14/07/2014. Disclosures and acknowledgements. Study Design (NCT ) Rationale for a QIV efficacy study in children

Transcription

1 /7/ Disclosures and acknowledgements I am employed by the GlaxoSmithKline group of companies and I own stocks/options of the GlaxoSmithKline group of companies; my travel to this meeting was funded by GlaxoSmithKline Biologicals SA EFFICACY OF INACTIVATED QUADRIVALENT INFLUENZA VACCINE (IIV) IN CHILDREN AND AN HI ANTIBODY CORRELATE OF PROTECTION The study I will present was funded by GlaxoSmithKline Biologicals SA Varsha K Jain, MD, MPH Vaccine Discovery and Development GlaxoSmithKline Vaccines Second WHO integrated meeting on development and clinical trials of influenza vaccines that induce broadly protective and long-lasting immune responses 7 May, Geneva, Switzerland Rationale for a efficacy study in children IIV efficacy estimates were not robust for children This study was done to provide direct evidence regarding IIV s clinical benefit in children using endpoints: The traditional endpoint (any influenza ) A more clinically relevant endpoint (moderate to severe influenza ) Design (NCT) Randomization : Age stratified and years N~, Surveillance for ~ days during the - influenza season Day Vaccination Blood sample Vaccination $ Blood sample* *Only for primed subjects $ Only for unprimed Day Day 6 Blood sample $ Blood sample Day Havrix TM is a trademark of the GlaxoSmithKline group of companies. ILI case count ~OCT Visit Havrix TM dose * or $ for control group

2 No. of Children /7/ Key objectives Case definitions for influenza Confirmed by RT-PCR in a nasal/throat swab Confirmatory Evaluate efficacy for the prevention of: Any RT-PCR confirmed influenza A/B (success criterion: LL 9% CI >%) Moderate to severe RT-PCR confirmed influenza A/B (success criterion: LL 97.% CI >%) Descriptive Descriptive immunogenicity (per strain GMT, SCR, SPR, SCF) Reactogenicity and safety Exploratory HI correlate of protection Any influenza is: Temperature 7. C, and One or more symptoms on the same day (cough, sore throat, runny nose or nasal congestion) Moderate to severe influenza is any influenza plus: Fever >9 C, or Physician-verified acute otitis media, or Physician-verified lower respiratory tract manifestations (shortness of breath, croup, wheezing, pulmonary congestion, bronchiolitis, bronchitis, pneumonia), or Physician-diagnosed serious extra-pulmonary complication of influenza (including myositis, myocarditis, seizure or encephalitis) (detects the more clinically relevant outcomes of influenza) 6 Countries and enrollment Subtype/lineage distribution Total vaccinated cohort - from day after vaccination HN PA DO TR LB BD TH PH Country N Dom Rep Philippines Thailand Bangladesh Honduras Panama Lebanon Turkey Influenza strains by country Overall distribution by strain B/Yam (.9%) A/HN B/Vic (.%) (.%) A/HN 7 (.7%) Demography similar between groups: mean age. ±.7 years; ~% female A/HN A/HN B/Victoria B/Yamagata Majority of children were vaccine unprimed received doses 7 Jain VK et al; N Engl J Med. Dec 6;69(6):-9

3 Geometric Mean Titer /7/ Vaccine efficacy - cases confirmed by RT-PCR Total vaccinated cohort for efficacy - from day after vaccination Immunogenicity HI antibodies Pre, post, & at least 6 month post vaccination Attack rate Vaccine efficacy (9% CI) Endpoint Group N n % % LL UL Any influenza, Havrix TM,.7 Attack rate Vaccine efficacy (97.% CI) Endpoint Group N n % % LL UL Moderate to severe influenza, Havrix TM, 6.6 N = number of subjects included in each group n = number of subjects reporting at least event in each group Vaccine efficacy assessed using Cox Regression model adjusted for age, region and priming status Jain VK et al; N Engl J Med. Dec 6;69(6):-9 9 Jain VK et al; N Engl J Med. Dec 6;69(6): Baseline Month End of Geometric Mean Titer Baseline Month End of Baseline Month End of Baseline Month End of A/HN A/HN B/Victoria B/Yamagata Control (Per-protocol immunogenicity subset) Methodology Case distribution, variables predicting outcome by logistic regression Preliminary Analysis of Correlate of Protection (unpublished data) 67 cases of HN occurred 6 to days after last vaccination matched controls per case, by age, gender, center Logistic regression w/ HI titer and treatment group as independent predictors of HN HI titer was statistically significant independent predictor

4 P(Protected) P(Protected) /7/ Methodology Dunning s scaled logit model Dunning s model used to quantify the relationship between an HI titer and the probability of developing (any & mod-severe) associated with a prevalence and an individual s chance of exposure to HN λ represent the probability that a susceptible individual develops and is estimated by the standard maximum likelihood method; α and β are parameters of the probability that an individual with titer t is protected and are estimated by the standard maximum likelihood method For cases, a linear decay model was assumed to estimate the HI titer at onset. For controls, the post-vacc HI titer was used unless the case occurred after D, when a D titer was used The Dunning model was used to estimate the HI titer associated with the probability of occurrence of influenza output was used to derive a plot of probability of protection [P(protection)] as a function of HN HI titer PDisease PExposure PProtected exp log t Protection against HN predicted by HI Titer Model predicts more protection per unit HI ab against mod to severe : Any influenza A/HN titer : A/HN titer Estimate Lower 9% CL Upper 9% CL Estimate Lower 9% CL Upper 9% CL Estimated Probability of Protection from Any and Influenza Afforded by a HI Titer of : at Time of Exposure, and HI Titers Predicted to Afford % to 9% Protection (TVC, excluding matched control subjects with MGI or confirmed influenza) Antibody Probability of protection titer Any : 6.% 67% Antibody titer Probability of protection Any :.% 77.9% :6.9%.6% :7 6.%.9% : 67.% 9.7% : 7.% 9.% : 6.7% 99.% Summary The evaluated was 7% protective against moderate to severe influenza and 9% protective against any influenza HI titer in children - YOA, whether elicited by IIV or not, is predictive of probability of HN influenza illness if exposed Immunity, correlated with HI antibody, prevents or attenuates illness, i.e low levels of HI antibody reflect immunity sufficient to prevent modsevere illness but higher levels of HI antibody (reflecting more complete immunity) are required to prevent mild illness Using the GSK HI assay: titer of : is predictive of % protection against any HN illness titer of : is predictive of % protection against mod to severe HN illness IIVs may have greater benefit for children age and older than appreciated Data limitations and next steps Data limitations Case definition of moderate to severe influenza illness needs to be clinically validated in other studies CoP model needs validation with another data set Relevance of CoP for children younger than age needs to be established Next steps Extend analysis to HN Extend analysis to B-Victoria Repeat analysis with N ab in place of HI ab Validate with an independent dataset down to 6 mos of age 6

5 /7/ BACK-UPs 7 Incidence of adverse events Total vaccinated cohort Influenza severity by subtype/lineage Total vaccinated cohort from days after vaccination Within 7 days of vaccination: (N=,) n % Havrix TM (N=,) 9% CI 9% CI n % LL UL LL UL Any AE (solicited and unsolicited) Any general AE (solicited and unsolicited) Any local AE (solicited and unsolicited) During entire study: Serious adverse events Medically attended events Grade medically attended events Medically attended events with causal relationship to vaccination Havrix TM Havrix TM Havrix TM Havrix TM n/% = number/percentage of subjects reporting at least event in each group 9