Summary Basis of Decision for Prolensa

Clinical Pharmacology

Bromfenac, the active ingredient of Prolensa, is a non-steroidal anti-inflammatory drug (NSAID). The mechanism of action is thought to be due to the drug's ability to block prostaglandin synthesis by inhibiting cyclooxygenase (COX) 1 and 2. Prostaglandins have been shown in many animal models to be mediators of certain kinds of intraocular inflammation.

Since Prolensa is for short-term topical ophthalmic administration, the assessment of bromfenac pharmacokinetics was minimal.

Ocular topical dosing (four times daily for 28 days in healthy human subjects) suggested that the bromfenac plasma concentration was very low (possibly ≤50 ng/mL). Therefore, systemic exposure following treatment with Prolensa is expected to be very low and of little clinical significance when Prolensa is used as directed. No major safety issues were detected after a 28-day topical ocular administration of 11 to 23 times the intended daily clinical dose for Prolensa. No other human pharmacology studies and no drug interaction studies were provided, however given the ophthalmic route of administration, this was found acceptable.

For further details, please refer to the Prolensa Product Monograph, approved by Health Canada and available through the Drug Product Database.

Clinical Efficacy

The clinical efficacy of Prolensa (bromfenac ophthalmic solution 0.07% w/v) was evaluated primarily in two similar multicentre, randomized, double-blinded, parallel-group and placebo (vehicle) -controlled, pivotal Phase III studies of identical design (Study 1 and Study 2). The efficacy of Prolensa was compared with placebo in the treatment of ocular inflammation and pain associated with cataract surgery.

Patients undergoing cataract surgery self-administered Prolensa or placebo once daily (QD), beginning 1 day prior to surgery, continuing on the morning of surgery and for 14 days after surgery. Complete clearance of anterior chamber (AC) inflammation was assessed on Days 1, 3, 8 and 15 post-surgery using slit lamp biomicroscopy. A pain score was self-reported.

The patient population included adults who required unilateral cataract surgery (phacoemulsification or extracapsular) with posterior chamber intraocular lens (PCIOL) implantation and for whom no other ophthalmic surgical procedures [for example (e.g), relaxing incisions, iridectomy, conjunctival excisions, et cetra (etc.)] were planned.

In the intent-to-treat (ITT) populations, a total of 440 patients were assessed. In Study 1, a total of 220 patients were randomized; 112 were treated with Prolensa, and 108 were treated with placebo. In Study 2, 110 were treated with Prolensa, and 110 were treated with placebo.

The primary efficacy endpoint for Study 1 and Study 2 was defined as the proportion of patients achieving a complete clearance of anterior ocular inflammation; that is, zero anterior chamber (AC) cell and no AC flare by Day 15 of the study. The main secondary efficacy endpoint (ocular pain) was determined by the proportion of patients who were pain free at Day 1.

Four analyses of efficacy were performed: an analysis of data based on last observation carried forward (LOCF), the primary efficacy analysis; an analysis of data based on observed cases (OC); an analysis of data based on baseline observation carried forward (BOCF); and an analysis based on multiple imputation. The latter three analyses were performed to assess the sensitivity of using different approaches for handling missing data.

Based on the primary efficacy endpoint of cleared inflammation using the ITT dataset with the last observation carried forward (LOCF) method, the responders' rate by Day 15 was statistically and clinically significantly higher with Prolensa (48% in Study 1, 49% in Study 2), as compared to placebo (17% in Study 1, and 32% in Study 2), for a difference between groups of 31% in Study 1 (p <0.0001), and 17% in Study 2 (p = 0.013). The treatment effect was apparent starting on Day 8.

The sensitivity analyses using the OC datasets captured only partially the benefit at Day 15 seen in the analysis using ITT/LOCF. The benefit of Prolensa was present in both studies at Day 8. The sensitivity analyses using BOCF datasets did not capture any of the differences between the treatment groups. This was likely due to the higher proportion of patients discontinuing the investigational product (IP) at Day 15 in the Prolensa group compared to the placebo group. Given that early IP discontinuation due to lack of efficacy was significantly more frequent with the placebo group (24%) as compared to the Prolensa group (3%), these sensitivity analyses were acceptable and were not considered to contradict the primary results using ITT with LOCF. In addition, two other sensitivity analyses for early study discontinuation, one assuming "treatment failures" and the other assuming "treatment success", supported the ITT/LOCF findings and their robustness.

One minor issue was noted in regard to the primary efficacy endpoint of cleared inflammation "at" each visit yielding results that were in line with the primary efficacy analysis "by" visit. Small differences were reported due to a few patients that may have had cleared inflammation before Day 15 but not at Day 15. This prompted Health Canada to use the results of the "at visit" primary efficacy analysis (ITT/LOCF) in the Prolensa Product Monograph. The "at visit" (or "at day") analyses are the most commonly used measures for this type of drug and condition. This did not change the overall results and efficacy conclusions.

For the proposed indication, it was imperative that Prolensa be efficacious on both ocular inflammation and pain. The measure of ocular pain was the main secondary efficacy endpoint in the efficacy analysis. The proportion of patients that were pain free at Day 1 [that is (i.e.), pain grade of None at Day 1] in the ITT/LOCF analyses was statistically and clinically significantly larger in patients treated with Prolensa (81% in Study 1, and 76% in Study 2), as compared to those receiving placebo (44% in Study 1, and 56% in Study 2), for a statistically and clinically significant difference of 38% in Study 1 (p <0.0001), and 21% in Study 2 (p = 0.002). The difference in response rates was maintained at Day 3, Day 8, and to a lesser extent at Day 15. The results from the two sensitivity analyses, and the multiple imputations analysis for the pain endpoint were in line with those from the ITT/LOCF analysis or did not raise any significant issue.

Rescue medications (mostly prednisolone or Prolensa) were statistically significantly more likely used by those treated with placebo (34% to 44%) as compared to Prolensa (5% to17%), which lent support to the efficacy findings. Many other efficacy analyses were performed but did not add significant information to the assessment of efficacy beyond the results of the two main efficacy endpoints (and associated sensitivity analyses).

For more information, refer to the Prolensa Product Monograph, approved by Health Canada and available through the Drug Product Database.

Clinical Safety

Two multicentre, randomized, double-blinded, parallel-group and placebo (vehicle) -controlled, pivotal Phase III studies of identical design (Study 1 and Study 2, described in the Clinical Efficacy section) evaluated the clinical safety of Prolensa. In the pooled studies, 212 patients received Prolensa alone and 204 received placebo once daily beginning one day prior to surgery, continuing on the day of surgery, and through the first 14 days post-surgery.

Drug compliance with Prolensa was acceptable (88% of patients had a compliance rate of ≥75%) and higher than that of patients with placebo (59% of patients had a compliance rate of ≥75%). In the pooled studies, the number of patients with adverse events (AEs), i.e., regardless of causal relationship to the drug, was higher with placebo (43%) as compared to Prolensa (29%). The rate of ocular AEs related to IP was also higher with placebo [21% versus (vs.) 7%].

Discontinuation of the IP due to adverse events (AEs) was reported in 14% of patients with placebo and 5% of patients with Prolensa. This difference may indirectly reflect the effects of Prolensa and outcomes of the surgery.

Most AEs were mild (18 to 19% in both groups), or moderate (17% with placebo and 8% with Prolensa). Severe AEs occurred in 7% of patients with placebo, and 3% of those with Prolensa. No particular severe AE occurred more frequently with Prolensa.

The ocular AEs related to IP most frequently reported were eye pain (2.8% with Prolensa vs. 7.8% with placebo), anterior chamber (AC) inflammation (2.4% vs. 5.4%), conjunctival hyperemia (0.9% vs. 3.9%), corneal edema (0.5% vs. 2.5%), increased lacrimation (0.5% vs. 2.5%), ocular hyperaemia (0 vs. 2.0%), foreign body sensation in eyes (0% vs. 2.5%), and photophobia (0.5% vs. 3.9%). For more details, see the Prolensa Product Monograph.

Systemic AEs were reported in approximately 6% of patients in each group, however none were considered to be related to the IP, except for headache (1 case with Prolensa and 2 with placebo). Some systemic AEs led to IP discontinuation but were not considered related to IP (e.g., hypoxia, myocardial infarction, high blood pressure, arrhythmia). Serious AEs were reported in 3 patients with Prolensa and 4 patients with placebo, all considered not related to the IP. No deaths were reported. Discontinuations due to AEs were significantly less frequent with Prolensa (5%) compared to placebo (14%), and were all considered not related to IP except for one case each of eyelid erythema/edema, uveitis, conjunctival hyperemia, and AC inflammation.

The mean intraocular pressure (IOP) in the Prolensa group was statistically significantly larger compared to the placebo group, but the differences were of modest magnitude (+ 1 to + 1.4 mmHg) and of little clinical relevance. Increases from baseline IOP ≥10 mmHg were recorded more frequently in patients treated with Prolensa [5.7%, number (n) = 12], as compared to those receiving placebo (2.5%, n = 5). This difference was only seen on Day 1, and not thereafter. Therefore, it was not a significant safety issue, and the sponsor was required to include this information in the Prolensa Product Monograph. Visual acuity (VA) change from baseline showed an average improvement of approximately 10 to 12 letters from baseline VA in both groups. Ocular comfort grades were in favour of Prolensa as compared to placebo. There was no evidence of a significant effect of age or gender on the AE profile of Prolensa as compared to placebo.

Based on the information provided, both of the two pivotal confirmatory clinical studies had an acceptable design and conduct, and Prolensa was shown to be efficacious with an acceptable safety profile in line with other topical ophthalmic NSAIDs in patients undergoing cataract surgery. Also, 20 other clinical studies used various ophthalmic solution formulations and concentrations, and the safety profile for the pooled bromfenac 0.09% QD studies in the post-surgery patients was in line with that seen in the pivotal Phase III studies for Prolensa.

Review of post-marketing data showed no findings that were unexpected for a topical ocular NSAID drug. Very rare cases of corneal effects (some of which serious and significant) were reported, however, corneal toxicity is a known and relatively uncommon potential effect of NSAIDs, and appropriate warnings are included in the Prolensa Product Monograph. The foreign post-marketing experience did not raise any specific significant issues at this time, and was therefore considered acceptable.

For more information, refer to the Prolensa Product Monograph, approved by Health Canada and available through the Drug Product Database.