The Role of Vitrectomy in the Management of Posterior Uveitis

An overview of indications and two case studies.

Angela K. Au, MD • David M. Hinkle, MD • Naomi S. Falk, MD, OD

Uveitis remains a devastating and sight-threatening group of diseases. Worldwide, uveitis is thought to be responsible for 10% of cases of all blindness.¹ The field of uveitis is one that has changed significantly over time and continues to evolve.

There has been a paradigm shift from the use of steroids as the mainstay of treatment for years to the current use of steroid-sparing immunomodulatory agents, which has significantly increased the safety and efficacy of pharmacologic treatment for uveitis.^2,3

There remains considerable disagreement and uncertainty regarding the role of vitrectomy in the management of uveitis and the complications of uveitis. In 2005, Matthias Becker and Janet Davis published a thorough review of the literature from 1981 to 2005 involving vitrectomy in the treatment of uveitis. They found that although vitrectomy was almost always recommended, strong evidence, in the form of randomized, controlled trials to indicate the utility of this approach, has been lacking.⁴

Since then, no large, prospective, controlled trials have shown clear benefits of vitrectomy, but many small studies have suggested possible benefits. In this article, we review the role of vitrectomy in posterior uveitis.

DIAGNOSTIC VITRECTOMY

Proper diagnosis is essential to the treatment of uveitis. Ophthalmologists look at a variety of factors to determine the etiology of ocular inflammatory disease, such as history, ophthalmic examination, systemic findings, laboratory tests, and therapeutic response to medications.

However, cases present atypically or are not responsive clinically to treatment. Such cases can pose a diagnostic dilemma. Specifically, primary intraocular lymphoma is a disease in which diagnosis is often delayed or missed. It has important implications as patients have been shown to develop CNS lymphoma in 65% to 90% of cases.⁵

Diagnostic vitrectomy has been shown to be especially useful in ruling out malignancy and infection.⁶ A review of the literature showed that diagnostic vitrectomy is a safe technique. However, the reported diagnostic yield varies widely between 14.3% and 61.5% depending on case selection, surgical technique, and testing used.⁷

The usual method of obtaining a specimen during vitrectomy involves placing ports, connecting a syringe to the cutter, and taking an undiluted sample of vitreous before the infusion is turned on. A specimen is collected until the eye visibly softens — approximately 2 mL of sample can safely be retrieved in this manner. Some surgeons infuse air or perfluorocarbon liquids to avoid hypotony and increase the volume of vitreous removed. The infusion is turned on, and an additional sample of diluted vitreous can be collected in the cassette.^6,7

CLINICAL PICTURE AS A GUIDE

It is important that the clinical picture guide the selection of tests to be carried out on a limited vitreous sample. Furthermore, the utility of diagnostic vitrectomy is dependent on the availability of these laboratory tests and experienced pathologists to interpret the results.⁷ We attempt to give our lab advance notice of a planned diagnostic vitrectomy to reduce the risk of processing errors, and some specimens are sent out to specialized labs for analysis.

Cytology remains the gold standard to identify malignant cells, and its sensitivity for malignancy ranges from 31% to 66.7%.⁶ The challenge with cytology is that lymphoma cells may be sparse and can easily degrade, so an undiluted specimen should be sent immediately to a pathologist.

Newer techniques allow for increased diagnostic yield. Flow cytometry is a useful tool to determine which clusters of differentiation (CD) are present in a specimen. Molecular analysis can be used to amplify immunoglobulin heavy chain genes in B lymphoctyes and to look for rearrangements and translocations in T lymphocytes, which can have prognostic and diagnostic implications in lymphoma.

Cytokines are another tool that can be suggestive of malignancy. Interleukin (IL)-10 is associated with B-cell malignancies, whereas IL-6 is associated with inflammatory disease, so primary intraocular lymphoma tends to have an IL-10-to-IL-6 ratio of greater than 1.

Cultures and polymerase chain reaction (PCR) are other standard approaches to diagnosing infectious etiologies for uveitis. Antibodies to organisms can also be detected. However, they may have less utility as they may not be present until late in the disease course.⁷

When vitrectomy fails to reach a diagnosis, retinal and choroidal biopsies can be helpful. In a series of 13 cases published by Johnston, the clinical diagnosis changed in five patients; in seven patients, treatment was guided based on the diagnosis; and in the six patients in whom diagnosis was not determined, malignancy was ruled out. The complications noted were intraoperative break, postoperative retinal detachment, subretinal hemorrhage, cataract, and phthisis.⁸

THERAPEUTIC VITRECTOMY

Patients with uveitis develop many posteriorsegment complications, including retinal detachments, epiretinal membranes, macular holes, and vitreous opacities. Specifically, vitreous opacities can significant limit a patient’s vision and the clinician’s view of the posterior segment.

The indications for vitrectomy in these cases are similar to eyes without inflammation. However, many retina specialists remain hesitant to perform retinal surgery in the setting of uveitis due to increased complications and lower rates of success.

One should avoid vitrectomy in the setting of active uveitis, as the retina can be very friable. Surgery by nature is proinflammatory, and even eyes with well-controlled uveitis will likely flare after vitrectomy. In the perioperative period (a few days before vitrectomy to a few weeks after), as with cataract surgery in a patient with uveitis, aggressive control of inflammation is essential and may help prevent complications.

Figure 1. Montage fundus photo of the left eye showing peripheral retinal whitening and vasculitis.

Figure 2. Montage fundus photo showing progression of the patient’s retinitis. The view is limited by vitritis.

In the case of autoimmune uveitis, we recommend treating with full therapeutic doses of either systemic prednisone/solumedrol or concurrent peribulbar/intravitreal triamcinolone acetonide at the time of surgery.

Figure 3. Montage fundus photo of the patient from Figures 1 and 2 two weeks after pars plana vitrectomy. The view is greatly improved, and the retinal whitening has progressed significantly compared to the initial lesion.

In the case of infectious etiologies, such as herpes, toxoplasmosis, or cytomegalovirus, we recommend increasing medications to full therapeutic levels perioperatively.

The view in uveitic eyes can be very poor during surgery due to posterior synechiae and media opacities. There is also a risk of ciliary body hyposecretion and resultant hypotony.⁹ The vitreous can be very adherent to the retina in uveitic eyes, and peeling of membranes can easily lead to central retinal tears.

Techniques to minimize complications in uveitic eyes include tangential peeling of membranes and the use of triamcinolone to help identify and remove vitreous.¹⁰ By restoring the posterior segment anatomically, pars plana vitrectomy can improve vision in patients with uveitis.

There have been reports suggesting that newer 25-gauge vitrectomy techniques may be better for uveitic vitrectomy. The rationale is that smaller ports and sutureless sclerotomies may result in less postoperative inflammation. Small case studies have shown 25-g vitrectomy to be safe in the setting of uveitis.^11,12

Removing the vitreous gel alone has been postulated to be therapeutic in uveitis. This was first proposed after studies by Diamond and Kaplan in 1979 showed improved outcomes in vision and macular edema after pars plana vitrectomy was performed, in conjunction with cataract removal via pars plana lensectomy.¹³

More recent papers have reported a decrease in inflammatory activity and a decrease in uveitis flares after vitrectomy, along with an increase in visual acuity. Investigators have observed a reduction in cystoid macular edema (either detected by angiography or by optical coherence tomography) after vitrectomy.⁴

The exact mechanism of improvement after vitrectomy is unknown. There are theories that inflammatory cells and mediators (including immune complexes and cytokines) are not easily cleared from the eye and remain in the vitreous, which can cause persistence of inflammation or reactivation.^9,10,14 Additionally, it has been hypothesized that vitrectomy can change oxygen gradients and remove the antigen in some cases of infection.

However, it is important to note that removal of the vitreous gel can increase the speed of clearance and decrease the duration of the therapeutic effect of intravitreal medications, such as anti-VEGF agents and corticosteroids.¹⁰

CASE STUDIES

CASE 1

A 79-year-old woman with a history of hypertension and hyperlipidemia was referred to our practice by her ophthalmologist for a retinal tear in the left eye. She had noted floaters in that eye for one to two weeks.

In the left eye, visual acuity was 20/40, and on exam, a large retinal tear was noted, with an adjacent area of retinal whitening and vasculitis (Figure 1). There was 2−3+ vitreous cells and debris noted. Fluorescein angiography showed optic nerve leakage and vascular leakage.

The patient had a recent history of cold sores, so empiric treatment was started with valacyclovir (Valtrex) for presumed herpetic acute retinal necrosis. A laser retinopexy was performed for the tear, and an initial anterior-chamber tap was performed, although it was inconclusive (CMV PCR negative, HSV/VZV PCR not done by the lab).

An initial systemic workup revealed only IgG positive for toxoplasmosis (IgM and all other labs and studies were negative). Prednisone was started to help decrease the inflammation, and sulfamethoxazole and trimethoprim (Bactrim) was added to cover for toxoplasmosis.

Initially the patient’s vitritis decreased. However, after a month on the same regimen, the vitritis worsened, and an additional anterior-chamber paracentesis was performed. The sample was negative for CMV/HSV/VZV by PCR, but given the clinical picture, it was thought that result could have been a false negative for HSV.

Therefore, the patient was given three intravitreal ganciclovir injections. After the injections, her inflammation appeared to subside. Her vision was 20/30, and her retinal whitening was beginning to look inactive. The decision was made to taper her prednisone while continuing the Bactrim and Valtrex.

A week after the patient was tapered off the prednisone, she again reactivated, and her vision decreased to 20/70. The patient was admitted to the hospital for a consultation with infectious disease. An anterior-chamber paracentesis was done, and toxoplasmosis PCR was found to be negative. The patient was thought to show ganciclovir resistance and was given a course of intravenous acyclovir and three intravitreal foscarnet injections.

Figure 4. FA showing diffuse choroidal lesions and optic nerve leakage. The view is hazy due to vitritis.

Figure 5. FA showed the vitritis had improved, and the choroidal lesions have now become inactive and scarred.

Despite intravitreal foscarnet, her retinitis and vitritis continued to progress, and her vision decreased to counting fingers (Figure 2). Off the prednisone, it was clear that the patient failed to respond clinically to treatment for both HSV and toxoplasmosis.

At that point, the differential diagnosis for the patient’s uveitis included HSV, toxoplasmosis, CMV, and malignancy. Given the lack of response to treatment and no clear diagnosis, the decision was made to submit her to diagnostic vitrectomy.

Cytology was diagnostic for primary intraocular lymphoma, and the IL-10-to-IL-6 ratio was 1.48, supporting this diagnosis. The patient was referred to oncology for a metastatic workup. Two weeks after the vitrectomy, the patient’s vision improved to 20/200, and the view improved significantly (Figure 3).

Figure 6. B scan revealing a large choroidal granuloma.

Figure 7. After treatment with isoniazid, rifampin, and prednisone, the choroid remains thickened, but the choroidal granuloma has resolved.

CASE 2

A 27-year-old woman from Bangladesh was referred by her ophthalmologist for uveitis. She had no past medical history but had been given the Bacillus Calmette–Guérin (BCG) vaccine in Bangladesh at the age of 10. The patient reported two months of redness, pain, and blurred vision in both eyes. She had been treated with topical cycloplegic and prednisolone drops.

On exam, her vision pinholed to 20/80 in the right eye and 20/40 in the left. She had fine KPs, 2+ anterior chamber cell/flare, and 3+ vitreous cell and debris. Initial FA showed patchy choroidal lesions and optic nerve leakage (Figure 4). Initial lab tests had been ordered by the referring ophthalmologist and showed only a positive Lyme Ab IgM. Additional quantiferon gold and CXR were ordered.

The patient did not follow up for a year and did not get the additional workup. In that time, her vision had dropped to 20/200 in both eyes. The patient’s topical steroid was changed to Durezol (difluprednate), after which her inflammation improved. The quantiferon gold came back positive, and the CXR was negative.

After consultation with infectious disease, an AC tap was performed for mycobacterium tuberculosis PCR and AFB culture/smear. All were negative. The decision was still made to initiate triple antibiotic therapy and prednisone.

However, after a month on the triple therapy, she had elevated liver enzymes, and her regimen was changed to single-agent isoniazid. Her retinal lesions and vitritis improved with treatment for five to six months (Figure 5). The prednisone was tapered, and the topical steroid and cycloplegic were stopped.

After these medications were stopped, the patient’s lesions appeared to reactivate. She again had significant inflammation, and her prednisone was restarted. At this time, she became pregnant, and therefore her isoniazid was stopped. With the confounding circumstance of pregnancy and difficulty with treatment, diagnostic vitrectomy was considered for a definitive diagnosis, but the patient chose to defer all treatment until after her pregnancy.

She returned seven months later after her baby was born, and her vision in the right eye had decreased to light perception with no view. The left eye vision was 20/150, and the lesions appeared inactive. B scan revealed a large granulomatous lesion in the right eye (Figure 6).

Diagnostic vitrectomy was ultimately performed, and a vitreous sample was sent for analysis. Cytology was negative for malignancy, noting only lymphocytes and histiocytes. B- and T-cell gene rearrangement showed a polyclonal pattern, again ruling out lymphoma. The Lyme, Treponema pallidum, and mycobacterium tuberculosis PCR were all negative.

Despite the inconclusive diagnostic vitrectomy, the decision was made to reinstate isoniazid, rifampin, and prednisone treatment. There was massive improvement in her uveitis on this regimen; her vision improved to counting fingers, and her granuloma resolved (Figure 7).

Despite the excellent response, she had to stop her tuberculosis treatment as her liver enzymes have been elevated and she was breastfeeding. Infectious disease consultation determined no safe alternative treatment for tuberculosis while the patient was nursing.

In this case, diagnostic studies and vitrectomy were unable to lead to a definitive diagnosis. While the vitrectomy was helpful in ruling out malignancy, it did not help in guiding treatment. We have had to rely on the patient’s clinical picture and her therapeutic response to antituberculosis medications. We continue to manage this patient as uveitis from presumed tuberculosis.

CONCLUSIONS

Pars plana vitrectomy has a role in the diagnosis and management of uveitis. Large randomized, controlled trials are needed to define and validate further the utility of vitrectomy. Diagnostic vitrectomy can help to differentiate among inflammation, infection, and malignancy and can help guide the treatment of uveitis. Newer lab tests are available to increase the diagnostic yield of vitrectomy and may require special handling and interpretation.

Therapeutic vitrectomy for posterior-segment complications of vitrectomy can certainly help to restore structure and function to the eye, and there are proposed theoretical benefits of removing the vitreous gel from the eye. Special considerations must be taken when performing vitrectomy on a uveitic eye. RP

REFERENCES

1. Nussenblatt RB. The natural history of uveitis. Int Ophthalmol. 1990;14:303-308.

2. Kruh J, Foster CS. The philosophy of treatment of uveitis: past, present and future. Dev Ophthalmol. 2012;51:1-6.

3. Foster CS. Uveitis—current standard of care. US Ophthalmic Rev. 2007;1:16-17.

4. Becker M, Davis J. Vitrectomy in the treatment of uveitis. Am J Ophthalmol. 2005;140:1096-1105.

5. Chan CC, Rubenstein JL, Coupland SE, et al. Primary vitreoretinal lymphoma: a report from an Int Primary Central Nervous System Lymphoma Collaborative Group symposium. Oncologist. 2011;16:1589-1599.

6. Margolis R, Brasil OF, Lowder CY, et al. Vitrectomy for the diagnosis and management of uveitis of unknown cause. Ophthalmology. 2007;114:1893-1897.

7. Margolis R. Diagnostic vitrectomy for the diagnosis and management of posterior uveitis of unknown etiology. Curr Opin Ophthalmol. 2008;19:218-224.

8. Johnston RL, Tufail A, Lightman S, et al. Retinal and choroidal biopsies are helpful in unclear uveitis of suspected infectious or malignant origin. Ophthalmology. 2004;111:522-528.

9. Turno-Krecicka A, Misiuk-Hojlo M, Grzybowski A, Oficjalska-Mlynczak J, Jakubowska-Kopacz M, Jurowska-Liput J. Early vitrectomy and diagnostic testing in severe infectious posterior uveitis and endophthalmitis. Medical Sci Monit. 2010;16:CR296-300.

10. Davis J. What are the indications for vitrectomy in a patient with uveitis? In: Foster CS, ed. Curbside Consultation in Uveitis: 49 Clinical Questions. Thorofare, NJ: SLACK Incorporated; 2012:137-139.

11. Androudi S, Praidou A, Symeonidis C, et al. Safety and efficacy of small incision, sutureless pars plana vitrectomy for patients with posterior segment complications secondary to uveitis. Acta Ophthalmologica. 2012;90:e409-410

12. Kamei RR, e Arantes TE, Garcia CR, Muccioli C. Twenty-five gauge vitrectomy in uveitis. Arq Bras Oftalmol. 2012;75:107-110.

13. Diamond JG, Kaplan HJ. Uveitis: effect of vitrectomy combined with lensectomy. Ophthalmology. 1979;86:1320-1329.

14. Scott RA, Haynes RJ, Orr GM, Cooling RJ, Pavesio CE, Charteris DG. Vitreous surgery in the management of chronic endogenous posterior uveitis. Eye (Lond). 2003;17:221-227.