Success and Failure Factors for Trabeculectomy in Glaucomatous Patients in Southwest China: A 325 Eyes Analysis

Purpose To evaluate the outcomes and elucidate the success and failure factors for trabeculectomy with mitomycin C (MMC) in Southwest Chinese patients. Methods A retrospective correlational study was conducted on the glaucomatous patients who underwent initial trabeculectomy with MMC in Southwest Hospital and had been followed up for 1-3 years. A complete success for surgery is dened as a postoperative intraocular pressure (IOP) > 5 and ≤ 21 mmHg and 20% reduction of IOP compared to preoperative, without IOP-lowering medications. A qualied success for surgery is dened as the abovementioned postoperative IOP with or without IOP-lowering medications. The primary outcomes were IOP, the number of IOP-lowering medications, and cumulative success rate. The secondary outcomes included best corrected visual acuity (BCVA), mean deviation (MD) of visual eld, major complications, and risk factors for surgical failure.


Conclusions
The quali ed success rate of trabeculectomy with MMC in glaucomatous patients in Southwest China is 92.0% at 1-year follow-up, and 77.7% at 3-year follow up. Hypertension, encapsulated bleb, and more preoperative topical medications are associated with surgical failure.

Background
Glaucoma is an irreversible optic neuropathy that considered to be one of the leading causes of blindness worldwide. More than 90% of patients are unaware of their vision loss and many risk factors are associated with physiopathology of glaucoma. Among them, intraocular pressure (IOP) remains the main modi able risk factor, and surgical treatment is the standard care for glaucomatous patient aiming to lower IOP if not controlled by pharmacological agents.
Nowadays, minimally invasive glaucoma surgeries are blooming [1,2]. Recent studies of the American Glaucoma Society have demonstrated a rise in the proportion of using tube shunts and a decline in the popularity of trabeculectomy [3,4]. Nevertheless, trabeculectomy still remains the most effective procedure 5 for glaucoma management in China since most patients suffer from primary angle-closure glaucoma (PACG). The success rate of trabeculectomy has been increasing along with innovation of intraoperative techniques such as wide application of anti brotics and placement of adjustable sutures.
It is believed that the postoperative interventions, for instance, removal of adjustable sutures or laser suture lysis, ocular massage, and bleb needling, have reduced complications and improved outcomes of trabeculectomy [6,7].
There are some studies report the e cacy and safety of trabeculectomy in Asian. The 5-year success rate was 87.3% according to the Japanese Collaborative Bleb-Related Infection Incidence and Treatment Study (CBIITS) [8]. A research in Taiwan described a cumulative quali ed success rate of 67.1% in a 10- year follow-up period [9]. A retrospective study on mitomycin C (MMC)-augmented trabeculectomy revealed that the quali ed success rate was 75% after a follow-up period of 21.8 months in Hong Kong [10]. However, studies on trabeculectomy in mainland Chinese patients are lacking. To evaluate the e ciency and safety for trabeculectomy, we retrospectively review 325 eyes of 261 glaucomatous patients who underwent trabeculectomy with MMC and elucidate the success and failure factors for trabeculectomy in Southwest Chinese patients.

Study design
The medical records of patients who were diagnosed as glaucoma and underwent trabeculectomy with MMC in Southwest Hospital of China were retrospectively reviewed from January 2016 to December 2018. This retrospective study was approved by the ethics committee of the Southwest Hospital (KY2020206) and adhered to the tenets of the Declaration of Helsinki.
The inclusion criteria were as follows. Patients were over 18 years old, diagnosed of glaucoma, and underwent initial trabeculectomy with MMC. The surgery includes trabeculectomy concomitantly performed with cataract surgery or anterior vitrectomy. Patients had no previous intraocular surgeries except laser peripheral iridotomy (LPI) and selective laser trabeculoplasty (SLT), and had been followed up for minimum 1 year. The exclusion criteria were patients of less than 18 years old, history of cataract surgery or trabeculectomy, and less than 1-year follow-up period.

Data collection
Preoperative data including the patients' demographic information, glaucoma type, IOP, number of IOPlowering medications, best corrected visual acuity (BCVA), mean deviation (MD) of visual eld, vertical cup to disc ratio, and surgery type were collected. IOP was measured by non-contact or rebound tonometer. At least 3 consecutive readings were recorded for each patient at each follow-up point. BCVA measurements were converted to logarithm of the minimum angle of resolution (logMAR) equivalents for the purpose of data analysis [11]. Postoperative data included BCVA, IOP, number of IOP-lowering medications, complications, and surgical interventions.
According to the World Glaucoma Association guidelines, a complete success for surgery is de ned as a postoperative IOP of > 5 and ≤ 21 mmHg and 20% reduction of IOP compared to preoperative without IOP-lowering medications. A quali ed success for surgery is de ned as the abovementioned postoperative IOP with or without IOP-lowering medications. The primary outcomes were IOP, the number of IOP-lowering medications, and cumulative success rate. The secondary outcomes included BCVA, MD of visual eld, major complications, and risk factors for surgical failure.

Surgical procedure
The surgeries were performed by highly skilled glaucoma specialists. Topical anesthesia or general anesthesia were tailored to the best experience for the patients and their speci c needs. After applying a corneal traction suture, a fornix-based conjunctival ap was created. The subconjunctival tissue was then dissected to expose the sclera in an appropriate size. After bipolar coagulation hemostasis and a rectangular partial-thickness scleral ap was made, a sponge soaked with MMC (0.4 mg/mL) was placed on the scleral bed and subconjunctival area for 2-5 min at the surgeon's discretion. The area was then irrigated thoroughly with balanced salt solution. A paracentesis was then performed to lower intraocular pressure and an anterior ostium was created, followed by peripheral iridectomy. After the scleral ap was sutured, the conjunctiva ap was closed separately to make a watertight closure. Prednisolone and antibiotics were administrated topically immediately post-surgery. Complications such as encapsulated bleb, shallow anterior chamber, hyphema, wound leak, malignant glaucoma, and choroidal effusion were closely observed. Personalized interventions such as removal of adjustable sutures, laser suture lysis, massage, bleb needling, and anti brotic injection were performed based on the surgeon's clinical judgement.

Statistical analysis
Statistical analysis was performed by SPSS software (version 20.0, IBM, USA). Continuous variables were presented as mean ± SD. T-test was used to compare the measurements of IOP, the number of topical medications, BCVA (logMAR), visual eld MD. Kaplan Meier survival analysis was applied to draw the survival curves and calculate success rates. Comparison of success rate between groups was performed using the Log rank test. Cox proportional hazards regression analysis was used to identify factors associated with surgical failure. A P ≤ 0.05 was considered statistically signi cant.

Eye characteristics
A total of 325 eyes of 261 patients were observed in the study, including 116 males (44.4%) and 145 females (55.6%). The mean follow-up time was 22 ± 12 months, ranging from 12 to 57 months. The average age was 59.3 ± 12.8 (19-90) years old, and 189 (58.2%) patients were over 60 years old. PACG was the main type of glaucoma, which accounted for 64.6% of patients. The mean preoperative IOP was 32.9 ± 12.0 mmHg, and the mean number of topical IOP-lowering medications before surgery was 3.0 ± 0.9. Of patients, 245 (75.4%) underwent trabeculectomy alone, and 80 (24.6%) received trabeculectomy combined with cataract surgery. The demographic characteristics of the patients and preoperative ocular characteristics are presented in Table 1. Information about visual eld MD was unavailable in 40 cases. CG, congenital glaucoma; BCVA, best corrected visual acuity; logMAR, logarithm of the minimal angle of resolution; IOP, intraocular pressure; SD, standard deviation; MD, mean deviation; dB, Decibels; C/D, vertical cup to disc ratio.

Surgical outcomes and success rate
The IOP and the number of topical IOP-lowering medications at each follow-up time are shown in Table 2.
The mean IOP and glaucoma medications were signi cantly decreased from 32.9 ± 12.0 mmHg to 16.4 ± 5.7 mmHg (P<0.0001) and 3.0 ± 0.9 to 0.9 ± 1.0 (P<0.0001), respectively, at 3-year follow-up. The IOP values and the number of topical medications were consistently lower than those observed at baseline throughout the follow-up period. We compared the surgical outcomes between PACG and POAG patients (Figure 2). Before the surgery, both the IOP values and the number of medications in PACG group were higher. However, after the surgery, both IOP and the number of glaucoma medications were comparable between PACG and POAG throughout 3-year follow-up. The 3-year complete success rates of PACG and POAG were 54.6%, 46.4%, while the quali ed success rates were 78.7%, 71.5%, respectively. The cumulative success rates of two groups were similar.
For PACG patients who underwent with trabeculectomy alone or phacotrabeculectomy, there was no signi cant difference in IOP control between the two surgical procedures. There was no signi cant difference with respect to IOP-lowering medications between the two surgical procedures at all follow up times except 1 month after surgery. The success rates of two surgical procedures were comparable ( Figure 3).

Visual acuity and visual eld test
The changes of BCVA (logMAR) and visual eld MD are shown in Table 3. The mean BCVA (logMAR) was 0.77 ± 0.94 before surgery and 0.67 ± 0.83 at the last follow-up. There was no signi cant difference between baseline and last follow-up (P=0.14). At the last follow-up, the mean visual eld MD was -16.80 ± 11.47 dB, which was slightly improved compared with that at baseline (P=0.04).  Table 4. Encapsulated bleb was treated with massage, needling, or anti brotic injection based on surgeon's clinical judgement.

Risk factors for surgical failure
The risk factors for surgical failure were investigated using Cox proportional hazard analysis (see Table  5). Hypertension (HR=1.904, P=0.011), encapsulated bleb (HR=2.756, P<0.001), and preoperative more topical medications (HR=2.475, P=0.008) were associated with a higher risk of surgical failure. Several lines of studies had evaluated the e ciency of trabeculectomy [12][13][14][15][16][17][18][19][20][21][22]. The success rate of trabeculectomy appeared to be lower in Asian than that of in Caucasians. A multicenter retrospective study conducted in the United Kingdom showed that the complete and quali ed success rates were 80% and 87% respectively at the 2-year follow-up [23]. Ng et al conducted a study on MMC-augmented trabeculectomy in Hong Kong and revealed that the complete and quali ed success rates were 47.9% and 75%, respectively, after a mean follow-up period of 21.8 months [10]. In the present study, the IOP decreased from 32.9 ± 12.0 mmHg to 16.4 ± 5.7 mmHg at 3-year follow up, with a mean 50.2% IOP reduction. The complete and quali ed success rates were 47.2% and 77.7% respectively at 3-year followup. The success rate is consistent with that of Hong Kong, but lower than that of European and American. This could be explained by the fact that the Chinese patients tend to show a more robust in ammatory response than Caucasians after trabeculectomy. This will lead to scarring and will therefore lower the success rate 24 . Nevertheless, it is di cult to compare the surgical results among different studies, due to differences in ethnicity, de nition of success, and follow-up periods. We assume that the differences in study design, intraocular surgical history, and glaucoma subtypes have contributed to the disparity in success rate among different studies.
There are limited studies to compare the surgical outcomes after trabeculectomy in PACG versus POAG. We suppose this is partly due to POAG is the main glaucoma type in European and American patients. While in China, PACG is the most common type of glaucoma. We nd that there is no signi cant difference in surgical outcomes between the two types of glaucoma. Although both types of glaucoma are characterized by aqueous humor out ow obstruction and progressive irreversible optic nerve degeneration, the pathophysiology of PACG is different from POAG. PACG is characterized by apposition of the peripheral iris against the trabecular meshwork and closure of an already narrow angle of the anterior chamber. While POAG is caused by increased resistance to aqueous out ow through the trabecular meshwork and the angle is open. For both types of glaucoma, the aim of trabeculectomy is to reconstruct the aqueous humor out ow pathway and will have similar outcomes to lower IOP.
It's known that trabeculectomy, cataract surgery, or combined surgeries of these two are the three options for PACG patients, depending on the patient's severity of glaucoma and visual compromise from a cataract. Previous studies have compared the effectiveness of these surgical modalities [15,[25][26][27]. It is noteworthy that cataract or clear lens removal has been advocated as initial treatment for PACG [28]. In China, the doctors prefer to perform cataract surgery only for PACG patients with peripheral angle synechiae less than 180 degrees. So, the patients who underwent solo cataract surgery were not included in this study. The results show that both IOP control and the success rates of trabeculectomy are comparable to those of phacotrabeculectomy, which indicate that cataract surgery has poor IOP-lowering e cacy in PACG patients with angle closed totally.
At the last follow-up, BCVA was slightly better than that at baseline. One possible reason is that some patients had acute attack of PACG before surgeries. The recorded preoperative vision is worse than their real vision. Another explanation is some patients underwent trabeculectomy together with cataract surgery which will result in better postoperative vision due to clear lens. The mean visual eld MD was improved at the last follow-up visit when compared with before surgeries. However, it's not conclusive that trabeculectomy can improve visual eld, because visual eld is di cult to restore once damaged.
Kashiwagi et al reported that visual function may deteriorate, despite effective control of IOP after trabeculectomy [29]. Studies have shown that IOP is not the speci c factor responsible for glaucomatous optic nerve damage, but rather some circulating and metabolic factors. We presume that visual function may be maintained when the IOP is controlled at the target level by surgery or topical medications [30].
The rate of postoperative encapsulated bleb was 7.7-12% in previous studies [21,23]. The incidence of postoperative encapsulated bleb in our study is 10.2%, similar to those reports. The incidences of shallow anterior chamber (3.1%) and hyphema (2.5%) were lower in our study. We think this is due to the retrospective nature of our study, which may lead to underestimation of these postoperative complications. In addition, interventions including removal of adjustable sutures or laser suture lysis were performed aforehand to reduce early postoperative complications. The low incidence of malignant glaucoma (1.2%) and choroidal effusion (0.6%) may be attributed to trabeculectomy combined with anterior vitrectomy for those high-risk patients, such as PACG with short axial length or several fundus diseases. Anterior vitrectomy can reduce these complications through balancing the pressure of anterior chamber and vitreous cavity.
In further, we did Cox proportional hazard plots and identi ed that hypertension, encapsulated bleb, and more preoperational topical medications were associated with high risk of surgical failure. To the best of our knowledge, this is the rst study to reveal the effect of hypertension on trabeculectomy. Previous studies had shown that hypertension may increase the risk of glaucoma because the common pathogenetic mechanisms in ciliary and renal tubular epithelia [31]. Patients with hypertension may have higher risk of intraoperative bleeding, which might result in localized in ammatory response and brous proliferation in the ltration zone [32]. Long-term topical IOP-lowering medications has been shown to induce subclinical ocular in ammation with proliferation of broblasts, lymphocytes, and macrophages. This will signi cantly lower trabeculectomy success rate [33]. Encapsulated bleb is the main cause of failure in glaucoma ltering surgery. For a successful ltration surgery, an appropriate amount of ltration might be necessary, especially in the early stage of post-operation. Ocular massage or needling procedures are necessary when the adhesion between the sclera and the conjunctiva are rm for those with encapsulated bleb[8].
There are also some limitations in the present study. First, the study is designed as a retrospective study.
Some key statistics cannot be measured. Researchers cannot control exposure or outcome assessments, but instead must rely on others for accurate recordkeeping. Second, patients are reviewed from a single tertiary referral center, which would limit the application of the data to other centers. Third, the operations are performed by three skilled surgeons. Different surgeons have their inherent deviations and might affect the surgical outcomes. Fourth, we exclude the patients who are followed up for less than 1 year, which would lead to selective bias of the cohort.

Conclusion
Our study reports the clinical outcomes of trabeculectomy with MMC in glaucomatous patients in Southwest China. Based on the criteria of 5 < IOP ≤ 21 mmHg and 20% reduction of preoperative IOP, the complete and quali ed success rates are 47.2% and 77.7% at 3-year follow up. This data is consistent with previous studies in Chinese population [9,10]. There are no signi cant differences in surgical outcomes between PACG and POAG. For PACG patients with peripheral angle synechiae more than 180 degrees, the success rates of trabeculectomy and phacotrabeculectomy are comparable. Hypertension, encapsulated bleb, and more topical preoperative medications are associated with surgical failure.
Therefore, glaucoma professionals should choose the appropriate glaucoma surgical approach based on the patient's disease condition. Intensive proactive preoperative and postoperative management is essential to prevent the deterioration of visual function. Figure 1 Kaplan-Meier survival plot of the cumulative probability of success until the last follow-up visit based on the criteria of complete success and quali ed success de nitions.

Figure 2
The mean IOP (A) and the number of glaucoma medications (B) at baseline and each follow-up visit of the PACG and POAG patients. There were no signi cant differences in IOP control and need for glaucoma medications between the PACG group and POAG group after surgery. Error bars represent SD. The complete success rate (C) and quali ed success rate (D) of the PACG and POAG patients are displayed by the Kaplan-Meier survival curve plot. The cumulative success rates of two groups were similar.