Clinical characteristics of patients with laboratory-confirmed influenza A(H1N1)pdm09 during the 2013/2014 and 2015/2016 clade 6B/6B.1/6B.2-predominant outbreaks

A novel pandemic influenza A(H1N1)pdm09 virus emerged in 2009 globally, and it continues to circulate in humans. The National Influenza Surveillance Network in Taiwan identified five A(H1N1)pdm09-predominant seasons, representing the 2009/2010, 2010/2011, 2012/2013, 2013/2014, and 2015/2016 outbreaks from 2009 to 2016. Independently, a retrospective cohort study (which enrolled 639 infected patients during the five seasons) was conducted at Chang Gung Memorial Hospital to explore the risk factors associated with influenza A(H1N1)pdm09-related complications. A phylogenetic analysis of hemagglutinin (HA) sequences showed that the circulating A(H1N1)pdm09 virus belonged to clades 1, 2, and 8 in 2009/2010; clades 3, 4, 5, and 7 in 2010/2011; clades 7 and 6C in 2012/2013; clades 6B in 2013/2014; and 6B/6B.1/6B.2 in 2015/2016. Compared to individuals infected in non-6B/6B.1/6B.2 seasons (2009/2010, 2010/2011, and 2012/2013), those infected in 6B/6B.1/6B.2 seasons (2013/2014 and 2015/2016) were at higher risk for influenza-related complications (adjusted odds ratio [aOR]: 1.6, 95% confidence interval [CI]: 1.0–2.8), pneumonia (aOR: 1.78, 95% CI: 1.04–3.04), mechanical ventilation (aOR: 2.6, 95% CI: 1.2–5.6), and acute respiratory distress syndrome (aOR: 5.5, 95% CI: 1.9–15.9). For the increased severity of infection during the influenza A(H1N1)pdm09 clade 6B/6B.1/6B.2 seasons, aspects related to the antigenic change of A(H1N1)pdm09 virus, immune response of the host, and environmental factors required further investigation.

SCIENTIfIC RepoRTs | (2018) 8:15636 | DOI: 10.1038/s41598-018-34077-4 number of severe cases and outcomes in the groups at risk and healthy young adults; these events were associated with A(H1N1)pdm09 clade 6B.1 infection 5,6 . Since 1999, the Taiwan Centers for Disease Control (CDC) has established a nationwide surveillance system requiring contract virologic laboratories to perform continuous virologic surveillance for respiratory viruses, particularly influenza and enteroviruses; this system was established after an epidemic of enterovirus 71 in 1998 7 . The long-term National Influenza Surveillance Network described the epidemiologic pattern of circulating viruses, and it has successfully identified the outbreaks of severe acute respiratory syndrome (SARS)-associated coronavirus and adenovirus 8,9 . Moreover, it also identified the novel H7N9 and H6N1 influenza viruses 10   were significantly different between each season (Table 1). A phylogenetic analysis of hemagglutinin (HA) sequences recovered in these epidemics, along with geographically diverse global influenza A(H1N1) pdm09 viral sequences, has revealed that the sequences are members of clades 1, seasons versus non-clades 6B/6B.1/6B.2 seasons, the patients were classified into two groups. The median (interquartile range, IQR) age of patients in the 6B/6B.1/6B.2 season was older than that in the non 6B/6B.1/6B.2 seasons ( Table 2). The number of infected individuals aged 50-64 years was higher in 6B/6B.1/6B.2 seasons than that in the non-clade 6B/6B.1/6B.2 seasons ( Table 2). The rate of underlying conditions; complications, including pneumonia and acute respiratory distress syndrome (ARDS); ICU admission; respiratory failure with mechanical ventilation; 30-day mortality; and in-hospital mortality in 6B/6B.1/6B.2 seasons were significantly higher than that in non-clade 6B/6B.1/6B.2 seasons (   Table 3). The rate of underlying conditions; complications, such as ARDS; ICU admission; and respiratory failure with mechanical ventilation in 6B/6B.1/6B.2 seasons was significantly higher than that in non-clade 6B/6B.1/6B.2 seasons (Table 3).

Risk factors of complications in patients with influenza A(H1N1)pdm09 virus infection.
The results of the logistic regression analysis on the risk factors associated with influenza A(H1N1)pdm09-related complications and pneumonia are shown in Table 4, and respiratory failure with mechanical ventilation and ARDS are also presented in Table 5. In the univariate analysis, 6B/6B.1/6B.2 season, age (50-64 years), onset to presentation, underlying conditions, obesity, smoking, alcoholism, and antiviral therapy were significant risk factors of complications, pneumonia, mechanical ventilation, and ARDS (Tables 4 and 5). In the multivariate logistic regression analysis, 6B/6B.1/6B.2 season, age (50-64 years and ≥65 years), underlying conditions, and antiviral therapy were significant independent risk factors of complications, pneumonia, and mechanical ventilation (Tables 4 and 5). Only 6B/6B.1/6B.2 season and obesity were considered as significant independent risk factors of ARDS ( Table 5). The effect of 6B/6B.1/6B.2 season on the total number of influenza-related complications was not significant in children aged ≤5 years. However, it was significantly stronger among individuals aged ≥6 years (Table S1).

Risk factors of mortality in patients with influenza A(H1N1)pdm09.
Among the hospitalized patients with laboratory-confirmed influenza A(H1N1)pdm09 infection, male patients and those with underlying conditions were significantly at risk for 30-day mortality (overall death within the first 30 days after hospital admission) and all-cause in-hospital mortality (overall death during hospital admission) as assessed using the multivariable Cox proportional hazard model ( Table 6). The same analysis showed that season was not associated with an increased risk for 30-day and all-cause in-hospital mortality (Table 6).

Discussion
The      During the 2013/2014 season, an unusually high hospitalization rate in adults aged 50-64 years was observed in the United States and Mexico 3, 16 . The increased morbidity in middle-aged adults during the 2013/2014 season had been attributed to the low vaccination rate in this age group 16 . However, that hypothesis cannot explain the unusual number of severe cases because the vaccination rate had already been low during the previous years 16,17 .
An interesting study has shown that up to 42% of middle-aged adults born between 1965 and 1979, who had been exposed to seasonal H1N1 viruses circulating in 1977, had reduced serologic reactivity with the 2013/2014 A(H1N1)pdm09; notably, the 2013/2014 virus harbors the distinctive K163Q HA antigenic mutation 18 . In the cohort of individuals born between 1965 and 1979, HA-specific antibodies with activity against A(H1N1)pdm09 must have been produced and shaped by exposure to prior-season H1N1 viruses (the so called "original antigenic sin") 19 . Nonetheless, the HA-specific antibodies in this cohort failed to recognize the 2013/2014 A(H1N1)    25 . Compared to the A/California/7/2009 vaccine virus, viruses of clade 6B harbor D97N, K163Q, S185T, K283E, and A256T substitutions in HA1. Viruses of subclade 6B.1 harbor further amino acid substitutions S84N, S162N, and I216T and 6B.2 and carry amino acid substitutions V152T and V173I 20 . More extensive studies must be conducted to identify the potential antigenic differences between clade 6B and subclade 6B.1/6B.2; such studies are expected to improve our understanding of how A(H1N1)pdm09 evolved (and continues to evolve) and how it affects and interacts with the human immune system. In the 2017/2018 season, the WHO has selected a new vaccine virus, which is the A/Michigan/45/2015 (H1N1)pdm09-like virus (a member of the 6B.1 subclade), as the influenza vaccine virus component for the northern hemisphere.
The National Influenza Surveillance Network coordinated by the Taiwan CDC was established more than 10 years ago. Policies favoring government funding for vaccines and antiviral agents have been consistent during the subsequent intervals. Between 2009 and 2015, government-funded vaccines have been administered primarily to those aged 6 months to 12 years, elderly individuals aged ≥65 years, healthcare workers, and individuals with underlying diseases. Individuals aged 13-64 years were not included in the government-funded vaccination program. Elementary school children aged 7-12 years had the highest influenza vaccination rate, with coverage reaching 60-70% annually 11 .
The present study is limited by its observational nature and the incorporation of a retrospective investigation. A potential bias may exist due to the exclusion of all cases with A(H1N1)pdm09 infection for 7 years. Nonetheless, no change was observed in terms of admission or management procedures during these outbreaks. The surveillance and reporting system in Taiwan has long been established. Taken together, the increased frequency of complications in 2013/2014 and 2015/2016 is unlikely due to detection bias. In addition, the major drawback of this study was the lack of documentation about the history of influenza vaccination in the records used to generate this study. However, a study by Taiwan CDC has reported that 95% of patients with complications in the 2015/2016 season had not received the influenza vaccine 26 . The vaccine coverage rate in non-elderly adults and elderly individuals would have been low during each of the outbreaks, particularly during the first wave, given that no A(H1N1)pdm09 vaccine was available in 2009/2010. Thus, the increased severity of influenza during the 2013/2014 and 2015/2016 seasons is unlikely to reflect a decreased rate of vaccination. The study has shown that Taiwan experienced the greatest burden of influenza-related complications due to A(H1N1)pdm09 clades 6B/6B.1/6B.2 in the sixth year of its circulation. The reasons for the increased impact of influenza-related complications remain uncertain. Aspects related to the antigenic change of A(H1N1)pdm09 virus, immune response of  the host, and environmental factors required further investigation. This report shows the importance of influenza disease surveillance and requires that the influenza A(H1N1)pdm09 virus should always be considered.

Methods
National influenza surveillance network. The network consists of eight regional commissioned laboratories located in the northern (n = 3), central (n = 2), southern (n = 2), and eastern (n = 1) parts of Taiwan. These laboratories have steadily collected more than 8000 respiratory specimens for surveillance per year, including more than 1000 influenza virus specimens annually, all of which are sent to the Taiwan CDC for the monitoring of influenza viral activity. The Taiwan Influenza Express, a weekly online influenza surveillance report, has been published by the Taiwan CDC from July to May of each year since 2005 (http://www.cdc.gov.tw/english/ submenu.aspx?treeid = 00ED75D6C887BB27&nowtreeid = 8F1E239D0FD8877A) 10,27 . This report includes the total number of respiratory specimens; isolate number of influenza A(H1N1), influenza A(H3N2), and influenza B; and case number of laboratory-confirmed influenza cases in intensive care units (ICUs), a class of events that is considered a category 4 nationally notifiable disease. However, data on weeks 19-37 are not available annually. A confirmed case involved a patient who had acute influenza-like illness (temperature ≥ 38 °C with either cough or sore throat) and nasopharyngeal/throat or bronchoalveolar lavage samples harboring influenza A(H1N1)pdm09 virus as detected using real-time (RT) reverse-transcription polymerase chain reaction (PCR) assay or via viral culture 11,28 . For the purposes of the present study, each season was defined as extending from July of the same year to May of the following year. The annual population figures provided by the Department of Household Registration Affairs of the Interior Ministry were used for the calculation of the incidence of laboratory-confirmed influenza A(H1N1)pdm09 cases in the ICU. , a 4000-bed, university-affiliated teaching hospital that is located in northern Taiwan and provides both primary and tertiary care. In addition, CGMH is one of the 8 regional commissioned laboratories of the Taiwan CDC. Patients who had acute influenza-like illness (temperature ≥ 38 °C with either cough or sore throat) and had influenza A(H1N1)pdm09 virus as detected using RT-PCR assay or via viral culture using respiratory specimens were included in the study. Patients whose data are not available were excluded. The institutional review board of CGMHT approved the study, and it was carried out in accordance with the relevant guidelines and regulations. Informed consent was waived due to the study's retrospective nature. All medical records of the enrolled patients were reviewed. Demographic characteristics, underlying medical conditions, clinical course, antiviral treatment (oseltamivir or zanamivir), mechanical ventilation, admission to an ICU, and death were recorded using a structured questionnaire. Body mass index (BMI), a measure of obesity, was calculated for patients whose height and weight data were available. Obesity was defined as follows: 1) body weight ≥ 95th percentile in children < 2 years of age; 2) BMI ≥ 25 kg/m 2 in patients aged between 2 and 18 years; and 3) BMI > 28 kg/m 2 (Chinese criteria) in patients > 18 years 29 . Medical conditions associated with a high risk for influenza complications were defined based on those listed by the United States Advisory Committee on Immunization Practices 30 . Patients with confirmed pneumonia on radiography, acute respiratory distress syndrome (ARDS), acute onset of cardiovascular, neurologic condition, respiratory failure with mechanical ventilation; those who were admitted in the ICU; and those who died were considered to have influenza-related complications. Pneumonia on radiography was diagnosed based on the presence of a consolidation, infiltrate, or opacity 31 . ARDS was defined according to the standard criteria 32 . The primary study outcome was the occurrence of (any) influenza-related complications. The secondary study outcomes were pneumonia, mechanical ventilation, ARDS, 30-day mortality, and in-hospital mortality.
Genetic characterization of the virus. A total of 82 isolated influenza A(H1N1)pdm09 virus were randomly selected for the analysis of viral hemagglutinin (HA) and neuraminidase (NA) genes across the five seasons. The RNA was extracted using the QIAamp Viral RNA mini kit (Qigen, Germany) according to the manufacturer's instructions. RT-PCR and primer pairs used for sequencing HA and NA genes were performed, as previously described 33 . Sanger sequencing of the viral HA and NA genes was performed to establish clade designation and to detect differences in amino acid 33 . The obtained amplicons were assembled into a full-length 1,701-bp span for HA and 1410-bp for NA using DNASTAR Lasergene (DNAStar, Madison, WI). Newly reported sequences in this study were deposited at the GenBank database under the accession numbers shown in Fig. S1 for HA and NA genes. The evolution history was inferred by the maximum likelihood method based on the Hasegawa-Kishino-Yano model 34 . The percentages of replicate trees (1,000 replicates) are shown next to the branches in which the associated taxa clustered together in the bootstrap test. Phylogenetic analysis in this study was conducted using MEGA7 35 .
Statistical analysis. Continuous variables were presented as medians and interquartile ranges (IQRs); categorical variables were presented as numbers and percentages. All analyses were performed using the Statistical Package for the Social Sciences software package version 22.0 (SPSS Inc., Chicago, IL, the USA). The incidence rate ratio (IRR) was generated using Poisson regression with 95% confidence intervals to compare the rates of laboratory-confirmed influenza A(H1N1)pdm09 cases in the ICU per 100,000 populations across different seasons; 95% confidence intervals for which the upper and lower bounds did not include 1 were considered as statistically significant. Differences in categorical variables were compared using the chi-square test or a Fisher's exact test. Continuous variables were compared using the Kruskal-Wallis one-way analysis of variance test. Multivariate logistic regression analysis and multivariate Cox proportional hazards model were used for outcome analysis. The variables included sex, season, age group, onset to presentation, underlying condition, obesity, smoking, alcoholism, and antiviral therapy. Variables with a P value < 0.1 in the univariate analysis were included in the multivariate model. The Hosmer-Lemeshow goodness-of-fit test was performed to assess the overall fit of the model. All statistical operations were two-tailed. P values ≤ 0.05 were considered statistically significant.