Clinical characteristics and disease burden of respiratory syncytial virus infection among hospitalized adults

The disease burden of respiratory syncytial virus (RSV) infection in the adult population has not been well characterized compared to children. Investigation of the clinical characteristics and disease burden of adult RSV infection would help to establish public health policy and a future vaccine strategy. We retrospectively collected medical records of hospitalized adult patients who were diagnosed with RSV infection from January 2012 to December 2015 from three tertiary hospitals. Baseline characteristics, clinical outcomes and economic charge during hospitalization were compared by age groups (19–49 years, 50–64 years, and ≥ 65 years) using Chi-square test. The odds of risk factors of RSV pneumonia were calculated using binary logistic regression. A total of 204 patients from three hospitals were enrolled. Patients who older than 65 years were 132 (64.7%). 118 (57.8%) patients had clinically confirmed pneumonia and 22 (10.8%) died in a hospital. The median medical cost of RSV pneumonia was 2,855.26 USD (interquartile range, 1,561.85–5,379.55) per each admission. Solid cancer (adjusted OR, 3.85; 95% CI, 1.65–9.02, p = 0.002) and hematologic malignancy (all patients had pneumonia) were shown to be risk factors for RSV pneumonia. RSV infection in South Korea seemed to have a significant burden among adults as pneumonia, care in the intensive care unit and mortality. Nationwide awareness and further effort to recognize the current burden, prepare specific treatment, and prevent adult RSV infection would be necessary.

Statistical analysis. Each nonparametric variable in baseline characteristics and clinical outcomes was compared using the chi-square test with Fisher's exact test. Continuous variable such as length of hospital stay was compared using the analysis of variance. The odds of risk factors of RSV pneumonia were calculated using binary logistic regression analysis. Any continuous independent variables were assumed to be a linear relationship with a dependent variable in the logistic regression analysis. A P value of < 0.05 was considered to indicate statistical significance. All statistical analyses were performed using SPSS Statistics version 20 for Windows (IBM Corp., Armonk, NY, USA). ethical approval. 2019GR0127

Results
A total of 204 patients were enrolled from the three hospitals: 146 from Guro Hospital, 44 from Ansan Hospital, and 14 from Kangnam Sacred Heart Hospital. Most patients underwent diagnostic tests for RSV because of their newly developed respiratory symptoms and/or lung infiltrations on chest radiograph, except four patients. These four participants were in chronic bedridden states and were incidentally diagnosed with RSV infection during work-up for altered states of consciousness. The demographic characteristics of the patients are shown in Table 1. Sore throat and nasal congestion/rhinorrhea was more dominant in 19-49 years-old group (sore throat: 5 (25.0%); nasal congestion/rhinorrhea: 7 (35.0%)) than in 50-64 years-old group (sore throat: 4 (7.7%); nasal congestion/rhinorrhea: 14 (26.9%)) and ≥ 65 years-old group (sore throat: 6 (4.5%); nasal congestion/rhinorrhea: 21 (15.9%)). All proportions of underlying diseases were not significantly different between age groups. There was no human immunodeficiency virus infection in the study population.
The general admission rate and proportions of mechanical ventilation, pneumonia and mortality were not significantly different between age groups. There were 12 (60.0%) cases of RSV pneumonia were in the young adult group, 31 (59.6%) in the middle-aged adult group, and 75 (56.8%) in the elderly group. All patients with hematologic malignancies had RSV pneumonia. The rate of ICU admission was higher among the elderly (25.0%) than among young adults (10.0%) and middle-aged adults (11.5%), but also showed no significant difference (p = 0.060) ( Table 1).
189 (92.6%) were diagnosed by multiplex real-time PCR and 15 (7.4%) were diagnosed by viral culture. The results of RSV serotype in 17 patients were undefined, including two missed data in PCR group. The dominant serotype of RSV changed annually and prevailed from winter to early spring. RSV serotype A was dominant in 2012-2013 and 2013-2014 winter seasons (Fig. 1). More than half of patients enrolled in the study were in the elderly group (65 years old or more, 64.7%) and the mean age ± standard deviation was 67.77 ± 13.94 (Fig. 2).
The median medical cost of RSV pneumonia per each admission seemed to be increasing by the age groups: 1,957.33 USD in the young adult group, 2,116.43 USD in the middle-aged group, and 2,933.17 USD in the elderly group. The differences in mean medical cost were not statistically significant between age groups and underlying diseases (data not shown). Admission rate, mortality and length of hospital stay also showed no statistical difference between age groups ( Table 2).
To evaluate which risk factors contribute to the occurrence of pneumonia in RSV infection, odds ratios (ORs) were calculated for each age group, sex, symptoms, RSV serotype and underlying disease (

Discussion
Our findings suggested that considerable mortality and economic burden of RSV infection in the adult population exist in South Korea. In addition, solid cancer and hematologic malignancy displayed significant relationships with RSV pneumonia.
Recently, the importance of RSV infection in the adult population is increasing globally. However, there have been insufficient studies about the clinical characteristics and disease burden of adult RSV infection. In contrast, studies of RSV infection in childhood are well known and actively conducted in various countries and groups. Hall et al. showed that 18% of children with acute respiratory infections under the age of 5 years had RSV infection. This high infection rate is a serious problem in pediatric public health because it leads to a substantial disease burden regardless of individual underlying illnesses 11,12 . On the other hand, RSV infection in adults seems to be associated with the elderly population and chronic diseases despite the lower rate than in childhood. In a study conducted in the U.S., RSV infection comprised 8.3% of acute respiratory illnesses among the elderly; this proportion was 4% in a French study 13,14 . In Korean studies, laboratory-confirmed RSV infection accounted for 2.8% of all respiratory viral infections and 12.7% of hospitalizations on account of respiratory Table 1. Baseline characteristics and clinical outcomes of RSV infected patients from 2012 to 2015 distributed by age groups. SD standard deviation, ICU intensive care unit. a Stroke includes both hemorrhagic and ischemic strokes. b Respiratory disease includes asthma, chronic obstructive pulmonary disease, bronchiectasis and interstitial lung disease. c Liver disease includes chronic hepatitis B and/or C, liver cirrhosis and autoimmune hepatitis.    15,16 . However, the percentage of RSV among all respiratory illnesses including bacterial infection in Korean adults had not been reported. The seasonal trend of RSV serotype predominance seemed to be alternated biannually, which was also shown in other countries 17 . This pattern might be explained by the acquisition of transient herd immunity of average duration 2 years 18 . It is controversial in children and infants that the serotype variability has an impact on clinical course and outcome of RSV infection 19 . However, none of studies in our knowledge presented precise difference of clinical manifestations between RSV serotype A and B in adults. In our study, the RSV serotype did not seem to be a risk factor of RSV pneumonia. Further clinical and physiologic research would be necessary to clarify the effect of RSV serotype to clinical outcomes.
Most RSV patients were over 65 years of age. Nonhospitalized RSV infected patients were excluded in our study, but they might impose a considerable burden on local clinics as influenza-like illnesses. Nevertheless, the high rates of ICU admission (25.0%), pneumonia (56.8%) and in-hospital mortality (10.6%) in the elderly population signify the severity of adult RSV infection. Lee et al. also reported that 71.9% of adults with RSV infection had lower respiratory complications, and 9.1% died within 30 days 7 . The rate of mortality in our study was elevated from 10.8 to 13.6% when the population was confined to that with pneumonia, which seems to be important in determining the prognosis of RSV infections.
The economic burden of adult RSV infection is not well known in contrast to that among children. In the U.S., RSV infection had similar burden and clinical outcomes to influenza infection in the adult population 2 . Han et al. calculated almost 20 years ago that the mean cost of RSV pneumonia during hospitalization was $11,000 20 . The relatively low cost in our study may be attributed to the discrepancy of medical charge and national insurance systems between the two countries. We set the median medical cost to amend the skewed distribution of individual medical costs. The median medical cost was similar between age groups which may reflect a high prevalence of underlying diseases in all groups. However, a larger and more intensive study of the economic burden of adult RSV infection would be necessary to clarify this pattern. The entire domestic cost of adult RSV infection would be much more increased because of easy diagnosis, aging of the population, and an increasing prevalence of concomitant chronic diseases.
The presence of underlying disease affected the poor outcome of RSV infection in previous studies, especially cancerous or immunocompromised status. Solid organ transplantation and hematopoietic stem cell transplantation were related to increased burden of RSV infection in many studies, and the outcomes worsened when the patients had lower respiratory diseases 21,22 . A study of South Korean adults also showed that chronic lung disease, bacterial co-infection and lower respiratory infection could be associated with increased mortality 23 . Our data revealed that solid cancer and hematologic malignancy were significantly related to the occurrence of RSV pneumonia; however, other underlying illnesses, RSV serotype, sex, and age were not related. This suggests that even young adults could suffer severe RSV infection and bear considerable disease burden. www.nature.com/scientificreports/ The study has several limitations. First, we could not obtain the general incidence of RSV infection in the entire population due to the retrospective study design and the limited area of residence among the participants. In addition, the unbalanced sampling from each participating hospital might lead to selection bias because most of cases were collected from one hospital [Guro hospital: 146 (70.2%)]. Differences of baseline characteristics between three hospitals were shown in Supplementary Table 1. Second, RSV infections with mild symptoms would be overlooked by low rate of PCR performance, clinicians' unawareness and lack of active antiviral agents 24 . This ignorance might underestimate the burden of RSV infection. Third, cases of some underlying diseases and other variables like liver disease and HIV infection were so few that related data might be unreliable. Nevertheless, our study would be helpful to recognize the magnitude of RSV infection in the hospital and to facilitate to make an appropriate countermeasure against adult RSV infection through vaccine development, novel antiviral agents and control of nosocomial transmission.
In conclusion, RSV infection in the adult population showed significant clinical and economic burden and largely related to lower respiratory tract complications and mortality. Solid cancer and hematologic malignancies contributed to the occurrence of RSV pneumonia. The substantial burden of adult RSV infection should be considered to establish public health policies and the target of the future RSV vaccine.