Since 1991, at least eight clinical trials of drugs to treat sepsis have failed to show efficacy: two by Centocor Inc., two by Xoma Corp., two by Synergen Inc., one by Cortech Inc. and one by Chiron Corp.

The latest failures, a Phase III study by Synergen Inc. and a Phase II trial by Cortech Inc., appear to have been well designed and carefully conducted, as both companies had built on the experiences from previous sepsis studies.

The results place a question mark both over the ability of any single drug to show a reduction in mortality from the syndrome and over the entry criteria and end points being used to assess sepsis drugs.

But the fact remains that sepsis patients continue to die and that at least some of the drugs tested probably have some biological activity. In fact, the potential market remains sufficiently attractive that Piper Jaffray analyst Barbara Hoffman estimates 25 companies are still working in sepsis. So the problem remains how to show that a drug or combination of drugs works.

What's been learned

"The question is why does something with biological activity fail?", said Timothy Rodell, executive vice president for operations and product development at CRTQ.

Experts in the field are focusing on two issues: the heterogeneity of the patient population and the specific points in the disease at which drugs are presumed to exhibit their effects.

"I think these drugs have a biological effect," said William Knaus, director of the ICU research unit at George Washington Medical Center. "But there's no way to show it if we keep using imprecise patient selection criteria and an end point of mortality. If other drugs had to use this threshold, we wouldn't have most drugs on the market."

Both SYGN and CRTQ tried to tackle the heterogeneity problem through careful enrollment criteria designed to isolate a subset of patients in whom previous trials had indicated the drugs would work.

SYGN had planned to enroll patients using the APACHE III scoring system developed by Knaus, which stratifies patients by risk. The company had used the prognostic system in parallel to its main enrollment criteria in its first Phase III study.