key: cord-288938-4bheqtk5
authors: Hönemann, M.; Martin, D.; Pietsch, C.; Maier, M.; Bergs, S.; Bieck, E.; Liebert, U.G.
title: Influenza B virus infections in Western Saxony, Germany in three consecutive seasons between 2015 and 2018: Analysis of molecular and clinical features
date: 2019-10-08
journal: Vaccine
DOI: 10.1016/j.vaccine.2019.08.027
sha: 
doc_id: 288938
cord_uid: 4bheqtk5

BACKGROUND: The impact of annual influenza epidemics and prevailing strains varies worldwide and regional. The majority of vaccines used contained two influenza A strains and only one influenza B strain (trivalent vaccine). AIM: The aim of the study was to compare laboratory confirmed influenza B cases during three consecutive years with respect to vaccination history, clinical symptoms and molecular virology. METHODS: Partial HA gene sequences were analyzed for lineage determination and complete HA sequence in cases with reported vaccination and in fatal cases. Clinical data were retrieved from patient charts. FINDINGS: During the 2015/16 season, 75 influenza B cases were retrieved; 11 in 2016/17, and 274 in 2017/18. The frequency of Yamagata-lineage strains increased from 7.6% to 100%. No difference was detected in the relative frequency of co-morbidities in season 2017/18. 37.7% of the adult patients and 4.5% of pediatric patients were vaccinated against influenza. INTERPRETATION: Phylogenetically, Yamagata strains clustered similarly in 2017/2018 when compared to the previous two influenza seasons. While the relative frequency of influenza B cases differed, the clinical symptoms remained similar. CONCLUSION: World Health Organization recommendations for the use of tetravalent vaccines that contain two influenza B strains (Yamagata and Victoria) in addition to the two influenza A strains (H1N1 and H3N2) should be implemented in national vaccination guidelines. FUNDING: This research was partially supported by the Association of Sponsors and Friends of Leipzig University.

Influenza B virus of the Orthomyxoviridae family [1] has a segmented single stranded, negative sense RNA genome. First isolated in 1940, influenza B virus diverged into two lineages, Victoria and Yamagata, in the late 1970s with similar clinical [2] but different phylodynamic properties [3] . Although infections in pigs and seals were observed, there is no known animal reservoir for influenza B [4] . Thus, while influenza A infections can be zoonotic, influenza B virus only circulates in the human population. The narrow host range and slower evolution [5] are thought to be the main contributing factors to the less frequent occurrence of influenza B epidemics [6] .

In 2017/18 influenza B predominance was reported in countries throughout the northern hemisphere. It was the leading influenza type in Europe and Canada [7] [8] [9] and the second most common type after influenza A/H3N2 in the USA [10] . In Germany, after the seasons of 2001/2, 2005/6 and 2015/16, it was only the fourth season of the 21st century with a dominance of influenza B [11] . As the prevailing lineage of influenza B changes frequently, the World Health Organization (WHO) recommended to include both lineages in a tetravalent vaccine in 2012 [12] .

The key component of the vaccine is the viral hemagglutinin (HA) that covers the viral surface in a trimeric form. The major antigenic regions are the 120-loop, 150-loop, 160-loop, and 190helix [13] and are located on the HA1 subunit which forms the globular head domain. It is responsible for recognition and binding of sialic acids on the surface of the target cells. The HA2 subunit is the main component of the HA stalk. Although some antibody 

Vaccine j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / v a c c i n e cross reactivity is observed [14, 15] , vaccine efficiency between the two lineages might be reduced in seasons in which the formulation of the trivalent influenza vaccine does not match the circulating strain due to antigenic differences [16, 17] . Additionally, there is growing evidence that vaccine effectiveness is already waning within a season thus reducing immunological protection even if the vaccine contained the matching strain in the previous season [18] .

The aim of this study was to analyze clinical and molecular features of laboratory-confirmed influenza B cases during three consecutive seasons between 2015 and 2018. A comparison between patient characteristics was done in relation to the combined seasons of 2015/16 and 2016/17 when appropriate. Additionally, molecular epidemiology of the viral hemagglutinin (HA) gene was performed on selected isolates and amino acid changes were mapped to the major antigenic domains of the HA protein.

Specimens and clinical data. 11,956 respiratory samples, including nasal aspirates, supernatants from nasal and pharyngeal swabs, throat rinsing fluid, tracheal secretions, and bronchoalveolar lavage fluids, of 7681 patients were tested for respiratory virus infections. Testing was initiated at the discretion of the Lineage determination. NA sequences of a 123 bp-long amplicon of the HA gene [19] were assessed (BigDye Terminator Sequencing Kit v1.1 and ABI 3500 Genetic Analyzer, Applied Biosystems, Foster City, USA). The sequenced region corresponds to nucleotide positions 966-1088 and 963-1085 of the complete HA gene of the Victoria lineage (B/Brisbane/60/2008) and the Yamagata lineage (B/Phuket/3073/2013), respectively. Sequences with nucleotides G16, G17, G28, G55, C64, G73, and A112 belonged to the Victoria-lineage. Viruses with nucleotides A16, A17, A28, A55, T64, T73, and G112 belonged to the Yamagata-lineage.

Phylogenetic analysis. NA sequencing of the complete HA was performed according to WHO guidance [20] . In cases of low influenza B concentration an alternative protocol generating three instead of two overlapping fragments was used [21] . Phylogenetic trees of the 1710 nucleotides large HA coding region were constructed at nucleotide level with the MEGA software version 6 using the Maximum-Likelihood method. Bootstrap analysis was performed with 1000 replicates [22] . Complete HA sequences were submitted to Genbank (accession numbers MK459565 -MK459656).

Statistical analysis. Statistical analysis was performed using IBM SPSS Statistics for Windows, Version 24.0 (Armonk, NY: IBM Corp.). Continuous values were expressed as mean or median (range) and categorical data as frequencies (percentages). Student's t-test or Mann-Whitney U test was performed to compare means. Chi-square or Fisher's exact test were performed for categorical variables. All tests were two-tailed. A p-level of < 0.05 was considered significant.

Relative and absolute numbers of influenza types detected between 2015 and 2018 are shown in Table 1 . For season 2015/16 the maximum number of cases occurred between weeks 10 and 14 of 2016. The peak occurrence in patients below 18 years was in week 12 while it was week 14 for the adult patients (Fig. 1A) . The eleven cases of the season of 2016/17 occurred Table 2 Study population and clinical features of influenza B infected cases. (Table 2) . For the pediatric patients there were no significant differences between season 2017/18 and 2015-2017 except for the combined rate of comorbidities and risk factors for a severe course of the influenza. However, this could not be attributed to a single analyzed factor. When looking at the adult population, there were no significant differences between season 2017/18 and the previous seasons except for a higher proportion of female patients and a lower proportion of viral coinfections in 2017/18. Most differences were noted between the age groups. 32.3% of the adult patients were vaccinated, mostly with the trivalent vaccine, whereas 95.5% of the pediatric patients were unvaccinated. The rate of comorbidities was significantly higher in the adult population. Furthermore, the clinical course of the disease was more severe in the adult population as higher rates of ICU admissions, invasive ventilation, and fatal outcomes were observed. HA phylogeny. Sequence analysis of the whole hemagglutinin gene was performed for 16 Victoria and 76 Yamagata strains and consistently confirmed the results of the initial lineage determination. Included were all strains of patients with a known trivalent (n = 33) or tetravalent (n = 7) vaccination, randomly selected patients without known vaccination (n = 51), and of all patients that had a fatal outcome within 30 days of the influenza B detection (n = 27). Except for one strain of 2017, all Victoria strains originated from the 2015/2016 season and belonged to the 1A clade. All of the strains were phylogenetically closely related and did not cluster based on vaccination history or fatal outcome (Fig. 3A) . All analyzed Yamagata strains belonged to clade 3. All of the strains were phylogenetically closely related, but strain phylogeny indicates a discrete drift from season 2015/16 to season 2017/18. Again, no clustering with regard to vaccination history or fatal outcome was observed (Fig. 3B) .

Amino acid changes. The detected amino acid changes (AAchanges) within the HA protein are depicted in Table 3 

The influenza season of 2017/18 was dominated by the Yamagata-lineage of influenza B in Europe. In Germany, it was the largest season in terms of case numbers since the establishment of the nationwide surveillance system in 2001 [11] . Large scale epidemics or pandemics are usually caused by influenza A. This is attributed mostly to the broader host range and faster rate of evolution of influenza A [6] . Influenza B infections usually occur with a temporal delay towards the end of an influenza Adominated season [23] . The exceptional impact of influenza B infections in 2017/18 with regard to public health was therefore surprising. However, seasons with a high proportion of influenza B infections occur infrequently. Specifically, in Germany, the mild seasons of 2001/2 and 2005/6 and the severe season of 2015/16 have been dominated by influenza B with a share of up to 78% of all influenza cases. The high severity of the season 2017/18 with an overall estimate of 9 million excess consultations in Germany [11] may be attributed to several factors.

According to the ECDC, the influenza season of 2017/18 started earlier than most of previous ones [9] . While this is true for the epidemic phase of the two influenza B heavy seasons in this study, the time frame for isolated cases was comparable. In the season of 2015/16 the epidemic phase started in week 5 of 2016 and lasted until week 15. In 2017/18 the epidemic threshold was crossed in week 52 of 2017 until week 14 of 2018 [11] .

Another factor contributing to the severity of the season might be the patient population itself. Both the attack rate and the relative illness ratio of influenza B are highest for children and decrease with age [24, 25] . However, the major difference that was seen for the season 2017/18 was the higher proportion of adult patients. The age-subgroups themselves, being pediatric patients and adult patients, were not different in their clinical characteristics between 2017/18 and the two previous seasons. However, the adult population that was seen had a high proportion of comorbidities and risk factors that might have contributed to a Fig. 3 (continued) severe outcome of the influenza disease, as it is known for example for cardiac insufficiency [26] , additionally to age itself [2, 27, 28] . A shift towards the adult population thus increases the impact on the public health system and may explain the severity of this season with a high number of excess consultations. An interpretation of the presented clinical data, however, needs to be done with caution as there may be a bias towards severe disease courses in a hospital setting. A fatality rate of 11.2% of the adult population of this study can thus not be transferred to the general population.

A third potential factor may be the circulating influenza type. The Yamagata-lineage of influenza B was predominantly found in the season of 2017/18. In contrast to the Victoria lineage, which has a faster evolution with nearly each season representing a phylogenetic bottleneck, the biology of the Yamagata lineage is different. In long term studies of the HA of the Yamagata lineage a long co-circulation of Yamagata variants was observed [3, 13, 29] . In accordance with these observations the analyzed strains of this study were phylogenetically closely related. Furthermore, the detected amino acid changes were scattered throughout the HA and to the best of our knowledge no substitution is known for increased virus pathogenicity.

Since 2000, changes in the antigenic sites are dominantly located in the 120-loop and 190-helix regions, indicating main targets for immune pressure. With regard to the vaccine strain B/Phuket/3073/2013 the main antigenic regions were only affected in 9 strains and occurred mainly in the 120-loop region, followed by the 160-loop region for the analyzed Yamagata strains. In contrast, all analyzed Victoria strains had at least one amino acid change in the 120-loop region with regard to the vaccine strain B/Brisbane/60/2008. It is therefore surprising that the Yamagatalineage that circulated in the population for three consecutive seasons was able to cause such a major epidemic. Table 3 Amino acid changes of the Victoria-and Yamagata-strains with regard to the respective vaccine strain.

Amino acids are numbered with regard to the HA0 protein, and HA1 as well as HA2 subunits are indicated. The vaccine strains to which the isolates were compared are given in red for each lineage. Isolates from this study are labeled as vaccinated with trivalent vaccine (triangle), vaccinated with tetravalent vaccine (square), randomly selected case without vaccination history (circle), fatal cases (black filled symbols). Symbols represent the main antigenic sites, 120-loop (A; amino acids 116-137, 38, 56, 75, 177, 179-181), 150-loop (B; amino acids 141-150), 160-loop (C; amino acids 162-167) and 190helix (D; amino acids 194-202). *For the Yamagata-lineage, 48 strains, that were identical at amino acid level, are depicted as a pool and contain patients with the following characteristics: 15xs; 15xd; 3xh, 1xj, 14x4.

A reason for this might be the interplay between the circulating influenza strains causing infections and the choice of the vaccine components or type of vaccine. In the season of 2015/16 Influenza A/H1N1 pdm09 and the Victoria-lineage of influenza B caused the majority of influenza infections. In 2016/17 influenza A/H3N2 was the main cause of flu disease. When looking at the influenza B component of the broadly used trivalent influenza vaccine there was a switch from B/Phuket/3073/2013-like (Yamagata-lineage) to B/ Brisbane/60/2008-like (Victoria-lineage) in the season 2016/17, effectively causing a mismatch in each of the three studied seasons. In 2017/18 the only strain that was not immunologically covered by recent epidemics or the current vaccine composition was the influenza B Yamagata-lineage. An increased proportion of non-immune individuals in the adult population thus may have contributed to the observed shift in the age distribution of the infected patients.

Cross-reactive antibodies can be isolated from vaccinated individuals [30, 31] but the contribution to the immunity against influenza B is not well defined. Some degree of lineage crossprotection has been reported upon vaccination with trivalent influenza virus vaccines. However, it seems to be limited to certain age groups and the vaccine efficacy is consistently superior for influenza B strains belonging to the vaccine lineage [32, 33, 16] . The capability of an individual to generate a broad antibody response against epitopes that include regions conserved between the two lineages, e.g. in the HA stalk, seems to be of major importance. Thus, the factors influencing lineage cross-protection are complex and not easy to predict as also the immune memory may have a substantial impact on the targeted epitopes [34, 35] .

In 76% of the patients with a vaccination history in this study the trivalent vaccine, with an expected moderate level of vaccine efficiency [7] , was used. The majority of the vaccinated patients were adults. However, a major improvement to influenza disease burden may only be accomplished by extending vaccine recommendations to younger age groups that are immunologically more naïve and thus more susceptible to infections [36, 25] . WHO recommendations for the use of tetravalent vaccines that contain two influenza B strains (Yamagata and Victoria) in addition to the two influenza A strains (H1N1 and H3N2) should be implemented in national vaccination guidelines as it was done in Germany as a result of the severity of the 2017/18 influenza season. 

The biology of influenza viruses

Clinical characteristics and severity of influenza infections by virus type, subtype and lineage: a systematic literature review

The contrasting phylodynamics of human influenza B viruses

The continual threat of influenza virus infections at the human-animal interface: what is new from a one health perspective

Comparison of the mutation rates of human influenza A and B viruses

The evolution of seasonal influenza viruses

Moderate influenza vaccine effectiveness in a B mismatch season: Preliminary results from the

Early season co-circulation of influenza A(H3N2) and B(Yamagata): interim estimates of 2017/18 vaccine effectiveness

18 influneza season. Last accessed 15.12

Update: Influenza Activity in the United States During the 2017-18 Season and Composition of the 2018-19 Influenza Vaccine

Seasonal Influenza Reports

The quadrivalent approach to influenza vaccination

Evolutionary dynamic of antigenic residues on influenza B hemagglutinin

Molecular-level analysis of the serum antibody repertoire in young adults before and after seasonal influenza vaccination

Three types of broadly reacting antibodies against Influenza B viruses induced by vaccination with seasonal influenza viruses

Efficacy of live attenuated influenza vaccine in children against influenza B viruses by lineage and antigenic similarity

Prediction of influenza B vaccine effectiveness from sequence data

Intra-season Waning of Influenza vaccine effectiveness

Simultaneous detection of influenza viruses A and B using real-time quantitative PCR

WHO information for the molecular detection of influenza viruses

Effect of a single mutation in neuraminidase on the properties of Influenza B virus isolates

MEGA6: molecular evolutionary genetics analysis version 6.0

Virological surveillance of influenza and other respiratory viruses during six consecutive seasons from 2006 to 2012 in Catalonia, Spain

Distribution of influenza virus types by age using case-based global surveillance data from twenty-nine countries

Estimation of influenza-attributable medically attended acute respiratory illness by influenza type/subtype and age

Influenza vaccine in heart failure: cumulative number of vaccinations, frequency, timing, and survival: A Danish Nationwide Cohort Study

Estimating influenza disease burden from population-based surveillance data in the United States

Age-and sexrelated risk factors for influenza-associated mortality in the United States between 1997-2007

Circulating pattern and genomic characteristics of influenza B viruses in Taiwan from 2003 to 2014

Human monoclonal antibodies broadly neutralizing against influenza B virus

Cross-lineage protection by human antibodies binding the influenza B hemagglutinin

Effectiveness of seasonal influenza vaccine for adults and children in preventing laboratory-confirmed influenza in primary care in the United Kingdom: 2015/16 end-of-season results

Influenza A/subtype and B/lineage effectiveness estimates for the 2011-2012 trivalent vaccine: cross-season and cross-lineage protection with unchanged vaccine

From original antigenic sin to the universal influenza virus vaccine

Immune history profoundly affects broadly protective B cell responses to influenza

Age, influenza pandemics and disease dynamics

This research was partially supported by the Association of Sponsors and Friends of Leipzig University (Vereinigung von Förderern und Freunden der Universität Leipzig).

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supplementary data to this article can be found online at https://doi.org/10.1016/j.vaccine.2019.08.027.