The immunogenicity and efficacy of pneumococcal vaccines: a systematic review and network meta-analysis of individual participant data

Proposal
5819

Title of Proposed Research

Lead Researcher

Merryn Voysey

Affiliation

University of Oxford

Funding Source

Potential Conflicts of Interest

Data Sharing Agreement Date

07 June 2019

Lay Summary

Decisions about which vaccines to use in the UK should be made based on good data to make sure we are protected against disease.Pneumococcal disease is a serious illness caused by a bacterial infection that can result in death. Children in the UK receive one type of pneumococcal vaccine to prevent this disease and the vaccine in use has been very successful. Another vaccine is used in other countries, and might be a good option for the UK as it is cheaper. From studies in this area, however, there is very little informationdirectly comparing these two vaccines, so it is difficult to know which is best.We plan to do a large analysis combining all pneumococcal vaccine studies. We will use information from studies published in medical journals, but we will also include data from studies done by the companies that own the vaccines. We have used this sort of company data before and are experienced in analysing it. We aim to show which vaccine produces the highest responses when measured in blood samples in infants and also to show which vaccine protects infants best against infection.Our results will provide the best comparison of pneumococcal vaccines done so far. This will help the UK to decide whether to keep the current vaccine or to change. In turn, this will help to continue to protect the population against disease.

Study Data Provided

[{ "PostingID": 259, "Title": "GSK-110808", "Description": "Primary vaccination course in children receiving the pneumococcal vaccine GSK 1024850A or Prevenar™ co-administered with Hiberix™" },{ "PostingID": 465, "Title": "GSK-105554", "Description": "Phase IIIa randomized, controlled study to assess the immunogenicity of GlaxoSmithKline (GSK) Biologicals’ 10-valent pneumococcal conjugate vaccine, when administered as a 3-dose primary immunization course before 6 months of age" },{ "PostingID": 470, "Title": "GSK-110142", "Description": "Non-inferiority of co-administration of GSK Biologicals’pneumococcal conjugate vaccine GSK1024850A with DTPa-IPV-Hib versus co-administration with DTPa-HBV-IPV/Hib." },{ "PostingID": 2209, "Title": "GSK-105553", "Description": "Assess lot-to-lot consistency of 3 lots (double blind design) of GlaxoSmithKline Biologicals' 10-valent pneumococcal vaccine and evaluate non-inferiority to Prevenar™ (single blind design) when administered as 3-dose primary immunization course before 6 months of age" },{ "PostingID": 2210, "Title": "GSK-107005", "Description": "To assess safety, reactogenicity and immunogenicity of GSK Biologicals’ 10-valent pneumococcal conjugate vaccine, when co-administered with DTPa-combined vaccines and MenC or Hib-MenC vaccines during the first 6 months of age." },{ "PostingID": 2211, "Title": "GSK-107007", "Description": "To assess the safety, reactogenicity and immunogenicity of GSK Biologicals' pneumococcal conjugate vaccine compared to Prevenar™, co-administered with DTPw-HBV/Hib & OPV or IPV vaccines as a 3-dose primary immunization course during the first 6 months of age" },{ "PostingID": 4289, "Title": "GSK-103488", "Description": "A randomized, controlled, phase II study to evaluate the safety and immunogenicity of different formulations of GlaxoSmithKline Biologicals’ 11-valent pneumococcal conjugate vaccine, when administered intramuscularly as a 3-dose primary immunization (2-3-4 month schedule) before 6 months of age" },{ "PostingID": 19770, "Title": "GSK-113994", "Description": "Safety, reactogenicity & immunogenicity of GSK Biologicals’ pneumococcal vaccine 2189242A when co-administered with DTPa-HBV-IPV/Hib vaccine in healthy infants" },{ "PostingID": 19918, "Title": "GSK-116485", "Description": "Immunogenicity and safety study of two formulations of GlaxoSmithKline (GSK) Biologicals’ pneumococcal vaccine (2830929A and 2830930A) when administered in healthy infants" }]

Statistical Analysis Plan

Network meta-analysisNetwork meta-analysis is a method of evidence synthesis whereby pair-wise comparisons of different treatments are combined and both direct and indirect treatment effects can be computed. Such methods combine data from a larger number of studies than traditional meta-analysis thereby increasing statistical power. In addition, it is possible to estimate treatment effects for pair-wise comparisons for which no data (or only limited data) are available (indirect effects). Pair-wise comparisons for which only a limited number of studies exist, are enhanced by ‘borrowing' statistical power from the other comparisons in the network.For licensed pneumococcal vaccines the main network is of the simplest format - a network with three pairwise comparisons: PCV7 vs PCV10, PCV7 vs PCV13, and PCV10 vs PCV13. Although there are few studies available for the main comparison of interest (PCV10 vs PCV13), there are many studies in which these vaccines were compared to the older PCV7 vaccine and the inclusion of all studies into the network adds substantial statistical power to the comparison of PCV10 vs PCV13. In total we are currently aware of at least 23 studies that would eligible for inclusion in the network meta-analysis.Network meta-analysis of immunogenicityWe will conduct a network meta-analysis of the three licensed pneumococcal vaccines (PCV7, PCV10, and PCV13) to compare immunogenicity of vaccine serotypes. The primary outcome will be antibody levels measured 4- 6 weeks after vaccination and these will be combined in a network meta-analysis to estimate differences between the three vaccines. The primary comparison of interest will be PCV10 vs PCV13. As the PCV7 vaccine is no longer licensed, comparisons with this vaccine will be included to enhance statistical power and precision of the primary comparison, and to allow for indirect comparisons to be made where no head-to-head data exist.Network meta-analyses will be conducted for each serotype and each time point (e.g. at 4 weeks post-primary or 4 weeks post-booster vaccine) separately.Each trial that has individual participant data available will be analysed separately using the online portals (for company data) or offline (for data obtained directly from non-industry trial study investigators). It is unlikely that online systems for different companies will ‘speak' to each other therefore combining all individual participant data into one file is unlikely to be possible. Geometric mean ratios and standard errors will therefore be computed for each serotype and time point of interest separately for each trial and each vaccine comparison. Trial-specific estimates and standard errors will be exported and saved into an offline database and used to conduct the network meta-analysis using Stata software. If individual participant data are unavailable for some studies, published estimates and standard errors will be included in the offline database and used for the network meta-analysis of immunogenicity.Network meta-analysis of efficacyWe will conduct a second set of network meta-analyses in a similar way as the comparison of immunogenicity, but the outcome of interest will instead be sero-efficacy. For each participant we will define their recent infection status from their antibody data. This will be a binary variable equivalent to 1 if antibody levels have increased between the one-month post-primary time point and the booster visit, or 0 otherwise. During this time period antibody levels typically decline substantially, therefore when antibody increases instead of declining sharply, it is evidence of exposure to the bacteria in the intervening period. Binary data can then be compared between vaccine groups to estimate odds ratios with corresponding standard errors. This will be done for each serotype for each trial separately using the online portal to access company data. These estimates will then be exported to the offline system and included in the network meta-analysis to be conducted offline.

Publication Citation

Shuo Feng, Julie McLellan, Nicola Pidduck, Nia Roberts, Julian P.T. Higgins, Yoon Choi, Alane Izu, Mark Jit, Shabir A. Madhi, Kim Mulholland, Andrew J. Pollard, Beth Temple, Merryn Voysey Immunogenicity and seroefficacy of 10-valent and 13-valent pneumococcal conjugate vaccines: a systematic review and network meta-analysis of individual participant data eClinicalMedicine, Volume 61, 2023.
https://doi.org/10.1016/j.eclinm.2023.102073.