Effect of Ruboxistaurin in Patients With Diabetic Macular Edema: Title and subTitle BreakThirty-Month Results of the Randomized PKC-DMES Clinical Trial FREE

Diabetic macular edema (DME), the leading cause of moderate visual loss in persons with diabetes,¹ can occur during either the nonproliferative or proliferative stages of diabetic retinopathy. Its prevalence increases with more severe retinopathy.² Approximately 10% to 15% of people with diabetes have DME, and the incidence increases with duration of diabetes.^{3 - 4} Diabetic macular edema is often treated with focal photocoagulation when it threatens vision. Focal photocoagulation of clinically significant macular edema reduces the occurrence of moderate visual loss by approximately 50%.⁵

Hyperglycemia is an important factor in the development and progression of diabetic retinopathy and DME.^{6 - 7} Hyperglycemia activates protein kinase C (PKC) by inducing de novo synthesis of diacylglycerol, a physiologic activator of PKC.⁸ Protein kinase C is a family of approximately 13 enzymes.⁹ Substantial preclinical and clinical data suggest that the β isoform may play an important role in the development of diabetic microvascular complications in the eyes, nerves, and kidneys.¹⁰ Increased PKC β-isoform activity induces retinal vascular permeability and neovascularization in animal models.^{11 - 12} Conversely, inhibition or genetic knockout of PKC β-isoform activity reduces diabetes-induced retinal permeability and ischemia-induced retinal neovascularization.^{13 - 14} Increased PKC β-isoform activity in patients with diabetes mediates early diabetes-induced alterations in retinal blood flow, an effect ameliorated by oral administration of a PKC β inhibitor.¹⁵

Ruboxistaurin (RBX) (compound identifier, LY333531) mesylate is a PKC β-isozyme–selective inhibitor with adequate bioavailability to permit oral administration once daily. In several animal models, RBX ameliorates hyperglycemia-induced diabetic microvascular complications, including diabetic retinopathy, DME, diabetic peripheral neuropathy, and diabetic nephropathy.^{11 ,13 ,16 - 19} Phase 1 and 2 clinical trials have demonstrated that RBX was well tolerated in persons with diabetes.^{15 ,20} The Protein Kinase C β Inhibitor Diabetic Macular Edema Study (PKC-DMES) was a multicenter, randomized, double-masked, parallel, placebo-controlled clinical trial that evaluated the effect of 3 doses of orally administered RBX on progression of DME and the need for laser photocoagulation.

The trial enrolled 686 patients with type 1 (18%) or type 2 (82%) diabetes mellitus who were aged 22 to 87 years and had a hemoglobin A_1c (HbA_1c) level of 5.1% to 13.1%. Participants were excluded if they had (1) a history of significant heart disease (including unstable angina, acute coronary syndrome, myocardial infarction, or a history of a coronary revascularization procedure) within the 6 months prior to the first visit; (2) significant hepatic disease (having aspartate transaminase, alkaline phosphatase, or total bilirubin levels greater than 2-fold the upper limit of normal), renal disease (having a serum creatinine level >2.5 mg/dL [221 μmol/L], having a history of renal transplantation, or undergoing dialysis at screening), or anemia (having a hemoglobin level <10 g/dL); (3) a systolic blood pressure of 180 mm Hg or higher, or a diastolic blood pressure of 105 mm Hg or higher; or (4) major surgery within the past 3 months.

The PKC-DMES was designed to test the primary hypothesis that RBX would delay either the progression of DME and/or the application of focal/grid laser photocoagulation in eyes with DME farther than 300 μm from the center of the macula, mild to moderately severe nonproliferative diabetic retinopathy, a visual acuity of 20/32 or better, and no prior laser photocoagulation for diabetic retinopathy or DME. Participants met all of the following ocular entry criteria in at least 1 eye: (1) retinal thickening within 2 optic disc diameters of the center of the macula (for the purposes of this article, an optic disc diameter will be defined as 1500 μm), with an area bigger than one sixth of an optic disc area but with no thickening or adjacent hard exudates within 300 μm of the center of the macula; (2) an Early Treatment Diabetic Retinopathy Study (ETDRS) retinopathy severity level between 20 and 47A; (3) a best-corrected visual acuity of 75 or more letters using ETDRS visual acuity protocol (a Snellen equivalent of approximately 20/32 or better); (4) no history of scatter (panretinal) or focal/grid photocoagulation for diabetic retinopathy; and (5) no evidence of glaucoma. Each patient had only 1 eligible eye designated as his or her study eye.

The PKC-DMES was a multicenter, double-masked, parallel, placebo-controlled study in which patients were randomized to 1 of 4 treatment groups using a block size of 8 patients. A sample size of 163 patients in each treatment arm was selected to permit comparison of each of the 3 treatment groups with the placebo group, with α = 0.0167 and power 80% to detect a 40% reduction in an anticipated event rate at 12 months of 0.33 (ie, 0.33-0.198 per 12 months), assuming 15% loss to follow-up in the first 12 months. The study was extended to 36 months because of a lower than expected event rate.

Patients received either placebo (n = 176); a dose of 4 mg of oral RBX per day (Eli Lilly and Company, Indianapolis, Ind), administered as a mesylate salt (n = 168); 16 mg of RBX per day (n = 174); or 32 mg of oral RBX per day (n = 168). From the 1460 patients who were screened, 686 patients were randomized (Figure 1). Of the 774 patients not randomized into the study, 687 (89%) were ineligible because they did not meet the ophthalmologic entry criteria. Eligibility was based on 2 screening visits that occurred within 6 weeks of the randomization visit. Following randomization, visits occurred at 1 and 3 months, every 3 months up to 30 months, and then every 6 months thereafter. All patients were observed until the last randomized patient completed 30 months of follow-up.

Decisions regarding application of photocoagulation resided with individual study investigators and patients, but study policy discouraged initiation of focal/grid photocoagulation in study eyes prior to the development of retinal thickening within 100 μm of the center of the macula (sight-threatening DME, the primary photographic end point). Panretinal photocoagulation for diabetic retinopathy was initiated at the investigator's discretion, but it was expected that it would not be applied before the development of level 65 proliferative diabetic retinopathy.

This trial was conducted in accordance with the Declaration of Helsinki, the guidelines on good clinical practice, and the regulations of the appropriate review boards at each center.²¹ Written informed consent was obtained from all participants.

All serious and nonserious adverse events were analyzed regardless of the investigators' assessments of causality. Adverse events that resulted in death, hospitalization, cancer, permanent disability, or threat to the life of the patient were classified as serious. The Medical Dictionary for Regulatory Activities was used to categorize reported adverse events. Laboratory evaluations were performed at each visit. Physical examinations and electrocardiograms were performed at screening and every 6 months thereafter. Electrocardiograms were also performed at 1 month postrandomization. For optimum assessment of safety and to account for less common adverse events, the safety data from the PKC-DMES trial were pooled with data from the Protein Kinase C β Inhibitor Diabetic Retinopathy Study (PKC-DRS), a trial of similar design and patient population. These pooled safety data represented 937 patients who were observed for periods ranging from 30 to 52 months and have been previously reported.²⁰ The studies were conducted in parallel and the vast majority of sites were identical. Except for the ophthalmologic entry criteria, the inclusion and exclusion criteria were also similar for each study. At baseline, no significant differences in the patient characteristics (eg, age, duration, type of diabetes) were evident between the 2 studies.

An ophthalmologic examination was performed at screening and at every visit during the treatment period. These examinations included a best-corrected visual acuity assessment using the ETDRS protocol,²² intraocular pressure determination, and a clinical lens grading (a simplified version of the Age-Related Eye Disease Study photographic system) every 3 months.²³ Diabetic retinopathy and DME were assessed by masked grading of 3-field or 7-field stereoscopic ETDRS fundus photographs.^{24 - 25} All photograph grading occurred at the University of Wisconsin Fundus Photograph Reading Center. Photographs were obtained at screening and at 3 and 6 months, and every 3 months thereafter (alternating between 3-field and 7-field assessments); photographs were graded using the ETDRS protocol, which was modified to include estimates of the area of retinal thickening in each subfield of the ETDRS grid and proximity of retinal thickening to the center of the macula. Gradings were carried out independently of each other for eyes and individual study visits.

The primary study outcome was a composite end point consisting of (1) progression to sight-threatening DME (defined as development of retinal thickening [or adjacent hard exudate] within 100 μm of the center of the macula, or development of retinal thickening within 300 μm of the center if the distance of retinal thickening [or adjacent hard exudate] was 1300 μm or farther from the center at baseline); or (2) application of focal/grid photocoagulation for treatment of DME in the study eye. The application of focal/grid photocoagulation prior to photographically documented progression to sight-threatening DME varied across clinical sites, and a secondary analysis was carried out evaluating progression to sight-threatening DME alone. In this analysis photographically documented progression to sight-threatening DME was considered an outcome whether it occurred before or after focal/grid photocoagulation, or whether it occurred in eyes that did not receive focal/grid photocoagulation during the study. Owing to the well-established effects of glycemic control on diabetic retinopathy and DME progression, as well as PKC β activity, further analyses were performed on patients subgrouped into a baseline HbA_1c tertile.

Baseline patient characteristics were compared across treatment groups by categorical tests or analysis of variance. All analyses were done using the intent-to-treat principle. The unadjusted effects of treatment on the occurrence of events (primary study outcome and sight-threatening DME progression alone) were analyzed by Kaplan-Meier time-to-event curves. In a secondary analysis, treatment effect (32 mg of RBX per day vs placebo) was assessed in a Cox proportional hazards model adjusting for baseline covariates identified from potential confounders, including use of angiotensin-converting enzyme inhibitors and/or angiotensin-receptor blockers, age, alcohol use, antibiotic use, antihypertensive use, body mass index (calculated as weight in kilograms divided by height in meters squared), urine protein level, diabetic retinopathy level, duration of diabetes, elevated low-density lipoprotein cholesterol or triglyceride levels, HbA_1c level, insulin use, mean arterial blood pressure, nitrates use, race, sex, tobacco use, type of diabetes, and visual acuity score. Patients with missing covariates (<10%) were excluded from the analysis using a Cox proportional hazards model.

Baseline demographic characteristics by treatment group are summarized in the Table. There were no clinically significant differences observed at baseline among treatment groups for demographic characteristics, laboratory values, or most concomitant medications, except possibly for the use of angiotensin-converting enzyme inhibitors and/or angiotensin-receptor blockers and use of antihypertensive medications.

Baseline ophthalmic characteristics by treatment group are summarized in the Table. There were no clinically relevant differences in baseline ophthalmic characteristics among treatment groups. In addition, there were no statistically significant differences in demographic or ophthalmic characteristics for DME study eyes among randomized patients who did not complete the PKC-DMES study, except for mean age (placebo group, 58.39 years; 32-mg dose of RBX group, 57.07 years; P = .05) and duration of diabetes (placebo group, 15.5 years; 32-mg dose of RBX group, 15.9 years; P = .04).

There were no statistical differences among treatment groups in the time of occurrence of the primary end point or in the cumulative percentage of patients who reached this end point (log-rank test of difference in survival curves, P = .22; Kaplan-Meier probabilities at 36 months for placebo, and 4, 16, and 32 mg of RBX per day: 55, 51, 53, and 47%, respectively). Progression to the primary end point in the 32-mg dose of RBX group was not statistically different from the placebo group (log-rank test, 32 mg of RBX vs placebo, P = .14; Cox proportional hazards model hazards ratio = 0.73, P = .06) (Figure 2).

Of the 305 primary end point events, 55 (18%) were caused by the application of focal/grid photocoagulation prior to study-defined DME progression. Ten of the 48 study sites (21%) with 163 of the 686 study patients (24%) accounted for 33 of the 55 focal/grid photocoagulations (60%) that occurred prior to photographically documented progression to sight-threatening DME. There was a suggestion of imbalance among study sites in the distribution of focal/grid photocoagulation occurring before photographically documented progression to sight-threatening DME (χ², P = .05; by bootstrap simulation). In light of this apparent imbalance, the secondary analysis of progression to sight-threatening DME was carried out, without consideration of the application of a laser. In an analysis of photographically documented progression alone (whether it occurred before, after, or in the absence of focal/grid photocoagulation during the study), there was a delay in progression to sight-threatening DME in the 32-mg dose of RBX treatment group (placebo vs 32 mg of RBX, P = .054; Kaplan-Meier probabilities at 36 months, 54% and 44%, respectively) (Figure 3). For both the primary end point (Figure 2) and the secondary end point of photographically documented progression alone (Figure 3), visual inspection of Kaplan-Meier curves showed that after 12 or more months in the study, patients receiving 32 mg of RBX per day appeared to have slower progression to sight-threatening DME than those taking placebo. Intermediate doses of RBX exhibited intermediate DME progression rates. In Cox proportional hazards analysis adjusting for important covariates (Figure 4), treatment with 32 mg of RBX reduced the risk of DME progression compared with placebo (hazards ratio = 0.66; 95% confidence interval [CI], 0.47-0.93; P = .016).

There were no differences observed among treatment groups for progression of diabetic retinopathy (3 steps on the ETDRS person scale or 2 steps on the ETDRS eye scale), development of level 65 proliferative diabetic retinopathy, development of proliferative diabetic retinopathy with high-risk characteristics (ETDRS levels 71 and 75), application of panretinal photocoagulation, doubling or halving of the area of DME, or change in visual function assessed using the Visual Function Questionnaire 25.

After 30 months, 22.7% of patients taking placebo and 18.0% of patients taking 32 mg of RBX per day had received at least 1 focal/grid photocoagulation treatment (P = .29). Kaplan-Meier curves did not show a difference between treatment groups for time to application of laser treatment. After 30 months, only 2.8% of patients taking placebo and 4.8% of patients taking 32 mg of RBX per day had experienced sustained moderate visual loss (defined as a decrease from baseline ETDRS visual acuity of 15 or more letters for the last 6 months of study participation; P = .42).

The Cox proportional hazards model analysis for DME progression (Figure 4) suggested that there was an increased risk of DME associated with an elevated HbA_1c level. Consequently, the effect of 32 mg of RBX per day on DME progression was assessed in the 3 groups of baseline HbA_1c (≤7.8%, >7.8%-10%, >10%). When using the middle group as the reference group, the interaction between the HbA_1c groups and treatment with 32 mg of RBX per day was evaluated, and significant interaction was observed for both the lower and upper groups (P = .02 and P = .09, respectively). Essentially, no effect was observed for the lowest HbA_1c group (odds ratio = 1.16) or the highest group (odds ratio = 0.84). Ruboxistaurin had the greatest effect in the middle group (odds ratio = 0.4).

Baseline covariates significantly associated with increased risk of progression to sight-threatening DME in an adjusted Cox proportional hazards model analysis were a body mass index greater than 30 (hazards ratio = 0.71; 95% CI, 0.50-0.99; P = .05), an HbA_1c level greater than 10% (hazards ratio = 1.76; 95% CI, 1.22-2.55; P = .003), and DME (severe) (hazards ratio = 1.41; 95% CI, 1.00-2.00; P = .05) (Figure 4).

The PKC-DMES was designed to test the hypothesis that RBX, a β-isoform–selective PKC inhibitor, would delay either the progression of DME and/or the application of focal/grid laser photocoagulation in eyes with DME farther than 300 μm from the center of the macula, mild to moderately severe nonproliferative diabetic retinopathy, a visual acuity of 20/32 or better, and no prior laser photocoagulation for diabetic retinopathy or DME. In unadjusted analysis, no statistically significant effect of RBX on this composite end point was observed among the 3 treatment doses after a minimum of 30 months of follow-up (Figure 2).

When considering progression of DME alone, without regard to laser photocoagulation, a possible positive effect of 32 mg of RBX vs placebo was observed in an unadjusted analysis, and there was a significant treatment effect in the Cox proportional hazards model adjusted for covariates (Figure 4). The effect of RBX on DME progression appeared to be dose responsive, with the highest dose uniformly exhibiting the greatest efficacy.

The mechanisms whereby PKC β activation can induce DME are discussed in detail in the PKC-DRS trial report.²⁰ In that study, RBX treatment did not have an effect on diabetic retinopathy progression but did appear to reduce the occurrence of moderate visual loss. The PKC-DRS population was too small (60-65 patients per treatment arm) and included too wide a range of DME at baseline to determine whether RBX affected DME progression. Diabetic macular edema in the PKC-DRS patients ranged from none to center involvement at baseline; visual loss occurred primarily in patients with DME at baseline. In PKC-DMES, no patients had DME at the center of the macula at baseline, and the occurrence of sustained moderate visual loss was more than 5-fold lower than that observed in PKC-DRS. With rates of sustained moderate visual loss in PKC-DMES below 5%, it would be nearly impossible to detect a treatment effect on this end point.

It has been previously reported that treatment of patients with DME for 3 months using a multitargeted kinase inhibitor, which also acts as a nonspecific PKC inhibitor, led to reductions in some measures of retinal thickening, as evaluated by optical coherence tomography.²⁶ Systemic applicability of this nonselective compound was limited by gastrointestinal side effects and dose-related problems with tolerability, glycemic control, and liver toxicity.

In contrast, treatment with RBX was well tolerated for the duration of this study. Safety data from this study have been previously published as part of a combined data set of 937 patients in 2 ocular studies of patients with diabetes.²⁰ Only first-degree atrioventricular block, asthma, and dysuria were statistically different among treatment groups, occurring more frequently in the 32-mg dose RBX group. These events have not been associated with RBX treatment in other 6- to 12-month trials of RBX. Expected associated diagnostic testing or reported events (eg, bronchospasm/wheezing) were not different between the RBX and placebo groups in the combined data set. In contrast to multitargeted kinase inhibitors, which were associated with more frequent adverse events,²⁶ an isoform-specific PKC inhibitor such as RBX might be expected to have a more favorable safety profile by virtue of its greater specificity. However, safety analyses from all ongoing studies using RBX are continuing to evaluate any effects that may become evident with greater patient exposure.

The PKC-DMES is the first clinical trial evaluating the effect of a PKC isoform-selective inhibitor on DME in patients with diabetes mellitus. In patients with DME farther than 300 μm from the center of the macula at baseline, the reduction in the composite end point of DME progression or application of focal/grid photocoagulation from 32 mg of RBX per day as compared with placebo was not statistically significant. However, when considering progression of DME alone, there was a significant reduction in those treated with 32 mg of RBX per day. Additional trials are under way to provide further information on the effects of RBX on DME.

Correspondence: Lloyd Paul Aiello, MD, PhD, Beetham Eye Institute, Joslin Diabetes Center, One Joslin Place, Boston, MA 02215 (lloydpaul.aiello@joslin.harvard.edu).

Submitted for Publication: October 31, 2005; final revision received May 22, 2006; accepted May 23, 2006.

Author Contributions: Dr L. P. Aiello has had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

PKC-DMES Study Group Members:

Financial Disclosure: Drs Hu, Sheetz, and Vignati were all employed by Lilly Research Laboratories during the time that this manuscript was written. Drs L. P. Aiello, Davis, and Milton are on the scientific advisory board for Eli Lilly and Co and have received consulting fees. Drs Lund-Andersen and Ross have also received consulting fees from Eli Lilly and Co.

Funding/Support: This study was funded by Eli Lilly and Co.

Acknowledgment: We thank all the patients who participated in the study. We also thank Rocky Johnson, MS, Keri A. Kles, PhD, Xin Zhi, PhD, and Tim Mason, MS, for their valuable assistance with the preparation of this manuscript.