The use of glaucoma drainage devices (GDDs) is becoming more frequent in the treatment of both refractory glaucoma and as the first filtration procedure.1 The Ahmed glaucoma valve (AGV; New World Medical, Rancho Cucamonga, CA), is a valved device designed to open at high intraocular pressure (IOP)2 and close at low pressure although significant variability occurs in the closing pressure of the FP7 model.3 The implanted tube is typically covered with a patch graft and superficially by local conjunctival reapproximation. Sclera, cornea, and tutoplast are most commonly used as patch tissues and less often dura mater or fascia.4 There is a long list of possible complications related to AGV insertion with the more prevalent complications including postoperative hypotony, tube-corneal touch, hypertensive phases, formation of an encapsulated bleb, and conjunctival problems.5
Intraoperative conjunctival problems include “button holes” or lacerations.6 Postoperative conjunctival complications include 2 distinct groups:
1. Conjunctival dehiscence and retraction leading to exposure of the underlying donor patch.
2. Erosion with exposure of the device through both the graft tissue and overlying conjunctiva.7 A literature review reveals mainly anecdotal reports of conjunctival problems related to AGV.
The purpose of this study was to retrospectively analyze the frequency, risk factors, management, and outcomes of postoperative conjunctival problems in a group of patients who had undergone AGV procedure.
MATERIALS AND METHODS
Ethical approval was obtained from the University Health Network Research Ethics Board. A retrospective review was undertaken to evaluate conjunctival complications recorded in subjects that underwent AGV procedures at the glaucoma service, Toronto Western Hospital between September 1, 1999 and March 31, 2007. The names of the subjects who underwent AGV procedure during the investigated time period was extracted from the computerized operation room system at the Toronto Western Hospital, Toronto, ON, Canada. The data were than collected from the surgical reports and from the patient’s charts. The data collection ended on October 30, 2008. All patients were operated by 1 of 2 surgeons (G.E.T. and Y.M.B.) using the S2 Ahmed implant until 2003 and the FP7 Ahmed implant thereafter.
Postoperative conjunctival complications were defined as follows:
1. Wound dehiscence: conjunctival retraction with exposure of underlying donor patch.
2. Device exposure: erosion of the device (tube, plate, or pars plana clip) through the graft tissue and overlying conjunctiva. Early exposure was defined as an exposure documented at the first 3 months after surgery.
This was 12-month postoperative follow-up study. Although all patients were followed up for at least 1 year, the data of patients who underwent device removal or replacement during the first year were extracted up to the point of resurgery as the results beyond that time point could no longer be attributed to the initial procedure. The following baseline data were extracted from the charts: age, sex, systemic and ocular medical history, best corrected visual acuity, number and type of preoperative glaucoma medications, previous glaucoma procedures, preoperative IOP, operative technique including conjunctival flap type (fornix/limbal based), implantation quadrant, and sutures type. Follow-up data were recorded at each consecutive year of follow-up, including: follow-up duration, time interval between surgery, complication diagnosis, and the specific management of the complication. Additional surgical interventions were also collected.
Risk factors for conjunctival complications were divided into 2 categories: ocular and systemic8,9 as follows.
Ocular Risk Factors
1. Previous conjunctival damage caused by ocular medications, trauma, and surgeries. Previous ocular surgeries and long-standing use of topical hypotensive medications may induce conjunctival scarring or thinning,10 previous ocular surgeries, and trauma and the number of preoperative topical hypotensive medications were recorded and compared between the 2 groups.
2. Postoperative use of topical corticosteroid drops. Corticosteroid drops are routinely used in the postoperative phase to control the inflammatory process.9 Exogenous corticosteroid administration is known to impair all phases of wound healing and the mechanism seems to be multifactorial.11 Corticosteroids retard inflammation that is necessary to activate the wound healing process. Secondly, they accelerate catabolism or protein breakdown. This process retards healing as a result of impairment in net protein synthesis required for healing. The dosage and duration of topical corticosteroid treatment were compared between the 2 groups.
3. Ocular inflammatory or surface diseases. All reported ocular chronic inflammatory and surfaces diseases requiring medical treatment were recorded, including allergies, chronic blepharitis, rosacea, and dry eye.
4. Surgical technique. The quadrant of implantation, the incision type (limbal or fornix based), type of sutures, and ocular antimetabolites use may affect the wound healing result. The investigators recorded each of the above factors and compared with the control group.
Systemic Risk Factors
1. Connective tissue diseases.12 All connective tissue diseases were collected and compared between groups.
2. Diabetes mellitus.10,13,14 All cases of diabetes were recorded based only on the diagnosis of the disease with no discrimination between the various types of diabetes, antidiabetic treatment, duration, or severity of the disease.
3. Systemic medications including corticosteroids and cytotoxins.9
The AGV was inserted using either fornix-based or limbal-based conjunctival incisions. When a fornix-based conjunctival incision was performed the tissues were ballooned up by a subconjunctiva/tenon injection of Lidocaine 2% anesthesia. A fornix-based conjunctival flap was then formed by incising the conjunctiva at the limbus. One or 2 relaxing incisions were made parallel to the upper border of the lateral or medial rectus depending on the quadrant of insertion. The conjunctiva was undermined with Wescott scissors and the dissection extended about 10 to 12 mm to create a pocket. When a limbal-based conjunctival incision was performed, a 10-mm-long incision was created between adjacent recti muscles 6 mm from and parallel to the limbus. The AGV plate was secured to the sclera by two 7-0 silk sutures (Synature, Norwalk, CT). Once the tube was inserted into the anterior chamber it was secured to the sclera with a 10-0 nylon suture (Ethicon Inc., Johnson & Johnson, Somerville, NJ) and a lamellar donor corneal patch graft was placed over the tube and secured to the sclera by two to four 10-0 nylon sutures.
Conjunctival Wound Closure Technique
With fornix-based conjunctival incision, the conjunctiva was closed with 4 wing 10-0 nylon sutures, 2 at each side, suturing the ends of the conjunctiva to their original position at the limbus while making sure the conjunctiva fully covered the donor cornea. The relaxing incisions were sutured with 8-0 vicryl sutures (Ethicon Inc., Johnson & Johnson). Limbal-based conjunctival incisions were closed in 2 layers using 8-0 vicryl sutures. The internal tenon’s layer was closed with 3 interrupted stitches and the conjunctival layer with a continuous suture. Antimetabolites were not used in any of the cases. All subjects were treated postoperatively with antibiotic drops for 1 week and corticosteroid drops for at least 4 weeks.
Categorical variables are presented by absolute number and proportion and continuous variables by arithmetic mean, SD, 95% confidence interval (CI), median, and minimum-maximum. χ2 or Fisher exact test was used for categorical variables where applicable. As continuous parameters were not normally distributed (Shapiro-Wilk test), Mann-Whitney and Kruskal-Wallis were used to compare 3 groups of patients for continuous variables. Bonferroni post hoc comparison was used to decide the statistical differences between any 2 groups. Survival was shown by Kaplan-Meier curve. P-value of 5% or less was considered statistically significant. The data were analyzed using SAS version 9.1 for Windows (SAS Institute, Cary, NC).
Two hundred sixty-eight subjects underwent AGV insertion during the studied period. One hundred fifty-eight eyes fulfilled the inclusion criteria and were analyzed. The mean follow-up duration was 43.46 (±24.9) months (range, 12 to 103 mo).
Sixty-seven (42.4%) eyes were diagnosed with postoperative conjunctival complications, including 53 eyes with wound dehiscence, and 14 with device exposure (11 tube exposures, 2 plates, and 1 Ahmed pars plana clip exposure). Three eyes that had both early dehiscence and later exposure were analyzed in the exposure group. The control group included 91 (57.6%) subjects with no recorded conjunctival complications.
Table 1 summarizes the demographic characteristics of the 3 groups. No significant differences were found between the 3 groups in demographics, glaucoma type, and baseline IOP.
The mean time interval between AGV insertion and the diagnosis of conjunctival complications was significantly shorter in the wound dehiscence group (31.6±35.7 d) compared with the device exposure group (996±735 d), P<0.001.
Two subjects (1.3%) were diagnosed with endophthalmitis. Both cases of endophthalmitis occurred in the device exposure group (2/14, 14.2%, 95% CI: 0.018, 0.428).
Risk Factor Analysis
No significant systemic risk factors for conjunctival dehiscence or erosion were found in this study, including age, diabetes mellitus (P=0.54), and connective tissue diseases (P=0.30) although the rates of connective tissue diseases were very low in all groups (0 to 3).
Table 2 summarizes the ocular risk factors. Most ocular risk factors did not reach significance, including previous ocular trauma (P=0.4), prevalence of ocular inflammatory and surface diseases requiring treatment (P=0.33). There were no significant differences in previous glaucoma procedures and the postoperative use of topical corticosteroids followed the same protocol in all cases. The mean preoperative number of glaucoma hypotensive medications was 3.3 (±1.1) for the control group versus 3.8 (±1.1) and 3.9 (±0.9) in the dehiscence and exposure groups, respectively (P=0.01).
Wound dehiscence occurred most commonly with inferior device implantation. The inferonasal quadrant was associated with the greater dehiscence rate (4/7, 57.1%), followed by the inferotemporal quadrant (30/65, 46.2%), the superotemporal (15/61, 24.6%), and the superonasal quadrant (4/25, 16%) (P<0.0073) (Table 2). Comparing valve position between dehiscence and control, there were significantly fewer cases of wound dehiscence in the superotemporal and significantly more cases of dehiscence in the inferotemporal quadrant (P=0.006).
No significant difference was found when the temporal half was compared with the nasal half (P=0.994) and when the superior half was compared with the inferior half. Unlike dehiscence, no statistically significant difference was found between insertion quadrants and AGV device erosions (superotemporal: 95% CI: 3.3, 18.2; superonasal: 95% CI: 0, 28.7; inferotemporal: 95% CI: 3.4, 18.7; inferonasal: 95% CI: 0, 51). There were also no significant differences between the groups regarding incision type and conjunctival closure technique (Table 2).
Forty-eight (90.6%) of the dehiscent wounds were managed conservatively whereas 5 (9.4%), with Siedel-positive wound leaks, were resutured a mean of 7.2±4.1 days after surgery.
The 14 exposed devices were treated as follows: 5 exposed tubes and 1 exposed plate were recovered with a lamellar donor corneal patch graft and an autologous conjunctival graft. One tube and 1 Ahmed pars plana clip were recovered by an oral buccal mucous membrane in combination with a lamellar corneal patch graft and autologous conjunctiva as reported elsewhere by us.13 Two exposed tubes and 1 exposed plate were removed and a second AGV inserted at a different quadrant, whereas 1 exposed tube was removed and not replaced.
There were no recorded cases of endophthalmitis in the dehiscence and the control groups whereas 2 subjects (14.3%) with tube exposure presented with endophthalmitis 3 and 6 years after surgery. One tube was inserted in the superonasal quadrant and the other in the inferotemporal quadrant. Once diagnosed both underwent urgent vitrectomy, tube removal, and intravitreal antibiotic injection. One subject ended up with a visual acuity of hand movements and the other developed phthisis and eventually underwent enucleation.
Conjunctival problems related to AGV have received relatively little attention in the literature, with little data regarding frequencies, risk factors, management, and outcomes. Two hundred sixty-eight charts of subjects who underwent AGV surgeries were reviewed. One hundred fifty-eight were included in the study and the rest (110 patients) were excluded as they did not complete 12 months follow-up. These subjects were not intended to be part of the study and specifically were not included in the protocol as a follow-up period of <1 year was likely not long enough to allow for erosion to occur. Toronto Western Hospital is a tertiary referral center that provides medical services to a wide range of geographical distances. Most of the patients are followed up at the hospital during the early postoperative period (up to 6 mo). Some patients prefer to continue follow-up with their general ophthalmologist in their home area and are lost to follow-up (although most of those patients are referred back when serious or refractory complications are diagnosed by their general ophthalmologist). This explains the high percentage of exclusions in this study.
Postoperative conjunctival complications include 2 distinct groups:
1. Conjunctival dehiscence and retraction leading to exposure of underlying donor patch. This complication was recorded in 53 subjects (33.5%) in this study and occurred early in the postoperative period (31.6±35.7 d after surgery). We could not find any reports documenting the occurrence rate of conjunctival dehiscence in the published peer-reviewed literature. The authors believe that as this complication is considered minor it is overlooked by many ophthalmologists. The same rates of dehiscence were recorded with both incision types (fornix-based and limbal-based incisions, 95% CI: 24.4, 45.7 and 24.8, 44.9, respectively) and with both types of sutures (nylon and vicryl). There was a higher rate of wound dehiscence in the inferior half. This may reflect the fact that inferior implantation is usually performed when the superior conjunctiva is not suitable because of tissue scarring which may predispose the tissue to retraction. In addition there is likely more exposure and lid rubbing over the inferior surgical location increasing the risk of wound complications. Conjunctival dehiscence can be managed conservatively in most cases.6 In case of a positive Seidel it is the practice of the surgeons involved in this study to resuture these rather than observe given the risk of endophthalmitis with an active Seidel leak. We are therefore unable to comment on the outcome of Seidel-positive cases managed with observation.
2. Device exposure due to erosion of the device through the graft tissue and overlying conjunctiva was recorded in 14 subjects (8.9%) after a mean of 33.2±24.5 months after surgery in this study (Fig. 1). No early exposures were documented in this study (ie, documented exposures occurring at the first 3 mo after surgery).
The same rates of exposure were recorded with both incision types (fornix-based and limbal-based incisions, 95% CI: 3, 45.1 and 5.9, 19.6, respectively).
Only 3 patients with a documented conjunctival dehiscence developed a tube or plate exposure. These subjects were analyzed in the device exposure group. As 53 subjects were diagnosed with conjunctival dehiscence only we do not feel that conjunctival dehiscence necessarily predisposes to device exposure. Various AGV exposure rates have been published in the literature. Ayyala et al15 published a 12-month follow-up report on 85 patients who underwent AGV implantations and recorded 11.7% end-plate exposure and 7% exposed tubes. Huang et al16 recorded a much lower exposure rate of 2.5% out of a group of 159 AGV procedures (mean duration of follow-up 13.4±0.7 mo). Wilson et al17 reported a 5.1% exposure rate in a group of 59 patients who were followed up for a mean of 31 months. The reported exposure rate in this study (8.9%) is within the range reported in the literature.
Exposure is a potentially catastrophic complication as it predisposes the eye to endophthalmitis.17–19 The fact that device exposure was recorded years after the initial procedure suggests that these patients must be examined regularly for device exposure for life. Patients must be made aware of the possible symptoms and signs of endophthalmitis and should be instructed to seek medical help urgently if symptoms arise.
Costa and colleagues17,20,21 reported endophthalmitis rates of 3.8%, 5%, and 12.5%, respectively. Despite the long follow-up duration in this study, we report an endophthalmitis rate of 1.3% (2/158). Both cases occurred in subjects with an exposed device (14.2%; 95% CI: 0.02, 0.43). Device exposure requires urgent surgical repair. If infection is diagnosed the device should be removed as soon as possible, whereas in noninfected cases, the exposure should be promptly repaired without removing the device.
A number of strategies have been proposed to cover noninfected GDDs, each of which has met with variable success. We have recently reported successful coverage of eroded devices using buccal mucous membrane.14 Other options include cornea or sclera donor patch grafts covered with conjunctival autograft.15
Risk factors for conjunctival damage include both ocular and systemic factors.
Ocular risk factors include long-standing ocular medical therapy,22,23 which may interfere with conjunctival healing and lead to postoperative complications.
Long-standing use of multiple hypotensive medications can change the nature of the conjunctiva and may lead to conjunctival complications.24 We found that a greater number of topical glaucoma medications was a risk factor for conjunctival complications (P=0.01). The mean difference in number of glaucoma medications was small (mean of 0.5 to 0.6 more medications in the group with conjunctival complications). The clinical significance of this result limited as the usual algorithm of attempting to control IOP with maximal medical therapy before considering a surgical intervention is unlikely to be affected by this outcome as the risk of surgery is greater than the risk of medical therapy. We believe that this is the first report suggesting that the preoperative use of a high number of ocular hypotensive medications is a risk factor for postoperative conjunctival dehiscence. Past trauma or surgery has also been reported as a risk factor.8 Some potential differences in the risk factors for conjunctival complications, such as systemic risk factors and others could have been missed in this study because of small numbers in some comparisons. The fact that no risk factors for device exposure were found in this study may also be related to the relatively low occurrence rate of this complication.
A literature review reveals different opinions regarding the preferred site for GDD implantation. Sidoti25 supports superior insertion claiming that inferior placement of GDD is considered technically more difficult, accompanied by more frequent wound dehiscence and anterior exposure of the patch graft owing to the relatively shorter lower fornix and poor healing. Pakravan et al26 found that the overall rate of complications, such as implant exposure necessitating removal, cosmetically unappealing appearance, and endophthalmitis, was higher in the inferior group. We previously suggested27 the superotemporal quadrant as the preferred surgical site because of easier surgical access, less postoperative exposure, maximal coverage of the tube by patch graft and plate by the eyelid and less diplopia. We report for the first time that wound dehiscence occurs most commonly in the inferonasal quadrant (95% CI: 18.4, 90.1), followed by the inferotemporal (95% CI: 33.1, 58.2), the superotemporal (95% CI: 12.9, 33.8), and the superonasal quadrants (95% CI: 10.9, 52.0), P<0.0073. In our practice AGVs are usually implanted superiorly whereas inferior placement is usually reserved for cases with poor superior conjunctiva owing to previous surgeries. This may have biased our results and lead to a higher prevalence of wound dehiscence in such cases.
AGV may be inserted through either limbal-based or fornix-based incisions. To the best of our knowledge there are no publications in the literature comparing these 2 options in terms of conjunctival complications. In this paper we report for the first time that no significant differences exist between limbal-based or fornix-based incisions with regard to conjunctival complications. Conjunctival closed with nylon or vicryl sutures were also accompanied by similar rates of conjunctival complications.
There are some limitations to this study; mainly its retrospective nature and the fact that some risk factors could have been missed because of small numbers in some comparisons. Despite the limitations, some of the findings have implications for clinical practice. Conjunctival wound dehiscence was found to be a common complication after AGV insertion (33.5%). As it is usually a benign condition it does not require surgical repair (specifically in cases where the tube and plate are well covered with no aqueous leak). Full-thickness erosions of the tube or plate are less common (8.9%), occurring much later and requiring prompt surgical repair to prevent endophthalmitis. The number of preoperative glaucoma medications is a significant risk factor for postoperative conjunctival complications. As delayed diagnosis of device exposure may lead to endophthalmitis, the surgeon should educate the patient regarding the signs and symptoms of device exposure and leakage (mainly tearing) to lessen the rate of this devastating complication.
1. Ayyala RS, Duarte JA, Sahiner N.Glaucoma drainage devices: state of the art.Expert Rev Med Devices.2006;3:509–526.
2. Ahmed AMChen TC.Ahmed valve surgery.Glaucoma Surgery.2008.Philadelphia, PA:Elsevier;55–74.
3. Moss EB, Trope GE.Assessment of closing pressure in silicone Ahmed FP7 glaucoma valves.J Glaucoma.2008;17:489–493.
4. Rosenberg LF, Krupin TRitch R, Shields MB, Krupin T.Implants in glaucoma surgery.The Glaucomas.1996.St. Louis:Mosby;;3:1789–1790.
5. Lim KS, Allan BDS, Lloyd AW, et al..Glaucoma drainage devices; past, present, and future.Br J Ophthalmol.1998;82:1083–1089.
6. Freedman J, Melamed S, Trope EGTrope GE.Management of glaucoma implant complications.Glaucoma Surgery.2005:75–81.Boca Raton:Taylor & Francis Group.
7. Smith M, Buys YM, Trope GE.Replacement of Ahmed aqueous drainage devices in eyes with device-related complications.J Glaucoma.2009;18:484–487.
8. Minckler DS, Vedula SS, Li TJ, et al..Aqueous shunts for glaucoma. Cochrane Database of Systematic Reviews 2006, Issue 2, Art. No.: CD004918.
9. Akyol N, Demir T, Kükner A, et al..Effects of systemic octreotide, local mytomycin-C and local corticosteroids on wound-healing reaction after glaucoma surgery.Int Ophthalmol.2001;24:235–241.
10. Byun YS, Lee NY, Park CK.Risk factors of implant exposure outside the conjunctiva after Ahmed glaucoma valve implantation.Jpn J Ophthalmol.2009;53:114–119.
11. Demling RH, DeSanti L.Scar management strategies in wound care.Rehab Manag.2001;14:26–30.
12. Hoan-Xuan T, Baudouin C, Creuzot-Garcher C.Inflammatory Diseases of the Conjunctiva. Chapter 9: Systemic Disease and Conjunctivitis.2001Stuttgart: Thieme;145–160.
13. Eming SA, Hammerschmidt M, Krieg T, et al..Interrelation of immunity and tissue repair or regeneration.Semin Cell Dev Biol.2009;20:517–527.
14. Rootman DB, Trope GE, Rootman DS.Glaucoma aqueous drainage device erosion repair with buccal mucous membrane grafts.J Glaucoma.2009;18:618–622.
15. 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.
16. Huang MC, Netland PA, Coleman AL, et al..Intermediate-term clinical experience with the Ahmed glaucoma valve implant.Am J Ophthalmol.1999;127:27–33.
17. Wilson MR, Mendis U, Paliwal A, et al..Long-term follow-up of primary glaucoma surgery with Ahmed glaucoma valve implant versus trabeculectomy.Am J Ophthalmol.2003;136:464–470.
18. Al-Torbak AA, Al-Shahwan S, Al-Jadaan I, et al..Endophthalmitis associated with the Ahmed glaucoma valve implant.Br J Ophthalmol.2005;89:454–458.
19. Bayraktar Z, Kapran Z, Bayraktar S, et al..Delayed-onset Streptococcus pyogenes
endophthalmitis following Ahmed glaucoma valve implantation.Jpn J Ophthalmol.2005;49:315–317.
20. Costa VP, Azuara-Blanco A, Netland PA, et al..Efficacy and safety of adjunctive mitomycin-C during Ahmed glaucoma valve implantation: a prospective randomized clinical trial.Ophthalmol.2004;111:1071–1076.
21. Lima FE, Magacho L, Carvalho DM, et al..A prospective, comparative study between endoscopic cyclophotoco-agulation and the Ahmed drainage implant in refractory glaucoma.J Glaucoma.2004;13:233–237.
22. Epstein SP, Ahdoot M, Marcus E, et al..Comparative toxicity of preservatives on immortalized corneal and conjunctival epithelial cells.J Ocul Pharmacol Ther.2009;25:113–119.
23. Baudouin C, Riancho L, Warnet JM, et al..In vitro studies of antiglaucomatous prostaglandin analogues: travoprost with and without benzalkonium chloride and preserved latanoprost.Invest Ophthalmol Vis Sci.2007;48:4123–4128.
24. Broadway DC, Chang L.Trabeculectomy, risk factors for failure and the preoperative state of the conjunctiva.J Glaucoma.2001;10:237–249.
25. Sidoti PA.Inferonasal placement of aqueous shunts.J Glaucoma.2004;13:520–523.
26. Pakravan M, Yazdani S, Shahabi C, et al..Superior versus inferior Ahmed glaucoma valve implantation.Ophthalmology.2009;116:208–213.
27. Rachmiel R, Trope GE, Buys YM, et al..Intermediate-term outcome and success of superior versus inferior Ahmed glaucoma valve implantation.J Glaucoma.2008;17:584–590.
© 2014 by Lippincott Williams & Wilkins.