Study Calls for Alternatives to Egg-Based Manufacturing of Influenza Vaccines

Because certain viral sub-types are associated with higher rates of morbidity and mortality than others, efforts to improve the effectiveness of influenza vaccines remain an important focus of future research.

The results of a systematic review and meta-analysis of test-negative design studies recently published in Lancet Infectious Diseases suggest substantial variation in vaccine effectiveness across influenza virus types and subtypes. Because certain viral sub-types are associated with higher rates of morbidity and mortality than others, efforts to improve the effectiveness of influenza vaccines remain an important focus of future research. Additionally, the study results indicate a need for alternatives to egg-based manufacturing, as egg-induced mutations in vaccine strains can result in antigenic mismatch.

The study was presented by Edward A Belongia, MD, from the Center for Clinical Epidemiology and Population Health, Marshfield Clinic Research Foundation in Marshfield, Wisconsin. Due to the increasing number of test-negative design studies reporting vaccine effectiveness estimates separately by viral type and subtype, Belongia et al conducted a systematic review and meta-analysis of published test-negative design studies to estimate seasonal vaccine effectiveness against illness caused by both influenza type A (H3N2, H1N1pdm09, H1N1 [pre-2009]), and type B strains.

Based on the literature review search strategy and meta-analysis inclusion criteria, 56 of the 3368 unduplicated publications identified were examined in detail. The majority of studies were conducted in the Northern Hemisphere, and the number of studies from Europe and North America were roughly equivalent. Additionally, most of the included studies (93%) were published after 2010. A total of 114 vaccine effectiveness estimates based on unrestricted age enrollment were collated from the 56 included publications. The 114 estimates consisted of 34 (30%) for seasonal vaccine against H3N2, 36 (32%) against type B, 29 (25%) against H1N1pdm09, five (4%) against H1N1 (pre-2009), and ten (9%) for a monovalent vaccine against H1N1pdm09. Taking patient age into account, 33, 28, and 13 estimates were identified for pediatric age groups, working-age adults, and older adults, respectively.

Through their research, Belongia and colleagues found that, "... relevant information about patient recruitment, symptom eligibility, and vaccine ascertainment was inconsistently reported...” Collectively, the results of the study indicated that currently available influenza vaccines can be expected to afford moderate to high protection against three influenza type A subtypes (H1N1pdm09, H1N1 [pre-2009]), and type B viral strains. The reviewed data also revealed substantially lower protection in terms of vaccine effectiveness against the influenza type A H3N2 strain. This lower level of protection is of particular concern, as the influenza type A H3N2 strain has been associated with higher morbidity and mortality than are other influenza type A subtypes.

Belongia et al also assessed differences in vaccine effectiveness across age groups, which were found to be minimal for the type A H1N1pdm09 subtype and type B influenza strains. Conversely, the effectiveness of vaccines against H3N2 was highest in pediatric age groups and lowest in older adults. Regardless of antigenic match or mismatch, the effectiveness of vaccines against H3N2 was found to be low; however, Belongia and colleagues stated, "...this comparison was limited by the absence of standardized antigenic characterization and information about antigenic distance."

According to Belongia et al, the vaccine manufacturing process may be a contributing factor to the low rate of vaccine effectiveness against H3N2 due to egg-induced mutations in the hemagglutinin that can affect the vaccine's antigenicity. Based on this assertion, Belongia and colleagues stated, "An accumulating body of evidence suggests that egg-based manufacturing is not optimal for H3N2 influenza viruses that are poorly adapted for growth in eggs. A crucial need exists for alternative vaccine technologies that generate greater protection against H3N2 than do current vaccines, and product-specific vaccine effectiveness studies will be needed to assess their effect after licensure."

Because of the variability in the vaccine effectiveness estimates obtained through their research, Belongia et al made recommendations to optimize vaccine effectiveness methods in the outpatient setting. These recommendations are consistent with those from a draft of recommendations being developed by the World Health Organization for the implementation and reporting of influenza vaccine effectiveness studies using test-negative design. These include:

  • Require and report specific symptom eligibility criteria corresponding to influenza-like illness or acute cough illness. Vaccine effectiveness analyses based on a convenience sample of clinical diagnostic tests could be biased and should be avoided.
  • Define and report standard procedures for collection of respiratory samples and RT-PCR testing.
  • Restrict enrollment to patients with a duration of illness of 7 days or fewer to minimize misclassification of influenza status.
  • Exclude patients vaccinated within 14 days before illness onset because of latent period between vaccination and serological response.
  • Report source of vaccination data. Use medical records or registries to confirm vaccine receipt, dates (including previous season vaccination), and manufacturer whenever possible. Describe influenza vaccine manufacturers and products used in the study population.
  • Include parameters for age group and calendar time in vaccine effectiveness logistic regression models; studies done in multiple sites should adjust for enrollment location. Other potential confounders should be individually assessed to establish whether they change the unadjusted odds ratio by 10% or more, although this threshold is arbitrary and can be adjusted up or down. Covariates that exceed this threshold are potential confounders and should be included in the adjusted model.
  • Report vaccine effectiveness estimates by type, subtype, and lineage whenever sample size is sufficient. Report age-stratified vaccine effectiveness estimates separately for pediatric and older adult age groups.
  • Restrict vaccine effectiveness analysis to periods of continuous local influenza circulation. One approach is to exclude controls with symptom onset before the week of the first influenza-positive case and those with symptom onset after the week when the last influenza case was identified.
  • When previous season vaccination data are available, analyze the independent and combined effect of current and previous season vaccination with classification of vaccine exposure into four groups: vaccinated current season and previous season, vaccinated current season only, vaccinated previous season only, and unvaccinated in both current and previous season (referent group).

William Perlman, PhD, CMPP is a former research scientist currently working as a medical/scientific content development specialist. He earned his BA in Psychology from Johns Hopkins University, his PhD in Neuroscience at UCLA, and completed three years of postdoctoral fellowship in the Neuropathology Section of the Clinical Brain Disorders Branch of the National Institute of Mental Health.