Weighing the burden of disease

How will diseases other than HIV affect AIDS vaccine trial design?

By Emily Bass, with reporting by Mark Boaz, Ph.D.*

For the better part of the 20th century, vaccine development and testing was the province of the industrialized world. Many of today’s vaccines, including those against polio and measles, were licensed based on data from efficacy trials in the United States and Europe. More recent vaccines, such as those targeting pneumococcal infections and hepatitis B virus, were evaluated in efficacy trials in developing countries including South Africa and Thailand. But there is still little precedent for the AIDS vaccine endeavor, which is focused on developing countries in sub-Saharan Africa, Asia and Latin America at every stage of vaccine testing—from small safety studies to large-scale efficacy trials. Throughout these regions, HIV is intertwined with other endemic infections, often called “diseases of poverty,” which include helminthic infections and intestinal parasites, malaria, tuberculosis and sexually transmitted diseases. These diseases add a layer of complexity to the already daunting task of finding an effective AIDS vaccine.

Licensed vaccines have a long history of being delivered in settings where healthcare is rudimentary. One dramatic example is the “Days of Tranquility” campaign in which El Salvador, Afghanistan and the Democratic Republic of Congo called ceasefires allowing the tens of millions of children to receive polio vaccine. However there is a world of difference between administering a licensed vaccine and testing one against the background of poverty and endemic, untreated disease. This is one reason why many clinical trials, which require a rigorously controlled environment, have been conducted in countries with high standards of healthcare and sanitation, and low levels of endemic infections.

But limiting studies to the industrialized world is not an option for AIDS vaccine trials. Although HIV is increasing dramatically in some US and European communities, the most severe epidemics are in the poorest countries of the world. Vaccine trials in these countries will provide crucial information on how HIV genetic diversity impacts on AIDS vaccine efficacy, and on the acceptance of AIDS vaccines in different populations. In order to conduct efficacy studies that meet the stringent standards of regulatory agencies like the US Food and Drug Administration, AIDS vaccine trial sponsors working in resource-poor settings must supply or strengthen a range of healthcare services, both as a benefit to the volunteers and to ensure that data on adverse events and efficacy can be generalized beyond the study community. “You can’t get good data without providing good care,” says IAVI Medical Affairs Director Pat Fast.

HIV-related services such as voluntary counseling and testing, prevention interventions, and treatment and care for HIV-infected people top the list of trial sponsor priorities—and the issue of whether or not to include ARVs (antiretrovirals) in trial-related healthcare has only recently been settled by several vaccine trial sponsors. But today, AIDS vaccine developers are paying increasing attention to diseases other than HIV. In doing so they are paying heed to warnings that have been sounded by researchers in other fields, notably parasitologists, who have warned that coinfection with common diseases of poverty could complicate analysis of trial data. This is a serious consideration, particularly for the AIDS vaccine candidates that will be evaluated solely based on efficacy trials in developing countries. Israeli parasitologist Zwi Bentwich (Rosetta Genomics, Tel Aviv) writes that, “Potentially good vaccines may fail in clinical trials if examined in the immune scenario presently existent in the developing world.”

Variable vaccine effects

One of the most powerful arguments for considering the interaction between AIDS vaccines and diseases of poverty comes from studies of other vaccines, including rotavirus, BCG, polio, and oral cholera, which have shown variable rates of immunogenicity in the developing, versus developed, world. For instance, a single dose of live oral cholera vaccine (CVD 103-HgR) induced vibriocidal antibody in 90% of vaccinees in the industrialized world but in only 16% of Indonesian children. In many instances, this effect can be overcome by increasing the dosage of the vaccine—but the underlying mechanism is not clear.

One hypothesis suggests that the variations in vaccine immunogenicity are caused by helminthic infections, which infect 1.5 billion people—one quarter of the world’s population—and are most common in the developing world.

The variety of small prospective studies that have tested the effects of helminth infection on immune responses to vaccines do show lower or less durable immune responses in people with intestinal parasites (see box). A small trial that examined responses to Hepatitis B vaccine in Egyptian children matched for age and other demographic characteristics found similar plasma antibody titers in children regardless of their Schistosoma mansoni status. However, by nine months post-vaccination the parasite-free control group had a significantly higher percentage of responders (97% versus 56%) and higher levels of antibody titers. Another trial randomized 60 helminth-infected Ethiopian adults to receive deworming therapy (albendazole) or a placebo prior to immunization with BCG vaccine, and found that the albendazole-treated participants had stronger cellular responses to tuberculin than placebo recipients. (as measured by stimulation index or IFN-g production).

The precise mechanism of the proposed helminth effect is not known, but one theory points to the fact that chronic helminthic infection leads to persistent activation of Th-2 type immune responses, which are broadly characterized as anti-inflammatory responses. This bias may hamper the ability to mount a robust vaccine-induced Th-1 type response. Data from human trials and mouse model experiments show that helminth infections may also cause anergy and hyporesponsiveness in immune cells.

Several studies of differential vaccine effects in developing and developed countries concern orally administered vaccines, such as cholera and polio, which are thought to work by inducing protective responses at mucosal surfaces such as the gut and lungs. This has led researchers to speculate that intestinal parasites could interfere with interactions between vaccine and antigen-presenting cells in the intestinal mucosa. However, other studies have shown an effect on parenteral vaccines such as BCG and tetanus toxoid. In one Kenyan study, schistosomiasis or filarial parasitic infection in pregnant women was shown to correlate with lowered responses to BCG vaccine among newborns.

These data are suggestive but by no means conclusive. Not all parasites have been linked to Th-2 type bias, and it appears that different types of worms have varying effects on the immune system. It’s difficult to forecast the implications for AIDS vaccines, since they could be tested in populations infected by or exposed to multiple parasites. To complicate matters further, in the absence of reliable correlates of protection it is impossible to know if changes in immunogenicity—such as those potentially caused by helminthic infections—will have a bearing on vaccine efficacy data.

Some researchers suggest that AIDS vaccine trials gather parasite epidemiology data during trials and keep the potential effects of parasites in mind. “I think it would be worth knowing people's helminth status in vaccine trials and, if the hypothesis is confirmed, doing some work on when is the best time to deworm people,” says Allison Elliott (Wellcome Trust), who has conducted several studies on the interaction between helminthic infection and HIV progression in HIV-infected Ugandans. Elliott explains that deworming immediately prior to immunization also has immune consequences—including sudden increases in anti-worm responses which might also impact on vaccine efficacy.

Zwi Bentwich agrees, saying that, “It is imperative to evaluate and follow the immune profile in vaccinees before and following vaccination, because of the potential interference of background immunity with the ability to generate specific responses to the vaccines.”

Taking stock of multiple factors

Although helminthic infections may be a contributing factor, it is unlikely that all differential vaccine effects can be attributed to a single facet of life in resource poor settings. Poor nutrition, chronic or recurring infection with other diseases like malaria, and host genetic background all play a role in immune profiles on an individual and community level. Sexually transmitted infections other than HIV could also compromise the efficacy of AIDS vaccines. Since these infections compromise the integrity of the genital mucosa, a vaccine which affords protection in the context of an intact mucosal membrane might be less effective if these physical barriers have already been compromised by other infections.

It’s impossible to predict which, if any, of these factors will affect AIDS vaccine efficacy, either in preventing HIV infection or in protecting against HIV-related disease in trial participants who become infected with HIV through high-risk behavior. But in order to identify potential confounding variables, sponsors need a clear picture of the common diseases and immune profiles of volunteers and potential recipients of a licensed vaccine.

Perhaps surprisingly, this information is hard to come by. One of the paradoxes of disease in the developing world is that, while it is omnipresent, it is also—from an epidemiological standpoint—ill-documented. Although some epidemiological studies have been done in potential AIDS vaccine trial sites, there are also sites where there is relatively little precise data on prevalence, incidence, recurrence, or immune effects of various coinfections.

Today many sites are undertaking studies designed to fill in these gaps. At the Soweto Vaccine Evaluation Unit at the Chris Hani Baragwanath Hospital, Guy De Bruyn of the HVTN and principal investigator Glenda Gray conducted a prevalence study of helminthic infections in over 100 potential AIDS vaccine trial volunteers. “There is very little recent prevalence data on helminthic infections among adults in the places in South Africa where trials will take place,” De Bruyn explains. “This is an initial look to see if we can even answer a question about helminth-vaccine interactions in this population.”

By mid-2004, De Bruyn and Gray will have preliminary prevalence data. If the prevalence is high, De Bruyn says the next step will be to screen people who are participating in trials to determine whether immune responses differ in people who are dewormed prior to vaccination compared to people who are parasite-free at the time of vaccination.

This paucity of data is also a stumbling block for evaluating vaccine efficacy in protecting against HIV-related disease. AIDS vaccine trial designers are now attempting to define a “composite endpoint” that could be used to evaluate vaccine impact on the course of HIV infection. This will be measured by changes in CD4+ cell count and viral load in vaccinees and placebo recipients. It may also be measured in terms of clinical conditions including AIDS-related illnesses. Although there is considerable data on AIDS defining illnesses in the industrialized world, much less is known about HIV-related illnesses in resource-poor settings, particularly in the early stages of infection.

“Actually, and incredibly, the data we need to answer these types of questions are extremely few. Good information on early clinical course of HIV infection in the developing world is one of the major gaps in our knowledge,” says HVTN statistician Steve Self.

Trial design considerations

How should information about the rates of various infections influence trial design? Coinfections raise different issues at each stage of clinical testing. In Phase I studies participants are rigorously screened prior to enrollment and asymptomatic conditions such as eosinophilia (which is associated with helminth infections) or anemia may be used as exclusion criteria. During the trial it is crucial to make precise rather than presumptive diagnoses in order to definitively rule out a vaccine-related side effect. So, for example, a feverish Phase I trial participant from a malaria-endemic area would undergo a confirmatory blood test before receiving treatment that might ordinarily be prescribed based on clinical symptoms.

In Phase II and III studies there are different issues. Participants in these trials will be clinically healthy, but will not be tested for asymptomatic conditions; there is also less intensive medical surveillance of trial participants. Jimmy Whitworth (London School of Hygiene and Tropical Medicine) thinks that this is the point at which trials should begin to gather data about interactions between co-infections and vaccine effects. Data on disease effects are “probably best accrued in later-stage Phase II trials,” he says, and suggests that substudies within Phase II trials could compare vaccine-induced immune responses in individuals with and without background infections. While small, these substudies would be able to detect a major effect on immune responses and could provide some guidance about inclusion and exclusion criteria for Phase III or expanded Phase II studies. While randomization safeguards against bias due to background infections in either trial arm, it will not protect against misinterpretations of data that could arise if a background infection skews immune responses to the vaccine itself.

“If coinfections have only minor effects on immune responses then I feel we should not screen them out from a Phase III or expanded Phase II study because of issues of generalizability,” Whitworth says, referring to the fact that excessively stringent exclusion criteria could limit the relevance of trial data to other populations with different disease burdens. “However if coinfections have major immune effects I think, though I am reluctant to say so, that we would have to screen them out at enrollment and possibly throughout the trial. Otherwise we may miss important vaccine effects.”

Practically speaking, it would be nearly impossible to incorporate treatment for all potentially confounding asymptomatic infections into large-scale trials. Taking helminthic infections and intestinal parasites alone, it would be a mammoth undertaking to collect and analyze stool samples and administer treatment to all infected participants in a large-scale efficacy trial. Different intestinal parasites respond to different medications; some can be eliminated with a single dose while others require multiple treatments to eliminate, and reinfection is common. This complexity precludes administering universal deworming treatments to all volunteers.

The challenges do not end with efficacy trials. Phase III studies still represent a controlled environment compared to the “real world” in which a vaccine will be used. “In the context of a clinical trial, conditions like ulcerative STDs that could change vaccine efficacy will be checked for and treated—at higher than average standard of care. But this may not be how people who eventually get the licensed vaccine will get care,” says Nzeera Ketter, head of efficacy trials at IAVI.

Once a vaccine has shown efficacy in a large-scale trial, vaccine developers will devise follow-up “bridging studies” to learn more about vaccine effects in populations with more complex health issues; bridging studies will also involve adolescents, who have different disease profiles and immune parameters than adults.

For now, scientists say that it is crucial to learn more about the potential confounding effects of coinfections. “It’ somewhat black box immunology,” says De Bruyn. “It’s very unclear what the immunological mechanism of helminth interference with vaccines might be, and whether this effect will be seen with AIDS vaccines. Then again, we're still uncertain about what AIDS vaccine effects will be protective against infection.”

 Parasitic Infections and Vaccines: A Bibliography

Bassily S, et al. Immunogenicity of recombinant hepatitis B vaccine among infants of mothers with active schistosomiasis. Am J Trop Med Hyg. 1997 Aug;57(2):197-9.
Study population: 385 Egyptian infants born to HB-negative mothers.
Key findings: Maternal schistosomiasis (n=191) had no effect on levels of anti-Hepatitis B antibodies in infants at 9 months following HB immunization at 2,4,6 mo.

Cooper, et al. Albendazole treatment of children with ascariasis enhances the vibriocidal antibody response to the live attenuated oral cholera vaccine CVD 103-HgR. J Infect Dis. 2000 Oct;182(4):1199-206. Epub 2000 Sep 08.
Study population: Ecuador (age range 6 – 13 years), n = 233.
Key findings: Significantly greater rates of seroconversion and geometric mean antinbody titer in the albendazole group (in subjects with non-O ABO blood groups) as compared to placebo group. (Loss to follow up exceeded 50% in both groups, complicating the analysis of this data).

Cooper, et al. Human onchocerciasis and tetanus vaccination: impact on the postvaccination antitetanus antibody response. Infect Immun. 1999 Nov;67(11):5951-7.
Study population: Ecuador (age range 5 – 80 years), n=193.
Key Finding: Concurrent infection with O. volvulus did not prevent the development of a protective antitetanus response, although heavier O. volvulus infections were associated with lower increases in TT-specific IgG levels.

Elias D, et al. Effect of deworming on human T cell responses to mycobacterial antigens in helminth-exposed individuals before and after bacille Calmette-Guerin (BCG) vaccination. Clin Exp Immunol. 2001 Feb;123(2):219-25.
Study population: Ethiopians (18-24 years old), n = 240.
Key findings: BCG immunization resulted in significantly improved PPD responses (measured by T cell proliferation and interferon gamma production) in participants who were de-wormed prior to vaccination but not in the control group.

Ghaffar YA et al. Hepatitis B vaccination in children infected with Schistosoma mansoni. Am J Trop Med Hyg 1990; 43(5): 516 – 9.
Study population: Egyptian male children (age range 8 – 12 years), n = 80
Key findings: Three months after vaccination, no difference in responders versus non-responders in otherwise matched children w/ and w/o Schistosoma-positive stool samples. At 9 months, control group had significantly higher percentage of responders, and higher levels of HBs antibody.

Kilian HD, Nielsen G. Cell-mediated and humoral immune responses to BCG and rubella vaccinations and to recall antigens in onchocerciasis patients. Trop Med Parasitol. 1989 Dec;40(4):445-53.
Study population: Liberian children, n = 629.
Key Findings: Conversion rate three months after BCG vaccination was significantly lower in children with onchocerciasis (48%) than in controls (85%).

Kilian HD, Nielsen G. Cell-mediated and humoral immune response to tetanus vaccinations in onchocerciasis patients. Trop Med Parasitol. 1989 Sep;40(3):285-91.
Study population: Liberians (2-29 years old), n = 326.
Key Findings: Diminished cell-mediated immune responses to tetanus toxoid immunization in matched vaccinees with and without onchocerciasis infection; no alteration in humoral responses between two groups

Malhotra, et al. Helminth- and Bacillus Calmette-Guerin-induced immunity in children sensitized in utero to filariasis and schistosomiasis. J Immunol. 1999 Jun 1;162(11):6843-8.
Study population: Kenya, n=25 infants.
Key findings: Mean level PPD-driven IFN- at 10–14 mo of age was greater in BCG-immunized infants who were not sensitized to helminthic infection in utero (p < 0.01).

Thanks to Guy de Bruyn, MBBCh and Allison Elliott, PhD for their assistance in compiling this bibliography.

*Emily Bass is Senior Writer of the IAVI Report; Mark Boaz, Ph.D., is Immunology Laboratory Manager at IAVI