Measurement of Gamma-Irradiated Corneal Patch Graft Thickness After Aqueous Drainage Device Surgery

Key Points

Question  Do gamma-irradiated sterile cornea patch grafts become thinner over time after aqueous drainage device surgery?

Findings  In a cross-sectional study including 107 patients, thinner grafts were observed as the time after surgery lengthened as assessed by anterior segment optical coherence tomography. The odds ratio per year that the graft was undetectable was 2.1.

Meaning  Because gamma-irradiated sterile cornea patch grafts thin and may become undetectable after surgery, their placement is no guarantee against exposure; better materials and strategies are needed for preventing tube exposure.

Importance  Exposure of the tube of an aqueous drainage device (ADD) through the conjunctiva is a serious complication of ADD surgery. Although placement of gamma-irradiated sterile cornea (GISC) as a patch graft over the tube is commonly performed, exposures still occur.

Objectives  To measure GISC patch graft thickness as a function of time after surgery, estimate the rate of graft thinning, and determine risk factors for graft thinning.

Design, Setting, and Participants  Cross-sectional study of graft thickness using anterior segment optic coherence tomography (AS-OCT) was conducted at the Wilmer Eye Institute at Johns Hopkins Hospital. A total of 107 patients (120 eyes, 120 ADDs) 18 years or older who underwent ADD surgery at Johns Hopkins with GISC patch graft between July 1, 2010, and October 31, 2016, were enrolled.

Intervention  Implantation of ADD with placement of GISC patch graft over the tube.

Main Outcomes and Measures  Graft thickness vs time after ADD surgery and risk factors for undetectable graft.

Results  Of the 107 patients included in the analysis, the mean (SD) age of the cohort was 64 (16.2) years, 49 (45.8%) were male, and 43 (40.2%) were African American. The mean time of measurement after surgery was 1.7 years (range, 1 day to 6 years). Thinner grafts were observed as the time after surgery lengthened (β regression coefficient, −60 µm per year since surgery; 95% CI, −80 µm to −40 µm). The odds ratio of undetectable grafts per year after ADD surgery was 2.1 (95% CI, 1.5-3.0; P < .001). Age, sex, race, type of ADD, quadrant of ADD placement, diagnosis of uveitis or dry eye, and prior conjunctival surgery were not correlated with the presence or absence of the graft.

Conclusions and Relevance  Gamma-irradiated sterile corneal patch grafts do not always retain their integrity after ADD surgery. Data from this cross-sectional study showed that on average, the longer the time after surgery, the thinner the graft. These findings suggest that placement of a GISC patch graft is no guarantee against tube exposure, and that better strategies are needed for preventing this complication.

Aqueous drainage devices (ADDs) are an effective surgical intervention for glaucoma.^1,2 Although ADDs have proven to be an important means of lowering intraocular pressure, 1% to 3% of eyes experience exposure of the tube through the conjunctiva within 5 years after ADD surgery³ and 5% of patients experience exposure in eyes that had previously undergone an operation.⁴ This complication can result in hypotony and endophthalmitis and required additional surgery to avoid these complications.⁵

To attempt to reduce the risk of exposure, various donor tissue grafts have been used to cover the tube of the ADD where it lies between the sclera and the conjunctiva. These materials include pericardium,⁶ sclera,⁷ dura,⁸ fascia lata,⁹ amniotic membrane,¹⁰ and various cornea preparations.^11,12 Each of these materials may differ in erosion rate.¹³ Recently, gamma-irradiated sterile cornea (GISC) patch graft preparations have become a popular choice for this surgery because the material minimizes the risk of disease transmission, is transparent after implantation and therefore cosmetically appealing, and can be stored at room temperature for up to 18 months.¹⁴

Previous studies have reported on the use of GISC patch grafts in a variety of ophthalmic procedures, including ADD surgery.^14-16 These studies have monitored patients for patch graft failure, with failure defined as exposure of the ADD through the graft and conjunctiva, for intervals of 16 months to 3.7 years after surgery.^13,17

In this cross-sectional study, we measured GISC patch-graft thickness over time, using anterior segment optic coherence tomography (AS-OCT) to determine whether graft thickness decreased with increasing time after surgery and whether the graft thinned to the point that it could not be visualized by AS-OCT.

All patients seen between July 1, 2010, and October 31, 2016, in the Johns Hopkins Hospital Glaucoma Center of Excellence who had undergone ADD surgery were potentially eligible. The following data were extracted from their medical records: age, sex, race uveitis diagnosis, dry eye diagnosis, prior conjunctival surgery (specifically, trabeculectomy, ADD surgery, retinal detachment, and vitrectomy), date of surgery, type of drainage device, and graft used to cover the ADD.

All protocols were approved by the institutional review board at Johns Hopkins Hospital, and methods were compliant with the Declaration of Helsinki.¹⁸ Patients provided written informed consent; there was no financial compensation.

From the pool of potentially eligible patients, individuals were excluded (Figure 1) according to the following criteria: (1) the patient underwent ADD surgery outside of Johns Hopkins Hospital, (2) a patch graft other than GISC (Tissue Banks International) was used, (3) the patient had undergone subsequent ADD revision surgery, (4) the patient was younger than 18 years, or (5) the ADD was placed in the pars plana. All persons included in this study had received either an Ahmed FP7 or S-2 (New World Medical Inc) or Baerveldt 250-mm² or 350-mm² (Pharmacia Lovision) drainage device.

Aqueous drainage device surgery was performed by 6 different surgeons (including one of us, H.D.J.). The surgical techniques were similar, but not standardized. After insertion of the ADD into the eye through a tract made with a 23-gauge needle 1 to 2 mm posterior to the limbus, a split-thickness piece of GISC was placed to cover the entry site of the tube into the eye and as much of the tube as possible since it extended back to the plate. The orientation of the graft was at the surgeon’s discretion. The graft was secured to the sclera with two to four 10-0 nylon sutures.

During the postoperative visit during which AS-OCT imaging was performed, the surgeon was asked to assess, on slitlamp examination, the status of the graft overlying the tube. Images were obtained using the Visante AS-OCT 3.0, model 1000 (Carl Zeiss Meditec Inc). The examiner attempted to image the ADD in the subconjunctival space 2 mm posterior to the limbus (Figure 2A, yellow arrowhead), adjusting the angle of the image so that it was perpendicular to the ADD (Figure 2A, white line). Two images were taken for each ADD by the examiner and the better-quality image was selected for analysis. A typical image is illustrated in Figure 2B.

Two examiners (2 of us, R.A.D. and A.M.) conducted this study and were responsible for obtaining patient images. A subset of ADDs was imaged by both examiners at the same appointment to assess reproducibility.

Measurements were performed directly on the AS-OCT and obtained in a standard manner: the examiner measured the external diameter of the ADD tube (Figure 2B, blue line) (measurement 1) and then the examiner measured the distance from the external surface of the tube closest to the sclera to the junction of the graft and the internal surface of the Tenon capsule and conjunctiva (measurement 2). The thickness of the graft (Figure 2B, yellow line with bars) was obtained by subtracting measurement 1 from measurement 2.

To compare the characteristics of patients whose grafts were detectable with those whose grafts were not detectable, a paired, 2-tailed t test was used for continuous variables and the χ² test was used for categorical variables.

To determine the rate of graft thinning, a linear regression model with robust variance estimates was carried out, regressing the thickness of the graft on the time since ADD surgery. Logistic regression models with robust variance estimates were carried out to determine the effect of predictors (ie, patient age, sex, race, time since ADD surgery, type of ADD, uveitis diagnosis, dry eye diagnosis, prior conjunctival surgery, and quadrant of ADD placement) on graft status. Mixed-effects regression models were considered but were not used because the correlation between eyes from the same patient was negligible. All statistical analyses were performed using Stata, version 14.2 (StataCorp), and P ≤ .05 was accepted as statistically significant.

Nine patients (10 eyes, 10 ADDs, 8.4% of patients of the study cohort) were imaged by both examiners on the same day. The 2 examiners independently measured the graft thickness (Pearson correlation coefficient, 0.84; P = .002) (eFigure in the Supplement). Among the 10 ADDs, there was only 1 disagreement of graft status (present vs absent); both examiners agreed on the graft status for the other 9 cases.

One hundred seventy-one patients (202 ADDs) were considered. After eliminating excluded patients and eyes and excluding eyes that could not be imaged, 107 patients (120 eyes, 120 ADDs) were included (Figure 1). The mean age of the patients was 64 (16.2) years (range, 24-96 years), with 49 (45.8%) men, 50 (46.7%) white patients, and 43 (40.2%) African American individuals (Table 1). The mean time since surgery was 1.70 (1.48) years (range, 1 day to 6 years). There were 63 (52.5%) ADDs implanted in the right eye and 57 (47.5%) implanted in the left eyes, 76 (63.3%) Baerveldt ADDs vs 44 (36.7%) Ahmed ADDs, and 102 (85.0%) patients had the ADD implanted in the superotemporal position.

We could not find manufacturer information describing the thickness of the partial thickness graft as it is supplied. We examined the graft thickness measurements that we made within 2 weeks after surgery as a surrogate for preoperative thickness, assuming that there was no change in thickness within the first 2 weeks. The median thickness (n = 5) was 410 µm (range, 200-560 µm). Figure 3 depicts the association between graft thickness and time after ADD surgery, using a Lowess smoothed curve. Each data point represents 1 of the 120 ADDs. When the time after ADD surgery was evaluated in a linear regression model with graft thickness as the outcome, a reduction of graft thickness was observed with a longer time after surgery (β regression coefficient, −60 µm per year since surgery; 95% CI, −80 µm to −40 µm; P < .001).

In 16.6% of the eyes (20 eyes, 20 ADDs), no GISC patch graft could be seen during the study period (Figure 3). When the time after ADD surgery was evaluated in the model with the presence or absence of the graft as the outcome, the odds ratio that the graft was undetectable for each year after ADD surgery was 2.1 (95% CI, 1.5-3.0; P < .001).

To determine whether any clinical or surgical factors showed an association with the presence or absence of the graft, multiple bivariate analyses were performed. Age, sex, race, type of ADD, position of ADD, uveitis diagnosis, prior conjunctival surgery, and dry eye diagnosis were not correlated with the presence or absence of the graft (Table 2).

Seventy of the 120 ADDs were evaluated using a slitlamp examination by the physician. The Pearson correlation between the surgeon’s assessment of the presence or absence of the graft and the detection of a graft by AS-OCT was low (Pearson correlation, 0.198; P = .004).

To our knowledge, this study is the first to quantify the thickness of patch grafts after surgery using AS-OCT. We studied patch graft thinning because we believe that such thinning is a critical step in the development of tube exposure and that, without patch graft thinning, tube exposure is unlikely to occur.

Other studies have examined the outcome of eyes that received patch grafts to cover ADDs, but none has done more than note the presence or absence of tube exposure through the conjunctiva¹⁰ or measured graft thickness in a subset of study patients for illustrative purposes at various time points.⁶ Data from the present cross-sectional study showed that the longer the time after surgery, the thinner the graft and, each year after surgery, the odds ratio of a graft being undetectable by AS-OCT increases by 2.1. At some point after surgery, 16.6% of eyes in this cohort had no detectable graft.

We found a rate of graft thinning of 60 µm/y. This finding should be interpreted in the context of the initial thickness of the graft used at the time of surgery. We could not find any information about the thickness of the grafts coming from the supplier, so we used in vivo measurements in eyes that had surgery within 2 weeks of measurement as a surrogate for initial thickness. Measurements of these 5 eyes yielded a mean (SD) thickness of 414 (139) µm. Therefore, given the initial thickness and our estimated rate of thinning, it is not surprising that 20 (16.6%) eyes that we studied did not have detectable grafts.

This study did not identify age, sex, race, type of ADD, position of ADD, prior conjunctival surgery, uveitis diagnosis, or dry eye diagnosis as a predictor for graft presence or absence. Our results lie in contrast to work from Chaku et al,¹⁹ who demonstrated that age was a significant risk factor for tube exposure. The different outcomes may be a result of the large number of pediatric glaucoma cases in the Chaku et al study, specifically in the tube exposure group, while our study had no pediatric patients.

In the present study, clinician estimation of graft thickness was poor (Pearson correlation, 0.198; P = .004). Pan et al¹⁷ recognized that it is difficult to assess GISC patch graft thickness using a slitlamp because of thick conjunctiva and the transparent nature of the GISC patch graft. Thus, AS-OCT imaging may be a useful tool to identify ADDs at risk for exposure.

We found 2 studies in the literature that address thinning of other materials used as patch grafts. Smith et al²⁰ reviewed 64 eyes that had received donor sclera, dura, or pericardium. Three instances of exposure—1 in each group—occurred after at least 2 years of follow-up. They also noted in 25% of the eyes (evenly distributed between the groups) that the tube was visible underneath intact conjunctiva, suggesting thinning or disappearance of these otherwise opaque graft materials. Raviv et al⁶ observed thinning without tube exposure of pericardial grafts in 5 of 44 eyes at a mean of 6 months after surgery assessed by visibility of the tube underneath intact conjunctiva. These incidences of thinning are in the range of what we are reporting.

Although it stands to reason that placement of a patch graft would reduce the likelihood of exposure, this has not been definitively demonstrated. One study in which surgeons did not use a patch graft had only 1 tube exposure 10 weeks after implantation in a cohort of 14 patients during a 3-year follow-up period,²¹ while another study had no tube exposure in a cohort of 128 ADD operations during a 6-month follow-up period.²² Other studies using either a technique in which the tube is placed underneath a partial thickness scleral flap²³ or is passed through a scleral needle track starting 4 to 6 mm behind the limbus (long-tunnel technique) report little to no exposure.^24,25

A randomized trial of a patch graft vs long-tunnel technique without a patch graft would be difficult to perform, because demonstrating a difference in prevalence of exposure of 5% (eg, 5% incidence of exposure in the patch-graft group and 10% incidence of exposure in the long-tunnel group) would require 435 patients in each group to achieve 80% power for detecting the difference between the group proportions. In the absence of definitive data, we recommend that efforts be made to decrease the likelihood of exposure either with a patch graft or with surgical techniques to bury the portion of the tube closest to the limbus in the sclera. GISC has advantages over other patch graft materials in terms of transparency and ease of use, but the literature does not allow us to conclude whether GISC differs from other materials in terms of preventing exposure.

One limitation of this study is that images were obtained and measured by more than 1 examiner. However, a reproducibility study of 9 patients (10 eyes) to determine whether the results were replicable was conducted. The study demonstrated that there was a high correlation between the measurements of graft thickness obtained by the 2 examiners (Pearson correlation coefficient, 0.84; P = .002). Another limitation of the study is that the examiners were not able to measure the graft at the same position for each patient. Using the AS-OCT, the examiner estimated 2 mm from the limbus. However, there was no method to ensure that the graft was measured at the same position for each patient. The examiners considered taking serial images and calculating a mean graft thickness measurement; however, given the time that it takes to acquire each image, it would have been difficult to ensure that the patient did not change positions during image acquisition. An additional caveat is that the ADD surgeries were not performed in a standard manner. The surgeon did not always place the graft on top of the tube in the same orientation, and there was likely variation in the initial thickness of the graft provided by the manufacturer. Finally, there may be some selection bias in the population because patients with prior tube exposure who underwent revision surgery were excluded. Therefore, graft thinning may be more frequent than reported in this study.

To our knowledge, this is the first study to evaluate thinning of the GISC patch graft after ADD surgery. Although the GISC patch graft is an attractive alternative to other forms of graft material for ADD surgery (eg, tutoplast, pericardium, and sclera) because of its cosmetic appeal, sterile nature, and ease of storage, its tendency to thin after surgery suggests that it is not the definitive solution for preventing tube exposure.

Accepted for Publication: June 12, 2017.

Corresponding Author: Henry D. Jampel, MD, Wilmer Eye Institute, Johns Hopkins Hospital, Maumenee B110, 600 N Wolfe St, Baltimore, MD 21287 (jampel@jhmi.edu).

Published Online: August 3, 2017. doi:10.1001/jamaophthalmol.2017.2628

Author Contributions: Ms de Luna and Dr Jampel had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis.

Study concept and design: Jampel.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: de Luna, Jampel.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: de Luna, Wang, Jampel.

Administrative, technical, or material support: de Luna, Jampel.

Supervision: Jampel.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.