Staff Writer

PSivida Corp. got a 42 percent share bump after it and partner Alimera Sciences Inc. reported statistically significant results from a pair of Phase III studies showing that eye drug Iluvien improved vision at high and low doses when data from both studies are combined.

Depending on how the data are analyzed, some 28-29 percent of treated study patients had a statistically significant improvement in vision, whether they were taking high- or low-dose Iluvien. Improved vision was based on patients' ability to identify letters on an eye chart.

But when the study results are analyzed separately, only one of the trials (Study B) hit the mark while the other (Study A) just missed it, acknowledged Paul Ashton, pSivida president and CEO.

Alimera plans to file for U.S regulatory approval of Iluvien as a potential treatment for diabetic macular edema (DME) in the second quarter of 2010.

Atlanta-based Alimera performed various analyses, including an analysis of the full dataset, which included all 956 patients randomized into Trial A and B, the FAME study. That group, referred to as the "intent-to-treat" population, did not exclude patients with a protocol violation, Ashton explained.

Using the full dataset to analyze the study was not pre-specified in the study protocol. However, pSivida said it believes that the FDA will consider the full analysis set, as described in guidance by the International Conference on Harmonization, to be the most relevant population for determining safety and efficacy in Trials A and B.

More detailed analyses will be presented in February 2010 at the Angiogenesis, Exudation and Degeneration meeting in Miami.

"For purposes of approval, the FDA wants to see two separate trials," Ashton said, noting that the 956 patients who participated in the study were divided into two treated groups and a control group that received a sham injection. "It comes down to how you analyze the data," he said, adding that he was pleased with the study results.

Ashton said that the FDA would consider all the data and likely would use multiple analyses in its review, and that could work in Iluvien's favor.

If Iluvien is approved, Watertown, Mass.-based pSivida would receive a $25 million milestone payment from Alimera as well as share in the sales profits.

In addition, pSivida would receive payment of a $15 million conditional note. If the note is not paid by April 2010, the annual interest rate increases to 20 percent (to be paid quarterly) and Alimera is to begin monthly principal payments of $500,000, according to pSivida.

Currently, the only FDA-approved treatment for DME is laser surgery. Bausch & Lomb's Retisert (fluocinolone acetonide) implant for uveitis has been prescribed off-label for DME, but it has a side effect of increased pressure in the eye.

In the Iluvien study, intraocular eye pressure (increases to 30 millimeters of mercury or greater at any time point) was a key adverse event studied in the trial. Such increases in eye pressure were seen in 21.6 percent of the high-dose Iluvien-treated patients and 16.3 percent of the low-dose patients.

Over the 24-month study period, some patients had to undergo surgery to relieve eye pressure. Some 5.1 percent of patients receiving the high dose and 2.1 percent of those receiving the low dose had received a trabeculectomy (filter surgery) to reduce their eye pressure.

Ashton contrasted those results with the 20 percent of patients on Retisert who undergo filter surgery to relieve eye pressure. And he said that the NDA filing would be for the low dose.

In related news, Alimera said that warrants were exercised for an additional $10 million in an extended Series C financing under which it had issued shares of preferred stock and warrants to purchase shares of its Series C preferred stock.

The warrants were to be exercised in full within 30 days of the delivery of top-line data from the company's Phase III trials for Iluvien. The warrant exercise will close in January.

"If approved, we believe this would be the first drug approved for the treatment of this condition [DME]," Dan Myers, president and CEO of Alimera Sciences, said in a statement.

In addition to the full analysis set, Alimera also conducted several other analyses of the 24-month data, as prospectively planned in the protocol. The other analyses included an All Randomized and Treated (ART) analysis of the 24-month data that excludes data collected subsequent to the use of treatments prohibited by protocol (such as intravitreal injections of Avastin, Lucentis or triamcinolone acetonide).

Under the Modified ART method, in Trial A 22.6 percent of patients on the low dose and 24.1 percent of patients on the high dose gained 15 or more letters compared with 12.6 percent of control patients (p = 0.057 and 0.026, respectively). Trial A was not statistically significant for either dose.

In Trial B using Modified ART, 29.7 percent of patients on the low dose and 29.3 percent of patients on the high dose gained 15 or more letters compared with 13.3 percent of control patients (p = 0.004 and 0.005, respectively). The results for both doses were statistically significant.

The FAME study protocol provides that the primary assessment of efficacy will be based on the Modified ART dataset and that the other datasets will be considered secondary. Though, the companies anticipate that the FDA will consider the full analysis set to be the most relevant population on which to determine safety and efficacy in Trials A and B.

Shares of pSivida (NASDAQ:PSDV) rose $1.48 to close at $5.