Main

Less-developed countries are affected by the bulk of the global infectious disease burden. Preventing infectious diseases is often cheaper than caring for their victims, which makes immunization a cost-effective strategy. International public-health programmes responded appropriately decades ago by focusing on the delivery of basic childhood immunizations — bacille Calmette-Guérin (BCG) vaccine against tuberculosis, live oral polio vaccine, diphtheria-whole cell pertussis-tetanus (DPT) vaccine, monovalent measles vaccine and hepatitis B vaccine — known as the Extended Program of Immunization (EPI). Unfortunately, the introduction of improved vaccines and vaccine formulations (for example, acellular pertussis, inactivated polio, and measles combined with mumps and rubella vaccines) into developing-country programmes — as well as additional childhood vaccines routinely used in industrialized countries (for example, Haemophilus influenzae type b or Hib and pneumococcal conjugate vaccines), and vaccines for diseases largely limited to the developing world (for example, Japanese B encephalitis, yellow fever, typhoid fever and cholera vaccines) — has been slow1,2,3.

Conducting vaccine trials in less-developed countries is essential for vaccines that are intended for use in these areas, for several reasons (Box 1)4. First, some vaccines, such as those against hookworm, target diseases that are geographically limited to developing countries and so data on protection against naturally occurring disease can only be obtained in these countries. Second, even for diseases that occur in both developing and industrialized countries, such as rotavirus diarrhoea, vaccine trials done in the latter setting will not necessarily yield information that can readily be generalized to developing countries. Third, vaccine trials are sometimes needed to evaluate the feasibility, acceptability and impact of vaccines in public-health programmes in developing countries. Fourth, some new vaccines are now being developed within institutions in developing countries, with the intent to license the vaccine in the country of origin (for example, an oral killed whole-cell cholera vaccine has been developed in Vietnam; a Group B meningococcal vaccine has been developed in Cuba; and several Japanese B encephalitis vaccines have been developed in China). In this paper, we discuss several issues pertinent to the design and conduct of vaccine trials in developing countries. To illustrate the points discussed, Figure 1 provides a geographical depiction of several completed trials of vaccines (which we refer to in this article) that have been undertaken in the developing world.

Figure 1: Examples of completed vaccine trials conducted in the developing world.
figure 1

The trials are cited according to the number of the publication in the references.

Clinical development of candidate vaccines

The research, production and clinical development of the EPI vaccines (BCG, DTP and polio vaccine) mainly took place in publicly funded research institutes in industrialized countries. At present, vaccines can be developed either in the developing or the industrialized world. Indeed, a growing number of qualified vaccine producers capable of developing new vaccines have emerged in developing countries in recent years2. Over the years, a standard paradigm for testing new vaccines in humans has been developed5. A key feature is the phased fashion of the assessment, which is analogous to the phased manner in which drugs are tested (Table 1; Fig. 2).

Table 1 Characteristics of the different phases of vaccine evaluations
Figure 2: The sequential stages of vaccine trials in relation to rational deployment in developing countries.
figure 2

A simplified depiction of the stages of clinical studies; often vaccine development to licensure is iterative due to changes necessitated by results that arise during clinical studies. *Instead of Phase III trials, bridging studies or Phase II-b trials may be conducted to obtain licensure in certain circumstances.

Phase I trials. Phase I trials are the first human studies to be conducted after the preclinical demonstration of suitable purity, safety, potency, immunogenicity and protectiveness in animal models. The primary goal of a Phase I trial is to evaluate the possibility of frequent vaccine side effects; preliminary estimates of immune responses to vaccines are also obtained, as are assessments of vaccine shedding for live vaccines. Phase I trials can also be designed to evaluate different doses of the vaccine to aid the selection of the optimal dose for evaluation in subsequent phases of clinical development. Preliminary formulations and regimens of the vaccine are often tested in these Phase I trials, which are typically small, usually enrol healthy adults, and can be designed either with or without control groups of individuals who do not receive the vaccine. Traditionally, and for ethical reasons, Phase I trials are generally conducted in the country where the vaccine is developed.

Phase II trials. Phase II trials evaluate vaccine safety and immunogenicity in a larger number of subjects, and ultimately in the target population for whom the vaccine is intended. They are typically designed as randomized, controlled trials in which the comparator group receives a placebo or an active agent that permits blinding. For live vaccines, Phase II trials can also assess vaccine shedding and transmissibility. Although Phase II trials are also done to evaluate different vaccine doses and regimens, Phase II trial data are crucial to the dossier for licensure, and so it is essential that the final formulation, route and immunization schedule be tested in these trials.

If a vaccine candidate is targeted for use in less-developed countries, it is essential that Phase II studies are carried out in these settings. It has been shown that the immunogenicity of vaccines can be much lower in populations in developing compared with industrialised countries. Poorer performance has been especially problematic for orally administered live vaccines. A frequently cited example is the finding that three doses of oral polio vaccine, as formerly used in the United States, resulted in sustained and probably lifelong immunity, whereas children in certain developing countries require more than three doses for adequate seroconversion6,7. Similarly, a live oral Shigella vaccine (SC602) and a live oral cholera vaccine (CVD-103HgR) that were shown to be immunogenic in North American volunteers were subsequently found to perform less well when tested in humans residing in developing country settings endemic for these targeted diseases8. The poor performance of these vaccines in populations in developing countries is not well understood, but could be due to several factors, including high levels of pre-existing natural immunity, poor nutritional status, tropical enteropathy and co-existing infections6.

The high costs and complexity of Phase III trials of vaccine efficacy have led to the increasing use of Phase II trials with relatively larger numbers of subjects to obtain preliminary estimates of vaccine protection before committing to large-scale Phase III trials. These larger Phase II trials are referred to as Phase IIb trials. For example, 2,022 Mozambican children participated in a Phase IIb trial of RTS,S/ASO2A, a malaria vaccine candidate, during 2003–20049; a Phase IIb trial of three formulations of an oral, human-derived monovalent rotavirus vaccine, strain RIX4414, was recently conducted in 2,155 infants in Brazil, Mexico and Venezuela10; and a Phase IIb study of an HIV-1 vaccine candidate that elicits anti-HIV-1 cellular immunity11 was recently launched in 3,000 individuals in the United States12. Although larger than conventional Phase II studies, Phase IIb trials do not have the statistical power of a typical Phase III efficacy trial to evaluate vaccine efficacy.

Confusingly, vaccine studies that evaluate clinical protection through the use of intentional challenge with the target pathogen are also sometimes referred to as Phase IIb trials5. In these studies, volunteers who have been allocated to receive vaccine or no vaccine are challenged with the target pathogen at a defined interval after vaccination. Estimates of vaccine protection in Phase IIb challenge studies, often conducted in small groups of about 20–30 subjects, can be helpful in triaging vaccines that are deserving of study in larger, more expensive Phase III trials13,14. In addition to providing estimates of vaccine protection, Phase IIb challenge trials can provide data on vaccine safety, immunogenicity, shedding and transmissibility (for live vaccines), and preliminary assessments of immunological responses that correlate with protection. These trials are generally limited to healthy adult subjects and to infections for which there is no risk of severe acute complications, significant sequelae, or chronicity when appropriate therapy is administered promptly if the challenge results in infection. Examples of pathogens for which Phase IIb challenge studies have been successfully carried out include cholera, diarrhoeagenic Escherichia coli, Shigella, Rocky Mountain spotted fever, malaria and influenza15. These studies are undertaken in research centres, mainly in industrialized countries, with the capacity for intensive monitoring and care of subjects who have been intentionally challenged with pathogens and, where appropriate, with biological containment facilities to prevent release of the challenge pathogens into the general population.

Phase III trials. When promising results have been obtained from Phase II trials, Phase III trials are undertaken to provide rigorous evidence about vaccine protection against naturally occurring infections, as well as to obtain additional data in larger numbers of individuals on vaccine safety. Phase III studies are designed as randomized, controlled trials with clear endpoints, and are conducted in a population that normally experiences the target infection and for whom vaccine licensure is desired. Phase III trials are usually large, sometimes enrolling tens of thousands of subjects who might need to be followed for several years post-vaccination. The large size, prospective conduct, lengthy duration and extensive quality-control and quality-assurance procedures needed for Phase III trials make them extremely expensive, sometimes costing tens of millions of dollars.

Bridging studies. Bridging studies can be conducted to achieve a number of goals: to show that a new vaccine is similar in performance to an existing alternative vaccine against the same pathogen; that a new formulation or new schedule of a vaccine is no worse than an existing formulation or schedule of the same vaccine in the same target population; or that a particular vaccine has acceptable performance in a new target population (defined by age, ethnicity, co-morbidity, geographical residence or some other feature). For vaccines for which there are immunological tests that measure responses that are regarded as immunological correlates of protection, bridging studies showing acceptable immunogenicity and safety can be sufficient to license the new vaccine or vaccine indication without resorting to a large-scale efficacy trial16.

Depending on the particular study question, these studies can be designed as non-inferiority or equivalence trials17,18. Such trials are designed to show that a new vaccine, vaccine formulation or vaccine regimen is similar to a licensed or standard vaccine in some measure(s) of safety, immunogenicity or efficacy. This means that the purpose, hypothesis and setting of non-inferiority and equivalence studies are different from trials that are constructed with control groups receiving agents, such as placebos, that are not expected to have a protective effect against the target pathogen. The distinction between non-inferiority and equivalence trials is that for the former, the question assessed is whether the new agent is no worse than the existing agent with respect to the outcome of interest; for equivalence trials, the question is whether the new agent is neither worse nor better than the existing agent. Non-inferiority trials therefore use one-tailed tests of statistical significance, whereas equivalence trials use two-tailed tests.

Non-inferiority rather than equivalence is the conventional design for bridging studies, because the general objective is to demonstrate that a new vaccine, new vaccine formulation or new vaccine regimen is no worse than a standard, available vaccine. As it is statistically impossible to demonstrate that two agents are the same, one can only assess whether the confidence interval for the observed difference between the compared groups exceeds a pre-specified margin or not, which is typically a difference that is considered clinically acceptable19. If the confidence interval excludes this difference, a conclusion of non-inferiority is reached. An essential requirement for a non-inferiority vaccine trial is that the licensed or standard comparator vaccine must have previously and consistently been shown to be safe, immunogenic or protective (depending on the endpoint), preferably in multiple placebo-controlled studies. International guidelines stipulate that if this condition is not met, one should consider a design other than non-inferiority19. Sample size requirements of non-inferiority trials vary according to the outcome of interest, the assumed true difference between the study groups and the maximally tolerated difference required to reach a conclusion of non-inferiority20.

As discussed below, bridging studies can be used to obtain vaccine licensure by showing comparable safety and immunogenicity profiles between a newly developed and similar licensed vaccine or between combination and individual component vaccines. With increasing information on immunological correlates of protection, bridging studies might be used more frequently in the future to obtain licensure of new-generation vaccines in developing countries.

Vaccine licensure pathways. In the clinical development of a vaccine candidate, the sequence of vaccine trials described above usually culminates in one or more Phase III trials as the basis for vaccine licensure5. Sometimes, proof of efficacy in a Phase III trial is not sought; instead, evidence of protection from Phase IIb challenge studies or from serological endpoints in bridging studies can be used in the application for licensure. For example, CVD103-HgR, a vaccine against cholera, was licensed in Europe for use in travellers on the basis of Phase I and II studies in the US and various countries in the developing world, together with Phase IIb challenge trials in North American volunteers that demonstrated protection against experimental cholera21. After licensure of other conjugate vaccines against Hib, PRP–tetanus toxoid conjugate vaccine was licensed for infants in the United States largely on the basis of proven safety and the attainment of protective serum antibody titres22,23. Similarly, combination vaccines are often submitted for licensure on the basis of non-inferiority trials that demonstrate safety and immunogenicity profiles similar to component vaccines administered separately24,25,26.

Licensure of vaccines in developing countries presents its own particular challenges27. Most current vaccines in wide use were developed and initially implemented in the industrialized world and regulatory decisions for these products were taken in their country of origin. Import, licensure or local production of these vaccines follows the legal provisions of individual developing countries under the auspices of a National Regulatory Agency (NRA).

Some vaccines targeting diseases that occur with appreciable frequency only in developing countries are now being developed by producers in industrialized countries (for example an oral live oral cholera vaccine is being developed by Avant Immunotherapeutics and a malaria vaccine is being developed by GlaxoSmithKline). If there is no anticipated use of these vaccines for populations in industrialized countries, regulatory authorities in the country of production might be reluctant to take responsibility for regulating all phases of licensure, production and use of the vaccine. At the same time, it is crucial that such products receive appropriate regulatory control from initial development to use in humans to ensure that they are safe, effective and of high quality27. Several possible regulatory options have been offered for consideration: licensing in the country of manufacture, with file review by the European Medicines Evaluation Agency on behalf of the World Health Organization (WHO); use of a contracted independent entity for global regulatory approval; shared manufacturing and licensing in a developing country with competent manufacturing and regulatory capacity; and export to a country with a competent NRA that could handle all regulatory functions for the developing world27.

In the past NRAs in many developing countries were considered to be weak with respect to one or more core competencies. The WHO has defined six functions that could be used to assess, monitor and design interventions to improve NRAs27,28. These include the publishing of requirements for licensing of products and manufacturers, surveillance of vaccine field performance (safety and efficacy), a system for batch or lot release, use of the laboratory when needed, regular inspections of manufacturers for Good Manufacturing Practice compliance, and evaluation of clinical performance through authorized clinical trials28. The WHO has been leading work in assessing and improving vaccine regulatory capacity in developing countries, and NRAs in several developing countries are now WHO-approved.

The presence of functioning NRAs and regulatory pathways in countries such as Indonesia and India allow the production in these countries of WHO pre-qualified EPI vaccines for United Nations' purchase and global use. Efforts are also currently underway to transfer the technology for the production of certain orphan vaccines to producers in developing countries with WHO-approved NRAs in an effort to expand the availability of these vaccines in developing countries. A recent example is provided by the programme of the International Vaccine Institute in Korea to transfer killed oral cholera vaccine to producers in Indonesia and India, where the presence of WHO-approved NRAs will facilitate the acceptance of the vaccine by other developing countries that might wish to deploy it.

Other information obtained from vaccine trials

Other than yielding data for licensure, pre-licensure vaccine trials can provide several additional types of important information, including immunological correlates of protection and evidence of indirect effects of vaccination. Alternatively, some of this information might be obtained post-licensure, as there is a rising concern that increased demands for pre-licensure data can delay registration of much-needed products.

Immunological correlates of vaccine protection. During Phase III trials immunological correlates of vaccine protection can be evaluated by comparing immunological responses to vaccination (for example, serum antibody levels) with the probability of being protected by the vaccine5. For vaccines shown to be protective in such trials, correlates of protection are best assessed by comparing immune responses in vaccinees who develop the target infection (sometimes referred to as 'breakthrough events') versus those who do not develop the target infection. These correlates of protection can be extremely useful for programmes that monitor the implementation of vaccines in routine public-health practice, for the development and evaluation of future-generation vaccines and for use in bridging studies. As discussed above, the availability of immunological correlates of vaccine protection can allow the assessment of protection afforded by new vaccine candidates compared with standard or licensed vaccines through small, short-term studies with immunological endpoints instead of conducting large, expensive Phase III efficacy trials with clinical infection endpoints.

Determining such immune markers of protection is logistically demanding, expensive and scientifically challenging for several reasons. First, to determine an immunological correlate of protection, it is necessary to contrast short-term, often wide-ranging immune responses in vaccinees with and without breakthrough events. However, at the time of vaccination and collection of blood or other specimens for immunological assessments, it is not known which and when vaccinees will have breakthrough events. Second, most Phase III trials have very few breakthrough events. Third, blinding of Phase III trials prevents knowledge of who received the vaccine and who received the control agent. The implication of these considerations is that it might be necessary to obtain suitable specimens from virtually every participant in the trial to ensure that immunological assessments will be available for a large enough number of vaccine breakthrough events that a statistically meaningful assessment of the putative immunological correlate can be made5. Even when the necessary specimens are collected from the required number of subjects, the measured immune responses among vaccines might not correlate with the probability of infection among vaccines, sometimes for obscure reasons29.

Measurement of indirect protective effects. Vaccines can protect in several ways30. Direct protection is protection conferred by immune responses elicited by a vaccine administered to an individual, irrespective of whether other persons living around the individual have also been vaccinated. For infections that are transmitted from person to person, either directly or via an intermediary (for example, insect vectors), vaccines can also reduce the intensity of transmission of an infection in a population if a sufficient proportion of the population is immunized, leading to the protection of non-vaccinated persons residing in that population. This is known as indirect protection. The reduced transmission in immunized populations can in turn lead to the enhanced protection of vaccinated individuals — that is, a combination of direct and indirect protection, and this is often termed total protection.

Traditionally, the indirect protective effects of vaccines have been inferred from observational studies, such as monitoring secular trends of disease incidence. Examples include a surveillance study in Israel that showed how a universal toddler hepatitis A immunization programme was followed by a decline in disease incidence in older age groups31, one in the Gambia that showed elimination of Hib disease after the introduction of routine childhood immunization with a Hib conjugate vaccine32, and another in the United States which suggested that routine vaccination of young children with a seven-valent pneumococcal conjugate vaccine prevented more than twice as many invasive pneumococcal disease cases through indirect than direct effects33.

Newer methodological techniques have also made it possible to assess indirect protective effects in vaccine trials conducted before licensure. The most straightforward approach is to conduct cluster randomization trials (CRTs), in which clusters of individuals, defined by consideration of contact between individuals within each cluster, serve as the units randomized to the vaccine and control arms of the trial34. In well-designed CRTs, transmission of the disease of interest occurs within geographical clusters, but inter-cluster transmission is minimal. Units can comprise geographic areas, schools, workplaces or families. Vaccine protection assessed by comparing disease incidence in persons receiving the experimental vaccine versus those receiving the control agent in a CRT is the total effect, a combination of direct protection (conferred on individuals because they themselves received the vaccine) plus any indirect protection (conferred on individuals because they live in a cluster in which a suitably large number of other individuals has received the vaccine). Furthermore, comparison of disease incidence in non-vaccinated members of clusters receiving the experimental vaccine versus those in clusters receiving the control agent enables measurement of indirect vaccine protection. The CRT design is illustrated by a recent pre-licensure trial of a conjugate pneumococcal vaccine in Native American children, which served as a pivotal trial for licensure of this vaccine in the United States35.

Until recently, it had been thought that individually randomized trials were capable only of measuring direct protection of the individual. The use of Geographic Information Systems (GIS) mapping provides a new method for assessing indirect effects within individually randomized trials, if the trials are conducted in study sites with discrete geographical subpopulations. In this design, illustrated by a recent analysis of a placebo-controlled, individually randomized trial of killed oral cholera vaccines in Bangladesh36, advantage is taken of the differing levels of vaccine coverage of geographically defined groups of individuals that can occur by chance in the randomization process and differing rates of participation. The risk of the target infection in non-vaccinated members of these discrete geographical areas is correlated with levels of vaccine coverage to assess indirect protection, whereas correlation of infection rates among vaccinees with levels of vaccine coverage in the discrete areas allows estimation of total vaccine protection.

Using vaccine trials to guide vaccine introduction

Traditionally, vaccine trials have been conducted simply to achieve vaccine licensure and their evaluation has mainly been limited to Phase I–III trials. More recently, trials have been used in innovative ways to provide information on disease burden, data on vaccine safety and evidence to support the introduction of licensed vaccines into public-health programmes of developing countries.

Assessment of disease burden. An estimation of disease burden is crucial to decisions about introducing new vaccines in developing countries. In determining accurate estimates of disease burden, the detection and confirmation of disease is often a major problem. The best available diagnostic tests can fail to detect many cases of illness caused by some pathogens. Vaccine trials have been found to be useful to 'probe' the burden of disease that is not detectable using diagnostic tests. For example, only a proportion of invasive Hib-associated illness is detectable through cultures or even more sensitive diagnostic testing (for example, polymerase chain reaction) of normally sterile body fluids. In a large, randomized trial of the Hib–tetanus protein conjugate vaccine in Gambian infants, protection was shown not only against culture-positive invasive disease, but also against culture-negative pneumonia, presumably because of the insensitivity of cultures in confirming Hib pneumonia37,38. Vaccination therefore demonstrated that the burden of Hib-associated disease is much greater than culture-proven disease. Particularly in less-developed countries, where laboratory confirmation of diagnoses can be difficult and where widespread over-the-counter use of antibiotics can create false-negative diagnoses, the use of highly protective vaccines has been advocated to probe the total burden of difficult-to-confirm illnesses such as Hib-associated39 and pneumococcal disease40. As an illustration, the Global Alliance for Vaccines and Immunization is currently supporting a large-scale, multi-site 'probe' trial of conjugate Hib vaccine in India to help resolve controversies about the burden of invasive Hib disease in that country and to assist policy decisions about the introduction of routine immunization against Hib41.

Assessment of vaccine safety. After licensure and introduction, vaccines continue to be monitored for safety in Phase IV observational studies. Phase IV studies typically compare outcomes of persons who had or had not received a vaccine. Post-licensure studies can lead to observations of serious adverse events not appreciated prior to introduction. For example, in 1998 a tetra-valent rhesus-based rotavirus vaccine (RRV-TV) was licensed in the United States and recommended for routine vaccination of infants. Subsequently, the country's Vaccine Adverse Event Reporting System detected 15 cases of intussusception among infants who had received RRV-TV42. After an association was established in several controlled studies, immunization was suspended and the vaccine removed from the market43. Likewise, a post-licensure study in Switzerland that showed a strong association between an inactivated, intranasal influenza vaccine and the occurrence of Bell's palsy led to withdrawal of this vaccine44.

Well-developed systems for conducting observational Phase IV studies of vaccine safety are rare in developing countries. Furthermore, these countries often lack stringent systems to detect adverse events following immunization (AEFI). Sometimes, however, Phase III trials conducted earlier can be used for this purpose. For example, in the late 1980s efficacy trials in sub-Saharan Africa of the Edmonston–Zagreb high-titre measles vaccine demonstrated that immunization of 6-month olds could prevent the disease in the latter half of infancy45,46. Selective use of the vaccine was recommended by the WHO for countries in which measles before the age of 9 months was a substantial cause of death47. Subsequently, long-term follow-up of participants in an already completed trial in rural Senegal showed significantly higher child mortality after immunization with high-titre than standard measles vaccine48, providing important evidence in WHO's decision to rescind its previous recommendation49.

Evaluation of feasibility, acceptability, costs and practical impact. Post-licensure trials can be used for an additional purpose. The decision to introduce a new vaccine into a developing country's public-health programme can have major programmatic and financial ramifications, and so policymakers are increasingly demanding more evidence on feasibility, acceptability, cost-effectiveness and practical impact of vaccines that are already licensed. In the past, this information was provided primarily by non-randomized studies, such as a study in Bangladesh which showed that measles immunization markedly improved overall survival of children50. Recently, more rigorously designed randomized trials have been proposed as a method to assess the effectiveness of a vaccine when given under routine conditions to a target population51. For example, the International Vaccine Institute has embarked on cluster-randomized effectiveness trials of the internationally licensed Vi polysaccharide typhoid vaccine in several Asian countries to provide evidence for wider-scale implementation4. In these trials, study populations are randomly allocated to receive the Vi or a control vaccine. Socio-behavioural and economic research studies are integrated into the trials. Expected outputs from these trials conducted in realistic public-health settings include evaluation of the feasibility and acceptability of vaccination from qualitative substudies; assessment of cost of vaccination and cost of typhoid illness from economic sub-studies; and measurement of protective impact against typhoid fever.

Sound and ethical clinical research

Vaccine trials in developing countries should comply with good clinical research practice to ensure that the rights, safety and well-being of participants are protected and that the trial data are credible5. The Good Clinical Practice (GCP) guidelines developed by the International Conference on Harmonization52 have been proposed as the gold-standard for the conduct of clinical research. However, these GCP guidelines are based on a consensus of experts and not on evidence. There have been increasing calls for the guidelines to be made more scientific, up to date, flexible and simple through collaborative and evidence-based efforts53,54,55. For vaccine trials in developing countries, rigid adherence to GCP standards as they are now formulated can be a double-edged sword. The complexity and expense of clinical trials has risen rapidly in recent years. A portion of this increased expense arises from the extensive documentation and auditing requirements demanded by GCP guidelines. Although this increased expense is accepted by producers anticipating major markets for their vaccines in affluent markets, it constitutes a disincentive to the clinical testing of vaccines against 'orphan diseases' affecting developing countries, for which lucrative markets are not foreseen5.

The ethics of clinical trials in developing countries has been the subject of intense discussion56. Many of the ethical concerns include choosing the appropriate research question and design, ensuring prior scientific and ethical review of the proposed protocol, minimizing the risk to participants, assuring a reasonable risk:benefit ratio, providing compensation for injuries directly related to the research, obtaining individual informed consent, providing equal consideration to participants, ensuring equal distribution of the burden and benefits of the research, and providing independent oversight of trials by properly constituted Data Safety Management Boards57. These concerns are consistent with principles embraced in many documents from governmental, non-governmental and international organizations that provide ethical guidelines for human research in developing countries. These include the World Medical Association's 1964 'Helsinki Declaration' (amended in 2000); the US National Bioethics Advisory Commission's document on 'Ethical and policy issues in international research: clinical trials in developing countries' (2001); the Nuffield Council on Bioethics report on 'The Ethics of Research Related to Healthcare in Developing Countries' (April 2002); the Council for International Organizations of Medical Sciences document on 'International Ethical Guidelines for Biomedical Research Involving Human Subjects' (September 2002); and the European Group on Ethics in Science and New Technologies opinion on the 'Ethical Aspects of Clinical Research in Developing Countries' (January 2003)58.

However, there are challenges to implementing the principles outlined in these documents in the developing world (Table 2)57. Many of these challenges can be addressed by increasing awareness and expertise of local investigators in the area of research ethics and their involvement in the development of updated international standards for the ethical and scientific conduct of clinical research59. The Strategic Initiative to Develop Ethical Review, the Fogarty International Center of the National Institutes of Health, and the WHO have begun targeting ethics training in less-developed countries57.

Table 2 Challenges in the implementation of ethical responsibilities in less-developed countries*

To enhance the benefits of vaccine trials in developing country settings, innovative designs have been advocated. For example, in randomized controlled trials, it is recommended that an active vaccine that might protect participants against a pathogen other than the pathogen targeted by the experimental vaccine be used as the control agent instead of placebo whenever possible. Another strategy is the CRT design. Mainly used to study therapeutic and behavioural interventions against infectious diseases in less developed countries60, the CRT design is now increasingly being used to evaluate vaccines61,62. Obtaining consent first at the community then at the individual level, cluster allocation of the intervention, and maximizing the potential impact of the vaccine through the combined effects of direct and indirect protection make the CRT design particularly suited to the settings of many developing countries. The stepwise introduction of a vaccine in a country's public-health programme could allow assessment of its impact while ensuring distribution of its protective benefits63. Finally, including negotiations during the planning stage on the post-trial obligations of sponsors to provide vaccine to the communities where trials are done could increase research benefits to impoverished communities participating in vaccine trials.

Summary and future perspectives

In recent years there has been a burgeoning of vaccine candidates against diseases that primarily affect developing countries, and future exploitation of whole-genome sequencing of pathogens is likely to sustain or even augment this trend. This trend will create exciting possibilities for disease prevention, and offer public-health policymakers attractive options for the cost-effective control of important diseases. Vaccine trials are increasingly being conducted in developing countries to evaluate vaccine performance in these settings. The profusion of vaccine candidates for the developing world has added a layer of complexity to the seemingly straightforward phased sequence of trials. There are numerous factors underlying this complexity, including considerations about when in the phased testing sequence a vaccine developed in an industrialized country should begin trials in developing countries; about the problems of vaccine licensing in the absence of standard regulatory pathways; about the additional studies that are needed to inform policymakers on rational introduction of vaccines; and about how to meet the scientific and ethical challenges of conducting clinical trials in developing countries.

In the future, more vaccines and vaccine candidates to help control infectious diseases in poor countries will become available, thereby increasing the need for vaccine trials in these settings. As experience is gained in this field, current innovative strategies will become common practice. As discussed, these include making more use of Phase IIb trials and bridging studies, obtaining vaccine licensure in developing countries, and conducting Phase IV trials to guide implementation. The design of trials in developing countries might more frequently be cluster-based, with an active agent as a control. Agreements of sponsors to provide vaccines to the communities in which trials are carried out should become the norm. In addition, with increased training of local scientists, it is hoped that future trials will be designed and conducted mainly by local scientists. It is also hoped that with capacity-building and infrastructure development, trials in developing countries will spread from the traditional research institutes where they tend to currently cluster. The manner in which these future developments take place will have a great bearing on the success of efforts to accelerate the introduction of new-generation vaccines into programmes for the poor.