H5N1 virus vaccine from human immunogenicity studies.

HHS Public Access Author manuscript Author Manuscript

Vaccine. Author manuscript; available in PMC 2017 July 19. Published in final edited form as: Vaccine. 2016 July 19; 34(33): 3796–3802. doi:10.1016/j.vaccine.2016.05.067.

Extrapolating theoretical efficacy of inactivated influenza A/ H5N1 virus vaccine from human immunogenicity studies Leora R. Feldsteina,b, Laura Matrajtb,c, M. Elizabeth Halloranb,c,d, Wendy A. Keitele, Ira M. Longini Jr.c,f, and H5N1 Vaccine Working Group1 aDepartment

of Epidemiology, School of Public Health, University of Washington School, Seattle,

WA

Author Manuscript

bVaccine

and Infectious Disease Division, Fred Hutchinson Cancer Research Center, Seattle, WA

cCenter

for Inference and Dynamics of Infectious Diseases, Fred Hutchinson Cancer Research Center, Seattle, WA dDepartment

of Biostatistics, School of Public Health, University of Washington, Seattle, WA

eDepartment

of Medicine, Department of Molecular Virology and Microbiology, Baylor College of Medicine, Houston, Texas

fDepartment

of Biostatistics, College of Public Health and Health Professions and College of Medicine, University of Florida, Gainesville, Florida

Author Manuscript

Abstract Influenza A virus subtype H5N1 has been a public health concern for almost 20 years due to its potential ability to become transmissible among humans. Phase I and II clinical trials have assessed safety, reactogenicity and immunogenicity of inactivated influenza A/H5N1 virus vaccines. A shortage of vaccine is likely to occur during the first months of a pandemic. Hence, determining whether to give one dose to more people or two doses to fewer people to best protect the population is essential. We use hemagglutination-inhibition antibody titers as an immune correlate for avian influenza vaccines. Using an established relationship to obtain a theoretical vaccine efficacy from immunogenicity data from thirteen arms of six phase I and phase II clinical trials of inactivated influenza A/H5N1 virus vaccines, we assessed: 1) the proportion of theoretical vaccine efficacy achieved after a single dose (defined as primary response level), and 2) whether

Author Manuscript

Corresponding author: P.O. Box 117450, 22 Buckman Drive, 452 Dauer Hall University of Florida, Gainesville, FL 32611, Tel: 352-294-1938; Fax: (352) 294-1930. [email protected] (Ira M. Longini Jr.). 1Keitel WA, Atmar RL, (Baylor College of Medicine, Houston, TX), Brady RC, Frenck RW, (Cincinnati Children’s Hospital Medical Center, Cincinnati, OH), Walter EB, Woods CW, (Duke University School of Medicine, Durham, NC), Mulligan MJ, Spearman P, (Emory University School of Medicine, Atlanta, GA), Jackson LA, (Group Health Cooperative, Seattle WA), Belshe RB, Frey SE, (Saint Louis University School of Medicine, St. Louis, MO), Winokur PL, Stapleton JT, (University of Iowa and Iowa City VA Medical Center, Iowa City, IA), Chen WH, Kotloff KL, (University of Maryland School of Medicine, Baltimore, MD), Edwards KM, Creech CB (Vanderbilt University School of Medicine, Nashville, TN). Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Disclaimer The views expressed here do not necessarily reflect the official policies of the Department of Health and Human Services; nor does mention of trade names, commercial practices, or organizations imply endorsement by the U.S. Government.

Feldstein et al.

Page 2

Author Manuscript

theoretical efficacy increases after a second dose, with and without adjuvant. Participants receiving vaccine with AS03 adjuvant had higher primary response levels (range: 0.48–0.57) compared to participants receiving vaccine with MF59 adjuvant (range: 0.32–0.47), with no observed trends in primary response levels by antigen dosage. After the first and second doses, vaccine with AS03 at dosage levels 3.75, 7.5 and 15 mcg had the highest estimated theoretical vaccine efficacy: Dose 1) 45% (95%CI: 36–57%), 53% (95%CI: 42–63%) and 55% (95%CI: 44–64%), respectively and Dose 2) 93% (95%CI: 89–96%), 97% (95%CI: 95–98%) and 97% (95%CI: 96–100%), respectively. On average, the estimated theoretical vaccine efficacy of lower dose adjuvanted vaccines (AS03 and MF59) was 17% higher than that of higher dose unadjuvanted vaccines, suggesting that including an adjuvant is dose-sparing. These data indicate adjuvanted inactivated influenza A/H5N1 virus vaccine produces high theoretical efficacy after two doses to protect individuals against a potential avian influenza pandemic.

Author Manuscript

Background

Author Manuscript

Influenza A virus subtype H5N1, an avian influenza strain, has been a serious public health concern for almost 20 years because of its virulence and potential to become transmissible among humans [1]. From 2003 to September 17, 2015, 844 confirmed human cases of H5N1 infection with 449 deaths occurred in 16 countries [2]. Because of the high case fatality ratio (53%) and negligible population immunity, a deadly pandemic could result if the virus becomes readily transmissible between persons [3,4]. If a pandemic were to start, a pandemic strain-specific vaccine would need to be produced and deployed rapidly [5]. Vaccines stockpiled for immediate deployment may not match the pandemic virus, however the stockpiled H5N1 influenza vaccines may provide some heterologous protection to similar clades [6–8], suggesting stockpiled vaccine could be a first-line intervention while well-matched vaccine is produced. To prepare for a possible A/H5N1 pandemic, phase I and II clinical trials have assessed the safety, reactogenicity and immunogenicity of inactivated influenza A/H5N1 virus vaccines. However, the efficacy of H5N1 vaccines in humans remains unproven [9]. Animal challenge studies suggest these vaccines are protective [10– 12], but extrapolating these results to efficacy in humans remains poorly understood. Generally two doses of vaccine are recommended to achieve full efficacy for an individual. However, at the population level, depending on the efficacy achieved with one dose, vaccinating a larger proportion of the population with a single dose could achieve a greater reduction in morbidity and transmission [13]. Currently, there are no estimates of vaccine efficacy for either one or two doses of avian influenza vaccines. Given the limited vaccine supply and time constraints during a pandemic, determining the efficacy is essential for optimal allocation of resources.

Author Manuscript

The aim of this study is to close this gap by providing theoretical efficacy estimates for avian influenza vaccines. Hemagglutination-inhibition (HAI) antibody titer is widely recognized as a correlate of protection against seasonal influenza infection. Currently, this is the only correlate of protection used for licensure in the US [14], although secretory IgA and anti neuraminidase antibodies have also been shown to correlate with protection [15,16]. Based on Hobson et al. [17] it is believed an HAI antibody titer of 40 is associated with 50% protection against seasonal influenza illness in healthy adults. However, little is known about

Vaccine. Author manuscript; available in PMC 2017 July 19.

Feldstein et al.

Page 3

Author Manuscript

the relationship between HAI antibody titer and protection against specific viral strains (or influenza A/H5N1 viruses), and the influence of host factors such as age [18]. In Coudeville et al. [19], data from 15 seasonal influenza vaccine studies (six challenge studies, five clinical trials and four cohort studies) reported between 1945 and 2006, were used to construct a continuous curve estimating level of protection at varying levels of HAI titers against illness caused by seasonal influenza strains. Fourteen of the studies included adults aged 18–60, and one included adults aged 60 years or greater. The protection measured was against a mixture of infection and illness. For the purpose of this analysis, we assume that avian influenza vaccine-induced protection is similar to that of seasonal influenza vaccines, and the HAI protection curve of Coudeville et al. provides a theoretical estimate of inactivated influenza A/H5N1 vaccine efficacy against avian influenza infection or illness. Under these assumptions, we analyzed data from 14 phase I and phase II clinical trials to estimate the proportion of theoretical efficacy achieved after the first and second doses of vaccine, and to assess the impact of antigen dosage and adjuvant on vaccine efficacy.

Author Manuscript

Methods

Author Manuscript

Data from 14 phase I and eight phase II randomized clinical trials assessing safety, reactogenicity and immunogenicity of inactivated influenza A/H5N1 virus vaccines were made available by the National Institute of Allergy and Infectious Diseases (NIAID), NIH. Healthy people between the ages of 18 and 99 years volunteered to participate in these trials. Within each trial, eligible subjects were randomized to receive two doses of varying dosages of vaccine antigen with or without one of two adjuvants: MF59 or Adjuvant System 03 (AS03). Some studies included a placebo arm. MF59 and AS03 adjuvants are both oil-inwater emulsions manufactured by Novartis Vaccines and GlaxoSmithKline (GSK), respectively. Only vaccine trial arms including A/H5N1 virus vaccine-naïve participants who received two intramuscular doses of 3.75, 7.5, 15, or 90 mcg of vaccine spaced 14 to 180 days apart were considered for inclusion in the analysis.

Author Manuscript

We included only trial arms with vaccines that had estimated theoretical efficacies greater than 60% after two doses, which corresponds to the curve given by Coudeville et al. to geometric mean HAI titers (GMT) >24. This decision was based on findings that recent influenza epidemics have had basic reproductive numbers (the expected number of secondary infections resulting from a typical infectious person in a completely susceptible population [20]), in the range of approximately 1.6 – 1.8 [21,22]. Assuming vaccine coverage in a targeted vaccination group during a future epidemic will be 70% at best, the estimated lower bound of vaccine efficacy to control the epidemic will be about 60% [23]. For example, if the basic reproductive number during an epidemic is 1.7, and vaccine coverage is 70%, vaccine efficacy would need to be at least 59% to control transmission in a homogeneously mixing population [24]. We did not exclude any data based on type of antigen used, type of adjuvant used, or the population of the study. Thirteen trial arms from six trials exhibited estimated theoretical efficacies greater than 60% and were therefore included in this analysis (Table 1). Seven of the trial arms used vaccines with either MF59 or AS03 [25–27]. The remaining six trial arms used an antigen dosage of 90 mcg and no adjuvant [6,28]. One of those six trial arms used


Feldstein et al.

Page 4

Author Manuscript

45 mcg of one clade (H5 A/Indo) and 45 mcg of another clade (5 A/Vietnam) at each dose. All but one of the thirteen trials arms (the Chiron antigen was used in trial #04-062) used a Sanofi antigen. Using a per protocol analysis, GMT and 95% confidence intervals were calculated for each trial arm 14 to 28 days after the first dose of vaccine was administered and again for each arm, 21 to 28 days after administration of the second dose. By extracting data points from the HAI protection curve (Figure 1) in Coudeville et al., theoretical vaccine efficacy and 95% credible intervals, the Bayesian analog of confidence intervals, were estimated from the GMT for each trial arm after one and two doses. We employed these same methods to conduct a sensitivity analysis using the HAI protection curve presented in Tsang et al. [29].

Author Manuscript

We define the primary response level as the proportion of the overall maximum efficacy obtained after a single dose of vaccine. For example, if a vaccine had an overall efficacy of 80% after two doses, a primary response level of 50% corresponds to obtaining half of the protection after one dose (40% vaccine efficacy after one dose). Primary response levels are useful measures because they allow us to parameterize mathematical and computer models that will compare vaccination strategies with multiple doses of vaccine [13]. We calculated the primary response level for each combination of vaccine, dosage and adjuvant using the ratio of the estimated theoretical vaccine efficacy after the first dose over the estimated theoretical vaccine efficacy after the second dose.

Author Manuscript

Using STATA 12, an unadjusted weighted linear regression was used to determine whether there was a significant absolute difference in GMT between adjuvanted vaccines (AS03 and MF59) and unadjuvanted vaccines, and to determine whether there was a significant absolute difference in GMT by antigen dosage (3.75, 7.5, 15, and 90 mcg). For the regression analysis, the weights chosen were based on the sample size of each arm and were inversely proportional to the variance of each GMT included in the regression model. Unpaired t-tests with unequal variances were used to determine mean GMT, mean estimated theoretical vaccine efficacy and mean primary response level by whether vaccines had adjuvants or not.

Results Geometric mean titers after the first and second doses

Author Manuscript

Fourteen to twenty-eight days after the first dose of vaccine was administered, GMTs were highest among individuals in the trial arm that used the IIVI formulated with AS03 at a dosage of 15 mcg (GMT=19.61, 95% CI: 14.00–27.47), followed by one of the trial arms that used 90 mcg of unadjuvanted IIVI, (GMT=19.15, 95% CI: 8.21–44.68) and by the trial arm that used the IIVI formulated with AS03 at a dosage of 7.5 mcg (GMT=18.73, 95% CI: 13.82, 25.40) (Table 1, Figure S1). Twenty-one to twenty-eight days after the second dose of vaccine, GMTs were highest among individuals in the three trial arms using the Sanofi antigen with AS03 at dosage levels 3.75, 7.5 and 15 mcg: 125.63 (95% CI: 85.13–185.40), 222.82 (95% CI: 177.49–279.71), and 259.50 (95% CI: 194.56–346.12), respectively (Table 1, Figure S2). The next highest GMT was from the unadjuvanted trial arm in which subjects


Feldstein et al.

Page 5

Author Manuscript

were given the second dose at 180 days after the first dose (GMT=58.74, 95% CI: 38.04, 90.72) (Table 1, Figure S2). Based on an unadjusted weighted linear regression, after the second dose, adjuvanted vaccines elicited an increase in GMT of 83.88 (95% CI: 2.42–165.34) compared to unadjuvanted vaccines despite a 6- to 24-fold higher HA content in the unadjuvanted formulations (p-value=0.034). Based on an unadjusted weighted linear regression, an antigen dosage increase from 3.75 to 7.5 mcg and from 3.75 to 15 mcg resulted in an increase in GMT; however, these increases in GMT were not statistically significant (Table 2). When antigen dosage increased from 3.75 to 90 mcg there was a decrease in GMT, likely because none of these 90 mcg dosage vaccines had an adjuvant, but this difference was also not statistically significant.

Author Manuscript

Estimated Theoretical Vaccine Efficacy Vaccine efficacy after the first dose—The trial arms using the Sanofi antigen with AS03 at dosage levels 3.75, 7.5 and 15 mcg had estimated theoretical efficacies of 45% (95% CI: 36–57%), 53% (95% CI: 42–63%) and 55% (95% CI: 44–64%), respectively, after the first dose (Table 3, Figure S3). Following these, the three trial arms adjuvanted with MF59 at 3.75, 7.5, 15 mcg had estimated theoretical vaccine efficacies of 25% (95% CI: 19– 32%), 37% (95% CI: 30–46%), and 32% (95% CI: 24–38%), respectively, after the first dose. The trial arm using 15 mcg of Chiron antigen with MF59 had an estimated theoretical efficacy of 32% (95% CI: 21–45%) after the first dose. The six trials arms using the unadjuvanted Sanofi antigen at a dosage of 90 mcg, after the first dose of vaccine, had estimated theoretical vaccines efficacies ranging from 18% (95% CI: 12–23%) to 54% (95% CI: 28–78%).

Author Manuscript Author Manuscript

Vaccine efficacy after two doses—The trial arms using AS03 adjuvant had the highest theoretical vaccine efficacy after two doses, estimated at 93% (95% CI: 89–96%), 97% (95% CI: 95–98%) and 97% (95% CI: 96–100%), for the 3.75, 7.5 and 15 mcg dosage levels, respectively (Table 3, Figure S4). The three trial arms adjuvanted with MF59 adjuvant had estimated theoretical vaccines efficacies of 78% (95% CI: 67–87%), 79% (95% CI: 68–88%), and 77% (95% CI: 63–87%) at dosage levels of 3.75, 7.5, and 15 mcg, respectively after two doses. The trial arm using 15 mcg of Chiron antigen together with MF59 adjuvant had an estimated theoretical vaccine efficacy of 69% (95% CI: 57–81%) after two doses. We next compare vaccines using Sanofi antigen adjuvanted with either AS03 or MF59. Adjuvanting with MF59 did not produce as high an estimated theoretical vaccine efficacy after two doses as AS03-adjuvanted vaccines (79% compared to 97%, respectively, p-value= 0.006). The six trial arms using unadjuvanted vaccine at a dosage of 90 mcg had estimated vaccine efficacies after two doses ranging from 61% (95% CI: 45– 75%) to 83% (95% CI: 74–90%). The highest efficacy of the unadjuvanted vaccine trial arms was from the arm in which subjects were given the second dose of vaccine 180 days after the first dose (Table 3, Figure S4). The mean estimated theoretical vaccine efficacy after two doses of adjuvanted vaccines was 17% higher than the mean estimated theoretical vaccine efficacy of unadjuvanted vaccines (p-value= 0.0083, Table 4). Adjuvanted vaccines


Feldstein et al.

Page 6

Author Manuscript

with an antigen dosage of 7.5 mcg had the highest average estimated theoretical vaccine efficacy of 88% after two doses (Table 5). Primary Response Level

Author Manuscript

The primary response level in the adjuvanted vaccines ranged from 0.48 to 0.57 in vaccines with AS03 adjuvant, and from 0.32 to 0.47 in those with MF59 adjuvant (Table 3). Participants receiving AS03-adjuvanted vaccines had higher primary response levels than participants who received vaccines adjuvanted with MF59 (primary response level mean of 0.53 versus 0.42, respectively). The primary response level in the unadjuvanted vaccines with 90 mcg ranged from 0.30 to 0.83. Primary response levels had a higher variance among the unadjuvanted formulations than among the adjuvanted formulations (p=0.016 based on an F-test). There was no observed trend in primary response level by antigen dosage (Table 3), but this might be due to the fact that we excluded formulations of unadjuvanted vaccines at lower dosages from the analysis, thereby missing marked dose-response effects. In several trials, high dosages of unadjuvanted vaccines tended to have higher primary response levels than vaccines with adjuvants (mean of 0.58 for unadjuvanted vaccines versus 0.47 for adjuvanted vaccines), but this difference was not significant (p-value=0.3290) and it is likely due to the small observed increase in GMTs from the first dose to the second dose among unadjuvanted vaccines. Sensitivity analysis

Author Manuscript

In a household study, Tsang et al. found that an HAI antibody titer of 40 was associated with considerably less than 50% protection against seasonal influenza illness [29].Using the HAI protection curve obtained from this article, we reanalyzed our data (Appendix). In brief, our results were consistent: the same three trial arms had the highest theoretical vaccine efficacy after one and two doses, however, the estimated theoretical vaccine efficacy was approximately half that of the estimated efficacy using the curve from Coudeville et al., (Table 1, Appendix). Efficacy after the first and second dose from the trial arms using AS03 adjuvant ranged from 21–24% and 45–52%, respectively. More importantly, this protection curve yielded similar primary response levels in both adjuvanted and unadjuvanted vaccines, with similar ranges (0.29–0.47 for adjuvanted vaccines and 0.27–0.83 for unadjuvanted vaccines).

Discussion

Author Manuscript

Based on the findings of these models, we infer that vaccinating individuals with two doses of inactivated subvirion influenza A/H5N1 virus vaccine formulated with AS03 could provide the highest level of protection against avian influenza infections, followed by vaccines adjuvanted with MF59 and finally by unadjuvanted vaccines. Although trial arms with antigen dosages of 90 mcg did not produce markedly different theoretical efficacies from trial arms with MF59 adjuvanted antigens, it may be unrealistic to use an antigen dosage of 90 mcg, because vaccine supplies are limited [30]. Among these arms, the highest theoretical vaccine efficacy was achieved in the trial arm where the second dose was given after 180 days, suggesting that a prepandemic priming vaccination strategy could be highly beneficial. Using the current standard dosage of 15 mcg would allow an increase in supply


Feldstein et al.

Page 7

Author Manuscript

of doses by at least 6-fold [26]. Interestingly, as long as an adjuvant was present, an increase in antigen dosage from 3.75 to 7.5 or from 7.5 to 15 mcg made little difference in increasing estimated theoretical vaccine efficacy. The option of using an antigen dose-sparing vaccination strategy during a pandemic emergency is particularly attractive, given the limited amount of vaccine supplies that may be available. Pandemic planners will also need to consider the accumulated tolerability and safety experiences for each adjuvant. Although adjuvanted influenza vaccines are generally well tolerated, the safety profiles of these adjuvants used in combination with novel influenza antigens warrants continued monitoring.


Dose-sparing strategies have been previously studied for seasonal influenza [31–33] and, more recently, for pandemic scenarios [11,12, 35–45]. Previous work suggests that, in a pandemic, an antigen-sparing strategy, where only a single dose of vaccine is given instead of two, is optimal to minimize illness attack rates for primary response levels in the vicinity of 0.50 [13]. Here, we find, on average, adjuvanted vaccines had a primary response level of 0.47, and unadjuvanted vaccines had a primary response level slightly higher (0.58, the difference between these groups was not significant), suggesting that vaccinating a larger proportion of the population with a single dose of vaccine (either adjuvanted or unadjuvanted) might be the optimal use of resources for minimizing illness attack rates [13]. Considering a different outcome could give rise to a different optimal use of vaccine. For example if two doses are needed to prevent infection, then vaccination campaigns with two doses of vaccine could be used if interest is in stopping transmission, but if a single dose of vaccine is enough to prevent death (even if it does not prevent infection), it would be advantageous to vaccinate more people with a single dose. Public health officials need to balance several outcomes such as mortality, morbidity, and financial loss when determining the use of available vaccines.

Author Manuscript

An important limitation of these findings is that they are based on several key assumptions. First, we assume geometric mean HAI titer corresponds to vaccine efficacy based on the protection curve produced in Coudeville et al. [19]. By using this curve we were able to provide an estimate of vaccine efficacy for any given value of HAI titer, rather than the currently used single-point threshold of HAI titer of 40 for a 50% protection in Hobson et al. [17]. The curve in Coudeville et al. has a steep slope, and associates a 60% protection with HAI titers of 24, much lower than the Hobson threshold, but that threshold is based on a single study from 1972 and has wide confidence intervals. More recent studies have shown a larger range of HAI titer to achieve protection, depending on the influenza strain and the population studied [45,46]. We used the HAI protection curve from one of those studies to show more conservative estimates of theoretical vaccine efficacy, but the primary response levels were conserved. Indeed, the vaccine efficacy estimates were approximately half that of the estimates extrapolated from the Coudeville et al. curve, however there are key differences in these two studies that may account for the substantial difference in vaccine efficacy estimates. Tsang et al., measured HAI titers from family members of influenzaconfirmed individuals. Protection may have been lower among this group because they may have been exposed to influenza virus for a longer duration or at a higher intensity than individuals in the general population [29]. Second, we assume the vaccine produced will be


Feldstein et al.

Page 8


antigenically close to the circulating H5N1 strain. To date, at least 20 different clades of H5N1 have been identified [47]. Therefore, predicting the correct antigenic strain or identifying it in time to produce and administer a vaccine to prevent a future pandemic would be challenging. Third, we assumed that increases in geometric mean HAI titers correlate with protection against avian influenza as they do for seasonal influenza. To date, few avian influenza infections have taken place in the humans, and none have occurred in vaccinated individuals, so characterization of the immune response induced by avian influenza vaccines and efficacy is not possible. Fourth, we combined data from trials using different antigens and adjuvants, and considered different time points following vaccination. Antibody levels could potentially continue to increase during that interval. We considered vaccines with a theoretical vaccine efficacy > 60%. This could result in overestimating the efficacy of unadjuvanted vaccines since only high-dosage arms (90 mcg) of unadjuvanted vaccine met this requirement and were analyzed. In addition, our cutoff of 60% was established assuming 70% vaccination coverage. This coverage is relatively high, but assuming a lower, more realistic coverage would result in an even higher theoretical vaccine efficacy cutoff to control a pandemic. However, it is reasonable to assume that if an avian influenza virus, with a high mortality rate, becomes transmissible among humans, the vaccine uptake will be higher than that for seasonal influenza. In this sense, our results are conservative. While these are strong assumptions, we believe that the present study used the best information available to provide theoretical vaccine efficacy estimates needed to quantify intervention strategies. HAI titers are the best-known correlate of protection against seasonal influenza. Extrapolating theoretical vaccine efficacy from GMTs provides crucial information about which vaccine to stockpile, as well as the quantity of doses needed to prevent a pandemic of avian influenza. The fact remains that suitable influenza vaccines are the best way to mitigate influenza pandemic.

Author Manuscript

Supplementary Material Refer to Web version on PubMed Central for supplementary material.

Acknowledgments This work was partially supported by National Institute of General Medical Sciences MIDAS grants U01GM070749 and U54-GM111274-01 and National Institute of Allergy and Infectious Diseases grant R37-AI032042. The analyses for this Research Project utilized data generated by the NIAID-funded Vaccine Treatment and Evaluation Units (VTEUs).

References Author Manuscript

1. Longini IM, Nizam A, Xu S, Ungchusak K, Hanshaoworakul W, Cummings DAT, et al. Containing pandemic influenza at the source. Science. 2005; 309:1083–1087. [PubMed: 16079251] 2. WPRO. [accessed September 21, 2015] Avian Influenza. 2015. http://www.wpro.who.int/ emerging_diseases/AvianInfluenza/en/ 3. Yang Y, Halloran ME, Sugimoto JD, Longini IM. Detecting human-to-human transmission of avian influenza A (H5N1). Emerg Infect Dis. 2007; 13:1348–1353. [PubMed: 18252106] 4. [accessed April 9, 2015] Highly Pathogenic Avian Influenza (H5N1) Human Cases and Deaths Since 2003 n.d. http://www.cdc.gov/flu/pdf/avianflu/avian-flu-human-world-summary.pdf 5. Germann TC, Kadau K, Longini IM, Macken CA. Mitigation strategies for pandemic influenza in the United States. Proc Natl Acad Sci. 2006; 103:5935–5940. [PubMed: 16585506]


Feldstein et al.

Page 9

Author Manuscript Author Manuscript Author Manuscript Author Manuscript

6. Belshe RB, Frey SE, Graham I, Mulligan MJ, Edupuganti S, Jackson LA, et al. Safety and immunogenicity of influenza A H5 subunit vaccines: effect of vaccine schedule and antigenic variant. J Infect Dis. 2011; 203:666–673. [PubMed: 21282194] 7. Gillard P, Caplanusi A, Knuf M, Roman F, Walravens K, Moris P, et al. An assessment of primeboost vaccination schedules with AS03A -adjuvanted prepandemic H5N1 vaccines: a randomized study in European adults. Influenza Other Respi Viruses. 2013; 7:55–65. 8. Belshe RB, Frey SE, Graham IL, Anderson EL, Jackson LA, Spearman P, et al. Immunogenicity of avian influenza A/Anhui/01/2005(H5N1) vaccine with MF59 adjuvant: a randomized clinical trial. JAMA. 2014; 312:1420–1428. [PubMed: 25291578] 9. Ferguson NM, Cummings DAT, Cauchemez S, Fraser C, Riley S, Meeyai A, et al. Strategies for containing an emerging influenza pandemic in Southeast Asia. Nature. 2005; 437:209–214. [PubMed: 16079797] 10. Liu G, Zhang F, Shi J, Tian G, Chen H, Yu K, et al. A subunit vaccine candidate derived from a classic H5N1 avian influenza virus in China protects fowls and BALB/c mice from lethal challenge. Vaccine. 2013; 31:5398–5404. [PubMed: 24055355] 11. Park S-J, Kim E-H, Pascua PNQ, Kwon H-I, Lim G-J, Decano A, et al. Evaluation of heterosubtypic cross-protection against highly pathogenic H5N1 by active infection with human seasonal influenza A virus or trivalent inactivated vaccine immunization in ferret models. J Gen Virol. 2014; 95:793–798. [PubMed: 24421115] 12. Song M-S, Moon H-J, Kwon H-I, Pascua PNQ, Lee JH, Baek YH, et al. Evaluation of the efficacy of a pre-pandemic H5N1 vaccine (MG1109) in mouse and ferret models. J Microbiol. 2012; 50:478–488. [PubMed: 22752912] 13. Matrajt L, Britton T, Halloran ME, Longini IM. One versus two doses: What is the best use of vaccine in an influenza pandemic? Epidemics. 2015; 13:17–27. [PubMed: 26616038] 14. [accessed October 30, 2015] Guidance for Industry: Clinical Data Needed to Support the Licensure of Pandemic Influenza Vaccines. US Dep Heal Hum Serv Food Drug Adm Cent Biol Eval Res. 2007. http://www.fda.gov/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/ Guidances/Vaccines/ucm074786.htm 15. Ghendon Y. The immune response to influenza vaccines. Acta Virol. 1990; 34:295–304. [PubMed: 1980401] 16. Couch RB, Atmar RL, Franco LM, Quarles JM, Wells J, Arden N, et al. Antibody correlates and predictors of immunity to naturally occurring influenza in humans and the importance of antibody to the neuraminidase. J Infect Dis. 2013; 207:974–981. [PubMed: 23307936] 17. Hobson D, Curry RL, Beare AS, Ward-Gardner A. The role of serum haemagglutination-inhibiting antibody in protection against challenge infection with influenza A2 and B viruses. J Hyg (Lond). 1972; 70:767–777. [PubMed: 4509641] 18. Hannoun C, Megas F, Piercy J. Immunogenicity and protective efficacy of influenza vaccination. Virus Res. 2004; 103:133–138. [PubMed: 15163501] 19. Coudeville L, Bailleux F, Riche B, Megas F, Andre P, Ecochard R. Relationship between haemagglutination-inhibiting antibody titres and clinical protection against influenza: development and application of a bayesian random-effects model. BMC Med Res Methodol. 2010; 10:18. [PubMed: 20210985] 20. Anderson, RM.; May, RM. Infectious Diseases of Humans. Oxford University Press; 1992. 21. Yang Y, Sugimoto JD, Halloran ME, Basta NE, Chao DL, Matrajt L, et al. The transmissibility and control of pandemic influenza A (H1N1) virus. Science. 2009; 326:729–733. [PubMed: 19745114] 22. Chao DL, Matrajt L, Basta NE, Sugimoto JD, Dean B, Bagwell DA, et al. Planning for the control of pandemic influenza A (H1N1) in Los Angeles County and the United States. Am J Epidemiol. 2011; 173:1121–1130. [PubMed: 21427173] 23. Matrajt L, Longini IM. Critical immune and vaccination thresholds for determining multiple influenza epidemic waves. Epidemics. 2012; 4:22–32. [PubMed: 22325011] 24. Hill AN, Longini IM. The critical vaccination fraction for heterogeneous epidemic models. Math Biosci. 2003; 181:85–106. [PubMed: 12421553]


Feldstein et al.

Page 10

Author Manuscript Author Manuscript Author Manuscript Author Manuscript

25. Chen WH, Jackson LA, Edwards KM, Keitel WA, Hill H, Noah DL, et al. Safety, Reactogenicity, and Immunogenicity of Inactivated Monovalent Influenza A(H5N1) Virus Vaccine Administered With or Without AS03 Adjuvant. Open Forum Infect Dis. 2014; 1:ofu091. [PubMed: 25734159] 26. Mulligan MJ, Bernstein DI, Frey S, Winokur P, Rouphael N, Dickey M, et al. Point-of-Use Mixing of Influenza H5N1 Vaccine and MF59 Adjuvant for Pandemic Vaccination Preparedness: Antibody Responses and Safety. A Phase 1 Clinical Trial. Open Forum Infect Dis. 2014; 1:ofu102. [PubMed: 25734170] 27. Bernstein DI, Edwards KM, Dekker CL, Belshe R, Talbot HKB, Graham IL, et al. Effects of adjuvants on the safety and immunogenicity of an avian influenza H5N1 vaccine in adults. J Infect Dis. 2008; 197:667–675. [PubMed: 18260764] 28. Treanor JJ, Campbell JD, Zangwill KM, Rowe T, Wolff M. Safety and immunogenicity of an inactivated subvirion influenza A (H5N1) vaccine. N Engl J Med. 2006; 354:1343–1351. [PubMed: 16571878] 29. Tsang TK, Cauchemez S, Perera RAPM, Freeman G, Fang VJ, Ip DKM, et al. Association between antibody titers and protection against influenza virus infection within households. J Infect Dis. 2014; 210:684–692. [PubMed: 24676208] 30. Rebmann T, Zelicoff A. Vaccination against influenza: role and limitations in pandemic intervention plans. Expert Rev Vaccines. 2012; 11:1009–1019. [PubMed: 23002981] 31. Engler RJM, Nelson MR, Klote MM, VanRaden MJ, Huang C-Y, Cox NJ, et al. Half- vs full-dose trivalent inactivated influenza vaccine (2004–2005): age, dose, and sex effects on immune responses. Arch Intern Med. 2008; 168:2405–2414. [PubMed: 19064822] 32. Falsey AR. Half-dose influenza vaccine: stretching the supply or wasting it? Arch Intern Med. 2008; 168:2402–2403. [PubMed: 19064820] 33. Treanor J, Keitel W, Belshe R, Campbell J, Schiff G, Zangwill K, et al. Evaluation of a single dose of half strength inactivated influenza vaccine in healthy adults. Vaccine. 2002; 20:1099–1105. [PubMed: 11803070] 34. Kenney RT, Frech SA, Muenz LR, Villar CP, Glenn GM. Dose sparing with intradermal injection of influenza vaccine. N Engl J Med. 2004; 351:2295–2301. [PubMed: 15525714] 35. Riley S, Wu JT, Leung GM. Optimizing the dose of pre-pandemic influenza vaccines to reduce the infection attack rate. PLoS Med. 2007; 4:e218. [PubMed: 17579511] 36. Wood J, McCaw J, Becker N, Nolan T, MacIntyre CR. Optimal dosing and dynamic distribution of vaccines in an influenza pandemic. Am J Epidemiol. 2009; 169:1517–1524. [PubMed: 19395691] 37. Winokur PL, Patel SM, Brady R, Chen WH, El-Kamary SS, Edwards K, et al. Safety and Immunogenicity of a Single Low Dose or High Dose of Clade 2 Influenza A(H5N1) Inactivated Vaccine in Adults Previously Primed With Clade 1 Influenza A(H5N1) Vaccine. J Infect Dis. 2015; 212:525–530. [PubMed: 25712967] 38. Mulligan MJ, Bernstein DI, Winokur P, Rupp R, Anderson E, Rouphael N, et al. Serological responses to an avian influenza A/H7N9 vaccine mixed at the point-of-use with MF59 adjuvant: a randomized clinical trial. JAMA. 2014; 312:1409–1419. [PubMed: 25291577] 39. Czajka H, Unal S, Ulusoy S, Usluer G, Strus A, Sennaroglu E, et al. A phase II, randomised clinical trial to demonstrate the non-inferiority of low-dose MF59-adjuvanted pre-pandemic A/ H5N1 influenza vaccine in adult and elderly subjects. J Prev Med Hyg. 2012; 53:136–142. [PubMed: 23362618] 40. Belshe RB, Frey SE, Graham I, Mulligan MJ, Edupuganti S, Jackson La, et al. Safety and immunogenicity of influenza A H5 subunit vaccines: Effect of vaccine schedule and antigenic variant. J Infect Dis. 2011; 203:666–673. [PubMed: 21282194] 41. Schwarz TF, Horacek T, Knuf M, Damman H-G, Roman F, Dramé M, et al. Single dose vaccination with AS03-adjuvanted H5N1 vaccines in a randomized trial induces strong and broad immune responsiveness to booster vaccination in adults. Vaccine. 2009; 27:6284–6290. [PubMed: 19856521] 42. Keitel W, Groth N, Lattanzi M, Praus M, Hilbert AK, Borkowski A, et al. Dose ranging of adjuvant and antigen in a cell culture H5N1 influenza vaccine: safety and immunogenicity of a phase 1/2 clinical trial. Vaccine. 2010; 28:840–848. [PubMed: 19835829]


Feldstein et al.

Page 11

Author Manuscript

43. Leroux-Roels I, Roman F, Forgus S, Maes C, De Boever F, Dramé M, et al. Priming with AS03 Aadjuvanted H5N1 influenza vaccine improves the kinetics, magnitude and durability of the immune response after a heterologous booster vaccination: an open nonrandomised extension of a double-blind randomised primary study. Vaccine. 2010; 28:849–857. [PubMed: 19835828] 44. Leroux-Roels G. Prepandemic H5N1 influenza vaccine adjuvanted with AS03: a review of the preclinical and clinical data. Expert Opin Biol Ther. 2009; 9:1057–1071. [PubMed: 19555313] 45. Fox JP, Cooney MK, Hall CE, Foy HM. Influenzavirus infections in Seattle families, 1975–1979. II. Pattern of infection in invaded households and relation of age and prior antibody to occurrence of infection and related illness. Am J Epidemiol. 1982; 116:228–242. [PubMed: 7114034] 46. de Jong JC, Palache AM, Beyer WEP, Rimmelzwaan GF, Boon ACM, Osterhaus ADME. Haemagglutination-inhibiting antibody to influenza virus. Dev Biol (Basel). 2003; 115:63–73. [PubMed: 15088777] 47. WHO. [accessed April 9, 2015] Influenza. 2011. http://www.who.int/influenza/gisrs_laboratory/ h5n1_nomenclature/en/#

Author Manuscript Author Manuscript Author Manuscript Vaccine. Author manuscript; available in PMC 2017 July 19.

Feldstein et al.

Page 12

Author Manuscript

Highlights 1.

Adjuvanted vaccines provided the highest vaccine efficacy after 1 and 2 doses.

2.

Vaccine efficacy of adjuvanted vaccines was 17% higher than unadjuvanted vaccines.

3.

Proportion of overall maximum efficacy obtained after 1 dose of vaccine was computed.

4.

This proportion ranged from 0.32 and 0.57 in adjuvanted vaccines.


Feldstein et al.

Page 13

Author Manuscript Author Manuscript Figure 1.

Author Manuscript

Hemagglutination-inhibition Protection Curve from Coudeville et al. [19], solid line, with 95% credible intervals, dashed lines *Probability of protection is the measure of vaccine efficacy

Author Manuscript Vaccine. Author manuscript; available in PMC 2017 July 19.

Author Manuscript

Author Manuscript

Author Manuscript

Phase II

Phase I/II

Phase I

Phase I

Phase I

Phase I

Phase I

Phase I

Phase II

Phase II

04-076

10-0016

10-0016

10-0016

10-0017

10-0017

10-0017

07-0019

07-0019

Phase I/II

04-062

04-063

Study Design

Trial #

Vaccine. Author manuscript; available in PMC 2017 July 19. Sanofi

Sanofi

Sanofi

Sanofi

Sanofi

Sanofi

Sanofi

Sanofi

Sanofi

Sanofi

Chiron

Vaccine Manufacturer

None

None

AS03, GSK

AS03, GSK

AS03, GSK

MF59, Novartis

MF59, Novartis

MF59, Novartis

None

None

MF59, Novartis

Adjuvant, Manufacturer

H5 A/Indo

H5 A/Indo & H5 A/Vietnam

H5 A/Indo

H5 A/Indo

H5 A/Indo

H5 A/Indo

H5 A/Indo

H5 A/Indo

H5 A/Vietnam

H5 A/Vietnam

H5 A/Vietnam

Strain

0, 28

0, 28

0, 21

0, 21

0, 21

0, 21

0, 21

0, 21

0, 28

0, 28

0, 28

Vaccine Interval (Days)

90

90

15

7.5

3.75

15

7.5

3.75

90

90

15

Dosage (mcg)

45–49

48–50

43–53

45–53

43–53

41–53

46–54

44–55

79

99

32

Sample Size

18 – 49

18 – 49

18 – 49

18 – 49

18 – 49

18 – 49

18 – 49

18 – 49

65 – 99

18 – 65

18 – 64

Age (years)

11.20 (7.86– 15.95)

5.86 (5.06– 6.79)

19.61 (14.00– 27.47)

18.73 (13.82– 25.40)

15.10 (11.05– 20.63)

9.68 (7.44– 12.59)

11.74 (8.91– 15.47)

7.48 (5.95– 9.41)

18.44 (13.12– 25.92)

13.53 (9.81– 18.66)

9.47 (6.25– 14.35)

GMT (95% CI) Post Dose 1

27.64 (16.79– 45.50)

24.30 (15.09– 39.16)

259.50 (194.56– 346.12)

222.82 (177.49– 279.71)

125.63 (85.13– 185.40)

43.90 (26.23– 73.46)

49.39 (30.18– 80.84)

46.46 (29.40– 73.41)

26.52 (18.95– 37.12)

28.14 (20.62– 38.39)

32.56 (20.02– 52.96)


Vaccines trial arms included in the analysis [6,25–28]. Geometric mean titer and 95% confidence intervals, 14–28 days after administration of dose 1 and 2.

Author Manuscript

Table 1 Feldstein et al. Page 14

Phase II

Phase II

07-0019

Author Manuscript 07-0019

Sanofi

Sanofi

Vaccine Manufacturer

None

None

Adjuvant, Manufacturer

H5 A/Indo

H5 A/Vietnam

Strain

0, 180

0, 14

Vaccine Interval (Days)

Author Manuscript

Study Design

90

90

Dosage (mcg)

46–52

24

Sample Size

18 – 49

18 – 49

Age (years)

7.26 (5.60– 9.42)

19.15 (8.21– 44.68)


Author Manuscript

Trial #

58.74 (38.04– 90.72)

27.88 (11.59– 67.05)


Feldstein et al. Page 15

Author Manuscript


Feldstein et al.

Page 16

Table 2

Author Manuscript

Unadjusted weighted linear regression by antigen dose Dosage (mcg)

Risk Difference in GMT

P-value

Std. Error

95% CI

3.75

-

-

-

-

7.5

49.59

0.517

73.61

−116.93, 216.11

15

36.89

0.608

69.48

−120.28, 194.07

90

−54.17

0.384

59.18

−188.05, 79.70


Author Manuscript

Author Manuscript

Author Manuscript

90

15

0.37 (0.30, 0.46)

MF59

10-0016

0.42 (0.33, 0.54) 0.52 (0.41, 0.64) 0.18 (0.16, 0.22) 0.36 (0.26, 0.47) 0.25 (0.19, 0.32) 0.54 (0.28, 0.78)

None None None None None

04-076

07-0019

07-0019

07-0019

07-0019

0.32 (0.24, 0.38)

MF59 None

0.32 (0.21, 0.45)

MF59

04-062

10-0016

04-063

0.55 (0.44, 0.64)

AS03

10-0017

0.53 (0.42, 0.63)

0.25 (0.19, 0.32)

AS03

MF59

10-0017

10-0016

7.5

0.45 (0.36, 0.57)

AS03

10-0017

VE Dose 1 (95% CI)

3.75

Adjuvant

Trial, Arm

Dosage

0.65 (0.37, 0.85)

0.83 (0.74, 0.90)

0.64 (0.49, 0.77)

0.61 (0.45, 0.75)

0.63 (0.53, 0.73)

0.66 (0.57, 0.74)

0.77 (0.63, 0.87)

0.69 (0.57, 0.81)

0.97 (0.96, 1.00)

0.79 (0.68, 0.88)

0.97 (0.95, 0.98)

0.78 (0.67, 0.87)

0.93 (0.89, 0.96)

VE Dose 2 (95% CI)

0.83 (0.33, 2.11)

0.30 (0.21, 0.43)

0.56 (0.34, 0.96)

0.30 (0.21, 0.49)

0.83 (0.56, 1.21)

0.64 (0.45, 0.95)

0.42 (0.28, 0.60)

0.46 (0.26, 0.79)

0.57 (0.44, 0.67)

0.47 (0.34, 0.68)

0.55 (0.43, 0.66)

0.32 (0.22, 0.48)

0.48 (0.38, 0.64)

Primary Response Level (95% CI)

Theoretical vaccine efficacies (VE) after dose 1 & 2 and primary response levels

Author Manuscript

Table 3 Feldstein et al. Page 17


Feldstein et al.

Page 18

Table 4

Author Manuscript

Mean theoretical vaccine efficacy, geometric mean titer and primary response level by adjuvant or no adjuvant (after the second dose, unweighted) Adjuvanted Vaccines (95% CI)

Unadjuvanted Vaccines (95% CI)

P-value

111.47 (24.23, 198.70)

32.20 (18.48, 45.93)

0.0683

Average Estimated Vaccine Efficacy

0.84 (0.74, 0.95)

0.67 (0.59, 0.75)

0.0083

Average Primary Response Level

0.47 (0.39, 0.54)

0.58 (0.32, 0.83)

0.3290

Average Geometric Mean Titer


Feldstein et al.

Page 19

Table 5

Author Manuscript

Average theoretical vaccine efficacy (VE), geometric mean titer (GMT) and primary response level (PRL) by antigen dosage (after the second dose) Dosage (mcg)

Average GMT

Average VE

Average PRL

3.75

86.05

0.86

0.40

7.5

136.11

0.88

0.51

15

111.99

0.81

0.48

90

32.20*

0.67

0.58

*

Decrease in average GMT is likely due to the lack of adjuvant present.


Safety, Reactogenicity, and Immunogenicity of Inactivated Monovalent Influenza A(H5N1) Virus Vaccine Administered With or Without AS03 Adjuvant.

Safety and immunogenicity of a vero cell culture-derived whole-virus H5N1 influenza vaccine in chronically ill and immunocompromised patients.

A new, surface-antigen-adsorbed influenza virus vaccine. I. Studies on immunogenicity in hamsters.

An oil-in-water nanoemulsion enhances immunogenicity of H5N1 vaccine in mice.

Immunogenicity and safety of AS03A-adjuvanted H5N1 influenza vaccine prepared from bulk antigen after stockpiling for 4 years.

H5N1 influenza virus-like particle vaccine protects mice from heterologous virus challenge better than whole inactivated virus.

Progress toward a universal H5N1 vaccine: a recombinant modified vaccinia virus Ankara-expressing trivalent hemagglutinin vaccine.

Immunogenicity of varicella zoster virus glycoprotein E DNA vaccine.

Safety and immunogenicity of a genetically engineered human immunodeficiency virus vaccine.

Immunogenicity trial of inactivated hepatitis A virus vaccine in human volunteers.

Immunogenicity and efficacy of a recombinant adenovirus expressing hemagglutinin from the H5N1 subtype of swine influenza virus in mice.

Potential human adaptation mutation of influenza A(H5N1) virus, Canada.

Immunogenicity of DNA Vaccine against H5N1 Containing Extended Kappa B Site: In Vivo Study in Mice and Chickens.

2005(H5N1) vaccine with MF59 adjuvant: a randomized clinical trial.

Full-genome analysis of avian influenza A(H5N1) virus from a human, North America, 2013.

Vaccine against human papilloma virus.

Enhancing poxvirus vectors vaccine immunogenicity.

Chimeric neuraminidase and mutant PB1 gene constellation improves growth and yield of H5N1 vaccine candidate virus.

H5N1 Vaccine Virus: Impact of AS03 Adjuvant.

Clinical cancer chemoprevention: From the hepatitis B virus (HBV) vaccine to the human papillomavirus (HPV) vaccine.

Vaccination with virus-like particles containing H5 antigens from three H5N1 clades protects chickens from H5N1 and H5N8 influenza viruses.

Immunogenicity of a prophylactic quadrivalent human papillomavirus L1 virus-like particle vaccine in male and female adolescent transplant recipients.

Development of broadly reactive H5N1 vaccine against different Egyptian H5N1 viruses.

Safety and Immunogenicity of a Single Low Dose or High Dose of Clade 2 Influenza A(H5N1) Inactivated Vaccine in Adults Previously Primed With Clade 1 Influenza A(H5N1) Vaccine.