Print page

Phase II Workshop

Les Etangs de Corot, Ville d'Avray 13-14 September 2004

Attendees:

Louis Moulineaux
Jean-Francois Trape, IRD
Christophe Rogier, IMTSSA
Pierre Druile, Paris Pasteur Institute
Brian greenwoord, LSTHM
Kevin Baird, US Navy
Robert Sauerwein, UMC-St Radboud
Aissatou Balde Touré, Dakar Pasteur Institute
Christian Roussilhon, Paris Pasteur Institute
Roma Chilengi, AMANET
Zarifah Reed, IVR-WHO
Blaise Genton, Swiss Tropical Institute
Sodiomon Sirima, CNRFP, Burkina Faso
Kevin Marsh, MRC-Kenya
Søren Jepsen, EMVI

Appologies
Rose Leke
Peter kremsner

Agenda:

  1. What is known about the patterns of infection? Data from various age groups and settings - Chair, Kevin Marsh
  2. What are the patterns of immune response to various potential vaccine antigens in various age groups and various epidemiological settings? What should be considered in assessing the protective immune response to blood stage vaccine? - Chair, Brian Greenwood
  3. Clinical development rationale - Chair, Robert Sauerwein
    Phase I
       Naive volunteers
       Exposed volunteers
    Phase II
       IIa: rationale for blood stage vaccine?
       IIb: rationale for the selection of the trial population; definition of evaluation criteria including duration of follow up.
  4. Summary and Conclusions

Welcome and Introduction

Søren Jepsen opened the meeting by welcoming the attendees. He then provided a brief summary of the EMVI portfolio, which includes several blood stage vaccines, one of which has already been tested in a malaria endemic population, and others that are approaching this stage in their evaluation. He reminded those attending the meeting of the potential of an effective malaria vaccine delivered through the EPI programme to reach all sections of the population, including the poorest.

Session 1: Chair Kevin Marsh

What is known about the patterns of infection? Data from various age groups and different settings?

Pierre Druilhe raised concerns that testing a blood stage vaccine into a population whose immune system had already been directed in a particular direction, as a result of previous natural exposure to malaria, could result in a negative result and the consequent rejection of a vaccine candidate that could be effective in other situations. He presented his views on the optimum epidemiological characteristics of a site for the testing of a blood stage vaccines. These included: a population with:

  1. some degree of anti-toxic immunity,
  2. an absence of anti-parasite immunity,
  3. a sufficiently high level of infection to allow a vaccine trial of reasonable size to be conducted.

He expressed the view that African children are not the best target for testing blood stage vaccines as they do not have the epidemiological characteristics described above, and that trials in less endemic areas, e.g. parts of Asia would be better (PD presentation).

During the discussion it was pointed out that it would be difficult to find a site where there had been no priming of the immune system, but where the attack rate was high enough to allow a vaccine trial to be done – compromises would have to be made. In later discussions the possibility of using migrant (including military) populations moving to endemic areas or populations at risk of epidemics was raised, but the practical difficulties of doing trials in these populations would be large, and it is uncertain how much more useful than an experimental challenge study such a trial would be.

Data were presented from a number of sites on the incidence of infection in the first year of life. In some situations only a relatively small proportion of the population has experienced an attack of malaria or an asymptomatic infection by the age of one year – 60% in a population in PNG, and these young children might meet the desired criteria for a vaccine study. However, there are complications over vaccination at the time that EPI vaccines are given.

Louis Molineaux presented data from the studies of malariotherapy1, which demonstrates that one infection lowered the likelihood of fever during subsequent infections, as well as the maximum parasite density, indicating that even one infection does influence the outcome of subsequent exposure, a view supported by the fact that very few infections are effective at preventing an African child from dying from malaria. As mentioned above, it will be difficult to find a population where the attack rate is high enough to conduct a vaccine trial, and in whom the immune system has not already been primed (or down-regulated) by prior exposure. Later in the meeting Louis Molineaux also presented data from the malariotherapy trials on variations in parasite growth rate, including the impact of strain (see Kevin Marsh’s summary).

Christian Roussilhon presented data, from an area of Senegal with seasonal malaria transmission, on the effects of season of birth, on age at first attack of malaria, and on parasite density during such attacks. No effect was seen during first, second or third attack.

Kevin Marsh raised the issue of fever threshold, and presented data derived using the Smith and Armstrong Schellenberg method2 for several populations living under different levels of transmission. Reasonably consistent results had been found – the threshold was low in those under one year, increased in those aged one – ten years and then declined. Data from Dielmo, Senegal presented by Jean-François Trape are consistent with this picture. In this population nearly all infants who became parasitaemic were symptomatic regardless of the level of the infection. The fever threshold is likely to be used to define clinical attacks of malaria in blood stage vaccine trials. A different value may be found in vaccinated and control groups, but it is unlikely that regulatory authorities would allow different definitions to be used in each arm of the trial, and that they will require a standard definition. Substantial attention has been paid to the importance of using a fever threshold in evaluating malaria interventions. However, using the protective effect of the genotype AS as an example, Kevin Marsh showed that use of different levels of parasite density to define a clinical attack of malaria had little effect on measurement of protective efficacy. Limited data from vaccine trials has shown a similar effect. Thus, the importance of parasite density in definition of a clinical attack of malaria, which should include a history of fever as well as observed fever, may have been over emphasised.

Session 2: Chair Brian Greenwood

What are the patterns of immune response to various potential vaccine antigens in different age-group and different epidemiological settings?

What should be considered in assessing the protective immune response to blood stage-vaccine?

Section 1: Continuation of comparison of incidence of clinical malaria under different transmission settings

Three further data sets were presented for young children under different transmission settings. Christophe Rogier presented data from Dielmo indicating that under an EIR of approx. two hundred, around 20% of children remained parasite negative (on the basis of weekly blood slides) after three hundred days of follow up. It was estimated that there was one clinical episode per sixty two infected bites, and that after two hundred infected bites, 80% of children had become positive on at least one slide (see Christophe Rogier’s presentation).

Kevin Baird indicated that in Northern Ghana under very high transmission conditions in a cohort of children six – twenty four months, the cumulative parasite incidence was 94% by nine weeks.

Blaise Genton presented data from Papua New Guinea under an EIR of between fifteen and forty indicating that by four hundred days in a cohort of one hundred and fifty infants, only 20% had experienced a clinical attack, though the cumulative parasite equivalents had reached 50% by two hundred days and 100% by four hundred days.

Conclusions

Taken together with data presented in the first session, it is clear that under high transmission settings, it is unlikely that it would be possible to identify a cohort of young children without considerable exposure to malaria and experience of clinical malaria. However, it seems likely that under low to moderate transmission conditions, a significant proportion of children (possibly between 25% and 50%) may reach their first birthday without having experienced an obvious clinical attack of malaria. Comparisons are hindered by slightly differing methodolgies, and if it is considered important to establish the exact degree of malaria experience, it will be necessary to conduct standardised studies, which incorporate active and passive clinical surveillance together with immunilogical measurements.

Section II: Modelling malariatherapy data

Louis Molineaux presented data from an analysis of malariatherapy data. It was noted that during malariatherapy it was rarely if ever necessary to give treatment on clinical grounds to patients with P. falciparum infection, other than in the first wave of parasitemia in any induced infection. It also appears, from limited numbers, that treatment on clinical grounds was rarely if ever necessary during any part of subsequent induced infections. Maximum morbidity therefore seems to be associated with the first wave of parasitisation, and analysis was restricted to these data. The kinetics of raise and fall of the parasitemia in the first wave were analysed for a hundred patients. A striking observation was the variability in the exponential multiplication rates, and these were related to the eventual maximum parasitemia achieved before parasite densities declined. The variation in multiplication rates appeared to be mostly related to individual patient variation, rather than to being a feature of the parasite. The data were modelled to consider the effects of innate parasite clearance mechanisms and subsequent acquired parasite clearance mechanisms. The innate ability to control the initial rise in parasite density is a critical factor limiting the eventual height of parasitisation.

Conclusions

  1. these data support the idea that delaying (reducing)parasite multiplication is the key aim of an anti-blood stage vaccine,
  2. individuals with the lowest innate response (per parasite) may be at the highest risk of morbidity,
  3. further modelling indicates that it may be easier to limit severe disease than to reduce the actual incidence of clinical symptoms per se by vaccination, which enhances the acquired immune response.

Section III: Discussion on whether it is possible to identify individuals with little malaria experience, but at high risk of clinical attacks

Pierre Druilhe presented data from Burma indicating that adolescents and young adults with relatively little evidence of previous exposure to malaria may be exposed to a clinical attack rate of around about one attack per year. Kevin Baird reported that this level of exposure was insufficient to lead to immunity in a cohort of individuals followed in Indonesia. However, in discussion it was clear that this was reflected in the failure to reduce the incidence of clinical attacks over a period of three years, and it was agreed that this was compatible with the experience of young African children in endemic conditions, who may not achieve the ability to limit clinical attacks over this period, while still developing substantial immunity to severe disease and death. Brian Greenwood presented data from the Amani mountains in Tanzania indicating that it was possible to define populations with continuous variation of exposure by altitude, these populations allow the possibility of dissecting the relative contribution of malaria exposure and age. It was agreed that the optimum populations with minimal exposure but high attack rates were to be found either in migrant populations, epidemic situations, or situations in which non-immunes such as military personnel are deployed in an endemic situation. However, it was considered that the logistic and ethical difficulties of working such populations would make vaccine studies extremely difficult.

Section IV: Immunological data and clinical attacks in Dielmo

Aissatou Balse Touré presented data on the dynamics of the immune response in a cohort of over fifty children followed up to eight years. Data on responses over time in individuals to MSP3, LSA3, R23, GLURP and crude parasite extract were presented. Responses to MSP3 showed high levels of response (IgG and IgM), which continued to rise over time (IgG). Other antigens also showed well established responses by the age of three, and in the case of LSA3, R23 and GLURP, the rise over time was less obvious. In the case of R23, responses declined over time. In an analysis of the relationship between antibody responses to individual antigens and experience of clinical malaria, responses to all antigens including the crude extract, were associated with significant protection ranging between 21% and 45%. A striking feature of the data was the considerable individual variation in responses to different antigens. The third analysis indicated that association with clinical protection was limited to IgG1 and 3 in the case of MSP3 and GLURP, to IgG1 in the case of R23 and IgG2 in the case of LSA3.

Section V: Requirements for a blood stage vaccine

In a discussion of what were the fundamental properties required of a blood stage vaccine, it was agreed that all such vaccines should limit parasite multiplication rates. It was therefore considered whether or not induced infection in non-immunes under tightly controlled conditions offered a critical screening process for potential vaccines. In principle it was agreed that such a screen would be desirable, and that a positive result (i.e. limitation of parasite multiplication rates and therefore delay to parasite appearance) would be a strong indication for proceeding with clinical trials of a vaccine candidate. However, it was accepted that this system has not yet been validated in relation to natural infection, and is limited in certain respects (the reliance on a single parasite strain and probably unnatural levels of challenge). It was therefore felt that at this stage, induced infections in non-immunes probably should not represent a GO / NO GO decision point, but that it is highly desirable to continue to gain experience with the use of this approach, and in particular to try to validate results in relation to subsequent clinical trials.

Session 3: Chair Robert Sauerwein

Clinical development rationale

Phase I

  • Naive volunteers
  • Exposed volunteers

Phase II

  • IIa: rationale for blood-stage vaccine?
  • IIb: rationale for the selection of the trial population; definition of evaluation criteria, including duration of follow-up

Robert Sauerwein emphasised that the objective of this session was to review and define the critical criteria pertaining to the various stages of clinical development of blood stage vaccines. For Phase Ia and Ib it was, besides the overriding importance of safety assessment, important to define for any vaccine, in which direction the immunisation should drive the immune response. Blaise Genton raised the issue of whether we really know what we want in terms of quality and quantity of the immune response. For certain antigens it appears that cytophilic antibodies are essential, but that this is not likely to be the case for all antigens. It is probably desirable to have a high titer of antibodies after immunisation, but an assessment of the quality, the functionality of these antibodies should be conducted. Brian Greenwood emphasised that for most vaccine candidates we have functional in vitro assays, GIA, ADCI, and although they may not have been sufficiently validated yet, they should serve as guidance for prioritising further clinical development. Kevin Marsh supported this point, arguing that with the several MSP1 vaccine constructs under development, it is essential to have a means of prioritising the many options. The influence of prior exposure to malaria on the immune response to a vaccine, and the need to differentiate the immune response as a result of prior exposure and the maturity of the immune system in infants was discussed at length. Christophe Rogier maintained that previous exposure, even ten years prior, might influence the immune response. It was noted that if an exposed population in a Ib trial has a different or marked inferior response compared to the response in a Phase Ia in a naïve population, it should raise concern and a re-assessment of the vaccine formulation.

Robert Sauerwein distributed a publication3 on the design and conduct of artificial challenge of Dutch non-immunised volunteers, and the place of this model in clinical development was debated briefly. The general opinion was that for vaccines like MSP1, AMA1 and others, which are believed to exert their protection via antibodies directly inhibiting parasite invasion of red blood cells, a Phase IIa challenge would be appropriate, provided adequate immune response quantitively and qualitatively in volunteers was elicited, preferably after a positive outcome of a Phase Ia in vitro GIA. It was, however, less obvious whether antigens thought to exert an indirect effect via an ADCI like mechanism would be suited for a Phase IIa challenge. It might very well be that parsitemia must reach a critical or threshold level before an ADCI like mechanism takes effect in vivo. This threshold could well be above the gold standard for treatment of volunteers, i.e. thick smear positivity.

Phase IIb trials were then discussed, and there were divergent opinions as to whether efficacy trials should initially be conducted in adults (where possible), or whether after age de-escalation, efficacy studies should only be conducted in toddlers and infants. Pierre Druilhe maintained that it was important to avoid co-stimulation of the immune system by existing parasitemia. Blaise Genton pointed out that in the Papua New Guinea study of combination B and an earlier Spf66 study, impairment of the immune response to the vaccine of pre-existing parasitemia was not demonstrated. Kevin Marsh recommended a stratified study with children treated prior to immunisation, and another arm with no prior treatment, to demonstrate whether or not existing asymptomatic parasitemia is a problem (this will also have implications for the practical application of any vaccine). Likewise, the possible unwanted immunological effect of concomitant worm infection could be solved by stratifying a child population to plus minus prior or concomitant de-worming. There was general agreement that presumed interference with the immune response to immunisation from concomitant infections/infestations etc. should be studied.

There was general agreement that the primary end point in the assessment of efficacy of blood stage vaccines would be clinical attacks of malaria. A clinical attack of malaria is the consequence of a new infection. A secondary end point could be anaemia, if the children were immunised during the dry season before transmission. It has been recorded in Ghana that 24% of children, age twelve – twenty four months, suffered from severe anaemia after the end of the rainy season, whereas only 2% of the children suffered from anaemia at the start of the rainy season, i.e. anaemia could be a good biological marker. Coming back to the issue of importance of prior exposure to malaria, it was suggested that, not only could one assess vaccine immunogenicity and subsequent efficacy in a "horizontal manner" by moving from high to low transmission areas, but that same could be studied in a "vertical manner" with genetically homogenious populations living at various levels above sea level. The results of the latter design might be more influenced by migration than the horizontal design.

Kevin Marsh emphasised that the end point should be reduction of expected clinical episodes. The sensitivity and specificity at various parasite densities in defining clinical attacks of malaria were debated. Kevin Marsh was of the opinion that the actual medium resting parasite density, instead of the number of positive children in a given population, would be more predictive of the risk of severe disease in the following months (results from six surveys). Christian Roussilhon emphasised that high specificity is essential to ensure a study with enough statistical power. Blaise Genton mentioned that the Phase IIb trial in Papua New Guinea was designed as large safety trial with exploration of the biological effects of immunisation. Robert Sauerwein summed up this session by concluding that:

Primary end point is reduction in the expected number of clinical malaria episodes, which might be reflected in the delay of first episodes after immunisation, or the number of recorded clinical episodes over a pre-defined period – the rainy season? – two rainy seasons? Secondary end points would be the reduction of anaemia, see above. Secondary end points could also be reduced parasite rate, or possibly even better reduced mean parasite density, and in public health terms, reduced work load on physicians/dispensaries, and reduced hospitalisation

The following specific studies were recommended:

  1. Immunogenicity study with and without concomitant worm infestation.
  2. Improved knowledge of the value of surrogate markers, and standardisation of these.
  3. Effect of asymptomatic parasitemia on immune response to the vaccine.
  4. Phase Ib studies in adults and children in high versus low transmission areas (age and exposure critical studies), vertical or horizontal trials.
  5. Optimisation of Phase IIa methodology.
  6. Validation of IIa trials after successful Phase IIb trials.
  7. Comparison of the safety profile in Phase Ia and Ib trials to be undertaken by WHO.

References:

  1. Molineaux L, Trauble M, Collins WE, Jeffrey GM, Dietz K, Malaria therapy reinoculation data suggests individual variation of an innate immune response and independent acquisition of antiparasite and antotoxic immunities. Trans R Soc Trop Med Hyg. 2002 Mar-Apr96(2):205-9
  2. Smith T, Schellenberg JA, Hayes R. Attributable fraction estimates and case definition for malaria in endemic areas. Stat Med. 1994 Nov 30;13(22):2345-58
  3. Hermsen CC, de Vlas SJ, van Gemert GJ, Telgt DS, Verhage DF, Sauerwein RW. Testing vaccines in human experimental malaria: statistical analysis of parasitemia measured by a quantitive real-time polymerase chain reaction. Am J Trop Med Hyg. 2004 Aug.71(2):196-201