XOMA's E5 FOLLOW-UP PHASE III STUDY SHOWS POSITIVE EFFECT IN REVERSING THREE KINDS OF ORGAN FAILURE, XOMA PRESIDENT SCANNON TESTIFIES IN PATENT TRIAL

Xoma's second Phase III clinical study of their monoclonal antibody E5 in the treatment of gram negative sepsis shows a positive effect in reversing three kinds of organ failure when the data for the three groups is pooled, Xoma President Patrick Scannon, MD, reported during the Xoma v. Centocor patent trial in San Francisco federal court on July 30. Scannon said the follow-up study showed E5 to be effective in reversing organ failures in the kidney, the intravascular coagulation system, and the lungs when the data from the three groups is combined. The trial in sepsis patients not suffering from shock is intended to support an earlier clinical trial submitted to FDA in Xoma's PLA for E5. Scannon testified that FDA allowed Xoma to unblind the second study on June 21. However, during cross-examination, Scannon acknowledged that the preliminary results of the second Phase III study yielded no statiscally significant results in overall reduction of mortality among nonshock patients, nor among subgroups of nonshock patients with documented gram-negative bacteria, documented gram-negative bacteria with major organ failure, patients with blood clotting failure alone, patients with kidney failure alone, or patients with liver dysfunction. Xoma is scheduled to appear before the FDA's Vaccines and Related Biologic Products Advisory Committee at their Sept. 4-5 meeting. Scannon said that the agency has notified Xoma that it would like the company to present "all of its data" at the upcoming advisory committee meeting. Scannon stressed the preliminary nature of the data in his testimony. "It normally takes two or three months at a minimum to complete all of the analyses that one has to do," Scannon told the court, "but we have information that we can share on our progress in that analysis." According to Scannon, the protocol for the second Phase III study was written in the fall of 1988 and began in February 1989. The multicenter, randomized, double-blind, placebo-controlled study looked at mortality in 811 patients suffering from gram- negative sepsis but excluded any patients in shock. Scannon said the study showed that E5 can reverse organ failure in patients with liver dysfunction and "disorders of the brain" when the data is pooled. He noted that "all of the organ systems, whether it was the clotting system, the kidney system, the lung system, the brain system, or the liver system . . . appeared to benefit from the treatment with E5." However, he acknowledged that the data supporting organ failure reversal only becomes statistically significant if data from different organ failure groups is combined. In addition, Scannon claimed that the study provides "the first evidence that E5 can actually prevent . . . some of these organ failures." Patients with gram negative sepsis but without organ failure were treated when entering the hospital, Scannon noted, "and what we were able to show is that . . . more placebo patients went on to get complications than the patient who got E5. . . . That says that E5 was able to prevent some of these complications from occurring." Centocor attorney Donald Dunner, from the D.C. law firm Finnegan, Henderson, Farabow, Garret & Dunner, conceded that E5 showed statistically significant improvement in the reversal of acute renal failure with or without disseminated intravascular coagulation, in hepatobiliary dysfunction in combination with central nervous system dysfunction, and in central nervous system dysfunction alone. He also agreed that the combined efficacy results in kidney, lung, and clotting systems was also significant. Scannon conceded that the second study did not confirm the positive results of the first in terms of overall mortality reduction for nonshock patients and overall organ failure. This second Phase III study is a follow-up to Xoma's first clinical study, the results of which were presented at the Interscience Conference on Antimicrobial Agents and Chemotherapy on Sept. 18, 1989, two days before Centocor filed a PLA for HA-1A (Centoxin). This multicenter, randomized, double-blind, placebo- controlled study looked at 468 patients over 30 days; the results showed that E5 had no effect in patients suffering from septic shock, but showed a statistically significant reduction in mortality among gram-negative patients not in shock both with and without bacteremia. Scannon, during cross examination, acknowledged that the first E5 efficacy study was turned down for publication in the New England Journal of Medicine in March 1990. Centocor attorney Dunner read a letter in which NEJM reviewers listed their "'specific concerns'" with the study -- "'Safety has only been partly tested. Re-exposure to subsequent murine antibodies may cause problems. Long term consequences have not [been tested]. Exposure group is small.'" Centocor's attorneys revealed in opening arguments of the trial the contents of a letter Xoma received from FDA in April 1990 questioning the E5 data submission. During the trial, Dunner read a passage from the letter in which FDA told Xoma that "the data presented in [the E5 PLA] are not sufficient for approval of this product." In the letter, FDA said it based its conclusion "on the limitations of the antibody itself, for example, its failure to react significantly with the type of endotoxin found in gram-negative sepsis, those of the smooth type; the lack of convincing preclinical protection studies in animals; and the very marginal at best efficacy data in man." The letter concluded: "If you intend to pursue this application for this indication, it will be necessary to submit an amendment addressing all issues raised in this review, including additional clinical data." Scannon said he interpreted the letter to mean "in this application the data as we presented it were not sufficient." Scannon claimed that FDA's concerns have been addressed by an October 1990 submission to the agency from Xoma, which answered what Scannon called "two categories" of inquiry: "those things that we wanted to supply additional information on and some things really that the FDA misunderstood." He acknowledged that "maybe we could have done a better job of explaining" in the initial submission. Scannon said that FDA's only communication with Xoma subsequent to the October submission was the notification of the Sept. 4 advisory committee. Dunner suggested that Xoma changed its primary efficacy endpoint midway through its first Phase III study, and in the second study looked only at patients not suffering from shock because this was the only area in which the drug showed efficacy. Dunner accused Scannon of having "dredge[d] through your data in an effort to find some efficacy" and suggested that FDA shared a similar skepticism about Xoma's first Phase III study by reading another passage from the April 10 letter: "The primary objective of the interim analysis in an ongoing trial is to look for early dramatic benefits or potential harmful effects of the treatment under study. However, it appears that in this case, you were analyzing various subgroups of patients to find some evidence of efficacy of Xomen-E5." Scannon called this a "misunderstanding" that had been corrected by Xoma's October 1990 submission. He distinguished the importance Xoma attached to different endpoint groups of patients by specifying whether they were defined as a "primary analysis" in the protocols or whether they were included as a supplemental analysis in an appendix to the analysis plan. The original efficacy endpoint proposed by Xoma for its first Phase III trial was "reducing the mortality rate of patients with gram-negative sepsis," Scannon said. Secondary endpoints were defined as "preventing or shortening the duration of major morbid complications of gram-negative sepsis," which were defined as "adult respiratory distress syndrome, ARDS; disseminated intravascular coagulation, DIC; and acute renal failure, ARF; and shock." Scannon defended the use of nonshock patients as a primary efficacy endpoint in the second trial. He explained: "when we define endpoints like shock . . . it is clear that the difference between all patients and the shock patients is the nonshock patients." Scannon admitted that "the FDA require[d] additional explanation" of this point. Xoma's October 1990 response to the letter, Scannon explained, makes clear that both shock and nonshock were primary efficacy endpoints, calling this difference of opinion "the source of the complication, because that April 10, 1990 letter assumes that nonshock is a secondary analysis." Scannon said that the company believes it has resolved the issue with FDA in its response to the April 10 letter by showing that the non-shock group was "a primary analysis." Xoma's October 1990 response to FDA conceded: "It is correct to conclude that gram-negative sepsis was a primary group for analysis, and this remained true throughout the study. This is the reason all groups of patients continue to be enrolled subsequent to the interim analysis." However, the letter went on to say that "the original study plan also included a provision for examining the efficacy in patients with and without shock." Xoma maintained in the letter that "the finding of a survival advantage in the group without shock is not limited to hypothesis generation, but is a statistically valid finding in itself." To further clarify this point, the October submission offers a breakdown of the patient groups. Scannon said that "the FDA had suggested that there were nine groups"; Xoma's response stated that "there were three rather than nine primary groups analyzed: (1) all patients with gram-negative sepsis; (2) patients with gram negative sepsis with shock at entry; and (3) patients with gram negative sepsis not in shock at study entry." Two class action lawsuits were filed by Xoma investors against the company on July 24, claiming that Xoma withheld vital information, including the April 10 letter, concerning the progress of FDA review on the E5 PLA. Salit v. Xoma and Miller v. Xoma, both requesting jury trials, were filed in the San Francisco federal court by the law firm of Milberg, Weiss, Bershad, Specthrie & Larach. Xoma did not disclose the specific language of the April 10 letter in its financial filings, but said in its quarterly report for mid-1990 that FDA "has raised various issues with respect to [the E5 PLA] and has requested additional information including data before further considering approval of [the] PLA." The disclaimer also applied to CD5, recently recommended for approval by the Biological Response Modifiers Advisory Committee ("The Pink Sheet" June 17, p. 19). Xoma Chairman Steven Mendell told the annual Robertson Stephens medical conference in November 1990, that the request was routine, and that FDA did "not request any additional clinical trials. They are just responding to the data previously submitted" ("The Pink Sheet" Dec. 3, 1990, T&G-2). On July 23, Xoma attorney Gerald Sobel, N.Y. law firm Kaye, Scholer, Fierman, Hays & Handler, revealed that Centocor received a letter from FDA Center for Biologics Evaluation and Research Acting Director Gerald Quinnan on Feb. 25 listing 129 questions "about the adequacy of Centocor's application for approval for HA- 1A." During the trial, Sobel summarized the questions, saying that some concerned deaths of patients on placebo not related to septic shock but "counted by Centocor in its study . . . such as if you fell down an elevator shaft." Many of the questions relate to "the mouse parts in HA-1A," specifically asking for "'assays . . . of higher sensitivity'" looking at serum antibodies of patients receiving HA-1A. Another question asks Centocor to "'comment on the effective different formulations of HA-1A on the pharmacokinetics and efficacy studies.'" The clinical study published on Centoxin in the Feb. 14 issue of the New England Journal of Medicine elicited at least eight letters to the journal published in the July 25 issue. These letters questioned the conclusions drawn by the authors of the study, particularly concerning the general use of HA-1A to treat sepsis, when the drug showed no statistically significant results in treating sepsis overall, but only in sepsis patients with bacteremia. Commenters were concerned that since there exists no efficient marker for the presence of bacteremia, many patients might be treated with HA-1A who were unlikely to benefit from it and who might in fact needlessly suffer from negative effects of the drug. Neither Xoma nor Centocor would comment on either the upcoming advisory committee meeting or on the testimony, citing a gag order entered by Judge Schnacke on the first day of the trial that forbids the two companies, their authorized representatives and all counsel from "orally or in writing discussing or commenting on the evidence presented, arguments of counsel, or rulings of the court in this proceeding with any members of the media, financial analysts, or other commentators."