Characterisation of respiratory syncytial virus activity in children and adults presenting with acute respiratory illness at primary care clinics in Singapore, 2014‐2018

Abstract Background Respiratory syncytial virus (RSV) is an important respiratory pathogen that affects people of all ages. Objectives We examined the patterns of RSV circulation in 2014‐2018, and investigated their age‐specific differences in tropical Singapore. Methods Nasopharyngeal and/or throat swabs were taken from outpatient attendees for the national influenza virological surveillance among those who presented with acute respiratory illness in the community. Specimens tested negative for influenza were then tested for RSV and other respiratory pathogens. Results Among 8436 influenza‐negative specimens tested during the five‐year period, 5.8% (95% confidence interval 5.3%‐6.3%) were positive for RSV. The peak of RSV activity occurred around middle of the year. The age‐specific proportion of RSV detections showed a reverse J‐shaped pattern; RSV positivity was the highest in young children ≤2 years of age (10.9%), followed by those aged 3‐5 years (6.4%) and persons aged ≥65 years (5.3%), while the nadir was observed in the age group of 15‐24 years (1.2%). RSV type A was predominantly circulating in children ≤5 years of age from 2014 to 2015 and 2017, whereas in 2016, they were more affected by type B. Conclusion Respiratory syncytial virus was more frequently detected among the two age groups that have been recommended for influenza vaccination; persons ≥65 years of age and children 6 months to <5 years of age. Characterisation of RSV activity in the community helps to better inform public health policies for effective prevention and control interventions.


| INTRODUC TI ON
Respiratory syncytial virus (RSV), a member of the Paramyxoviridae family, is recognised as an important respiratory pathogen that affects people of all ages. Lower respiratory tract infection (LRTI) due to RSV is a leading cause of paediatric hospitalisations worldwide. [1][2][3][4] There is also substantial disease burden from RSVassociated acute respiratory illness (ARI) among persons ≥65 years of age. 5 Most studies that have attempted to quantify the burden of LRTIs were based on hospital-based surveillance, and they focused on children <5 years of age or subpopulations such as immunocompromised children and adults. 6 The most common causes of respiratory viral infections are RSV and rhinoviruses in these studies, including those in Asia. [7][8][9][10] Active surveillance of healthy children aged 6 months to 10 years with influenza-like illness (ILI) enrolled in a randomised trial at 17 centres in eight countries (including Singapore) between February 2010 and August 2011 demonstrated considerable burden of RSV-associated illness in the community. 11 Despite the considerable burden of respiratory viral infections, there are few effective pharmacologic interventions to mitigate the health impact of these pathogens other than for influenza. 12 In 2016, the World Health Organisation (WHO) piloted a two-year project to test the feasibility of implementing RSV surveillance based on the Global Influenza Surveillance and Response System (GISRS) in 14 countries across all six WHO regions. 13 The initiative arose from recognition of the need to provide the evidence base such as seasonality, disease burden and risk groups, to guide future RSV vaccination and other prevention programmes. 13 As of end 2019, none of the RSV vaccine candidates have reached licensing, and treatment of RSV infection is primarily supportive.
While surveillance of influenza and other respiratory viruses may include testing for RSV, there has been limited data on the epidemiology of RSV, particularly in tropical settings. The aim of this study was to describe the pattern of RSV circulating in the community based on sentinel surveillance of outpatient attendees who presented with ARI in Singapore, a globally connected city-state in the tropics in Southeast Asia. We further sought to investigate the age-specific differences in RSV activity.

| Virological surveillance and laboratory methods
The Ministry of Health (MOH) conducts the National Surveillance Programme for Influenza (NSPI) throughout the year. For the virological surveillance in the community, nasopharyngeal and/or throat swabs are taken from outpatient attendees with an ARI and measured fever of ≥38°C and cough at government-funded primary care clinics and sentinel private general practitioner (GP) clinics after obtaining verbal informed consent. 14,15 The decision to test is based on clinical judgement, and outpatient attendees are not systematically enrolled based on symptoms. For consenting outpatient attendees who fulfil the inclusion criteria, information such as the date of symptom onset, travel history in the past 10 days and influenza vaccination in the past 6 months, if any, are recorded on a data collection form. The specimens are stored in viral transport medium and despatched by courier delivery to the National Public Health Laboratory (NPHL), Singapore, on the same day for nasopharyngeal and/or throat swabs taken on weekdays, and on the following Monday for those taken on weekends.
Testing of the specimens collected for the virological surveillance under NSPI was conducted by NPHL. Real-time reverse transcription polymerase chain reaction was used to determine influenza virus types and subtypes. 14
In view of the smaller numbers in the age groups of 6-14 years and ≥65 years, we did stratification by four broad age groups for the analysis on monthly RSV detections and distribution of RSV types among RSV-positive specimens: ≤5 years, 6-24 years, 25-44 years, ≥45 years. We determined the proportion of influenza-negative specimens tested positive for RSV as a measure of RSV activity in the community. The 95% confidence intervals (CI) for binomial proportions were calculated using Wilson's method. The Mantel-Haenszel linear-by-linear association chi-square test was used to evaluate whether there was linear trend in annual proportion of RSV detections across the five-year study period. A two-proportion z test was used to assess the difference in the proportions of RSVpositive specimens between any 2 years. A RSV type was deemed to be predominant if its proportion was 10% points or higher than that of the other type among RSV-positive samples. All statistical tests were two-sided, and P-values <.05 were considered statistically significant.

| RE SULTS
During the five-year period from 2014 to 2018, 16 877 specimens from outpatients who presented with ARI at outpatient setting were collected and 8436 (49.4%) tested negative for influenza. The annual number of influenza-negative specimens that were tested for RSV increased from 1281 in 2014 to 1892 in 2018 (Table 1). About 45.7% of the influenza-negative specimens were from children aged ≤5 years, 32.7% from adults aged 25-64 years and 5.2% from the age group of ≥65 years. The median age was 7 years (interquartile range 2-38).
Among the 8436 influenza-negative specimens, 487 (4.8%, 95% CI: 5.3%-6.3%) tested positive for RSV. There was no significant linear trend in the annual proportion of RSV-positive samples across the study period (P = .122), and in the seven age groups (all P > .10).
The proportion of RSV detections dropped significantly from 7.6% in 2014 to 4.7% in 2015 (P = .001) ( Table 1) The RSV positivity was consistently highest in infants and toddlers aged ≤2 years over the five-year study period (Table 1)    Based on the national virological surveillance in Singapore, seasonal influenza epidemics usually occur around the start or end and middle of the year. 15 Our study revealed that the peak of RSV activity among outpatient attendees in the community occurred mostly around middle of the year as well, which coincided with that of influenza. The monthly proportion of RSV detections exceeded the threshold of 10% in two years, 2014 and 2018, which was set by the US Centers for Disease Control and Prevention to define a RSV season. 23 In temperate climates, the timing of epidemics associated with influenza often overlap with that of RSV. 24,25 Unlike temperate regions where RSV exhibits distinct seasonality with onset starting in late fall or early winter and ending in late spring, 26 the epidemiological features of influenza and RSV in the tropics are more varied.

| D ISCUSS I ON
We observed three spikes in the monthly RSV positivity in each year from 2014 to 2016, followed by a unimodal peak in the ensuing two years (Figure 2).
The age-specific RSV positivity in Singapore depicted a reverse J-shaped pattern; the RSV positivity was highest in infants and toddlers ≤2 years of age, followed by children aged 3-5 years and persons aged ≥65 years (Figure 1). RSV is a common infection in infancy; it has been estimated that over 95% of children have been infected with RSV by 2 years of age. 27,28 A prospective study conducted during the peak of five influenza seasons in Canada revealed a higher burden and severity of infections due to RSV compared with influenza in hospitalised children <2 years of age. 29 While RSV has been primarily seen as a cause of illness in infants and children, studies in hospitalised adults have led to increasing recognition of the RSVassociated burden among elderly persons. 30 While RSV predominately affect young children, 35 there has been more evidence indicating that the impact on older adults, particularly the elderly, is similar to that of seasonal influenza. 20,36,37 In Hong Kong, RSV patients were more likely to have underlying chronic lung disease and major systemic co-morbidities when compared with influenza patients of similar age, while the mortality rate of RSV patients with severe lower respiratory complications was similar to that of seasonal influenza. 38 In the United States, RSVattributable healthcare resource use (hospital stays, emergency room/urgent care visits, ambulatory visits and outpatient visits) and costs across age groups were substantial, with the highest burden in the elderly aged ≥65 years. 39 In our study, RSV type A was predominantly circulating in children ≤5 years of age from 2014 to 2015 and 2017, whereas in 2016, they were more affected by type B (Figure 3B). The antigenic differences between the two RSV types have been found to affect susceptibility to infection or disease. 40 However, differences in the clinical severity between the two RSV types remain unclear. In a prospective study on infants hospitalised with bronchiolitis, RSV type A infection resulted in greater severity, 41 which was in agreement with most published studies. [42][43][44] Other studies had either found infection due to RSV type B to be more severe, 45 or no significant association between clinical severity of RSV infection and type. 46,47 Possible differences in disease severity between RSV types A and B might have important implications for prevention and treatment strategies, including future vaccination and clinical management. 43 As the RSV activity in this study was reflected by the proportion of RSV-positive specimens that had been tested negative for The data for our study originated from a nationwide network of In conclusion, RSV was most frequently detected among age groups that have been recommended for influenza vaccination; young children aged ≤5 years followed by persons aged ≥65 years who presented with ARI in the local outpatient setting. Given the varied patterns of RSV circulation shown in this study, there is a need for continuous surveillance throughout the year. The characterisation of RSV activity provides insights into the epidemiological aspects for review of public health strategies to prevent and reduce transmission of RSV in the community.

ACK N OWLED G EM ENTS
The authors thank the staff at government primary care clinics and