Epidemiological and clinical characteristics of community‐acquired and nosocomial influenza cases and risk factors associated with complications: A four season analysis of all adult patients admitted in a tertiary hospital

Abstract Background Information on the characteristics of patients with nosocomial influenza and associated complications is scarce. This study compared epidemiological and clinical characteristics of patients admitted with hospital‐acquired influenza (HAI) to those with community‐acquired influenza (CAI) and analyzed risk factors associated with complications. Methods This retrospective, observational study included all adult patients with confirmed influenza virus infection admitted to Son Espases University Hospital during the influenza season in Spain (October to May) from 2012‐2013 to 2015‐2016. Symptom onset before admission was included as CAI, and 2 days after admission or within 48 hours after previous discharge were considered as HAI. Results Overall, 666 patients with laboratory‐confirmed influenza were included; 590 (88.6%) and 76 (11.4%) had CAI and HAI, respectively. Baseline characteristics and vaccination rates were similar in both groups. Patients with HAI had significantly fewer symptoms, less radiological alterations, and earlier microbiological diagnosis than those with CAI. Eighty‐five (14.4%) and 20 (27.6%) CAI and HAI patients, respectively, experienced at least one complication, including septic shock, admission to the intensive care unit, mechanical ventilation or evolution to death (any one, P = .003). Univariate and multivariate binary logistic regression was performed to assess independent risk factors associated with the occurrence of complications: nosocomial infection, diabetes, oseltamivir treatment, having received no vaccination, microbiological delay, dyspnea, and the state of confusion were the most important significant factors. Conclusions Our study shows the need to implement microbiological diagnostic measures in the first 48 hours to reduce HAI frequency and associated complications.


| INTRODUC TI ON
The influenza virus contributes significantly to the number of hospital admissions during the annual outbreak months. Among the existing viral pathogens, influenza remains the most clinically significant viral cause of community-acquired pneumonia in adults, and it is a key cause of hospital-acquired pneumonia. 1 However, the impact of each influenza epidemic on morbidity and mortality varies considerably depending on the prevalence of the influenza virus in each season. Since 2009, following the new influenza A (H1N1) pdm09 virus pandemic, the use of rapid polymerase chain reaction (PCR) tests on respiratory samples has been extended in hospitals. 2 Outbreaks of nosocomial infection usually occur during the seasons of community infection, predominantly in units with hematological and immunosuppressed patients. [3][4][5][6] Comorbidities such as chronic bronchitis, immunosuppression, 7 diabetes, chronic kidney disease, nosocomial infection, 8 and malnutrition 9 have been previously described as a risk factors associated with complications like septic shock, need for mechanical ventilation, admission in intensive care unit (ICU) and death. However, only few studies have been conducted on the sporadic nosocomial cases of influenza and their associated complications, with complete clinical and virologic data. 10,11 Identifying and understanding risk factors of hospital-acquired influenza (HAI) infection is important to implement control measures to reduce morbidity and mortality.
In Spain, the epidemiological surveillance of influenza is conducted through sentinel networks that are integrated into the influenza surveillance system (ScVGE). Surveillance on cases of confirmed influenza requiring hospitalization has been conducted since 2010; however, the data for all patients admitted for influenza have not been systematically collected. The main strategy for the prevention and control of influenza in hospitals is providing vaccination to people at risk and healthcare workers. This is, coupled with precautions against transmission through infected droplets, and early treatment with antivirals for people admitted for influenza. 12,13 However, adoption of these measures is highly variable between hospitals. A recently published study showed a reduction in nosocomial infection and length of hospital stay after the application of a real-time surveillance system. 14 This study therefore aimed to compare the epidemiological and clinical characteristics of patients who required hospital admission and diagnosed with community-acquired influenza (CAI) and HAI or nosocomial. Additionally, we aimed to determine risk factors associated with complications and analyze whether nosocomial infection is also a risk factor for the development of complications. The secondary aim was to compare the rate of nosocomial infection in different seasons.

| Data collection
The medical charts for positive cases at the microbiological department of HUSE were reviewed. Data were extracted from the medical charts using a case report form; the data collected included the date of admission and discharge, date of any previous discharges, signs and symptoms related to flu, date of symptoms onset and diagnosis, epidemiological variables, comorbidities, vaccination status for the current influenza season, analytical and radiological findings used to make a diagnosis, treatment with antiviral and antibiotics, and complications during the admission.

Conclusions:
Our study shows the need to implement microbiological diagnostic measures in the first 48 hours to reduce HAI frequency and associated complications.

K E Y W O R D S
community-acquired influenza, complications, epidemiology, hospital-acquired influenza

| Procedures and settings
Detection of influenza viruses was performed using a commercial real-time PCR (RT-PCR: Allplex Respiratory Full Panel Assay; Seegen) which enables simultaneous and differential detection of 16 respiratory viruses, including influenza A and B and subtypes H1N1 and H3N2. It was performed 5 days a week during the influenza season. HUSE is an 800-bed tertiary care center and is the reference hospital for the follow-up of serious influenza cases in the Balearic Islands. An infection prevention policy is followed in the hospital, and respiratory hygiene precautions are taken for people treated for respiratory symptoms in the emergency department. Additionally, respiratory sampling was performed (oropharyngeal smear or other respiratory samples) in people admitted for suspected influenza. Up until 2018, only confirmed cases were isolated. Based on the hospital protocol, patients with confirmed influenza or patients with influenza-like illness with comorbidities (immunosuppression, heart disease, lung disease or chronic kidney disease) were treated with oseltamivir.

| Description of complications
Complications associated with flu included septic shock, admission to the ICU, need for mechanical ventilation and death. Other complications, less commonly described as associated with influenza, are described in the results section of this paper.

| Statistical analysis
Numerical variables are expressed as medians and interquartile ranges (IQR) and categorical variables as frequencies and proportions. Differences between CAI and HAI were analyzed using the

| RE SULTS
Overall 666 patients with laboratory-confirmed influenza were included over the four seasons; 590 (88.6%) and 76 (11.4%) patients were diagnosed with CAI and HAI, respectively. The distribution of the subtypes by seasons and the rate of nosocomial cases are shown in Figures 1,2, and Table 1. More nosocomial cases were diagnosed in the last seasons of the study with an incidence of 12% (P = .036).
No differences in the rate of HAI were observed based on subtype when subtypes A and B were compared. Median time from hospital admission until HAI diagnosis was 10 days (IQR 5-14), and median time between respiratory symptom onset and microbiological diagnosis was 1 vs 4 days for those with HAI and CAI, respectively

| Demographic characteristics and comorbidities
Upon comparison of patients with HAI and CAI, no differences were observed in mean age, sex, or in the main comorbidities, including immunosuppression (Table 1). A small proportion of chronic

| Clinical characteristics and initial treatment
Patients with HAI presented with cough, myalgia, and a breathless sensation less frequently, but a higher proportion of patients experienced confusional symptoms (15.8% vs 8.8%, P = .052) (

| Risk factors associated with influenza complications
In univariate analysis, pregnancy, obesity, diabetes, interval between symptoms and microbiological diagnosis, oseltamivir treatment, not having been vaccinated, nosocomial infection, lower oxygen saturation, and symptoms of dyspnea and confusion were significantly associated with complications (

TA B L E 3 Complications and prediction scales in community-and hospitalacquired influenza patients
The use of rapid diagnostic testing that allows a shorter turnaround time to receive the results might aid in the earlier initiation of oseltamivir treatment; unfortunately, such tests were not used in our institution.
Additionally, this might allow for a reduction in the use of antibiotics and allow for more rapid initiation of preventive measures. 25 There were no significant differences in the frequency of influenza subtypes A and B among those with CAI and HAI. These findings differ from those previously reported, 4 implying that the risk of transmission is in hospitals and thus precautions need to be similarly implemented. Hemagglutinin, neuraminidase gene sequencing, or whole -genome sequencing of the influenza virus have been suggested for detecting nosocomial influenza outbreaks more quickly and accurately; however, we did not have the means to implement these set tests in this study. 27 This study has several limitations. There was a risk of loss of information regarding the prognosis of patients who were transferred to another hospital from the emergency department.
Additionally, we did not have full access to the medical charts for patients who were previously admitted to other hospitals.
Moreover the number of patients included was limited, and there might have been some differences in baseline pathologies or previous surgeries that were not evaluated in this study. These limitations have an impact on the validity and generalizability of our findings.

| CON CLUS IONS
Hospital-acquired influenza is common and it is associated with a greater frequency of complications and death. A significant proportion of CAI patients are admitted to the hospital and do not receive treatment or early isolation. The early diagnosis and treatment, and implementation of microbiological measures, will help to reduce the complications. Also, the implementation of preventive measures is the tools that will help to reduce the frequency of nosocomial transmission.

ACK N OWLED G EM ENTS
To Francisco de Borja Yuste Valero for his computer help in the development and explotation of the tables. To Meryl Jones for her help in editing and revising the manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

E TH I C A L A PPROVA L
As an observational study, no intervention was performed in patients and no written informed consent was given to the patients. Medical charts and information were anonymized before the analysis.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request. Melchor Riera Jaume https://orcid.org/0000-0002-0887-8710