PEER REVIEWED

Steroids in Diabetic Macular Edema

In sustained-release formats, these drugs have shown impressive evidence.

AARON M. YEUNG FRCOphth, PhD • MARK GILLIES FRANZCO, PhD

Diabetic retinopathy is one of the most common causes of visual impairment in developed nations. It has been estimated that there are approximately 93 million people that are affected with DR: 17 million with proliferative DR, 21 million with diabetic macular edema, and 28 million with sight-threatening retinopathy.¹

The exact pathogenesis of DR has yet to be determined; however, numerous biochemical mechanisms have been proposed that are associated with hyperglycemia and have been implicated in the pathogenesis of the disease. These mechanisms include activation of protein kinase C (PKC) isoforms, oxidative stress, advanced glycation end-product formation, and increased activity of the polyol and hexosamine pathways.

Unfortunately, treatments that have been developed to influence these biochemical pathways, such as the specific PKC beta inhibitor ruboxistaurin (Eli Lilly, Indianapolis, IN), have so far not been translated into clinical practice.

THE ROLE OF INFLAMMATION

There has in recent years been an increasing realization that inflammation is a prominent feature of DR. It is associated with increased levels of inflammatory mediators including interleukin-1β, tumor necrosis factor (TNF)-α, intercellular adhesion molecules, and angiotensin II,² as well as activation of microglial cells.³

Aspirin, meloxicam (Mobic, Boehringer Ingelheim (Ingelheim-Rhine, Germany), and etanercept (a soluble TNF-α receptor Fc hybrid; Enbrel, Amgen, Thousand Oaks, CA) have all been shown to normalize vascular permeability, leukostasis, CD18 and ICAM expression, and nuclear factor κB activation in animal models.⁴

The well-known anti-inflammatory effect of corticosteroids is the likely mechanism for their clinical effects on DME discussed below. DME is associated with loss of endothelial tight junction proteins^5,6 and increased levels of VEGF production.

By altering these pathways, steroids help to reduce retinal edema and may also inhibit neovascularization.^7,8 Finally there is some evidence that steroids may have a neuroprotective effect on the retina.^9-11

CLINICAL EVIDENCE

Jonas and Sofker¹² were the first authors to report on the successful use of intravitreal triamcinolone acetonide (IVTA) injection for the treatment of DME, in a single case report of a 73-year-old patient who had persistent DME despite previous macular laser treatment.

Following treatment, the patient’s visual acuity improved from 20/200 to 20/80 over a five-month follow-up period. The authors reported an increase in intraocular pressure, which was managed with topical medication.

The first randomized clinical trial of IVTA for DME that persisted following laser treatment was the Triamcinolone for Diabetic Macular Oedema (TDMO) study,¹³ a two-year, prospective, double-masked, placebo-controlled, randomized trial. A total of 69 eyes of 43 patients were included, with 34 eyes randomized to receive IVTA (4.0 mg) and 35 receiving placebo.

The primary outcome was the improvement of ≥5 logMAR letters of VA, which 19 of 34 (56%) in the IVTA group achieved, compared to nine of 35 (26%) in the placebo group (P=.006). The secondary outcome was the mean improvement in VA, which was 5.7 letters more in the IVTA-treated eyes (95% confidence interval [CI], 1.4-9.9).

After 2 years, foveal thickness had decreased by 59 µm more in the IVTA group (95% CI: 15-104) compared to placebo (P=.009). The authors concluded that IVTA improved VA and reduced macular thickness in eyes with DME that was refractory to laser treatment and that this effect persisted for up to two years with repeated treatment. The authors also noted that patients with DME could still spontaneously improve over a two-year period.

DRCRnet Enters the Picture

Next, the Diabetic Retinopathy Clinical Research Network¹⁴ (DRCRnet) completed a large, multicenter, randomized clinical trial over a two-year period, comparing 1.0-mg and 4.0-mg doses of intravitreal triamcinolone acetonide (Kenalog, Allergan, Inc., Irvine, CA) with focal or grid photocoagulation.

The study included 840 eyes of 693 patients with DME involving the fovea and VA of 20/40 to 20/320. The eyes were randomized to focal/grid photocoagulation (n=330), 1 mg IVTA (n=256), or 4 mg IVTA (n=254). In addition, retreatment was provided for persistent or new edema at four-month intervals.

The primary outcome measurement, change in mean VA at four months, was better in the 4-mg IVTA group than either the laser group (P<.001) or the 1-mg IVTA group (P=.001). At the end of the first year, there were no significant differences among the groups.

However, at two years, mean VA was better in the laser group than in the other two groups. Retinal thickness measured on OCT was a secondary outcome, and the results were consistent with their associated VA outcomes.

Three-year outcomes were reported as a follow-up to the first trial.¹⁵ A total of 306 eyes were available, and the authors concluded that the three-year follow up results were consistent with the previously published two-year results. They suggested that IVTA did not appear to be beneficial over the long term compared to focal or grid photocoagulation for patients with DME.

The results of the DRCRnet study^14,15 and the TDMO study¹³ were contradictory. One reason for this contradiction may be the different recruitment criteria for the two studies. Patients recruited into the TDMO group had to have previously received laser and were deemed to have “refractory” DME. These eyes were excluded from the DRCRnet study because the investigator had to believe that further laser would be beneficial so the DRCRnet study included patients with milder disease.

A small subgroup analysis of the patients with the greatest visual impairment (ie, VA of 6/60 to 6/96) at baseline in the DRCRnet study¹⁴ demonstrated that the group receiving 4 mg of IVTA had the best outcomes. Seventy-seven percent of the 4-mg IVTA vs 46% of the 1-mg IVTA group and 42% of the laser group experienced improvement of 10 or more letters. Worsening of 10 letters or more was found in 0%, 15%, and 17% of these groups, respectively, after two years. These data suggest that 4 mg of IVTA was beneficial for DME with severe visual impairment.

Fluocinolone Acetonide

The FAME studies were two identically designed phase 3 clinical trials — FAME A and FAME B — which compared two doses of a nonbioerodable intravitreal implant (Iluvien, Alimera Sciences, Alpharetta, GA) that released submicrogram doses of fluocinolone acetonide (FAc), compared with a sham injection over a three-year period.^16,17

Study participants who had persistent DME despite previous macular laser treatment were randomized to 0.5 µg/d (high dose) (n=393), 0.2 µg/d (low dose) (n=375), or sham injection (n=185). All of the participants were eligible for rescue laser after six weeks, and based on retreatment criteria, additional study drug or sham injections could be given after one year. The primary outcome measurement was the percentage of patients with improvement of ≥15 letters from baseline.

At three years, using the last observation carried forward, the primary outcome was achieved in 27.8% (0.5 µg/d), 28.7% (0.2 µg/d), and 18.9% (sham) of patients (P=.018). When only considering patients still in the trial at three years, the primary outcome was achieved in 31.9% (0.5 µg/d), 33.0% (0.2 µg/d), and 21.4% (sham) (P=.03).

Finally, a preplanned subgroup analysis demonstrated a doubled benefit compared with sham injections in patients who reported a duration of DME ≥3 years at baseline, with 28.8% (0.5 µg/d; P=.002), 34.0% (0.2 µg/d; P<.001), and 13.4% (sham) of this group achieving the primary outcome.

The authors concluded that FAc inserts could provide substantial visual benefit for up to three years in the treatment of patients with DME.

Chronic vs Nonchronic Disease

Chronic and “nonchronic” DME likely have distinct pathological features. As DME initially develops, it is characterized by acute inflammation and vascular dysfunction. In contrast, chronic DME may be associated with chronic inflammation and neuronal damage.

The long-term benefit of sustained delivery of FAc in patients with chronic DME was reported by the FAME study group.¹⁸ Nonchronic DME responded to intermittent laser and off-protocol intravitreal therapy (IVTA or anti-VEGF); however, low-dose corticosteroid did not add any significant benefit. Chronic DME did not respond to laser or off-protocol intravitreal therapy (IVTA or anti-VEGF) but did benefit from continuous low-dose corticosteroid therapy.

This change from nonchronic DME to chronic DME was also reported in two phase 3 clinical trials of monthly ranibizumab (Lucentis, Genentech, South San Francisco, CA). It was observed in both trials that sham control subjects who were crossed over to monthly injections of ranibizumab after two years did not show as great a visual improvement as those participants who received ranibizumab for the first two years,¹⁹ suggesting that there may be a shift in the pathophysiology of DME once patients have chronic DME. One potential bias of the FAME subgroup analysis, however, was that there was no strict guidance for determination of the duration of DME.¹⁸

Dexamethasone

The MEAD study²⁰ examined the safety and efficacy of a dexamethasone intravitreal implant (Ozurdex, Allergan) in DME. There were two randomized, multicenter, masked, phase 3 clinical trials. A total of 1,048 patients with DME were recruited, with BCVA of 20/50 to 20/200 and central retinal thickness of ≥300 µm.

The patients were randomized 1:1:1 to dexamethasone implant 0.7 mg, dexamethasone implant 0.35 mg, or sham, and they were followed for three years. Patients who met retreatment criteria could be retreated no more often than every six months. The main outcome measurement was improvement of ≥15 logMAR letters from baseline to study end.

Significantly more participants who received steroid treatment achieved a 15-letter gain: 22.2% in the 0.7-mg group and 18.4% of the 0.35-mg group, compared with 12.0% of the sham group (P≤.018). A secondary outcome, the mean reduction in central retinal thickness from baseline, was greater in the groups receiving dexamethasone implants with 0.7 mg (-111.6 µm) and 0.35 mg (-107.9 µm) vs sham (-41.9 µm; P<.001).

The BEVORDEX study²¹ reported the 12-month results of a randomized head-to-head clinical trial that compared bevacizumab (Avastin, Genentech) with dexamethasone implant.

Eighty-eight eyes of 61 patients with center-involving DME were enrolled, with 42 eyes receiving bevacizumab every four weeks and 46 eyes receiving a dexamethasone implant as frequently as every 16 weeks, both PRN.

The primary outcome was the proportion of eyes that improved in VA by 10 logMAR letters, which was achieved in 40% in the bevacizumab group vs 41% of the dexamethasone-treated eyes (P=.83).

None of the 42 eyes treated with bevacizumab lost 10 letters or more, but 11% of the eyes given the dexamethasone implant did, mostly due to cataract. Bevacizumab-treated eyes received a mean of 8.6 injections, compared with 2.7 injections for dexamethasone-treated eyes.

ADVERSE EFFECTS OF INTRAVITREAL CORTICOSTEROIDS

Cataract and elevated IOP are the most well-documented side effects associated with intravitreal corticosteroid use.^22,23

In the TDMO study, 15 of 28 (54%) subjects required cataract surgery, compared to 0 of 21 in untreated eyes (P=.0001). Elevated IOP of more than 5 mm Hg affected 23 of 34 (68%) of IVTA-treated patients compared to three of 30 (10%) in the placebo group.

DRCRnet reported that the cumulative probability of cataract surgery was 83% of the 4-mg IVTA group, compared with 46% of the 1-mg IVTA group and 41% of the laser group. IOP increased by more than 10 mm Hg in 33% of the 4.0-mg IVTA group, compared to 18% of the 1.0-mg IVTA group and 4% of the laser group.

Practically all phakic patients in the FAME study developed cataracts; however, the authors concluded that the visual benefits postcataract surgery were similar to those in pseudophakic patients. In addition, the incidence of incisional glaucoma surgery at three years was 8.1% in the 0.5 µg/day group and 4.8% in the 0.2 µg/day group.

The MEAD study also reported the greatest incidence of cataract in the higher 0.7-mg group (67.9%), compared with the 0.35-mg (64.1%) and sham groups (20.4%). The percentages of clinically significantly increased IOP were 36%, 34%, and 18%, respectively, and the total number of patients who required trabeculectomy were two and one in the 0.7-mg and 0.35-mg dexamethasone implant groups, respectively.

Finally, BEVORDEX reported after 12 months that 2.4% of bevacizumab-treated and 6.5% dexamethasone-treated eyes required cataract surgery; however steroid-induced cataracts usually take at least one year to develop. No eyes receiving bevacizumab, compared with 19.6% of the dexamethasone group, developed IOP elevation of ≥10 mm Hg.

Different intravitreal corticosteroids have different pharmacologic and pharmacokinetics that may account for their differing safety profiles. Triamcinolone, fluocinolone, and dexamethasone have been demonstrated to activate different patterns of gene expression in human trabecular meshwork cell lines.²⁴ Dexamethasone is not as lipophilic as triamcinolone or fluocinolone and therefore may not accumulate to the same extent in the trabecular meshwork and lens, which may be the reason why there is a somewhat reduced risk of IOP increase and cataract progression with dexamethasone.²⁵

There is no evidence that injection-related complications, such as endophthalmitis, retinal detachment, vitreous hemorrhage, and lens damage, are more common after injections of steroids compared with other drugs.^14,26,27

CONCLUSION

There is increasing evidence that demonstrates the efficacy of intravitreal steroids in the treatment of DME. While they appear to have a long and effective therapeutic effect, they are associated with increased risk of cataract and increased IOP.

Classifying DME into early and chronic types and discovering their underlying pathogeneses will continue to throw light on how best to approach our patients. In combination with anti-VEGF agents and selective laser photocoagulation, steroids appear to be a useful adjunct in the treatment of DME. RP

REFERENCES