DRUG CONCENTRATION-CONTROLLED TRIAL BENEFITS OVER DOSE RESPONSE BEING CONSIDERED BY FDA; BLOOD LEVEL DATA SEEN AS VALUABLE PHARMACOKINETIC SCREEN

FDA is evaluating the benefits of drug concentration- controlled trials vis-a-vis traditional dose-response studies, FDA Center for Drug Evaluation and Research, Research and Methodology Planning Staff Director Lilly Sanathanan, PhD, told an April 24-26 conference on integrating pharmacodynamics, pharmacokinetics and toxicokinetics into drug development. "The reason we are considering concentration-controlled trials is that we believe it leads to greater efficiency," Sanathanan said, with efficiency defined "in terms of sample sizes required" for randomized concentration -controlled and randomized dose- response trials. In one scenario she worked out, a dose-response trial would require about two-and-one-half times as many patients than would a concentration-controlled trial. CDER Director Carl Peck, MD, noted: "We have realized that there is some variability in the exposure to patients . . . when controlling the dose level alone." Peck continued: "Several of us at the agency in collaboration with . . . industry and academia have begun to think about how to refine that control in the form of concentration control." Sanathanan added that "we at FDA are quite interested in developing this further and in collaborating with people in industry in terms of maybe drafting some guidelines for concentration-controlled trials." FDA is planning to integrate ideas from the meeting on PK/PD studies into the clinical guidelines for drug development (see related story). She pointed out that "for drugs with high [pharmacokinetic] variability, you will see some substantial gains" in efficiency when using concentration-controlled trials. "In going from dose to concentration, what you're doing essentially is minimizing the inter-individual variability in [pharmacokinetics]." Concentration-controlled trials would be particularly useful for narrow therapeutic range drugs, Sanathanan said. "The higher the concentrations the more . . . efficiency. This particular point has a message in terms of drugs with narrow therapeutic windows because those are the types of drugs where you will need to go to higher concentrations in order to see some drug effect." FDA believes that concentration-controlled studies are "best suited for the early phase of drug development, Phase II ideally, in some cases, Phase I, especially for drugs with narrow therapeutic windows, primarily for safety reasons," Sanathanan said. Investigators will need "some information from Phase I in order to be able to start [the trials]. Certainly you need the PK information and hopefully some knowledge of what [are] the maximum safe concentrations." One design for a concentration-controlled trial under consideration by FDA "basically consists of two parallel arms, one with maximum safe concentration and one a placebo group that gets titrated upward," Sanathanan said. This design "lets you evaluate efficacy early on. Plus it gives you some idea of what the concentration response is, so that you can have some idea of what the dose should be for Phase III." Both Sanathanan and Peck pointed out that concentration- controlled trials have been done and are ongoing. Peck said, "one of the most complicated ones I know of" underway now with FDA input is at "NIDA [the National Institute on Drug Abuse]." That trial is testing carbamazepine for the treatment of cocaine addiction ("The Pink Sheet" Dec. 3, T&G-12). The agency is also looking at blood level data to explain unexpected results during drug development. FDA Office of Drug Evaluation I Director Robert Temple, MD, told the scientists that blood level data collected from Phase II and Phase III trials can be used as "a kind of screen for unexpected pharmacokinetic variability -- a check on some of the basic PK assumptions that have been made in the development of the drug." Blood level data along with the clinical results can serve as "a possible explanation for surprising outcomes, or to put it slightly differently, a source of testing hypotheses about how to develop the drug and how to use it," Temple said. One current example where blood level data is serving as a wellspring of hypothesis-making is the analysis of data collected from the trials of Warner-Lambert's Alzheimer's drug Cognex (tacrine). Temple noted that data from the first trial presented at the March 15 FDA advisory committee on Cognex, "blood levels were obtained in 77 out of about 100 patients who received tacrine." Most of the patients who had blood concentrations of 15 or 20 nanograms per ml showed improvements in the ADAS cognitive scale, Temple noted, whereas the numbers of patients with concentrations below 15 ng per ml were split in terms of getting better or worse. Temple said that "in 65 patients, with less than 15 ng per ml, 35 worsened, nine had no change and only 21, that's 32% improved. In the placebo group, 21% improved." For the "12 patients with blood levels of at least 15 ng per ml, however, all but one improved for a 92% response," Temple noted. When movements of at least two units in a favorable direction on the scale were measured, "a lot of the action and a fairly consistant rate of improvement, 60% for the 15 to 20 group and almost 80% for the greater than 20 group, seems to occur when you get the blood levels up higher." In the clinical global scale assessment, Temple added, "you can see that there's essentially nothing going on until you get to the six patients out of 77, who managed a blood level of at least 20 [ng/ml], and 100% of them had an effect on global." "One possible explanation here is that there is a plasma concentration below which not much happens and above which something quite interesting occurs," Temple said. "It's pretty obvious that this kind of data isn't an answer in and of itself, but is an interesting path to some questions." One question that the data raises is: "Can you reliably say that patients with concentrations of 15 [ng/ml] or 20 [ng/ml] respond much more often to a significant degree than others?" A second question, Temple posed, is: Why did some people have higher concentrations than others? He noted that of "42 patients with blood level measurements, who got 80 mg of tacrine, only 12, that's 29%, had at least a 15 [ng] level and only 15% had at least a 20 [ng] level." Temple pointed out that the dose of tacrine used in the trials had been lowered to 80 mg because of hepatotoxicity seen at higher doses. So a third question is: "Can higher blood levels be obtained, if indeed they really are useful without unacceptable toxicity." A fourth question is: "How do the well-known active metabolites fit into all this?" Temple emphasized that "without the blood level data there would have been very little to explain the very modest effects that were seen." He added that the blood level information is "probably one of the most important things to be found out about the drug."