USA
  1. Correspondence to Dr Mark B Sherwood, Department of Ophthalmology, University of Florida, 2000 SW Archer Rd, Gainesville, FL 32608, USA; Sherwood{at}ufl.edu

Abstract

Background/aims To evaluate long-term efficacy of a second glaucoma drainage device (GDD) versus cyclophotocoagulation (CPC) after failure of primary drainage implant.

Methods This is a non-randomised, retrospective cohort study. A chart review was conducted of patients who underwent GDD surgery between July 1986 and November 2012 requiring further glaucoma procedures for intraocular pressure (IOP) control. An additional GDD was placed in 15 eyes, while 32 eyes underwent CPC. The main outcome measurement was IOP control and/or time to failure of secondary intervention (IOP >18 mm Hg on two sequential measurements).

Results Mean follow-up after the second procedure was 63±65.8 months (range 6–254 months) in the CPC group and 132±91.8 months (range 12–254 months) in the GDD group. Thirty-four per cent (11/32 eyes) undergoing CPC later required further treatment at a mean of 13.6±10.7 months with 10/11(91%) of additional interventions occurring within 2 years. Despite an initially high success rate for IOP control in the first 5 years, eventually 60% (9/15 eyes) that underwent a second tube required additional treatment at a mean of 73.4 months with only 2/9(22%) requiring this within the first 2 years. The risk of visual acuity worsening by 2 Snellen lines or more at 12 months was 5/14 for the GDD group (36%) and 4/23(17%) for the CPC group.

Conclusions After failure of an initial drainage implant to control IOP, a sequential tube had a high initial rate of success but a relatively high likelihood of long-term failure, generally after 6 years. Eyes that received CPC tended to fail earlier, often within the first year, but had relatively few late failures.

  • Glaucoma
  • Treatment Lasers
  • Treatment Surgery
  • Intraocular pressure

https://doi.org/10.1136/bjophthalmol-2015-306725

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Drainage implant surgery is becoming an increasingly popular treatment for glaucoma that is uncontrolled with medications alone.1–4 The Tube versus Trabeculectomy (TVT) study1 showed similar intraocular pressure (IOP) results at 5 years in a trial that prospectively compared Baerveldt 350 mm2 glaucoma drainage devices (GDDs) with trabeculectomy with mitomycin-C (MMC) in patients with glaucoma with uncontrolled IOPs following either failed trabeculectomy or prior cataract surgery. The most common cause of medium-term to long-term GDD failure is inadequate IOP control despite the re-introduction of medication, due to encapsulation of the fibrous capsule surrounding the episcleral plate.1

    The long-term incidence of this complication is not well documented in the literature. In the prospective TVT study at 5 years, 13/107 of the eyes that received a Baerveldt 350 mm2 implant were recorded as having inadequate IOP control, with seven (7%) requiring additional procedures. We presented an earlier study as a poster at ARVO (Association for Research in Vision and Ophthalmology) that retrospectively analysed the results of 104 consecutive patients that underwent either double-plate Molteno or Baerveldt 350 mm2 glaucoma drainage implant surgery at the University of Florida with a single surgeon (MBS) between July 1986 and July 1992. With a mean follow-up of nearly 10 years (115 months) 18 patients (17%) developed inadequate IOP control despite maximum tolerated medical therapy and required additional procedures. Shah et al2 found comparable results with 33/281 (11.7%) of their patients with a GDD (Molteno, Baerveldt and Ahmed) at a mean follow-up of 35 months requiring an additional tube surgery secondary to inadequate IOP control.

    Management of this problem has included needling over the plate, excision of the capsule over the plate,5 surgical revision of the implant, the addition of a new drainage implant and cyclophotocoagulation (CPC).2 ,4–9 Generally, the most successful of these options in the literature have been placement of an additional GDD implant or the performance of laser CPC. However, current literature describes only small numbers of patients with mixed diagnoses and limited follow-up.

    This study compares the long-term results of a consecutive group of patients at a single centre with inadequate IOP control following initial GDD surgery who received either an additional drainage implant or laser photocoagulation.

    Methods

    A retrospective chart review was conducted including patients who had received either a double plate Molteno (Molteno Ophthalmic, New Zealand) implant or a Baerveldt 350 mm2 (Advanced Medical Optics, USA) implant between July 1986 and November 2011 by two glaucoma specialists (MBS and MFS) at the University of Florida. Fifty-two eyes of 48 subjects who have previously undergone an aqueous tube shunt implant followed by a secondary interventional treatment with either CPC or a sequential GDD, including both the paediatric and adult populations, were identified. Five eyes of five patients had less than 6 months of follow-up from their secondary glaucoma procedure and were excluded from the comparison part of the study if still successfully meeting IOP criteria, because it was not felt possible to adequately judge success of the procedure at this early postoperative stage. Thus a total of 47 eyes of 43 patients were analysed. Treatment with CPC was performed on 32 eyes (68%) while placement of a second GDD had been performed on 15 eyes (32%). The factors regarding the surgical decision for which treatment was performed for each individual patient are difficult to access as this is a retrospective study, but may include the general health of the patient and the preference of the surgeon and/or patient. Despite this difficulty determining patient surgical assignment, when the demographic data of the groups were compared the only statistically significant difference between the two groups was found to be time to final follow-up. Specifically, there were no significant differences noted between the groups’ initial IOP, medication use, diagnosis, visual acuity (VA) and age (see table 1).

    View this table:
    Table 1

    Demographic comparison

    Patients in the GDD group received as a second implant either a double plate Molteno drainage implant (n=4) or a Baerveldt 350 mm implant (n=11). The tube was occluded with an internal stent+external dissolvable ligature and generally became functional around 6 weeks after implantation. Venting slits were made in the tube for early IOP control as previously described.10

    Treatment with CPC was performed using either a neodymium-doped yttrium aluminium garnet (Nd:YAG) laser (n=9) or later a diode (n=23) laser. Treatment was applied to 270° of the ciliary body with approximately six spots being applied per quadrant and avoiding the 3 and 9 o'clock meridians to spare the long ciliary arteries. An average of 18–20 spots were treated in total with the power adjusted to produce only an occasional small gas bubble formation or ‘pop’.

    In this study data on the IOP measurement, VA, number of hypotensive glaucoma agents and all ocular procedures and complications were collected from the patients’ medical records at the time of first implant and time of secondary procedure, and the data were recorded for every visit for the duration of their follow-up. Post-procedural follow-up data were recorded at 1, 7 and 14 days, 1, 3, 6 and 12 months, and annually thereafter. We determined the time from primary procedure to secondary procedure and from secondary procedure to either secondary intervention failure or, for those successful, to the most recent follow-up. Patient demographics, type of glaucoma, prior surgery, IOP, anti-glaucoma medications, VA and complications as well as the duration of follow-up were noted. IOP failure was defined as a pressure of greater than 18 mm Hg on two sequential measurements and VA failure was defined as a loss of at least 2 lines of vision on Snellen testing or no light perception (NLP).

    For the purpose of further analysis, all Snellen VAs were converted to their logarithm of the minimal angle of resolution (logMAR) equivalent. Paediatric patients who were unable to perform initial Snellen acuity assessment and patients with VAs of light perception (LP) or NLP were excluded from data analysis with the use of logMAR at that time point due to a lack of correlation with logMAR equivalent values.

    Kaplan–Meier survival analyses were performed to compare the duration of intervention success. Paired t tests were used to compare IOP and VA for variance of continuous variables. χ2 or Fisher exact test were used for categorical variables. Analysis was considered statistically significant with a p value of less than 0.05.

    Results

    The mean age of the population at the time of the secondary procedure was 51.60±25.80 (SD) years (range 1–80 years), of which 53% were male patients and 47% female patients in the CPC group, and 45.07±23.47 (SD) years (range 8–82 years) with 47% male patients and 53% female patients in the sequential GDD group. The diagnoses of glaucoma and further demographic comparisons of each group are represented in table 1.

    No significant demographic differences were found between the two groups. The mean time to failure of initial aqueous tube shunt placement was just over 3 years (37±28.6 months; range 5–156 months) with 9 of the 47 eyes (19%) failing after 5 years. When calculated according to group affiliation, time to initial tube failure was 35±31.4 months (range 5–156 months) in the CPC group and 42±21.7 months (range 17–87 months) in the GDD group with a p value of 0.771.

    The average follow-up time in our study after the secondary procedure was 7 years (85±80.9 months; range 6–254 months). This was significantly different (p<0.005) for the two groups with just over 5 years (63±65.8 months; range 6–254 months) in the CPC group and 11 years (132±91.81 months; range 12–254 months) in the sequential GDD group.

    Prior to the secondary procedure, the mean IOP preoperatively was 22.4±5.7 mm Hg (range 13–42) in the CPC group and 22.6±6.5 mm Hg (range 14–40) in the GDD group. The scattergram (figure 1) compares with the mean IOP at 12 months. Seven eyes in the CPC group required further intervention before their 12-month follow-up and were excluded. The remaining 25 eyes in the CPC group had a mean IOP of 16.0±6.13 mm Hg (range 4–31) with a 6.13 mm Hg (27%) reduction at 1 year. In the GDD group at 12 months, the mean pressure was 15.3±6.51 mm Hg (range 6–33) with a reduction of 7.31 mm Hg (33%) with only one eye excluded for requiring further treatment before this date.

    Figure 1
    Figure 1

    Scattergram of intraocular pressure (IOP) at 12 months. (Data from patients which failed prior to 12 months were carried forward from the last recording prior to additional intervention and are denoted by a coloured circle.) CPC, cyclophotocoagulation.

    Eleven of the 32 (34.4%) eyes in the CPC group, despite the re-introduction of maximal tolerated medical therapy, required additional treatment following secondary procedure at a mean of 13.5 months (range 5–40 months) while 9 of the 15 (60%) in the sequential GDD group underwent additional treatment at a mean of 73.4 months (range 5–173 months) (see Kaplan–Meier survival proportions analysis shown in figure 2). Of the 11 eyes that failed after receiving CPC, four received an additional tube and seven received further CPC laser treatment. In three of these 11 eyes, a total of four procedures were required, and for two eyes a total of five procedures were required. All nine eyes that failed treatment with an additional tube initially received further treatment with CPC. Three of the nine eyes required a total of four procedures, two required five, and one required seven procedures (including a third drainage tube).

    Figure 2
    Figure 2

    Kaplan–Meier analysis of time in months for either cyclophotocoagulation (CPC) or secondary tube, to fail.

    At final follow-up for the CPC patients (mean 63 months), the pressure was 15.9±6.64 mm Hg (range 6–31) with a 6.47 mm Hg (27%) IOP reduction. The final mean IOP at the most recent examination for all eyes in the GDD group (mean 132 months) was 14.80±6.84 mm Hg (range 8–34) with a 7.8 mm Hg or 30.5% IOP reduction.

    The number of hypotensive drops used preoperatively before the second procedure was 2.66±1.07 and 2.60±1.35 in the CPC and GDD groups, respectively. Twelve months after the secondary procedure these were significantly reduced to 1.76±1.01 in the CPC group and to 1.50±1.28 in those eyes that underwent a sequential tube placement. At the most recent study follow-up, those in the CPC group (mean 63 months) were found to have a small but significant reduction in glaucoma drops to 2.09±1.20 while the GDD group (mean 132 months) had an insignificant reduction to 2.20±1.01 medications.

    An insignificant worsening of VA from logMAR 1.22±0.97 to 1.27±0.97 (p=0.618) and from 1.01±0.89 to 1.11±0.90 (p=0.070) was noted in both the CPC and GDD groups from the preoperative visit prior to the secondary procedure to the 12-month follow-up evaluation. For those with better than hand motion (HM) vision preoperatively and who were old enough to perform VA testing, 5/14 of the GDD group (36%) had worsening of VA of 2 Snellen lines or more versus 4/23 (17%) of the CPC group. At the most recent follow-up, there was a significant decrease in the logMAR to 1.66±1.20 (p=0.006) in the CPC group (mean 63 months) and to 1.67±1.14 (p=0.040) in the GDD group (mean 132 months). Thirty-two per cent of eyes in the CPC group (8/25) progressed to HM or less by last visit (four eyes were excluded due to this being their initial VA and three paediatric patients were excluded due to inability to obtain initial VA) compared with 38% of eyes (5/13) in the GDD group, with two eyes excluded due to this being their initial VA. One eye of a paediatric patient in the CPC group progressed to NLP. The scattergram and bar graph (figures 3 and 4) compare visual outcome at 12 months and final follow-up in non-paediatric patients. Table 2 outlines the results of IOP, VA and glaucoma medications between the CPC and sequential GDD groups.

    View this table:
    Table 2

    Mean IOP, VA and glaucoma drops following secondary procedure in the comparing groups

    Figure 3
    Figure 3

    Scattergram of visual acuity (VA) status for both groups at 12 months (A few patients who were too young to perform acuity testing were excluded from this scattergram, n=4). CPC, cyclophotocoagulation; logMAR, logarithm of the minimal angle of resolution.

    Figure 4
    Figure 4

    Vision outcomes, at final follow-up, in non-paediatric patients with initial better than hand motion vision. The average time to final follow-up after the second procedure was 60 months in the cyclophotocoagulation (CPC) group and 132 months in the tube group.

    Corneal decompensation occurred in 7 of 32 eyes (21.9%) in the CPC group (three eyes with pre-existing penetrating keratoplasty, PKP) and 4 of 15 eyes (26.7%) in the GDD group (one eye with prior Descemet stripping endothelial keratoplasty). There was an existing corneal graft in 5/32 eyes in the CPC group (15.6%) and 2/15 eyes in the GDD group (13.3%). Difference in rates of corneal decompensation between the two observed groups was not found to be statistically significant (p=0.7252).

    Discussion

    Although there are several common complications of GDD surgery, including erosion of the tissues overlying the implant, drainage tube obstruction and corneal decompensation,9 ,11 the major factor leading to partial or complete tube failure is inadequate IOP control despite re-addition of glaucoma medications related to excessive fibrotic response occurring around the drainage plate.2 ,5 ,7 Management of a patient that has had failure of a glaucoma drainage implant is controversial. Several different strategies are recommended to reduce the elevated pressure including needle revision over the implant's drainage plate, capsule excision, the addition of a sequential drainage implant and/or CPC.12 There are only a dozen or so papers in the literature, most of which are non-comparative and generally report only small numbers of patients with mixed diagnoses and more importantly a limited follow-up of less than 40 months from the intervention. Although previous studies have shown success rates of 25%–75% with capsule excision over the plate, further treatment by means of an additional tube implant or CPC has generally demonstrated greater rates of success.5 ,7

    Table 3 summarises the published results of many studies on sequential tubes, capsule excision and CPC after failure of initial GDD for comparison with our results.

    View this table:
    Table 3

    Success rates in the literature

    We identified seven previous publications, all of which retrospectively reported their findings following a sequential second GDD implantation.

    Burgoyne et al12 looked at the outcomes of sequential tubes in 22 of their patients and reported a 33% reduction of IOP from the second tube shunt procedure with a mean follow-up of 35 months. Within this time, 13.6% of eyes were considered to have failed with this treatment. Another study, which retrospectively reviewed the use of additional tube implants in 21 eyes, was conducted by Shah et al2 with a mean follow-up of 35 months. Sixty-two per cent of the eyes (in comparison with 42% in their tube revision group) were considered to have achieved success with a significant reduction of pressure from 29.8 mm Hg preoperatively to 17.7 mm Hg. Godfrey et al11 reported success in 83% and 37% of their subjects at 1 and 3 years who received a secondary GDD implant, with a resulting reduction in IOP from 29.5 mm Hg preoperatively to 19.6 mm Hg in 18 eyes with a mean follow-up of 19.6 months.

    Smith et al13 described a 16% failure rate in 19 eyes at a mean of 39 months of follow-up with a 42% pressure reduction. A series of paediatric cases, of which eight eyes underwent implantation with a second tube, were analysed by Sood and Beck.14 They found a 62.5% success rate with their patients at 26 months of follow-up with a mean IOP of 32.3 mm Hg initially and 19.6 mm Hg at the final follow-up. In the most recent published series by Anand et al,15 43 eyes were followed up for a mean of 33 months. Of these, 83% were concluded to be successful with a mean preoperative IOP of 24.7 mm Hg having reduced to 13.6 mm Hg postoperatively at the most recent follow-up.

    More recently, Tung et al16 evaluated the outcome of a subsequent GDD in those with refractive paediatric glaucoma. Failure of this second tube occurred at a mean of 26.1±32.3 in 42% of eyes. The success rates of this secondary device were found to be 81% at 1 year, 62% at 2 years and 50% at 3 years.

    In our study, the mean follow-up of 11 years in patients with sequential tube placement was appreciably longer than in all these previous studies. Nine of 15 patients eventually failed in the sequential tube group in our study but only three (20%) had failed before 4 years. At this time point our results were similar to those of prior published studies. With longer follow-up of up to 21 years from the second tube insertion, however, our study has found that 60% of the sequential tubes eventually went on to fail due to inadequate IOP control. The majority of these failures (56%) occurred more than 5 years after insertion of the second GDD implant. This suggests that the same aqueous and tissue factors that led to excessive fibrosis and failure of the original implant may, over time, lead to loss of IOP control with the additional implant, despite the increased total surface area available for diffusion over the plates. As with any glaucoma surgery, the tendency for a sequential GDD to fail was more apparent through longer follow-up.14 The number of failures may continue to rise with yet longer follow-up and our study highlights that it is important clinically to follow up these patients for many years following sequential GDD surgery.

    Three of the previous studies2 ,12 ,15 reported statistically significant decrease in the number of hypotensive drops used while one study11 showed no significant reduction in the amount of medications required. Our final outcome of medications required agreed with Godfrey et al11 in that the reduction in medications from 2.60 at the time of secondary implantation to 2.20 at the final visit was not found to be statistically different.

    It is often thought that cyclodestructive procedures are more convenient for both the patient and the surgeon as well as financially less burdensome.13 However, CPC as a standalone procedure is recognised to have a higher risk of hypotony, phthisis, uveitis and vision loss when compared with GDD implantation.12–19

    We identified three previous publications, two retrospective and one prospective, which reported their findings of treatment efficacy with CPC following failure of a primary GDD implantation (see table 3).

    Semchyshyn et al4 describe a series in which they reviewed 21 eyes for a mean duration of 26.9 months following treatment with CPC. Significant reductions of pressure from 35.7 mm Hg preoperatively to 13.6 mm Hg postoperatively were found but 28.6% were considered to have therapy failure. Another retrospective study, which observed the paediatric population, was conducted by Sood and Beck,14 and of their series of 17 eyes, nine underwent CPC as their second procedure. Reduction of pressure from 31.7 mm Hg preoperatively to 22.5 mm Hg was noted with a mean follow-up time of 19.8 months and a 33% failure rate at 24 months.

    Recently, Francis et al17 described a prospective study of 25 patients who had undergone endoscopic CPC following GDD implantation failure, making assessments for 24 months. Success was found to occur in 88% of their patients with a 25.5% reduction of IOP.

    The percentage reduction of pressure in the above study groups was 61.9%, 29% and 25.5% and this is generally comparable with our finding of a 27% reduction at 1 year and 21% IOP reduction at final follow-up. Owing to the advanced nature of glaucoma in many of our cases, our criteria for IOP success was selected as 18 mm Hg or less, which is a stricter definition than that in most previously published reports. Variations in the definition of success may lead to differences when our results are compared with those reported in the previous literature.

    The failure rates were noted to be 28.6%, 33.3% and 12% respectively, which compares with our finding of 34.4%. In contrast to an approximate 2 years of follow-up in these studies, the mean follow-up period in our study for the laser CPC group was just over 5 years. Unlike our study sequential tube group, the majority of the laser CPC failures occurred during the first year, and indeed after 24 months only one further eye went on to fail. We found adequate control of IOP achieved in most patients (66%) with only one laser CPC treatment. To prevent overtreatment with the CPC we only lasered approximately 9 clock hours of the ciliary body. If we change our success criteria to allow for two laser treatments following the failed GDD surgery, given 5–7 months apart, a further two eyes were successful long term (total 72%).

    In the case of a secondary glaucoma surgery failure, further intervention is often pursued, especially in patients with a reasonable life expectancy. Thirty-four per cent of the eyes in our CPC group required additional treatment following their secondary procedure compared with 60% in the sequential GDD group. Although less number of patients in the CPC group required additional treatment, the time in which the additional intervention failed was considerably sooner when compared with the GDD group (mean 13.5 months vs 73.4 months).

    Postoperative corneal oedema and decompensation has been reported in previous studies as a common complication and is therefore important to compare. Ayyala et al20 found that there were no differences between patients undergoing mitomycin C trabeculectomy, a GDD implantation, or Nd:YAG CPC in regards to incidence of PKP graft failure or control of IOP. However, more recent literature meta-analysis review of post-keratoplasty eyes suggests a correlation between glaucoma drainage implants and a higher rate of graft failure when compared with trabeculectomy or CPC.21

    Shah et al2 noted that corneal oedema was more common in patients receiving an additional tube shunt than in those who underwent revision of the original shunt suggesting that the presence of an additional tube in the anterior chamber may be detrimental to corneal function. Of note, we did not find any significant difference in the rate of corneal decompensation between our two observed groups with decompensation occurring in 22% in the CPC group and 27% in the GDD group. Our finding with a sequential tube implant is comparable with the results found in previous studies. However, as our study, like that of Shah et al,2 is a retrospective, non-randomised study, with important factors such as tube position in the anterior chamber and initial endothelial count of the cornea not recorded, it is hard to draw definite conclusions regarding relative risks of corneal decompensation with each procedure.

    Potential limitations of our study include the wide array of primary and secondary glaucoma diagnoses of the patients in this study and with this the possible bias that the diagnosis may have an effect on the surgical outcome. Of course, this is the real-world situation in clinical practice. This was a retrospective study and therefore not randomised. A longer follow-up was noted with the GDD group compared with the CPC group and this difference in length of follow-up could potentially affect the rates of success determined. In both groups, the range of follow-up was up to 254 months but it is possible that there initially may have been some reluctance for the surgeons to perform CPC until a clearer idea of outcome with this technique was apparent. The patient demographics and diagnoses, however, were not significantly different between the two groups (see table 1). Specifically, there were no significant differences between the two groups in relation to age, gender, corneal health, time to primary tube failure, number of operations before the secondary glaucoma procedure or the VA measured in logMAR units at the start.

    In addition, this study reviewed the results of patients under the care of two different surgeons at the same university clinic to allow for a more generalisable analysis of the results. The majority of the patients were operated on by one surgeon (85% of eyes), but both surgeons used each secondary procedure at an equal percentage rate. The numbers of patients operated by the second surgeon were not large enough to allow statistical comparison of outcomes between the two surgeons but there were no obvious differences between the surgical techniques.

    Also, to increase the generalisability of the study, paediatric cases, although small in number, were included in the study population. The number of paediatric cases in each group was found to be of equal percentage, and the inclusion of this population resulted in no significant difference in the outcome between the two groups.

    Quality of life is an important factor for the determination of which surgical intervention would be most beneficial for the patient. With this being a retrospective study, the impact of the procedure on the patient's quality of life could not be determined but might be assessed in future prospective studies.

    Two different models of GDDs and two different lasers (Nd:YAG and diode), which may have affected surgical outcomes, were used throughout this study. No differences in success rate or VA outcome or reduction of more than 2 lines on Snellen testing or those progressing to HM or worse vision were noted between the different implants or laser modalities at the 1-year time point although the numbers are small. The decrease in logMAR VA in the GDD group was greater at final follow-up than in the CPC group but the mean follow-up was significantly longer for the GDD group (11 vs 5 years).

    Conclusion

    The purpose of this study is to try to provide longer-term information to support the patient and surgeon's plan of action when faced with the challenge of failure of an initial GDD. Based on the findings in this study, we suggest that although a greater success rate is predicted initially with the implantation of a sequential GDD, in the long term there is continued attrition and a higher rate of failure. Treatment with CPC has a higher failure rate in the first 2 years but for those who are successful at this point, most seem to then retain good IOP control.

    As this was a non-randomised study and the mean follow-up is longer in the second tube group than the CPC group it is difficult to directly compare the success rate, VA and corneal decompensation outcomes of the two treatment modalities. A prospective, randomised controlled study comparing GDD use with CPC use in patients who fail an initial glaucoma drainage implant surgery is recommended to provide this information. Irrespective of the treatment performed, this study emphasises the importance of long-term follow-up of these patients so that early intervention may be considered at the first sign of failure.

    References

    1. ↵
      1. Gedde SJ,
      2. Schiffman JC,
      3. Feuer WJ, et al
      ., The Tube Verses Trabeculectomy Study Group. Treatment outcomes in the tube versus trabeculectomy (TVT) study after five years of follow-up. Am J Ophthalmol 2012;153:789–803. doi:10.1016/j.ajo.2011.10.026
      OpenUrlCrossRefPubMed
    2. ↵
      1. Shah AA,
      2. WuDunn A,
      3. Cantor LB
      . Shunt revision versus additional tube shunt implantation after failed tube shunt surgery in refractory glaucoma. Am J Ophthalmol 2000;129:455–60. doi:10.1016/S0002-9394(99)00410-9
      OpenUrlCrossRefPubMed
    3. ↵
      1. Gedde SJ,
      2. Heuer DK,
      3. Parrish RK II.
      , Tube Versus Trabeculectomy Study Group. Review of results from the tube versus trabeculectomy study. Curr Opin Ophthalmol 2010;21:123–8. doi:10.1097/ICU.0b013e3283360b68
      OpenUrlCrossRefPubMed
    4. ↵
      1. Semchyshyn TM,
      2. Tsai JC,
      3. Joos KM
      . Supplemental transcleral diode laser cyclophotocoagulation after aqueous shunt placement in refractory glaucoma. Ophthalmology 2002;109:1078–84. doi:10.1016/S0161-6420(02)01019-9
      OpenUrlCrossRefPubMed
    5. ↵
      1. Valimaki J,
      2. Tuulonen A,
      3. Airaksinen PJ
      . Capsule excision after failed Molteno surgery. Ophthalmic Surg Lasers 1997;28:382–6.http://dx.doi.org/10.2147/OPTH.S34848
      OpenUrlPubMedWeb of Science
    6. ↵
      1. Rashad MA
      . Efficacy of repeated 5-fluorouracil needling for failing and failed filtering surgeries based on simple gonioscopic examination. Clin Ophthalmol 2013;7:15–22. http://dx.doi.org/10.2147/OPTH.S34848
      OpenUrlPubMed
    7. ↵
      1. Eibschitz-Tsimhoni M,
      2. Schertzer RM,
      3. Musch DC, et al
      . Incidence and management of encapsulated cysts following Ahmed glaucoma valve insertion. J Glaucoma 2005;14:276–9. doi:10.1097/01.ijg.0000169391.94555.c1
      OpenUrlCrossRefPubMed
    8. ↵
      1. Rosentreter A,
      2. Mellein AC,
      3. Konen WW, et al
      . Capsule excision and PlogenTM implantation for revision after glaucoma drainage device surgery. Graefer Arch Clin Exp Ophthalmol 2010;248:1319–24. doi:10.1007/s00417-010-1385-y
      OpenUrlCrossRef
    9. ↵
      1. Tsai JC,
      2. Grajewski AL,
      3. Parrish RK II.
      . Surgical revision of glaucoma shunt implants. Ophthalmic Surg Lasers 1999;30:41–6.
      OpenUrlPubMedWeb of Science
    10. ↵
      1. Sherwood MB,
      2. Smith MF
      . Prevention of early hypotony associated with Molteno implants by a new occluding stent technique. Ophthalmology 1993;100:85–90. doi:10.1016/S0161-6420(93)31688-X
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Godfrey DG,
      2. Krishna R,
      3. Geenfield DS, et al
      . Implantation of second glaucoma drainage devices after failure of primary devices. Ophthalmic Surg Lasers 2002;33:37–43.
      OpenUrlPubMed
    12. ↵
      1. Burgoyne JK,
      2. WuDunn D,
      3. Lakhani V, et al
      . Outcomes of sequential tube shunts in complicated glaucoma. Ophthalmology 2000;107:309–14. doi:10.1016/S0161-6420(99)00039-1
      OpenUrlCrossRefPubMed
    13. ↵
      1. Smith M,
      2. Buys YM,
      3. Trope GE
      . Second Ahmed valve insertion in the same eye. J Glaucoma 2009;18:336–40. doi:10.1097/IJG.0b013e318182edfb
      OpenUrlCrossRefPubMed
    14. ↵
      1. Sood S,
      2. Beck AD
      . Cyclophotocoagulation verses sequential tube shunt as a secondary intervention following primary tube shunt failure in pediatric glaucoma. J AAPOS 2009;13:379–83. doi:10.1016/j.jaapos.2009.05.006
      OpenUrlCrossRefPubMed
    15. ↵
      1. Anand A,
      2. Tello C,
      3. Sidoti PA, et al
      . Sequential glaucoma implants in refractory glaucoma. Am J Ophthalmol 2010;149:95–101. doi:10.1016/j.ajo.2009.07.019
      OpenUrlCrossRefPubMed
    16. ↵
      1. Tung I,
      2. Marcus I,
      3. Thiamthat W, et al
      . Second glaucoma drainage devices in refractory pediatric glaucoma: failure by fibrovascular ingrowth. Am J Ophthalmol 2014;158:113–17. doi:10.1016/j.ajo.2014.03.017
      OpenUrlCrossRefPubMed
    17. ↵
      1. Francis BA,
      2. Kawji AS,
      3. Vo NT, et al
      . Endoscopic Cyclophotocoagulation (ECP) in the management of uncontrolled glaucoma with prior aqueous tube shunt. J Glaucoma 2011;20:523–7. doi:10.1097/IJG.0b013e3181f46337
      OpenUrlPubMed
    18. ↵
      1. Topouzis F,
      2. Yu F,
      3. Coleman AL
      . Factors associated with elevated rates of adverse outcomes after cyclodestructive procedures versus drainage device procedures. Ophthalmology 1998;105:2276–81. doi:10.1016/S0161-6420(98)91229-5
      OpenUrlCrossRefPubMed
    19. ↵
      1. Noureddin BN,
      2. Wilson-Holt N,
      3. Lavin M, et al
      . Advanced uncontrolled glaucoma. Nd:YAG cyclophotocaogulation or tube surgery. Ophthalmology 1992;99:430–7. doi:10.1016/S0161-6420(92)31956-6
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Ayyala RS,
      2. Pieroth L,
      3. Vinals AF, et al
      . Comparison of mitomycin C trabeculectomy, glaucoma drainage device implantation, and laser neodymium: YAG cyclophotocoagulation in the management of intractable glaucoma after penetrating keratoplasty. Ophthalmology 1998;105:1550–6. doi:10.1016/S0161-6420(98)98046-0
      OpenUrlCrossRefPubMed
    21. ↵
      1. Tandon A,
      2. Espandar L,
      3. Cupp D, et al
      . Surgical management for postkeratoplasty glaucoma: a meta-analysis. J Glaucoma 2014;23:424–9. doi:10.1097/IJG.0b013e31827a0712
      OpenUrlCrossRefPubMed

    Footnotes

    • Data presented in part at ARVO 2012 as a poster and the Zimmerman Lecture ICGS conference September 2012.

    • Contributors Concept and design: JLS, MAL, GM, HK, MFS and MBS; provision of materials, patients or resources: MFS and MBS; data collection: JLS, MAL, GM, HK and MBS; analysis and interpretation: JLS, GM and MBS; writing of article: JLS and MBS; critical revision of article: JLS, MAL, GM, HK, MFS and MBS.

    • Funding Supported in part by an unrestricted departmental grant from RPB.

    • Competing interests None declared.

    • Ethics approval All research was performed in accordance with the University of Florida's Institutional Review Board (IRB) with protocol approval prior to initiation of the study.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Linked Articles

    • At a glance
      Highlights from this issue
      Keith Barton James Chodosh Jost Jonas
      British Journal of Ophthalmology 2015; 99 i-i Published Online First: 20 Nov 2015. doi: 10.1136/bjophthalmol-2015-308032