June 4 (Bloomberg) -- Telik Inc. said women with ovarian cancer who took its experimental drug Telcyta in a trial died five months earlier than those on standard chemotherapy, driving shares down as much as 23 percent.

The findings, presented at a medical meeting in Chicago, showed that patients who took Telcyta lived for a median time of 8.5 months, compared with 13.6 months for patients on standard chemotherapy, the company said in a statement. Another study, in lung cancer, also showed patients died more rapidly on the drug.

Telik, with no marketed drugs, has lost 72 percent of its value since Dec. 26, 2006, when it announced that Telcyta, its most advanced experimental drug, didn't prolong life in patients with ovarian and lung cancers. After Telik released detailed data last night, it faced questions about how patients on a standard drug could have lived so much longer.

``This is nightmare data,'' said David Miller, president of Biotech Stock Research, an independent equity research firm in Seattle, in an interview yesterday.

Telik's shares fell $1.22, or 21 percent, to $4.59 at 4 p.m. New York time in Nasdaq Stock Market composite trading, after reaching $4.47. The findings were released last night at the annual meeting of the American Society of Clinical Oncology.

A Telik spokeswoman, Carol DeGuzman, said an independent group of safety monitors met quarterly to oversee the trial, and never recommended it be halted because of adverse events. The company can't explain why patients on chemotherapy lived longer than those on its drug, she said today in a telephone interview.

`Serious Enough to Report'

``You have to wonder why they did so poorly, and whether it was serious enough to report to the medical community,'' said Len Lichtenfeld, deputy chief medical officer of the American Cancer Society, based in Atlanta. Since Telik relied on its independent experts for advice, company officials ``did what they were supposed to do.''

Michael Bookman, an ovarian cancer specialist at Fox Chase Cancer Center in Philadelphia, said he worked on an earlier, smaller study on Telcyta, and that researchers at his center declined to participate in this trial because of concerns it was poorly designed. Telik didn't have enough scientific evidence to move forward so quickly, he said.

``Small companies with a small portfolio of products that they are aggressively trying to promote often overreach in their studies,'' Bookman said in an interview. ``It's an important question that needs to be asked.''

`Stop the Trial'

Ren Benjamin, an analyst at Rodman and Renshaw in New York, said, ``If something's not right in a trial, they are supposed to stop it. I've never seen data like this.''

Karen Riley, an FDA spokeswoman, declined to comment.

Yesterday's statement by the company added further data about the ovarian cancer study, known as Assist-1, than was available in December.

The trial enrolled 232 women who took Telcyta and 229 women on either Doxil or topotecan, chemotherapy drugs. The patients had failed on prior therapies. Patients who entered the study had run out of treatment options and were expected to live six months, Telik's DeGuzman said.

The company first learned of the failed studies in December 2006, DeGuzman said. Patients in both the Telcyta group and the chemotherapy group stayed on therapy for about three to four months, then left the trial and received other treatments or hospice, DeGuzman said.

Drug Well Tolerated

The deaths ``absolutely'' didn't appear related to Telcyta side effects, and the drug was well-tolerated, she said. The FDA reviewed the trial design before it began, DeGuzman said.

``Patient welfare is of the absolute utmost concern to us,'' DeGuzman said.

The company still has a Telcyta trial actively enrolling patients with ovarian cancer, and it is expected to continue, DeGuzman said.

``We were disappointed with the results,'' DeGuzman said. ``We believe very much in what we're doing with Telcyta and are continuing development.''

Another clinical trial on Telcyta, called Assist-2, enrolled 530 patients with lung cancer, the company statement said. Patients who took Telcyta lived a median time of 4.6 months, compared with 6.1 months for those who took another drug, AstraZeneca Plc.'s Iressa.

The study failed its primary and secondary goals of prolonging lives and slowing the spread of the disease, the company said in a statement. The company said the two patient groups had similar disease prognosis.

Telintra

Telcyta is a conventional chemical compound designed to be active inside tumors, Telik said in a statement. The company has another drug in development, Telintra, for myelodysplastic syndrome.

Bookman, at Fox Chase, said that after he participated in the smaller study, he offered to do tumor biopsy tests to confirm that Telik's drug worked the same way in humans that the company suggested it did in animal tests.

Telik refused, saying it was concerned such data might interfere with its ability to get regulatory approval, he said.

Patients who entered the trial had failed on platinum-based chemotherapy, meaning they were likely to die within months. Doctors typically give another round of chemotherapy to such patients, such as doxorubicin or topotecan, which has been shown to prolong lives.

Because Telik had so much confidence its drug would work, it designed the study to directly compare Telcyta to one of the proven chemotherapy agents, Bookman said. The 50 percent risk patients had to take, that they might not get a proven drug, made doctors reluctant to join the study, and caused patient enrollment to be slow, Bookman said.

Telik, founded in 1988, had 118 employees at the end of February, and an accumulated deficit of $393 million at the end of the year, according to its most recent annual report with the Securities and Exchange Commission.

Eastbourne Capital Management LLC, Oppenheimer Funds Inc. and Icahn Associates Corp. are the largest holders of Telik shares, according to data compiled by Bloomberg.

To contact the reporter on this story: Luke Timmerman in San Francisco ltimmerman@bloomberg.net.

Last Updated: June 4, 2007 17:55 EDT