Shares of Northfield Laboratories Inc. lost more than half their value after top-line data showed that its blood substitute product, PolyHeme, failed to meet its primary endpoints of superiority and non-inferiority at Day 30 mortality in a Phase III trauma study.

Despite the miss, Evanston, Ill.-based Northfield said it plans to proceed with a biologics license application filing, hoping that the FDA will consider that the product met non-inferiority criteria once a subset of patients who violated protocol was removed from the analysis.

To show non-inferiority, the upper limit of the confidence interval (CI) needed to be 7 percent or less. Analysis of the entire 712-patient modified intent to treat population exceeded the limit by 0.3 percent; however, once the population was limited to the 586 patients who were enrolled or randomized to proper protocol (the per-protocol population), the CI level fell to 5.8 percent. Patients in the trial were randomized to receive either PolyHeme or the standard treatment of saltwater plus blood.

The per-protocol (PP) population results represent "the purest opportunity to observe the treatment effects between the two groups," Northfield's chairman and CEO, Steven Gould, said during a conference call, adding that, based on data from the PP subset, "we will continue to move toward submission of a BLA." Investors, however, were skeptical. Northfield's shares (NASDAQ:NFLD), which traded at more than 20 times their normal volume, lost $6.57 Wednesday, or 57.5 percent, to close at a 52-week low of $4.85.

"I think the chances are pretty slim" that the FDA will grant approval based on that subset data, said analyst Eugene Trogan, of New York-based Morgan Joseph & Co. Inc., since that subset analysis "deviates from the pre-specified" endpoints spelled out in the company's special protocol assessment. Furthermore, that 7 percent upper limit of CI might no longer fall in line with the FDA's requirements for blood substitute products, Trogan said, adding that the agency "seems to be looking at those products more carefully."

Trogan reiterated his "sell" rating on Northfield and lowered the price target from $6 to $4. The stock also was downgraded from "outperform" to "neutral" by New York-based Cowen and Co.

Blood substitute products have been under increasing scrutiny by the FDA, especially in regard to the issue of patient consent. Though Northfield was permitted to conduct its trial on trauma patients under federal regulations allowing consent to be waived in life-threatening conditions, the agency has since been re-examining those rules.

The FDA's Blood Products Advisory Committee voted 11-8 last week against a proposed 1,100-patient Phase IIb/III study of Biopure Corp.'s blood product, Hemopure (hemoglobin glutamer - 250 [bovine]). The Cambridge, Mass.-based company said the Naval Medical Research Center, which is sponsoring the trial, is working with the FDA on plans for a smaller Phase II trial.

Northfield's Phase III trial was designed to evaluate PolyHeme, a human hemoglobin-based oxygen-carrying red blood cell substitute, in severely injured and bleeding patients when a blood transfusion is needed but blood is not immediately available. Of the total 712 patients in the modified intent-to-treat population (MITT), 349 were in the PolyHeme group and 363 were in the control group.

Northfield reported that 20 percent of patients in the PolyHeme group and 15 percent in the control arm represented protocol violations, which included errors related to eligibility and treatment regimen.

Data did not demonstrate superiority to the standard regimen, but "what is most concerning," Trogan told BioWorld Today, is that the preliminary results do show "an increase in the rate of mortality with PolyHeme."

Results from the 586-patient PP population showed 30 deaths (10.8 percent) in the PolyHeme arm vs. 28 deaths (9.1 percent) in the control arm. The larger MITT group reported 46 deaths (13.2 percent) for patients receiving PolyHeme and 35 deaths (9.6 percent) for the control group. Of the 126 patients who violated protocol, 16 (22.8 percent) died in the PolyHeme arm and seven (12.5 percent) died in the control arm.

During the conference call, Gould reported that two discrepancies in the dates of death for patients were discovered while reviewing the top-line data, so the database will be unlocked and corrected prior to finalizing statistical analysis. He said further analysis should be available in the next four to six weeks. In the meantime, the company has a strong balance sheet, Gould said, with "just under $52 million" as of Nov. 30.

PolyHeme has encountered several hurdles in its development history. Northfield's original BLA for the product was turned down by the FDA in 2001. Most recently, the company was named in a class-action lawsuit alleging that it failed to disclose material adverse facts relating to PolyHeme development. The complaint followed news released in February reporting that 10 of the 81 patients who received the product suffered heart attacks within seven days of treatment, and two of those resulted in death.

In response to that news, Northfield said there was no evidence linking PolyHeme treatment to those serious adverse events.