Predictive value of early postoperative IOP and bleb morphology in Mitomycin-C augmented trabeculectomy

Background: To determine the predictive value of postoperative bleb morphological features and intraocular pressure (IOP) on the success rate of trabeculectomy. Methods: In this prospective interventional case series, we analyzed for one year 80 consecutive primary open angle glaucoma patients who underwent mitomycin-augmented trabeculectomy. Bleb morphology was scored using the Indiana bleb appearance grading scale (IBAGS). Success was defined as IOP ≤15 mmHg at 12 months. We applied a multivariable regression analysis and determined the area under the receiver operating characteristic curve (AUC). Results: The mean age of participants was 62±12.3 years in the success and 63.2±16.3 years in the failure group (P= 0.430) with equal gender distribution (P=0.911). IOPs on day 1, 7 and 30 were similar in both (P= 0.193, 0.639, and 0.238, respectively.) The AUC of IOP at day 1, day 7 and 30 for predicting a successful outcome was 0.355, 0.452, and 0.80, respectively. The AUC for bleb morphology parameters of bleb height, extension, and vascularization, on day 14 were 0.368, 0.408, and 0.549, respectively. Values for day 30 were 0.428, 0.563, and 0.654. IOP change from day 1 to day 30 was a good predictor of failure (AUC=0.838, 95% CI: 0.704 to 0.971) with a change of more than 3 mmHg predicting failure with a sensitivity of 82.5% (95% CI: 68 to 91%) and a specificity of 87.5% (95% CI: 53 to 98%). Conclusions: IOP on day 30 had a fair to good accuracy while bleb features failed to predict success except bleb vascularity that had a poor to fair accuracy. An IOP increase more than 3 mmHg during the first 30 days was a good predictor of failure.


Introduction
For several reasons, the number of trabeculectomies has sharply declined in the last decades in many developed countries 1,2 . More surgeons are now trained in microincisional glaucoma surgeries 3,4 which can be performed sooner due to a superior risk profile 5 ; laser trabeculoplasty devices have become more affordable and are used as the first line of treatment 6,7 ; and prostaglandin analogues are now available as generic, less expensive eye medications, further reducing the number of trabeculectomies 6 . However, in economies categorized as developing by the International Monetary Fund 8 and countries not in the upper quartile of the Human Development Index 9 , trabeculectomies remain the leading surgeries for moderate and severe glaucomas 10,11 . Although trabeculectomies have a considerable complication rate 12 , they remain a common primary procedure for advanced glaucoma because they lower intraocular pressure (IOP) effectively and are not as expensive as epibulbar glaucoma drainage devices 13 , which involve a comparable surgical effort and risk profile.
Careful patient selection and postoperative management are key to making trabeculectomies successful. Recognized risk factors are, among others, younger age, a higher baseline IOP, and inflammation 14,15 . While meticulous preoperative planning, intraoperative use of antifibrotics, and the surgical technique impact the outcome, the postoperative period plays an equally important role 16,17 including adjustment of the frequency of postoperative visits, adjustment of steroid frequency, release or lasering of scleral flap sutures, bleb massage, and bleb needling. Yet, even with the most personalized postoperative care, the individual wound healing and tissue remodeling make predicting a successful outcome difficult. Several recent studies have tried to capture the specific wound healing by correlating the early postoperative IOP or the bleb morphology with long-term outcomes, and found a low early postoperative IOP produces favorable long-term outcomes [18][19][20][21] . In comparison, bleb grading systems are more complex, less objective but hold promise to describe data otherwise difficult to quantify; they have not been widely adopted 22,23 .
The purpose of our study was to validate in our patient population the accuracy of the early postoperative IOP and bleb morphology (Indiana Bleb Appearance Grading Scale (IBAGS)) as a test to predict a failure of trabeculectomy at 12 months. In contrast to prior studies, we also examined the role of an IOP increase during the first 30 postoperative days in predicting such failure.

Study design
Eighty eyes of 80 consecutive patients diagnosed with primary open angle glaucoma were included in this prospective observational study. In patients who had trabeculectomies in both eyes, the eye was chosen by randomization 24 . The study was performed at the glaucoma clinic of the Labbafinejad Medical Center of the Shahid Beheshti University of Medical Sciences, Tehran, Iran, from September 2013 to March 2015. It was approved by the ethics committee and the institutional review board at the Ophthalmic Research Center (protocol number: IR.SBMU. ORC.REC.1393.2) and followed the tenets of the Declaration of Helsinki. Informed written consent was obtained from each participant. Inclusion criteria were equal to or above 30 years of age and progressive primary open angle glaucoma (glaucomatous optic neuropathy, retinal nerve fiber layer loss and corresponding visual field defect on standard automated perimetry) that could not be controlled medically. Exclusion criteria consisted of prior ocular surgery, manipulation of the conjunctiva, a necessity for combining trabeculectomy with cataract extraction, ocular or systemic comorbidities that could affect the procedure and study outcomes including immunodeficiency, connective tissue disease, and uncontrolled diabetes. At baseline, all patients underwent a comprehensive ophthalmic examination including determination of best-corrected visual acuity (BCVA), slit lamp examination, Goldmann applanation tonometry, gonioscopy, and fundus examination.
Surgical technique and postoperative management All trabeculectomies were performed by two surgeons in equal numbers (HE and MP), as described in the following 25 . After administration of intravenous sedation, a peribulbar block was placed using 2 milliliters (ml) of 2% lidocaine (Lignidic 2%, Caspian Tamin Pharmaceutical Co., Rasht, Iran). Following lid speculum insertion and cul de sacs irrigation with povidoneiodine and normal saline solution, a 7-0 silk traction suture was placed through the superior peripheral cornea. A superior peritomy was fashioned over 1-½ clock hours, and Tenon's capsule was dissected with Westcott scissors. A 3.0 × 4.0 mm trapezoidal half-thickness scleral flap was fashioned using a crescent knife followed by lamellar dissection of the flap 1mm into the clear

Amendments from Version 1
The reviewer's suggestions helped us improve this manuscript by including the following changes: • The discrepancy between early postoperative visits (1 week versus 2 weeks) has been corrected. As we have confirmed by going back to the original SPSS files, intraocular pressures were checked at the 1-week postoperative visit mistakenly listed as week 2 in Table 2. The discrepancy was corrected throughout the manuscript.
• We clarified the criteria when sutures were released, bleb needling performed and that prostaglandin analogs were avoided during the postoperative period. Releasable sutures were removed within the first month of the surgery based on the IOP and the surgeons' assessment of bleb function. Highly vascularized as well as contracting blebs were indications for needling. The same medications that were effective and safe before the surgery were resumed if necessary except prostaglandin analogs which were avoided.
• We updated our conclusion in the Abstract by adding the importance of a vascularized bleb in predicting the failure and the need for a prompt intervention depending on the vascularization.
• We also now state in the Abstract the exact average of the age of analyzed patients and p-value for gender differences: the mean age of the study participants was 62±12.3 years in the success and 63.2±16.3 years in the failure group (P= 0.430) with equal gender distribution (P=0.911).

REVISED
cornea. After creating the scleral flap, pieces of 0.02% MMC (Mitomycin-C, Kyowa, Kogyo Company, Tokyo, Japan) soaked cellulose sponges were placed under the Tenon and conjunctiva over the scleral flap for 3 minutes. After removing the sponges and vigorously irrigating the area with 50 ml of balanced salt solution (BSS), the anterior chamber was entered with a small keratome, and a block of clear cornea was removed with a 1 mm Kelly punch followed by a peripheral iridectomy using Vannas scissors. The scleral flap was then secured using two 10-0 nylon releasable sutures. The anterior chamber was formed by injection of BSS through a sideport, and the filtration was titrated by adjusting the tension of the scleral flap sutures. The conjunctiva was closed using 10-0 nylon wing sutures. Cefazolin 50 mg (Cefazolin 500, Exir Pharmaceutical Co, Tehran, Iran) and betamethasone 2 mg (Betazone, Caspian Tamin Pharmaceutical CO., Rasht, Iran) were injected subconjunctivally into the inferior fornix, and the eye was patched.
The postoperative regimen consisted of chloramphenicol 0.5% eye drops (Sina Darou Lab. Co., Tehran, Iran) 4 times a day for 1 week and betamethasone 0.1% eye drops (Sina Darou Lab. Co., Tehran, Iran) 6 times a day, which was tapered over 8 to 12 weeks depending on the inflammation. The releasable sutures were removed as needed after 72 hours depending on bleb appearance to achieve IOPs near 10 mmHg. In most cases, the releasable sutures were removed within 4 weeks after the surgery. One removed first and if the IOP was not within the target range the other one was released during the follow up one week later. Needling was done during the postoperative period for impending failure from highly vascularized and contracting blebs. Thirty minutes after injecting 0.1 ml of 0.02% Mitomycin-C into the bleb-adjacent subtenon space a 27-gauge needle was used to reform the bleb and dissect adhesions at the slit lamp.
Before the surgery patients used glaucoma medications from all four major classes, including beta-blockers, prostaglandin analogous, alpha-2 agonists, and carbonic anhydrase inhibitors. Postoperatively, patients received the same drops that had proved effective and safe before surgery if the target IOP was not achieved. The only exception was prostaglandins analogous that we tried to avoid during the first three postoperative months.
We saw our patients on a weekly basis for one month and then monthly for 12 months. At each postoperative visit, BCVA, IOP, bleb morphology, glaucoma medications, and complications were noted. A slit lamp photograph of the blebs was obtained at each visit (Haag-Streit BX 900 slit lamp, Haag-Streit Diagnostics, Köniz, Switzerland). Two glaucoma surgeons, masked to surgeon and patient, graded the bleb images on the Indiana Bleb Appearance Grading Scale (IBAGS). The IBAGS variables included the bleb height (H0-H3), extent (E0-E3), vascularity (V0-V4), and leakage (S0-S2) 23 . A trabeculectomy was considered successful if IOP was equal or less than 15 mmHg and more than 5 mmHg with or without glaucoma medications 18 .

Results
A total of 80 patients were enrolled in this study. This sample size and ratio of success and failure provided an above 90% power to detect an AUC difference of 0.5 26 . Forty-four eyes would have provided a power of 80% to detect an IOP change of 3 mmHg while 80 eyes allowed for a 97% power. The mean age of study participants was 62.3±13.1 years, and 51.3% of the patients were male (P=0.911). The characteristics of participating patients are presented in Table 1. The procedure was considered successful in 64 patients (80%) at 12 months follow-up.
Preoperative medications in the no success group were 1.4±1.9 and in the success group were 1.2±1.6 (p=0.785), while the postoperative medications were 1.4 ± 0.5 in the no success group and 1.0±0.0 in the success group (p=0.04). There were no significant differences between the success and failure groups in terms of age, the number of preoperative glaucoma medications, preoperative IOP, visual field mean deviation or vertical cup-disc ratios ( Table 1). The postoperative IOP was not different between these groups at any time (Table 2) because an IOP of lower than 5 mmHg was also counted as no success in accordance to

Discussion
Trabeculectomy remains a common surgery to lower IOP in glaucoma patients and has a long history. In 1958, Grant was the first one to describe a guarded filtering procedure in enucleated eyes with subsequent facility measurements which he termed trabeculectomy 27 . Sugar followed by performing the first such trabeculectomy in patients in 1961 28 and Cairns finally popularized the term and technique by providing detailed drawings and figures 29 . The procedure was eventually made more effective by adding Mitomycin-C as an antifibrotic 30 and somewhat titratable by using sutures that can be lasered or manually released 31,32 . The variable success rate of this procedure can be improved by careful patient selection, surgical technique 33 , and postoperative management 15 . Identifying patients at risk is critical because a timely intervention can improve the outcomes of trabeculectomy 16 ; preoperative factors including high baseline IOP, number of glaucoma medications, young age, aphakia or pseudophakia and prior surgery with conjunctival scarring are known risk factors for failure 33 . The success and failure groups in our study have similar baseline IOP, medications, visual field loss and optic nerve damage indicating that any risk factors related to those were relatively equal in both.
Several recent studies found a low early postoperative IOP produces favorable long-term outcomes 18-21,34 but also causes more frequent and more severe complications that include a shallow anterior chamber, choroidal effusion, hypotony maculopathy, and vision loss 35 . Although we did observe that the IOP change during the first month was a good predictor of failure, we found only a weak relationship between absolute IOP at day 1, 7, and 30 and long term success. This is similar to Polikoff et al. 36 who also reported that early post-trabeculectomy IOP is a poor predictor of success at one year follow-up. The accuracy of a test depends on how well it separates groups under study into those with and without the issue in question. Accuracy is measured by the area under the ROC curve and is commonly classified into .90-1 = excellent, .80-.90 = good, .70-.80 = fair, .60-.70 = poor and .50-.60 = fail 26 . In our study, the AUC for IOP was the highest on day 30 with 0.8, making it a fair to good predictor 26 . Early post-trabeculectomy IOP is affected by various factors that include reduced aqueous production, inflammation, the amount of TGF-beta and other growth factors 37 , and breakdown of the blood-aqueous barrier, over filtration, and choroidal detachment. In our study, the IOP appears to summarize these and other factors reasonably well to a single variable that can be used as a predictor.
An advantage of this study is that approximately 75% more patients participated than in prior bleb grading studies 22,23,38,39 . Different from those, we also examined the change in IOP during the first 30 days and found that it served as a good predictor of success that performs better than the IOP itself. The cutoff for an at-risk IOP elevation was only 3 mmHg. Our study was not designed to discover causality, however. It is likely that this finding identifies patients at risk of failure who experienced an IOP increase despite an intensified treatment. It could also indicate that these were patients who received a suboptimal postoperative management. An elevated IOP will increase the stretch of the bleb wall and incite a reactive fibrosis. Stretch is a well-established activator of fibroblast proliferation and fibrosis 40,41 . This may lead to a cycle of auto-enhancing IOP increase with further increase of stretch and fibrosis that again elevates pressure as the bleb wall becomes thicker and the bleb contracts. In this model, IOP drives the morphological changes of the bleb. This order matches our AUC findings better (the AUC was better for IOP change than for bleb morphology) than a primary morphological change that is followed by increased IOP. Our observation is supportive of the practice to inject or needle with postoperative 5-fluorouracil or Mitomycin-C to reduce the fibrosis 42,43 .
Like IOP, a bleb's morphology is a reflection of a patient's specific wound healing that is directly observable but more difficult to quantify. The development of bleb grading scales was an attempt to approach this problem systematically 22,23 . While grading systems can be useful teaching tools of bleb features, its inherently subjective nature and categorical data weaken the predictive power. Accordingly, we found only low AUC values for bleb morphology that ranged from 0.368 to 0.549 on day 14 and would be interpreted as a test failure. The day 30 features would be interpreted only slightly better, as poor.
Of the variables bleb height, extension and vascularization, vascularization had the highest accuracy. These findings match Picht et al. 44 who concluded an early vascularization indicates a poor

Supplementary Table 1: Intraocular pressure at each pre-and postoperative visit when success included IOPs below 5 mmHg without complications.
Click here to access the data.
prognosis with higher IOP at 12 months and may support use of anti-VEGF agents 45,46 . The above reasons and recent reports of ab interno trabeculectomy results with a reduced complication rate, considerable range of glaucoma indications, and results that can rival traditional glaucoma surgery 47-55 suggest a conventional trabeculectomy with Mitomycin-C may be more appropriately used as a secondary procedure when microincisional surgical modalities are available or only be used in advanced glaucoma with a rather low IOP target.
In conclusion, we found that the IOP change during the first month, rather than the IOP at each visit, was a good predictor of failure. The bleb morphological features did not predict failure except for bleb vascularity which performed poorly to fairly. These finding highlight how important it is to carefully observe for IOP and bleb vascularity changes and to intervene swiftly if necessary. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

2.
3. The study is carefully done and the topic is interesting for clinicians.

Comments/remarks:
You write the exact number of preoperative and postoperative medications the patients used. As there are differences in the IOP-lowering effects of the different types of these drugs, it would be interesting to see what type of IOP-lowering drugs the patients applied and if there was a difference in the type of drugs used pre-and postoperatively.
You state that you used postoperative needling with Mitomycine C, but you do not clarify when exactly ("impending failure from a contracting bleb" is not specific). Please give a comment in which patients you used the needling and add also in how many patients a needling was necessary.
In the "Conclusions" of the abstract you state that "the bleb features failed to predict a successful 3. In the "Conclusions" of the abstract you state that "the bleb features failed to predict a successful outcome" while in the "Conclusions" at the end of the manuscript you say that an augmented vascularity in the bleb should be carefully observed. Please add this sentence in the "Conclusions" of the abstract.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 01 Dec 2017 , University of Pittsburgh Medical Center, USA

Nils Loewen
The study is carefully done and the topic is interesting for clinicians. Comments/remarks: You write the exact number of preoperative and postoperative medications the patients used. As 1. there are differences in the IOP-lowering effects of the different types of these drugs, it would be interesting to see what type of IOP-lowering drugs the patients applied and if there was a difference in the type of drugs used pre-and postoperatively.
We very much appreciate the reviewer's time and care reviewing our manuscript.

Authors:
Thank you for allowing us to clarify this point. We now state in Methods: "Before the surgery patients used glaucoma medications from all four major classes, including beta-blockers, prostaglandin analogous, alpha-2 agonists, and carbonic anhydrase inhibitors. Postoperatively, patients received the same drops that had proved effective and safe before surgery if the target IOP was not achieved. The only exception was prostaglandins analogous that we tried to avoid during the first three postoperative months." You state that you used postoperative needling with Mitomycin C, but you do not clarify when 2. exactly ("impending failure from a contracting bleb" is not specific). Please give a comment in which patients you used the needling and add also in how many patients a needling was 1.

2.
which patients you used the needling and add also in how many patients a needling was necessary.
Thank you for allowing us to clarify this point. We needled only highly vascularized Authors: and contracting blebs when the IOP was uncontrolled. We now state: "A total of ten eyes (12.5%) underwent needling, including two eyes in the success group and eight eyes in the failure group (P=0.001)." In the "Conclusions" of the abstract you state that "the bleb features failed to predict a successful 3. outcome" while in the "Conclusions" at the end of the manuscript you say that an augmented vascularity in the bleb should be carefully observed. Please add this sentence in the "Conclusions" of the abstract. This is a good point. We have modified the Conclusions in the Abstract based on the Authors: reviewer's comments: "Conclusions: The postoperative IOP on day 30 had a fair to good accuracy while the bleb features failed to predict a successful outcome except for bleb vascularity that predicted success with poor to fair accuracy. An IOP increase by as little as more than 3 mmHg during the first 30 days was a good predictor of failure. It is important to carefully observe for IOP changes and bleb vascularity to intervene if necessary." No competing interests were disclosed. This is a carefully done study containing interesting results which merit indexing. It is very useful and can provide a proof for ophthalmologist to decide how long the patient should be re-visit postoperation according to IOP change. For the benefit of the reader, however, a number of points need clarifying and certain statements require further justification. A few minor revisions are list below.
There are mistakes in the study results and we listed some examples in the following and you should revised it in your paper carefully and discreetly.
1.1 In the article, they describe that collect IOP outcomes on Day 7, but it was Week 2 in table 2. In the results part of Line 3 Page 1 (All P≥0.2), but it was 0.193 in day 1 in table 2.
1.2 IOP change more than 3mmHg in Page 1, but it was less than or equal than 3 mmHg in Page 6.
There are some misunderstanding in the article. The releasable sutures were removed as needed after 72 hours depending on bleb appearance to achieve IOPs near 10 mmHg. (Page 3, Marked in yellow and underline). In what time the releasable suture were removed for most of them (for example 7-10 days or something else for 70%(or ?%) patient), and you removed the two releasable suture for one time or two time respectively. If you removed the suture for two times, how much time will pass between them.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. on day 1, 7 and 30 (P= 0.193, 0.639, and 0.238, respectively.)" IOP change more than 3mmHg in Page 1, but it was less than or equal than 3 mmHg in Page 1.2 6.
That is correct but there is no real contradiction. The reason for this is that we talk Authors: about that a change in IOP less than or equal to 3 can predict the at month 12. In contrast, success we talk about when a change of more than 3 mmHg was observed. failure There are some misunderstanding in the article.

2.
The first sentence in the results of page 1. (Age and gender were equally distributed 2.1 (62.3±13.1 years, P=0.911).) Thank you for pointing this out. We changed the sentence "Age and gender were Authors: equally distributed" to "The mean age of the study participants was 62±12.3 years in the success and 63.2±16.3 years in the failure group (P= 0.430) with equal gender distribution (P=0.911)".
The anterior chamber was formed by injection of BSS a paracentesis, and the filtration was 2.2 titrated by adjusting the tension of the scleral flap sutures.
We apologize and have changed the sentence to state "The anterior chamber was Authors: formed by injection of BSS a paracentesis, and the filtration was titrated by adjusting the tension of the scleral flap sutures "to "The anterior chamber was formed by injection of BSS through a sideport, and the filtration was titrated by adjusting the tension of the scleral flap sutures".
The releasable sutures were removed as needed after 72 hours depending on bleb appearance 3. to achieve IOPs near 10 mmHg. (Page 3, Marked in yellow and underline). In what time the releasable suture were removed for most of them (for example 7-10 days or something else for 70%(or ?%) patient), and you removed the two releasable suture for one time or two time respectively. If you removed the suture for two times, how much time will pass between them.
We changed Methods to state: "In most cases, the releasable sutures were removed Authors: within 4 weeks after the surgery. One removed first and if the IOP was not within the target range the other one was released during at the follow up one week later. Needling was done during the postoperative period for impending failure from highly vascularized and contracting blebs".
No competing interests were disclosed.

Competing Interests:
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage