key: cord-0868103-ati3xouu authors: Jayaweera, J. A. A. S.; Morel, A. J.; Abeykoon, A. M. S. B.; Pitchai, F. N. N.; Kothalawela, H. S.; Peiris, J. S. M.; Noordeen, F. title: Viral burden and diversity in acute respiratory tract infections in hospitalized children in wet and dry zones of Sri Lanka date: 2021-12-17 journal: PLoS One DOI: 10.1371/journal.pone.0259443 sha: 74663a428f029f06db7c439760dff0d87bab6cdb doc_id: 868103 cord_uid: ati3xouu The present study was done to identify the viral diversity, seasonality and burden associated with childhood acute respiratory tract infection (ARTI) in Sri Lanka. Nasopharyngeal aspirates (NPA) of hospitalized children (1 month—5 years) with ARTI were collected in 2 centers (wet and dry zones) from March 2013 to August 2014. Respiratory viral antigen detection by immunofluorescence assay (IFA) was used to identify the infecting viruses. IFA negative 100 NPA samples were tested for human metapeumovirus (hMPV), human bocavirus and corona viruses by polymerase chain reaction. Of the 443 and 418 NPAs, 37.2% and 39.4% were positive for any of the 8 different respiratory viruses tested from two centers studied. Viral co-infection was detected with respiratory syncytial virus (RSV) in both centers. Peak viral detection was noted in the wet zone from May-July 2013 and 2014 and in the dry zone from December-January 2014 suggesting a local seasonality for viral ARTI. RSV showed a clear seasonality with a direct correlation of monthly RSV infections with rainy days in the wet zone and an inverse correlation with temperature in both centers. The case fatality rate was 2.7% for RSV associated ARTI. The overall disability adjusted life years was 335.9 and for RSV associated ARTI it was 241.8. RSV was the commonly detected respiratory virus with an annual seasonality and distribution in rainy seasons in the dry and wet zones of Sri Lanka. Identifying the virus and seasonality will contribute to employ preventive measures and reduce the empirical use of antibiotics in resource limited settings. Acute respiratory tract infection (ARTI) represents one of the most common acute illnesses in childhood. ARTIs range from common cold, a mild, self-limited catarrhal syndrome to life threatening lower respiratory tract infections [1] . Viruses account for most ARTIs and several located in the North Central Province (Dry zone), and the Teaching Hospital, Gampola (THG) located in the Central Province (Wet zone) of Sri Lanka (Fig 1) . To study the year-round incidence of different viral ARTI, recruitment of participants for the study was done on consecutive days throughout the year for 18 months from March 2013 to August 2014. Hospitalized children were recruited with ARTI with severe acute respiratory illness (SARI) using the case definition of the World Health Organization (WHO) [19] . The clinical diagnosis of ARTI was done by the consultant pediatricians using the ICD-10 2012 guidelines [20] . A child was included in the study after obtaining the guardian's or parents' consent following the objectives explained in English or the local languages. Patient's history and data on clinical signs and symptoms, demography and risk factors for acquiring ARTI were collected using an investigator administered questionnaire by interviewing the parent or guardian with the informed written consent. Then the nasopharyngeal aspirates (NPA) were taken using the recommended mucous aspirator with minimal discomfort to the child by the pediatrician or his trained medical staff to test for the infecting respiratory virus. NPA was collected with sufficient number of respiratory nasopharyngeal epithelial cells (NPC) as these cells have the virus infection. NPA was then released into 8 mL phosphate buffered saline (PBS) and transported on ice to the laboratory (4˚C) within a few hours of collection. Indirect immunofluorescence assay (IFA) was performed using the D3 (USA) [21] with comparable sensitivity with culture and a required specificity as stated by the manufacturers. IFA covered screening for 8 respiratory viruses including RSV, AdV, PIV-1, PIV-2 and PIV-3, Inf-A and Inf-B, and hMPV and then the IFA screening positive NPA were subjected to identify the infecting virus using monoclonal antibodies provided by the D3 Ultra DFA Respiratory Virus Screening and ID Kit (USA). Stained slides were examined under UV-Epi fluorescence microscope (Leads, Germany) and intracellular nuclear and/or cytoplasmic apple green fluorescence emitting cells were recorded as virus infected. IFA negative samples were then vortexed, transferred into 1.5 mL Eppendorf tubes and stored at -70˚C for polymerase chain reaction (PCR) to detect hMPV, hBoV, and coronaviruses (CoV) [22] . To detect respiratory viral co-infections, a qualitative PCR was performed for hMPV, hBoV and CoVin100 RSV positive children. Moreover, 100 selected IFA negative children's samples were subjected to PCR to detect hMPV [23] , hBoV [24] and CoV [25] as primary etiological agents in ARTI using previously published protocols. All the laboratory tests for the study were done in the Diagnostic and Research Virology Laboratory of the Department of Microbiology, Faculty of Medicine, University of Peradeniya, Sri Lanka. Data obtained was double entered into a spreadsheet in Microsoft 1 Excel and cleaned for double or wrong entries. Associations between categorical and continuous variables were analyzed using the Chi square tests or Fisher's exact tests, and the 2-tailed Student's t test or Wilcoxon rank-sum test as appropriate. Demographic data, clinical features, disease spectrum, mean hospital stay and use of antibiotics in RSV, PIV-1, -2 and -3, AdV, Inf-A, Inf-B, and hMPV associated ARTIs were expressed in measures of central tendency (Table 1 and Supplementary 1 in S1 File). Multivariable analyses were performed using a step wise logistic regression analysis to assess the risk factors for THG THA THG THA THG THA THG THA THG THA THG THA THG THA THG THA Number of children with coinfections 13 Age at presentation, mean months ± SD 10±9.1 11±9. Un specified lower ARTI developing RSV associated ARTI vs other tested individual viral ARTIs. Initially, a univariate analysis was done and p<0.1 was selected for the multivariate analysis and modeling. Modeling was done for RSV positive vs. positive for any of other seven viruses. Variables such as age at hospitalization (nearest month), duration of the disease (days), gender, ethnicity (Sinhala, Tamil, Moor, and other), weight (kg), and Hb%, chronic malnutrition (height-for-age � 2SD was taken while weight-for-age �2SD also considered to assess malnutrition), gestational age considering the maturity of the mother, mode of delivery (vaginal or lower segment cesarean section), presence of underlying medical conditions (congenital heart disease-CHD; chronic lung disease-CLD, asthma, cystic fibrosis) and genetic disorders: Down's syndrome-trisomy 21; neuromuscular disorders and pre-existing respiratory tract morbidity, passive smoking (the involuntary inhalation of cigarette smoke from other smokers or the child's father), having house hold pets, presence of indoor smoke through fire wood used for cooking / houses without chimney and outdoor air pollution through construction activities outdoor or industries), crowding (child's living area <24 m 2 ) [26], guardian's/parent's education (1 child) and occupation. Odds ratio was calculated for these variables with consideration of significance at p value of <0.05. Multivariate odds ratios with 95% CIs that did not include 1.0 were considered as significant and the continuous variables were expressed as measures of central tendency. Children with viral co-infections were excluded in clinical data analysis. A "peak" in virus activity was defined when the monthly proportion was �10% during the study period. The correlation between the monthly incidence of overall viral ARTI and climatic factors (mean atmospheric temperature in a given month (mTm in˚C), mean relative humidity in a given month (mRH in %), and mean number of rainy days in a given month, windspeed and direction and atmospheric pressure) was determined using Spearman correlation coefficient and multiple linear correlation coefficient. Further a separate analysis was done between the monthly incidence of RSV associated ARTI, and climatic factors (Supplementary 4 in S1 File). In addition, the effects of climatic factors and their interaction were calculated using the geographical detector method. The geo-detector q-statistic measures the degree of special heterogeneity of a particular variable (Y) and the determinant power of an explanatory variable (X) of Y; the value of q-statistic is strictly within 0, 1 [27] . The q-statistic was calculated to evaluate the spatially stratified heterogeneity (Supplementary 5 in S1 File) as described in Xu et al. [28] . Disability adjusted life years (DALYs) was calculated using the Health Statistics and Information Systems WHO, 2000 WHO, -2011 . The statistical analysis was done using the Statistical Analysis System (SAS), Version 9.1 [30]. The viral, demographic and clinical characteristics data of the study are summarized in Table 1 and Supplementary 1 in S1 File. A total of 418 and 443 children were included over the period of 18 months in THA and THG, respectively. Of the tested, 165 children were positive for respiratory viral antigen by IFA from each of the center for one or more viruses. In both study centers, the number of boys with ARTI was significantly higher compared to girls with ARTI (p = 0.03). Considering the ethnicity in THA and THG study samples collectively, the number of Sinhalese (ethnic majority in Sri Lanka) children with ARTI was significantly higher (p = 0.02, 0.01) compared to Tamil (an ethnic minority in Sri Lanka) and Muslim (another ethnic minority in Sri Lanka) children. Compared to THA, in the THG study sample, number of Tamil children with ARTI was significantly higher than other ethnic groups (p = 0.03). Considering the residential areas of THA and THG study samples, the number of children with ARTI from the rural areas (away from the urban and suburban limits) (p = 0.03) was higher than that from the semi-urban and urban areas, respectively. In the THG study sample, many children with ARTI were from semi-urban areas when compared to the THA study sample (p = 0.04). Viral ARTI was more common in 12-24 months-old children in both study centers compared to 1-�12 and � 24-� 60-month-olds. No significant differences were noted in the distribution of viral ARTI in different age categories between the two study centers. Body weight was in median to -1 SD to -2 SD in THA and -1 SD in THG study samples. Blood hemoglobin (Hb)% was 9.1 ± 1.9 g/dl in the THA and it was 9.3±2.1 g/dl in the THG. Body weight and Hb% were not significantly different among children with ARTIs or viral ARTIs between the two study centers. All children have had the routine vaccine coverage from the national immunization program, which does not include immunization against influenza. Moreover, none of the study participants had vaccination against influenza. Clinical characteristics of respiratory virus positive children are summarized in Table 1 . Multiple viral infections were detected in 29 children. In both centers, co-infection with RSV was significantly higher compared to other viral co-infections (p = 0.03). The most important symptoms in children for mothers to seek medical assistance were rapid respiratory rate (60%), fever (48%), wheezing (33%), and anorexia (20%). When compared to PIV-1, PIV-2, PIV-3, Inf-A, Inf-B, and hMPV infections, fever was less frequently observed in children with RSV or AdV infections (p<0.02). When compared to PIV-1, PIV-3, AdV, Inf-A, Inf-B, and hMPV infections, respiratory rales were frequently auscultated in children with RSV or PIV-2 infections (p<0.03). Chest recession was frequently observed in children with RSV or PIV-2 infection (p<0.03). Compared to RSV, PIV-1, PIV-2, PIV-3, Inf-A, Inf-B, and hMPV infections, conjunctivitis (p<0.04) and diarrhea (p<0.03) were frequently observed in children with AdV infection. Tachypnea was frequently observed in children with RSV or hMPV infections compared to PIV-1, PIV-2, PIV-3, AdV, Inf-A, and Inf-B infections (p<0.03). Bronchiolitis was frequently observed in children with RSV infections compared to PIV-1, PIV-2, PIV-3, AdV, Inf-A, Inf-B, and hMPV infections (Supplementary 2 in S1 File) (p<0.02). Children in all age category had bronchiolitis, however, bronchiolitis was frequently noted in children between 1 and 24 months (p = 0.03). There was no difference in the duration of illness in different types of viral ARTI. Past history of ARTI predisposed the children to a current viral infection (p = 0.04). Monthly distribution of ARTIs in THA and THG are shown in Fig 2. In both centers, ARTI was detected throughout the study period. ARTI peaked in the THA from December to January and in the THG from May to July. Overall, ARTI peaked during the rainy seasons. In the dry zone (THA), the peak ARTI was associated with north-east monsoon (from December to March) (Fig 3) and in the wet zone (THG), the peak ARTI was associated with south-west monsoon (from May to September) (Fig 4) . In the dry zone (THA), two peaks of RSV infection were noted between December-January in 2013 (major peak) and in April in 2013 and 2014 (minor peak) (Fig 3) . RSV infections were detected throughout the year in both centers. In the wet zone (THG), the peak of respiratory viral infections was noted from April to June in 2013 and 2014 (Fig 4) . In Overall, the explanatory power of the climatic factors (atmospheric temperature, relative humidity and mean number of rainy days) was greater in boys than girls for each age group (1-�12, 12-24 and � 24-� 60 months). However, no significant difference was observed in the explanatory power of the climatic factors on viral ARTI between THA and THG (Supplementary 6 in S1 File). Moreover, Spearman's rank correlation and multiple linear regression showed a direct correlation of monthly viral ARTI cases with number of rainy days in both THA and THG centers and an inverse correlation with the mean atmospheric temperature. Additionally, regression analysis identified a further significant inverse correlation of viral ARTI cases with relative humidity in both centers. Monthly increase of one rainy day was associated with a 0.78 increase in monthly viral ARTI cases in THA and 0.88 increases in monthly viral ARTI cases in THG based on the multiple regression. However, in THA and THG, an increase of 1% of relative humidity and 1˚C of temperature were associated with 1.91 and 1.83 and 2.82 and 2.76 decrease in viral ARTI cases, respectively. In THA, a total of 32.3% of explained variance (R 2 = 0.323) in the number of monthly viral ARTI cases was attributed to number of rainy days in a month, relative humidity and temperature. In THG, a total of 33.3% of explained variance (R 2 = 0.333) in the number of monthly viral ARTI cases was attributed to the climatic factors. Other climatic factors such as windspeed and wind direction and atmospheric pressure did not have an influence on viral ARTI. Since RSV infection showed a clear seasonality with the highest number of cases, the association of RSV with climatic factors was further analyzed. For RSV infection in the THA study sample, the explanatory power of the atmospheric temperature was 0.49 (p<0.05) and that of the relative humidity was 0.42 (p<0.05). For RSV infection in the THG study sample, the explanatory power of the atmospheric temperature and relative humidity were 0.51 (p<0.05) for both variables. For RSV infection in the THG study sample, the explanatory power of the mean number of rainy days was 0.52 (p<0.05). Spearman's rank correlation and multiple linear regressions showed a direct correlation of monthly RSV cases with number of rainy days in the THG study sample and an inverse correlation with mean atmospheric temperature in both study samples. Additionally, regression analysis identified a significant inverse correlation of RSV cases with relative humidity in THA and THG study samples. Based on the multiple regression analysis, a monthly increase of one rainy day was associated with a 0.58 increase in monthly RSV cases in the THG. However, in THA and THG, an increase of 1% of relative humidity and 1˚C of temperature were associated with 1.21 and 1.03 and 1.92 and 1.86 decreases in RSV cases, respectively. In the THA study sample, a total of 20.3% of explained variance (R 2 = 0.203) in the number of monthly RSV cases was attributed to number of rainy days in a month, relative humidity, and temperature. In the THG study sample, a total of 23.3% of explained variance (R 2 = 0.233) in the number of monthly RSV cases was attributed to the same. Other climatic factors like the wind speed and wind direction and atmospheric pressure were not significantly associated with RSV infections. The pattern of distribution of RSV associated ARTI with climatic factors (mean monthly relative humidity (%), mean monthly temperature, mean monthly number of rainy days) in THA and THG study samples are shown in Fig 5 and Supplementary 4 in S1 File. The supplementary 7 in S1 File shows the interactions of the paired climatic factors on viral ARTIs and RSV associated ARTI at THA and THG. The dominant interaction is in bold letters where the q value is higher than the sum of each individual factor. The dominant interactors on overall ARTI (q = 0.71) at THA (q = 0.58) and THG (q = 0.72), viral ARTI and RSV associated ARTI at THA (q = 0.72) was atmospheric temperature and atmospheric pressure. The dominant interactor on RSV associated ARTI at THG was the mean number of rainy days and the mean atmospheric pressure (q = 0.74) (Supplementary 7 in S1 File). Viral co-infection was not included in risk factor analysis ( Table 2) . Male sex was a significant risk factor for acquiring RSV infection (Odds ratio = 2.4; p = 0.03) and other 7 respiratory viral infections (Odds ratio = 2.3; p = 0.03). Malnutrition was a significant risk factor for acquiring PIV-2 compared to RSV infection in THA (Odds ratio = 1.8; p = 0.02) and THG (Odds ratio = 2.1; p = 0.02). Birth weight <2500 g was a significant risk factor for acquiring AdV infection compared to RSV in children with ARTI in THA (Odds ratio = 1.7; p = 0.04) and THG (Odds ratio = 1.8; p = 0.04). Living in crowded conditions was a significant risk factor for acquiring AdV infection compared to RSV in children from THA (Odds ratio = 2.8; p = 0.03) and THG (Odds ratio = 3.2; p = 0.03). Birth through LSCS was a significant risk factor for acquiring hMPV infection compared to RSV in children from THA (Odds ratio = 1.9; p = 0.05) and THG (Odds ratio = 2.4; p = 0.05). Considering the co-morbidities, having congenital heart disease was a significant risk factor for acquiring RSV and AdV infection compared to 6 No. of co-infections a 13 13 3 3 3 4 0 0 4 4 1 1 0 1 4 4 Study site THG THA THG THA THG THA THG THA THG THA THG THA THG THA THG THA Risk factors Malnutrition (weight-for-age z-score Outdoor air pollution e -3.2 (0.04) Breastfeeding (during the first 4 months) - - - - - - - - - - - - - - - - Crowded living - - - - - - - - 3.2 (0.03) 2.8 (0.03) - - - - - - Concurrent/ Congenital heart diseases 3.4 (0.02) 3.2 (0.02) - - 2.4 (0.03) 2.3 (0.03) - - - - - - - - - - Asthma - - 1.4 (0.04) 1.4 (0.04) - - - - - - - - - - 2.1 (0.03) 2.1 (0.03) Immunodeficiency - - - - - - - - - - - - - - - - Epilepsy - - - - - - - - - - - - - - - - Viral co-infections were excluded from the analysis. Only significant odds ratio and P value are given. P value was calculated in multivariate analysis and P< 0.05 was considered significant and given in parenthesis in relevant columns. Viral burden and diversity in acute respiratory tract infections in hospitalized children other respiratory viral infections in THA and THG. Having asthma was a significant risk factor for acquiring PIV-2 and hMPV infections compared to RSV infection in THA and THG. Considering the socio-demographic factors, poor educational level of the mother (up to primary level only) was a significant risk factor for acquiring RSV and hMPV infections compared to 6 other viruses in THA and THG. Day care attendance was a significant risk factor for acquiring PIV-1 or PIV-2 or AV infection compared to RSV in THA and THG. In addition, outdoor air pollution was a significant risk factor for acquiring RSV and PIV-2 infection in THA compared to 6 other viruses. Indoor air pollution was a significant risk factor for acquiring PIV-3 infection compared to RSV in both THA and THG. Considering the climatic factors, high relative humidity (not the mean monthly relative humidity) was a significant risk factor for acquiring RSV infection in children compared to all seven viruses in THA and THG. Similarly, high atmospheric temperature was a significant risk factor for acquiring RSV infection. Furthermore, having more than 15 rainy days in a month was a significant risk factor for acquiring RSV infection. Empirical antibiotic use was noted in all children with ARTI, in the study sample. Viral incidence/100,000 person years for ARTI are given in Table 3 and Fig 6. Overall, RSV associated ARTI incidence was 29.76/100,000 person years and it was significantly higher for RSV than the rest of the viral ARTI (p = 0.001). In the THA study sample, 6 deaths were recorded for viral ARTI. Two of these children had RSV infection alone, 2 had RSV and hMPV co-infection (Supplementary 3 in S1 File) and 2 had RSV and PIV-1 co-infection (Table 4 ). In the THG study sample, 3 deaths were recorded for viral ARTI during the study period. Two of these children had RSV infection alone and one had RSV and PIV-1 co-infection. Case fatality rate for childhood ARTI at THA and THG was 1.43% (6/418) and 0.67% (3/443), respectively. Overall RSV associated case fatality rate in THA and THG study samples was 3.63% (6/165) and 1.81% (3/165), respectively. Case fatality rate for RSV mono-infection for both centers was 1.21% (2/165). DALYs in THA and THG were 429.77 and 242.06, respectively. DALYs for RSV infection were 322.33 and 161.37 for THA and THG, respectively. Overall, period prevalence of ARTI in the total study population was 21.5% while period prevalence of viral ARTI was 8.25%. Period prevalence of RSV associated ARTI in THA was 4.25% and in THG was 4.7%. Overall, ARTI incidence in the total study population was 287/ 100,000 person years for children �5 years. In THA, RSV associated ARTI incidence was 28.3/ 100,000 person years for children �5 years. In THG, RSV associated ARTI incidence was 31.3/ 100,000 person years for children � 5 years. Age related incidence of ARTI in infants <12 months of age in THA and THG was 9.6/100000 and 9.3/100000 person years, respectively. Hospitalization following RSV infection in children aged <5 years in the rural population was significantly (p = 0.03) higher than that in the urban population. The risk of viral ARTI associated case fatality for THA and THG was 3.63% (6/165) and 1.81% (3/165 (Table 4 ). All deaths were associated with RSV associated pneumonia or bronchiolitis in mono-infection or coinfection with other respiratory viruses. Viral ARTI associated case fatality rate was 2.4% (4/ 165) and 1.2% (2/165), respectively for THA and THG in the presence of congenital cardiac disease. As with most parts of the world, ARTIs in children under 5 years is one of the most common causes of morbidity in Sri Lanka [31] . The distribution of respiratory viruses in Sri Lanka appears to be like those in developing countries and other tropical regions [32] . The age distribution and the clinical picture of respiratory viral infections were also like those previously described in tropical countries [33] . As noted in the other regions [32, 33] , the present study demonstrated an annual pattern for respiratory viral infections in Sri Lanka. Information on the causes of respiratory illness in tropical countries is limited compared to temperate countries. The available data indicates that about one third of the respiratory tract infections are caused by viruses based on antigen detection using IFA [34, 35] . The present study identified viruses in 37.1% ± 0.75 and 36.4% ± 0.12 of children hospitalized for ARTI in THA and THG, respectively. Compared to previous studies, antigen detection by IFA in our study shows a higher detection rate [36] [37] [38] [39] , and it may be due to higher detection indices of the current assay compared to those used in the previous studies. In the present study, RSV was the most predominant virus detected in both centers with an incidence of 29.8/100,000 person years. This RSV incidence of the present study is lower than that estimated in children based on hospital-surveillance done in Thailand (46/100,000 person-years in children) [40] and Kenya (293/100 000 person-years in children aged <5 years) [41] . In the USA and UK, it was 300 per 100,000 person-years in children aged<5 years [42, 43] . According to the current study, RSV associated ARTI incidence in Sri Lanka is somewhat like Thailand and it might be due to low prevalence of RSV infection in the Asian region compared to UK, USA and Kenya. Another reason for low prevalence of RSV infection in Sri Lanka might be due to low level of hospitalization of children following ARTIs as urban as well as rural areas of Sri Lanka has general practitioners. Conversely, with high degree of parental concern towards childhood illnesses, parents seek medical advice early in the illness. RSV predominance in Sri Lankan children agrees with the fact that this virus is the single most frequent respiratory pathogen detected in infants and young children worldwide. When considering the clinical characteristics of RSV associated ARTI, fever was less frequently noted in the present study. Rhonchi, rhinorrhea, chest recession and tachypnea were observed commonly. Overall, these signs and symptoms are related to the any known pathogen causing ARTIs and not characteristically confined to a specific virus as noted by many other studies as well [43] [44] [45] . Disease spectrum in RSV associated ARTIs ranged from common cold to life threatening severe bronchiolitis and pneumonia, suggesting the role of RSV in most of the cases with exacerbation of bronchiolitis [46, 47] . Moreover, the disease spectrum noted by the current study agrees with that noted in many studies done elsewhere [7, 48, 49] . RSV is associated with many viral co-infections and this finding is also in agreement with previous studies [50] [51] [52] . The association between the occurrence of co-infection and development of severe disease is unclear [53] , and the present study also noted morbidity and mortality in children with viral co-infections with RSV. Overall, the present study demonstrated an annual seasonality of viral ARTI in Sri Lanka in two different climatic zones especially for RSV infections. There is a marked seasonal variation in viral incidence between the temperate and tropical countries [54] . In tropical countries, the seasonality varies according to the temperature-dependent local weather pattern such as humidity or/and rainfall [17] . Epidemiological studies from tropical regions indicate that factors such as rainy seasons [55] or low air humidity [56] and temperature may influence outbreaks of viral ARTIs. Global warming has a profound influence in the earth's climate and changes in the climate has an impact on the seasonality and burden of respiratory viral infections. In addition to climatic factors, population density and degree of exposure to risk factors predisposing to the acquisition of ARTI in childhood also have a role in the overall burden of viral ARTI [57] [58] [59] . RSV infection was detected throughout the year in the current study. The seasonality correlated with the rainy season in both wet (South-west monsoon from May to September) and dry zone (North-east monsoon from December to March) and this explains the positive correlation with number of rainy days per month instead of monthly rain fall, which has a less bearing on the viral incidence [57, 60] . A slightly similar RSV seasonality from May to July has been reported by other studies conducted in the southern part of Sri Lanka, which experiences almost similar climatic conditions to the wet zone study center of the current study [61, 62] . The peak occurrence of RSV infections in the rainy season might be due to indoor overcrowding. At THG, the dominant interaction was detected between mean number of rainy days and atmospheric pressure (q = 0.74). At THA, the dominant interaction was detected between the mean atmospheric temperature and atmospheric pressure (q = 0.58). Moreover, in a study conducted in China [28], the dominant interaction was observed between the mean wind speed and the mean atmospheric pressure (q = 0�836). According to that study, a low atmospheric pressure would enhance the spread of airborne viruses. Our findings and the findings of the Xu et al. [28] suggest that these interactions might differ in different geographical areas and that the interpretations would differ accordingly. There have been a few studies on PIV infections in developing countries and most of these studies do not differentiate the subtypes of the virus. Our study recorded a 3.1% period prevalence of PIV infections with the predominance of PIV-3. Similar hospital-based studies have been reported in other developing countries [63, 64] . PIV infections have been strongly associated with croup based on the findings of some previous studies [65, 66] . However, the present study recorded a smaller number of cases with croup among PIV infected patients. In the THA study sample, an absence of a well demarcated seasonality was seen for PIV associated ARTI. In the THG study sample, PIV associated ARTI was scattered in the second quarter of the year. This distribution cannot be explained using climatic or other risk factors. Furthermore, the number of PIV cases was also less in the present study and thus extracting a clear association is difficult. AdV was the third highest prevalent virus causing ARTI in the study sample. The period prevalence for AdV associated ARTI was 5.66% and this was similar in both study centers. In the tropics, AdV has been reported to be responsible for 5-10% of the ARTI affecting infants [65, [67] [68] [69] . In our study, AdV was evenly distributed in different age groups of children <5 years. Furthermore, diarrhea and conjunctivitis were significantly associated with AdV infection. However, AdV infections showed a year-round distribution in the wet zone including the rainy season with no seasonality. Period prevalence for Inf-A was 3.66 and 2.66 and Inf-B was 2.66 and 1.66 in THA and THG, respectively. Inf-V infections were associated with a smaller number of morbidity and mortality and this may be due to scattering of antigenically stable Inf-A and Inf-B strains during the inter-epidemic period. In 2015, an influenza epidemic was reported in Sri Lanka with significant mortality and morbidity, however, no obvious seasonality was noted with these outbreaks either [70] , supporting the findings of the present study. hMPV infection is more prevalent in countries with temperate climates. We detected children with hMPV associated ARTI in both study centers. This is one of the large-scale studies detecting hMPV infection in Sri Lanka. Co-infections with RSV and hMPV were also detected. The disease spectrum and the severity of hMPV infection are like that of RSV associated ARTI [71] [72] [73] . Although the number of cases identified with hMPV associated ARTI were less, hMPV associated ARTI peak overlapped the RSV peak and thus the same climatic conditions might influence the distribution of RSV and hMPV. Taken study findings together, identifying the pattern of distribution of viral ARTI in a location will help implementing the early preventive measures, which can be communicated to the public with awareness on RSV infection including the RSV dominant times in the forth coming year and in rainy seasons. The RSV burden is high and present throughout the year and the Inf-V related morbidity is less and confined only to the epidemic period with low prevalence during the inter-epidemic period, however, introducing influenza vaccination will help to prevent Inf-V associated ARTI in high risk children. Now no specific evidence based preventive measures are practiced in Sri Lanka to control childhood ARTI. The current data on RSV burden and the pattern of distribution will contribute to health promotion activities in Sri Lanka among parents, guardians and childcare providers. Considering the viral burden, DALYs is less compared to other studies done in the tropics [74, 75] . DALYs is a measure of mortality and associated burden following viral ARTI. DALYs for ARTI in THA and THG were 429.77 and 242.056, respectively. DALYs for RSV infection was 322.33 and 161.37 for THA and THG, respectively. These DALYs values of the current study are significantly lower when compared to DALYs associated with ARTI in South-East Asian region [76] . According to the WHO, Sri Lanka lies in <25 deaths/1000 live births for under 5 years of age [77] and this reflects the low mortality rate and DALY associated with ARTIs in the island. On the other hand, it may also be due to high level of parental care, including seeking early medical advice for childhood illnesses leading to lesser chances for worsening of the illness. Although specific antiviral treatment is not available for many viral ARTI, symptomatic management eases the severity and the duration of the illness. Malnutrition has generally been considered as a risk factor for viral ARTI in developing countries [78, 79] , however, the increased mortality could occur especially when bacterial infections are involved during post-viral convalescence. In our study, malnutrition was identified as a potential risk factor for acquiring PIV-2 and AdV infections. In addition to tested nutrition indicators, the low level of Hb% reflected the level of malnutrition and this finding has been shown by many other studies as well [80] . All patients with anemia were screened for a complete blood evaluation (CBE) and found to have iron deficiency anemia. Viral infections further deprive the nutritional state by reducing the appetite in a child with ARTI, leading to increased risk for recurrent viral infections [81] . Viral testing on respiratory secretions like NPA is the accurate way of identifying the viral etiology and lack of viral diagnostics will raise the use of antimicrobials irrationally. The present study also noted the empiric use of antibiotics in children with ARTI in accordance with the pediatric emergency treatment protocol for managing moderate to severe bronchiolitis. In these instances, antibiotic treatment does not contribute positively to clinical improvement as noted by previous studies [80, 81] . On the other hand, use of antibiotics in children with viral ARTI contributes to an enormous health cost in developing economies like Sri Lanka. Moreover, rational use of antibiotics will contribute to minimize the development of antimicrobial resistance. Health care policy makers must consider reviewing the clinical algorithm for patient management with virologists and clinicians considering the cost incurred by the irrational use of antibiotics for viral infections and this expenditure can be diverted to establish viral diagnostics. Future cross-sectional and longitudinal studies of viral ARTI in out-patients and communities will contribute to identify the time bound trends on the respiratory viral burden. This will alarm the clinicians about common, uncommon and newly emerging respiratory viruses. This study also shows the number of viral ARTI associated hospitalization in two climatic zones of Sri Lanka. Overall, period prevalence of ARTI in both centers was 21.5% while the period prevalence of virus identified ARTI was 8.25%. Overall, ARTI incidence in the study sample was 287/ 100,000 person years for � 5-year-old children. In addition to the morbidity, the management cost and the loss of parent/guardian manpower used for the childhood ARTI have economic consequences. This is the first large study to detect hMPV associated ARTI in Sri Lankan children. Our data shows the respiratory viral burden on childhood ARTI and this will help the health care providers and policy makers to re-evaluate the existing infection control and prevention strategies for viral ARTI. Awareness on local respiratory viral dynamics is important to implement early preventive measures like use of respiratory precautions and health education for different categories of care providers including parents. Prevention plays a significant role in reducing the burden of childhood viral ARTIs, however, adherence to preventive methods must be promoted by the health care providers and the policy makers. Recruiting only the hospitalized children with SARI led to miss a fair number of patients who reported to the outpatient department with viral ARTIs. Moreover, Sri Lanka has a well spread network of general practitioners and parents seek paid medical care following ARTI and this would have also reduced hospitalizations leading to drop in the actual number of ARTI and this would under-estimate the true burden of viral ARTI. This study was confined to children between 1 month to 5 years and thus missed the pediatric ARTI cases from >5 to <12 years. Since detection sensitivity of IFA for viral antigen is low compared to molecular methods, the study would have missed some more true viral ARTI cases. We have also not tested for rhinoviruses and enteroviruses which contribute to >10% of all viral ARTIs. In addition to detecting hMPV using IFA, only 100 RSV positive and 100 virus negative NPA were further tested using PCR for hMPV, hBoV, and CoV. Overall, 38.33% of the hospitalized children had virus identified ARTI based on antigen detection by IFA. The level of viral detection was similar in both study samples representing dry and wet zones of Sri Lanka. RSV was the commonly detected respiratory virus with an annual pattern of distribution and RSV associated ARTI was distributed during the rainy seasons. In the wet zone, RSV associated ARTI was seen around May to July along with the South-West monsoon. In the dry zone, RSV associated ARTI was seen around December to January along with the North-East monsoon. Furthermore, monthly RSV cases negatively correlated with mean monthly atmospheric temperature and mean monthly relative humidity, and positively correlated with the number of rainy days present in a month instead of the mean rain fall. Other respiratory viruses tested in the study have not shown such a correlation. Moreover, this study has demonstrated that respiratory viruses are associated with a considerable number of hospitalizations in Sri Lanka. Future cross-sectional and longitudinal studies of viral-ARTI including out-patients will contribute to assess the true burden and the time bound diversity of viral ARTI in Sri Lanka. Upper respiratory tract infection Manual of clinical microbiology Prevalence of respiratory virus in symptomatic children in private physician office settings in five communities of the state of Veracruz, Mexico Identification of new respiratory viruses in the new millennium. Viruses Microbial interactions in the respiratory tract Antibiotic selection pressure and resistance in Streptococcus pneumoniae and Streptococcus pyogenes Respiratory viral infections in infants: causes, clinical symptoms, virology and immunology Risk factors in children hospitalized with RSV bronchiolitis versus non-RSV bronchiolitis Respiratory syncytial virus infection in high-risk infants-an update on palivizumab prophylaxis Lung infection-a public health priority World Health Organization. Burden of disease project Antimicrobial resistance: risk associated with antibiotic overuse and initiatives to reduce the problem Reducing antibiotic use in influenza: challenges and rewards The current epidemiology and clinical decisions surrounding acute respiratory infections How contagious are common respiratory tract pathogens? Respiratory viruses in the pediatric intensive care unit: prevalence and clinical aspects Contribution of RSV to bronchiolitis and pneumonia-associated hospitalizations in English children Epidemiology of respiratory syncytial virus infection in northern Taiwan Global epidemiological surveillance standards for influenza. World Health Organization Diseases of the respiratory system. ICD codes. Available at D3 Ultra DFA Respiratory virus screening and ID Kit (USA) Viral pathogens of acute lower respiratory tract infection in China Use of monoclonal antibodies for rapid diagnosis of respiratory viruses in a community hospital Lack of sensitivity of rapid antigen tests for the diagnosis of respiratory syncytial virus infection in adults Cluture vs direct antigen assays for detection of microbial pathogens from lower respiratory tract specimens suspected of containing the respiratory syncytial virus The burden of hospitalized lower respiratory tract infection due to respiratory syncytial virus in rural Thailand Incidence and severity of respiratory syncytial virus pneumonia in rural Kenyan children identified through hospital surveillance Population-based surveillance for hospitalizations associated with respiratory syncytial virus, influenza virus, and parainfluenza viruses among young children Importance of global surveillance for respiratory syncytial virus Lower respiratory tract infections due to adenovirus in hospitalized Korean children: epidemiology, clinical features, and prognosis Epidemiology of respiratory viral infection using multiplex RT-PCR in The role of respiratory virus infections in childhood asthma inception The WHO child health epidemiology reference group. WHO estimates of the causes of death in children Indoor air pollution in developing countries and acute respiratory infection in children Does viral co-infection influence the severity of acute respiratory infection in children? RSV, adenovirus most common viral co-infections among young children Respiratory pathogens in children with and without respiratory symptoms Respiratory syncytial virus and other viral infections among children under two years old in Southern Vietnam 2009-2010: Clinical characteristics and disease severity How contagious are common respiratory tract pathogens? Seasonality of viral respiratory infections in Southeast of Brazil: the influence of temperature and air humidity Seasonal trends of viral respiratory tract infections in the tropics Climate change impacts in the United States: The Third National Climate Assessment Contribution of working groups i, ii and iii to the fourth fifth assessment report of the intergovernmental panel on climate change Incidence of bacterial coinfection with respiratory syncytial virus bronchopulmonary infection in pediatric inpatients Respiratory syncytial virus and parainfluenza virus Burden and seasonality of viral acute respiratory tract infections among outpatients in Southern Sri Lanka' Respiratory viral infection: An underappreciated cause of acute febrile illness admissions in Southern Sri Lanka' Parainfluenza viruses Diagnosis of childhood pneumonia in the tropics Association of invasive pneumococcal disease with season, atmospheric conditions, air pollution, and the isolation of respiratory viruses Epidemiology and clinical presentation of respiratory syncytial virus infection in a rural area of southern Mozambique Clinical presentation and severity of viral community-acquired pneumonia in young Nepalese children Molecular epidemiology of human adenovirus isolated from children hospitalized with acute respiratory infection in Sao Paulo A newly discovered human pneumovirus isolated from young children with respiratory tract disease Human metapneumovirus associated pneumonia and severe bronchiolitis in a 9-month old infant admitted to a Sri Lankan hospital Viral burden of acute respiratory tract infections in hospitalized children in wet and dry zones of Sri Lanka Disability adjusted life years. Health statistics and information systems A case series on common cold to severe bronchiolitis and pneumonia in children following human metapneumovirus infection in Sri Lanka Levels and trends in child mortality Burden of disease in DALYs by sex and mortality stratum in WHO regions, estimates for Viral etiology and epidemiology of acute respiratory infections in children in Etiology of acute lower respiratory tract infections in Gambian children: I. Acute lower respiratory tract infections in infants presenting at the hospital Micronutrients in childhood nutrition. Sri Lanka J Child Health Acute respiratory infections. Last updated in Risk of secondary bacterial infection in infants hospitalized with respiratory syncytial viral infection Better tests, better care: improved diagnostics for infectious diseases We would like to acknowledge the supportive staff at Teaching hospitals Anuradhapura and Gampola, Sri Lanka for the support given for the sample collection.