Purpose
To identify factors associated with the successful treatment of malignant glaucoma (MG).
Design
Retrospective case series.
Methods
Setting: single institution; study population: 64 eyes (55 subjects) with MG; observation procedure(s): chart review; main outcome measures: anatomy, intraocular pressure (IOP), best visual acuity (BVA).
Results
87.5% (n=56/64) of eyes with MG required surgical intervention. Vitrectomy was more likely to be successful in eyes with a history of <3 incisional surgeries, <3 glaucoma drops, or IOP <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='≤’>≤≤
≤
30 mm Hg ( P < .05). If vitrectomy was performed within 30 days, recovery of anatomy, BVA, and IOP occurred sooner ( P < .05). IOP reduction was greater in subjects treated with oral carbonic anhydrase inhibitors ( P = .016) or Nd:YAG laser hyaloidotomy ( P = .007), and without a history of MG ( P = .007). Time to maximal improvement was significantly longer for IOP and BVA than anatomy ( P < .001). Treatment of MG with an oral carbonic anhydrase inhibitor hastened anatomic recovery ( P = .01). Time to improvement in BVA was significantly faster in men and African Americans ( P < .05). Time to maximal reduction in IOP occurred sooner in eyes that underwent anterior chamber reformation in clinic ( P < .002). Trabeculectomy surgery prior to MG was associated with prolonged recovery of anatomy, BVA, and IOP ( P < .05).
Conclusions
Earlier vitrectomy may shorten recovery times for MG. Nd:YAG laser hyaloidotomy and oral carbonic anhydrase inhibitors may lead to greater IOP reduction. The time to maximal improvement in IOP and BVA may be longer than the time to anatomic resolution. Although trabeculectomy may impede time to recovery from MG, oral carbonic anhydrase inhibitors may shorten the time to anatomic recovery and anterior chamber reformation may hasten IOP recovery.
Malignant glaucoma (MG) is an uncommon, vision-threatening form of secondary angle closure glaucoma that typically develops following incisional cataract or glaucoma surgery in eyes with chronic narrow angle glaucoma, plateau iris, and nanophthalmos. Initial medical management of MG includes both aqueous suppression and cycloplegia via a combination of phenylephrine, anticholinergic, and antihypertensive agents. In pseudophakic eyes, neodymium-doped yttrium-aluminum-garnet (Nd:YAG) laser anterior hyaloidotomy also may be attempted, although successful disruption of the hyaloid face can be unsuccessful or yield only a temporary benefit. Definitive surgical treatment typically involves vitrectomy by either an anterior or pars plana approach to disrupt the anterior hyaloid face and maneuvers that establish a unicameral chamber. A majority of the publications on MG have described the efficacy of a single procedure for small case series or cohorts with short follow-up. Several larger reviews with longer follow-up have recently been published about cohorts in China (N=50 eyes), India (N=58 eyes), and Saudi Arabia (N=69), all of which demonstrated anatomic and functional recovery from MG following pars plana vitrectomy.
As a tertiary referral center for the southeastern United States, our institution treats a large white and African American population. Factors associated with successful management of MG in this population have not been previously examined. Moreover, to our knowledge, no study has considered which factors impact the time to recovery from MG or whether time to anatomic resolution may differ from time to maximal improvement in intraocular pressure (IOP) or best visual acuity (BVA). The purpose of this study was to describe the management patterns of MG at a single institution over a 10-year time period, and identify factors associated with functional and anatomic recovery as well as the time to recovery.
Methods
This is a 10-year retrospective case series of patients who were treated for MG by the Duke Glaucoma Service between 2007 and 2017. The study was approved by the Institutional Review Board and was conducted in accordance with the Tenets of Helsinki. A waiver of informed consent was provided because of the retrospective nature of this study.
Subjects were identified by an ICD-9/ICD-10 code search for “malignant glaucoma” and “aqueous misdirection” using the Duke Enterprise Data Unified Content Explorer (DEDUCE) portal. Included subjects were diagnosed with MG, which was defined as an elevation of IOP and a shallowing of the peripheral and central anterior chamber (AC) that persisted despite the presence or placement of a patent peripheral iridotomy and in the absence of an alternative etiology, such as a posterior segment mass, suprachoroidal hemorrhage, ciliary effusion, or pupillary block. B-scan, ultrasonographic biomicroscopy, and anterior segment–optical coherence tomography images were reviewed when available. All cases identified by the ICD code search demonstrated an elevation in IOP at the time of MG. Included subjects had at least 1 month of follow-up. Only 4 subjects were excluded owing to inadequate follow-up and documentation. Data were collected on history of glaucoma, history of prior episode of MG in either eye, number of prior intraocular surgeries, type of surgery that immediately preceded the onset of MG, and number of glaucoma medications prior to onset of MG. Fixed-combination dose topical glaucoma medicines (eg, dorzolamide-timolol) were counted as 2 medications. Also, characteristics at presentation with MG and during follow-up, such as Snellen best visual acuity converted to logarithm of the minimum angle of resolution (logMAR), IOP, and lens status were recorded. BVA was recorded as the best corrected visual acuity on refraction, unless the patient had very low vision (eg, counting fingers) and could not be refracted, in which case the pinhole visual acuity was recorded. The management of MG was extracted from operative reports and progress notes. Medical management was defined as any treatments or interventions that were performed in the clinic on diagnosis of MG in an attempt to resolve the episode without returning to the operating room. They included use of topical and oral glaucoma medications, topical cycloplegics (eg, atropine, cyclopentolate), and topical phenylephrine, with or without Nd:YAG laser posterior capsulotomy with anterior hyaloidotomy and AC reformation with cohesive viscoelastic at the slit lamp. The use of an oral carbonic anhydrase inhibitor to treat MG either medically or preoperatively up until the time of vitrectomy, if performed, was recorded. The doses of medications and performance of these procedures were at the discretion of the attending surgeon.
Incisional surgical management was categorized into cases that included anterior vitrectomy (AV) and/or pars plana vitrectomy (PPV). AV was performed using a coaxial 21-gauge vitrector in eyes that were either pseudophakic or that underwent same-day phacoemulsification with intraocular lens placement. After inflation of the capsular bag with cohesive viscoelastic, a posterior capsulotomy was created either with a cystotome and Utrata forceps or with the vitrector, and the anterior hyaloid face was disrupted followed by core vitrectomy by an anterior approach. Irido-zonulo-hyaloidotomy was also performed in eyes without a patent iridotomy or the vitrector was advanced through the existing iridotomy, then through the peripheral zonules and into the anterior hyaloid. When PPV was pursued the decision to perform a full or core PPV was at the discretion of the retina surgeon. All PPV in this study used a standard 3-port technique, with placement of a pars plana infusion cannula in the inferotemporal quadrant followed by either a 23-gauge or 25-gauge vitrector and light pipe in the 2 superior quadrants. In all cases, care was taken to disrupt the anterior hyaloid face, and the posterior capsule was also opened in pseudophakic patients or those undergoing simultaneous lensectomy. Hyaloido-zonulo-iridectomy was then performed to establish a unicameral chamber by passing the vitrector through the hyaloid, up through the zonules and through the peripheral iris or an existing iridotomy. All sclerotomies were sutured. Vitrectomy could also be combined with anterior chamber reconstruction, glaucoma drainage device (GDD) implantation with placement of the tube at the level of the pars plana, or endocyclophotocoagulation (ECP), at the discretion of the surgeon.
Every patient was examined by a glaucoma attending with the examination and note updated and signed by the attending. Anatomic resolution of MG was defined as deepening of the central and peripheral AC on slit lamp examination that was maintained off cycloplegia. Complete resolution was defined as anatomic resolution with reduction in IOP to <22 mm Hg. The lowest IOP, number of glaucoma medications, and logMAR BVA achieved following treatment of MG was recorded. Maximal reduction in IOP was defined as the difference in the IOP pre-intervention and the nadir of the IOP post-intervention.
Statistical Analysis
Statistical analysis was performed in Stata 15.1 (StataCorp, Inc., College Station, Texas, USA). Univariate and multivariate regression analysis was used to evaluate factors associated with the change in IOP, logMAR best visual acuity, and number of glaucoma medications following successful management of MG. Logistic regression was also performed to determine factors that predicted complete resolution of MG (ie, improvement in anatomy and sustained reduction of IOP to <SPAN role=presentation tabIndex=0 id=MathJax-Element-2-Frame class=MathJax style="POSITION: relative" data-mathml='≤’>≤≤
≤
22 mm Hg) following vitrectomy (AV and/or PPV). Independent variables that reached a P value of <.05 in univariate analysis were tested in multivariate analysis. Both the linear and logistic regression analyses were performed with generalized estimating equations using an exchangeable correlation structure to account for more than 1 eye within subjects.
Time-to-event analysis was also performed. Univariate and multivariate Cox proportional hazards models with cluster-robust standard errors to account for eyes nested within subjects were constructed to evaluate which demographic and clinical factors were associated with faster time to anatomic resolution of MG, time to maximal reduction in IOP, and time to greatest improvement in best visual acuity. Kaplan-Meier curves of time to improvement in anatomy, BVA, and IOP were plotted for significant factors.
Results
Sixty-four eyes in 55 subjects presented with MG over the 10-year time period. Nine subjects (16.4%) developed MG in both eyes. Table 1 shows the characteristics and management of eyes with MG. Approximately three-quarters of eyes had a history of angle closure glaucoma, and glaucoma surgery was the most common surgery preceding the onset of MG. However, 2 eyes had undergone pars plana vitrectomy in the past, one to repair a retinal detachment and the other as part of a combination penetrating keratoplasty with placement of a glaucoma drainage device tube in the pars plana. Most eyes (60.94%, n=39/64) were diagnosed with MG within 30 days of the precipitating surgery. All eyes underwent initial medical treatment except for 1 eye that underwent immediate PPV due to intraoperative diagnosis of MG. Clinic-based medical and/or laser interventions were successful in managing 12.5% (n=8/64) of MG. Five of these eyes had chronic narrow angle glaucoma and 3 had primary open angle glaucoma. The preceding surgeries varied and included 5 cataract-glaucoma combination surgeries, 1 trabeculectomy, 1 Ahmed, and 1 DSEK; also, 1 patient underwent a YAG capsulotomy and 2 had AC reformation in clinic.
| Variable (N=64 eyes in 55 Subjects) | |
|---|---|
| Age, y, mean ± SD | 65.3 ± 16.8 |
| Sex, n (%) | |
| Female | 46 (71.9) |
| Male | 18 (28.1) |
| Race, n (%) | |
| African American | 32 (50) |
| White | 32 (50) |
| Glaucoma diagnosis, n (%) | |
| Angle closure glaucoma | 46 (71.9) |
| Open angle glaucoma | 16 (25) |
| No glaucoma | 2 (3.1) |
| Lens status, n (%) | |
| Pseudophakic | 47 (73.4) |
| Phakic | 17 (26.6) |
| History of <3 incisional surgeries, n (%) | 54 (84.4) |
| History of MG, n (%) | 14 (21.9) |
| IOP > 30 mm Hg prior to surgery on glaucoma medication, n (%) | 22 (34.4) |
| Number of glaucoma drops prior to MG, mean ± SD | 2.6 ± 1.56 |
| <2 glaucoma drops prior to MG, n (%) | 19 (29.7) |
| Diagnosis of MG in <SPAN role=presentation tabIndex=0 id=MathJax-Element-3-Frame class=MathJax style="POSITION: relative" data-mathml='≤’>≤≤ ≤ 30 d of preceding surgery, n (%) | 39 (60.9) |
| Surgery preceding MG, n (%) | |
| Cataract surgery | 5 (7.8) |
| Corneal transplant ± cataract surgery | 5 (7.8) |
| Penetrating keratoplasty + cataract | 1 (1.6) |
| DSEK/DMEK + cataract | 1 (1.6) |
| DSEK/DMEK | 3 (4.7) |
| Glaucoma surgery + cataract surgery | 16 (25.0) |
| Trabeculectomy + cataract | 5 (7.8) |
| Glaucoma drainage device + cataract | 10 (15.6) |
| Cyclophotocoagulation + cataract | 1 (1.6) |
| Glaucoma surgery | 36 (56.3) |
| Trabeculectomy | 10 (15.6) |
| Glaucoma drainage device | 23 (35.9) |
| Cyclophotocoagulation | 3 (4.7) |
| Other surgery a + pars plana vitrectomy | 2 (3.1) |
| Treatment of MG, n (%) | |
| Medical | 8 (12.5) |
| Surgical | 56 (87.5) |
| Pars plana vitrectomy | 47 (73.4) |
| Anterior vitrectomy | 7 (10.9) |
| Anterior and Pars plana vitrectomy | 2 (3.1) |
| Number of glaucoma drops to treat MG, mean ± SD | 2.42 ± 1.38 |
| Anterior chamber reformation in clinic to treat MG, n (%) | 27 (42.2) |
| Nd:YAG laser hyaloidotomy in clinic to treat MG, n (%) | 14 (21.9) |
| Oral carbonic anhydrase inhibitor to treat MG, n (%) | 18 (28.1) |
a One patient underwent retinal detachment repair with pars plana vitrectomy and C3F8 gas. The other patient underwent a limited pars plana vitrectomy to allow tube placement in the pars plana during a combined glaucoma drainage device and penetrating keratoplasty surgery.
However, 87.5% (n=56/64) of eyes required additional incisional surgical intervention. Complete resolution of MG occurred in 98.4% of eyes (n=63/64). The single case of MG that persisted despite medical and surgical interventions eventually underwent enucleation. In 1 eye that had a remote history of PPV, prior to the onset of MG, the MG was successfully managed with medical therapy without further surgical intervention. In a minority of cases, PPV or AV were combined with cataract extraction (CE; n=5 and n=7), GDD placement (n=6 and n=1), or ECP (n=1 and n=1, respectively). Also, surgical reconstruction of the anterior chamber was performed at the time of PPV and/or AV in 26 eyes. Initial surgical treatment was successful in 100% (n=5/5) of eyes that underwent PPV with CE, 66.7% (n=4/6) of phakic eyes that underwent PPV without CE, 81.1% (n=30/37) of pseudophakic eyes that underwent PPV, 83.3% (n=5/6) of eyes that underwent AV with CE, and 100% (n=2/2) of pseudophakic eyes that underwent AV. If MG failed to resolve following the first surgery, a second surgery was attempted. MG successfully resolved in 3 pseudophakic eyes that had a repeat PPV, 1 eye that underwent PPV after failed AV with CE, 1 eye that underwent AV with CE after a failed PPV, and 5 eyes that underwent a combination of other surgical procedures. MG did not recur following successful intervention.
Factors Associated With Resolution of Malignant Glaucoma After Vitrectomy
When MG was managed with AV and/or PPV, vitrectomy was more likely to result in complete resolution of MG (n=49) if the eye had a history of <3 prior incisional surgeries (odds ratio [OR] 9.5, P = .01), had been on fewer than 3 glaucoma medications (OR 10.6, P = .036), or had an IOP ≤30 mm Hg (OR 11.8, P = .031) with medical treatment at baseline, prior to the surgery that precipitated MG. Pseudophakic lens status was not predictive of vitrectomy success (OR 0.37, P = .38). Prior Nd:Yag hyaloidotomy also did not increase the likelihood of successful resolution with vitrectomy (OR 0.09, P = .008). Intraoperative AC reconstruction was associated with a nonstatistically significant increased chance of success (OR 5.4, P = .14). There was no significant association between vitrectomy success and whether the onset of MG occurred within 30 days of precipitating surgery (OR 1, P = .9). If the first vitrectomy was performed within 30 days of the diagnosis of MG (n=41/56, 73.2%), there was a higher likelihood of successful resolution of MG, but this association did not reach statistical significance (OR 3.35, P = .17). In a multivariate model, both a history of fewer than 3 incisional surgeries (OR 12.2, P = .014) or fewer than 3 glaucoma medications (OR 13.2, P = .036) remained statistically significant.
Factors Impacting the Change in Best Visual Acuity, Intraocular Pressure, and Number of Topical Glaucoma Medications
Treatment of MG resulted in a significantly improved BVA, lower IOP, and decreased number of glaucoma medications (all P < .001) ( Table 2 ). The change in IOP, BVA, or number of glaucoma medications after treatment did not significantly vary by age, race, gender, whether onset of MG occurred within 30 days of prior surgery, or whether the eye was managed medically or surgically (all P > .10) ( Table 3 ). Improvement in BVA was significantly associated with a history of being on <2 glaucoma medications prior to the onset of MG ( P = .002). Eyes that underwent Nd:YAG laser hyaloidotomy in clinic experienced a smaller improvement in BVA ( P = .048) but a greater improvement in IOP ( P = .007). Reduction in IOP was greater in eyes without any prior history of MG ( P = .007) and in subjects treated with oral carbonic anhydrase inhibitors (CAIs) ( P = .016). In multivariate analysis, use of <2 glaucoma medications prior to MG and Nd:YAG laser hyaloidotomy were independently associated with the change in BVA (all P < .05). A prior history of MG, Nd:YAG laser hyaloidotomy, and use of oral CAI were independently associated with the change in IOP (all P < .05).
| Variable | Before Treatment, Mean ± SD | After Treatment, Mean ± SD | Beta Coefficient (95% CI) | P Value a |
|---|---|---|---|---|
| Best visual acuity, logMAR | 1.73 ± 0.86 | 1.14 ± 0.99 | –0.58 (–0.80, –0.36) | <.001 |
| Intraocular pressure, mm Hg | 32.5 ± 18.5 | 11.5 ± 4.2 | –21.4 (–26.2, –16.6) | <.001 |
| Number of glaucoma drops | 2.42 ± 1.38 | 1.46 ± 1.32 | –0.96 (–1.32, –0.60) | <.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
