Discussion
In studying the efficacy of the AGV M4, our results showed a cumulative probability of failure of 32% at 6 months and 72% at 1 year. These results are higher than previous reports on the AGV. For the FP7 model at 1 year, the Ahmed Baerveldt Comparison (ABC) study reported a cumulative failure rate of 16.4%, and the Ahmed versus Baerveldt (AVB) study reported a cumulative failure rate of 43%.[19,20] In addition, Kim et al[21] reported a 20% cumulative probability of failure using the M4 model at 1 year. The high rate of failure in our study may be explained in several ways. First, many eyes that failed used a low number of ancillary glaucoma medications following surgery. Eyes that failed at last follow-up had a mean number of glaucoma medications of 1.3±1.3, whereas those who were considered qualified success had a mean number of glaucoma medications of 2.5±0.8 (P<0.01). With high IOP being the reason for 68% of the failed eyes, this indicates that the number of glaucoma medications postsurgery may be a factor in differentiating the failure group from the qualified success group. Moreover, the number of medications at 1 year (1.2) was lower than the above cited studies (~1.5 to 1.8).[19–21] This may be due to the surgeons' decision to occasionally not give adjunctive medication to eyes that had an IOP>21 mm Hg postoperatively, as these eyes had different IOP needs, were already visually impaired eyes that stabilized, and/or were known to be allergic to the glaucoma medications. Overall, there was a significant increase in IOP 1 month following surgery as IOP stabilized at high levels thereafter. This suggests that to improve IOP control in eyes requiring a low-target IOP, it may be necessary to administer adjunctive medication such as aqueous suppressant treatment in the early (<1 mo) postoperative period.[4,20,21,27] This approach, which was suggested by Pakravan et al,[27] showed that eyes which received topical timolol-dorzolamide fixed-combination drops when IOP exceeded 10 mm Hg had a higher success rate and reduced the HP compared with eyes which did not. Although earlier usage of aqueous suppressants in the current study would lower the percentage of complete success, it may also potentially decrease the chance of failure.
In addition, operative characteristics (ie, number of surgeons, quadrant placement, type of sutures, and type of patch graft) were different from the study by Kim and colleagues. Our study included patients who had a previous glaucoma implant surgery and may be at greater risk for failure, compared with other studies that did not include prior glaucoma implant surgery with the AGV.[1,21] Prior intraocular surgery is a well-established risk factor for failure,[28–30] and many of our patients had failed superotemporal tube shunts or scarring that prevented surgery at the desired location. While the inferotemporal AGV M4 implants were found to have a higher percentage of reoperation for glaucoma, the increase was not statistically different (P=0.53) which may be due to low sample size and power.
Our clinical impression was that adhesions likely formed between the Ahmed M4 plate and Tenon's capsule, which prevented aqueous from flowing freely from the anterior chamber to form a low diffuse bleb. The porous plate was designed to allow blood vessels to course through the pores and to have a lower outflow resistance.[21,31] However, this did not seem to be the case at the 3- and 6-month follow-up. In the eyes that failed, the outline of the plate was clearly visible with the conjunctiva and Tenon's capsule adhering to it. This compression may lead to less fluid egress and a subsequent higher IOP. While other porous biomaterials such as expanded polytetrafluoroethylene have been incorporated in animal models,[32,33] polyethylene which composes the plate of the M4 is a new material and the dynamics of postoperative wound healing is not well understood. In addition, the M4 presents a smaller surface area plate of 160 mm2, compared with the AGV S2 and FP7 counterparts which have a surface area of 184 mm2. GDDs with a larger plate produce a larger encapsulating surface area and a generally greater degree of pressure reduction.[34,35] With a lower encapsulating surface area, the M4 may have a lower capacity to reduce IOP compared with other AGVs and GDDs.
The HP was defined as a rise in IOP up to 6 months after adequate IOP control (<22 mm Hg) in the first postoperative week. The pathophysiology of a HP is unknown. However, this phenomenon generally occurs due to a secondary to a thick-walled bleb, which develops over the plate in the first few weeks. With a reduction of bleb congestion and inflammation, a HP was not expected to be present.[7,13,17,18] However, we found a HP incidence of 63%. While this contrasts with the results of Kim et al,[21] it is comparable with other AGV studies with a reported HP range from 56% to 82%.[13–15,27,36] The reasons behind a HP are speculative, with a possible change in the regulatory pathways leading to greater congestion and edema, as well as an increased fibrotic response due to early exposure of aqueous humor.[13,18,37] However, similar to Nouri-Mahdavi and Caprioli,[18] we believe the term "hypertensive phase" to be misleading as the increase in IOP does not seem to be transient. IOP stabilized at high levels at 3 to 12 months, and required the need for adjunct medications.
The observed number of complications was comparable with other studies as most complications, particularly hyphema, were transient and resolved. Transient hyphema is more common in neovascular glaucoma and has been reported to have an incidence of 65% after a Molteno implant.[23,38,39] However, only 1 of 4 neovascular glaucoma eyes had hyphema, as other differences such as surgical technique or detecting hyphema postoperatively may account for the increased incidence. An advantage of the AGV M4 is the low rate of hypotony and flat anterior chamber in the early postoperative period compared with other GDDs.[7,8,20] Another advantage of the AGV M4 was the low rate of diplopia, which was similarly noted by Kim et al.[21] While previous AGVs have lower incidences of diplopia (6%) than other GDDs, the low-profile bleb of the AGV M4 may further decrease the involvement of extraocular muscles and rate of diplopia.[13,17,20]
A major limitation of this study was its retrospective nature and missing data points, which could underestimate the number of complications at 6 and 12 months and affect the success and failure rates. One of the criteria for inclusion in this study was a follow-up period of >3 months, which is a relatively short period. In addition, the inclusion of eyes with a previous GDIs may have biased our results, as a previous failed tube implant may be associated with increased failure rate of the second implant. In addition, we did not exclude eyes with a history of allergic reaction to IOP-lowering medications or eyes with end-stage glaucoma where a goal of >21 mm Hg might have been acceptable, and this could have affected our results. Moreover, the inclusion of uveitic glaucoma patients may introduce other variables such as inflammation, increased steroid use, or aqueous hyposecretion, which may all affect the results of the study.
Another limitation is the absence of a comparison group. While our results were compared with similar studies such as that of Kim and colleagues, the ABC study, and the AVB study, the criteria for success and failure for each study were different. Kim et al[21] defined success as IOP between 5 and 18 mm Hg and >20% reduction of IOP at all consecutive visits. The ABC study defined failure as IOP≤5 mm Hg and >21 mm Hg, and a <20% reduction of IOP from baseline.[19] The AVB study defined failure as IOP between 5 and 18 mm Hg with ≥20% reduction from baseline for 2 consecutive visits after 3 months.[20] Certainly, the difference in criteria should be taken into consideration when comparing our results with previous studies.
The search for GDDs with a high complete success rate and minimal complications is still ongoing. We wanted to assess whether the new AGV M4 lowers the cumulative probability of failure relative to other GDDs. There is a reported cumulative probability of failure of 31% to 51% for the AGVs and 15% to 34% for the Baerveldt implants at 3 years.[1,7,8] Unfortunately, the M4 did not show marked improvement with a 32% and 72% cumulative probability of failure at 6 and 12 months, respectively, in our study. The AGV M4 effectively lowered IOP during the first postoperative month, but was not able to maintain that degree of IOP lowering thereafter. The surgeons tried to limit the number of glaucoma medication to maintain the quality of life in patients, but the M4 may ultimately require supplementary medication to maintain IOP control. Additional prospective and long-term studies are required to determine the safety and efficacy of the Ahmed M4.
Disclosure
M.R.M. has a nonrestricted research grant from New World Medical Inc. and has lectured for the company. The other authors declare no conflict of interest.
J Glaucoma. 2016;25(5):e475-e480. © 2016 Lippincott Williams & Wilkins
