Viral aetiology of bronchiolitis in hospitalised children in Qatar

Background Bronchiolitis is considered one of the earliest and most common causes of hospitalisation in young children. Development of molecular technologies allowed a better understanding of bronchiolitis aetiology. Results from cohort studies evaluating the association between single, multiple viral infections and clinical outcomes are conflicting. Data on viral bronchiolitis in children were found to be limited in Qatar. This study aimed to determine frequency and seasonal trends of viral pathogens causing acute bronchiolitis, and to explore association between viral pathogens, disease severity and length of stay (LOS). Methods This is a retrospective descriptive study, including children admitted in 2010 and 2011 with acute bronchiolitis. Presenting history, physical examination and respiratory viral co-infections as detected by molecular assays were analysed. Results At least one virus was detected in 315/369 (85.4%) of included children with single and multiple viruses in 67 and 33% of cases respectively. Respiratory syncytial virus (RSV) was the most detected virus, accounting for 51.2% followed by rhinovirus (RV) in 25.5% of cases. Fall and summer admissions were associated with longer LOS. On multivariate logistic regression analysis, retraction (OR 3.96; 95% CI 1.64,9.59) and age group 1–3 months (OR 3.09; 95% CI 1.06,9.05) were associated with longer LOS. Crepitation (OR 9.15; 95% CI 1.58,53.13), retraction (OR 4.10; 95% CI 1.05,16.12) and respiratory rate (OR 1.46; 95% CI 1.28,1.66) were associated with moderate to severe bronchiolitis. Identifying the viral agent did not influence disease severity or LOS. Conclusion Clinical presentation is of more relevance to LOS and disease severity than the detected viruses. Future studies should investigate the interplay between climate characteristics, population’s factors and the most detectable circulating viruses. Electronic supplementary material The online version of this article (doi:10.1186/s12879-017-2225-z) contains supplementary material, which is available to authorized users.


Background
Bronchiolitis is considered one of the earliest and most common causes of hospitalisation among young children during their first 2 years of life [1]. Although the causative agents include bacteria, fungi and viruses, this acute infection is mainly attributed to respiratory syncytial virus, accounting for 50-90% of the cases [2,3]. With the current utilisation of reverse transcriptase real-time polymerase chain reaction (PCR), the detection of specific viral nucleic acids facilitated a better understanding of the viral aetiology of the infection. The 2014 American Academy of Paediatrics bronchiolitis guideline recommends against the routine use of radiographic or laboratory studies on the basis that knowing the infecting pathogen would rarely alter the clinical management [4]. However, a growing body of literature has identified an association between specific infecting pathogens with short and long-term outcomes [5]. Association between viral co-infection and disease severity have been assessed in several studies but with conflicting findings [5][6][7]. Although bronchiolitis is a self-limiting condition, hospitalisation rate has increased during the last two decades [3]. Assessment of bronchiolitis severity, through a combination of clinical symptoms and physical signs, remains a standard measure in daily practice though its impact on clinical outcomes, such as length of hospital stay, has yet to be confirmed. Data has been reported on seasonal variation of viral activity with conflicting evidence on its significance on disease severity and clinical outcomes [8,9]. In Qatar, viral aetiology of bronchiolitis in children has been limited. Therefore the aims of this retrospective, descriptive study were: to determine the frequency and seasonal trends of viral pathogens causing acute bronchiolitis, and to explore the association between specific viral pathogens, disease severity and length of stay (LOS).

Study design
This retrospective descriptive study took place at the paediatric wards of Hamad General Hospital (HGH), a 603-bed, tertiary-care facility in Doha, for two consecutive years, 2010 and 2011. The institutional review board at Hamad Medical Corporation approved the study, IRB number 12054.

Study population
The study population included children aged between 2 weeks-2 years admitted to the paediatric ward at HGH with a diagnosis of acute bronchiolitis defined according to International Classification of Disease (ICD) code 466.1 in the 9 th revision of the ICD. Based on recorded clinical diagnosis, an episode of acute bronchiolitis was determined by a constellation of clinical signs and symptoms including fever, rhinitis, tachypnoea, cough, wheezing, crackles, use of accessory muscles and possible chest X ray findings of hyperinflation of the lungs, peri-bronchial thickening, collapsed segment or a lobe of the lung and increase interstitial markings. Due to the potential for an atypical natural history of bronchiolitis, children with the following diagnoses were excluded: born prematurely, defined as <37 weeks of gestation, those with low birth weight, defined as birth weight <2.5 kg, with underlying chronic lung disease or neurological disease or congenital heart disease, and those immunocompromised or with hospital-acquired bronchiolitis.

Data collection
Medical charts of the children admitted to the paediatric ward during the study period were retrieved from the medical records department at HGH, a tertiary teaching hospital located in Doha, Qatar. A trained research staff screened patients' discharge summaries for diagnosis of bronchiolitis. A data collection form (DCF) was developed to capture the necessary information. For every child who met the inclusion criteria the following data were collected: age, gender, family history of asthma and/or atopy, clinical presentation and vital signs upon admission and length of hospital stay. Bronchiolitis severity was based on a score ranging between 1 and 5 that assessed wheezing, retraction, respiratory rate and oxygen saturation (see Additional file 1: Table S1). Patients with score of 1 or 2 were considered to have mild bronchiolitis and to have moderate to severe bronchiolitis if score was between 3 and 4, with 5 indicating respiratory failure. Wheezing on physical exam was categorised as none, mild, moderate or severe. Retraction was categorised as none, intercostal, intercostal and subcostal or intercostal and subcostal with seesaw chest motion. Viral identification laboratory results, medical therapy and need of mechanical ventilation during hospital stay were recorded. Hospital admissions due to bronchiolitis were presented on an episode base.

Nasopharyngeal sample collection and analysis
Nasopharyngeal samples were collected by inserting Dacron™ polyester tipped or FLOQSwab™ in both nostrils till the nasopharynx and rotating at 360°. The swabs were placed into 3 ml of transport medium (Copan, UTM-RT™, CA, USA). Following collection, samples were kept on wet ice and transported to central processing lab of HGH during the working hours. Specimens that were not delivered immediately to the lab were stored in a refrigerator (4 -8°C) for no more than 24 h. Beyond that, samples were stored at -70°C and transported on dry ice. EZ1 virus mini kit v 2.0, was used for nucleic acid extraction (Qiagen®, Hilden, Germany). Multiplex, real-time, polymerase chain reaction (RT-PCR) using FTDResp21 kit (Fast Track Diagnostics, Silema, Malta) was used for the detection of respiratory pathogens on Applied Biosystem™ 7500 instrument (ThermoFisher Scientific Inc, MA, USA). The kit detects the following respiratory pathogens: influenza A and B; coronaviruses NL63, 229E, OC43 and HKU1; parainfluenza viruses 1, 2, 3 and 4; human metapneumovirus A/B (hMPV); rhinovirus (RV); respiratory syncytial virus A/B (RSV); adenovirus (AdV); enterovirus (EV); parechovirus (PeV); bocavirus (BoV) and Mycoplasma pneumoniae (Mpneu).

Statistical analysis
The focus of the data analysis in our study was to determine frequency of each virus that is causing acute bronchiolitis in young children in our community and link it to the clinical presentation, length of hospital stay, severity of symptoms and need for intensive care support including invasive and non invasive ventilation. Categorical and continuous values were expressed as frequency (percentage) and mean ± SD or median and interquartile range (IQR) as appropriate. Descriptive statistics were used to summarise demographic, medical history, frequency of each virus and clinical characteristics of the patients. Associations between two or more qualitative variables were assessed using chi-square test and Fisher Exact test as appropriate. Quantitative data between the two independent groups were analysed using unpaired 't' and Mann Whitney U tests as appropriate.
Univariate and multivariate logistic regression methods were used to assess and compute the predictive values of each predictor or risk factors for severity of bronchiolitis and hospital length of stay. For multivariate regression models, variables were considered if statistically significant at the P <0.10 level in univariate analysis or if determined to be clinically important. The results of logistic regression analyses were reported as odds ratio (OR) with 95% confidence intervals (CIs). Pictorial presentations of the key results were made using appropriate statistical graphs. A two-sided P value <0.05 was considered to be statistically significant. All statistical analyses were done using statistical packages SPSS 22.0 (SPSS Inc. Chicago, IL) and Epi Info 2000 (Centres for Disease Control and Prevention, Atlanta, GA).

Patient identification and demographics
A total of 369 children were admitted to the hospital with a clinical diagnosis of bronchiolitis documented in their medical charts during the study period with 261 (75.2%) aged 3 months or younger and 247 (67%) were male sex as illustrated in Table 1. Of the 369 cases, 145 (39.3%) had a family history of atopy or asthma. On Quantitative data were presented in mean ± SD. For some parameters the sum is not equal to total n = 369 due to some missing observations and respective % were computed based on non-missing observations

Seasonal distribution
The seasonal pattern of RSV and RV were generally distributed all year around although incidence of RSV infection peaked during the fall season whereas RV peaked during the spring. The findings of the current study revealed more adenovirus (ADV) infections in the summer months than spring (Fig. 1). Coronavirus (CoV-OC43) circulated predominantly during the winter months while parainfluenza virus 3 (PIV3) during the fall but sporadically across the year. Dual peaks were observed for human metapneumovirus (hMPV) during the spring and winter seasons.
The median LOS was longer during the summer (7 days) compared to the winter (6 days) and shortest in the fall and spring (5 days), Fig. 2a. Our findings show a peak in hospital admission rates during the winter season, 154 (41.7%), followed by the fall, 138 (37.4%) with least during the summer season, 20 (5.4%) as presented in Fig. 2b.

Viral aetiologies and multiple co-infections
We used polymerase chain reaction (PCR) to establish the viral aetiology of acute bronchiolitis. At least one virus was detected in 315/369 (85.4%) of children with single virus in 211/315 (67%) and multiple viruses in 104/315 (33.0%). RSV was the most common virus identified, accounting for 189 cases (51.2%) followed by RV in 94 cases (25.5%), 23 cases (6.2%) of hMPV and ADV each. Influenza A and B were the least detected viruses accounting for less than 1% of the cases, as shown in Table 1.
Out of the 315 samples analysed, 27.3% had dual infections and three viruses were detected in 5.4% of the sample with a quadruple infection found in one case. There were no marked differences in the management of single and multiple infections. β-agonists were used in the majority of cases, followed by antibiotics, hypertonic saline and adrenaline, presented in Table 2.
Respiratory viruses distribution by age-group, gender, medical history and clinical presentation Children were grouped into three age groups with different rates of single and multiple viral infections, Table 3. In children aged <1 month, the most common detected virus was RSV (47.9%); for those between 1 and 3 months of age, RV (47.2%) was the most common virus. Children aged >3 months were found to be co-infected with RSV + RV (40%). Co-infection with RSV+ non-RV (75.8%) had the highest incidence in males whereas in females, RV+ non-RSV co-infection (40.9%) dominated. Family history of atopy or asthma, P = 0.008, and high respiratory rate, P = 0.026, were more common in RV single group, whereas crepitation was more frequent in the RSV+ non-RV co-infection group, P = 0.039. More children in the RV single group (91.7%) had LOS ≥4 days compared to the RSV single group (82.4%). Similar pattern was observed in terms of bronchiolitis severity with more children in the RV single group (65.7%) with moderate to severe bronchiolitis compared to children in the RSV single group (57.6%).   Table 5. On multivariate logistic regression analysis, crepitation (adjusted OR 9.15; 95% CI 1.58,53.13; P = 0.014), retraction (OR 4.10; 95% CI 1.05,16.12; P = 0.043), and increased respiratory rate (adjusted OR 1.46; 95% CI 1.28,1.66; P < 0.001) were significant predictors associated with bronchiolitis severity, provided in Additional file 1: Table S3. Complications including atelectasis, collapse, pneumonia and sepsis were

Discussion
To our knowledge this is the first comprehensive report on viral aetiology in hospitalised children with bronchiolitis in the State of Qatar. Our study revealed that RSV and RV were the most commonly detected causative agents. This finding is in contrast to a previous report with a small sample size of 56 children with respiratory tract infection (RTI) in Qatar in which hMPV was the most frequently isolated virus [10]. Our findings are, however, in line with a number of large studies conducted in the Middle East and elsewhere [11][12][13][14]. In Jordanian children of less than 2 years of age, RSV was a major cause of hospitalisation [11]. Other studies conducted in Saudi Arabia and Turkey reported the same [15,16]. Geographic variations in seasonal activity of the most commonly detected respiratory viruses have been reported in the literature. In temperate countries, RSV peaks during the winter season, but with variation in the tropics. This diversity is mainly attributed to the region's climate characteristics, i.e. relative humidity, average monthly temperature and rainfall [17][18][19]. In their study, Haynes et al. assessed RSV seasonality in seven countries with diverse climate characteristics and found that RSV onset and peak timings were inconsistent with climate characteristics in all assessed countries apart from Thailand. RSV offset was found to be consistent in Guatemala [20]. Rhinovirus A was detected all year around with high proportion by the onset of summer in Japan but during the fall and spring in Cyprus while Rhinovirus C was mainly detected in the winter but during the fall and spring in Japan and Cyprus, respectively [18,21]. In this study, RSV exhibited a strong seasonal pattern with peak activity during the fall and summer months, in contrast to RV, which showed minimal circulation during the fall.
In comparison to children with RSV only infections, children infected with RSV in combination with non-RV or RV alone were more likely to have longer LOS. However, we found no significant association between single or multiple infections and disease severity or LOS. The findings are consistent with two large studies conducted in North America and Europe [13,14], but in contrast with previous reports where RV was associated with shorter LOS [12,22]. These discrepancies might be related to sample size, study design and variation in host risk factors.
In this study, we noted an association between individual components of the clinical severity score and bronchiolitis outcome after adjusting for clinical and demographic factors. Similar to Ricart et al., McCallum et al. and Weisgerber et al., the data indicate that, in the Qatari community, children's clinical characteristics are more relevant than the specific infecting viruses in determining the duration of LOS and disease severity [23][24][25].
Detection of the presence of multiple infections has become common in practice although its clinical impact on disease management and predicting outcomes remains unclear [26]. By using RT-PCR to quantify viral load, a number of studies have concluded that RSV genomic load is correlated with disease severity [13,14]. Similarly, two investigational anti-RSV drugs showed promising results in reducing viral load and clinical disease severity [27,28].
Some studies have determined severity based on a modified Wood-Downes score or the need for paediatric intensive care unit (PICU) admission [7,23]. In this study we analysed severity based on a clinical score that allows children's classification according to respiratory rate, oxygen saturation and physical examination, mirroring the severity of the respiratory burden.
Previous studies have reported males to be more susceptible than females to lower RTIs across different age groups [29]. In this study the percent of males with severe bronchiolitis was as twice as that to females, 70.8 vs. 29.2%. An explanation was suggested two decades ago by Gupta et al. in which the disproportionally narrower peripheral airways in the younger males could be a possible mechanism of the observed difference [30].
Childhood respiratory infection may influence the development of chronic respiratory diseases. In a cohort of 47 Swedish hospitalised infants with RSVbronchiolitis, asthma was prevalent in 30% of the group when aged 7 years increasing to 37% at the age of 13 [31]. While Jackson et al. noted an increase in asthma risk in children with RSV associated wheezing; RV was associated with a 10-fold increase of asthma [32]. An earlier retrospective study from Qatar noted an increase in recurrent wheezing by 44% in hospitalised infants due to RSV bronchiolitis 2 years after admission compared to 12% in the control group [33]. In the current study, 63% of the admitted children had wheezing of which 51.2% had RSV associated bronchiolitis. Compared to RSV, RV associated bronchiolitis is suggested to occur in older infants, 13 versus 5 months, who present more often with atopic dermatitis and blood eosinophilia [34].
This study has potential limitations. First, it is a retrospective study at a single hospital. Another limitation is that data on concomitant bacterial infection, viral genomic load and palivizumab prophylaxis were not collected.

Conclusions
In summary, RSV was the most common virus identified, followed by RV with peaks during the fall and spring seasons respectively. Our data show that the clinical presentation is more related to the duration of hospital stay and disease severity than the detected viruses.
Conducting a prospective, multi site surveillance of viral bronchiolitis in the warm, desert climate of the Gulf Cooperation Council (GCC) countries will allow for proper timing of preventable measures. Future studies within the GCC countries should investigate the interplay between climate characteristics, population's factors and the most detectable circulating viruses.

Additional file
Additional file 1: Table S1. Bronchiolitis severity score; Table S2. Predictors of hospital length-of-stay ≥ 4 days among children with mild to severe bronchiolitis associated with different viruses: multivariate logistic regression analysis; Table S3. Predictors of moderate (1 to 2) to severe (3 to 4) ICS score among children with viral aetiology of acute bronchiolitis: multivariate logistic regression analysis.