Panel Recommends FDA Approval of Gliadel for GBM

July 1, 1996

GAITHERSBURG, Md--An FDA advisory panel has recommended approval of Guilford Pharmaceutical's Gliadel Wafer (polifeprosan 20 with carmustine) for use as an adjunct to surgery to prolong survival in recurrent glioblastoma multiforme (GBM).

GAITHERSBURG, Md--An FDA advisory panel has recommended approvalof Guilford Pharmaceutical's Gliadel Wafer (polifeprosan 20 withcarmustine) for use as an adjunct to surgery to prolong survivalin recurrent glioblastoma multiforme (GBM).

The Oncologic Drugs Advisory Committee (ODAC), however, urgedthat FDA approve the use of the drug delivery system only forGBM patients until further study resolves efficacy issues.

The Gliadel system consists of a wafer containing polifeprosan20, a biodegradable, polyanhydride co-polymer, and carmustine(BiCNU), said Ross S. Laderman, Guilford's vice president forregulatory affairs. About dime-size in diameter, the wafers measure1 mm thick. Up to eight wafers, and almost always a minimum ofseven, are placed in the brain cavity created by the tumor's removal.

The wafers release most of their car-mustine within a few days.About 50% is released in 3 days and "greater than 80% isreleased within 7 days after wafer implantation," said Dr.Craig R. Smith, Guilford's CEO. "Tumoricidal concentrationshave been found adjacent to the wafers up to 21 days after implantation."

Dr. Earl Webb Henry, Guilford's vice president for clinical research,reviewed two phase III randomized, controlled, double-blind Gliadeltrials. Study 8802 involved 222 patients undergoing surgery forrecurrent malignant glioma. The primary endpoint for this US-Canadianstudy was 6-month survival.

Study CL-0190, which used 12-month survival as its primary endpoint,was carried out in Finland and Norway. Its protocol called forup to 100 patients undergoing initial surgical resection for malignantglioma. But because of an inadequate supply of wafers, researchersenrolled only 32.

In trial 8802, 110 patients received carmustine-containing wafersand 112 got drugless wafers. "At 6 months, 60% of all Gliadelpatients and 47% of all placebo patients were still alive,"Dr. Henry said. "The survival difference was at the marginof significance (P = .06)." Gliadel patients showed an earlysurvival benefit, "but nearly all of this was within thefirst 12 months," he noted.

Sixty-five percent in each group had GBM. At 6 months, 56% ofthe Gliadel patients and 36% of the placebo recipients were alive.Mean survival reached 7.2 months in the treatment group and 5.4months in the placebo group. "This 33% increase in mediansurvival is clinically meaningful," Dr. Henry noted.

In CL-0190, all 16 of the Gliadel patients had GBM, but only 11of 16 in the placebo group had this diagnosis. "Expectedmedian survival in the placebo group was 12 months," Dr.Henry said. "Gliadel treatment was to be considered effectiveif it prolonged survival by 33%."

At the 12-month time point, 63% of the Gliadel patients and 19%of those in the placebo group were still live. "The P valuefor this survival was .03," he said.

Safety Data

Dr. Henry reviewed safety data derived from 337 patients in NDAclinical trials--209 Gliadel-treated and 128 placebo patients.The wafer was "not judged a solo or contributory factor"in the death of any patient in the studies, he said.

Frequent adverse events were "nearly identical" in thetwo groups, Dr. Henry said. However, fever, pain, abnormal healing,and urinary tract infection occurred more frequently in Gliadelpatients; anemia affected the placebo patients more often. Intracranialinfections (meningitis and abscess) proved low for both groupsin the controlled trials, Dr. Henry added. The frequency of newor worsened seizures in the controlled trials was similar--20%in the treated group and 19% in the placebo patients.

In their discussion of the data, panel members debated such issuesas the few patients in CL-0190 and its truncated enrollment, andthe statistical value of any findings related to malignant gliomasother than GBM.

The committee voted unanimously that 8802 was adequate and well-controlled.But it split 4-to-4 whether the study provided convincing evidenceof the wafer's efficacy. The division stemmed in part from questionsof whether the two patient groups were adequately balanced andthe lack of statistical significance. "This is somewhat disappointingin what I think was an excellent study," said Judith Ochs,MD, of Arkansas Children's Hospital, Little Rock.

On the question of whether Gliadel's toxicity was acceptable forpatients with recurrent malignant gliomas, the panel agreed 7-to-0,with one abstention. Finally, ODAC voted 7-to-1 to recommend approvalof Gliadel solely for use in people with recurrent GBM.

Paul Bunn, MD, of the University of Colorado Cancer Center, Denver,acknowledged the small magnitude of the difference between thetwo patient groups in 8802. But he said that "the consistencyof the evidence" across the NDA trials persuaded him. "Wedon't have much to offer these patients," Dr. Bunn said."I think it's reasonable for Gliadel to be out there as anoption."