Tweet
he Lancet Infectious Diseases 2006; 6:522-528 DOI:10.1016/S1473-3099(06)70552-7
Early clinical trials with a new tuberculosis vaccine, MVA85A, in tuberculosis-endemic countries: issues in study design
Hannah B IbangaFWACP a, Roger H BrookesPhD a, Philip C HillFRACP a, Patrick K OwiafeHND a, Helen A FletcherPhD c, Christian LienhardtMD b, Adrian VS HillFMedSci c, Richard A AdegbolaFRCPath a and Dr Helen McShaneMRCP
c 
Summary
Introduction
New vaccine development
Tuberculosis specific issues
HIV infection: prevalence and acceptability of testing within a clinical trial
Trial awareness and recruitment
Ethical issues
Results of screening and reasons for exclusion
Conclusions
Search strategy and selection criteria
References
Summary
Tuberculosis remains a substantial global health problem despite effective drug treatments. The efficacy of BCG, the only available vaccine, is variable, especially in tuberculosis-endemic regions. Recent advances in the development of new vaccines against tuberculosis mean that the first of these are now entering into early clinical trials. A recombinant modified vaccinia virus Ankara expressing a major secreted antigen from Mycobacterium tuberculosis, antigen 85A, was the first new tuberculosis vaccine to enter into clinical trials in September 2002. This vaccine is known as MVA85A. In a series of phase I clinical trials in the UK, MVA85A had an excellent safety profile and was highly immunogenic. MVA85A was subsequently evaluated in a series of phase I trials in The Gambia, a tuberculosis-endemic area in west Africa. This vaccine is the only new subunit tuberculosis vaccine to enter into clinical trials in Africa to date. Here, we discuss some of the issues that were considered in the protocol design of these studies including recruitment, inclusion and exclusion criteria, reimbursement of study participants, and HIV testing. These issues are highly relevant to early clinical trials with all new tuberculosis vaccines in the developing world.
Back to top
Introduction
In 1993, WHO declared the current tuberculosis epidemic ?a global emergency?.1 More recently, in August 2005, WHO has again declared the tuberculosis epidemic in Africa an emergency situation.2 Worldwide, there are 8 million new cases and 2 million deaths each year from this disease.3 98% of these deaths occur in the developing world and countries of the former Soviet Union. The global prevalence of latent Mycobacterium tuberculosis infection is estimated to be 32% (1?86 billion people).3 Such individuals are at risk of reactivation of this latent infection should they become immunosuppressed for any reason. Globally, coinfection with HIV is the most important risk factor for progression of latent M tuberculosis infection to tuberculosis disease and the current HIV pandemic has fuelled the tuberculosis epidemic. The emergence of multidrug-resistant strains of M tuberculosis has further compounded the problem and made the need to control this disease even more urgent.
The best way to control the tuberculosis epidemic would be with an effective vaccine. BCG, the only available vaccine against M tuberculosis, is a live attenuated strain of Mycobacterium bovis that was first used in 1921.4 Since then it has become a well established part of the WHO Expanded Programme on Immunisation and is widely administered at birth throughout the developing world. When administered at birth it confers consistent and appreciable protection against disseminated disease, including tuberculous meningitis.5 However, adult pulmonary disease is responsible for the huge global public-health burden. The protection imparted by BCG against pulmonary disease has been evaluated in a number of large randomised controlled trials and observational studies, and has been shown to vary greatly, from 0 to 80%.4,6 Furthermore, a recent large cluster randomised trial estimating the effect of BCG revaccination in Brazilian school children who had been vaccinated with BCG at birth demonstrated that BCG revaccination did not confer any additional protection and should therefore not be recommended.7
There are several explanations for the variable efficacy in protection against pulmonary disease conferred by BCG. The one best supported by the available data is that exposure to environmental mycobacteria?eg, Mycobacterium avium, Mycobacterium marinum, and Mycobacterium intracellulare?interferes with BCG ?take?. There are two possible mechanisms for this interference: masking and boosting. First, exposure to environmental mycobacteria induces an antimycobacterial immune response, and subsequent vaccination with BCG does not increase or boost that response. Evidence of such masking comes from a series of studies done in parallel in adolescent school children in the UK and in Malawi.8 In the UK, baseline cellular immunity to mycobacterial antigens was low, and was increased after BCG vaccination. By contrast, in Malawi, the baseline cellular immune response to mycobacterial antigens was high and BCG vaccination did not result in a substantial incremental rise in these responses. The second possible mechanism for environmental mycobacterial interference is that antimycobacterial immunity induced by environmental mycobacteria abrogates the replication of BCG and subsequent induction of an immune response. This blocking hypothesis is supported by data from murine experiments.9 Importantly in this murine study, the protective efficacy of the subunit vaccines tested was not abrogated by prior exposure to environmental mycobacteria.
A second explanation for the variable efficacy of BCG is the age at administration. Trials done in young, mycobacteria-naive children have been generally positive. Other possible explanations for the variability in BCG efficacy across different trials include differences between BCG strains, host genetics, and nutritional factors.10?12 Several of these factors might contribute to the variability seen.
Back to top
New vaccine development
There is an urgent need for a better vaccination strategy. Since BCG does confer substantial protection against disseminated disease in childhood, ideally any new vaccination strategy should include BCG to retain this protective effect. Over the past decade there has been a resurgence of interest in the development of new tuberculosis vaccines and some of the most promising of these are now entering into early clinical trials.
M tuberculosis is an intracellular pathogen and protective immunity is entirely dependent on an intact cellular immune response.13 A Th1 type CD4 T-cell response is essential for protective immunity, and MHC class I restricted CD8+ T cells might also be required. Although a validated correlate of protection has not yet been identified, the secretion of interferon γ from antigen-specific T cells is essential for the protective immune response to this pathogen.14,15
There are two main strategies for developing an improved tuberculosis vaccine. The first is to develop a subunit vaccine (DNA, recombinant viral vector, or recombinant protein) expressing an immunodominant antigen(s) from M tuberculosis. Such vaccines are largely being developed to boost BCG rather than replace it. The second approach is to use the whole organism, either by developing a recombinant strain of BCG or by developing an attenuated strain of M tuberculosis. This strategy aims to replace BCG. Any candidate tuberculosis vaccine is evaluated in a series of animal models, and the most promising candidates are then considered for evaluation in early clinical trials. To date, only four such candidates have entered into clinical trials. The first, a recombinant modified vaccinia virus Ankara expressing antigen 85A known as MVA85A, entered into clinical trials in the UK in September 2002.16 Phase I studies have demonstrated that MVA85A is safe and induces high levels of cellular immunity when used to vaccinate BCG-naive adults. When used in BCG-primed adults, significantly higher levels of cellular immunity are induced, which persist for longer (p=0?015).17 In May 2003, when sufficient safety data was obtained from the UK studies, MVA85A entered into phase I clinical trials in The Gambia (figure 1). MVA85A is currently in phase II clinical trials in the Western Cape in South Africa.
Click to enlarge image
Figure 1. Members of the MRC Gambia vaccine trial team celebrate with the first person in Africa to be vaccinated with MVA85A at the end of his follow-up
The second vaccine to enter into clinical trials was a recombinant strain of BCG over-expressing antigen 85B, rBCG30.18 This vaccine entered into clinical trials in the USA in early 2004.19 The third vaccine that has entered into clinical trials is a recombinant fusion protein comprising the 32 and 39 kDa proteins from M tuberculosis, M tuberculosis 72F, used in conjunction with an adjuvant (AS02).20 This vaccine also entered into clinical trials in the USA in early 2004.21 In November 2005, a fusion protein expressing ESAT6 and antigen 85b entered into phase I studies in Europe.22 In addition to these four vaccines, there is an ongoing phase III efficacy trial of an inactivated whole-cell mycobacterial vaccine in HIV-infected individuals in Tanzania.23
Here, we discuss the issues that arose in the design, setting up, and implementation of the clinical trials with MVA85A in The Gambia.
Back to top
Tuberculosis specific issues
The Koch reaction
The Koch reaction describes the development of immunopathology in a person or animal with tuberculosis when an exaggerated immune response to M tuberculosis is stimulated.24 Such pathology occurs both at the site of infection and at the site of vaccination. In 1891, Koch observed that 4?6 weeks after the establishment of M tuberculosis infection in guineapigs, intradermal challenge with either the whole organism or culture filtrate resulted in necrosis and subsequent healing of the injection site and similar lesions in the original tuberculous lesion.25 In human beings, Koch found that injection of larger quantities of culture filtrate subcutaneously into tuberculosis patients evoked necrosis in established tuberculous lesions at distant sites, with disastrous results in the spine and lungs.
As the first new candidate tuberculosis vaccines enter into clinical trials, there is a concern that highly immunogenic vaccines could induce such a reaction. The available animal data suggest that such reactions correlate with bacterial load and are unlikely to occur in animals or human beings with latent infection, where the bacterial load is low.26 To minimise the risk of a Koch reaction occurring, we began the clinical trials with MVA85A in individuals with minimal pre-existing mycobacterial immunity.
The evaluation of pre-existing antimycobacterial immunity
For any new tuberculosis vaccine entering into clinical trials, it is necessary to determine the degree of pre-existing mycobacterial infection in individuals at screening. This assessment allows early trials to be done in individuals with minimal pre-existing exposure, and thus minimise the risks of inducing a Koch reaction. The clinical studies with MVA85A in the UK and in The Gambia were designed to allow for vaccination of patient groups sequentially with a step-wise increase in mycobacterial exposure to minimise this risk.16 In addition, studies were done sequentially in the UK and then in The Gambia, where exposure to both environmental mycobacteria and M tuberculosis is higher.27,28 In these studies, we used a combination of tuberculin skin testing, together with measurement of cellular immune responses using an ex-vivo interferon γ Elispot assay, to determine the pre-existing antimycobacterial immunity.
Tuberculin skin test
The tuberculin skin test is the standard test used throughout the world to determine the level of antimycobacterial immunity, and therefore mycobacterial infection. This test was first used in 1890, and is the oldest diagnostic test in clinical use.29 It involves an intradermal injection with tuberculin purified protein derivative (PPD). The size of the delayed-type hypersensitivity response is measured 2?3 days later. The biggest limitation of the tuberculin skin test is that PPD lacks specificity for M tuberculosis infection because many of the proteins present have homologues in a number of non-tuberculous mycobacteria. A positive tuberculin skin test response?ie, a Mantoux response of greater than 15 mm in an individual who has been BCG vaccinated, or greater than 10 mm in someone without BCG vaccination?can therefore result from infection with M tuberculosis, prior vaccination with BCG, or exposure to environmental mycobacteria.30,31 A further limitation of the tuberculin skin test is that it can lack sensitivity?eg, in immunosuppressed individuals.32 Despite these limitations, the tuberculin skin test remains a commonly used test in determining infection with M tuberculosis throughout the world.33
In-vitro T-cell assays: the ex-vivo interferon γ Elispot assay
In-vitro assays of cellular immunological function are increasingly being used in parallel with the tuberculin skin test to obtain greater sensitivity and specificity with respect to the level of pre-existing immunity to mycobacteria. These assays are based on the measurement of the secretion of interferon γ after stimulation with mycobacterial antigens.34 In the clinical studies with MVA85A, the ex-vivo interferon γ Elispot assay was used as the main cellular immunological assay.
In these early vaccine studies, we were interested in quantifying Elispot responses to both PPD and two M tuberculosis-specific antigens, ESAT6 and CFP10. The responses to these two antigens allowed the diagnosis and exclusion of individuals who were latently infected with M tuberculosis.35 ESAT6 and CFP10 are highly immunogenic and are among the early secreted antigens in M tuberculosis infection. Their genes are not present in any strain of BCG and homologues of these genes are absent from a number of relevant non-tuberculous mycobacteria.36?38 Individuals with active tuberculosis were excluded on the basis of clinical history, ESAT6 and CFP10 Elispot response, and chest radiograph findings. Furthermore, to identify individuals who were as mycobacterially naive as possible for these early studies, we set a relatively low threshold for the upper limit to the PPD Elispot response at screening. Individuals with a PPD response greater than 100 spot forming cells per million peripheral blood mononuclear cells were excluded from the study.
Elispot is one of the most sensitive cellular immunological assays available.39 The assay does not require expensive equipment, requires relatively low technology, and has been used in larger scale epidemiological studies and therefore has a potential role in larger scale vaccine studies.40 For these reasons, in the clinical studies with MVA85A, the ex-vivo interferon γ Elispot assay was used as the main cellular immunological assay. The assay was done at screening to determine pre-existing antimycobacterial immunity and repeated after vaccination to monitor vaccine-induced changes in immune responses.
Back to top
Recruitment of BCG-naive subjects: reliability of BCG scar detection
In the first clinical trial with MVA85A in The Gambia, the aim was to recruit individuals who had not been BCG vaccinated and who had minimal pre-existing mycobacterial exposure. BCG-vaccinated individuals were subsequently recruited once the safety of MVA85A had been demonstrated in BCG-naive individuals. BCG was not widely used in The Gambia until 1985. The most reliable method to determine if a person has been vaccinated with BCG is to examine the infant welfare card where vaccinations given in infancy are documented or the mothers are asked about this vaccination. In the absence of such source data, we had to rely on the presence or absence of a scar. Absence of scar is an imperfect indicator of whether or not someone has received BCG vaccination. In the UK, it has been reported that about 26% of individuals may not produce a scar 2 years after BCG vaccination.41 A study in Malawi showed that BCG scar was a highly sensitive indicator of vaccination status when vaccine was properly handled and given at over 3 months of age, but in infants under 1 month of age vaccinated in health centres, sensitivity of BCG scar had declined to under 80% by 4 years post-vaccination.42 Despite these limitations, presence of a BCG scar was an exclusion criterion in the BCG-naive study.
Back to top
HIV infection: prevalence and acceptability of testing within a clinical trial
In The Gambia, the prevalence of HIV-1 infection is 1?0%, and the prevalence of HIV-2 infection is 0?8%.43 There is a high level of public awareness about HIV infection and people are encouraged to come for voluntary counselling and testing. However, the stigma surrounding HIV is still high. Full pretest and post-test counselling was provided for all of the individuals screened for this study. Individuals were informed that they would be tested for HIV, but had the option of not being told what the result was. In a Zambian study in which 475 people were HIV screened, 41 (8?6%) people did not come back for their results,44 including six people who did not want to know the results, and others who gave reasons such as distance from the centre, death, or illness.
In total, 228 people were screened for the MVA85A vaccine trials, and there were no positive HIV results, indicating the low HIV prevalence in this population. If any individuals had tested positive for infection with HIV, they would have been referred to the local genitourinary medicine clinic within the Medical Research Council (MRC) unit, where they would be provided with ongoing medical care and considered for free antiretroviral therapy. In the studies with MVA85A, none of the study participants objected to having an HIV test, after suitable counselling. This study demonstrates the acceptability of HIV testing in this study population, although this may be different elsewhere.
Back to top
Trial awareness and recruitment
Before the study started, officers in the government Departments of State for Health and Social Welfare were informed about the trials. Meetings with community and religious leaders followed these visits to sensitise the community about the trial. In conjunction, there was a media campaign educating the population on tuberculosis. The media campaign included radio and television programmes in the major local languages and English. These programmes were both live phone-in and educational programmes. Media releases were done in the major newspapers. Additionally, the study team distributed T-shirts and caps printed with calls for people to be involved in the tuberculosis vaccine trial.
At the mosques or institutions visited, the people present were encouraged to visit the MRC and obtain more information on the trial. Others responded to the media campaign by either making phone calls or coming to the MRC unit. At the MRC unit, more detailed information was given on the trial. Information sheets, in English, were translated verbally into the relevant local language (Wolof or Mandinka).
Back to top
Ethical issues
Consent
Informed consent is an essential component of any research involving human patients. This consent must be from the individual taking part in the research, but could also involve the family or community in which that individual lives. In The Gambia, the family and community might be involved in the decision to participate, as well as the individual. Community awareness meetings conducted in the mosques and educational institutions enabled discussion of the rationale and study requirements with the community and religious leaders. Leaders who gave consent to the trial then called a meeting of their community and the study staff explained the study in detail to the community. Without the participation and consent of the leaders, it would have been difficult to reach the individuals within these communities.
After information was given to the interested individuals, many consulted with their families on whether or not to participate. Some interested volunteers had to withdraw their consent because of refusal of their families to allow them to participate in the trial. Two of the 21 volunteers who agreed to participate asked that their families would not be made aware of their participation in the study.
Potential volunteers were all given written information about the background to the study, numbers of individuals vaccinated in the UK, expected side-effects, and potential risks to taking part.
Written, informed consent was obtained from all individuals at screening. The results of the screening blood tests were given to all volunteers and eligible participants were given a date for enrolment. All the volunteers were given copies of their laboratory results. All the research participants who were vaccinated had certificates given to them in a public ceremony such as the World TB Day 2004 or ceremonies at the MRC unit.
Reimbursement
Reimbursement of study participants is a difficult ethical issue in any research involving human beings, but is perhaps more so in the developing world where the potential for undue coercion is greatest. It is generally agreed that study participants should be paid or compensated as long as the inducement is not ?undue?. The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research report45 defined undue inducement as an excessive, unwarranted, inappropriate, or improper reward or other overture to obtain compliance. This report acknowledges that what might be acceptable in one setting could be considered undue influence in some vulnerable populations. The National Commission for the Protection of Human Subjects states that ?limiting remuneration to payment for time and inconveniences of participation and compensation for any injury resulting from participation is the best way to keep inducements due?. It has recently been suggested that, providing the research is ethical and presents a favourable risk-benefit ratio, there is no need to worry about undue coercion.46
In the UK studies with MVA85A, participants were compensated for their time at an hourly rate that was equivalent to the minimum wage. In addition, participants were compensated for their travelling expenses. In The Gambia, participants were reimbursed for their travel expenses and provided with a meal. The total amount reimbursed in The Gambia was approximately ?10 for nine visits over the course of the study period (6 months). The minimum wage in The Gambia is approximately ?7?9 per month. Furthermore, participants were eligible for free medical treatment at the MRC clinic (figure 2). In The Gambia, medical treatment is relatively expensive and so the provision of this facility is considered to be of great benefit to the volunteers, who might not otherwise be able to afford it.
Click to enlarge image
Figure 2. A routine clinic at the MRC laboratories, Fajara
Back to top
Results of screening and reasons for exclusion
228 volunteers were screened for eligibility for the studies. The panel shows the general inclusion and exclusion criteria for the trials. 23 (10%) of these individuals were eligible to participate and 21 eventually did so. Table 1 shows the reasons for exclusion, and table 2 shows the number of individuals excluded for one or more of these reasons. 11 people were vaccinated in the ?mycobacterially and BCG naive? group and ten were vaccinated in the ?BCG vaccinated? group.
Panel: General inclusion and exclusion criteria for the MVA85A vaccine studiesInclusion criteria
?Healthy adult male aged 18?45 years
?Normal medical history and physical examination
?Normal haematological and biochemical indices, negative HIV antibody test, and no serological evidence of HBV infection
?Normal chest radiograph
Exclusion criteria
?Positive Elispot response to ESAT6 or CFP10
?PPD Elispot greater than 100 spots per million peripheral blood mononuclear cells on ex-vivo Elispot assay
?BCG scar (for the first study only)
?Certain clinical conditions?eg, skin disorders (eczema, psoriasis, etc), allergy, immunodeficiency, cardiovascular disease, respiratory disease, endocrine disorder, liver disease, renal disease, gastrointestinal disease, or neurological illness
?History of splenectomy
?Haematocrit <30%, serum creatinine >130 mmol/L, serum ALT>80 IU/L
?Recent (1 month) blood transfusion
?Previous vaccination with any experimental poxvirus vaccine
?Recent (2 weeks) administration of any other vaccine or immunoglobulin
?Positive HIV or hepatitis B surface antigen test
?Recent (3 months) participation in another clinical trial
Table 1. Reasons for exclusion from the study
Table 2. Number of volunteers excluded for one or more reason
The most frequent reason for rejection was a strong PPD Elispot response, probably indicating the prevalence of exposure to environmental mycobacteria, as only 24% of the volunteers had responses to ESAT6 or CFP10. Approximately 65% of individuals screened had positive PPD Elispot responses, whereas only approximately 36% of them had a positive tuberculin skin test, illustrating the increased sensitivity of the PPD Elispot assay. A chest radiograph was done in all individuals, and anyone with an abnormal chest radiograph was referred to the general medical clinic at the MRC unit for further assessment. Antituberculous chemoprophylaxis was not given to anyone identified as being latently infected, since this is not the normal ?standard of care? in The Gambia.
The aim of these early studies was to obtain safety data on this vaccine in an endemic area. Future studies will be done in individuals more representative of the general population.
Back to top
Conclusions
The first clinical trials with the candidate tuberculosis vaccine MVA85A have now been successfully completed in a tuberculosis-endemic setting in west Africa. A detailed report on safety and immunogenicity will be presented elsewhere. Entry criteria were established that minimised the risks of any tuberculosis-specific Koch reaction occurring. Participants were stratified according to mycobacterial load. Once safety was demonstrated in the ?most mycobacterially naive group?, the next trial, which recruited individuals with prior BCG vaccination, was done.
Tuberculosis lags behind the other major pathogens of the developing world like HIV and malaria in terms of numbers of candidate vaccines evaluated in clinical trials. Any serious adverse event in a new tuberculosis vaccine trial would not only put back the development of that vaccine candidate, but would impact on the clinical progression of other tuberculosis vaccine candidates. As experience is gained with the new tuberculosis vaccines currently entering into early clinical trials, we will be able to set less stringent entry criteria for subsequent studies. However, these studies involved the first of the new generation vaccines to enter into clinical trials in a tuberculosis-endemic area, and MVA85A had been shown to induce strong cellular immune responses in the UK studies.17
Once safety has been demonstrated in early clinical trials, trials can then be done in individuals who are more representative of the population in tuberculosis-endemic areas, including people who are latently infected with M tuberculosis, individuals who are infected with HIV, and individuals who are coinfected with M tuberculosis and HIV.
Back to top
Search strategy and selection criteria
To ensure that there were no vaccines currently in clinical development of which we were not aware, and which might inform this article, we searched PubMed (1996 to March 2006) using the terms ?tuberculosis? and ?vaccination? and limited the search to clinical trials. We reviewed the 251 titles and, where appropriate, the abstract. Only English language papers were reviewed.
Back to top
Conflicts of interest
AH is a cofounder of, and consultant to, Oxxon Therapeutics Ltd and HM is a shareholder. None of the other authors have any conflicts of interest to declare.
Acknowledgments
The phase I vaccine trials with MVA85A in The Gambia were funded by an EU 5th Framework Programme grant number QLK2-CT-2002-01613, with additional support from the Wellcome Trust and the UK MRC. We thank the many contributors to this study in the unit in The Gambia including the participants, field workers, and Simon Donkor, who created the trial database.
Back to top
Early clinical trials with a new tuberculosis vaccine, MVA85A, in tuberculosis-endemic countries: issues in study design
Hannah B IbangaFWACP a, Roger H BrookesPhD a, Philip C HillFRACP a, Patrick K OwiafeHND a, Helen A FletcherPhD c, Christian LienhardtMD b, Adrian VS HillFMedSci c, Richard A AdegbolaFRCPath a and Dr Helen McShaneMRCP
c Summary
Introduction
New vaccine development
Tuberculosis specific issues
HIV infection: prevalence and acceptability of testing within a clinical trial
Trial awareness and recruitment
Ethical issues
Results of screening and reasons for exclusion
Conclusions
Search strategy and selection criteria
References
Summary
Tuberculosis remains a substantial global health problem despite effective drug treatments. The efficacy of BCG, the only available vaccine, is variable, especially in tuberculosis-endemic regions. Recent advances in the development of new vaccines against tuberculosis mean that the first of these are now entering into early clinical trials. A recombinant modified vaccinia virus Ankara expressing a major secreted antigen from Mycobacterium tuberculosis, antigen 85A, was the first new tuberculosis vaccine to enter into clinical trials in September 2002. This vaccine is known as MVA85A. In a series of phase I clinical trials in the UK, MVA85A had an excellent safety profile and was highly immunogenic. MVA85A was subsequently evaluated in a series of phase I trials in The Gambia, a tuberculosis-endemic area in west Africa. This vaccine is the only new subunit tuberculosis vaccine to enter into clinical trials in Africa to date. Here, we discuss some of the issues that were considered in the protocol design of these studies including recruitment, inclusion and exclusion criteria, reimbursement of study participants, and HIV testing. These issues are highly relevant to early clinical trials with all new tuberculosis vaccines in the developing world.
Back to top
Introduction
In 1993, WHO declared the current tuberculosis epidemic ?a global emergency?.1 More recently, in August 2005, WHO has again declared the tuberculosis epidemic in Africa an emergency situation.2 Worldwide, there are 8 million new cases and 2 million deaths each year from this disease.3 98% of these deaths occur in the developing world and countries of the former Soviet Union. The global prevalence of latent Mycobacterium tuberculosis infection is estimated to be 32% (1?86 billion people).3 Such individuals are at risk of reactivation of this latent infection should they become immunosuppressed for any reason. Globally, coinfection with HIV is the most important risk factor for progression of latent M tuberculosis infection to tuberculosis disease and the current HIV pandemic has fuelled the tuberculosis epidemic. The emergence of multidrug-resistant strains of M tuberculosis has further compounded the problem and made the need to control this disease even more urgent.
The best way to control the tuberculosis epidemic would be with an effective vaccine. BCG, the only available vaccine against M tuberculosis, is a live attenuated strain of Mycobacterium bovis that was first used in 1921.4 Since then it has become a well established part of the WHO Expanded Programme on Immunisation and is widely administered at birth throughout the developing world. When administered at birth it confers consistent and appreciable protection against disseminated disease, including tuberculous meningitis.5 However, adult pulmonary disease is responsible for the huge global public-health burden. The protection imparted by BCG against pulmonary disease has been evaluated in a number of large randomised controlled trials and observational studies, and has been shown to vary greatly, from 0 to 80%.4,6 Furthermore, a recent large cluster randomised trial estimating the effect of BCG revaccination in Brazilian school children who had been vaccinated with BCG at birth demonstrated that BCG revaccination did not confer any additional protection and should therefore not be recommended.7
There are several explanations for the variable efficacy in protection against pulmonary disease conferred by BCG. The one best supported by the available data is that exposure to environmental mycobacteria?eg, Mycobacterium avium, Mycobacterium marinum, and Mycobacterium intracellulare?interferes with BCG ?take?. There are two possible mechanisms for this interference: masking and boosting. First, exposure to environmental mycobacteria induces an antimycobacterial immune response, and subsequent vaccination with BCG does not increase or boost that response. Evidence of such masking comes from a series of studies done in parallel in adolescent school children in the UK and in Malawi.8 In the UK, baseline cellular immunity to mycobacterial antigens was low, and was increased after BCG vaccination. By contrast, in Malawi, the baseline cellular immune response to mycobacterial antigens was high and BCG vaccination did not result in a substantial incremental rise in these responses. The second possible mechanism for environmental mycobacterial interference is that antimycobacterial immunity induced by environmental mycobacteria abrogates the replication of BCG and subsequent induction of an immune response. This blocking hypothesis is supported by data from murine experiments.9 Importantly in this murine study, the protective efficacy of the subunit vaccines tested was not abrogated by prior exposure to environmental mycobacteria.
A second explanation for the variable efficacy of BCG is the age at administration. Trials done in young, mycobacteria-naive children have been generally positive. Other possible explanations for the variability in BCG efficacy across different trials include differences between BCG strains, host genetics, and nutritional factors.10?12 Several of these factors might contribute to the variability seen.
Back to top
New vaccine development
There is an urgent need for a better vaccination strategy. Since BCG does confer substantial protection against disseminated disease in childhood, ideally any new vaccination strategy should include BCG to retain this protective effect. Over the past decade there has been a resurgence of interest in the development of new tuberculosis vaccines and some of the most promising of these are now entering into early clinical trials.
M tuberculosis is an intracellular pathogen and protective immunity is entirely dependent on an intact cellular immune response.13 A Th1 type CD4 T-cell response is essential for protective immunity, and MHC class I restricted CD8+ T cells might also be required. Although a validated correlate of protection has not yet been identified, the secretion of interferon γ from antigen-specific T cells is essential for the protective immune response to this pathogen.14,15
There are two main strategies for developing an improved tuberculosis vaccine. The first is to develop a subunit vaccine (DNA, recombinant viral vector, or recombinant protein) expressing an immunodominant antigen(s) from M tuberculosis. Such vaccines are largely being developed to boost BCG rather than replace it. The second approach is to use the whole organism, either by developing a recombinant strain of BCG or by developing an attenuated strain of M tuberculosis. This strategy aims to replace BCG. Any candidate tuberculosis vaccine is evaluated in a series of animal models, and the most promising candidates are then considered for evaluation in early clinical trials. To date, only four such candidates have entered into clinical trials. The first, a recombinant modified vaccinia virus Ankara expressing antigen 85A known as MVA85A, entered into clinical trials in the UK in September 2002.16 Phase I studies have demonstrated that MVA85A is safe and induces high levels of cellular immunity when used to vaccinate BCG-naive adults. When used in BCG-primed adults, significantly higher levels of cellular immunity are induced, which persist for longer (p=0?015).17 In May 2003, when sufficient safety data was obtained from the UK studies, MVA85A entered into phase I clinical trials in The Gambia (figure 1). MVA85A is currently in phase II clinical trials in the Western Cape in South Africa.
Click to enlarge image
Figure 1. Members of the MRC Gambia vaccine trial team celebrate with the first person in Africa to be vaccinated with MVA85A at the end of his follow-up
The second vaccine to enter into clinical trials was a recombinant strain of BCG over-expressing antigen 85B, rBCG30.18 This vaccine entered into clinical trials in the USA in early 2004.19 The third vaccine that has entered into clinical trials is a recombinant fusion protein comprising the 32 and 39 kDa proteins from M tuberculosis, M tuberculosis 72F, used in conjunction with an adjuvant (AS02).20 This vaccine also entered into clinical trials in the USA in early 2004.21 In November 2005, a fusion protein expressing ESAT6 and antigen 85b entered into phase I studies in Europe.22 In addition to these four vaccines, there is an ongoing phase III efficacy trial of an inactivated whole-cell mycobacterial vaccine in HIV-infected individuals in Tanzania.23
Here, we discuss the issues that arose in the design, setting up, and implementation of the clinical trials with MVA85A in The Gambia.
Back to top
Tuberculosis specific issues
The Koch reaction
The Koch reaction describes the development of immunopathology in a person or animal with tuberculosis when an exaggerated immune response to M tuberculosis is stimulated.24 Such pathology occurs both at the site of infection and at the site of vaccination. In 1891, Koch observed that 4?6 weeks after the establishment of M tuberculosis infection in guineapigs, intradermal challenge with either the whole organism or culture filtrate resulted in necrosis and subsequent healing of the injection site and similar lesions in the original tuberculous lesion.25 In human beings, Koch found that injection of larger quantities of culture filtrate subcutaneously into tuberculosis patients evoked necrosis in established tuberculous lesions at distant sites, with disastrous results in the spine and lungs.
As the first new candidate tuberculosis vaccines enter into clinical trials, there is a concern that highly immunogenic vaccines could induce such a reaction. The available animal data suggest that such reactions correlate with bacterial load and are unlikely to occur in animals or human beings with latent infection, where the bacterial load is low.26 To minimise the risk of a Koch reaction occurring, we began the clinical trials with MVA85A in individuals with minimal pre-existing mycobacterial immunity.
The evaluation of pre-existing antimycobacterial immunity
For any new tuberculosis vaccine entering into clinical trials, it is necessary to determine the degree of pre-existing mycobacterial infection in individuals at screening. This assessment allows early trials to be done in individuals with minimal pre-existing exposure, and thus minimise the risks of inducing a Koch reaction. The clinical studies with MVA85A in the UK and in The Gambia were designed to allow for vaccination of patient groups sequentially with a step-wise increase in mycobacterial exposure to minimise this risk.16 In addition, studies were done sequentially in the UK and then in The Gambia, where exposure to both environmental mycobacteria and M tuberculosis is higher.27,28 In these studies, we used a combination of tuberculin skin testing, together with measurement of cellular immune responses using an ex-vivo interferon γ Elispot assay, to determine the pre-existing antimycobacterial immunity.
Tuberculin skin test
The tuberculin skin test is the standard test used throughout the world to determine the level of antimycobacterial immunity, and therefore mycobacterial infection. This test was first used in 1890, and is the oldest diagnostic test in clinical use.29 It involves an intradermal injection with tuberculin purified protein derivative (PPD). The size of the delayed-type hypersensitivity response is measured 2?3 days later. The biggest limitation of the tuberculin skin test is that PPD lacks specificity for M tuberculosis infection because many of the proteins present have homologues in a number of non-tuberculous mycobacteria. A positive tuberculin skin test response?ie, a Mantoux response of greater than 15 mm in an individual who has been BCG vaccinated, or greater than 10 mm in someone without BCG vaccination?can therefore result from infection with M tuberculosis, prior vaccination with BCG, or exposure to environmental mycobacteria.30,31 A further limitation of the tuberculin skin test is that it can lack sensitivity?eg, in immunosuppressed individuals.32 Despite these limitations, the tuberculin skin test remains a commonly used test in determining infection with M tuberculosis throughout the world.33
In-vitro T-cell assays: the ex-vivo interferon γ Elispot assay
In-vitro assays of cellular immunological function are increasingly being used in parallel with the tuberculin skin test to obtain greater sensitivity and specificity with respect to the level of pre-existing immunity to mycobacteria. These assays are based on the measurement of the secretion of interferon γ after stimulation with mycobacterial antigens.34 In the clinical studies with MVA85A, the ex-vivo interferon γ Elispot assay was used as the main cellular immunological assay.
In these early vaccine studies, we were interested in quantifying Elispot responses to both PPD and two M tuberculosis-specific antigens, ESAT6 and CFP10. The responses to these two antigens allowed the diagnosis and exclusion of individuals who were latently infected with M tuberculosis.35 ESAT6 and CFP10 are highly immunogenic and are among the early secreted antigens in M tuberculosis infection. Their genes are not present in any strain of BCG and homologues of these genes are absent from a number of relevant non-tuberculous mycobacteria.36?38 Individuals with active tuberculosis were excluded on the basis of clinical history, ESAT6 and CFP10 Elispot response, and chest radiograph findings. Furthermore, to identify individuals who were as mycobacterially naive as possible for these early studies, we set a relatively low threshold for the upper limit to the PPD Elispot response at screening. Individuals with a PPD response greater than 100 spot forming cells per million peripheral blood mononuclear cells were excluded from the study.
Elispot is one of the most sensitive cellular immunological assays available.39 The assay does not require expensive equipment, requires relatively low technology, and has been used in larger scale epidemiological studies and therefore has a potential role in larger scale vaccine studies.40 For these reasons, in the clinical studies with MVA85A, the ex-vivo interferon γ Elispot assay was used as the main cellular immunological assay. The assay was done at screening to determine pre-existing antimycobacterial immunity and repeated after vaccination to monitor vaccine-induced changes in immune responses.
Back to top
Recruitment of BCG-naive subjects: reliability of BCG scar detection
In the first clinical trial with MVA85A in The Gambia, the aim was to recruit individuals who had not been BCG vaccinated and who had minimal pre-existing mycobacterial exposure. BCG-vaccinated individuals were subsequently recruited once the safety of MVA85A had been demonstrated in BCG-naive individuals. BCG was not widely used in The Gambia until 1985. The most reliable method to determine if a person has been vaccinated with BCG is to examine the infant welfare card where vaccinations given in infancy are documented or the mothers are asked about this vaccination. In the absence of such source data, we had to rely on the presence or absence of a scar. Absence of scar is an imperfect indicator of whether or not someone has received BCG vaccination. In the UK, it has been reported that about 26% of individuals may not produce a scar 2 years after BCG vaccination.41 A study in Malawi showed that BCG scar was a highly sensitive indicator of vaccination status when vaccine was properly handled and given at over 3 months of age, but in infants under 1 month of age vaccinated in health centres, sensitivity of BCG scar had declined to under 80% by 4 years post-vaccination.42 Despite these limitations, presence of a BCG scar was an exclusion criterion in the BCG-naive study.
Back to top
HIV infection: prevalence and acceptability of testing within a clinical trial
In The Gambia, the prevalence of HIV-1 infection is 1?0%, and the prevalence of HIV-2 infection is 0?8%.43 There is a high level of public awareness about HIV infection and people are encouraged to come for voluntary counselling and testing. However, the stigma surrounding HIV is still high. Full pretest and post-test counselling was provided for all of the individuals screened for this study. Individuals were informed that they would be tested for HIV, but had the option of not being told what the result was. In a Zambian study in which 475 people were HIV screened, 41 (8?6%) people did not come back for their results,44 including six people who did not want to know the results, and others who gave reasons such as distance from the centre, death, or illness.
In total, 228 people were screened for the MVA85A vaccine trials, and there were no positive HIV results, indicating the low HIV prevalence in this population. If any individuals had tested positive for infection with HIV, they would have been referred to the local genitourinary medicine clinic within the Medical Research Council (MRC) unit, where they would be provided with ongoing medical care and considered for free antiretroviral therapy. In the studies with MVA85A, none of the study participants objected to having an HIV test, after suitable counselling. This study demonstrates the acceptability of HIV testing in this study population, although this may be different elsewhere.
Back to top
Trial awareness and recruitment
Before the study started, officers in the government Departments of State for Health and Social Welfare were informed about the trials. Meetings with community and religious leaders followed these visits to sensitise the community about the trial. In conjunction, there was a media campaign educating the population on tuberculosis. The media campaign included radio and television programmes in the major local languages and English. These programmes were both live phone-in and educational programmes. Media releases were done in the major newspapers. Additionally, the study team distributed T-shirts and caps printed with calls for people to be involved in the tuberculosis vaccine trial.
At the mosques or institutions visited, the people present were encouraged to visit the MRC and obtain more information on the trial. Others responded to the media campaign by either making phone calls or coming to the MRC unit. At the MRC unit, more detailed information was given on the trial. Information sheets, in English, were translated verbally into the relevant local language (Wolof or Mandinka).
Back to top
Ethical issues
Consent
Informed consent is an essential component of any research involving human patients. This consent must be from the individual taking part in the research, but could also involve the family or community in which that individual lives. In The Gambia, the family and community might be involved in the decision to participate, as well as the individual. Community awareness meetings conducted in the mosques and educational institutions enabled discussion of the rationale and study requirements with the community and religious leaders. Leaders who gave consent to the trial then called a meeting of their community and the study staff explained the study in detail to the community. Without the participation and consent of the leaders, it would have been difficult to reach the individuals within these communities.
After information was given to the interested individuals, many consulted with their families on whether or not to participate. Some interested volunteers had to withdraw their consent because of refusal of their families to allow them to participate in the trial. Two of the 21 volunteers who agreed to participate asked that their families would not be made aware of their participation in the study.
Potential volunteers were all given written information about the background to the study, numbers of individuals vaccinated in the UK, expected side-effects, and potential risks to taking part.
Written, informed consent was obtained from all individuals at screening. The results of the screening blood tests were given to all volunteers and eligible participants were given a date for enrolment. All the volunteers were given copies of their laboratory results. All the research participants who were vaccinated had certificates given to them in a public ceremony such as the World TB Day 2004 or ceremonies at the MRC unit.
Reimbursement
Reimbursement of study participants is a difficult ethical issue in any research involving human beings, but is perhaps more so in the developing world where the potential for undue coercion is greatest. It is generally agreed that study participants should be paid or compensated as long as the inducement is not ?undue?. The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research report45 defined undue inducement as an excessive, unwarranted, inappropriate, or improper reward or other overture to obtain compliance. This report acknowledges that what might be acceptable in one setting could be considered undue influence in some vulnerable populations. The National Commission for the Protection of Human Subjects states that ?limiting remuneration to payment for time and inconveniences of participation and compensation for any injury resulting from participation is the best way to keep inducements due?. It has recently been suggested that, providing the research is ethical and presents a favourable risk-benefit ratio, there is no need to worry about undue coercion.46
In the UK studies with MVA85A, participants were compensated for their time at an hourly rate that was equivalent to the minimum wage. In addition, participants were compensated for their travelling expenses. In The Gambia, participants were reimbursed for their travel expenses and provided with a meal. The total amount reimbursed in The Gambia was approximately ?10 for nine visits over the course of the study period (6 months). The minimum wage in The Gambia is approximately ?7?9 per month. Furthermore, participants were eligible for free medical treatment at the MRC clinic (figure 2). In The Gambia, medical treatment is relatively expensive and so the provision of this facility is considered to be of great benefit to the volunteers, who might not otherwise be able to afford it.
Click to enlarge image
Figure 2. A routine clinic at the MRC laboratories, Fajara
Back to top
Results of screening and reasons for exclusion
228 volunteers were screened for eligibility for the studies. The panel shows the general inclusion and exclusion criteria for the trials. 23 (10%) of these individuals were eligible to participate and 21 eventually did so. Table 1 shows the reasons for exclusion, and table 2 shows the number of individuals excluded for one or more of these reasons. 11 people were vaccinated in the ?mycobacterially and BCG naive? group and ten were vaccinated in the ?BCG vaccinated? group.
Panel: General inclusion and exclusion criteria for the MVA85A vaccine studiesInclusion criteria
?Healthy adult male aged 18?45 years
?Normal medical history and physical examination
?Normal haematological and biochemical indices, negative HIV antibody test, and no serological evidence of HBV infection
?Normal chest radiograph
Exclusion criteria
?Positive Elispot response to ESAT6 or CFP10
?PPD Elispot greater than 100 spots per million peripheral blood mononuclear cells on ex-vivo Elispot assay
?BCG scar (for the first study only)
?Certain clinical conditions?eg, skin disorders (eczema, psoriasis, etc), allergy, immunodeficiency, cardiovascular disease, respiratory disease, endocrine disorder, liver disease, renal disease, gastrointestinal disease, or neurological illness
?History of splenectomy
?Haematocrit <30%, serum creatinine >130 mmol/L, serum ALT>80 IU/L
?Recent (1 month) blood transfusion
?Previous vaccination with any experimental poxvirus vaccine
?Recent (2 weeks) administration of any other vaccine or immunoglobulin
?Positive HIV or hepatitis B surface antigen test
?Recent (3 months) participation in another clinical trial
Click to view table
Table 1. Reasons for exclusion from the study
Click to view table
Table 2. Number of volunteers excluded for one or more reason
The most frequent reason for rejection was a strong PPD Elispot response, probably indicating the prevalence of exposure to environmental mycobacteria, as only 24% of the volunteers had responses to ESAT6 or CFP10. Approximately 65% of individuals screened had positive PPD Elispot responses, whereas only approximately 36% of them had a positive tuberculin skin test, illustrating the increased sensitivity of the PPD Elispot assay. A chest radiograph was done in all individuals, and anyone with an abnormal chest radiograph was referred to the general medical clinic at the MRC unit for further assessment. Antituberculous chemoprophylaxis was not given to anyone identified as being latently infected, since this is not the normal ?standard of care? in The Gambia.
The aim of these early studies was to obtain safety data on this vaccine in an endemic area. Future studies will be done in individuals more representative of the general population.
Back to top
Conclusions
The first clinical trials with the candidate tuberculosis vaccine MVA85A have now been successfully completed in a tuberculosis-endemic setting in west Africa. A detailed report on safety and immunogenicity will be presented elsewhere. Entry criteria were established that minimised the risks of any tuberculosis-specific Koch reaction occurring. Participants were stratified according to mycobacterial load. Once safety was demonstrated in the ?most mycobacterially naive group?, the next trial, which recruited individuals with prior BCG vaccination, was done.
Tuberculosis lags behind the other major pathogens of the developing world like HIV and malaria in terms of numbers of candidate vaccines evaluated in clinical trials. Any serious adverse event in a new tuberculosis vaccine trial would not only put back the development of that vaccine candidate, but would impact on the clinical progression of other tuberculosis vaccine candidates. As experience is gained with the new tuberculosis vaccines currently entering into early clinical trials, we will be able to set less stringent entry criteria for subsequent studies. However, these studies involved the first of the new generation vaccines to enter into clinical trials in a tuberculosis-endemic area, and MVA85A had been shown to induce strong cellular immune responses in the UK studies.17
Once safety has been demonstrated in early clinical trials, trials can then be done in individuals who are more representative of the population in tuberculosis-endemic areas, including people who are latently infected with M tuberculosis, individuals who are infected with HIV, and individuals who are coinfected with M tuberculosis and HIV.
Back to top
Search strategy and selection criteria
To ensure that there were no vaccines currently in clinical development of which we were not aware, and which might inform this article, we searched PubMed (1996 to March 2006) using the terms ?tuberculosis? and ?vaccination? and limited the search to clinical trials. We reviewed the 251 titles and, where appropriate, the abstract. Only English language papers were reviewed.
Back to top
Conflicts of interest
AH is a cofounder of, and consultant to, Oxxon Therapeutics Ltd and HM is a shareholder. None of the other authors have any conflicts of interest to declare.
Acknowledgments
The phase I vaccine trials with MVA85A in The Gambia were funded by an EU 5th Framework Programme grant number QLK2-CT-2002-01613, with additional support from the Wellcome Trust and the UK MRC. We thank the many contributors to this study in the unit in The Gambia including the participants, field workers, and Simon Donkor, who created the trial database.
Back to top
</TD><TD class=sibtbg>
The later strains were selected because they had the least side effects but maybe the efficiency has become less and les 
Comment