Docket No. 2004N-0432From: Innis, Robert (NIH/NIMH) [innisr@mail.nih.gov] Sent: Wednesday, November 17, 2004 3:53 PM To: 'fdadockets@oc.fda.gov' Subject: Docket No. 2004N-0432 RECEPTOR OCCUPANCY: LIKELIHOOD OF PHARM EFFECTS In my opinion, the barriers for First-in-Human use of PET radiotracers should be reduced because: 1) The compounds are typically injected at low mass doses. 2) The tracers will help to explore pathophysiology. 3) The tracer can be useful biomarkers for the development of therapeutic agents. On the other hand, the FDA is charged to protect safety of research subjects. I think all would agree that if a compound is injected at tracer pharmacological doses, then the regulatory barrier should be much lower. Why would any animal tox data be needed, if we truly know that the radiopharmaceutical has no pharmacological effects? Thus, a critical issue is what constitutes adequate knowledge that the dose of a particular compound lacks pharmacological effects? For those situations in which the target is a receptor, I think that measurement of receptor occupancy may be the best method to demonstrate lack of pharmacological effects. This argument is based on the reasonable assumption that for very low doses of a non-biological compound, both effects and side effects would be expected to be receptor mediated. At low doses, the only significant effects on the body would be mediated by a high affinity receptor. That is, at low concentrations, only a high affinity receptor (by definition) can "sense" in any significant way the presence of the drug. [Please note that some biological compounds (e.g., foreign proteins) could induce an allergic response at very low doses - and my comments deal with non-biological compounds.] Thus, if the pharmacology of the target receptor is reasonably well known from animal and/or human studies, then "receptor occupancy" can predict the likelihood of pharmacological effects / side effects. For example, if pharmacological effects are first seen with minimal receptor occupancy of 10-20% - and if animal studies can demonstrate that that the tracer occupies only 1% of receptors - then one could be highly confident by a 10-20 fold factor that the radiopharmaceutical would lack pharmacological effects. In summary, I recommend that the FDA include wording in revised regulations for the "Exploratory IND" and for RDRC that would allow (or even recommend) the use of receptor occupancy data (when they exist) to assess the likelihood of pharmacological effects or side effects. "RDRC SHOULD BE ALLOWED TO DO FIRST-IN-HUMAN STUDIES" The current RDRC regulations state that the radiotracer "shall be known not to cause any clinically detectable pharmacological effect in human subjects." I've had difficulty with this phrasing for many years for two reasons: 1) What constitutes adequate knowledge? 2) What does clinically detectable mean? For example, if we knew from animal studies that the first observable effect occurred at a dose of 5 mg/kg, then would one have enough "knowledge" to inject just one molecule? How about 2 molecules? How about 1 ng/kg? This regulation places greater value on data in humans than in animals, and I doubt this is true for low doses. Many radiopharmaceuticals are injected in humans at truly tracer doses, with no reproducible pharmacological effects and are essentially equivalent to a placebo. However, even placebos have "clinically detectable" effects. How large a sample is required to determine with adequate power that the effects from the radiopharmaceutical are not different from those of placebo injection? For example, I've seen patients become anxious while entering a PET camera, with increase in heart rate and blood pressure. But was this the camera - or the radiopharmaceutical? Since we're dealing with low doses, what is the best way to know a lack of pharm effects? As stated, I think human data are flawed at low doses because of the difficulty to distinguish effects from those of placebo. High dose animal data are probably more reliable than low dose human data. I have more confidence in the existence of a true pharmacological effect if it occurs in a dose -related fashion in animals (often "high" doses), than low/tracer doses studies in humans. In light of these concerns, I recomend that the wording be changed. 1) Change "clinically detectable" to "clinically significant." This will allow some medical judgment to be applied. 2) Change the criteria for knowledge to allow animal data. "shall be known with reasonable certainty not to cause ..." "The expectation of no pharmacological effect can be based on animal pharmacological studies (at many multiples of the human equivalent dose) or the use of the compound in human subjects (at either high or low doses)." In summary, I think the RDRC could be safely used for First-in-Human use - especially if a pharmacologist is added to the required membership. If the FDA decides not to allow RDRC to perform First-In-Human use, then some of this phrasing might be incorporated in the Exploratory IND draft guidance. ************************************ Robert B. Innis, MD, PhD Chief, Molecular Imaging Branch NIMH; Bldg. 1 Rm. B3-10 1 Center Drive MSC 0135 Bethesda, MD 20892-0135 Tel: 301-594-1368 Secretary: 301-594-1089; alzonae@irp.nimh.nih.gov Fax: 301-480-3610 Email: robert.innis@nih.gov http://intramural.nimh.nih.gov/mood/proginfo/mib/ ************************************