Role of age and birth month in infants hospitalized with RSV‐confirmed disease in the Valencia Region, Spain

Abstract Background RSV is the leading cause of hospital admissions in infants and the principal cause of bronchiolitis in young children. There is a lack of granular data on RSV‐associated hospitalization per season using laboratory confirmed results. Our current study addresses this issue and intends to fill this gap. Methods The study was conducted from 2014 through 2018, in 4 to 10 hospitals in the Valencia Region, Spain. Infants included in this study were admitted in hospital through the Emergency Department with a respiratory complaint and tested by RT‐PCR for RSV in a central laboratory. Results Incidence rates of RSV‐associated hospitalization varied by season and hospital. Overall, the highest incidence rates were observed in 2017/2018. RSV‐associated hospitalization was highest in infants below 3 months of age and in those born before or at the beginning of the RSV season. Almost 54% of total infants hospitalized with laboratory confirmed RSV were found to be born outside the season, from April to October. The RSV positivity rate by ICD‐10 discharged codes varied by season and age with results from 48% to 57% among LRI (J09‐J22). Conclusion The study was instrumental in bringing forth the time unpredictability of RSV epidemics, the critical impact of age, and the comparable distribution of RSV‐associated hospitalization in infants born on either side of the RSV season. These data could help in better characterization of the population that drives the healthcare burden and is crucial for the development of future immunization strategies, especially with upcoming vaccines in against RSV.


| INTRODUCTION
Each year, nearly 33 million cases of acute lower respiratory tract infection (ALRTI) associated with Respiratory Syncytial virus (RSV) are diagnosed in children under 5 years old. 1,2 RSV is the principal cause of bronchiolitis in young children 3 and is globally the leading cause of hospital admissions in infants. 4,5 Although mortality due to RSV is very low among children in high-income countries, 6 yet it plays an important role in the hospital resources utilization. 7,8 RSV is responsible for nearly 16 times more infant hospitalizations than influenza. 9 In temperate climates, RSV circulation starts generally around November and end in March-April. 2,10 Age, prematurity, birth close to the start of the RSV season, and presence of chronic conditions have been commonly identified as potential risk factors for RSV-associated disease. 7,11,12 However, by the age of two, almost all children are infected by RSV. 13 Despite the high unmet medical need, a solution to protect all infants at risk to develop an RSV-associated disease is not yet available, but several candidates are expected to be licensed in coming years. Therefore, understanding the disease burden by month of age and by month of birth to determine who will benefit the most from these vaccination or monoclonal antibodies strategies is valuable to support future immunization policies and recommendations. Usually, burden-of-disease data are estimated using International Statistical Classification of Diseases and Related Health Problems (ICD) diagnosis codes on syndromic surveillance 14 or hospital discharges data. 15 Despite strengths of these studies, they usually do not report laboratory confirmed RSV information. To overcome this issue, we conducted a prospective active-surveillance study during 4 years in Valencia Region, Spain. The objective of this study was to better quantify the incidence of RSV-associated hospitalized disease by season, age, and month of birth. In addition, we aimed to describe the clinical presentation of the RSV-associated disease and to determine the risk factors of infants hospitalized with laboratory confirmed RSV. area of these hospitals was well defined; they covered 21% (4 hospitals) to 46% (10 hospitals) of total inhabitants of the Valencia Region. From November to March/April every year, except in 2017/2018 season (from September to June), the active surveillance of RSV was set up. RSV circulation period was defined as the weeks between the first of at least two consecutive weeks with two or more RSV cases and the week prior to the first of at least two consecutive weeks without RSV cases. Hospitalized patients from all age groups meeting inclusion criteria were enrolled in the study. For the current publication, we considered only infant population aged <1 year.

| Study design
The methodology of the active-surveillance network has been already described in previous publications. [16][17][18] Full-time dedicated nurses screened consecutive hospitalized patients discharged from the Emergency Department with a diagnosis possibly related to a respiratory infection (Table S1). Patients were included in the study if they were resident in the catchment area of one of the participating hospitals, non-institutionalized, and not discharged from a previous admission in the last 30 days. The onset of symptoms that led to hospitalization was required to be 7 days prior to admission, and patients had to be in hospital between 8 and 48 h before their inclusion in the study. Infants under 1 month of age who left the hospital after delivery with no incidents (who were not admitted after birth in the neonatal unit) and who were subsequently hospitalized after a period of 1 week in the community were susceptible to be included in the study.
The Ethics Research Committee of the Direcci on General de Salud Pública-Centro Superior de Investigaci on en Salud Pública (DGSP-CSISP) approved the protocol of the study. All caregivers signed written informed consent prior to inclusion of their infants in the study.

| Laboratory procedures
Nasopharyngeal and nasal (FLOQSwabs, Copan, Italy) swabs were obtained within the first 48 h of admission from each patient fulfilling the inclusion criteria. Both swabs were combined in a tube of viral transport media (Copan, Italy) and frozen between À50 C and À20 C until shipped refrigerated to a centralized Virology laboratory at FISABIO-Public Health. One third of the viral transport media volume was used to extract total nucleic acids using an automated silica-based method (Nuclisens Easy-Mag, BioMérieux, Lyon, France).
Extracted nucleic acids were tested for RSV, influenza, and other respiratory viruses (a total of 19) by multiplex real-time reverse transcription-polymerase chain reaction (RT-PCR), following WHO protocols 19 with the qScript XLT One-Step RT-qPCR ToughMix (Quanta BioSciences, MD, USA) in a Lightcycler 480II apparatus (Roche Diagnostics, Spain).

| RSV-associated hospitalization incidence rates by season
We calculated the RSV hospitalization incidence rates per 100,000 infants-season overall and by hospital. The catchment area of each participating hospital along the different seasons was considered as the denominator. The RSV circulation period was estimated by epidemiological weeks, using EPIWEEK STATA module.
2.4.2 | RSV-associated hospitalization incidence rates by age and birth month According to age (0 to 11 months) and birth months (January to December), we calculated the RSV hospitalization incidence rates per 100,000 infants-season. The numerator was the number of RSV cases by age or by birth month for each season. The denominator was estimated by dividing the catchment population under 1 year old by 12. Due to the different duration of the seasons and to allow comparisons between them, the RSV hospitalization incidence rates were provided per 100,000 infants-week and per 100,000 infants-month (restricted to the RSV circulation period).

| Risk factors of infants hospitalized with laboratory confirmed RSV
Comparison between RSV positive and RSV negative hospitalizations was conducted based on the following parameters: (1) birth month,

| Impact of age on RSV-associated hospitalization
A negative binomial regression, a generalization of the Poisson regression model that addresses the over-dispersion issue, was performed to assess the impact of age (≤3 months, >3 months) on RSV hospitalization incidence rates, after adjusting by calendar month at hospital admission (restricted from November to March), hospital and season.
The population denominator was included as an offset. The adjusted relative risks (aRR) and their 95% confidence intervals (CIs) were provided.
2.4.5 | RSV-associated disease-RSV positivity rate according to ICD-10 discharge diagnoses by season and age Hospital discharge information using the 10th revision of ICD diagnosis codes (ICD-10) was retrieved for each infant included in the study.
RSV-associated disease was defined based on the following ICD-10 codes recorded at hospital discharge: Lower Respiratory Infection (LRI): J09-J22, bronchiolitis: J21 and pneumonia: J12-J18. These outcomes were described by season and age group (<3 months, 3 to 5 months, and 6 to 11 months). Laboratory confirmed RSV results were used to determine RSV positivity rates for each disease outcome.
All statistical analyses were carried out in Stata version 14 (StataCorp, College Station, Texas) and R (Viena, Austria). All probabilities were two-tailed, and p values <0.05 were considered significant.

| RSV-associated hospitalization incidence rates by season, age, and birth month
The overall RSV-associated hospitalization incidence rates ranged   percent of RSV positive cases were exposed to tobacco versus 39% of RSV negative cases (Table 2).   Figure 5).
3.6 | RSV-associated disease-RSV positivity rate according to ICD-10 discharge diagnoses by season and age Over the four seasons, 601 infants with an ICD-10 LRI code (J09-J22) were RSV laboratory confirmed, which represents 51% of the infants included in the study with this discharge code. This positivity rate var-  Table 3.

| DISCUSSION
In this study, we explored RSV laboratory-confirmed hospitalizations in infants <1 year old in a prospective, active-surveillance, multicenter infants with a defined efficacy period 22 and, therefore, should be continuously monitored to identify the best timing to implement these strategies.
Overall, the RSV A subtype was most detected, although subtype B prevailed in 2016/2017. Similar studies have also detected the RSV A subtype predominance over time. 11 As the RSV circulation period did not last the same for all seasons (from 16 to 20 weeks), we calculated the RSV hospitalization incidence Similar observation was made in other studies, 6 demonstrating the unpredictability of the severity of RSV-associated disease.
In the effort to better characterize infants hospitalized due to RSV, we found that majority of infants are full terms and otherwise healthy, which is in agreement with previous publications. 23,24 In addition, we observed that the month of birth and age were critical factors for hospitalization due to RSV. Highest rates were detected in young ones and in those born before or at the beginning of the RSV season.
Previous studies also reported the same findings. 6,20,[25][26][27][28][29] RSVassociated hospitalization risk in infants born at the end of RSV season (January to March) appeared to be lower than the ones born between August and September. Transfer of maternal antibodies certainly has a role here. 30 However, this protection appears to be short with the RSV-associated hospitalization incidence peak in infants at 1 month old, regardless of the season. The incidence rates started to decrease for infants >2 months old, although it remained quite high in infants with 3 months of age in 2016/2017.
Another important finding of this study is the percentage of laboratory-confirmed RSV per ICD-10 code. RSV has been reported as the main cause of bronchiolitis in infants in several publications. 28,31,32 We detected RSV in 58% of the infants discharged from hospital with a bronchiolitis ICD-10 code, in 51% of LRI and in 39% of pneumonia, with variations based on season and age. Hospital data registries are commonly used to estimate RSV-associated hospitalizations at national or regional level. 5,9,33 However, hospitalized patients are tested for RSV at clinician's discretion. Therefore, some RSV cases could be missed, leading to an RSV misclassification. 9 By contrast, the VAHNSI network tested by RT-PCR any admitted patient with a suspicion of a respiratory infection. Then, our laboratory-confirmed information could support better ascertainment of RSV-associated hospitalization by adjusting results from ICD-10 code to the positivity rates described in our data. Another method of adjustment was performed by Arriola et al. 23 in United-States that account for testing practices and test method sensitivities.
The role of other respiratory viruses on RSV infections has been previously analyzed. 11,34 We detected the presence of co-infections due to other respiratory viruses among 13% of RSV-positive cases.
Other studies reported similar percentages, ranging between 10% and 16%. 23,35 In our study, the commonest mixed infection was the combination of RSV and rhinovirus/enterovirus. This is in agreement with other publications, 11,36 although different combinations were more frequent in other studies, as Bonzel et al. 35 reported, RSV and human bocavirus as the most common detected coinfection. No agreement has been reached regarding the severity of the RSV mixed infections.
Whereas some studies reported a more severe course in patients with RSV coinfection, 37 others did not find differences in the clinical severity between RSV single infections and coinfections. 38 This work presented some limitations. The different durations of the seasons and the variability among participating hospitals policies can make comparisons hard to interpret. We approached the denominators for each month of age by dividing the total population under 1 year old by 12 months. Although it was a satisfactory approach, we lost some precision in the estimates.
Despite these limitations, our study was robust by using laboratory confirmed data and a well-defined catchment population. We provided recent data from four consecutive seasons from a hospital network with large experience in observational studies, using laboratory-confirmed RSV cases and the same testing practice in a centralized laboratory. Very few networks presented these methodologies 24,39 and, in some studies, different specimens and diagnostic assays were used at site level, mixing commercial and institutionspecific in-house RT-PCR in the same network. 24 Our catchment population was very well defined, so we had the possibility of calculating hospitalization incidence rates to estimate the burden of the disease, providing more precise estimates than those obtained when using non-population-based study. 39,40 Finally, we were able to link clinical, demographic and laboratory data to improve the characterization of RSV-associated hospitalizations in infants.

| CONCLUSIONS
There is a lack of prospective, hospital-based, active-surveillance epidemiological studies on RSV thereby limiting the knowledge on how this pathogen impacts infant health. The VAHNSI network is a crucial component required to inform and to support prospective immunization that will occur in near future to protect infants at risk of RSV. This study brings forth the unpredictability of RSV epidemics, the critical impact of age, and the comparable distribution of RSV-associated hospitalization among infants born outside or during the RSV season.
There is critical need for solutions to decrease the burden of RSV on health resources consumption which is driven by otherwise healthy infants. This study provides some views to support decision makers to decrease hospitals overcrowding associated with RSV infection.

ACKNOWLEDGMENTS
We gratefully acknowledge the staff of Hospital General Universitario

FUNDING INFORMATION
This collaborative study has been partly funded by FISABIO-Public Health (Fundaci on para el Fomento de la Investigaci on Sanitaria y Biomédica de la Comunitat Valenciana), Sanofi Pasteur, and CIBER-ESP (ISCIII).

PATIENT CONSENT STATEMENT
All patients signed a written informed consent before their inclusion in the study.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.