Shares of Panacos Pharmaceuticals Inc. fell 61.7 percent on Tuesday due to preliminary data from the latest cohort of a Phase IIb trial with HIV drug bevirimat.

The data showed a lower mean viral load reduction and fewer responders with the 350-mg solution of bevirimat compared to the 300-mg dose previously tested. During a conference call, Panacos CEO Alan Dunton called the finding "unexpected" and said it "suggests we may be at or near the top of the dose-response curve."

On the plus side, the finding means Panacos needs to analyze just one more cohort in its Phase IIb trial to confirm there is no benefit to further increasing the dose. The company expects to have data from the final, 400-mg cohort by the end of the first quarter of 2008 and then will proceed to a Phase III trial.

But the new data also raised issues - such as the possible existence of responder and nonresponder populations - that could complicate the Phase III trial. Those issues prompted investors to push Panacos' stock (NASDAQ:PANC) down to new 52-week lows before it closed at 80 cents, a loss of $1.29 for the day.

Bevirimat is the first in a new class of HIV drugs called maturation inhibitors. Discovered by Panacos, these small molecules are designed to target a step in the HIV lifecycle that allows them to act against a broad range of viral strains, including drug-resistant strains.

Back in 2005, a Phase IIa trial of bevirimat 200-mg solution administered to HIV patients as a monotherapy for 10 days resulted in a median 1 log10 viral load reduction. Panacos moved into a 14-day Phase IIb trial using a 400-mg tablet formulation of bevirimat on top of whatever HIV regimen the patients were failing to respond to, but the pill resulted in lower plasma levels and a mean viral load reduction of just 0.36 log10.

After meeting with the FDA, Panacos decided to complete its Phase IIb trial using the solution rather than the tablet. A 250-mg cohort achieved a mean 0.68 log10 viral load reduction, with three of seven patients (43 percent) achieving a 1 log10 reduction, and a mean bevirimat trough plasma concentration of 38.3 mcg/ml. The 300-mg cohort tested next looked good, with a mean 1.02 viral load reduction, five of eight patients (63 percent) achieving a 1 log10 reduction, and a mean bevirimat trough plasma concentration of 48.4 mcg/ml.

Then came the data from the 350-mg cohort, which might have been expected to keep increasing, but instead showed a mean 0.62 log10 viral load reduction, with just three of nine patients (33 percent) achieving a 1 log10 reduction and a mean bevirimat trough plasma concentration of 43.8 mcg/ml.

Panacos Chief Operating Officer Graham Allaway concluded that absorption was not the issue and pointed to a potential peak for maximum efficacy. He added that the low number of patients achieving a 1 log10 reduction suggest there may be distinct responder and nonresponder populations. As further evidence, he pointed to the fact that viral load reduction tended to fall either above 1 log10 or below 0.5 log10, and a subgroup analysis of just responders showed a linear dose-response relationship.

In a research note, analyst Brian Abrahams of CIBC World Markets agreed that the data support the responder/nonresponder theory. However, he added that "without known predictors of response . . . development and commercialization will be more challenging."

Allaway said Panacos has not yet figured out how to identify responders. Analyses of patient history, background medications and viral genotypes have failed to provide the answer, and he proposed that variations in protein binding could be involved. But until Panacos knows for sure, the company is considering a Phase III trial design that would involve treating all patients initially for a few weeks and rolling responders into the broader trial.

Eric Schmidt, analyst with New York-based Cowen and Co., downgraded Panacos from outperform to neutral. In a research note, he said that while "the totality of Phase II data support maturation inhibition as a viable therapeutic strategy and indicate bevirimat may be an approvable drug, a variable dose response curve adds further uncertainty to the bevirimat development program."

Beyond the Phase III trial, Panacos may face commercialization hurdles as well. Although Dunton estimated that about half the patients in the trials were responders, Abrahams predicted that "given the crowded landscape among HIV therapies, the requirement to assess after the start of dosing could be somewhat of a commercial liability."

Beyond bevirimat solution, Watertown, Mass.-based Panacos anticipates beginning clinical trials in 2008 with another try at a solid formulation of the drug. The company also initiated Phase I trials earlier this year with PA-040, another maturation inhibitor for HIV that may offer broader activity than bevirimat, and it plans to enter the clinic next year with its small-molecule HIV fusion inhibitor program.