A strategy for accelerating the development of preventive AIDS vaccines
Excler, Jean-Louisa; Rida, Wasimab; Priddy, Francesa; Fast, Patriciaa; Koff, Waynea
From the aInternational AIDS Vaccine Initiative, New York, USA
bIndependent Consultant, Arlington, Virginia, USA.
Received 10 April, 2007
Revised 27 May, 2007
Accepted 15 June, 2007
Correspondence to Patricia Fast, International AIDS Vaccine Initiative, 110 William Street, 27th Floor, New York, NY 10038-3901, USA. E-mail: email@example.com
An effective vaccine against HIV-1 is urgently needed. Many candidate vaccines are currently in or are approaching clinical trials. The selection of vaccines for advanced development is now based on safety and immunogenicity in humans and on the ability of analogous vaccines to protect macaques against SIV. In the next few years, however, many candidates will complete phase II testing and must be prioritized, because phase III efficacy trials are costly, lengthy, and the capacity to conduct them is limited. Given the complexity of HIV pathogenesis and the unconfirmed predictive value of vaccine-induced immune responses in non-human primate challenge studies, AIDS vaccine clinical trials conducted in at-risk populations with clinically relevant endpoints may be the best way to judge the true biological significance of human immune responses to vaccines as well as guiding, accelerating, and prioritizing efficient vaccine development. Phase IIB test of concept trials have recently been proposed to evaluate vaccine candidates for preliminary evidence of efficacy in trials smaller than pivotal efficacy trials [1–3]. The co-primary endpoints are the prevention of HIV infection and a reduction of viral load in vaccinated individuals who acquire HIV . Phase IIB test of concept trials are still large (numbers approximately 2500–8500) and time-consuming (3.5–5 years).
None of the current vaccine candidates is likely to prevent infection completely, because they do not induce potent neutralizing antibodies against primary HIV isolates [4,5]. The failure to protect by antibodies against the HIV envelope may be the result of HIV antigenic variation, a lack of neutralization potency, or additional factors . Most of the current candidate vaccines  are designed to induce cell-mediated immunity, which is expected to decrease the HIV viral load in vaccinated individuals who become infected, thereby potentially slowing or preventing progression to AIDS and decreasing infectiousness [8–11]. In HIV-infected individuals, the time to key clinical events in HIV-1 disease progression was significantly longer for those with viral load setpoints 0.5–1.25 log10 copies/ml lower than the reference group [12,13]. Studies in non-human primates have demonstrated that vaccine-induced viral load reduction accompanies increased survival . By quantifying the anticipated clinical benefits of reducing the viral load at setpoint, Gupta and colleagues  have supported the use of virological end points in HIV-1 vaccine trials.
Given the hypothesis that the primary effect of current vaccines will be to reduce viral load, we propose a more efficient strategy for determining the value of various vaccine approaches: screening test of concept (STOC) trials. The primary endpoint of a STOC trial is plasma HIV-RNA viral load at setpoint in individuals who acquire HIV. Approximately 30 total HIV infections are required to detect a 1.0 log10 reduction in viral load with adequate statistical power. Data from these trials can demonstrate biological plausibility that a cell-mediated immunity-based vaccine has a meaningful effect, although if the reduction were only 1.0 log10, the next step might be to begin an interactive process, improving the vaccine and conducting another trial. Along with evidence of efficacy in animal models, safety and immunogenicity data, and feasibility of vaccine manufacture and distribution, these trials could serve to accelerate promising candidates to phase III trials, or limit the use of resources on less effective candidates.
The basic STOC trial design is a two-arm, double-blind, randomized study comparing a single vaccine candidate with placebo. The primary endpoint will be the plasma HIV-RNA viral load at setpoint for participants who become HIV infected.
Participants will be tested for HIV-1 infection every 3 months after vaccination. For individuals who become infected, viral load will be measured at two timepoints, approximately 3–6 months after the estimated date of HIV infection (defined as the midpoint between the last negative HIV test and first positive HIV test) with the geometric mean of the measurements used as the viral load setpoint. Infected participants will be offered enrollment into a long-term follow-up study to assess trends in such postinfection outcomes as viral load, CD4 T-cell count, and disease progression.
The null hypothesis of no difference in viral load at setpoint between infected vaccine and placebo recipients will be tested against the alternative hypothesis that vaccination reduces viral load using the Wilcoxon rank sum test. With 30 total infections, a one-sided 0.05 level test will have better than 80% power to detect a decrease in viral load when the true decrease on the logarithmic scale is 1.0 log10. Power calculations assume that vaccination has little or no effect on HIV acquisition. Also, the HIV infection and viral load endpoints are assumed to be independent, such that the effect of vaccination on incident infection does not cause a selection bias on viral load.
Table 1 displays the required sample size for a STOC trial in which 30 HIV infections are expected during the postvaccination follow-up period for various incidence rates and postvaccination follow-up periods. We recommend enrolling sufficient participants so that 18 months of postvaccination follow-up would suffice to complete the trial in 3 years. For a population with a 3% annual incidence rate, approximately 643 individuals would need to be enrolled.
The basic STOC trial design can be extended to include two vaccine candidates, to select the better candidate for further testing. For each vaccine candidate, the Wilcoxon rank sum test determines if there is a significant reduction in viral load compared with placebo. For those candidates with a significant reduction in viral load, we prioritize the candidate with the smaller mean, other factors being equal. As extreme values for viral load may unduly influence the mean, we may consider calculating trimmed means whereby outlying values are excluded.
A three-arm STOC trial could also assess the need for a prime-boost regimen. In this trial, volunteers would be randomly assigned in a 1: 1: 1 ratio to a prime-boost regimen, a placebo prime plus active boost regimen, or a placebo prime plus placebo boost regimen. With approximately 15 infections in each arm, there is a 90% chance of selecting the correct regimen when the true difference in mean viral load is at least 0.5 log10 and the standard deviation is no more than 1.0 log10 . A STOC trial only has adequate power to demonstrate superiority of the prime-boost regimen if the true difference in viral load is at least 1.0 log10. Also, a STOC trial is not likely to have adequate power to demonstrate the non-inferiority of the boost-only regimen unless the variation in viral load is small.
The STOC trial design is similar to that of a phase IIB test of concept trial, but accrues approximately 50% of the statistical information (number of HIV infections observed). Unlike the phase IIB test of concept trial, the STOC trial treats HIV infection as a secondary endpoint. If substantial heterogeneity in the circulating virus or the immune response to vaccination is anticipated, neither a STOC trial nor a simple phase IIB test of concept may have adequate power. An example of a phase IIB test of concept trial designed to account for anticipated heterogeneity in response is the STEP trial of Adenovirus 5 (Ad5) vaccine, conducted by Merck in collaboration with the HIV Vaccine Trials Network [1,17]. This trial is designed to detect a 60% reduction in the risk of HIV infection or a 1.0 log10 reduction in viral load at setpoint among vaccinated participants who become infected, stratified by low or high preexisting immunity to the Ad5 vector. It will capture approximately 50–60 incident infections per stratum. Of note is the fact that the Merck STEP trial plans an interim analysis after 30 HIV infections occur in the low Ad5 titer group . If a positive result is observed, the vaccine could be advanced to a phase III efficacy trial approximately 9–15 months sooner. To declare a positive result in an interim analysis, however, the effect of vaccination on HIV infection or viral load would have to be significant at a level much less than 0.05.
Given that the current cell-mediated immunity-based vaccines do not induce potent neutralizing antibodies, STOC trials can test more limited hypotheses, that of biological plausibility, by measuring the effect of vaccination on the viral load setpoint, which is postulated to predict disease progression [15,17,18] and secondary transmission [19,20]. Therefore, STOC trials can rapidly select, among the leading cell-mediated immunity-based vaccine approaches, those worthy of advancement to larger trials. As with phase IIB test of concept trials, STOC trials should be designed within a coherent product development plan . STOC trials could support the ‘go/no go’ decision point in product development (assuming the manufacturing process is in place). If a single STOC trial of a cell-mediated immunity-based vaccine in a particular population were to confirm a clinically significant reduction in viral load, the trial could serve as the ‘go’ decision point to proceed to a phase III pivotal trial.
Enhancement of susceptibility to infection or disease progression by vaccination has been a concern, but has not been demonstrated to date . With approximately 650 volunteers, a STOC trial will provide limited safety data and will not have power to detect a modest change in the risk of HIV infection. A negative point estimate for protective vaccine efficacy (vaccine effect on susceptibility;VEs) might suggest enhancement. The confidence interval for VEs will, however, be quite wide (±45%) and is likely to include zero. If a STOC trial demonstrates a significant reduction in viral load, then a subsequent phase III trial with approximately 400 HIV infections will be required to obtain a precise estimate of VEs as well as information about the duration of viral suppression and clinical endpoints.
STOC trials require fewer financial and infrastructural resources, expose fewer participants, and require less vaccine than phase IIB test of concept trials. A STOC trial that yielded promising results could lead directly to a pivotal phase III trial, shortening the overall development time (Fig. 1). STOC trials are particularly appealing in the situation of limited access to populations with a higher HIV incidence. The involvement of fewer high-risk populations for a shorter duration offers fewer logistical constraints and is likely to be accepted more easily by communities, provided the purpose, value and limitations of the trial are clearly communicated to stakeholders. Risk-taking behaviour during the trial should be carefully monitored [2,23]. Populations with HIV incidence greater than 3% per annum have been described in Africa [24,25] and China , and among populations of men who have sex with men in the Americas [27,28].
STOC trials are not designed for licensure, and pivotal phase III efficacy trials will still be necessary when results warrant. A STOC trial of representative vaccine candidates may, however, if successful, galvanize the field and lead to efficacy trials for vaccine candidates when consensus is now lacking. A small but detectable effect on viral load might indicate that the vaccine approach has promise but requires improvement. If such cell-mediated immunity-based vaccine candidates fail to alter viral load in the first 6 months of infection, they are unlikely to merit further development.
In conclusion, STOC trials provide a flexible approach to screen various cell-mediated immunity-based AIDS vaccine candidates and to orient further vaccine development in a shorter timeframe. This will accelerate the development of the most promising vaccine candidates or approaches, better utilizing the limited resources available for large efficacy trials. We consider the STOC trial concept to be integral to accelerating AIDS vaccine development.
The authors would like to thank Steve Self and Peter Gilbert for their helpful and constructive comments. They also thank many colleagues at IAVI for useful discussion, Arielle Ginsberg for her excellent editorial assistance, Lisa Gieber for help with literature searches.
Sponsorship: This work was supported by the International AIDS Vaccine Initiative.
1. Mehrotra DV, Li X, Gilbert PB. A comparison of eight methods for the dual-endpoint evaluation of efficacy in a proof-of-concept HIV vaccine trial. Biometrics 2006; 62:893–900.
2. WHO/UNAIDS/IAVI Expert Group. Executive summary and recommendations from the WHO/UNAIDS/IAVI expert group consultation on ‘Phase IIB-TOC trials as a novel strategy for evaluation of preventive HIV vaccines’, 31 January–2 February 2006, IAVI, New York, USA. AIDS
3. Rida W, Fast P, Hoff R, Fleming T. Intermediate-size trials for the evaluation of HIV vaccine candidates: a workshop summary. J Acquir Immune Defic Syndr Hum Retrovirol 1997; 16:195–203.
4. Emini EA, Koff WC. AIDS/HIV. Developing an AIDS vaccine: need, uncertainty, hope. Science 2004; 304:1913–1914.
5. Pantophlet R, Burton DR. GP120: target for neutralizing HIV-1 antibodies. Annu Rev Immunol 2006; 24:739–769.
6. Dhillon AK, Donners H, Pantophlet R, Johnson WE, Decker JM, Shaw GM, et al
. Dissecting the neutralizing antibody specificities of broadly neutralizing sera from HIV-1 infected donors. J Virol 2007; 81:6548–6562.
7. Fast PE. Recent trends in clinical trials of vaccines to prevent HIV/AIDS. Curr Opin HIV AIDS 2006; 1:267–271.
8. Duerr A, Wasserheit JN, Corey L. HIV vaccines: new frontiers in vaccine development. Clin Infect Dis 2006; 43:500–511.
9. Follmann D, Duerr A, Tabet S, Gilbert P, Moodie Z, Fast P, et al
. Endpoints and regulatory issues in HIV vaccine clinical trials: lessons from a workshop. J Acquir Immune Defic Syndr 2007; 44:49–60.
10. Koff WC, Johnson PR, Watkins DI, Burton DR, Lifson JD, Hasenkrug KJ, et al
. HIV vaccine design: insights from live attenuated SIV vaccines. Nat Immunol 2006; 7:19–23.
11. Letvin NL. Progress and obstacles in the development of an AIDS vaccine. Nat Rev Immunol 2006; 6:930–939.
12. Mellors JW, Munoz A, Giorgi JV, Margolick JB, Tassoni CJ, Gupta P, et al
. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection. Ann Intern Med 1997; 126:946–954.
13. Lyles RH, Munoz A, Yamashita TE, Bazmi H, Detels R, Rinaldo CR, et al
. Natural history of human immunodeficiency virus type 1 viremia after seroconversion and proximal to AIDS in a large cohort of homosexual men. Multicenter AIDS Cohort Study. J Infect Dis 2000; 181:872–880.
14. Letvin NL, Mascola JR, Sun Y, Gorgone DA, Buzby AP, Xu L, et al
. Preserved CD4+ central memory T cells and survival in vaccinated SIV-challenged monkeys. Science 2006; 312:1530–1533.
15. Gupta SB, Jacobson LP, Margolick JB, Rinaldo CR, Phair JP, Jamieson BD, et al
. Estimating the benefit of an HIV-1 vaccine that reduces viral load set point. J Infect Dis 2007; 195:546–550.
16. Bechhofer R, Santner T, Goldsman D. Design and analysis of experiments for statistical selection, screening and multiple comparisons. New York: J. Wiley & Sons; 1995.
17. Tarwater PM, Gallant JE, Mellors JW, Gore ME, Phair JP, Detels R, et al
. Prognostic value of plasma HIV RNA among highly active antiretroviral therapy users. AIDS 2004; 18:2419–2423.
18. Arnaout RA, Lloyd AL, O'Brien TR, Goedert JJ, Leonard JM, Nowak MA. A simple relationship between viral load and survival time in HIV-1 infection. Proc Natl Acad Sci U S A 1999; 96:11549–11553.
19. Fideli US, Allen SA, Musonda R, Trask S, Hahn BH, Weiss H, et al
. Virologic and immunologic determinants of heterosexual transmission of human immunodeficiency virus type 1 in Africa. AIDS Res Hum Retroviruses 2001; 17:901–910.
20. Quinn TC, Wawer MJ, Sewankambo N, Serwadda D, Li C, Wabwire-Mangen F, et al
. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group. N Engl J Med 2000; 342:921–929.
21. Self S. Vaccine efficacy trial design: role of phase IIB v. PhIII trials. In: AIDS Vaccine 2005 Conference
. Montreal, Canada, 6–9 September 2005. Abstract 56.
22. Gilbert PB, Ackers ML, Berman PW, Francis DP, Popovic V, Hu DJ, et al
. HIV-1 virologic and immunologic progression and initiation of antiretroviral therapy among HIV-1-infected subjects in a trial of the efficacy of recombinant glycoprotein 120 vaccine. J Infect Dis 2005; 192:974–983.
23. Van Griensvan F, Keawkungwal J, Tappero JW, Sangkum U, Pitisuttithum P, Vanichseni S, et al
. Lack of increased HIV risk behavior among injection drug users participating in the AIDSVAX B/E HIV vaccine trial in Bangkok, Thailand. AIDS 2004; 18:295–301.
24. Kaul R, Kimani J, Nagelkerke NJ, Fonck K, Ngugi EN, Keli F, et al
. Monthly antibiotic chemoprophylaxis and incidence of sexually transmitted infections and HIV-1 infection in Kenyan sex workers: a randomized controlled trial. JAMA 2004; 291:2555–2562.
25. Sanders EJ, Graham S, Van der Elst E, Mwangome M, Mumba T, Mutimba S, et al.
Establishing a high risk HIV-negative cohort in Kilifi, Kenya. In: AIDS Vaccine 2006 Conference
. Amsterdam, the Netherlands, 29 August–1 September 2006. Abstract P13a09.
26. Ruan Y, Qin G, Liu S, Qian H, Zhang L, Zhou F, et al
. HIV incidence and factors contributed to retention in a 12-month follow-up study of injection drug users in Sichuan Province, China. J Acquir Immune Defic Syndr 2005; 39:459–463.
27. CDC. Epidemiology of HIV/AIDS – United States, 1981–2005. MMWR Morb Mortal Wkly Rep
28. Cohen J. HIV/AIDS: Latin America & Caribbean. Overview: the overlooked epidemic. Science 2006; 313:468–469.
© 2007 Lippincott Williams & Wilkins, Inc.