Surgical outcomes of Baerveldt glaucoma implant versus Ahmed glaucoma valve in neovascular glaucoma: A multicenter study

doi:10.21203/rs.3.rs-4148174/v1

Download PDF

Article

Surgical outcomes of Baerveldt glaucoma implant versus Ahmed glaucoma valve in neovascular glaucoma: A multicenter study

https://doi.org/10.21203/rs.3.rs-4148174/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

This multicenter retrospective study compared the surgical outcomes of Baerveldt glaucoma implant (BGI) surgery with those of Ahmed glaucoma valve (AGV) surgery in patients with neovascular glaucoma (NVG). This study included patients with NVG aged ≥ 20 years who had undergone BGI (223 eyes) or AGV (146 eyes) surgery between April 1, 2012, and December 31, 2021, across 10 clinical centers in Japan. Surgical success or failure was the primary outcome measure of this study. We defined surgical failure as a reduction of < 20% in the pre-operative intraocular pressure (IOP) or criterion A (IOP > 21 mmHg), criterion B (IOP > 17 mmHg), or criterion C (IOP > 14 mmHg). In addition, we considered a requirement for reoperation, loss of light perception, and hypotony as surgical failure. The surgical success rate of the BGI surgery group was significantly higher than that of the AGV group for criteria A (P = 0.01) and B (P = 0.01). Multivariate analysis revealed that AGV surgery showed significant associations with surgical failure for criteria A (hazard ratio, 1.74), B (hazard ratio, 1.72), and C (hazard ratio, 1.33). The overall incidence of postoperative complications was comparable between the two groups. The requirement for reoperation in the AGV surgery group was significantly higher than that in the BGI surgery group (12.3% vs. 5.8%, P = 0.03). BGI surgery yielded a higher success rate than AGV surgery in patients with NVG for a target IOP of < 21 or < 17 mmHg. No significant differences were observed between the two procedures in terms of the incidence of postoperative complications. Additional glaucoma surgery was required more frequently following AGV surgery.

Health sciences/Medical research

Health sciences/Medical research/Outcomes research

Neovascular glaucoma (NVG), a major type of refractory glaucoma, is caused by severe ocular ischemic diseases such as ocular ischemic syndrome, proliferative diabetic retinopathy, and retinal vein occlusion.^1,2 Glaucoma surgery is chosen in cases of NVG wherein adequate management of intraocular pressure (IOP) cannot be achieved with glaucoma medications. Trabeculectomy was the most frequently performed surgery for the management of NVG in the past;³ however, the surgical outcomes of trabeculectomy for NVG are poorer than those of other types of glaucoma.^4–6 Thus, the management of IOP in patients with NVG is challenging.

Tube-shunt surgery has become the most commonly performed surgery for the management of IOP elevation in patients with NVG in the United States in recent years.⁷ Similarly, tube-shunt surgery is being favored increasingly over trabeculectomy for the management of NVG in Japan.³ Previous studies have reported the efficacy of tube-shunt surgery in the management of NVG.^8–11 Our previous study demonstrated the superiority of Baerveldt glaucoma implant (BGI) surgery over trabeculectomy in the management of IOP in eyes with NVG.¹² The use of BGI and the Ahmed glaucoma valve (AGV) in tube-shunt surgery has been approved in Japan, and BGI and AGV surgery have been performed at an almost equal frequency at the discretion of the surgeon.³ However, few studies have compared the surgical outcomes of BGI surgery with those of AGV surgery for the treatment of NVG.

Previous studies that compared the surgical outcomes of BGI surgery with those of AGV surgery for the treatment of NVG have revealed comparable outcomes in both groups; however, the number of postoperative glaucoma medications administered postoperatively in the BGI group was lesser than that in the AGV group.^13–15 Notably, the sample sizes of these studies were small. Furthermore, the pooled data analysis of the Ahmed–Baerveldt comparison study and the Ahmed–Baerveldt comparison study including NVG (approximately 30%) revealed that the failure rate of BGI surgery was lower than that that of AGV surgery. Similarly, the mean IOP and the number of glaucoma medications administered postoperatively reported following were lesser than those reported following AGV surgery.¹⁶

Previous studies have identified AGV and NVG as risk factors for the failure of tube-shunt surgery.¹⁷ BGI surgery has exhibited a better IOP-lowering effect with fewer glaucoma medications than AGV surgery for refractory glaucoma. However, the type of tube-shunt surgery more suitable for the treatment of NVG remains unclear. Therefore, this study aimed to compare the surgical outcomes of BGI surgery with those of AGV surgery for NVG by analyzing the retrospective data of patients with NVG who underwent BGI or AGV surgery across ten clinical centers in Japan.

Patient selection

This multicenter, retrospective, clinical cohort study was approved by the Institutional Review Boards of the Fukui University Hospital, Kumamoto University Hospital, Kanazawa University Hospital, Juntendo University Hospital, Toho University Ohashi Medical Center, Shimane University Hospital, Nagoya City University Hospital, Ehime University Hospital, Gunma University Hospital, and Kochi University Hospital in Japan. The study protocol adhered to the tenets of the Declaration of Helsinki. The requirement for obtaining informed consent was waived owing to the retrospective nature of the study.

We included patients with NVG aged ≥ 20 years who had undergone BGI (BG101-350 or BG102-350; Johnson & Johnson Vision, CA, USA) or AGV (FP7; New World Medical, CA, USA) surgery between April 1, 2012, and December 31, 2021, across 10 clinical centers. We excluded patients with no light perception (NLP), patients who had undergone tube-shunt surgery (BGI or AGV), and patients who had not been followed up for ≥ 3 months postoperatively. However, we included patients who met the failure criteria owing to reoperation or NLP within 3 months postoperatively. We evaluated the eye that was treated first in cases where both eyes of the same patient satisfied the inclusion criteria. We evaluated only the right eye in cases where both eyes of the same patient were operated on simultaneously.

Data collection

We collected data regarding the age, sex, best corrected visual acuity (BCVA), pre-operative IOP, postoperative IOP, history of intraocular surgeries, etiology of NVG, and lens status of the patients. In addition, we also collected data regarding the history of panretinal photocoagulation (PRP), intravitreal injection of anti-vascular endothelial growth factor (VEGF) administered within 1 month before surgery, neovascularization of the iris or anterior chamber angle (ACA), totally closed ACA by peripheral anterior synechiae (PAS), and combined surgery, as well as the incidence of any postoperative complication and the number of glaucoma medications administered. We converted the BCVA into the logarithm of the minimal angle of resolution (logMAR) using the logarithm of the reciprocal of the decimal BCVA and classified eyes without form vision into the following four low-vision categories based on the decimal equivalents: 0.00500, counting fingers; 0.00250, hand motions; 0.00125, light perception; and 0.00010, NLP.⁴ We defined history of totally closed ACA as the formation of > 75% PAS. Oral carbonic anhydrase inhibitors were considered glaucoma medications.

Primary outcome measure

Surgical success or failure, defined according to the three IOP criteria, was the primary outcome of this study. We defined surgical failure as the following IOP levels at ≥ 3 months postoperatively with or without the use of glaucoma medications at two consecutive follow-up visits: criterion or a reduction of < 20% from the pre-operative IOP; criterion A, IOP > 21 mmHg; criterion B, IOP > 17 mmHg; and criterion C, IOP > 14 mmHg. We used IOP measurements acquired every within 6 months during the follow-up visits to determine surgical success. In addition, surgical failure was declared for all criteria in cases that required reoperation for glaucoma (including removal of the tube-shunt due to complications) or presented with NLP or with hypotony of ≤ 5 mmHg. We did not consider needling as glaucoma reoperation. We considered the surgical outcomes of cases that did not meet these criteria as successful and compared the probability of success of the eyes that underwent BGI surgery with that of those that underwent AGV surgery.

Secondary outcome measures

The secondary outcome measures of this study were IOP, number of glaucoma medications administered postoperatively, incidence of early- or late-onset postoperative complications, type of reoperation for glaucoma, visual acuity, and postoperative intervention.

Statistical analysis

We used the chi-square test, Fisher’s exact test, and the Mann–Whitney U nonparametric test to perform univariate comparisons between the groups and applied Bonferroni correction when repetitive analyses were performed. Log-rank test was used for statistical analyses of Kaplan–Meier survival curves. Statistical significance was set at P-values of < 0.05. We performed multivariate analysis using Cox proportional hazards models to determine the prognostic factors of surgical failure subsequently and selected important prognostic factors that may be associated with surgical failure based on the findings of previous reports and clinical experience. We used JMP Pro statistical package, version 17.2.0 (SAS Institute Inc. Cary, NC, USA) for all statistical analyses.

Patient characteristics

This study included 369 eyes (369 patients) comprising 223 and 146 eyes that had undergone BGI surgery (BGI group) and AGV surgery (AGV group), respectively. Table 1 summarizes the pre-operative characteristics of the patients. The pre-operative IOP in the BGI group was significantly higher than that in the AGV group (38.2 mmHg and 34.3 mmHg, respectively, P < 0.01). The mean follow-up period in the BGI and AGV groups was 34.5 months and 26.1 months (P < 0.01), respectively. The number of previous intraocular surgeries in the BGI group was significantly higher than that in the AGV group (P < 0.01). The number of eyes that had undergone pars plana vitrectomy previously in the BGI group was higher than that in the AGV group (P < 0.01). The incidence of retinal vein occlusion was more frequent in the AGV group than in the BGI group (P = 0.04). Previous neovascularization of the iris or ACA and previous totally closed ACA was observed significantly more frequently in the AGV group than in the BGI group (P < 0.01 and P < 0.01, respectively). No other statistically significant differences were observed between the groups in terms of the pre-operative status. Supplemental Table 1S presents the surgical conditions. Combined surgery was performed more frequently in the AGV group than in the BGI group (P = 0.02). Pars plana insertion was performed more frequently in the BGI group than in the AGV group (P < 0.01).

Table 1

Preoperative Characteristics.
Characteristic	BGI (n = 223)	AGV (n = 146)	P value
Age (years), mean ± SD	61.6 ± 12.9	63.5 ± 14.8	0.10
Sex, n (%)			0.73
Men	156 (70.0)	105 (71.9)
Women	67 (30.0)	41 (28.1)
Visual acuity (LogMAR), mean ± SD	1.3 ± 0.9	1.3 ± 0.9	0.53
IOP (mmHg), mean ± SD	38.2 ± 11.6	34.3 ± 11.5	< 0.01
Glaucoma medicines, n, mean ± SD	4.0 ± 1.3	3.7 ± 1.7	0.22
Follow-up period (months), mean ± SD	34.5 ± 27.4	26.1 ± 21.9	< 0.01
History of intraocular surgeries, n, mean ± SD	2.2 ± 1.6	1.7 ± 1.1	< 0.01
History of filtering surgery, n (%)	52 (23.3)	36 (24.7)	0.80
History of pars plana vitrectomy, n (%)	138 (61.9)	64 (43.8)	< 0.01
Etiology of NVG, n (%)			0.04
PDR	175 (78.5)	105 (71.9)
RVO	22 (9.9)	28 (19.2)
OIS	19 (8.5)	12 (8.2)
Others	7 (3.1)	1 (0.7)
Lens status, n (%)			0.93
Phakia	34 (15.3)	21 (14.4)
Pseudophakia	188 (84.3)	124 (84.9)
Aphakia	1 (0.4)	1 (0.7)
PRP, n (%)	191 (85.7)	117 (80.1)	0.20
Intravitreal injection of anti-VEGF, n (%)	78 (35.0)	59 (40.4)	0.32
Neovascularization of the iris or ACA, n (%)	100 (44.8)	90 (61.6)	< 0.01
Totally closed ACA by PAS, n (%)	51 (22.9)	64 (43.8)	< 0.01
ACA, anterior chamber angle; AGV, Ahmed glaucoma valve; BGI, Baerveldt glaucoma implant; IOP, intraocular pressure; logMAR, logarithm of the minimum angle of resolution; NVG, neovascular glaucoma; OIS, ocular ischemic syndrome; PAS, peripheral anterior synechiae; PDR, proliferative diabetic retinopathy; PRP, panretinal photocoagulation; RVO, retinal vein occlusion; SD, standard deviation; VEGF, vascular endothelial growth factor.

Primary outcome measure

Figure 1 depicts the Kaplan–Meier survival curves comparing the surgical outcomes of the BGI and AGV groups according to the failure criteria: A, B, and C. Surgical failure occurred within 3 months postoperatively in eight eyes (NLP, six eyes; reoperation, one eye; and removal of tube-shunt, one eye) in the BGI group. Surgical failure occurred within 3 months postoperatively in five eyes (NLP, three eyes; reoperation, two eyes) in the AGV group. The probability of success in the BGI group was significantly higher than that in the AGV group for criteria A (P = 0.01) and B (P = 0.01). However, no significant differences were observed between the groups for criterion C (P = 0.13). The probability of success at 1, 3, and 5 years postoperatively in the BGI group for criterion A was 86.1%, 72.4%, and 66.3%, respectively. The probability of success at 1, 3, and 5 years postoperatively in the AGV group for criterion A was 73.5%, 57.3%, and 52.8%, respectively. The probability of success at 1, 3, and 5 years postoperatively in the BGI group for criterion B was 75.4%, 53.0%, and 46.3%, respectively. The probability of success at 1, 3, and 5 years postoperatively in the AGV group for criterion B was 57.1%, 40.5%, and 29.5%, respectively. The probability of success at 1, 3, and 5 years postoperatively in the BGI group for criterion C was 60.2%, 28.9%, and 23.4%, respectively. The probability of success at 1, 3, and 5 years postoperatively in the AGV group was 45.0%, 26.6%, and 22.8%, respectively.

Table 2 presents the reasons for treatment failure. Inadequate reduction in IOP was the most common reason for failure in both groups (IOP > 21 mmHg or < 20% of reduction in the pre-operative IOP at two consecutive follow-up visits after 3 months). Inadequate reduction in IOP was observed in six and five patients in the BGI and AGV groups, respectively, and these patients underwent reoperation for glaucoma subsequently. No significant differences were observed between the two groups in terms of the classification of the reasons for treatment failure (P = 0.31). However, reoperation for glaucoma was performed more frequently in the AGV group than in the BGI group (8.9% vs. 3.1%).

Table 2

Reasons for treatment failure.
	BGI^a (n = 60)	AGV^a (n = 50)
Inadequate IOP reduction^b,c	32 (14.3)	25 (17.1)
Reoperation for glaucoma	7 (3.1)	13 (8.9)
Hypotony^d	3 (1.3)	2 (1.4)
Removal of tube shunt	6 (2.7)	2 (1.4)
Loss of light perception	12 (5.4)	8 (5.5)
^aData are presented as number of patients (percentage).
^bIOP > 21 mmHg or < 20% reduction in the preoperative IOP at two consecutive follow-up visits after 3 months.
^cSome patients underwent reoperation for glaucoma subsequent to failure owing to inadequate management of IOP.
^dIOP ≤ 5 mmHg at two consecutive follow-up visits after 3 months.
AGV, Ahmed glaucoma valve; BGI, Baerveldt glaucoma implant; IOP, intraocular pressure.

Subgroup analyses of the primary outcome

We conducted subgroup analyses by extracting eyes with no history of undergoing filtering surgery, including trabeculectomy and Ex-PRESS (Alcon Laboratories, TX, USA) (BGI group, 171 eyes; AGV group, 110 eyes). Supplemental Fig. 1S presents the comparison of the Kaplan–Meier survival curves for the surgical outcomes of the BGI and AGV groups as the primary filtering surgery for criteria A, B, and C. The probability of success in the BGI group was significantly higher than that in the AGV group for criteria A (P < 0.01) and B (P < 0.01). However, no significant differences were observed between the groups for criterion C (P = 0.06). The probability of success at 1, 3, and 5 years postoperatively for criterion A in the BGI group was 88.9%, 78.7%, and 71.6%, respectively. The probability of success at 1, 3, and 5 years postoperatively for criterion A in the AGV group was 73.8%, 54.9%, and 54.4%, respectively. The probability of success at 1, 3, and 5 years postoperatively for criterion B in the BGI group was 79.4%, 63.2%, and 55.3%, respectively. The probability of success at 1, 3, and 5 years postoperatively for criterion B in the AGV group was 63.1%, 40.5%, and 26.3%, respectively. The probability of success at 1, 3, and 5 years postoperatively for criterion C in the BGI group was 63.8%, 33.8%, and 27.8%, respectively. The probability of success at 1, 3, and 5 years postoperatively for criterion B in the AGV group was 47.2%, 25.0%, and 20.0%, respectively.

Secondary outcome measures

We compared the IOP values and number of glaucoma medications administered in the BGI group with those in the AGV group at various follow-up time points (Table 3). We subsequently excluded eyes that had undergone additional glaucoma surgery from the analysis after reoperation. The pre-operative IOP in the BGI group was significantly higher than that in the AGV group (P = 0.01). Both surgical procedures resulted in a significant and sustained reduction in IOP, and the mean IOP in the BGI and AGV groups 5 years postoperatively was 14.2 ± 4.9 mmHg and 15.6 ± 4.4 mmHg, respectively. A significant reduction in the use of glaucoma medications was observed in both groups. The number of glaucoma medications administered in the BGI and AGV groups 5 years postoperatively was 2.1 ± 1.7 and 2.1 ± 1.1, respectively. No significant differences were observed between the groups at any time point postoperatively in terms of the IOP or number of glaucoma medications administered.

Table 3

Intraocular pressure and glaucoma medicines at preoperative and follow-up visits.
	BGI	AGV	P value
Preoperative
IOP (mmHg)	38.2 ± 11.6	34.3 ± 11.5	0.01
Number of medicines	4.0 ± 1.3	3.7 ± 1.7	> 0.99
Number of patients	223	146
6 months
IOP (mmHg)	14.9 ± 6.2	15.6 ± 5.7	> 0.99
Number of medicines	1.7 ± 1.6	1.6 ± 1.4	> 0.99
Number of patients	199	130
1 year
IOP (mmHg)	14.1 ± 5.2	13.8 ± 4.2	> 0.99
Number of medicines	1.7 ± 1.6	1.9 ± 1.4	> 0.99
Number of patients	180	107
2 years
IOP (mmHg)	14.8 ± 5.1	14.8 ± 4.1	> 0.99
Number of medicines	1.9 ± 1.7	2.2 ± 1.5	> 0.99
Number of patients	132	66
3 years
IOP (mmHg)	14.3 ± 4.4	15.1 ± 4.1	> 0.99
Number of medicines	2.0 ± 1.6	2.5 ± 1.5	0.77
Number of patients	99	44
4 years
IOP (mmHg)	14.4 ± 4.7	14.4 ± 3.8	> 0.99
Number of medicines	2.1 ± 1.7	1.9 ± 1.3	> 0.99
Number of patients	70	28
5 years
IOP (mmHg)	14.2 ± 4.9	15.6 ± 4.4	> 0.99
Number of medicines	2.1 ± 1.7	2.1 ± 1.1	> 0.99
Number of patients	56	14
Data are presented as mean ± standard deviation.
AGV, Ahmed glaucoma valve; BGI, Baerveldt glaucoma implant; IOP, intraocular pressure.

Table 4 presents the early-onset postoperative complications that occurred within 3 months postoperatively. The overall incidence of early-onset postoperative complications was similar in the two groups: a total of 100 early-onset postoperative complications were observed in 66 patients (29.6%) in the BGI group, whereas a total of 41 early-onset postoperative complications were observed in 33 patients (22.6%) in the AGV group (P = 0.15). The incidence of choroidal detachment was significantly more frequent in the BGI group than in the AGV group (8.1% vs. 2.1%, P = 0.02). No statistically significant differences were observed between the groups in terms of the incidence of early-onset postoperative complications.

Table 4

Early-onset postoperative complications (onset ≤ 3 months).
	BGI^a (n = 223)	AGV^a (n = 146)	P value
Hyphema	28 (12.6)	19 (13.0)	1.00
Shallow or flat anterior chamber	6 (2.7)	4 (2.7)	1.00
Choroidal detachment	18 (8.1)	3 (2.1)	0.02
Vitreous hemorrhage	33 (14.8)	12 (8.2)	0.07
Hypotony maculopathy	4 (1.8)	0 (0.0)	1.00
IOL dislocation	1 (0.4)	0 (0.0)	1.00
Tube obstruction	7 (3.1)	4 (2.7)	1.00
Wound dehiscence	2 (0.9)	0 (0.0)	0.52
Removal of the tube shunt	1 (0.4)	0 (0.0)	1.00
Suprachoroidal hemorrhage	0 (0.0)	0 (0.0)	NA
Endophthalmitis/blebitis	0 (0.0)	0 (0.0)	NA
Total number of patients with early-onset postoperative complications^b	66 (29.6)	33 (22.6)	0.15
^aData are presented as number of patients (percentage).
^bMore than one complication was reported in some patients.
AGV, Ahmed glaucoma valve; BGI, Baerveldt glaucoma implant; IOL, intraocular lens; NA, not applicable.

Table 5 presents the late-onset postoperative complications that occurred > 3 months postoperatively. We excluded eyes that met the failure criteria within 3 months postoperatively (BGI group, n = 8; AGV group, n = 5) from the analysis of late-onset postoperative complications. The overall incidence of late-onset postoperative complications was similar in both groups: a total of 55 late-onset postoperative complications were observed in 34 patients (15.8%) in the BGI group, whereas a total of 27 late-onset postoperative complications were observed in 21 patients (14.9%) in the AGV group (P = 0.88). No significant differences were observed between the groups in terms of the incidence of late-onset postoperative complications.

Table 5

Late-onset postoperative complications (onset > 3 months).
	BGI^a (n = 215)	AGV^a (n = 141)	P value
Hyphema	7 (3.3)	4 (2.8)	1.00
Vitreous hemorrhage	16 (7.4)	13 (9.2)	0.56
Hypotony maculopathy	3 (1.4)	0 (0.0)	0.28
Endophthalmitis/blebitis	4 (1.9)	1 (0.7)	0.65
Retinal detachment	1 (0.5)	2 (1.4)	0.57
IOL dislocation	1 (0.5)	0 (0.0)	1.00
Tube obstruction	1 (0.5)	0 (0.0)	1.00
Tube erosion	14 (6.5)	4 (2.8)	0.14
Removal of the tube shunt	7 (3.3)	3 (2.1)	0.75
Suprachoroidal hemorrhage	1 (0.5)	0 (0.0)	1.00
Corneal decompensation	0 (0.0)	0 (0.0)	NA
Total number of patients with late-onset postoperative complications^b	34 (15.8)	21 (14.9)	0.88
^aData are presented as number of patients (percentage).
^bMore than one complication was reported in some patients.
AGV, Ahmed glaucoma valve; BGI, Baerveldt glaucoma implant; IOL, intraocular lens; NA, not applicable.

Table 6 presents the reoperations performed for glaucoma. The number of reoperations required in the AGV group was significantly higher than that in the BGI group (12.3% vs. 5.8%, P = 0.03). AGV surgery was the most frequently performed reoperation surgery (6.2%) performed for glaucoma in the AGV group. AGV surgery was performed significantly more frequently as a reoperation surgery in the AGV group than in the BGI group (P < 0.01).

Table 6

Reoperation for Glaucoma.
	BGI (n = 223)	AGV (n = 146)	P value
Bleb revision	4 (1.8)	6 (4.1)	0.20
Trabeculectomy	1 (0.4)	1 (0.7)	1.00
BGI	2 (0.9)	0 (0.0)	0.52
AGV	0 (0.0)	9 (6.1)	< 0.01
TSCPC	2 (0.9)	1 (0.7)	1.00
MP-TSCPC	4 (1.8)	1 (0.7)	0.65
Total number of patients who underwent reoperation for glaucoma	13 (5.8)	18 (12.3)	0.03
AGV, Ahmed glaucoma valve; BGI, Baerveldt glaucoma implant; MP-TSCPC, micropulse transscleral cyclophotocoagulation; TSCPC, transscleral cyclophotocoagulation.

The visual acuity (LogMAR) at the final follow-up visit in the BGI group was significantly poorer than that in the AGV group (1.6 ± 1.2 vs. 1.3 ± 1.2, P = 0.03). The visual acuity (logMAR) deteriorated significantly from 1.3 ± 0.9 at the pre-operative visit to 1.6 ± 1.2 at the final follow-up visit in the BGI group (P < 0.01). The final visual acuity deteriorated in 124 eyes (55.6%), including 27 eyes (12.1%), wherein visual acuity decreased to NLP. The final visual acuity improved in 64 eyes (28.7%) and was maintained in 35 eyes (15.7%). The visual acuity (logMAR) in the AGV group was 1.3 ± 0.9 and 1.3 ± 1.2 at the pre-operative and final follow-up visits, respectively (P = 0.98). The final visual acuity deteriorated in 61 eyes (41.8%), including 11 eyes (7.5%), wherein visual acuity decreased to NLP. The final visual acuity improved in 67 eyes (45.9%) and was maintained in 18 eyes (12.3%). Supplemental Fig. 2S shows comparison of the Kaplan–Meier survival curves for the NLP progression rates of the BGI and AGV groups. No significant differences were observed between the groups about the NLP progression rates (P = 0.45).

Supplemental Table 2S presents the postoperative interventions. The overall incidence of postoperative interventions was comparable between the two groups: a total of 70 interventions were performed in 54 patients (24.2%) in the BGI group, whereas a total of 44 interventions were performed in 37 patients (25.3%) in the AGV group (P = 0.81).

Multivariate analysis to determine prognostic factors of surgical failure

We evaluated patient characteristics, such as the type of surgery, age, number of intraocular surgeries performed previously, history of intravitreal injection of anti-VEGF, history of neovascularization of the iris or ACA, history of totally closed ACA, pre-operative IOP, and number of pre-operative glaucoma medications, as possible determinants of surgical failure. Table 7 presents the results of the analyses using the multivariate Cox proportional hazards regression models. AGV surgery was a significant prognostic factor for surgical failure for criteria A (hazard ratio [HR], 1.74; P < 0.01), B (HR, 1.72; P < 0.01), and C (HR, 1.33; P = 0.045). Younger age was associated with poorer surgical outcomes for criteria A (HR, 0.97; P < 0.01), B (HR, 0.97; P < 0.01), and C (HR, 0.98; P < 0.01). Totally closed ACA was associated with an increased risk for criteria A (HR, 1.83; P < 0.01), B (HR, 1.55; P = 0.02), and C (HR, 1.39; P = 0.04). The number of previous intraocular surgeries was associated with an increased risk for criterion B (HR, 1.17; P < 0.01). Higher pre-operative IOP was also associated with surgical failure for criterion B (HR, 1.01; P = 0.04).

Table 7

Multivariable analysis performed using Cox proportional hazards regression models to identify the prognostic risk factors for failure.
			Criterion
	A		B		C
	HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
Type of surgery (AGV/BGI) Age per year Number of previous intraocular surgeries per each Intravitreal injection of anti-VEGF Neovascularization of the iris or ACA Totally closed ACA Preoperative IOP per mmHg The number of preoperative glaucoma medication per 1-medication	1.74 (1.14–2.64) 0.97 (0.96–0.99) 1.11 (0.97–1.25) 0.91 (0.60–1.38) 1.08 (0.71–1.64) 1.83 (1.21–2.77) 1.02 (0.99–1.03) 1.09 (0.96–1.25)	< 0.01 < 0.01 0.11 0.67 0.73 < 0.01 0.08 0.20	1.72 (1.23–2.41) 0.97 (0.96–0.99) 1.17 (1.06–1.28) 0.99 (0.71–1.40) 1.15 (0.81–1.62) 1.55 (1.09–2.22) 1.01 (1.00–1.03) 1.07 (0.97–1.19)	< 0.01 < 0.01 < 0.01 0.99 0.43 0.02 0.04 0.19	1.33 (1.01–1.77) 0.98 (0.97–0.99) 1.06 (0.96–1.16) 1.02 (0.76–1.35) 1.09 (0.81–1.46) 1.39 (1.02–1.89) 1.01 (0.99–1.02) 1.06 (0.97–1.15)	0.045 < 0.01 0.24 0.91 0.56 0.04 0.11 0.21
ACA, anterior chamber angle; AGV, Ahmed glaucoma valve; BGI, Baerveldt glaucoma implant; HR, hazard ratio; IOP, intraocular pressure; VEGF, vascular endothelial growth factor.

This retrospective study compared the surgical outcomes of BGI surgery with those of AGV surgery in patients with NVG. The probability of success of BGI surgery was significantly higher than that of AGV surgery for criteria A (P = 0.01) and B (P = 0.01). The potential risk factors identified by the Cox proportional hazard model revealed that AGV surgery contributed to surgical failure for criteria A (hazard ratio [HR] 1.74), B (HR, 1.72), and C (HR 1.33). Furthermore, reoperation for glaucoma was performed significantly more frequently in the AGV group than in the BGI group (12.3% vs. 5.8%, P = 0.03).

Three previous studies have compared the surgical outcomes of BGI surgery with those of AGV surgery for the treatment of NVG. Shalaby et al. compared the long-term effects of BGI surgery (61 eyes) with those of AGV surgery (91 eyes) in patients with NVG in their single-center, retrospective study.¹⁵ Suda et al. compared the effects of BGI (10 eyes) and AGV (11 eyes) surgery with those of pars plana tube insertion in patients with NVG at 1 year postoperatively in their retrospective study.¹³ Maeda et al. compared the outcomes of BGI (26 eyes) and AGV (15 eyes) surgery with those of pars plana tube insertion in patients with NVG at 3 years postoperatively in their retrospective study.¹⁴ Our study is unique, in that, it evaluated 369 eyes with NVG across multiple centers and directly compared the surgical outcomes of BGI surgery with those of AGV surgery in these eyes.

The success rate of BGI surgery was significantly higher than that of AGV surgery in our study, which is consistent with the results in the previous studies that included different types of glaucoma.^16,17 A higher degree of fibrous encapsulation has been observed in the AGV surgery than in the BGI surgery, which may elevate IOP.¹⁸ This may represent the immediate filtration of the aqueous humor rich in inflammatory mediators by the AGV, which promotes a fibrovascular reaction and forms encapsulation around the end plate.¹⁹ A larger endplate yields a greater reduction in IOP after tube-shunt surgery.^16,20–22 The smaller end plates of AGV compared with those of BGI may be another factor leading to the failure of AGV surgery.

The incidence of early-onset and late-onset postoperative complications was similar in both groups. However, a significant difference was observed between the groups in terms of the incidence of choroidal detachment, an early-onset postoperative complication. Choroidal detachment occurs more frequently after BGI surgery as BGI is a non-valved implant prone to transient hypotony during the early postoperative period, which causes early postoperative choroidal detachment.

The number of reoperations performed for glaucoma in the AGV group was higher than that in the BGI group in our study, which is similar to the findings of previous studies.^15,18,23 AGV surgery was the most commonly performed reoperation for glaucoma in the AGV group. This finding may be attributed to the small plate size of AGV, which facilitates easier placement of additional AGV in other quadrants during reoperation.

Previous reports of tube-shunt surgery showed that the visual acuity decreased significantly after surgery.^16,18,23 However, in this study, the visual acuity deteriorated significantly at the final follow-up visit in the BGI group (P < 0.01), but not in the AGV group (P = 0.98). The reason for maintaining visual acuity after AGV surgery compared with BGI surgery may stem from differences in the patient backgrounds (especially the severity of NVG and the etiology of NVG), follow-up durations, and the reduction rates of IOP. Actually, the previous study demonstrated that the cumulative proportion of NVG eyes that progressed to NLP vision in the AGV group was significantly lower than in the BGI group.²³ We could not evaluate the cause of vision loss because of the complex disease for NVG, which is a limitation of our study.

Multivariate analysis revealed four other risk factors for surgical failure. Younger age increased the risk of surgical failure for all criteria in our study, which is consistent with the findings of previous studies on tube-shunt surgery.^17,24–27 Age-related impairments in wound healing have been associated with fibroblast dysfunction, whereas younger age has been associated with a more robust wound healing response.^28–30 The wound healing process involving active fibroblasts may play a role in the surgical failure of filtering surgery. Thus, aging may impair wound healing and increase the surgical success rate of tube-shunt surgery. A history of undergoing a higher number of intraocular surgeries was associated with surgical failure for criterion B. Repeated intraocular surgeries have been associated with poor surgical prognosis.^{12,24–26,31,32} Non-functioning blebs express more extracellular matrix components and activated fibroblasts than functioning blebs after tube-shunt surgery.³³ Furthermore, fibroblast activation may occur in the subconjunctival tissue owing to the intraoperative conjunctival incisions made during phacoemulsification or vitrectomy.^34,35 Thus, repeated ocular surgery, including filtering surgery, may result in disturbed bleb formation after tube-shunt surgery.³² Our study identified totally closed ACA as a risk factor for surgical failure for all criteria. As the NVG progresses, neovascularization appears over the angle structures, leading to complete closure of the angle during the final stages.¹ Therefore, eyes with totally closed ACA may have exhibited greater disease severity of NVG. A higher pre-operative IOP is associated with an increased risk of surgical failure for criterion B. Previous studies have reported an association between high pre-operative IOP and surgical failure for filtering surgery.^12,25,26,36 This result indicates that it is difficult to reduce IOP in patients with NVG, which is a typical refractory glaucoma, regardless of the surgical procedure.

This study has certain limitations, which can be attributed to its multicenter and retrospective nature. First, significant differences were observed between the two groups in terms of pre-operative characteristics, which may represent a difference in the severity of NVG. Second, a selection bias for the type of surgery for NVG may have affected the surgical outcomes because multiple surgeons performed the surgeries at multiple centers. Third, we were unable to standardize the surgical procedures (e.g., number of Sherwood slits, combined procedures, and administration of mitomycin C or triamcinolone acetonide at the time of surgery) or postoperative procedures (e.g., ocular massage, removal of the rip cord, laser suture lysis, and medications). Our study exhibited a significant difference in the probability of success in the BGI group across different center for criteria B and C, and in AGV surgery for criterion C. Fourth, we were unable to collect some of clinical data. Perioperative conjunctival scarring and postoperative inflammation of the anterior chamber can affect bleb formation. Moreover, repeated intraocular surgeries affect the outcome of filtering surgery. Aqueous fluid flow into the subconjunctival space during the early postoperative period accelerates surgical failure of tube-shunt surgery.^19,37 Furthermore, the visual field and corneal endothelial cell density play a crucial role in the evaluation of the efficacy and complications of tube-sunt surgery. Further multicenter, randomized, prospective studies are required to address these limitations.

In conclusion, the surgical success rate of BGI surgery was higher than that of AGV surgery among patients with NVG in our study. Reoperation for glaucoma was performed more frequently after AGV surgery. No significant differences were observed between the two procedures in terms of the incidence of postoperative complications and requirement for interventions. BGI surgery was more effective than AGV surgery in the management of IOP in eyes with NVG.

Competing Interests

M.I.’s work has been partially funded by Johnson & Johnson Vision. The other authors declare no conflict of interest regarding any of the products mentioned in this article.

Funding

This work was supported in part by the Japan Society for the Promotion of Science (KAKENHI) (Grant number 22K16967, 20H03840, 22K19582).

Author Contribution

M.I. had full access to all the data in the present study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Ke.I. and M.I. designed study concept. Acquisition, analysis and interpretation of data were performed by Ke.I., S.K., R.W., A.M., K.Y., A.T., M.K., M.N., Ko.N., Ke.N., Y.M., T.I., T.H., Ky.I., and M.T. Ke.I. and M.I. drafted the manuscript. All authors reviewed and approved the manuscript.

Acknowledgement

The authors thank CACTUS (http://www.cactus.co.jp/) for English language review.

Data Availability

Data is fully available upon reasonable request to corresponding author.

Rodrigues, GB. et al. Neovascular glaucoma: a review. Int J Retin Vitr. 2, 26 (2016).
Hayreh, SS. Neovascular glaucoma. Prog Retin Eye Res. 26, 470–485 (2007).
Iwasaki, K., Arimura, S., Takamura, Y. & Inatani, M. Clinical practice preferences for glaucoma surgery in Japan: a survey of Japan Glaucoma Society specialists. Jpn J Ophthalmol. 64, 385–391 (2020).
Takihara, Y. et al. Trabeculectomy with mitomycin C for neovascular glaucoma: prognostic factors for surgical failure. Am J Ophthalmol. 147, 912–918, 918.e1 (2009).
Kiuchi, Y., Sugimoto, R., Nakae, K., Saito, Y. & Ito, S. Trabeculectomy with mitomycin C for treatment of neovascular glaucoma in diabetic patients. Ophthalmologica. 220, 383–388 (2006).
Takihara, Y. et al. Combined intravitreal bevacizumab and trabeculectomy with mitomycin C versus trabeculectomy with mitomycin C alone for neovascular glaucoma. J Glaucoma. 20, 196–201 (2011).
Vinod, K. et al. Practice Preferences for Glaucoma Surgery. J Glaucoma. 26, 687–693 (2017).
Dong, Z., Gong, J., Liao, R. & Xu, S. Effectiveness of multiple therapeutic strategies in neovascular glaucoma patients: A PRISMA-compliant network meta-analysis. Medicine (Baltimore). 97, e9897 (2018).
Tojo, N., Nakamura, T., Ueda, TC., Yanagisawa, S. & Hayashi, A. Results of the Baerveldt® Glaucoma Implant for Neovascular Glaucoma Patients. Nihon Ganka Gakkai Zasshi. 121, 138–145 (2017).
Kolomeyer, AM. et al. Combined pars plana vitrectomy and pars plana Baerveldt tube placement in eyes with neovascular glaucoma. Retina. 35, 17–28 (2015).
Nishitsuka, K., Sugano, A., Matsushita, T., Nishi, K. & Yamashita, H. Surgical outcomes after primary Baerveldt glaucoma implant surgery with vitrectomy for neovascular glaucoma. PLoS One. 16, e0249898 (2021).
Iwasaki, K. et al. Comparing Surgical Outcomes in Neovascular Glaucoma between Tube and Trabeculectomy: A Multicenter Study. Ophthalmol Glaucoma. 5, 672–680 (2022).
Suda, M. et al. Baerveldt or Ahmed glaucoma valve implantation with pars plana tube insertion in Japanese eyes with neovascular glaucoma: 1-year outcomes. Clin Ophthalmol. 12, 2439–2449 (2018).
Maeda, M. et al. Three-Year Outcomes of Pars Plana Ahmed and Baerveldt Glaucoma Implantations for Neovascular Glaucoma in Japanese Eyes. J Glaucoma. 31, 462–467 (2022).
Shalaby, WS. et al. Outcomes of Valved and Nonvalved Tube Shunts in Neovascular Glaucoma. Ophthalmol Glaucoma. 4, 182–192 (2021).
Christakis, PG. et al. Five-Year Pooled Data Analysis of the Ahmed Baerveldt Comparison Study and the Ahmed Versus Baerveldt Study. Am J Ophthalmol. 176, 118–126 (2017).
Bowden, EC. et al. Risk Factors for Failure of Tube Shunt Surgery: A Pooled Data Analysis. Am J Ophthalmol. 240, 217–224 (2022).
Christakis, PG. et al. The Ahmed Versus Baerveldt Study: Five-Year Treatment Outcomes. Ophthalmology. 123, 2093–2102 (2016).
Freedman, J. & Iserovich, P. Pro-Inflammatory Cytokines in Glaucomatous Aqueous and Encysted Molteno Implant Blebs and Their Relationship to Pressure. Investig Opthalmology Vis Sci. 54, 4851–4855 (2013).
Gedde, SJ., Panarelli, JF., Banitt, MR. & Lee, RK. Evidenced-based comparison of aqueous shunts. Curr Opin Ophthalmol. 24, 87–95 (2013).
Heuer, DK. et al. Which is better? One or two? A randomized clinical trial of single-plate versus double-plate Molteno implantation for glaucomas in aphakia and pseudophakia. Ophthalmology. 99, 1512–1519 (1992).
Goulet, RJ 3rd., Phan, AD., Cantor, LB. & WuDunn, D. Efficacy of the Ahmed S2 glaucoma valve compared with the Baerveldt 250-mm2 glaucoma implant. Ophthalmology. 115, 1141–1147 (2008).
Budenz, DL. et al. Five-year treatment outcomes in the Ahmed Baerveldt comparison study. Ophthalmology. 122, 308–316 (2015).
Seah, SK., Gazzard, G. & Aung, T. Intermediate-term outcome of Baerveldt glaucoma implants in Asian eyes. Ophthalmology. 110, 888–894 (2003).
Krishna, R. et al. Intermediate-term Outcomes of 350-mm 2 Baerveldt Glaucoma Implants. Ophthalmology. 108, 621–626 (2001).
Iwasaki, K. et al. Long-term outcomes of Baerveldt glaucoma implant surgery in Japanese patients. Sci Rep. 13, 14312 (2023).
Iraha, S. et al. Factors associated with the surgical outcomes of Baerveldt glaucoma implant for open-angle glaucoma, an age-related eye disease. Sci Rep. 12, 1359 (2022).
Uitto, J. A method for studying collagen biosynthesis in human skin biopsies in vitro. Biochim Biophys Acta. 201, 438–445 (1970).
Sussman, MD. Aging of connective tissue: physical properties of healing wounds in young and old rats. Am J Physiol. 224, 1167–1171 (1973).
Fujiwara, T. et al. Age-associated intracellular superoxide dismutase deficiency potentiates dermal fibroblast dysfunction during wound healing. Exp Dermatol. 28, 485–492 (2019).
Matsushita, K., Kawashima, R., Kawasaki, R. & Nishida, K. Prognostic factors for successful Baerveldt glaucoma implant surgery for refractory glaucoma after multiple surgeries. Jpn J Ophthalmol. 65, 820–826 (2021).
Iwasaki, K. et al. Evaluation of Bleb Fluid After Baerveldt Glaucoma Implantation Using Magnetic Resonance Imaging. Sci Rep. 7, 11345 (2017).
Välimäki, J. & Uusitalo, H. Immunohistochemical analysis of extracellular matrix bleb capsules of functioning and non-functioning glaucoma drainage implants. Acta Ophthalmol. 92, 524–528 (2014).
Gozawa, M. et al. Prospective observational study of conjunctival scarring after phacoemulsification. Acta Ophthalmol. 94, e541-e549 (2016).
Gozawa, M. et al. Comparison of subconjunctival scarring after microincision vitrectomy surgery using 20-, 23-, 25- and 27-gauge systems in rabbits. Acta Ophthalmol. 95, e602-e609 (2017).
Kawabata, K. et al. Efficacy and safety of Ex-PRESS® mini shunt surgery versus trabeculectomy for neovascular glaucoma: a retrospective comparative study. BMC Ophthalmol. 19, 75 (2019).
Iwasaki, K. et al. Filtering Blebs after Baerveldt Glaucoma Implantation Using Magnetic Resonance Imaging: A Prospective Investigation. Ophthalmol Glaucoma. 3, 221–224 (2020).

Competing interest reported. M.I.’s work has been partially funded by Johnson & Johnson Vision. The other authors declare no conflict of interest regarding any of the products mentioned in this article.

Supplemental.docx

Download PDF

Editor invited by journal
05 Apr, 2024
Submission checks completed at journal
05 Apr, 2024
First submitted to journal
22 Mar, 2024

You are reading this latest preprint version

Surgical outcomes of Baerveldt glaucoma implant versus Ahmed glaucoma valve in neovascular glaucoma: A multicenter study

Status:

Version 1

Abstract

Figures

Introduction

Methods

Patient selection

Data collection

Primary outcome measure

Secondary outcome measures

Statistical analysis

Results

Patient characteristics

Primary outcome measure

Subgroup analyses of the primary outcome

Secondary outcome measures

Multivariate analysis to determine prognostic factors of surgical failure

Discussion

Declarations

Competing Interests

Funding

Author Contribution

Acknowledgement

Data Availability

References

Additional Declarations

Supplementary Files

Status:

Version 1