Long-Term Surgical Outcomes of Glaucoma Drainage Implants in Eyes with Preoperative Intraocular Pressure Less than 19 mmHg

Background This retrospective review reports on patients who underwent glaucoma drainage implant (GDI) surgery and had baseline intraocular pressure (IOP) of ≤18 mmHg with at least one year of follow-up. Methods Clinical data of 67 eyes of 67 patients were collected from patients' charts, and the outcomes of GDI were evaluated until 7 years. GDI failure was defined as IOP reduction of less than 20% from the baseline at two consecutive visits three months after surgery, decline to no light perception, or if additional glaucoma surgery was performed. Results The average age was 65.9 ± 13.2 years. Most cases were male (52.2%), White (53.7%), and had primary open-angle glaucoma (62.7%). Forty-four eyes had prior glaucoma surgery (68.6%) and 46 (68.6%) had severe glaucoma. Though postoperative (postop) IOP changes were insignificant, the average postop number of medications dropped from 2.4 ± 1.4 to 1.9 ± 1.2 medications two years after surgery (p = 0.0451). Postop complications (23.9%) included GDI exposure (7.5%), inflammation (4.5%), shallow anterior chamber (4.5%), and strabismus (1.5%). Hypotony was observed in 4 eyes (5.9%), none of which developed hypotony maculopathy. The cumulative one-year failure rate was 56.7%, most of which were due to failure to lower IOP. Conclusion In patients with baseline IOP ≤18 mmHg who had GDI surgery, though the change in IOP was not statistically significant, the number of medications dropped and visual field progression slowed in a subset of patients with adequate perimetric data. Due to a relatively high rate of complications and limited effectiveness in lowering IOP, GDI should be cautiously used in these eyes.


Introduction
Glaucoma is a progressive optic neurodegenerative disease that causes visual feld (VF) loss and can lead to signifcant visual impairment [1].When the patient presents with glaucomatous optic nerve damage and preoperative (preop) intraocular pressure (IOP) < 21 mmHg, managing the condition becomes challenging.In some of these cases, the VF continues to worsen despite clinically controlled IOP.Te exact mechanism of disease progression in these cases is not entirely understood.However, several factors, including systemic hypotension, sleep apnea, vascular pathologies, and IOP fuctuations, have been implicated [2][3][4][5].
Despite signifcant advancements in glaucoma treatment, IOP reduction remains the primary therapeutic approach in such cases.Tis reduction is achieved by using hypotensive glaucoma medications or undergoing glaucoma surgeries.Trabeculectomy, the gold-standard surgical method for glaucoma management, has shown clinical effectiveness in slowing down disease progression in eyes with low to normal baseline IOP in several studies [6][7][8][9][10].However, this procedure is not without complications and these eyes are prone to postoperative complications, such as hypotony maculopathy and choroidal efusion [11,12].
In recent years, glaucoma specialists have increasingly turned to glaucoma drainage implant (GDI) surgery as an alternative option for controlling disease progression in refractory glaucoma [13].Similar to trabeculectomy, GDI surgeries aim to lower IOP by providing an alternative outfow for aqueous humor.Studies, like the primary tube versus trabeculectomy (PTVT) study, have shown that GDI surgeries carry a lower risk of postoperative complications compared to trabeculectomy (29% vs. 41%) [14].However, GDIshave a higher reported 1-year failure rate compared to the trabeculectomy group (17.3% vs. 7.9%) [15,16].
Despite the potential advantages of GDI surgeries, there is a lack of comprehensive data on their efectiveness in cases with preop IOP ≤ 18 mmHg.Tis study aims to investigate the long-term outcomes of GDI surgery in managing glaucoma eyes with preop IOP ≤ 18 mmHg by providing clinical and functional data.Herein, we provide insights into the surgical efectiveness and safety of GDIs in this specifc patient population from a tertiary care center, which will inform treatment strategies and clinical decision-making in this population.

Materials and Methods
Tis retrospective study reportson adult glaucoma patients who underwent GDI surgery at the glaucoma service of Wills Eye Hospital (a tertiary care center in Philadelphia, USA) between January 1, 2005, and January 1, 2021.Sixtyseven eyes of 67 patients who had an average preop IOP ≤ 18 mmHg, regardless of the number of hypotensive glaucoma medications, were included in the analysis.Te patients were required to have at least one year of follow-up to be included in the study.Teir clinical data were recorded up to 7 years after surgery.Patients with insufcient pre-and postoperative (postop) data or coexisting ocular conditions that could signifcantly afect visual acuity, IOP, or glaucoma management were excluded from the study.Te following demographic and clinical data were collected: age, sex, race/ethnicity, medical history, medications, surgical history, glaucoma type, and stage, best-corrected visual acuity (BCVA), IOP, lens status, type of GDI, VF data, surgical complications, and need for reoperation.Glaucoma severity was assessed using the Hodapp-Parrish-Anderson criteria [17].For eyes with missing VF data in the electronic medical records system or picture archiving and communication system (PACS), documented judgments of the physicians based on the criteria were recorded as disease severity.GDIs were considered to have failed if there was a need for IOPlowering glaucoma reoperation and/or the occurrence of failure criteria.Tese criteria included progression to no light perception (NLP), less than a 20% IOP reduction, and postop IOP of 18 mmHg or higher at two consecutive visits three months after surgery.Te patients' VF mean deviation/defect (MD) were also collected, and the reliability cutof for false positive and false negative was 33% [18].).Categorical variables were expressed as the frequency and percentage (%) of the total, whereas continuous variables were expressed as average ± standard deviation (SD) in the text and average ± 95% confdence interval in the graphs.Statistical analysis was performed using R Statistical software (version 4.0.5;R Foundation for Statistical Computing, Vienna, Austria) and GraphPad Prism for Windows (version 9.3.1)(GraphPad, La Jolla, CA, USA).Te Kolmogorov-Smirnov test was used to identify the normality of the distribution.For numerical variables, t-test, Wilcoxon rank-sum, or Wilcoxon signedrank tests were employed.For categorical variables, χ 2 or or Fisher exact tests were utilized.Linear and binary logistic regression models wereto identify independent predictors of one-year failure.p values less than 0.05 were considered signifcant.

Institutional Review Board
Approval.Tis retrospective cohort study was approved by the Institutional Review Board (IRB) at Wills Eye Hospital (Philadelphia, Pennsylvania, USA) and included patients seen by the Glaucoma Service.Te research was conducted in accordance with the Health Insurance Portability and Accountability Act of 1996 and adhered to the tenets of the Declaration of Helsinki.

Outcomes.
Te average preop logMAR BCVA of all eyes in this study was 0.9 ± 0.7 [20/150] and ranged from 0 [20/ 20] to 2.7 (light perception).Te follow-up logMAR BCVA ranged from 0 to 3 (NLP), and no signifcant change was observed in the average logMAR BCVA at each study point 2 Journal of Ophthalmology compared to the preop visit (p � 0.1502).One progressed to NLP three months after the surgery.Figure 1 illustrates the trend in pre-and postoperative logMAR BC VA.Te average preop IOP was 14.6 ± 3.3 mmHg, ranging from 10 to 18 mmHg.No signifcant diference was found between the average pre-and postop IOP(p � 0.0946) (Figure 2).Almost one-third (32.8%, n = 22) of the eyes experienced IOP > 18 mmHg at least once during the followup period after the frst three months (Figure 2).No signifcant diference was found between the average or median IOPs of valved and nonvalved GDI groups at each study point (p > 0.05).Figure 2 demonstrates the trend in postop IOP changes during the follow-up period.Te average number of preop prescribed medications was 2.4 ± 1.4 and signifcantly decreased to 1.9 ± 1.2 after two years of followup at a p � 0.0451 (Figure 2).
Among the 67 included eyes, 35 had preop perimetry records in theEMR or PACS, and their average MDs were 15.6 ± 8 dB (n = 27) and −16.3 ± 8.5 dB (n = 8) as measured by Octopus G TOP and Humphrey SITA 24-2 perimetry, respectively.Te average MDs signifcantly worsened within 2 years before surgery; however, no signifcant diference was observed in the subsequent follow-up tests.In 16 eyes with at least three visual felds to calculate pre-and postop MD slopes (only G TOP), mean defect slopes were calculated using all available VF data points.Te average slopes signifcantly improved from −2.2 ± 3.1 to −0.2 ± 0.7 dB/year after the surgery at p � 0.0176 (Figure 3).Tis improvement was observed in 81.2% (n = 13) of these 16 eyes.4

Journal of Ophthalmology
Te cumulative one-year failure rate was 56.7%.GDI failed to control the IOP in 52.3% of the eyes, 2 eyes (2.9%) needed further glaucoma surgery to control IOP, and 1 eye progressed to NLP (1.5%) within the frst year (Figure 4).Tough only lower age was signifcantly correlated with a higher failure rate in the univariable analysis (β = −0.258,R 2 � 0.066, p � 0.034), no signifcant independent predictor of one-year failure was found in the multivariable analysis.

Discussion
In this retrospective cohort study, we evaluated the longterm clinical outcomes of GDI surgeries in eyes with preoperative IOP less than 19 mmHg.To the best of our knowledge, this is the frst report on the use of GDIs to manage glaucoma in eyes with the low baseline IOP.
Several previous studies have reported reduced visual feld progression following a 30% reduction of IOP [9,11,12,19,20].Notably, the Collaborative Normal-Tension Glaucoma Study group demonstrated that lowering IOP, either through medications or trabeculectomy, signifcantly slowed down the progression rate of visual feld loss in a dose-dependent manner [21].Similarly, Naito et al. reported a signifcant enhancement of mean deviation (MD) slope following trabeculectomy surgery in eyes with preoperative IOP less than 15 mmHg [22].In our study, we observed that GDI surgery did not have a signifcant efect on IOP reduction, unlike the previous reports on trabeculectomy.However, GDI surgery demonstrated signifcant efectiveness by preserving visual acuity, reducing the number of medications, and slowing down VF loss progression in a subset of patients with adequate data.
In a subanalysis of the PTVT study, Gedde et al. [23][24][25] compared the outcomes of these two interventions on patients with the baseline IOP <21 mmHg.In their study, GDI failed to control patients' IOP in 63% of the cases after three years.In contrast, our study observed a cumulative one-year failure rate of 56.7%, and the failure rate increased to 86.7% after three years.Tis diference in failure rates may be attributed to the diferent baseline characteristics of the study populations, with our study's patients having a signifcantly lower average baseline IOP compared to theirs.
Regarding postoperative complications, we observed a cumulative complication rate of 23.9% and an overall reoperation rate of 19.4%.In comparison, the PTVT study reported postoperative complications as high as 29% and 34% after one and three years in the GDI group, respectively, compared to 41% and 48% in the trabeculectomy group [16].However, our study had a signifcantly higher rate of reoperations due to complications, particularly GDI exposure, which was observed in 7.5% of eyes and recurred in three eyes at subsequent follow-up visits.Tese diferences in complication rates may be related to the unique characteristics of our study population, which consisted of more advanced glaucoma cases with a history of prior hypotensive interventions.
One of the main limitations of our study is the limited number of repetitive visual feld tests, which allowed us to evaluate only 16 MD slopes.Tis limitation raises concerns for unintended selection bias or issues of regression to the mean after rapid progression.Terefore, further investigation with prospective cohort studies may further elucidate our fndings.In addition, the majority of the included eyes in our study were at severe glaucoma stages and the surgeries were done by diferent surgeons with diferent techniques, which may limit the generalizability of our fndings to less advanced cases.However, this is the frst study providing data on the use of GDIs in this patient population.Tese data can be used to inform clinical decision-making for this specifc population.In conclusion, GDI surgeries may be used in patients with lower preop IOP to reduce the number of required hypotensive medications.However, our study suggests that GDIs did not lead to signifcant IOP reduction and were associated with a considerably high complication rate.Terefore, careful consideration should be exercised when using GDIs in eyes with lower preop IOP.Further research may help further elucidate the fndings of this study and the role of GDIs in managing glaucoma in these cases.

6
Journal of Ophthalmology

Figure 3 :
Figure 3: Postoperative changes in the number of prescribed medications among included eyes, with fndings presented as the average ± 95% confdence interval (CI).* and * * p values less than 0.05 and 0.01, respectively.

Figure 4 :
Figure 4: Kaplan-Meier survival curve compares diferent failure defnitions in our population.Censored data points are demonstrated as + in each curve.
Analysis.Snellen acuities were converted to logMAR values for statistical purposes.Average preop IOP was calculated including IOP measurements of three visits before the surgery.Data and analysis are demonstrated in tables and graphs with colorblind safe colors.GDIs were categorized as valved (i.e., Ahmed FP7 [new world medical, Edison CourtRancho Cucamonga, California, USA]) and nonvalved (including Baerveldt [Abbott Medical Optics, Abbott Park, Illinois, USA], Molteno [Molteno Ophthalmic Limited, Dunedin, New Zealand], and Ahmed valveless Clearpath [ACP, New World Medical, Edison Rancho Cucamonga, California, USA]

Table 1 :
Demographic profle and clinical characteristics of the included glaucoma patients.
Line graph illustrating the trend in postoperative visual acuity of the eyes, with fndings presented as the average ± 95% confdence interval (CI).