Vitrectomy for Diabetic Macular Edema

It reduces DME, is as safe as series of injections, and costs much less.

MAURICE B. LANDERS, III, MD • MICHAEL W. STEWART, MD • DAVID J. BROWNING, MD, PhD • TOSHINORI MURATA, MD, PhD • MATT OHJI, MD • HUGO QUIROZ-MERCADO, MD • JAY CHHABLANI, MD • JERZY NAWROCKI, MD, PhD • VERONICA KON GRAVERSON, MD • MARIE S. VOGEL, BS, MS

The role of pars plana vitrectomy in the management of early diabetic macular edema has not been well established. Dozens of retrospective and prospective studies have generated conflicting data, and even the most recent large studies addressing this issue^1,2 have produced equivocal, and possibly even misleading, results.

This confusion has contributed to a significant management problem, inasmuch as DME is one of, if not the leading cause of significant loss of vision in working age adults in the Western world.³

Moreover, the incidence of diabetes is increasing rapidly in the developing world, particularly in China and India. Approximately 115 million Chinese adults currently have diabetes, with nearly another 500 million having prediabetes.⁴ The prevalence of diabetes is growing at an alarming rate in India, which is now home to more than 65 million people with the disease.⁵

THE PHARMACOLOGIC LANDSCAPE

The current “gold standard” or “standard of care” for DME in the United States has become intravitreal pharmacotherapy, primarily with one of the three commonly used drugs that block the actions of VEGF: bevacizumab (Avastin, Genentech, South San Francisco, CA), ranibizumab (Lucentis, Genentech), or aflibercept (Eylea, Regeneron, Tarrytown, NY).

Treatment of DME with intravitreal anti-VEGF drugs, although dramatically better than the natural course of DME and even much better that the previous “gold standard” of laser photocoagulation in and around the macula, still leaves much to be desired.

Intravitreal injections of anti-VEGF drugs, and even intravitreal injections of corticosteroids, can significantly reduce DME and improve visual acuity. However, these treatments are not without their shortcomings.

Intravitreal corticosteroids — triamcinolone acetonide, dexamethasone, and fluocinolone acetonide — are often quite effective at reducing DME by modulating the chronic inflammation that contributes to the formation of edema.

However, steroids are generally regarded as second-line therapy because of their associated adverse effects. Repeated steroid injections frequently induce glaucoma, which on occasion is sufficiently severe to require surgical intervention to control the intraocular pressure. Further, intravitreal steroids are a common cause of posterior subcapsular cataracts. Triamcinolone and dexamethasone are not long lasting, and fluocinolone is extremely expensive.

Complexities of VEGF Blockade

Intravitreal injections of anti-VEGF drugs are currently the most frequent treatment for DME in the United States. They are largely effective, although not universally so. They rarely cause major complications, although when multiple injections are required, as is usually the case with DME, the incidence of sight-threatening complications becomes significant, approaching the complication rates of more invasive forms of treatment, such as vitrectomy.

The use of intravitreal anti-VEGF drugs becomes even more problematic when one considers problems of regulatory compliance, compounding, and counterfeiting. The use of the least expensive drug, bevacizumab, is challenged by many physicians and government agencies because its intravitreal use is “off label” due to its lack of FDA approval for use within the eye.

Because it is sold by the manufacturer in a 4-mL single-use vial, it must be fractionated by a compounding pharmacy, which increases the risks of microbial contamination and endophthalmitis. Bevacizumab ranks seventh in worldwide drug sales, and because of its clear liquid formulation, it has unfortunately become the world’s most commonly counterfeited drug.

Compared to bevacizumab, the two anti-VEGF drugs approved by the FDA for the treatment of DME, ranibizumab and aflibercept, were recently deemed by the DRCRnet study group to be “not cost effective,” as measured with the Incremental Cost Effectiveness Ratio (ICER) per quality-adjusted life year (QALY).⁶

The DRCRnet research group found that these two drugs would have to be reduced in price from 69% to 80% to be labeled “cost effective in terms of $100,000 per QALY” (Table 1).

This leaves only bevacizumab, which is actually only available for use in the eye because of the serendipitous discovery that the oncologic formulation worked essentially just as well to treat neovascular age-related macular degeneration as did its purposely modified sister drug, ranibizumab.⁷

Unfortunately, many national regulatory agencies prevent the intraocular use of bevacizumab, leaving many low-income patients without an affordable pharmacologic treatment. Therefore, the role of PPV in the management of early DME must be re-evaluated in a prospective, randomized, multicenter, controlled trial, with the goal of developing high-level evidence regarding whether or not this form of treatment is effective.

METHOD OF ACTION OF PPV

All of the previously mentioned treatments for DME — laser, steroids, and anti-VEGF drugs — improve vision in the same manner: they decrease edema within the central retina. There have been no reports of significant improvements in the vision of patients with DME in the absence of resolving edema, as demonstrated by a decrease in the central subfield thickness (CST) of the retina on optical coherence tomography.

Some treatments decrease the CST without concurrent improvement in vision, possibly because the retina has already been irreversibly damaged, but within a study cohort, vision never improves without an associated decrease in CST.

Pars plana vitrectomy usually decreases the macular edema (CST) in DME patients. The DRCRnet’s Vitrectomy Study found that, in both the primary cohort of 87 patients with vitreomacular traction¹ and the entire cohort of 241 patients,² CST decreased by an average of 33%, from 412 µm to 278 µm.

In the May 2013 issue of Retinal Physician,⁸ we reported that the vast majority of more than 1,800 patients with DME from multiple studies experienced decreased macular thickness after vitrectomy and that this thinning was usually sustained throughout the duration of the studies (Figures 1 and 2; Table 2).

Figure 1. Vitrectomy reduces central subfield thickness in DME.

Figure 2. Vitrectomy decrease of CST is long lasting.

MISCONCEPTIONS AND MISDIRECTIONS

Even with considerable evidence that PPV reduces macular edema in patients with DME and with evidence that reduction in macular edema is the mechanism by which vision is improved in all successful treatments of DME, there remains a considerable reluctance by many surgeons to use vitrectomy in the management of early DME. Several factors may be responsible for this reluctance.

The History of Vitrectomy

Vitreous surgery has improved dramatically since its introduction more than 45 years ago, when it was used primarily to treat nonresolving vitreous hemorrhages and recurrent retinal detachments secondary to proliferative vitreoretinopathy.

Surgical techniques and instruments have improved to the point that vitrectomy is now routinely used to repair primary retina detachments and to remove epiretinal membranes and repair macular holes in eyes with 20/50 or better vision, and several significant series have described the successful removal of floaters.⁹

Perceived Risk

There is the perceived risk of an “invasive procedure” to treat an eye disease that could otherwise be managed “medically.” Surely these are relative terms. Both the “invasive” PPV and the “safe” series of intravitreal injections (required over three years or longer) to treat DME would seem to have similar rates of major complications, approximating a 1% incidence of a sight-threating RD in the case of vitrectomy and an approximately 1% cumulative three-year incidence of endophthalmitis in the case of intravitreal injections.¹⁰

Induction of Cataracts

There is good evidence that vitrectomy hastens the development of nuclear sclerotic cataracts, particularly in patients older than 50. However, this is a relatively small problem because large series of cataract surgeries have reported that excellent functional rehabilitation occurs very rapidly, often in a few days, in more than 98% of patients.

Kaiser Hospitals in Colorado have even reported a large series of successful cataract surgeries performed in minor surgery rooms of eye clinics, without the use of formal, dedicated operating rooms.¹¹

Doubts About Success

The mistaken concept persists that vitrectomy is not a successful means of reducing DME. This concept has been supported in part by the reports from the DRCRnet’s Vitrectomy for DME study because the patients in that study did not often obtain improved vision, although their 241 study patients experienced a decrease in mean CST from 412 µm to 278 µm. The DRCRnet reported that more than 90% of operated patients experienced decreased CST.

Concerns About More Harm Than Good

It is often stated but possibly mistaken that vitrectomy will disadvantage the eye, if and when additional intravitreal injection treatment is needed. Limited available evidence^12,13 suggests that postvitrectomy eyes respond to intravitreal anti-VEGF and sustained-release corticosteroid treatments in approximately the same manner as do eyes with an intact vitreous.

Doubts About Improved VA

There persists the idea that reduced CST, achieved by vitrectomy, as reported in the DRCRnet’s Vitrectomy Study, is not likely to improve VA. It would seem likely, however, that this outcome may simply be the result of selection bias in that study.

The only patients recruited into this study were DME patients who had VMT (87 of the 241 patients in the study) or patients “who in the absence of traction [had] edema judged to persist despite previous non-surgical treatment, and [were] considered by the investigator unlikely to respond to further macular laser photocoagulation.”

In short, only severely damaged eyes were entered into this study — eyes that were otherwise felt to be beyond repair.

Lack of Health Evaluation of Retinas

The inability at the time of the DRCRnet’s Vitrectomy Study to meaningfully evaluate the health of the retinas of the recruited DME patients occurred because spectral-domain OCT was not yet available, and the investigators had only time-domain OCT images with which to evaluate the retinas of their patients. They had no way to determine the integrity of the outer retinas of the patients whom they recruited into the study.

However, several studies have since shown a significant correlation between the integrity of the outer retinal layers (external limiting membrane and ellipsoid zone [EZ]) and the current level of VA in DME patients.^14-17

Two studies^18,19 have now shown a correlation between the previtrectomy ELM and EZ (IS/OS junction) and the potential for VA improvement once vitrectomy has improved or resolved the DME (Figure 3).

Figure 3. Previtrectomy and postvitrectomy DME eyes with ELM overwritten in light blue; EZ overwritten in yellow. Eyes with intact outer retinal layers (ORL) have better prognosis for improved vision post reduction in DME (top) than do eyes with damaged or absent ORL.

CONCLUSIONS

The incidence rates of diabetes and DME are increasing worldwide. In addition, current treatments for DME are suboptimal because they pose a challenge to compliance, have short duration (thereby requiring numerous, repeated injections), are time-consuming for patients, caregivers, and physicians, are extremely expensive even for regimens with bevacizumab (where they are available), and are not completely risk free.

Pars plana vitrectomy has been shown to reduce or resolve DME in a large percentage of patients. The sight-threatening complications of vitrectomy occur with approximately the same frequency as those with an extended series of intravitreal injections.

The cost of vitrectomy to the patient or third-party payer (insurance companies or government) is approximately the same as that of a single year of off-label injections of bevacizumab (Table 1), but it is generally only a small fraction of the cost of ranibizumab (approximately $15,000 for the initial year) or aflibercept (close to $21,000). Whereas the edema-resolving effects of vitrectomy are durable, intravitreal injection regimens usually must be continued for years.

To quote from the DRCRnet study itself²:

The role of vitrectomy compared with other approaches in the management of DME remains uncertain as the potential benefits and risks have not been clearly defined in the context of long-term, adequately sized randomized clinical trials. However, with respect to patient management at present, the findings [of this study] should be viewed with caution since it is unknown how these eyes would have fared in the absence of vitrectomy or with other treatment modalities.

Pars plana vitrectomy treatment of early DME must be studied carefully. Performed in the appropriate patient, it is reasonably likely that vitrectomy will prove to be as effective, safe, and efficient and considerably less expensive than the intravitreal injection therapies that are currently the standard of care in the United States for this widespread, growing problem. RP

REFERENCES