Photo by Todd McNaught
To parents who visit a pediatrician seeking routine immunizations for their child, the payoff is well worth the pain: A quick needle stick in the arm protects against the itchy bumps, fevers and swollen glands of common childhood diseases.
Experts say the outcome from the first promising vaccines developed against AIDS will be different.
Preventing infection may not be the way initial AIDS vaccinations will work. What's more, the nature of disease progression in HIV-infected individuals is strikingly different from many diseases for which successful vaccines exist.
Those complications require scientists to evaluate candidate HIV vaccines in novel ways, said Dr. Peter Gilbert, a biostatistician in the center's Statistical Center for HIV/AIDS Research and Prevention and an investigator in the Public Health Sciences Division.
"Whether or not a vaccinated person develops clinically significant infection historically has been the main endpoint for determining whether a vaccine is effective and whether it should be licensed," Gilbert said.
"For HIV, infection itself is considered clinically significant, so indeed HIV infection would be an endpoint that could be used for licensure. The distinction is that with past diseases, serious illness or mortality is observable very soon after infection. But for HIV, if serious disease or death is what we are measuring, then one must wait five to 10 years after infection to see the endpoint, and because of the drug treatment that many who become infected will initiate, the interval may be 10 to 20 years or even longer."
Partially protective vaccines could have a significant impact on the AIDS epidemic-which thus far has claimed 29 million lives-by keeping viral levels low enough in infected individuals to slow HIV progression to AIDS and death and to prevent HIV transmission to others. Yet to merit licensure of such vaccines by the U.S. Food and Drug Administration (FDA), researchers will need to demonstrate that measuring virus levels is an appropriate surrogate indicator for true clinical endpoints such as the onset of AIDS or death.
Criteria for licensure
The FDA has never licensed vaccines based primarily on surrogate markers, so experts worry that it may take a very long time to license and distribute partially effective vaccines to those in need.
In a series of papers published this summer, Gilbert and colleagues outline how to evaluate such vaccines in order to demonstrate their effectiveness for licensure. Despite the possible risks of licensing vaccines based on trials that do not assess traditional endpoints-prevention of death or disease symptoms-the authors argue that the potential cost to human life will be greater if new vaccines are slow or unlikely to hit the market.
Using mathematical models, the researchers evaluated whether outcomes such as a delay in time until the initiation of drug therapy or the suppression of the level of virus in the blood can serve as useful surrogate endpoints for assessing the long-term effects of the vaccines. They conclude that vaccines that demonstrate strong and moderately durable effects on these endpoints merit consideration for licensure.
The studies appear in the July 15 issue of the Journal of Infectious Diseases, the July 30 issue of Statistic in Medicine and the September issue of Biometrics. Collaborators include Dr. Michael Hudgens, staff scientist in SCHARP; Dr. Steven Self, director of SCHARP and a PHS investigator; Dr. Larry Corey, principal investigator of the HIV Vaccine Trials Network and investigator in the Clinical Research Division; and colleagues at the Harvard School of Public Health and Columbia University.
Gilbert noted three broad areas that pose challenges for HIV-vaccine development and that might necessitate modification of the current requirements for product licensing. The first, he said, is the variable nature of the virus and its effect on the immune system.
"Multiple strains of HIV are in circulation, and the parts of the virus that are shared among all or most strains are buried deep within the virus, making them poor targets for an immune response," he said.
Because researchers have not yet been able to create a vaccine that elicits broad, neutralizing antibodies to live circulating HIV strains, he said, the field has shifted toward the development of vaccines that stimulate a protective response from immune-system cells known as cytotoxic T cells (as opposed to another class of T cells called CD4 cells, which are infected by HIV and whose levels drop dangerously as infection worsens).
"These are post-infection vaccines that are designed to attack HIV after the virus has infected CD4 cells, so we must analyze whether or how the disease progresses in vaccinated individuals who acquire HIV," Gilbert said.
A second complication is the lengthy period between viral infection and onset of disease symptoms or death.
"If HIV vaccines are to be licensed based on historical criteria-reduction in death or disease-we'll have a very difficult time designing trials," Gilbert said. "You would have to follow people until a substantial fraction develop AIDS or die from the disease, and this can take a long time. Another strategy is to enroll sexual partners in which one individual is infected and the other is not and monitor viral transmission, but this is hard to do."
Ethical considerations also have hampered the interpretation of clinical trials because vaccine recipients who acquire HIV cannot be denied drug therapy. Those who participate in trials sponsored by the Fred Hutchinson-based HIV Vaccine Trials Network may request to be treated with antiviral agents if their CD4 cell levels drop below a certain threshold.
"We're hoping to measure the vaccine's effect on viral load (level of virus in the blood)," said Hudgens, a lead author of one of the studies. "If a participant goes on drug therapy, that will also reduce their viral load and complicate the ability to gauge the effect of the vaccine by itself."
As a possible means of addressing these complications, the researchers used mathematical modeling based on published clinical data to determine whether measurement of the time until the initiation of drug therapy or until CD4 levels drop to a certain threshold could be used to infer clinically significant vaccine effects. They conclude that a vaccine that delays therapy, delays CD4 progression and reduces viral levels in the blood in a moderately durable fashion should be considered for licensure.
Gilbert said that adopting such criteria potentially could enable vaccine candidates to be tested faster and with less expense than larger, long-term trials.
But he also said that even the most promising vaccine candidates as measured by these criteria would need long-term follow-up.
"A major issue is whether viral strains will emerge over time that are vaccine resistant," Gilbert said. "In primate studies, resistance has popped up years after vaccination."
He and others acknowledge that researchers have no way of knowing whether vaccines will affect the long-term evolution of the virus, possibly causing it to become even more dangerous by selection for strains that elude the vaccine's protective effect.
"When the first vaccine is developed that has a good effect on suppressing viral load, we'll be holding our breath to see how long it will it last," he said. "Ultimately, we may need multiple interventions-behavioral education, preventive vaccines, therapeutic vaccines and antiviral drugs-to control the epidemic."