VaxGen's AidsVax: company admits giving "erroneous information" in trial data

The failure of AidsVax to prevent infection with HIV - in clinical trial results published in 2003 - triggered an intense debate about the controversial product and its manufacturer, VaxGen Inc of Brisbane, California. Mail to this website, maintained by Brian Deer, shows that existing material on a VaxGen-AidsVax index is read by significant numbers. This page seeks to inform the discussion

Dow Jones Business News
VaxGen Says Aidsvax Trial Results 'Remain Accurate'
Thursday February 27, 9:12 pm ET

BRISBANE, Calif. -- VaxGen Inc. (NasdaqNM:VXGN - News) said results from the Phase III clinical trial of its Aidsvax AIDS vaccine candidate "remain accurate as stated" and its analysis continues.

As reported in Wednesday's edition of The Wall Street Journal, VaxGen admitted that an important piece of statistical evidence supporting its claims for the efficacy of the vaccine appears to be weaker than the company had initially asserted.

Yet the company insisted that the problem doesn't appear to seriously degrade the statistical significance of the results.Data from the trial, which VaxGen released on Monday, indicated that Aidsvax failed to protect volunteers from infection by HIV, the AIDS virus. But in an ethnic subgroup of 498 nonwhite, non-Hispanic volunteers, VaxGen said the vaccine appeared to provide protection in the range of 30% to 84%. That analysis, the company said, had less than a 1% chance of being due to random chance, making it highly statistically significant. VaxGen President Donald Francis touted the results as evidence that Aidsvax can protect against HIV infection, although he also acknowledged they reflected preliminary analysis and could turn out to be a "statistical fluke."

Outside scientists and AIDS activists have criticized the claim of partial efficacy, largely because it was based on an analysis of just 29 HIV infections distributed between vaccinated volunteers in that subgroup and those who received a placebo.

VaxGen has said it followed good statistical practice by taking "penalties" related to its analyses of multiple subgroups. Such penalties are designed to reduce the statistical significance of results obtained from slicing a body of data into many smaller pieces. But on Tuesday, Lance Ignon, VaxGen's vice president for communications, admitted that the company hadn't taken those penalties after all. Mr. Ignon said he didn't know how the erroneous information was released, adding that the company was "still trying to figure that out."

In a press release Thursday evening, the company said its analysis of data from the Phase III trial followed a statistical analysis plan that the U.S. Food and Drug Administration (News - Websites) agreed on in advance. That plan included analyses of various subgroups, including racial backgrounds.The company said the number of required adjustments for its subgroup analyses is subject to interpretation and there are a variety of methods to calculate those adjustments. VaxGen said it can't predict the impact these adjustments may have on the findings since that determination ultimately will rest with regulatory authorities. In addition, the potential utility of the candidate vaccine will be evaluated based on the full range of observations, including primary and secondary endpoints, correlates of protection and other aspects of the immune response, the company noted.

VaxGen added that it is currently focused on analyzing these factors, as well as discussing the data and the need for any additional studies with the U.S. Centers for Disease Control and the FDA. The company plans to report additional results at the Keystone Symposia, March 29 to April 4. VaxGen shares, which have fallen sharply all week after the release of the data on Monday, closed Thursday at $4.25, down 57 cents or 11.8%, for the
day.

VaxGen Issues Statement on Clinical Trial Data Analysis
Thursday February 27, 7:02 pm ET

BRISBANE, Calif., Feb. 27 /PRNewswire-FirstCall/ -- In response to media inquires, VaxGen, Inc. (Nasdaq: VXGN - News) issued the following statement:

VaxGen's analysis of data from its Phase III clinical trial of AIDSVAX B/B followed a statistical analysis plan that was agreed on in advance with the U.S. Food and Drug Administration (FDA). The statistical analysis plan included analyses of various subgroups, including racial backgrounds. The results VaxGen reported on Monday remain accurate as stated, and the analysis continues.

The number of required adjustments for VaxGen's subgroup analyses is subject to interpretation and there are a variety of methods to calculate those adjustments. The company cannot predict the impact these adjustments may have on the findings since that determination will ultimately rest with regulatory authorities.

Furthermore, potential utility of the candidate vaccine will be evaluated based on the full range of observations including primary and secondary endpoints, correlates of protection and other aspects of the immune response. VaxGen is currently focused on analyzing these factors, as well as discussing the data and the need for any additional studies with the U.S. Centers for Disease Control and the FDA. VaxGen plans to report additional results at the Keystone Symposia, March 29 - April 4th