SPERMICIDE MADE BY COLUMBIA LABS
FAILS TO STOP HIV IN CLINICAL TRIAL
By Martha Brannigan
Wall Street Journal 13 June 2000
A clinical trial sponsored by the United Nations group UNAIDS, using 700
prostitutes in South Africa and other developing countries, has ended in
bitter disappointment after about 100 women included in the study of
Columbia Laboratories Inc.'s spermicide contracted HIV.
Researchers had hoped the spermicide would prevent transmission of the
virus that causes AIDS. But in a bewildering development for
researchers, the group of patients in the experiment whose members
received a placebo, the vaginal cream Replens, appeared to show a
somewhat lower HIV infection rate than those using the once-promising
"Experts in the field of AIDS didn't foresee this," William Bologna,
Columbia's chairman and chief executive officer, said on a conference
call Monday. "This was obviously unexpected and unanticipated." It isn't
clear how the HIV transmission rate for the test group compares with the
infection rate among prostitutes outside the study. The company's
spermicidal gel, called Advantage-S, in earlier Phase I and Phase II
studies had held the promise of blocking HIV transfer.
Officials of UNAIDS involved in the study couldn't be reached.
The clinical study was conducted on 700 women working as prostitutes in
South Africa, Thailand, Ivory Coast and a fourth African country between
1996 and May 2000, Mr. Bologna said in an interview. The women were
randomly assigned to eight groups in the double-blind test, with half
getting the placebo and half receiving the Advantage-S spermicide, which
uses a lower-than-normal concentration of a drug known as Nonoxynol-9.
Mr. Bologna said all women in the UNAIDS study were provided with
condoms and encouraged to get clients to use them. He said the women
were asked to keep diaries and overseen by monitors to track how they
used the drugs and condoms.
Nonoxynol-9 is a common ingredient in spermicides in the U.S., but
Columbia used technology to increase the efficacy so it could be used at
a lower dose, Mr. Bologna said. According to the World Health
Organization, Nonoxynol-9, which has been available for many years and
was generally considered safe, kills HIV in the test tube and provides a
physical, "gooey" barrier that sperm and microbes cannot penetrate.
However, higher concentrations of the drug have recently been found to
cause vaginal lesions, which can then be entry points for HIV. Health
experts have warned against overuse of the drug, particularly by
prostitutes who might apply it repeatedly during a single evening.
The clinical trial was ended recently after the number of women testing
positive for HIV reached 100 of the 700 women in the study. "It is
unfortunate for the poor women in Africa, because we thought we had a
product that worked," Mr. Bologna said. "They need a product like this
because men aren't going to use condoms."
Mr. Bologna defended the ethics of using a group of largely illiterate
prostitutes for the test. "There was a lot of debate about the issues
involved. The reality is, if you are going to help these women, you have
to conduct the tests on them. Mary Jane in Toledo, Ohio, is not at high
risk for the transfer of HIV. It's these poor women in Africa."
Moreover, he said a drug-safety monitoring board had worked closely with
UNAIDS to make sure the patients' interests were observed. "Obviously,
everyone tried to do this to the highest ethical standards," Mr. Bologna
In 4 p.m. American Stock Exchange trading Monday, Columbia fell $7.1875,
or 56%, to $5.75. In March, shares of Columbia, a Miami pharmaceuticals
company, soared after news that an interim analysis of the trial
indicated one group of patients in the double-blind trial was doing
better than the other. Mr. Bologna said "people assumed" that the group
doing better in the study was the one using the spermicide. As it turned
out, the placebo group appears to have done better, although Mr. Bologna
cautioned that those results may not be scientifically meaningful.
Barbara Martinez in New York contributed to this article