Endophthalmitis associated with XEN stent implantation

Abstract

Purpose

The aim of this study is to report the clinical characteristics, causative organisms, and treatment outcomes in patients presenting with endophthalmitis related to XEN stent implants

Design

This is a retrospective, noncomparative consecutive case series.

Methods

Clinical and microbiological review was performed for eight patients presenting to the Bascom Palmer Eye Institute Emergency Room from 2021 to 2022 with XEN stent related endophthalmitis. Data collected includes clinical characteristics of patients at presentation, organisms identified in ocular cultures, treatments received, and visual acuity at last follow up.

Results

The current study included eight eyes from eight patients. All cases of endophthalmitis occurred greater than 30 days after implantation of the XEN stent. At time of presentation, there was external exposure of the XEN stent in four out of the eight patients. Five of the eight patients had positive intraocular cultures, all of which were variants of staphylococcus and streptococcus species. Management included intravitreal antibiotics in all patients, explantation of the XEN stent in five patients (62.5%), and pars plana vitrectomy in six patients (75%). At last follow-up, six of the eight patients (75%) had a visual acuity of hand motion or worse.

Conclusions

Endophthalmitis in the setting of XEN stents results in poor visual outcomes. The most common causative organisms are staphylococcus or streptococcus species. At time of diagnosis, prompt treatment with broad spectrum intravitreal antibiotics is recommended. Consideration can be made to explant the XEN stent and perform early pars plana vitrectomy.

Precis/Table of Contents:

Patients receiving XEN stents have a risk of endophthalmitis which may be greater than other minimally invasive glaucoma procedures. When endophthalmitis occurs, it may result in poor visual outcomes despite early and appropriate treatment.

Introduction

Intraocular pressure (IOP) is the main modifiable risk factor in the management of glaucoma. When medical management is inadequate, incisional surgeries are traditionally performed to create a non-physiologic pathway of drainage of aqueous into the subconjunctival space. This is typically accomplished via trabeculectomy or insertion of a glaucoma drainage device.¹ In the past decade, many minimally invasive glaucoma surgery (MIGS) techniques have been introduced. The goal of MIGS surgeries are to provide an alternative to traditional incisional surgery that follows the following tenants: minimally invasive/traumatic, ab-interno delivery, efficacious decrease in intraocular pressure, a favorable safety risk profile and early recovery from surgery for the patient.² Earlier MIGS strategies involved implanting drainage devices into or directly incising the trabecular meshwork to improve aqueous outflow via Schlemm’s canal. Subsequently, implantable devices to create new non-physiologic aqueous drainage pathways into the suprachoroidal and subconjunctival spaces were developed.³ The XEN stent was designed to create a new aqueous drainage pathway to the subconjunctival space - similar to traditional trabeculectomies and glaucoma tube surgeries.

The XEN stent is a flexible, 6-mm long hydrophilic tube composed of porcine gelatin. It comes in three sizes with inner diameters of 140, 63, and 45 μm. The 45 μm size (XEN45) is the version that is preloaded in the XEN Glaucoma Injection System that is approved by the United States Food and Drug Administration (FDA). One benefit of the XEN45 stent over traditional trabeculectomies is that there is a relatively safer theoretical IOP floor of six to eight mmHg, which is dictated by the Hagen-Pouiselle equation.⁴

One of the most feared complications of traditional trabeculectomies or glaucoma drainage devices is the long-term risk of endophthalmitis which a previous review from 2005 found to range from 0.8% to 6.3% (mean of 2.0%).⁵ More recent studies demonstrate a similar rate of roughly 1 to 2%.^6–8 Given that XEN stents also lead to bleb formation, it has been predicted to have a higher rate of endophthalmitis than other non-bleb forming MIGS procedures. The aim of the present study was to report endophthalmitis associated with XEN stent implantation in a consecutive series.

Methods

This consecutive retrospective case series consists of patients with endophthalmitis in the setting of XEN stents. All patients were seen and evaluated at the Bascom Palmer Eye Institute from May 2021 to September 2022. Data collection included baseline demographic information, systemic comorbidities, ocular history, presenting characteristics, intraocular culture results, treatments, and final clinical outcomes. Institutional Review Board (IRB) approval was obtained by the IRB at the University of Miami School of Medicine. The research performed adhered to the tenets of the Declaration of Helsinki.

XEN stent exposure was defined as direct exposure of the XEN stent itself or complete erosion of conjunctiva and Tenon’s capsule within 1 mm of the stent. To calculate mean BCVA, Snellen visual acuity was converted to LogMAR and then back to Snellen equivalents. When visual acuity was count fingers (CF), hand motion (HM), light perception (LP), and no light perception (NLP), LogMAR equivalents of 1.90, 2.30, 2,70, and 3.00, respectively, were assigned based on previously reported estimates.^9,10 Regarding initial treatment, early pars plana vitrectomy (PPV) was defined as receiving a PPV within the first three days of presentation and delayed PPV was defined as receiving a PPV four days or more after initial presentation.

Results

A total of eight eyes from eight patients were identified. The clinical features of these eyes are summarized in Table 1. The average age of these patients was 74.5 years (range 63 to 89 years). One patient had a history of diabetes mellitus. Otherwise, there were no other patients with a history of immunosuppression or other risk factors for endogenous endophthalmitis such as recent infections or hospitalizations, indwelling lines or catheters, other implanted devices, or a history of intravenous drug abuse. Six of the eight individuals had XEN stent placement concurrently with cataract surgery. In all but one case, placement of the XEN stent was the last surgery that was performed on the eye prior to the diagnosis of endophthalmitis. All cases of endophthalmitis occurred at least 30 days after the most recent surgery. The average time elapsed from insertion of the XEN stent to diagnosis of endophthalmitis was 330 days (range 33 to 1498). In four out of eight cases, there was exposure of the XEN stent or erosion of adjacent conjunctiva and Tenon’s capsule within 1 mm of the stent (Figure 1). The mean presenting best corrected visual acuity (BCVA) was HM. On average, eyes were symptomatic for 3.5 days (range 1 to 14) prior to presentation with endophthalmitis.

Table 1.

Presenting clinical features of patients in this series with XEN stent related endophthalmitis.

Case	Gender	Age	Eye	Systemic comorbidities / immunosuppression	XEN implanted with CEIOL	Implant duration (days)	Days of symptoms prior to presentation	VA at presentation	IOP at presentation	XEN exposure†	Conjunctival injection	Hypopyon	Corneal edema	Bleb status	Culture results‡
1	M	76	OS	Diabetes mellitus	Yes	33	6	HM	10	No	Yes	Yes	Yes	Unidentifiable	Negative
2	F	72	OS	None	Yes	131	1	HM	40	Yes	Yes	Yes	Yes	Avascular	Streptococcus oralis, Streptococcus pseudoporcinus
3	M	70	OD	None	Yes	110	14	NLP	18	Yes	Yes	No, but cell/flare	No	Unidentifiable	Negative
4	M	76	OD	None	No	319	1	HM	57	No	Yes	Yes	Yes	Unidentifiable	Staphylococcus aureus
5	M	86	OS	None	Yes	1498	1	LP	42	Yes	Yes	Yes	No	Blebitis	Streptococcus oralis, Streptococcus pseudoporcinus
6	M	64	OS	None	No	159*	1	NLP	58	Yes	Yes	Yes	Yes	Unidentifiable	Granulicatella adiacens, Streptococcus oralis
7	M	89	OS	None	Yes	246	2	20/40	22	No	Yes	Yes	Yes	Small	Negative
8	F	63	OD	None	Yes	144	2	LP	39	No	Yes	Yes	Yes	Blebitis	Streptococcus pneumoniae

^*This was the only patient whose last surgery did not involve XEN implantation. The patient reported he was told he had a retinal detachment at an outside provider and underwent pars plana vitrectomy, although no retinal detachment was identified. The last surgery was 113 days prior to diagnosis with endophthalmitis.

^†XEN exposure defined as direct exposure of stent or complete erosion of adjacent conjunctiva and Tenon’s capsule within one mm of the stent

^‡Culture results shown here are from samples obtained via vitreous taps or cassette fluid from the pars plana vitrectomy if performed

OD = right eye; OS = left eye; CEIOL = Cataract extraction with intraocular lens implantation; PPV = pars plana vitrectomy; VA = visual acuity; IOP = intraocular pressure

Figure 1.

Slit lamp photo of case two that presented with endophthalmitis in the setting of a XEN stent. There is significant purulence and the stent can be seen superonasally (arrow). Immediately superior to the stent, there is erosion of the conjunctiva and Tenon’s capsule down to bare sclera.

All eyes in this series had vitreous biopsies taken (Table 2). All but one eye received intravitreal vancomycin (1 mg) and ceftazidime (2.25 mg) at initial presentation (Table 2). Case four received intravitreal antibiotics the day after initial presentation. This patient initially presented with 4+ cell and fibrin but no hypopyon in the anterior chamber. Additionally, there were few vitreous opacities and no membranes on ocular ultrasound. On follow-up the next day, an aqueous sample that was sent for culture and gram stain was positive for staphylococcus species and a diagnosis of endophthalmitis was made. Five of the eight eyes also received intravitreal dexamethasone (0.4 mg) at least once. Out of these five eyes, one case (case one) received intravitreal dexamethasone at the time of presentation due to a high suspicion for bacterial rather than fungal etiology based off of the clinical exam. The other four eyes received intravitreal dexamethasone at a subsequent visit after preliminary cultures were positive for streptococcus species. One eye received two injections of intravitreal voriconazole (100 mcg) during the treatment course.

Table 2.

Culture results, treatments and final visual outcomes of patients with XEN stent related endophthalmitis.

Case	Vitreous culture source	Culture results	Intravitreal injections (days after presentation)	PPV	Days from presentation to PPV	XEN explanted	Follow-up (days)	Final VA	Final IOP
1	Vitreous tap	Negative	vanc/ceftaz/dex (0) vanc/ceftaz (2)	No	N/a*	No	319	20/60	15
2	Vitreous tap and PPV sample	Streptococcus oralis, streptococcus pseudoporcinus	vanc/ceftaz (0) vanc (1) vanc (3) vanc/dex (36)	Yes	36*	Yes	189	HM	9
3	Vitreous tap and PPV sample	Negative	vanc/ceftaz/vori (0) vanc/ceftaz/vori (3) vanc/ceftaz (46)	Yes	3	Yes	214	HM	34
4	Vitreous tap	Staphylococcus aureus	vanc/ceftaz (1) vanc/ceftaz (4) vanc/dex (12)	Yes	27*	No	215	HM	8
5	Vitreous tap and PPV sample	Streptococcus oralis, streptococcus pseudoporcinus	vanc/ceftaz (0) vanc/ceftaz (1) vanc/ceftaz/dex (4)	Yes	1	Yes	83	NLP	13
6	Vitreous tap and PPV sample	Granulicatella adiacens, streptococcus oralis	vanc/ceftaz (0) vanc/ceftaz/dex (1)	Yes	1	Yes	312	NLP&	-
7	Vitreous tap and PPV sample	Negative	vanc/ceftaz (0) vanc/ceftaz (2)	Yes	2	Yes	98	20/200	19
8	Vitreous tap	Streptococcus pneumoniae	vanc/ceftaz (0) vanc (3)	No	N/a†	No	90	NLP&	-

VA = visual acuity; PPV = pars plana vitrectomy; PPL = pars plana lensectomy, vanc = vancomycin; ceftaz = ceftazidime; vori = voriconazole; HM = hand motion; LP = light perception; NLP = no light perception

^*PPV was not performed within three days of presentation as there was an insufficient view to the posterior pole.

^†PPV was not performed as the anesthesia team was unable to obtain intravenous access for adequate sedation.

^&Patient was enucleated

A total of four eyes underwent early PPV within three days of presentation with endophthalmitis, and two eyes underwent delayed PPV. Cases one, two and four did not undergo early PPV due to an inadequate view secondary to severe corneal edema. Case two underwent a delayed PPV at 36 days due to the identification of a possible retinal detachment on a follow-up ocular ultrasound. Intraoperatively, it was determined to be a pre-retinal post-inflammatory membrane. Case four underwent a delayed PPV and pars plana lensectomy at 27 days due to an anterior dislocation of the lens leading to pupillary block. In one eye (case eight), the anesthesiology team was unable to obtain intravenous access at the time of initial surgery so adequate sedation could not be administered and early PPV was not tolerated. A total of five of the eight eyes underwent explantation of the XEN stent, four of which were exposed on initial exam. Case one did not have XEN stent explantation due to significant improvement following intravitreal antibiotics alone. In the remaining two eyes, removal of the XEN stent was attempted but the XEN stent could not be identified intraoperatively because of corneal opacity (case four) or patient intolerance of the surgery due to inadequate anesthesia (case eight).

The average length of follow-up after initial presentation with endophthalmitis was 190 days (range 83 to 319 days)(Table 2). Five of the eight eyes had positive intraocular cultures. All cases with positive cultures contained a variant of staphylococcus (one of five) or streptococcus (four of five) species, with multi-organism infections occurring in three of the five positive cultures. The mean BCVA by final follow-up was HM. A total of six of the eight eyes had HM or worse vision, which included one eye which was enucleated, one eye which was eviscerated, and one eye that was NLP. Only two out of the eight eyes in this series had an improvement in visual acuity at the last follow up examination. Of note, both of these eyes had negative vitreous cultures.

Discussion

The XEN stent differs from most other MIGS procedures on the market in that it creates a new subconjunctival drainage pathway which leads to the formation of a bleb, similar to glaucoma tube implants and trabeculectomies. The presence, location, and characteristics of a bleb not only poses a risk for blebitis, but also increases the risk of endophthalmitis.¹¹ A major risk factor for endophthalmitis is likely tube exposure after erosion of the overlying conjunctiva and Tenon’s capsule which allows microorganisms to enter the stent and subsequently the eye.

The other currently FDA approved implantable MIGS devices, the Hydrus (Ivantis, Inc) and the iStent (Glaukos Corp), both internally stent open additional outflow pathways via Schlemm’s canal but do not lead to bleb formation. To the best of our knowledge, there are not yet any reported cases of endophthalmitis associated with the Hydrus implant.¹² Cases of endophthalmitis in currently FDA approved non-XEN stent implantable MIGS devices are summarized in Table 3. The largest published case series of iStent related endophthalmitis included nine cases that were all performed with concomitant cataract surgery.¹³ Out of all known reported cases of endophthalmitis in the setting of an iStent device, all but one occurred in the immediate postoperative period which makes it difficult to attribute the endophthalmitis entirely to the iStent rather than to the cataract surgery itself. Of note, Starr et al. also reported one case of endophthalmitis associated with the Cypass (Alcon) device, but this device has been pulled from the market in the United States due to safety concerns regarding higher than expected endothelial cell loss after implantation.¹³

Table 3.

Reported cases of endophthalmitis in non-XEN stent implantable minimally invasive glaucoma surgery devices approved by the United States Food and Drug Administration

Authors	Gender	Age	Type of implant	Duration of implant	BCVA at presentation with endophthalmitis	IOP at presentation with endophthalmitis	Intravitreals received	Pars plana vitrectomy performed	Culture results	Stent explanted	Last recorded BCVA
Chaves et al.32	F	87	iStent inject	4	LP	-	vanc/ceftaz	N	Negative	No	NLP
Lam et al.33	M	60	iStent inject	10	HM	20	vanc/ceftaz/dex	Y	Rothia mucilaginosa	Yes	HM
Starr et al.12	M	75	iStent	135	20/200	18	vanc/ceftaz	Y	Staphylococcus epidermidis	No	20/100
	M	86	iStent	16	20/125	22	vanc/ceftaz	N	Staphylococcus aureus	No	20/25
	F	75	iStent	21	20/200	20	vanc/ceftaz	N	Negative	No	20/200
	M	62	iStent	6	20/100	20	vanc/ceftaz	N	Negative	No	20/25
	F	84	iStent	9	HM	12	vanc/ceftaz	N	Streptococcus sanguinis	No	HM
	M	68	iStent	10	LP	17	vanc/ceftaz	Y	Streptococcus salivarus	No	LP
	F	82	iStent	3	HM	18	vanc/amik	N	Negative	No	20/20
	F	91	iStent	3	HM	31	vanc/ceftaz	N	Streptococcus gordonii	No	NLP
	M	66	iStent	4	CF	14	vanc/ceftaz	N	Streptococcus epidermidis	No	20/25

BCVA = visual acuity; vanc = vancomycin; ceftaz = ceftazidime; amik = amikacin; dex = dexamethasone; CF = count fingers; HM = hand motion; LP = light perception; NLP = no light perception

In contrast with the iStent related endophthalmitis cases, two of the three XEN related endophthalmitis cases identified by Starr et al. occurred greater than 30 days after implantation and were attributed to stent exposure, similar to the cases from the current series.¹³ Another series of XEN stent related endophthalmitis reported five cases with a rate of 1.7%.¹⁴ Also similar to cases in the current study, all five eyes presented with endophthalmitis greater than 30 days after implantation and had an exposed stent at the time of endophthalmitis diagnosis.

XEN stents differ from most MIGS procedures regarding their delivery method. MIGS procedures are typically performed via an ab-interno approach. The XEN stent was also originally FDA approved to be delivered via an ab-interno approach, although many investigators are evaluating off-label ab-externo implantation of the device.^15–19 These studies suggest comparable IOP reduction whether an ab-interno or ab-externo approach is used, but they are underpowered to determine differences in complication rates between the two groups. One key difference between ab-interno and ab-externo approaches is the intentional disruption of the conjunctiva in the ab-externo approach. In the ab-externo approach, the XEN45 injector is inserted through the sclera into the anterior chamber either directly through the conjunctiva or after cutting down the conjunctiva for better exposure.¹⁶ Especially when combined with the use of mitomycin C to promote bleb longevity in either primary implantation or during needling post-operatively, this may lead to an increased risk of endophthalmitis due to poor wound healing. In this case series, all XEN stents were implanted at outside institutions and operative reports were not available. Thus, it is unclear which implantation techniques were used.

XEN stents may also have a relatively higher rate of post-operative needling compared to other traditional bleb forming glaucoma surgeries. In prior studies, rates of needling of XEN blebs were reported to be between 16% and 45%.^15,20–24 This is in contrast with a much lower needling rate of 14% of trabeculectomies in the one year follow-up data from the primary tube versus trabeculectomy (PTVT) study.²⁵ Other studies have also corroborated a higher needling rate in XEN stents when compared head-to-head with trabeculectomies.^23,24 Needling may represent another opportunity for microbes to enter the eye and also increase the risk of endophthalmitis.

In this cohort of patients, the relatively high rate of stent exposure suggests that, like traditional glaucoma drainage devices, exposure of the XEN stent is a major risk factor for endophthalmitis.²⁶ The exclusively staphylococcus and streptococcus positivity in this series is also consistent with other studies evaluating patients with endophthalmitis in the setting of glaucoma that demonstrate high rates of skin flora as the causative organism.^8,27,28 The relatively high rate of streptococcus positivity in this cohort may contribute to the poor visual outcomes. A prior study from Staropoli et al. demonstrated particularly poor visual prognosis for patients with streptococcal endophthalmitis despite early and appropriate antibiotic treatment.²⁹ In this current series, all three eyes which had a final visual acuity of no light perception had cultures positive for streptococcus species.

Regarding treatment, this series is too small to draw any conclusions regarding specific management decisions including the utility of dexamethasone, XEN stent explantation, or early PPV. However, prior work in endophthalmitis secondary to other causes can help guide management decisions. A Cochrane review reported weak evidence supporting the use of dexamethasone in cases of acute onset endophthalmitis.³⁰ The decision to inject intravitreal steroids in patients with XEN stent related endophthalmitis can be made on a case-by-case basis, and intravitreal steroids should be avoided in cases where there is a suspicion for fungal etiologies. Of note, the case in this series that had the largest improvement in visual acuity from HM to 20/60 received intravitreal dexamethasone at presentation. However, more evaluation of early dexamethasone administration is necessary in this patient population prior to defining any clear treatment recommendations. Regarding the utility of stent explantation, prior work from Islam et al. suggests that explantation of traditional glaucoma drainage devices in the setting of endophthalmitis reduces rates of phthisis, evisceration, and enucleation.³¹ The same group also reported that, among a total of 91 eyes reviewed, no eye that received immediate PPV became phthisical or underwent evisceration/enucleation.³¹ A large retrospective chart review comparing patients with bleb-related endophthalmitis that were treated with PPV versus vitreous biopsy and injection of intravitreal antibiotics also suggested a benefit in treatment with PPV.³² In this series, there was no clear association between XEN stent explantation or management with early PPV and final visual outcomes. These results are likely limited by the small sample size. Finally, given that all patients in this series experienced endophthalmitis after the first 30 days following surgery, results from this case series do not apply to cases of acute endophthalmitis following combined cataract extraction and XEN stent implantation.

Conclusion

Compared to other published reports on MIGS procedures, XEN stents may have a higher risk of endophthalmitis due to the creation of a new non-physiologic aqueous drainage pathway and resulting bleb formation. As with any glaucoma surgery, pre-operative counseling on risks of endophthalmitis is recommended. Care should be taken when planning on inserting a XEN stent in patients with history of poor wound healing or difficulty with follow up.

Similar to other bleb-creating procedures, the most common organisms in XEN stent related endophthalmitis are staphylococcus and streptococcus species. Endophthalmitis in the setting of XEN stents has a very poor visual prognosis. Based on the currently available experience with these cases, treatment of these patients is similar to other patients with endophthalmitis secondary to traditional glaucoma drainage devices. Prompt injection of intravitreal antibiotics is warranted for broad spectrum coverage of common staphylococcus and streptococcus species. Intravitreal dexamethasone can also be used. Early pars plana vitrectomy and removal of XEN stent can be considered, although there is mixed evidence that these are beneficial in patients with XEN stent related endophthalmitis.

Biographies

Elena Bitrian biosketch:

Elena Bitrian, M.D., is an Associate Professor of Ophthalmology at Bascom Palmer Eye Institute in Miami, FL, who specializes in adult and pediatric glaucoma surgery. She is one of the editors of the textbook Surgical Techniques in Childhood Glaucoma and the compendium on Childhood Glaucoma and is also the Director of Education for the Balkan International Pediatric Glaucoma Center at Bascom Palmer.

Benjamin Lin biosketch:

Benjamin Lin, M.D., is an ophthalmology resident at the Bascom Palmer Eye Institute in Miami, FL. He is the co-founder and executive director of EyeGuru.org, a nonprofit website dedicated to basic ophthalmic education. He will be staying at Bascom Palmer for the next two years to complete his fellowship training in surgical retina as well as a year as chief resident.