SURROGATE ENDPOINTS FOR AIDS DRUGS: LONG-TERM STUDIES NEEDED

SURROGATE ENDPOINTS FOR AIDS DRUGS: LONG-TERM STUDIES NEEDED to determine association between surrogate markers and clinical outcome until the mechanism of action of a drug is understood, Thomas Fleming, University of Washington, maintained at a recent Institute of Medicine meeting. "It's entirely possible that ]a drug[ could have a positive effect on T4 ]cell count[, which we know is associated with an outcome," Fleming stated, "but there could be other mechanisms of action that would give us overall detrimental effect on the outcome of interest." Fleming addressed his comments to panelists at IoM's Sept. 11-12 meeting on the use of surrogate markers to evaluate the effectiveness of AIDS drugs. The key issue is whether surrogate markers are correlated with a definitive clinical outcome," Fleming maintained. "What we need to establish is that when we have a positive benefit on the surrogate we also have a positive benefit on the clinical marker, or vice versa." To illustrate his point, Fleming noted the findings from the Cardiac Arrhythmia Suppression Trial (CAST). The trial was designed to determine if anti-arrhythmic drugs reduce sudden cardiac death in post-myocardial infarction patients with non-life-threatening arrhythmias. The study found that two drugs, encainide and flecainide, may actually increase the risk of death or non-fatal cardiac arrests. "Clearly ]encainide and flecainide[ do have the desired effect on the potential surrogate, which is arrhythmia suppression," Fleming stated. "But a clinical study showed that they were actually detrimental on survival, and in that setting arrhythmia suppression is not a surrogate for death." National Institute of Allergy and Infectious Diseases Director Anthony Fauci, MD, emphasized that there are currently no markers in HIV infection that correlate with survival. However, positive CD4 cell count emerged as the best indicator of drug efficacy. "The one thing we absolutely know for certain is that the ]clinical[ event rate is very clearly related to the level of CD4," Fauci said. "And the other thing we know is that if an event ]leads to[ AIDS . . . if you don't have AIDS you have a better ]chance of[ survival than if you do have AIDS." Furthermore, Fauci continued, "if you don't get an event based on the level of your CD4 cells, common sense tells you that if you delay the drop in CD4, that will prevent you from getting the event, ]and[ you are going to have an increase in survival." The two surrogate markers most frequently mentioned at the IoM meeting were CD4 cell counts and p24 antigen levels. Other potential surrogate markers mentioned include viral cultures, HIV antigenemia, serum tumor necrosis factor levels, lymphocyte count, and such clinical markers as weight, fine motor skills, and mental/cognitive function. Stephen Lagakos, PhD, Harvard University School of Public Health and the Dana-Farber Cancer Institute, outlined three potential uses for surrogate markers. He said that in Phase I and Phase II trials they could be used to "screen" new drugs for subsequent use in large efficacy trials or to determine "optimal doses/schedules for Phase III trials." In Phase III trials, surrogate endpoints could suplement more definitive clinical endpoints. Surrogate markers also could be used in the clinical management of patients, such as indicators of when and how to change treatments, Lagakos said. Summarizing the panel's discussion on surrogate endpoints, FDA Antiviral Division Director Ellen Cooper said that "there was basic agreement that at this time there is no single surrogate marker that is sufficiently well validated to substitute by itself for the more definitive clinical endpoints in controlled efficacy trials." Although "CD4 seems to be the furthest along in this regard," Cooper questioned "whether CD4 counts have the same prognostic value in treated patients as they do untreated patients."