Literature Review on Pediatric Fluad® Influenza Vaccine Use in Children 6-72 Months of Age

Preamble

The National Advisory Committee on Immunization (NACI) provides the Public Health Agency of Canada (hereafter referred to as the Agency) with ongoing and timely medical, scientific, and public health advice relating to immunization. The Agency acknowledges that the advice and recommendations set out in this statement are based upon the best current available scientific knowledge and is disseminating this document for information purposes. People administering the vaccine should also be aware of the contents of the relevant product monograph(s). Recommendations for use and other information set out herein may differ from that set out in the product monograph(s) of the Canadian manufacturer(s) of the vaccine(s). Manufacturer(s) have sought approval of the vaccine(s) and provided evidence as to its safety and efficacy only when it is used in accordance with the product monographs. NACI members and liaison members conduct themselves within the context of the Agency's Policy on Conflict of Interest, including yearly declaration of potential conflict of interest.

Table of Contents

• Executive Summary
• Introduction
• Literature Review Methods
  • Search strategy
  • Eligibility Screening
  • Quality assessment of studies
  • Data Extraction
• Fluad^®in paediatric population (6-72 months)
  • Burden of influenza illness in children aged 6 to 72 months
  • Efficacy and Effectiveness
  • Immunogenicity
  • Safety
  • Evidence Gaps
  • Discussion
• List of Abbreviations
• Acknowledgments
• References
• Appendix A: Search Strategy
• Appendix B: Attrition Flow diagram
• Appendix C: Summary of evidence related to the efficacy/effectiveness of Fluad^®in children 6 to 72 months of age
• Appendix D: Summary of evidence related to the immunogenicity of Fluad^®in children 6 to 72 months of age
• Appendix E: Summary of evidence related to the safety of Fluad^®in children 6 to 72 months of age

Executive Summary

Despite the significant disease burden associated with influenza infections in children younger than 6 years of age, currently used inactivated unadjuvanted trivalent influenza vaccines (UTIVs) have limited effectiveness, especially after a single dose in unprimed children, during seasons with a mismatched strain and against Influenza B types. There are a limited number of studies on the effectiveness of UTIVs among children <2 years. Although FluMist^®, a nasally administered live attenuated vaccine (LAIV), is reasonably effective among younger children, it is not recommended for children <2 years. New strategies are needed for effective influenza disease control in this age group.

Until recently, seasonal influenza vaccines approved for use in Canada did not include adjuvants. In 2011, Fluad^®, an inactivated subunit adjuvanted trivalent influenza vaccine (ATIV), containing MF59™ as the adjuvant, was approved for use in older (≥65 years) adults for active immunization against influenza. Adjuvanted vaccines were used extensively during the 2009 H1N1 pandemic; a monovalent AS03 adjuvanted vaccine was used in Canada. However, the discussion of adjuvanted pandemic vaccines is beyond the scope of this review and has been excluded.

In anticipation of the availability of Fluad™ Pediatric vaccine, a systematic review of the literature was conducted to provide evidence to inform recommendations on using Fluad^®for the prevention of influenza in children aged 6 to 72 months. On November 28, 2014, after completion of this review of the literature, Fluad^®Pediatric Influenza Vaccine was approved for use in children 6 months to less than two years of age. The NACI review was not limited to children age 6 months to less than 2 years, as the product indications for Canada were not known when the review was being conducted, and there were some data available in the literature for children older than 24 months receiving ATIV.

A systematic review of the literature was conducted to provide evidence to inform recommendations on using ATIV for the prevention of influenza in children aged 6 to 72 months. Only 8 articles met all pre-specified inclusion criteria, of which 7 articles were from clinical trials that compared ATIV with a UTIV comparator and one from a study that looked at ATIV in healthy children compared to children with juvenile arthritis.

Based on this review, information on comparative efficacy is available only from a single trial suggesting a higher efficacy for ATIV (86% vs. 40% for comparator UTIV). The bulk of the data was obtained during one mild season that was dominated by influenza A/H3N2. The comparator vaccine induced a poorer immune response compared to equivalent UTIVs. There were also concerns raised by a European Medicine Agency inspection about the quality of diagnostic laboratory testing and validity of ascertainment of influenza cases.

There is limited but consistent evidence that ATIV is more immunogenic than comparable UTIVs against influenza A types. In particular, a single dose of ATIV is more immunogenic than a single UTIV dose. However, two doses of ATIV are still necessary to achieve satisfactory immune response against influenza B. As immunogenicity is an intermediate outcome, it is unclear what clinical protection is conferred.

ATIV was not compared directly to LAIV or to the quadrivalent influenza vaccine (QIV). One study using a dose-ranging factorial design with adjuvanted and unadjuvanted versions of seasonal TIV and QIV was identified, but a comparison of ATIV and QIV was not the primary objective of the study and the data were not grouped appropriately for such a comparison. Clinical trials comparing efficacy, immunogenicity, and safety of a single- and two-dose regimes of ATIV, LAIV and QIV are needed.

Safety data for ATIV in children is consistent with what is known about ATIV safety in adults. ATIV results in 10-15% more solicited local and systemic reactions compared to UTIV. However, most reactions are mild and tend to resolve quickly. Severe reactions are rare, but several of the reviewed studies were too small to detect clinically significant but rare adverse events. In particular, the safety information is limited for ATIV in children with immunodeficiencies and other chronic illnesses.

Taken together, the limited body of evidence identified in this review suggests that ATIV is likely both more immunogenic and more reactogenic than UTIV among children 6-72 months of age. There are insufficient data to assess whether ATIV is more effective than UTIV or LAIV in practice or to make an informed risk-benefit analysis.

Introduction

Influenza remains a significant cause of morbidity in children. Each year, 20-30% of children become infected with influenza.^{Footnote 1} Most of these infections are typically asymptomatic or associated with a mild self-limiting illness.^{Footnote 2} However, influenza can cause severe illness leading to hospitalization and death, especially among infants and children with underlying chronic conditions.^{Footnote 2,Footnote 3}

The burden of pediatric influenza on the healthcare system is illustrated by the fact that up to 20% of pediatric acute care visits with fever or respiratory symptoms during influenza season are due to influenza.^{Footnote 4} In addition, children shed the virus for longer and more prolifically than adults, playing a major role in spreading influenza in their communities during annual influenza outbreaks.^{Footnote 5} Therefore, prevention of influenza infection among children can bring significant health benefits to both children and their communities.

Vaccination is the primary public health strategy for preventing influenza and reducing the impact of influenza epidemics.^{Footnote 6} Most provinces and territories in Canada offer universal influenza vaccination programs to all residents. Provinces without universal vaccination programs target groups at higher risk for infection including children.^{Footnote 7}

There are several seasonal influenza vaccine preparations currently authorized for pediatric use in Canada. The National Advisory Committee on Immunization (NACI) recommends that FluMist^®, a nasally administered trivalent live attenuated influenza vaccine (LAIV), be used for healthy children and adolescents 2-17 years of age. There is evidence supporting preferential use in young children (<6 years of age), based on superior efficacy of LAIV compared to TIV, with weaker evidence in older children. It is anticipated that superior efficacy of LAIV extends beyond age 6 years, although the age at which efficacy between LAIV and TIV become equivalent is unknown. If LAIV is contraindicated or is unavailable, unadjuvanted trivalent inactivated vaccine (UTIV) should be used. For children 6-23 months of age, only UTIV is recommended. Use of currently available influenza vaccines is not recommended for infants <6 months of age, because of limited efficacy.^{Footnote 6}

Compared to UTIVs, LAIV, in children, induces a stronger systemic antibody response, as measured by serum levels of antibodies against hemagglutinin (HA) antigen of targeted strains, and a strong mucosal response, as measured by levels of nasal mucosal IgA antibodies.^{Footnote 8} However, LAIV is not currently recommended for children aged less than two years because of higher incidence of wheezing (4%) within 42 days of administration, compared to about 2% following UTIVs noted among children less than 12 months of age.^{Footnote 9} For the same reason, LAIV is not recommended for children with severe or unstable asthma. Because it contains a live virus, the vaccine is also contraindicated for children with immune compromising conditions.^{Footnote 6} There is also a theoretical risk of reassortment with wild type influenza and possible spread to immunocompromised individuals.^{Footnote 10}

The evidence for the effectiveness of the more commonly used conventional UTIVs (either inactivated split virus or subunit TIVs) is limited. Osterholm et al found no papers that met their strict inclusion criteria for UTIV effectiveness in children 2-17 years old.^{Footnote 11} The findings of other reviews with less stringent criteria are difficult to interpret because the reviewed studies varied significantly in terms of their methods, intensity of case finding, outcomes assessed, age ranges of participants, circulating strains etc.^{Footnote 10,Footnote 12,Footnote 13} In an extensive review of the literature conducted as part of a recently published Cochrane review, Jefferson et al found, based on the analysis of 2 small clinical trials, that UTIVs were about 39% (95%CI: 0-66%) efficacious in preventing influenza among children younger than 6 years of age.^{Footnote 13} Corresponding estimates from observational studies varied by study design ranging from 37 to 55%. No usable evidence was identified for children younger than 2 years.^{Footnote 13} There was no evidence that UTIVs reduced the risk of lower respiratory tract disease or otitis media and weak evidence that UTIVs may reduce school absenteeism.^{Footnote 13} In addition, there is evidence that UTIVs perform particularly poorly among young children in seasons with significant strain mismatch.^{Footnote 10}

Prior to 2011, seasonal influenza vaccines approved for use in Canada did not include adjuvants. In 2011, Fluad^®, an inactivated and MF59™ adjuvanted subunit TIV, was approved for use in older (≥65 years) adults for active immunization against influenza.^{Footnote 14} MF59™ refers to a group of squalene-in-water adjuvants manufactured by Novartis Vaccines and Diagnostics. The specific adjuvant used in Fluad^®(MF59C.1) is an emulsion composed of squalene as the oil phase, stabilised with two non-ionic surfactants (polysorbate 80 and sorbitan trioleate), in a citrate buffer.^{Footnote 14} Squalene, a highly unsaturated hydrocarbon synthesized in the liver, is a precursor for all mammalian sterols, including cholesterol, steroid hormones and vitamin D, and principal hydrocarbon of human cell surface lipids.^{Footnote 15} Squalene used in the production of MF59C.1 is derived from shark liver oil. Squalene is also the primary ingredient of another adjuvant (AS03), manufactured by GSK, which was used in several monovalent 2009 pandemic influenza vaccines (e.g., Pandemrix^®and Arepanrix^®) as well as in candidate H5N1 vaccines.^{Footnote 16}

The adult formulation of Fluad^®was first authorized in Italy in 1997 for use in the prevention of influenza in older adults and is now authorized in >30 countries for the same indication. The immunogenicity and safety of Fluad^®was evaluated using clinical trials in different age populations, and there is considerable post-marketing experience with its use in adults (reviewed in a previous NACI statement^{Footnote 15}). NACI recommended that Fluad^®can be used for the prevention of influenza in older (≥65) adults, but the evidence was insufficient to make a recommendation for its preferential use over other influenza vaccines available in Canada.^{Footnote 15}

This systematic review of the literature was conducted to provide evidence to inform recommendations on using Fluad^®for the prevention of influenza in children aged 6 to 72 months. Fluad™ Pediatric received a Notice of Compliance from Health Canada on November 28, 2014, after the initiation of this review, for use in children 6 months to <2 years. Authorization for use in Canada was based on immunogenicity and safety data.

Literature review methods

Search Strategy

The literature search was conducted in three primary electronic databases: Medline, Embase and Web of Science, employing a strategy based on using Medical Subject Headings (MeSH) for Medline, Emtree indexing for Embase, and keyword searching for all three databases. Specifically, the Ovid interface was used to search the following databases: Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations, Ovid MEDLINE(R) 1946 to Present, Embase 1974 to 2014 and the Web of Science Core Collection.

The general search strategy was (Fluad^®) OR (influenza vaccine AND adjuvant) AND (Pediatric), limited to human studies (See Appendix A or details of the search algorithm). All searches were limited to literature published between June 1st, 2004 and June 30th, 2014. The search strategy was applied to all 3 primary electronic databases on July 18th, 2014, yielding 1,529 articles.

Three additional databases were searched: clinicaltrials.gov, CINAHL and Cochrane Library. The general search terms for these databases consisted of "Fluad^®OR MF59", with no additional restrictions to maximize yield from these small databases. This search was conducted July 21st, 2014 yielding 232 articles.

To capture a wider breadth of articles, particularly concerning the burden of disease, an additional search was conducted using the following general search strategy: "influenza" and "pediatric" and "Canada", limited to papers published after 2004. This search was applied to the same three primary databases: Medline, EMBASE and Web of Science. Various source of "grey literature" were also searched; these included the databases of the Canadian Agency for Drugs and Technologies in Health (CADTH), PHAC (e.g. FluWatch) and the WHO Tables on the Clinical Evaluation of Vaccines. In addition, forward (citing articles) and backward (cited articles) searching was performed to identify possible relevant articles. This was done by checking Google Scholar, Scopus and the citations of the included studies. Forward and backward searching yielded 1 additional study. One additional published poster was provided by the manufacturer through a project contact at the International Centre for Infectious Disease (ICID).

If possible, email alerts were set in databases to notify the literature review group of newly published citations which met the search strategy criteria set for each database. No relevant literature was identified by email alert. The total number of records identified through bibliographic database searching was 1,761 and 2 records were identified through other sources. After removing duplicates, 1,173 citations were identified for screening (Appendix B).

Eligibility Screening

Two independent reviewers screened all titles and abstracts for relevance based on the following predetermined eligibility criteria: (i) primary research studies, regardless of design, where the vaccine assessed is Fluad^®; (ii) the population age range overlaps with the age group of interest (6 to 72 months); and (iii) publication date between June 1st, 2004 and June 30th, 2014. Case reports, case series, and opinion papers were excluded.

Upon applying these criteria, 1,116 records were excluded. Of the remaining 57 articles, 29 were Fluad^®-related articles and were retrieved for full-text review. The rest (28) comprised articles on other MF59 adjuvanted influenza vaccines (e.g., the monovalent pandemic H1N1 vaccine).

Articles retrieved for full-text review were only excluded if they were assessed as ineligible by two independent reviewers. Any disagreements were resolved by consulting a third reviewer. Articles were excluded if, after reviewing the full text, they were deemed not to meet the inclusion criteria (outlined above). On this basis, 13 articles were excluded for the following reasons: age group outside range (n=1); intervention not Fluad^®(3); no outcome of interest (1) and article was neither primary research nor a systematic review (8). Eight articles were included in the detailed review, of which 7 articles described controlled clinical trials that compared Fluad^®with one or more UTIV comparators and one from a uncontrolled trial that looked at Fluad^®in healthy compared to arthritic children. All studies included at least one age group which overlapped with the 6-72 months of age criterion. In addition, several articles (relevant primary research, reviews etc.) were also retained to aid in writing the literature review but were not included in the summary tables (Appendix B).

Quality Assessment of Studies

Articles retained for review were critically appraised independently by two reviewers in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and the design-specific criteria outlined in Harris et al.^{Footnote 17} On the basis of the information included in the articles, all included studies were rated "fair" or better and no studies were excluded due to poor quality.

Data Extraction

Two reviewers independently extracted data from the included articles using a common abstraction form designed to capture data for each of the outcomes of interest: vaccine efficacy, measures of immunogenic responses (e.g., seroconversion, seroprotection and geometric mean titre) and safety outcomes (solicited, unsolicited, and serious adverse events). The form also captured factors that might influence study findings such as different vaccine formulations, influenza seasons, participants' age groups and inclusion criteria. Any disagreements or discrepancies between abstractors were resolved by discussion or by consulting a third reviewer.

Fluad^®In Paediatric Population (6-72 Months)

Burden of influenza illness in children aged 6 to 72 months

The burden of disease due to influenza in young children is substantial, although rarely precisely known. Nair et al analyzed data from 43 population-based studies that collectively recruited more than 8 million children younger than 5 years from countries representing all WHO regions.^{Footnote 18} They estimated that about 90 million influenza infections, 20 million cases of influenza-associated acute lower respiratory infection (ALRI) and 1 million cases of severe ALRI occur worldwide each year among children <5 years of age.^{Footnote 18} They also estimated that influenza was responsible for about 13% of all ALRIs and 7% of all severe ALRIs in this age group.

Influenza incidence rates vary from year to year, depending on the circulating strain, population susceptibility and other environmental factors. In developed countries, the pooled incidence rate (95%CI) of influenza episodes was estimated by Nair et al to be around 550 (280-1000) per 10,000 children younger than 5 years per year, whereas the rates of ALRI and severe ALRI due to influenza were around 120 (70-180) and 10 (10-20) per 10,000 person-years, respectively.^{Footnote 18} These rates were approximately 2-3 times higher in developing countries. These figures likely underestimate the true burden of influenza, as most studies were based on passive hospital-based surveillance systems and employed relatively insensitive rapid assays for influenza detection. The WHO estimates that during interpandemic seasons, 20-30% of children become infected with influenza.^{Footnote 1} In certain populations, e.g., daycare attendees, attack rates may exceed 50%.^{Footnote 19,Footnote 20}

A child's age is a very important determinant of the risk of influenza infection. Evidence from seroprevalence studies suggests that most children would have been exposed to at least one influenza A virus by age 6.^{Footnote 19} By contrast, only a small percentage (20-30%) of 0-6 year-olds have influenza B virus-specific IgG antibodies,^{Footnote 21} supporting the notion that unvaccinated children accumulate natural immunity to influenza B more slowly than they do to influenza A.^{Footnote 22}

In Canada, and depending on the season (2003/04 through 2011/12), children under 4 years of age comprised between 10 and 33% of all laboratory confirmed influenza cases reported through the sentinel laboratory-based Respiratory Virus Detections Surveillance System (RVDSS), whereas those aged 5-9 years comprised between 5 and 12%. By age group, children under 4 years and children 5-9 years each comprise approximately 5-6% of the Canadian population. In addition to higher infection rates, these figures may also reflect increased utilization of acute care services and higher likelihood of laboratory testing among younger children. During influenza seasons, children under 4 years of age with ILI represent between 3 to 7% of all primary care visits made by patients to sentinel clinicians participating in the national FluWatch influenza surveillance program, reflecting the burden influenza (and other respiratory pathogens) imposes on the healthcare system. However, the high rates of laboratory confirmed influenza in young children observed in the RVDSS data are consistent with observations from active surveillance prospective studies conducted elsewhere.^{Footnote 18,Footnote 19}

Young age is also an important risk factor for complications, serious illness and hospitalization among those infected with influenza. Generally, children younger than 2 years are at an increased risk of serious influenza illness although older (5-9 years) children are more likely to contract influenza.^{Footnote 1} For instance, in developed countries including Canada, estimates of the rates of hospitalizations attributed to influenza among infants ranged from 10 to over 100 per 10,000.^{Footnote 19,Footnote 23,Footnote 24,Footnote 25} In one Canadian study, the highest rate of hospitalization due to influenza was among infants 6 to 11 months of age (20 per 10,000 infants per year), a rate that was comparable to the admission rate among 60-65 year-olds.^{Footnote 24} Nevertheless, respiratory syncytial virus and para-influenza infections were still more frequent among hospitalized infants than influenza, even during periods of peak influenza transmission.^{Footnote 24}

Complications of influenza, such as acute otitis media (25-75%), are very common especially among children younger than 2 years old.^{Footnote 19} Influenza can also cause croup and bronchiolitis and exacerbate asthma symptoms. More severe complications include bacterial co-infections, e.g., staphylococcal pneumonia, acute respiratory distress syndrome (ARDS), encephalitis and Guillain-Barre's syndrome.^{Footnote 19,Footnote 25,Footnote 26} Generally, rates of specific complications are not precisely known, but it is widely accepted that they are more common among very young children and those with underlying chronic diseases.^{Footnote 10,Footnote 25,Footnote 27}

Rates of severe respiratory illness due to influenza tend to decline with increasing age in children.^{Footnote 19} For instance, in the systematic review by Nair et al, the pooled rate of influenza-related severe ALRI was twice as high among infants (20/10,000) compared to all < 5 years-old.^{Footnote 18} In children above 5 years, the rate of hospitalization is only about 5/10,000 and a higher proportion of these children have underlying medical conditions. Over the period from 2003/04 to 2011/12, the number of influenza-associated pediatric hospitalizations, as reported by the Immunization Monitoring Program Active network (IMPACT) from participating Canadian centres ranged from 370 to 948 patients per season. Between 11-23% of all reported pediatric hospitalizations were among 0-5 month-old infants, 8-32% were among 6-12 month-olds and a further 20 to 27% were reported for 2-4 year-olds. Overall, about 70-75% of all pediatric hospitalizations occurred among children under the age of 5, except during the 2009 pandemic where this age group represented less than 60% of all pediatric hospitalizations.

In the IMPACT data, the majority (60-99%) of pediatric hospitalizations were due to influenza A, except for the 2011/12 season when influenza B was identified in about 60% of pediatric hospitalizations. Consistently (excluding during the 2009 pandemic), about 50% of admitted children had an underlying medical condition. About 10-15% of all children admitted to hospital were also admitted to an Intensive Care Unit.

Excluding the pandemic year, the number of deaths attributed to influenza in Canadian children was small, ranging from 2-6 per season, and representing <0.8% of all children admitted with confirmed influenza-associated illness at participating IMPACT sites. About half of these deaths were among 6-23 month-olds, corresponding to about 1.5% of influenza hospitalizations in this age group. The average case fatality ratio (relative to all laboratory confirmed infection among children) is much smaller (<0.15%) and likely many folds smaller relative to all influenza infections. These figures are consistent with estimates from other developed countries where mortality rates were generally < 1 per 100,000 person-years and the pooled case fatality relative to influenza associated severe ALRI was 0.17% (95%CI 0.08-0,26%).^{Footnote 19} These figures are likely 10-20 fold higher in developing countries.^{Footnote 19}

In a country-wide surveillance study in the United States during the 2003/04 season, Bhat et al identified 153 influenza associated deaths among children. Mortality rates were highest among children < 6 months of age (about 9 per 100,000 children per year) and decreased with increasing age to about 0.11/100,000 among 5-17 year-olds.^{Footnote 26} About 50% of children had one or more underlying chronic conditions, most often respiratory or cardiac conditions. About 40% of fatalities occurred at home, in transit to hospital or in an emergency department, highlighting the sudden and unexpected nature of these deaths.^{Footnote 26} Only 16% of the children who died and had known vaccination status had received at least one dose of influenza vaccine during the 2003/04 season.

Efficacy and Effectiveness

A randomized control trial (Novartis identifier: V70P5) involving children aged 6 to 72 months conducted over two consecutive seasons (2007/08 and 2008/09) in Germany and Finland, was the only study that provided data on Fluad^®(henceforth ATIV) efficacy against a disease (non-serological) endpoint (PCR-confirmed influenza).^{Footnote 28} Children (N=4707) were randomized 2:2:1 into those receiving 2 doses, 28 days apart, of either ATIV, an UTIV or a control non-influenza vaccine (a meningococcal C conjugate or encephalitis vaccine). The ATIV and UTIV used contained comparable doses of the WHO-recommended vaccine strains. For children aged 6 to < 36 months, a half dose (0.25mL) of either vaccine was administered, whereas a full dose (0.5mL) was administered for children aged 36 to < 72 months. The three study groups were comparable in demographic characteristics and attrition rates.

The attack rates of PCR-confirmed influenza among the control groups were relatively low in both seasons: 2.5% in 2007/08 and 5.2% in 2008/09. Vaccine efficacy (VE) estimates were calculated for the 2008/09 season which was dominated by a vaccine-matched A/H3N2 strain (90% of all isolates; 10% were lineage-mismatched B strains). VE was not calculated for 2007/08 due to a small case count. In the 2008/09 season, ATIV was 86% (95%CI: 73-92%) effective (relative to the control non-influenza vaccines) against all strains compared to 40% (11-60%) for UTIV (relative to the control non-influenza vaccines). The efficacy of ATIV relative to UTIV was 76% (55-87%). ATIV had a lower VE (79%) in 6-36 month-olds compared to 92% in children 36 to 72 months of age. The corresponding values for UTIV were 40% and 45%.

VE estimates were very similar when data for both seasons were combined and when detection of vaccine-matched strains was the outcome. Because of the predominance of A/H3N2, the authors noted that no conclusions could be drawn with regard to VE against A/H1N1 or influenza B. Furthermore, UTIV used in 2008/09 was a split-virion product manufactured by GSK Biologicals, as opposed to the subunit UTIV manufactured by Novartis Vaccines that was used in the prior (non-contributory) season. The strain-specific antibody titres of the GSK UTIV formulation were shown to be 48-63% lower among children 6 to 35 months of age than a comparable Sanofi Pasteur formulation in an unrelated study conducted during the 2006/07 season.^{Footnote 30,Footnote 31}

The European Medicines Agency (EMA) raised concerns regarding trial V70P5 during the regulatory approval process in 2011. Critical and major flaws in good clinical practice (GCP) were identified following an inspection of the sites of the main study.^{Footnote 31} Some of these flaws included:

• insufficient quality assurance system at the laboratory site in Germany,
• reliability of patient data collection (including recording of adverse events and tracking of suspected influenza cases) and data handling, and
• issues with sample storage and transport, including lack of temperature monitoring (some samples may have reached temperatures that resulted in sample degradation).

As evidenced in the product monograph, data from the efficacy trial was also not considered in granting product authorization in Canada. NACI considers the findings from the report of the EMA should be taken into account when assessing the results from the study.

Immunogenicity

In all trials, the European Union (EU) Committee's for Medicinal Products for Human Use (CHMP) licensing criteria (CPMP/BWP/214/96) for interpandemic influenza vaccines (as set for 18-60 years-old adults) were used to benchmark observed immune responses. Regulators in Canada, the US and Europe accept immunologic data (HI antibody titers) as acceptable surrogates for clinical protection, and the same criteria are used for adults and children. An HI titre of ≥1:40 has been recognized as an immunologic correlate that corresponds to a 50% reduction in the risk of contracting influenza in adults (on the basis of a study by Hobson et al).^{Footnote 34} However, evidence is lacking on the HI titre required to confer the same level of protection in children.^{Footnote 35}

Recent evaluation of immunologic correlates of protection in children suggests that higher titers may be necessary to confer the same level of protection as in adults. Using data from the immunogenicity cohort of the V70P5 trial, Black et al estimated that an HI titre of ≥1:40 was associated with a risk reduction of only 22% in previously unvaccinated healthy 6-72 months-old children.^{Footnote 35} A much higher cutoff (1:110) was associated with 50% clinical protection rate, whereas titres of 1:330 and 1:629 were associated with 80% and 90% clinical protection, respectively. Similar to the Hobson study^{Footnote 34}, the analysis by Black et al was based on protection against infection with a vaccine matched A/H3N2 strain. It is unclear whether the findings would be applicable to protection against A/H1N1, B types or mismatched strains. It is also unclear whether these estimates are equally applicable to adjuvanted and unadjuvanted vaccines as there could be significant differences in the avidity of the generated antibodies or other aspects of the immune response due to the presence of the adjuvant.^{Footnote 35} Finally, the estimates were imprecise, with relatively wide confidence intervals, because they were based on a very small number of cases (n=22).

In addition to the abovementioned trial (V70P5),^{Footnote 28} two Phase 2 studies (V70P2^{Footnote 32} and V70P6^{Footnote 33}) assessed the immunogenicity of ATIV compared with UTIV in healthy previously unvaccinated children aged 6-36 months. V70P6 also provided data on children 36-59 months of age. In all three studies, two doses of vaccine were given 28 days apart. Children 6-35 months of age received 0.25mL of vaccine per dose and children 36-72 months of age received 0.5mL per dose. All studies used a hemagglutination inhibition (HI) assay to measure immune response before each dose (Day 1 and Day 29) and approximately 3 weeks after the second dose (Day 50).

Using the standard CHMP immunogenicity criteria set for adults 18-60 years of age, V70P5, V70P2 and V70P6 were consistent in showing that ATIV met at least two CHMP criteria for protection against Influenza A strains after a single dose but a second dose was necessary for a sufficient response to the B strain. Generally, ATIV induced a weaker HI response against B strains compared to A strains. However, ATIV induced a stronger immune response against B strains when compared with UTIVs. Similarly, ATIV induced a stronger response to heterologous A strains than to heterologous B strains, and in both cases, response was more robust than that observed for comparator vaccines.

A similar pattern was seen in a dose- and schedule-finding study that compared adjuvanted and unadjuvanted formulations of both trivalent and quadrivalent influenza vaccines (V70P6).^{Footnote 33} This study was conducted using a factorial design, with varying doses of both HA and the MF59 adjuvant. The study found that the addition of adjuvant conferred a clear increase in Geometric mean titres (GMT) with further small increases with increasing MF59 dose. Children receiving one dose of ATIV, or the quadrivalent equivalent, were well protected against influenza A strains, but few (17%) were seroprotected against B strains by day 29 (just before the second dose). Between 79-97% of children who received two doses of a vaccine that contained either 7.5 or 15µg HA (per strain) and 50% of the adult dose of the MF59 adjuvant were seroprotected against the B strains by day 50. By comparison, UTIV with 7.5 and 15µg of HA, did not meet the adult CHMP criteria for seroprotection for the B strains even after a second dose. The study also found that the addition of a second B strain did not affect the immunogenicity of the vaccine against other strains.

The extension study (V70P2E1^{Footnote 36}) assessed the immunogenicity of a third ("booster") dose of either ATIV or a split virus UTIV given in the following season, one year after the two priming doses given to the children participating in V70P2, who previously received MF59-adjuvanted TIV. In this small trial, both ATIV and UTIV induced 100% seroprotection (HI ≥1:40) against influenza A strains, but only ATIV induced 100% seroprotection against influenza B (compared to 68% for UTIV). Generally, ATIV induced higher GMTs than UTIV especially among younger children.

In a recently published phase 2 trial (V70_34)^{Footnote 37}, Zedda et al, applying more stringent criteria of seroprotection (HI ≥1:330), reported a higher seroprotection rate for ATIV (80-92%), compared to UTIV (28-53%), against influenza A strains among healthy 6-36 month-olds receiving two 0.25ml doses four weeks apart. Comparable figures for B/Brisbane were much lower: 40% in ATIV and 10% in UTIV. In addition, ATIV induced the expansion of CD4+ T cells specific for both homologous and heterologous strains, whereas UTIV induced a significant expansion of CD4+ T cells for homologous strains only. There was however no difference in the cytokine profile of influenza specific CD4+ T cells, which was dominated by the production of IL-2 and TNF-a for both vaccines.

One published poster describing a multi-site clinical trial (V70_29) conducted in Argentina, Australia, Chile, Philippines, and South Africa (n=6100) was provided by Novartis^®.^{Footnote 38} The information provided included immunogenicity data comparing children who received a sub-unit type TIV (UTIV-1) or split-virion type TIV (UTIV-2) with children who received ATIV. Using more stringent seroprotection criteria (HI ≥1:330), a higher percentage of participants were seroprotected after a second dose (Day 50) against homologous influenza A/H1N1 strains if they received ATIV (91%) compared to UTIV-1 (56%) and UTIV-2 (61%). Comparable figures were reported for the homologous A/H3N2 (ATIV vs UTIV-1 vs UTIV-2: 95% vs 66% vs 74%) but the corresponding figures were much lower for B/Brisbane (56% vs 28% vs 25%) although still higher for ATIV than either UTIV. In addition, ATIV recipients maintained higher GMTs than UTIV-1 and UTIV-2 recipients against the three homologous strains on Day 209. This study was subsequently published after the literature review was conducted, and it was reported that re-defined non-inferiority criteria were met by ATIV for geometric mean titers (GMTs) and seroconversion rates when compared to UTIV-1 and UTIV-2 at Day 50. ATIV also met the superiority criteria compared to UTIV-1, but not compared to UTIV-2.^{Footnote 39}

Only one study investigated immunogenicity of ATIV in children and adolescents (mean age 9±5 years) with an autoimmune disease, juvenile idiopathic arthritis (JIA).^{Footnote 40} This study compared healthy boys with boys diagnosed with JIA being treated with anti-rheumatic drugs. The results indicate that immunogenicity of ATIV in JIA patients treated with Disease-Modifying Anti-Rheumatic Drugs (methotrexate or sulfasalazine) was not significantly different from that of healthy children of the same age. However, among those who received the TNFa antagonist etanercept alone or in combination with methotrexate, immunogenicity was good against influenza A strains (albeit with lower GMTs), but significantly impaired in the case of influenza B strains compared to children who were not on etanercept.

Safety

Clinical safety data were provided in all identified pediatric ATIV efficacy and immunogenicity trials. In addition, Black et al pooled safety data from several pediatric ATIV trials and other trials an MF95-adjuvanted influenza vaccine was used.Footnote 41 No observational studies were found. "Serious" adverse events (AEs) were rare and occurred at a similar rate among the recipients of ATIV and comparator vaccines. However, of the seven studies with safety data, five had fewer than 500 participants and limited power to detect rare but important AEs. There is no post-market experience with ATIV in children; however, there is considerable post-marketing experience with ATIV use in older adults (reviewed in a previous NACI statementFootnote 14).

Across all pediatric ATIV trials, and regardless of age group or comparator vaccine type, additional 10-15% of ATIV recipients experienced solicited local and systemic reactions compared to UTIV recipients. Most reactions were described as "mild to moderate" in severity and self-resolved in 2-3 days. There was no significant difference in the incidence of local AEs following first or second vaccinations. On the other hand, the incidence of systemic AEs was slightly lower following the second vaccination. Severe reactions were rare (<1%) in all vaccine groups.

Among 3-36 month-olds, the most frequently reported local reactions were: tenderness (range: 18-33% ATIV, 15-26% comparator) and erythema (15-25% ATIV, 11-22% comparator). Severe local reactions including severe ecchymosis, induration, erythema or tenderness, occurred in <1% of subjects regardless of vaccine group. There was no increased incidence of local reactions after a third dose given in the following year in the extension study (V70P2E1). Similar trends were reported for older (>36 months) children, although, as expected, pain at the injection site was the most frequently reported local reaction (36-48% ATIV; 27-35% comparator). Severe local reactions were slightly more frequent (1-3%) in this age group compared to younger children.

Among 3-36 month-olds, the most frequently reported systemic reaction was irritability (21-32% ATIV, 19-26% comparator). Unusual crying and changes in eating habits were also common (about 15-25% and slightly more common among ATIV recipients). Fever (≥38°C) was reported for 5-25% of children, regardless of vaccine group. Body temperature of ≥40°C was reported for <1% of children in V07P5 with similar incidence in each vaccine group. Among older children, the most frequently reported systemic reactions were: malaise (about 15% in all groups) and fatigue (21-33% ATIV, 17-20% comparator). Fever occurred at a comparable rate to that among younger children.

Incidence of unsolicited AEs was comparable between ATIV and active control vaccine groups, with most reported events representing the usual common childhood illnesses (e.g., cough, otitis media, rhinitis, URTI, diarrhea etc.). Incidence of unsolicited AEs deemed to be possibly or probably related to vaccination (e.g., fever, rash, etc.) was also similar (collectively about 10% within 3 weeks of vaccine administration in each vaccine group).

Severe AEs (AEs that prevent a person from performing normal daily activities) were infrequent in both vaccine groups and were very rarely (<1%) linked to vaccine administration. These include very few (1-2 each) cases of cyanosis, abnormal behaviour, asthma, autoimmune thyroiditis, ITP and epilepsy. In particular, there was no evidence of increased incidence of febrile or other convulsions in the ATIV group. One death occurred in the comparator group of trial V70P6 and was not related to vaccine administration.

Nolan et al provided safety data from the abovementioned multi-site trial (V70_29).^{Footnote 38,Footnote 39} Solicited AEs possibly related to vaccination occurred with similar frequency in all three vaccine groups (5%-7% of subjects), with events occurring slightly more frequently following the first vaccination than the second. Rates of unsolicited AEs were very low and similar in the ATIV and subunit and split UTIV groups (<1% of subjects), and rates of serious AEs (fatal or life-threatening AEs or AEs resulting in hospitalization or prolongation of hospitalization or major disability) were similar across all three vaccine group (4-5% of subjects in each group). Rates of fever (≥38°C) in 24 to <72 month-olds were higher among ATIV recipients (25%) than UTIV recipients (13%), but were similar for all three vaccine groups in the younger age group 6 to <24 month-olds (19%-24%).

Only one study provided safety data on ATIV use in children with an underlying autoimmune disease.^{Footnote 40} Local reactions occurred in about 40% of the subjects, and systemic reactions in about 30%, with no significant difference between healthy children (n=30) and patients with juvenile idiopathic arthritis (JIA) (n=60). One serious but transient adverse event (fever and coxalgia) was reported in a 6-year-old JIA patient treated with Disease-Modifying Anti-Rheumatic Drugs but resolved spontaneously. A follow up 3 months after vaccination determined that none of the JIA patients showed any significant clinical or laboratory changes in disease activity.

One study (V70P6) employed a dose-ranging factorial design and included both adjuvanted and unadjuvanted versions of seasonal TIV and QIV administered to children 6-36 months old.^{Footnote 33} Overall, there was no indication of an increasing risk of AEs associated with increasing MF59 dose, antigen dose, or the addition of a second B strain. However, reactogenicity of 15µg formulations were slightly higher for both adjuvanted and unadjuvanted vaccines compared to the corresponding 7.5µg formulations.

Evidence Gaps

Evidence on efficacy and effectiveness of ATIV in children aged 6 to <72 months is limited to a single clinical trial (V70P5) with concerns around the reliability of the results.

There is limited but consistent evidence that ATIV is more immunogenic than comparable UTIVs against influenza A types. However, the CHMP standard vaccine licensing criteria were used to benchmark vaccine immunogenicity in all these trials, and it is unclear whether these criteria are generalizable to young children, and the HI titre cutoff predictive of clinical risk reduction among children remains unknown. It is also unclear whether these criteria are applicable to protection against influenza B strains.

Although there is reasonable evidence regarding safety of the vaccine, no information is available on the long-term effects of repeated ATIV administration, and evidence on the safety (and effectiveness) among children with autoimmune diseases, cancer, organ transplant, or chronic medical conditions is limited.

No studies were found directly comparing trivalent ATIV to QIV or LAIV, which are the most relevant vaccine comparators for children <2 years of age and ≥ 2 years of age respectively. In most trials, children aged 6-36 months receive half a dose (0.25mL) as per regulatory guidelines. In Canada, the standard of care is a full dose (0.5mL) of unadjuvanted inactivated vaccine for all age groups including children 6 to <36 months of age, based on evidence of improved immunogenicity without increase in reactogenicity.^{Footnote 5} One study using a dose-ranging factorial design with adjuvanted and unadjuvanted versions of seasonal TIV and QIV was identified, but a comparison of ATIV and full dose QIV was not the primary objective of the study and the data were not grouped appropriately for such a comparison. Further trials that directly compare trivalent 0.5mL ATIV to 0.5mL dose of trivalent or quadrivalent influenza vaccine in children 6 to <36 months as is recommended in Canada would be beneficial.

All the identified primary clinical trials were sponsored and conducted by the manufacturer with the aim of obtaining licensure for pediatric use. No data provided by independent researchers were available for review.

Discussion/Summary

Information on the efficacy of ATIV versus UTIV is only available from a single trial. The bulk of the data was obtained during one mild season that was dominated by A/H3N2. The comparator vaccine induced a poorer immune response compared to equivalent UTIVs. There were also concerns raised by an EMA inspection about the quality of laboratory testing and validity of ascertainment of influenza cases especially at the main German site.

There is limited but consistent evidence that ATIV is more immunogenic than comparable UTIVs against influenza A types. In particular, a single dose of ATIV is more immunogenic than a single dose of UTIV. However, two doses of ATIV are still necessary to achieve a satisfactory immune response against influenza B. ATIV was not compared directly to LAIV in children ≥2 years of age, and it is unclear whether ATIV would generate a smaller, greater or equivalent immunogenic response than LAIV. Clinical trials comparing the immunogenicity and efficacy of a single- and two-dose regimes of ATIV and LAIV are needed.

It is unclear whether the stronger humoral immune response induced by ATIV translates into an appreciable advantage over UTIVs in terms of preventing influenza or its complications. Evidence is growing that the conventional CHMP adult immunogenicity criteria may not be applicable to young children, although it is far from clear what alternative criteria should be used for children.

Safety data in children are consistent with what is known about ATIV safety profile in adults. ATIV is more reactogenic than UTIV, with ATIV recipients experiencing 10-15% more solicited local and systemic reactions compared to UTIV recipients. However, most reactions were mild and resolved quickly. Severe reactions were rare, but all reviewed studies were too small to detect clinically significant but rare adverse events. Like all influenza vaccines, there are few data on safety and immunogenicity of ATIV in children with immunodeficiency and other chronic illnesses.

Taken together, ATIV is likely both more immunogenic and more reactogenic than UTIV among children 6-72 months of age. There are insufficient data to assess whether ATIV is more effective than UTIV or LAIV in practice or to make an informed risk-benefit analysis. More research is needed to inform these issues.

List of Abbreviations

AEs

Adverse events

AQIV

Adjuvanted quadrivalent influenza vaccine

ATIV

Adjuvanted trivalent influenza vaccine

CHMP

EU Committee for Medicinal Products for Human Use

CI

Confidence Interval

DMARDS

Disease-modifying anti-rheumatic drugs

GMT

Geometric mean titre

HI

Hemagglutination inhibition

IM

Intramuscular

JIA

Juvenile Idiopathic Arthritis

QIV

Quadrivalent influenza vaccine

SAEs

Serious adverse events

SD

Standard Deviation

UTIV

Unadjuvanted trivalent influenza vaccine