Cover Image
close this bookWHO-UNAIDS Consultation on Issues Related to Access to Future HIV Vaccines (UNAIDS, 2000, 30 p.)
View the document(introduction...)
View the documentSummary
View the document1. Introduction
View the document2. Framing the issues
View the document3. Different scenarios for future vaccine use
View the document4. Other logistical issues
View the document5. Conclusions and recommendations from the Working Groups
View the document6. The way forward
View the documentSelected references
View the documentAppendix

2. Framing the issues

This session, chaired by Bhamarapravati Natth, provided a review of the major issues to be considered in planning use of future HIV vaccines. Formal presentations were made by Drs. Esparza, Longini, Goldenthal, Clemens, Nkowane, Mastro, Bishai and Stanton.

2.1. HIV vaccines in the pipeline and timelines for decisions (Esparza) [1].

The first phase I trial of an HIV candidate vaccine was conducted in the United States in 1987. Since then, more than 60 phase I/II trials have been conducted, with a total of approximately 30 different HIV candidate vaccines. Most of these trials have been conducted in the US and Europe, but since 1993 trials have also been conducted in developing countries (Brazil, China, Cuba, Thailand and Uganda). At the present time 19 preventive HIV candidate vaccines are at different levels of clinical evaluation in the US, including recombinant proteins, synthetic peptides, nucleic acid vaccines and different recombinant live vectors.

The only HIV candidate vaccine that has entered phase III efficacy evaluation is a gp120 product from VaxGen (Brisbane, California, USA). Two different versions of that product are being tested in the United States and in Thailand. A bivalent candidate vaccine (based on two subtype B strains) entered a phase III trial in the US in June 1998, involving 5.500 volunteers, mostly men-who-have-sex-with-men (MSM). A bivalent BE gp120 candidate vaccine entered phase III evaluation in Thailand in March 1999, involving 2.500 recovering intravenous drug users (IDU) in Bangkok. The interim efficacy analysis of the US and Thai trials will take place in November 2001 and August 2002, respectively, with final results becoming available one year thereafter.

There are plans to initiate a second phase III trial in the United States and in several countries in the Caribbean and South America, using a prime-boost strategy including two different subtype B products: a canarypox-HIV recombinant vector (Aventis Pasteur) followed by gp120 (VaxGen). This phase III trial could start sometime in 2002, with efficacy results becoming available 3-4 years later. Another prime-boost phase III trial is being discussed for implementation in Thailand, using a subtype E canarypox-HIV recombinant vector followed by a gp120 BE boost.

In summary, the earliest that an HIV vaccine could become available would be late 2002, or in 2003, depending on the results of the ongoing VaxGen trials.

2.2. Defining HIV vaccine efficacy from phase III trials (Longini) [2-4].

HIV vaccines could have at least three important protective effects:

· They could reduce the susceptibility to infection in vaccinated people, i.e., vaccine efficacy for susceptibility (VES);

· They could reduce the rate of infection and/or disease progression in vaccinated people who get infected, i.e., vaccine efficacy for infection and/or disease progression (VEP); and

· They could reduce the level of infectiousness of infected vaccinated people, i.e., vaccine efficacy for infectiousness (VEI).

Vaccine trials can be designed to measure all or some of these effects through a number of primary and secondary endpoints. The primary endpoint is usually how well the vaccine protects against HIV infection, comparing the infection rate in the vaccinated versus the unvaccinated.

A number of secondary endpoints are of interest. A good measure of progression of the infection could be the level of HIV-RNA in plasma. In this case, measures such as virus loads in the vaccinated and unvaccinated infected participants are compared after the set point has been established, four to six months following infection. Differences in virus loads between the two groups could suggest a protective effect of the vaccine. Another measure of the VEP could be to compare the percentage of infected unvaccinated and vaccinated participants with a predefined low virus load level after the set point.

The VEI measures another important secondary endpoint. For sexual transmission this can be measured through the augmented sexual partners design. Additionally, virus load measurements can indirectly indicate decreased transmission potential, with levels below 1500 RNA copies per ml associated with very low transmissibility.

All of the above mentioned VE measurements could be stratified by the circulating HIV subtypes, although it is recognized that the significance of HIV genetic subtypes in terms of potential vaccine efficacy is not known, and it could be different with candidate vaccines based on different vaccine concepts.

Further secondary endpoints could involve finding potential immune correlates of protection, such as antibody levels and CTL function following vaccination.

2.3. Regulatory considerations in relation to Phase III HIV vaccine trials (Goldenthal) [5-7].

Prior to the initiation of a phase III efficacy trial, the US Food and Drug Administration (FDA) would expect to review information pertaining to the following areas: i) recent epidemiological data (e.g., recent seroincidence, endemic subtypes) from the intended trial population; ii) data supporting the safety and immunogenicity of the product (including the basis for selecting the proposed formulation, dose and schedule), and; iii) the "scientific rationale" for conducting the trial. The "scientific rationale" for a phase III trial includes human immunogenicity data from phase I/II trials. Animal challenge/protection data may also play a prominent role. Issues related to the vaccine product including stability are also important. Because of the potential differences in safety, immunogenicity (and potential efficacy) between populations, safety and immunogenicity data should also be obtained using the candidate vaccine in the specific population in which the efficacy trial will be performed.

Appropriate laboratory assays should be available to detect vaccine-elicited immune responses, and to identify and characterize HIV strains from infections occurring in the trial population. The validation of these laboratory assays should include relevant data, e.g., on specificity, sensitivity, ruggedness, and reproducibility.

The efficacy trial protocol should describe the inclusion/exclusion criteria for the study population, the control group, the randomization schema and study masking, and the parameters (safety, immunogenicity, efficacy) to be monitored with the time schedule. The vaccine efficacy trial protocol must include information regarding surveillance plans and length of follow-up. Surveillance for efficacy should be performed from the time of randomization. Prototype case report forms, subject diaries and consent forms should be submitted along with the protocol. Information should also be provided regarding logistics (such as specimen collection and shipping).

Possible outcomes that might be observed in an HIV vaccine phase III efficacy trial can be summarized as follows: i) prevention of infection; ii) prevention of chronic infection (transient infection); iii) occurrence of infection, but AIDS is prevented or delayed (assessed by candidate surrogate markers such as virus loads, or by clinical findings such as AIDS or mortality); iv) occurrence of infection but vaccinee is less infectious; and v) combinations of above. The primary endpoint for both ongoing VaxGen efficacy trials is prevention of infection.

The statistical section of the protocol should include prospective and detailed information, especially for the primary endpoint, and 95% confidence limits of efficacy estimates. Both intent-to-treat and "per protocol" estimates of vaccine efficacy using the primary endpoint are of interest. Plans for any interim analysis must be described.

At least for the initial HIV vaccine efficacy trials, FDA would convene an advisory committee meeting to review and comment on protocol design and relevant data. Both FDA staff and sponsors would present information and issues at such meetings.

An important outcome of efficacy trials is the possibility of identifying immune correlates of protection, defined as particular type and quantity of immune response(s) associated with protection from infection or disease. Immune correlate(s) of protection could be useful for interpreting future trials with immune response endpoints, such as bridging studies. However, identification of correlates is not a requirement for US licensure. Examples of vaccines licensed without an identified immune correlate of protection include acellular pertussis, typhoid, and tuberculosis (BCG).

Future clinical bridging studies could be needed to i) address concerns that manufacturing changes might have resulted in a "different" vaccine no longer clinically equivalent to the previous version used in the efficacy trial; ii) provide evidence that efficacy data can be extrapolated to different populations; and iii) support new dosing schedules.

Foreign efficacy trials have been used to support licensure in the United States, including vaccines against typhoid fever, Japanese encephalitis, pertussis, and hepatitis A. However, in this situation, bridging studies for safety and immunogenicity could be needed, at a minimum, for licensure in the United States.

In conclusion, HIV vaccines present unique considerations for product and clinical development. Overall careful planning is needed to permit timely development. In this regard, important areas include: i) product characterization and manufacturing; ii) anticipating needs of future trials (e.g., developing and validating critical assays); iii) accumulating sufficient safety, immunogenicity and efficacy data during clinical development. The latter includes planning and conducting clinical bridging studies (for example, in relation to use in different populations and product scale-up) needed for approval. Sponsors are encouraged to utilize FDA resources and documents to facilitate these activities.

2.4. Potential bridging and effectiveness trials with HIV vaccines (Clemens)[8].

Bridging studies are conducted to address uncertainties about biological generalizability of vaccine performance. Effectiveness (phase IV) trials are conducted to address uncertainties about practical generalizability of vaccine performance.

After a candidate vaccine has shown efficacy in well-controlled phase III trials, additional clinical evaluation would be needed to address the biological vagaries and novelties that could challenge the generalizability of a vaccine's performance, including: i) changes in manufacture, formulation, dosage and administration of vaccine; and ii) changes in the target population for the vaccine, including changes in the epidemiology of the target infection (route of transmission, intensity of transmission, and antigenic variation).

Vaccine bridging studies may need to be conducted to support approvals for marketing. Examples of bridging studies that could be needed for the first generation of effective HIV vaccines include: different schedules of administration, different routes of transmission, and protection against different strains.

Additional questions that may remain after vaccine licensure include: i) how well will the vaccine work under realistic conditions (expanded spectrum of vaccine recipients, administration of the vaccine under routine conditions, co-administration of other vaccines or drugs, and against outcomes of pragmatic interest to decision-makers)? ii) how well will the vaccine be accepted? iii) how logistically feasible will it be to use the vaccine? iv) how cost-effective will the vaccine be? v) what will be the total impact of the vaccine (direct and indirect effects)?

Effectiveness trials are conducted to answer some of the above questions, especially the impact of the vaccine on practical health outcomes, assessed under ordinary conditions of a public health programme. The research question posed by effectiveness trials is: "What are the practical health outcomes, both beneficial and not beneficial, when the vaccine is administered under the ordinary conditions of a public health programme?". Phase IV observational studies of vaccine effectiveness are performed after licensure and rely upon the comparative occurrence of outcomes in persons who are or are not vaccinated in routine practice, using cohort or case-control study designs.

The conduct of effectiveness trials with future HIV vaccines will be justified only when there is a "decisional equipoise", which will be influenced by numerous factors including political considerations.

2.5. Potential HIV immunization strategies (Longini) [9-11].

It is expected that soon after the efficacy of an HIV vaccine is demonstrated in phase III trials, there will be limited quantities of vaccine available for administration. When a limited supply of vaccine is available, its distribution may involve determining the proportion of the various population groups that should be vaccinated in order to minimise the impact of HIV. The solution to this problem depends on a number of factors including the following:

i) vaccine efficacy;
ii) HIV subtypes circulating;
iii) important "risk groups", including "core" transmitters;
iv) mixing behaviour of "risk groups";
v) quantity of vaccine available;
vi) vaccine acceptance and possible distribution levels;
vii) objectives of HIV control.

For a particular population, the vaccine deployed should be effective against the major immunotypes of HIV circulating in that population. The prime candidates for receiving vaccine would be the important "risk" groups, and "core" transmitters within those risk groups. If the quantity of vaccine is not sufficient to slow transmission substantially in the important risk groups, then it may be best to use these limited quantities of vaccine in the most vulnerable people in the population.

Once the above seven factors have been determined for a particular population, the optimal distribution of a limited quantity of vaccine can be determined that achieves the objective specified in item (vii) above (HIV control). The optimal distribution can be studied by constructing a mathematical model of HIV transmission for the population in question and then minimising the objective function subject to the constraints on vaccine availability and distribution possibilities. This modelling solution provides qualitative guidelines for potential vaccine distribution.

The seven factors listed above will vary for each population under study, and mathematical models will be used to analyse different scenarios.

2.6. Operational issues for HIV immunization delivery systems (Nkowane).

In relation to operational issues for future HIV immunization delivery systems, much can be learned from the experience with other vaccines being delivered through the Expanded Programme on Immunization (EPI).

Any immunization delivery system should take into consideration: i) the characteristics of the vaccine; ii) the target or at risk group or population, and; iii) the programme objectives (individual protection, disease prevention, disease control or disease elimination or eradication).

The EPI is a delivery system primarily targeting infants, and heavily dependent on a functional health system. Routine programmes include primary immunization of infants, children and adolescent, and booster immunization. In addition, regular "catch-up" or targeted mass campaigns are also implemented for routine vaccine delivery, to improve coverage, or for disease prevention and control (in case of epidemics). Existing immunization schemes for different vaccines target infants, pre-school children, school children, adolescents or adults, which are accessed in different places (hospitals, health centres, pre-school organizations, schools, and/or workplaces).

Mass immunization campaigns are time-limited activities done once or twice a year, and they are most effective if targeted, and when objectives are achieved after a limited number of doses of vaccines are deployed. Campaign fatigue is a major problem. Mass campaigns require mobilization of communities and partners, and the most successful campaigns use volunteers. The role of health care workers is often limited to supervision of activities.

Factors which are critical for potential immunization delivery of future HIV vaccines are service level, logistics, and vaccine supply and quality. The level of service will depend on the coverage obtained in the target group, the drop out rate (if more than one dose is required), and the quality of the service (i.e., injection safety issues). Logistics will depend on the proportion of days in which service can be realistically offered, the systems for disposal of wastes and used equipment, and the communication between the various levels of the system. Finally, factors related to vaccine supply and quality are related to the regularity of supplies and equipment, wastage of vaccine, and systems for monitoring adverse events.

Different options can be considered for potential immunization delivery of future HIV vaccines: i) integration into routine EPI programmes; ii) targeting groups outside routine EPI; iii) immunization campaigns, and; iv) combinations of the above. Integration of future HIV vaccines into routine/existing EPI programmes has a number of limitations. They would only reach infants or women and more importantly, would depend on the duration of immunity and the ability to provide booster injections during childhood and early adulthood. The immediate option to deliver an HIV vaccine would be to target immunization programmes to individuals at higher risk. This could be an expensive option, but it should be viewed in the context of health systems development. Private service delivery could be a critical factor for success. Finally, targeting groups through mass immunization campaigns may be necessary when an HIV vaccine becomes available. Targeting will be required for the initial years of the campaign to achieve the immediate objectives (especially if insufficient vaccine is available for general use). Additional resources for operations will be needed for the many years that the campaign may last.

In summary, the nature of future HIV vaccines will determine the critical operational issues for delivery. Currently available immunization delivery systems in many priority countries may not meet the immediate objectives of a new HIV immunization strategy. Targeted approaches outside of the existing delivery systems will be needed in the initial phases of HIV vaccine delivery. There is, however, adequate experience in immunization programmes to develop appropriate delivery systems for HIV vaccines, even in the most difficult settings. Advocacy, political support and long term funding will be critical for delivery of future HIV vaccines to those who need it the most. A sustainable delivery system should be based on infrastructure strengthening rather than in the development of a parallel delivery system.

2.7. Targeting populations for HIV vaccination (Mastro) [12,13].

Possible strategies for the use of future HIV vaccines could include vaccination of everyone at risk of HIV infection, of those at highest risk, of targeted groups, or some combination of the above. The selection and implementation of any strategy will be highly dependent on existing national systems, availability of funds, and a possibility of phased introduction of the vaccine in different populations.

The information needed to set an HIV immunization strategy is: the status of the epidemic and the identification of people at risk. Data sources for such information are AIDS case reports (where the epidemic was), HIV surveillance (where the epidemic is), HIV incidence (where the epidemic is going), behavioral surveillance (where the epidemic might go), sexually transmitted diseases (STD) reports and surveys, and ad hoc research studies. The quality of those data sources, however, varies greatly from country to country.

In the United States in the mid 1990s, the estimated HIV prevalence was about 700.000 persons living with HIV/AIDS, and the estimated incidence was about 41.000 new HIV infections per year. The most affected populations are injecting drug users (IDU), men who have sex with men (MSM), and at risk heterosexuals (ARH) (Table 1).

Table 1. Estimated HIV incidence and prevalence in the United States
(96 metropolitan areas, mid 1990s)

Population at risk

Size of population

Estimated HIV prevalence

Estimated HIV incidence per year


1.5 million

204.000 (14%)

19.000 (1.5/100PY)


1.7 million

314.000 (18%)

9.800 (0.7/100PY)


2.1 million

47.000 (2.3%)

9.300 (0.5/100PY)


5.2 million

565.000 (11%)

38.000 (0.8/100PY)

It must be noted, however, that there are other epidemiological and demographic aspects that will need to be considered when identifying potential target populations for a possible HIV immunization strategy in the United States. In recent years, the number of cases of AIDS has been steadily decreasing among white non-Hispanic individuals, and increasing among black non-Hispanic persons. In addition, the distribution of cases is not uniform among the different states.

In Thailand, the best source of epidemiological data is the HIV sentinel surveillance programme, which is conducted annually in all 76 provinces, including blood donors, pregnant women, female sex workers, drug users, and male STD patients. In addition, biannual random surveys are conducted among 21-year-old military conscripts. HIV prevalence among IDU has been increasing in the country, with median provincial seroprevalence in 1999 of more than 50%. Prevalence among brothel female commercial sex workers (CSW) has declined slightly, with a median provincial seroprevalence in 1999 close to 20%. Median seroprevalence among other female CSW and male STD patients is in the order of 10%, and the prevalence among women attending antenatal clinics is close to 2%. Among the military conscripts, the highest HIV prevalences were recorded from 1990 to 1993 in the northern provinces (6-8%); these rates decreased to less than 2% in 1999, to a level similar to that in other regions.

Epidemiological data from most African countries are often less comprehensive. AIDS case reports are very incomplete and HIV surveillance data vary greatly in quality. However, available data do allow for general characterization of the severity of the epidemic. In one hypothetical "representative" sub-Saharan African country, HIV seroprevalences in different populations are represented as follows: pregnant women (20% urban, 14% rural), STD patients (22% urban), blood donors (18%), factory workers (23% urban), military recruits (15%,) female CSW (56%), tuberculosis patients (60%).

2.8. Modeling the benefits of an AIDS vaccine (Bishai).

Two models of the economic benefits that could be gained through the provision of HIV vaccines to various populations could be considered. One model, called the health sector model, considers only the benefits achieved from preventing the need for medical spending on behalf of vaccine recipients and the people they may secondarily infect. The other model, called the societal model, considers prevented medical spending as well as prevented losses of productive capacity for vaccine recipients and the people they may secondarily infect.

Different assumptions can be made in both models, including different levels of vaccine efficacy. In these models, the non-budgetary (intangible) benefits of HIV vaccines, such as avoided pain, suffering, and grief are likely to be large, but are ignored, not because they are insignificant, but because the current way that typical financial decisions are made ignores intangible costs.

The structure of the models is such that the economic benefits are not necessarily highest where incidence is highest, but where incidence and medical spending and GDP are high. Data on HIV incidence, medical spending, and GDP are used to calculate the benefits of AIDS vaccines in different age groups. Table 2 shows some examples of the health sector benefits in adult men and women in selected developed and less-developed regions of the world. The essential result is that the economic benefits of prevented medical spending per patient vaccinated are highest in developed countries and lowest in less developed countries. This is in opposition to where the greatest epidemiological benefit should be.

Table 2. Health sector perspective estimates among adults in selected regions:
Net expected benefit of vaccination by group and region (US$):




Western Europe



Australia and New Zealand



North America






North Africa and Middle East



Sub-Saharan Africa



South and South East Asia



East Europe and Central Asia









Latin America



From the health sector perspective, the global demand for HIV vaccines would be greatly influenced by price, with a significant potential increase in vaccine use if the price drops below US$ 50 per course.

Low production capacity will not be the only factor that could keep the supply low following vaccine discovery. Patents and licenses provide monopolistic incentives that can also keep the supply of vaccine low.

Anticipating possible obstacles to expanding the vaccine supply, policy leaders are proposing tax credits for a supplier who builds capacity in order to offer vaccines to impoverished countries. Policy makers may also need to contemplate the genuine possibility of several years of high spending with an HIV vaccine whose production costs are high. Even at a price of US$ 10 per course of vaccine, health minister in poor countries could still be reluctant to purchase vaccine, because US$10 per person would exhaust their average health budget. The models discussed predict that at this price it would require a roughly US$ 9 purchase subsidy per person to make the most optimal vaccine purchase affordable for health ministries in sub-Saharan Africa. That would require a vaccine purchase fund on the order of US$ 10 billion. An additional source of financing for the subsidies could be realised under a tiered pricing regime. Co-operative bargaining between the manufacturers and international agencies could maintain tiered pricing and potentially offer the manufacturer a guaranteed share of the consumers's surplus in exchange for subsidising below marginal cost prices for nations whose ability to pay is less than marginal costs.

To the extent that world markets are used to distribute vaccines, they will allocate vaccines to populations and regions based on ability to pay just like any other commodity. Achieving equity with less reliance on political support for taxation would suggest the alternative of tiered pricing.

2.9. Behavioral issues related to future HIV vaccine use (Stanton).

Future use of an HIV vaccine will need to seriously consider a number of behavioural issues, for the following reasons: i) initial vaccines might not be fully effective; ii) existing interventions can already reduce HIV infection on a public health scale: and iii) at-risk populations often tend to be those with least access to health care.

The two main questions to be addressed are related to vaccine acceptance and vaccine effect on behaviour.

There is little available information in relation to the acceptability of future HIV vaccines. On the other hand, there are several studies in relation to willingness to participate in HIV vaccine trials. For example, a study conducted in Uganda showed that 88% of the military would participate in trials. Studies conducted in three cities in the United States among MSM revealed that willingness to participate in trials declined from 37% at baseline to 21% at 12 and 18 months, underscoring the importance of the informed consent process. Willingness to participate in trials change depending on the educational level, perception of risk, and age. Altruism and a desire for protection are common motivators for participation, and vaccine safety is usually the major concern. Interest in participating in trials also declines as the hypothetical regime becomes more demanding.

Studies of acceptability of a hypothetical HIV vaccine conducted among college students in the United States have suggested that universal vaccine acceptance can not be assumed, and that certain health beliefs and previous experiences will influence acceptability. Safety and high vaccine efficacy will have strong influence on acceptability, followed by vaccine type and cost. Low efficacy vaccines (50%) were largely unacceptable.

There is conflicting data in relation to changes in risk behaviour during participation in HIV vaccine trials. Some studies have documented an increase in risk behaviour, whereas others have suggested that intensive counselling can prevent such changes. In any case, this is an important concern in the planning of HIV vaccine trials, and will be a key element of any future HIV immunization strategy.

Relevant behavioural data applicable to future HIV vaccine use could be obtained from the present experience with post-exposure prophylaxis (PEP) and the expanding access to highly active antiretroviral therapy (HAART).

Also relevant to future HIV immunization strategies is data indicating that intensive (but not brief) counselling and HIV testing alone can change behaviours. This suggests that the behavioural interventions accompanying future HIV vaccination programmes may have to be intensive to be effective.