Glaucoma is a progressive optic neuropathy, often associated with raised intraocular pressure (IOP).1 It is one of the leading causes of blindness in the USA, most often occurring in people above 40 years of age.1 An estimated 67 million people worldwide have glaucoma, with 6.7 million bilaterally blind.2 The mainstay of treatment is reduction of IOP, which has been shown to preserve vision.1,3
Glaucoma drainage devices (GDD) are used frequently to control IOP in complicated or refractive glaucomas.4 One type of valved GDD is the Ahmed glaucoma drainage implant (New World Medical, Rancho, Cucamonga, CA) which was introduced to the market in the early 90s. The Ahmed Flexible Plate Model FP7 has a scarab-shaped end plate made from medical silicone, a surface area of 184 mm2, and a mean closing pressure of 7.1 mm Hg5 to minimize hypotony postoperatively.
After Ahmed glaucoma valve (AGV) surgery, the IOP typically goes through an initial hypotensive phase between 1 and 4 weeks after surgery6,7 followed by a hypertensive phase (HP) around weeks 3 to 6.5,8,9 The HP has been reported in 56% to 82% of patients4–6 and can last for up to 4 to 6 months.4,8 The elevated IOP during this phase may result in further damage to the optic nerve.5
The reason for this transient period of high pressure is poorly understood. After GDD implantation, topical corticosteroid medications such as dexamethasone are used to reduce the inflammation that occurs as a result of the surgery.5 A common side effect of steroid eye drops is elevated IOP.10 As the onset of the HP coincides with the timing of a steroid-induced pressure rise,10 it is possible that postoperative topical steroids may initiate or worsen this phase. However, no studies have investigated whether the HP after AGV surgery is caused by topical steroids.
Topical nonsteroidal anti-inflammatory drug (NSAID) medications such as ketorolac is an alternative to topical steroids after eye surgery.11 Topical NSAIDs do not elevate IOP and have been used as a safe alternative after other types of ocular inflammation and ocular surgery.11,12 A prospective randomized study showed no difference in success rates after trabeculectomy when topical NSAID diclofenac was used in place of topical steroids.13
The purpose of this study was to compare the effect of topical steroids versus NSAIDs on the HP after AGV surgery. We hypothesized that the NSAID group would not have a HP and therefore a lower mean IOP than the topical steroid group.
PATIENTS AND METHODS
This was a prospective, randomized, double-masked controlled trial. The study was approved by the Toronto Western Hospital Research Ethics Board and registered with the US National Institutes of Health (Clinicaltrials.gov identifier NCT00956462). The study was carried out at a university-based tertiary care glaucoma practice between October 1, 2008 and September 30, 2009. Patients scheduled for AGV surgery were considered for the study. Inclusion criteria included age of 18 years or above. Exclusion criteria included the following: undergoing combined glaucoma and cataract surgery; ocular condition that may have required more topical anti-inflammatory therapy (eg, uveitic glaucoma, previous penetrating keratoplasty, etc); pregnant or planning to become pregnant during the study period; breast-feeding; and known allergy to ketorolac or any other NSAIDs.
Two surgeons (G.E.T. and Y.B.) performed the AGV surgery using Model FP7, with anterior chamber tube placement through a standard technique described elsewhere.14 Surgical technique was similar between the surgeons with the main difference being fornix-based (G.E.T.) versus limbal-based flaps (Y.B.) and peribulbar (G.E.T.) versus retrobulbar anesthesia (Y.B.).
Postoperatively, patients were randomized to receive either topical 0.1% dexamethasone or 0.5% ketorolac. Both investigators and study participants were masked to the treatment arm. The study medication was prepared and dispensed by the pharmacy department using identical medication bottles with all bottles labeled as “ketorolac or dexamethasone.” The study drug was perscribed 4 times daily for 6 weeks followed by a slow taper depending on the inflammatory response. Antiglaucoma medication and digital massage by the patient15 were instituted when the IOP was considered to be too high by the treating physician. In addition, all patients received 1% atropine twice daily and 0.3% tobramycin 4 times daily postoperatively for 1 week. There was a total of 5 study visits (postoperative weeks 1, 2, 4, 6 to 8, and 10 to 12) and total follow-up time was 3 months.
Main outcome measure was IOP. Secondary measures included incidence and severity of the HP, mean time to appearance of the HP, visual acuity, number of glaucoma medications (topical and oral agents), postoperative complications, and subsequent procedures. HP was defined as IOP greater than 21 mm Hg after initial postoperative reduction to less than 22 mm Hg, which is consistent with definitions used in published literature.4–6 Severity of the HP was defined as the mean IOP at the appearance of the HP. Hypotony was defined as IOP less than 5 mm Hg in keeping with definitions used by other study groups.4–6
A sample size calculation indicated that a minimum of 12 subjects in each arm (a total of 24 subjects) would be required to detect a difference of 2 mm Hg at a SD of 2.5 mm Hg with a power of 80%. Statistical analyses were performed using SAS 9.1 (SAS Institute, Cary, NC) software packages. The demographic and preoperative data of the 2 groups were compared using the Student t test or the Mann-Whitney-Wilcoxon test for continuous data. For categorical data, the Fisher exact test was used. Differences in IOP between the 2 groups at each time period were analyzed with the profile analysis to take into account the correlation of repeated IOP readings over time from the same individuals.
Twenty-eight consecutive consenting patients fulfilled the inclusion criteria. There were 13 patients in the ketorolac arm and 15 patients in the dexamethasone arm. All patients completed the study. Table 1 summarizes the demographics of the study sample. The mean age of the patients was 64.2±17.7 y in the ketorolac arm versus 62.9±10.9 y in the dexamethasone arm (P=0.817). The mean preoperative IOP was 27.7±11.4 mm Hg versus 31.3±12.6 mm Hg in the ketorolac and dexamethasone groups, respectively (P=0.438). The mean number of glaucoma medications required preoperatively was 4.5±0.8 in the NSAID group and 4.0±0.8 in the dexamethasone group (P=0.075). The median preoperative logarithm of the minimum angle of resolution (logMAR) visual acuity was 0.30 in the ketorolac arm and 1.30 in the steroid arm (P=0.112). The most common indication for surgery was primary open-angle glaucoma (36%), followed by neovascular glaucoma (25%) and secondary open-angle glaucoma (21%).
Figure 1 compares the mean IOP between the 2 groups at each study visit. There was a significant reduction in IOP from baseline for both groups at every visit. The mean postoperative IOP (in mm Hg) in the ketorolac versus dexamethasone arms respectively was as follows: 8.8±4.7 versus 10.0±4.5 at week 1 (P=0.500); 10.7±6.7 versus 17.5±10.4 at week 2 (P=0.053); 11.0±6.5 versus 18.0±7.3 at week 4 (P=0.013); 14.8±8.6 versus 17.5±5.2 at week 6 to 8 (P=0.323); and 14.8±9.6 versus 17.8±7.5 at week 10 to 12 (P=0.374).
Four patients (31%) in the ketorolac arm versus 8 patients (53%) in the dexamethasone arm exhibited the HP (P=0.276). Figure 2 compares the incidence of the HP between the 2 groups at each study time point. The incidence of the HP in the ketorolac arm versus the dexamethasone arm, respectively, was as follows: nil for both groups at week 1; 0 of 13 versus 4 of 15 (27%) at week 2 (P=0.106); 1 of 13 (8%) versus 2 of 11 (18%) at week 4 (P=0.576); 2 of 11 (18%) versus 2 of 9 (22%) at week 6 to 8 (P=0.100); and 1 of 10 (10%) versus 0 of 7 at week 10 to 12 (P=1.000). The severity of the HP, determined by the mean IOP at the appearance of the HP, was 25.8±3.5 mm Hg versus 26.4±8.9 mm Hg in the ketorolac and dexamethasone groups, respectively (P=0.897). The mean time to appearance of the HP was 6.3±1.5 weeks in the ketorolac arm versus 3.8±2.2 weeks in the dexamethasone arm (P=0.069).
At the last study visit, the median logMAR visual acuity was 0.30 in the ketorolac arm and 1.30 in the dexamethasone arm (P=0.375). No study patient lost greater than 5 lines of vision. At the conclusion of the study, 8 patients (62%) in the ketorolac arm and 5 patients (33%) in the dexamethasone arm required glaucoma medications (P=0.255). Among those that required medications, the mean number of glaucoma medications used was 3.1±1.7 for the ketorolac group and 2.4±1.5 for the dexamethasone group (P=0.458).
Table 2 illustrates the postoperative complications encountered in each group. Wound leak was the most severe complication and there were 3 cases (23%) in the ketorolac group versus nil in the steroid group (P=0.087). Conjunctival retraction or wound dehiscence was observed in 8 patients (62%) in the ketorolac arm versus 2 patients (13%) in the dexamethasone arm (P=0.016). Hypotony occurred in 6 patients (46%) in the ketorolac group and 3 patients (20%) in the dexamethasone group (P=0.228). There was 1 patient in each group who experienced rebound inflammation upon discontinuation of the anti-inflammatory eye drop.
Table 3 shows the postoperative procedures required in the 2 arms. Of the 3 patients with wound leak in the ketorolac group, 1 required an autologous conjunctival transplant, and the other 2 required buccal grafts. The latter 2 patients required resuturing of the wound after the buccal graft, and subsequently required a second buccal graft in the same area. One of the patients who received a buccal graft also required shortening of the tube which was carried out at the time of the first buccal graft. Digital massage of the eye performed by patients was considered a procedure and was implemented in 4 patients (31%) in the ketorolac arm and 6 patients (40%) in the dexamethasone arm (P=0.706).
There have not been any published studies to date comparing NSAIDs versus steroids after AGV surgery. The evidence on the HP after AGV surgery is also sparse, and all studies to this point have been retrospective in nature.4–6 Our study is the first prospective randomized double-masked controlled trial evaluating the role of steroids as an etiologic factor in the HP.
In this study, the steroid group consistently had higher mean IOP in comparison with the NSAID group with this difference being statistically significant at week 4 (P=0.013). This corresponds to the timing of steroid-induced pressure rise, which can occur as early as week 2, although typically it occurs at week 4.10 Topical steroids have been shown to produce a pressure response over a period of weeks in both normal11,16 and glaucomatous eyes17,18 although patients with glaucoma are much more susceptible to steroid-induced pressure rise.10 Our study supports the hypothesis that steroids are at least partially responsible for this HP after AGV surgery.
Although the HP is probably not an all or none effect, for the purposes of this study, we chose to define HP as IOP greater than 21 mm Hg after initial postoperative reduction to less than 22 mm Hg. The incidence of the HP was 53% in the steroid group, which is in line with the rates from the published literature (56% to 82%).4–6 The incidence of HP was much lower at 31% in the ketorolac group although the difference between the 2 groups were not statistically significant (P=0.276). In the steroid group, the HP appeared between 2 and 8 weeks with a mean of approximately 4 weeks, which is similar to the literature reports of 3 to 6 weeks.5,6,8,9 In the ketorolac group, the HP appeared between 4 and 12 weeks with a mean of approximately 6 weeks. Nouri-Mahdavi and Caprioli19 group reported a mean peak IOP of 30.1±7.5 mm Hg during the HP. The duration of the HP was not a parameter that our study was designed to evaluate, but Ayyala et al4 reported that the HP stabilized at 4 to 6 months after AGV surgery. Interestingly, there is some evidence that suggests that Ahmed implants are associated with a greater incidence of HP compared with double-plate Molteno's and Baerveldt's.20,21
Whether or not the HP leads to more difficult long-term IOP control is a contentious issue. We performed a subanalysis comparing mean IOP at the end of the study at 3 months between the group that had a HP versus the group that did not show a HP. Mean IOP was 23.2±9.2 mm Hg versus 11.8±4.4 mm Hg, respectively, and this difference was statistically significant (P=0.006). Mean number of glaucoma medications at the end of the study period between the 2 groups was 1.5±1.8 versus 1.4±1.9, respectively, which was not statistically significant (P=0.930). Our data suggests that the presence of HP did lead to more difficult lOP control as the group that showed a HP had a higher mean IOP at 3 months compared with the group that did not exhibit a HP, although the number of medications required were similar at the end of the study. However, it would be helpful to have a study with a longer duration of follow-up.
How and why a steroid response occurs after trabeculectomy or AGV surgery is controversial, as both procedures bypass the trabecular meshwork. Thomas and Jay22 demonstrated that after trabeculectomy approximately 24% of patients with primary open-angle glaucoma exhibited a steroid-induced ocular HP 4 weeks after surgery. Wilensky et al's23 group also found steroid-induced ocular hypertension in patients with filtering blebs. The mechanism for steroid-related IOP elevation after glaucoma surgery is speculative, but perhaps some measure of meshwork outflow still exists in affected patients. With GDD's specifically, the bleb surrounding the plate could also be negatively impacted by the corticosteroids in the immediate postoperative period.
The preoperative versus the postoperative median logMAR visual acuity was similar in the ketorolac arm (P=0.087) and in the dexamethasone arm (P=0.254). AGV surgery was effective in reducing the number of glaucoma medications required postoperatively in both groups, and there was a statistically significant difference between preoperative and postoperative usage of glaucoma medications in the ketorolac arm (P<0.0001) and in the dexamethasone arm (P<0.0001).
It is concerning that all 3 wound leaks (2 with fornix-based flap and 1 with limbal-based flap) were in the ketorolac group, and they all developed at postoperative week 4. As there was representation from both types of flap construction, it seems unlikely that the wound leaks were attributable to technique. However, to better ascertain this, we performed a subanalysis of rate of wound leak based on surgical technique. The rationale was that if one technique was more prone to wound leak, there should be a difference in the rate of wound leak between the 2 groups. We found that the rate of wound leak was 2/11=0.18 with fornix-based flaps versus 1/17=0.06 with limbal-based flaps, and this difference was not statistically significant (P=0.543).
Moreover, the ketorolac group had a greater incidence of conjunctival retraction or wound dehiscence than the dexamethasone arm, and this difference was statistically significant. Although conjunctival retraction is a benign complication, 7 of 8 patients in the ketorolac group and 1 of 2 patients in the steroid group who developed conjunctival retraction had fornix-based flaps. It is possible that flap construction may have been an additional factor in conjunctival retraction.
There is evidence substantiating the association between topical NSAID use and delayed wound healing, conjunctival, corneal, and scleral melts, and even corneal perforation.24–26 The precise mechanism of the ocular surface problem is unclear but several theories have been postulated. NSAIDs enhance the ability of neutrophils to infiltrate, and neutrophil degranulation during inflammation releases certain collagenases.26 NSAIDs may also affect matrix metalloproteinases, which are implicated in tissue degradation.26 In addition, NSAIDs have a topical analgesic effect which may lead to neurotrophic ulceration.26 In our cohort, none of the patients with wound-healing problems had underlying systemic or ocular collagen vascular diseases that would predispose them to wound leak or dehiscence. The wound healing complications encountered in this study may be related to the longer duration and increased frequency of NSAID use, as compared with the relatively quick taper regimen that is commonly adhered to after cataract surgery. The patient in the ketorolac arm who required a conjunctival transplant developed plate exposure at 6 months postoperatively, which necessitated removal of the entire drainage device.
With regards to hypotony, although there seemed to be more cases of this in the ketorolac arm, 3 of 6 cases were because of wound leak and if these cases were removed from the analysis, the rates of hypotony between the 2 groups were similar.
The number of patients with rebound inflammation upon discontinuation of the topical anti-inflammatory was not any higher in the NSAID group.
Our study is unique in that it is the first prospective study examining the role of topical steroids as an initiator or cause of the HP after AGV surgery, and it is the first study to compare the effect of NSAIDS versus corticosteroids on IOP after this type of surgery. Limitations include the small sample size, 2 surgeons with slight variation of surgical technique, and a follow-up time limited to 3 months.
It would be worthwhile comparing the effect of dexamethasone versus less potent steroids (such as rimexalone, lotemax, fluoromethalone, or prednisolone 0.12%) on IOP and the HP after AGV surgery.
In conclusion, post-AGV surgery, mean IOP was greater at all time points in the steroid group with the difference between groups statistically significant at week 4, coinciding with the usual onset of steroid-induced ocular hypertension. This suggests that steroids are at least partially responsible for the higher IOP in the steroid arm. The NSAID group showed greater wound healing problems and as a result, we do not recommend the utilization of NSAIDs after AGV surgery.
1. Ayyala RS, Duarte JL, Sahiner N. Glaucoma drainage devices: state of the art. Expert Rev Med Devices. 2006;3:509–521
2. Quigley HA. Number of people with glaucoma worldwide. Br J Ophthalmol. 1996;80:389–393
3. The AGIS investigators. . The Advanced Glaucoma Intervention Study (AGIS) 7. The relationship between control of intraocular pressure and field deterioration. Am J Ophthalmol. 2000;130:429–440
4. Ayyala RS, Zurakowski D, Smith JA, et al. A clinical study of the Ahmed glaucoma valve implant in advanced glaucoma. Ophthalmology. 1998;105:1968–1976
5. Moss EB, Trope GE. Assessment of closing pressure in silicone Ahmed FP7 glaucoma valves. J Glaucoma. 2008;17:489–493
6. Wu SC, Huang SC, Lin KK. Clinical experience with the Ahmed glaucoma valve implant in complicated glaucoma. Chang Gung Med J. 2003;26:904–910
7. Coleman AL, Hill R, Wilson MR, et al. Initial clinical experience with the Ahmed glaucoma valve implant. Am J Ophthalmol. 1995;120:23–31
8. Hong CH, Arosemena A, Zurakowski D, et al. Glaucoma drainage devices: a systematic literature review and current controversies. Surv Ophthalmol. 2005;50:48–60
9. Lieberman MF, Ewing RH. Drainage implant surgery for refractory glaucoma. Int Ophthalmol Clin. 1990;30:198–208
10. Kersey JP, Broadway DC. Corticosteroid-induced glaucoma: a review of the literature. Eye. 2006;20:407–416
11. Flach AJ. Topical nonsteroidal antiinflammatory drugs in ophthalmology. Int Ophthalmol Clin. 2002;42:1–11
12. Kim A, Stark WJ. Are topical NSAIDs needed for routine cataract surgery? Am J Ophthalmol. 2008;146:483–485
13. Kent AR, Dubiner HB, Whitaker R, et al. The efficacy and safety of diclofenac 0.1% versus prednisolone acetate 1% following trabeculectomy with adjunctive mitomycin-C. Ophthalmic Surg Lasers. 1998;29:562–569
14. Freedman J, Trope GETrope GE. How to insert a glaucoma implant. Glaucoma Surgery. 2005 Boca Raton Taylor and Francis:63–74
15. McIlraith I, Buys Y, Campbell RJ, et al. Ocular massage for intraocular pressure control after Ahmed valve insertion. Can J Ophthalmol. 2008;43:48–52
16. Armaly MF. Effect of corticosteroids on intraocular pressure and fluid dynamics: I. The effect of dexamethasone in the normal eye. Arch Ophthalmol. 1963;70:482–491
17. Armaly MF. Effect of corticosteroids on intraocular pressure and fluid dynamics: II. The effect of dexamethasone on the glaucomatous eye. Arch Ophthalmol. 1963;70:492–499
18. Cantrill HL, Palmberg PF, Zink HA, et al. Comparison of in vitro potency of corticosteroids with ability to raise intraocular pressure. Am J Ophthalmol. 1975;79:1012–1017
19. Nouri-Mahdavi K, Caprioli J. Evaluation of the hypertensive phase after insertion of the Ahmed glaucoma valve. Am J Ophthalmol. 2003;136:1001–1008
20. Siegner SW, Netland PA, Urban RC Jr, et al. Clinical experience with the Baerveldt glaucoma drainage implant. Ophthalmology. 1995;102:1298–1307
21. Ayyala RS, Zurakowski D, Monshizadeh R, et al. Comparison of double-plate Molteno and Ahmed glaucoma valve in patients with advanced uncontrolled glaucoma. Ophthalmic Surg Lasers. 2002;33:94–101
22. Thomas R, Jay JL. Raised intraocular pressure with topical steroids after trabeculectomy. Graefes Arch Clin Exp Ophthalmol. 1988;226:337–340
23. Wilensky JT, Snyder D, Gieser D. Steroid-induced ocular hypertension in patients with filtering blebs. Ophthalmology. 1980;87:240–244
24. Flach A. Topically applied nonsteroidal anti-inflammatory drugs and corneal problems: an interim review and comment. Ophthalmology. 2000;107:1224–1226
25. Gaynes BI, Fiscella R. Topical nonsteroidal anti-inflammatory drugs for ophthalmic use: a safety review. Drug Saf. 2002;25:233–250
26. Guidera AC, Luchs JI, Udell IJ. Keratitis, ulceration, and perforation associated with topical nonsteroidal anti-inflammatory drugs. Ophthalmology. 2001;108:936–944
© 2011 Lippincott Williams & Wilkins, Inc.