A Clinical Trials Primer
Clinical drug trials of any type are not easy to run. Extra care must be taken in community-based trials to guarantee that patients' needs are met, which is, after all, at least half the reason for community trials in the first place. The following is a primer of some of the more pressing ethical and technical issues involved in trials of AIDS drugs.
Controls
Controls are as basic to clinical research as drugs are to medicine. They are usually necessary for collecting accurate data, because they give investigators something with which to compare effects of a given substance. But are these controls always essential when the patients are facing life and death situations? Is it possible to investigate a new drug, without denying many of the test patients access to that drug? Most AIDS activists believe so, and have stepped up debate over several typs of controls.
PLACEBOS: The notorious sugar pill everyone learns about in biology class. Because they are inert, placebos are ideal in many clinical trials: half the subjects are given a drug, the other half receive placebo. These two groups are chosen at random and patients are seldom told which substance they are taking (a process known as "blinding"). But how ethical is it to let one group receive treatment, and live, while others receive a placebo, and die? A 1986 trial of the antiviral AZT gave 20 patients the drug and 20 patients placebos. Six months later, 19 people on placebo had died, while only two on AZT died.
In life or death cases, placebos are immoral, and unnecessary, according to AIDS activists. When CRI and CCC successfully tested aerosolized pentamidine to fight PCP, a pneumonia associated with AIDS, no controls were used. Everyone took the drug, and most people benefitted. There was no need — nor time — to test one group against another.
One other problem with placebos is that patients can find out if they received the drug or the inert substance by having their medication analyzed in a lab. No one knows exactly how many patients manage to find out, but doctors estimated that 10–20 percent in any experiment somehow discover the truth.
ACTIVE CONTROLS: Currently the preferred method among most AIDS activists, including ACT UP, active controls give group A one type of drug, and group B another type. No placebos are used and everyone receives some form of potential therapy. These controls are nearly as effective as placebos in testing the efficacy of a substance. Most active-control experiments today use AZT and ddI (dideoxyinosine), a promising anti-HIV drug.
CROSSOVER CONTROLS: Again, no placebo is used. In this case, group A would switch from, say, three months on AZT to three months on ddI. Information is then obtained by comparing patient reactions to the two drugs.
WAITING LIST CONTROLS: In an experiment testing a popular drug under great demand, doctors can enroll their patients on waiting lists, and compare their data with those who are receiving the drug. Although this is less preferable to active controls, it is often more humane than using placebos, because everyone will receive treatment as soon as possible.
End Points
How long should an experiment last? If there is a clear and proven benefit (or adverse effect) on patients, how much clinical evidence is needed before taking the next step toward approval? Again, the ’86 AZT case is illustrative. Dr. Donald Armstrong of Sloan-Kettering Cancer Center said doctors should look for an "early warning of significance" during trials. "Yes, we have to show efficacy. But did we have to go the full six months to know that 19 patients on placebo would die?" Armstrong asked.
The design of any protocol, then, must include trial "end points." But when can doctors know if an experiment has succeeded or failed? Each case is different, and trial directors must closely watch for these "early warnings," especially if placebos are involved.
Parallel Tracks
Another concept promoted by ACT UP and others — and gaining favor with the NIH and the FDA — is the so-called "parallel track" system. Under this system, drugs that have been proven safe in clinical trials wolud be made available to AIDS, ARC and HIV-positive patients, even while the drugs are still being tested for effectiveness. Thus, parallel track patients could "provide a wealth of information on how a drug worked in the real world," according to a statement released by ACT UP's Treatment and Data Committee.
The system is ideal for many patients who do not fit into any of the rigorous protocols used in clinical trials. Often, people are excluded from trials because they are taking another type of drug. With the parallel track they would not be excluded. This list of advantages is a long one. But using the parallel track would clearly benefit thousands of people while providing AIDS researchers with vital data on promising new drugs.
Avoiding Tragedies
In early July, a man with AIDS died while he was enrolled in an underground study of the antiviral GLQ223, better known as "Compound Q." Federal officials are investigating whether the drug, made from Chinese cucumber root, led to the man's death. Drug safety is essential, and must take precedence over even the most desperate demands of those who would try anything at all.
"We must be very careful. Even though the [compound Q] case is not surprising," said Dr. Armstrong. "PWAs often had to empower themselves every step of the way ... right up to testing drugs themselves.
"But there must be strict guidelines, no coercion and totally informed consent in every protocol," he said.
— Sidebar in OutWeek Magazine No. 5, July 24, 1989, p. 9.