In a recent study published in Vaccines, researchers review the effectiveness of seasonal vaccinations against influenza, in which almost 50% of vaccinations were found to prevent disease successfully. A match between influenza strains present in vaccines and strains in local circulation was identified as the most important factor in vaccine efficacy.
Study: Seasonal Influenza Vaccine Effectiveness in Persons Aged 15–64 Years: A Systematic Review and Meta-Analysis. Image Credit: Numstocker / Shutterstock.com
Influenza and tests of vaccine effectiveness
Seasonal influenza is an acute respiratory infection caused by influenza viruses that circulate globally. Influenza is a highly infectious disease, with World Health Organization (WHO) global estimates of up to one billion cases each year, 650,000 of whom succumb to the illness.
Influenza infections peak between November to April in the Northern Hemisphere and June to October in the Southern Hemisphere. This seasonal infection pattern has prompted many nations to invest in seasonal flu vaccination campaigns.
Influenza viruses mutate at an extremely fast rate, with multiple strains simultaneously in circulation and more discovered annually. Mismatches between strains used in vaccine production and those in circulation may significantly affect vaccine effectiveness (VE), thus making measures of location-specific vaccine performance imperative.
A large body of literature exists on the VE of seasonal influenza vaccines (SIVs), predominantly comprising observational studies and randomized controlled trials (RTCs). RCTs use a metric called vaccine efficacy (VER) to measure vaccine performance.
Given the controlled test environment wherein RCTs are conducted, VERs are accurate snapshot performance measures. However, these studies are expensive, time-consuming, and rarely repeated after the introduction of SIVs, thus making them unideal in the real-time assessment of SIV effectiveness.
Clinical annual monitoring of SIV performance is predominantly conducted using the test-negative design (TND) approach, wherein individuals with infection are considered cases, and those without infection are treated as controls. Individuals reporting influenza-like illnesses (ILIs) are laboratory tested for the disease and categorized as influenza-positive individuals or controls. The specificity and sensitivity of influenza diagnostic tools, combined with easily accessible patient vaccination data, allow TNDs to be more useful than RCTs for real-time vaccine performance evaluation.
RCTs and TNDs are used in different contexts. RCTs are the gold standard for licensing of use and TNDs are the main tool for monitoring the annual effectiveness of the SIVs.”
Unlike RCTs, TNDs use VE, which hitherto has not been scaled to VER. Thus, results observed from these different sampling designs have not been compared and the relationship between VE and VER has yet to be established.
About the study
The aim of the present study was to evaluate vaccine performance in RCTs and TNDs using VE as a standard measure. To this end, the researchers performed a systematic review of available literature on the topic, as well as a meta-analysis of the data derived from studies that met their inclusion criteria.
All included studies were published between 2013 and 2023, with all study participants between 15-65 years of age. All participants in TND studies were required to have been vaccinated at least 14 days before ILI symptoms, and influenza in all cases was confirmed using laboratory methods.
Researchers first used a search strategy to query two online databases, Cochrane’s library and MEDLINE through PubMed. Publications were processed through multiple rounds of title, abstract and full-text screening, which culminated in 73 publications for the analysis.
Studies employing live attenuated (LAIV), trivalent inactivated (TIV), or tetravalent inactivated (QIV) vaccines were considered for the analysis. Noncommercial and monovalent vaccines were excluded due to their limited use, especially in seasonal vaccine campaigns.
Studies with low heterogeneity scores were analyzed using fixed effects models, while those with high scores were processed through mixed effect models. Data from RCTs and TNDs were normalized using a restricted estimation maximum likelihood (ML) methodology, followed by logistic regression. This allowed for presenting RCT and TND data as pooled results expressed as both VER and VE.
Of the 2,993 publications that matched the researchers’ search strategy, only 123 studies from 73 publications were used in the final review and meta-analysis. Of these, nine studies were RCTs and 114 were TNDs. The analysis dataset was globally encompassing, with representation from both the Northern and Southern Hemispheres.
RCT analyses revealed VER between -2% to 70% for sample sizes ranging from 85 to 7,515 in the nine included studies. The most commonly used vaccine was TIV, with QIV and LAIV used in one study each. Between-study heterogeneity was low, with six studies showing matches between vaccine strains and those in local circulation.
TND results revealed sample sizes between 62 and 59,150, with TIV the most used vaccine, followed by QIV in 59 and 19 studies, respectively. VE in these studies ranged from -2% to 70%; however, unlike in RCT studies, between-study heterogeneity was very high.
VE/VER normalization revealed that RCT studies overestimated VE by 10% more than matched TND studies. Nevertheless, this result was not significant, which suggests that TND is a viable and cost-effective alternative to RCT in SIV performance evaluation.
Due to the low number of available RCTs, pooled estimates for RCT studies could not be calculated. However, the analyses highlight that the match between strains used in vaccine development and those in local circulation is the most crucial factor for determining vaccine performance. Performance improvements of almost 25% were noted between matched vaccine-circulating strain pairs when compared to unmatched pairings.
TIV vaccines performed better than QIV, despite the higher number of strains included in the latter.
This was a surprising result as VE should increase with the number of strains included in the vaccine, although this was already observed in previous work on children. Our understanding is that a match between the strains included in the vaccine and those that are predominantly circulating is the most influential factor. Hence, it is not relevant to have a high number of strains in a vaccine if they do not match the strains the vaccine aims to prevent.”
LAIV vaccines were associated with extremely low effectiveness values, irrespective of the study methodology. However, these results are not generalizable, as LAIV vaccines were only documented in three of the 123 studies analyzed.
In the present study, researchers evaluated vaccine performance in preventing seasonal influenza infection using data from RCTs and TNDs, with VE as a normalization between these otherwise incomparable approaches.
While RCTs had VE values that were 10% higher as compared to TND studies, these results are insignificant, thus highlighting the substitution of TNDs as a cost-effective real-time performance monitoring alternative to RCTs. TIVs were the most commonly used commercial vaccine, with better efficacy than QIVs and LAIVs.
Pooled results could not be obtained for RCTs due to the lack of available studies; however, the current study identified vaccine strain matching with strains in local circulation as the most critical factor in VE, as this improved vaccine performance by up to 25% over unmatched pairs. This finding supports the WHO decision to establish influenza surveillance and monitoring systems to identify local strains for future vaccine development.