|
Epidemiology of human respiratory viruses in children with acute respiratory tract infection in a 3-year hospital-based survey in Northern Italy ☆
Abstract
Acute respiratory tract infections (ARTIs) are among the leading causes of morbidity and mortality in children. The viral etiology of ARTIs was investigated over 3 years (October 2012-September 2015 in 2575 children in Parma, Italy, using indirect immunofluorescent staining of respiratory samples for viral antigens, cell culture, and molecular assays. Respiratory viruses were detected in 1299 cases (50.44%); 1037 (79.83%) were single infections and 262 (20.17%) mixed infections. The highest infection incidence was in children aged N6 months to ≤3 years (57.36%). Human respiratory syncytial virus (27.12%) and human adenovirus (23.58%) were the most common viruses identified. The virus detection rate decreased significantly between the first and third epidemic season (53.9% vs. 43.05%, P b 0.0001). The simultaneous use of different diagnostic tools allowed us to identify a putative viral etiology in half the children examined and to provide an estimate of the epidemiology and seasonality of respiratory viruses associated with ARTIs.
Acute respiratory tract infections (ARTIs) are a persistent public health problem (Lu et al., 2013) .
ARTIs affect the upper respiratory tract, leading to colds, rhinosinusitis, pharyngitis, laryngitis, tracheitis, and otitis media, as well as the lower respiratory tract, causing tracheitis, bronchitis, bronchiolitis, and pneumonia (Bicer et al., 2013) . Although the majority of ARTIs remain confined to the upper tract (Tregoning and Schwarze, 2010) , they can cause severe manifestations when affecting the lower tract (Zappa et al., 2008) .
Viruses are frequently the cause of ARTIs characterized by high risks of morbidity and mortality (Taylor et al., 2017) . The viruses most frequently detected are: influenza viruses A and B (IAV, IBV), human parainfluenza virus (HPIV), human adenovirus (HADV), human metapneumovirus (HMPV), and human respiratory syncytial virus (HRSV). This list is constantly evolving along with improvement in the performance of diagnostic tests (Ljubin-Sternak et al., 2016; Tregoning and Schwarze, 2010) and the discovery of more recent viruses, for example, human bocavirus (HBOV) (Allander et al., 2005) , human coronaviruses (HCOVs) NL63 and HKU1, and new human enterovirus (HEV), parechovirus, and rhinovirus strains (Berry et al., 2015) .
The main diagnostic methods for respiratory virus infections are: virus isolation in cell culture, viral antigen/nucleic acid detection, and virus-specific serology. Virus isolation is labor-intensive (Landry, 1997) , making it impractical to determine a diagnosis during the acute infection phase. Shell vial cultures and viral antigen detection reduce the time for diagnosis (Engler and Preuss, 1997; Gardner and McQuillin, 1974) . Nearly all these methods are being replaced by highly sensitive multiplex PCR (Kim et al., 2009; Tregoning and Schwarze, 2010) .
ARTI prevalence in children varies by geographic area, season, and year (Goktas and Sirin, 2016; Moura et al., 2009; Panda et al., 2017) . Recent Italian data are limited, apart from those relating to influenza (Pariani et al., 2015; Trucchi et al., 2017) .
This three-year (October 2012-September 2015) hospital-based survey in Parma (Northern Italy) aimed to determine the prevalence of respiratory virus infections, their seasonality, and any patterns of mixed infections in children with ARTIs by using indirect immunofluorescent staining of respiratory samples for viral antigens, cell culture, and molecular assays.
The study was conducted at the Virology Unit of the University Hospital of Parma (Northern Italy), a 1300-bed tertiary care center with more than 50,000 admissions per year from the city and surroundings with approximately 450,000 inhabitants. Laboratory diagnosis was performed upon medical request. Patients' identities and medical information were protected.
The inclusion criteria were: patients ≤14 years old, acute fever, respiratory symptoms, and illness onset within 2 days.
From October 1st, 2012 to September 30th, 2015, 2892 samples from 2575 children (1148 females, 44.58%, and 1427 males, 55.42%; 2311 inpatients, 89.75%, and 264 outpatients, 10.25%) were collected by bronchoalveolar lavages (BALs), bronchial aspirates (BASs), nasopharyngeal aspirates (NPAs), nasal swabs (NSs), sputum (SP), and throat swabs (TSs). NPAs, NSs and TSs were stored in viral transport medium (De Conto et al., 2018) and all samples were kept at 4°C until submitted to laboratory (within 2 h of sampling).
Age was available for 2570 children (99.8%) and ranged from ≥2 days to 14 years (mean age: 3 years 6 months; median age: 2 years 1 month). Children were divided into four age groups: 0 to ≤6 months (21.98%), N6 months to ≤3 years (37.78%), N3 to ≤6 years (18.95%), and N 6 to 14 years (21.28%). Of note, 236 children (9.16%) were examined at least twice during the survey, with a time interval between care provider visits of 3 days to 8 months, presumably for either worsening symptoms of a current infection or infection recurrence.
NPAs, NSs, and TSs were centrifuged (1000 rpm, 10 min, 4°C) and the pellets resuspended in phosphate-buffered saline (PBS, pH 7.4; 7 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 , 137 mM NaCl, 2.7 mM KCl), before cytocentrifugation (1000 rpm, 5 min). Cell sediments on slides were dried in a laminar-flow hood and fixed in acetone (10 min, −20°C). These were then hydrated with PBS, blocked with 1% bovine serum albumin (BSA) in PBS, and the slides were incubated (1 h, 37°C) with anti-HRSV monoclonal antibodies (mAbs) (1:40; Argene/BioMérieux, France) in 0.2% BSA in PBS. After washes with PBS, they were incubated with fluorescein isothiocyanate-conjugated anti-mAbs (1:200; Argene/ BioMérieux) in 0.0001% Evans Blue in PBS (45 min, 37°C). After further washes with PBS, the sediments were mounted in glycerol and analyzed by epifluorescence microscopy (DMLB Leica, Germany).
Madin-Darby Canine Kidney cells with 6-linked sialic acids enhanced expression (MDCK-SIAT1) came from Sigma-Aldrich (Italy), with human lung fibroblasts (MRC-5) from LGC Standards (Italy). Human larynx epidermoid carcinoma (HEP-2), embryonic human intestine (Intestine 407), rhesus monkey kidney (LLC-MK2), and African green monkey kidney (Vero) cells came from the "Istituto Zooprofilattico Sperimentale della Lombardia e dell'Emilia-Romagna" (Italy). Cells were grown in Earle's Modified Minimum Essential Medium (E-MEM) supplemented with 2 mM L-glutamine, 10% fetal bovine serum, and antibiotics (100 U/ml penicillin, 100 μg/ml streptomycin), and maintained at 37°C in a 5% CO 2 atmosphere. Reagents were from Life Technologies (Italy).
After aspiration of the growth medium, MDCK-SIAT1, HEP-2, and Intestine 407 cells grown in 8-well chamber slides (Fisher Scientific, Italy) were incubated with samples (1 h, 37°C). After restoring the E-MEM, incubation was carried out in E-MEM for 24 h before IIF with mAbs against HADV, HPIV, HRSV, and influenza virus (IV) (Argene/ BioMérieux), as described above.
HADV, HEV, HPIV, HRSV, and IV isolation was performed in LLC-MK2, MDCK-SIAT1, HEP-2, Intestine 407, MRC-5, and Vero cells grown in 24-well plates (Sigma-Aldrich). After aspirating the E-MEM, cells were incubated with samples (1 h, 37°C). After restoring the E-MEM, incubation in E-MEM was continued until the appearance of cytopathic effect (CPE). If the cells showed no CPE after 7 days, they were scraped, and cell suspension was collected for a second inoculation. When CPE was observed, cells were scraped in PBS and cytocentrifuged, before IIF. When CPE suggested HEV infection, neutralization assays were performed.
HEVs serotype identification was performed with type-specific antisera (Statens Serum Institut, Denmark), according to Lim and Benyesh-Melnick (1960) .
Samples were screened for HADV, HBOV, HCOV, HMPV, HPIV, HRSV, and IV nucleic acids with real-time (RT)-PCR or reverse transcription (rt)-RT-PCR assays. Nucleic acid extraction was performed with the NucliSENS® easyMAG® kit on the EasyMAG extractor (BioMérieux, Italy) and rt with the RT-kit plus (ELITechGroup Molecular Diagnostics, France) on the GeneAmp PCR System 9700 thermocycler (Applied Biosystems, Thermo Fischer Scientific-Waltham, USA). Extracted nucleic acids were tested with Influenza A/B Q-PCR Alert AmpliMIX kit (ELITechGroup) on the 7300 RT-PCR system (Applied Biosystems). The Respiratory Multi Well System (MWS) r-gene™ assay (Argene/ BioMérieux) was used to detect ADV/HBOV, HCOV/HPIV, and RSV/ HMPV on the Rotor-Gene Q system (Qiagen, Germany). Assays were performed according to the manufacturer's instructions.
A chi-square test was performed using GraphPad Prism software. P b 0.05 was considered statistically significant.
A total of 2892 respiratory secretions from 2575 children were analyzed as described in the Methods section; 1408 samples (48.69%) from 1299 (50.44%) children (714 males, 54.97%, and 585 females, 45.03%) were virus-positive (Table 1 and Supplementary Table 1) .
Single infections were detected in 1128 (80.11%) samples from 1037 (79.83%) children. Mixed infections were detected in 280 (19.89%) samples from 262 (20.17%) children. There was a significant difference in the frequency of single vs. mixed infections among positive samples (80.11% vs. 19.89%; P b 0.0001) as well as among total examined samples (39% vs. 9.68%; P b 0.0001). Among mixed infections, two viruses were observed in 252 (90.00%) samples, three viruses in 26 (9.29%) samples, and four viruses in 2 (0.71%) samples. The percentage difference in the number of mixed infections with two viruses (90.00%) and four viruses (0.71%) was significant (P b 0.0001).
Overall, 1718 viruses were detected. HRSV (466/1718: 27.12%) and HADV (405/1718: 23.58%) were the most common viruses identified, followed by HCOV (262/1718: 15.25%), IV (198/1718: 11.53%), HBOV (161/1718: 9.37%), HPIV (114/1718: 6.64%), HMPV (76/1718: 4.42%), and HEV (36/1718: 2.09%).
HRSV and HADV were found in single (30.14% and 23.23%, respectively) and mixed infections (126/280: 45% and 143/280: 51.07%, respectively) ( Table 1)
The virus detection rate decreased from 53.9% (463/859) 34 (12.14) 2 (0.71) 0 (0) 0 (0) 0 (0) 0 (0) 36 (12.85) HRSV + HCOV 13 (4.64) 22 (7.85) 0 (0) 0 (0) 0 (0) 0 (0) 35 (12.5) HADV + HBOV 20 (7.14) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 20 (7.14) HRSV + HBOV 11 (3.93) 5 (1.78) 0 (0) 0 (0) 0 (0) 0 (0) 16 (5.71) HRSV + IAV 6 (2.14) 3 (1.07) 0 (0) 0 (0) 1 (0.36) 0 (0) 10 (3.57) HCOV + IAV 7 (2.5) 3 (1.07) 0 (0) 0 (0) 0 (0) 0 (0) 10 (3.57) HADV + HPIV 8 (2.85) 1 (0.36) 0 (0) 0 (0) 0 (0) 0 (0) 9 (3.21) HADV + HMPV 6 (2.14) 2 (0.71) 0 (0) 0 (0) 0 (0) 0 (0) 8 (2.85) HCOV + HMPV 5 (1.78) 3 (1.07) 0 (0) 0 (0) 0 (0) 0 (0) 8 (2.85) HBOV + HCOV 6 (2.14) 1 (0.36) 0 (0) 0 (0) 0 (0) 0 (0) 7 (2.5) HBOV + IAV 6 (2.14) 1 (0.36) 0 (0) 0 (0) 0 (0) 0 (0) 7 (2.5) HRSV + IBV 7 (2.5) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 7 (2.
For the 1295 children with viral ARTIs and age available, the mean and median age were 2 years 9 months and 1 year 7 months, respectively.
The infection rate was highest in children between N6 months to ≤3 years (57.36%; 43.01% among positive cases), followed by the 0 to ≤6 months (58.05%; 25.33% among positive cases) and N 3 to ≤6 years (49.9%; 18.76% among positive cases) groups (Fig. 3A ).
Significant differences in the infection rate were evidenced among children aged 0 to ≤3 years and N 3 to ≤6 years (P b 0.05), among children aged 0 to ≤3 years and N 6 to 14 years (P b 0.0001), and among children aged N3 to ≤6 years and N 6 to 14 years (P b 0.0001), but not among children 0 to ≤6 months and N 6 months to ≤3 years (P N 0.05).
HRSV had the highest prevalence (223/328: 67.98%) in children aged 0 to ≤6 months, followed by HCOV (66/328: 20.12%) and HADV (42/328: 12.8%), while HADV prevailed in children N6 months to ≤3 years (218/557: 39.13%), followed by HRSV (167/557: 29.98%) and HCOV (103/557: 18.49%) (Fig. 3B) . In children N3 to ≤6 and N 6 to 14 years, HADV prevailed (94/243: 38.68% and 48/167: 28.74%, respectively), followed by IV (55/
Of the 1408 virus-positive samples, 1373 were detected by molecular assays and 426 by culture (97.51% vs. 30.25%, P b 0.0001) ( Table 2) . Most of the culture-positive samples (421, 29.9%) were also positive by molecular methods, although different virus combinations were detected in 9 mixed infections. In discordant culture-negative samples molecular assays detected 262 HCOV (18.6%), 161 HBOV (11.43%), and 76 HMPV (5.4%). Finally, of the 466 samples positive for HRSV by PCR, 233 (50%) were also detected by IIF (100% vs. 50%, P b 0.0001).
This survey carried out in Parma (Northern Italy) from October 2012 to September 2015 describes the detection of viruses associated with respiratory samples from 2575 patients presenting with symptoms consistent with acute respiratory tract infections. Overall, 50.44% of cases were positive by the simultaneous employment of different diagnostic assays. Other studies have reported similar rates, although they refer to a single epidemic season and fewer number of cases examined exclusively by PCR (Annan et al., 2016; Pratheepamornkull et al., 2015; Zuccotti et al., 2011) . However, different results have also been described (Do et al., 2011; Venter et al., 2011) . It must be considered that many factors could lead to prevalence variations, such as the virus panel examined, diagnostic methods, study setting, and geographic area.
HRSV (27.12%) and HADV (23.58%) prevailed; HCOV (15.25%), IV (11.53%), and HBOV (9.37%) showed a higher frequency than HPIV (6.64%), HMPV (4.42%), and HEV (2.09%). Single infections predominated over mixed infections (80.11% vs. 19.89%, P b 0.0001), which was consistent with previous observations (Martin et al., 2012; Paranhos-Baccalà et al., 2008) . Of multiple infections, dual infection (90%) was predominant, in accordance with other authors (Martin et al., 2012; Peng et al., 2009) . HRSV, HADV, and HCOV were mainly involved in mixed infections, as would be expected from their overlapping seasonal distribution. Although HRSV and HADV are among the viruses most commonly involved in mixed infections (Cantais et al., 2014; Harada et al., 2013) , the relevance of rhinoviruses, HPIV, and IV has also been assessed (Lu et al., 2013) . Very complex combinations of mixed viral infections have been observed (Lu et al., 2013; Martin et al., 2012; Paranhos-Baccalà et al., 2008) and certain pathogens appear to have higher multiple infection frequency (Brunstein et al., 2008) . In this scenario, superinfection exclusion mechanisms preventing a secondary infection with the same or a closely related virus should be envisaged (Folimonova, 2012) . Although mixed infections have been widely assessed in ARTIs (Moesker et al., 2016; Seo et al., 2014; Turner et al., 2013) , there is still no clear relationship with prolonged virus shedding, asymptomatic virus persistence, and disease severity. However, there is evidence that genome quantification could clarify the role of each virus in mixed infections (Franz et al., 2010; Martin et al., 2012) . Accordingly, high viral loads have been related to severe diseases (Franz et al., 2010; Hasegawa et al., 2014; Jansen et al., 2011) , but conflicting results have been reported (Adams et al., 2015; Wu et al., 2012) . In addition, some viruses have been detected in healthy children (Advani et al., 2012; Moe et al., 2016) . In this study, molecular assays showed greater sensitivity when compared to culture assays (97.51% vs. 30.25%, P b 0.0001) and HRSV antigen detection (100% vs. 50%, P b 0.0001). Conversely, culture assays allowed the detection of infectious viruses in 426 (30.25%) of the 1408 positive samples, eventually providing a clue to discriminate the main causative agent in co-infections. Nevertheless, cautious conclusions should be drawn, since the growth difficulties of some viruses may have decreased the number of those found by culture assays, while the long-term viral excretion after clinical recovery in children with frequent ARTIs may have increased the mixed infection rate.
The epidemic season of respiratory viruses ranged from 5 to 6 months, with peaks in February (2013 and 2014) or January (2015). The virus detection rate decreased significantly between the first and third seasons (53.9% vs. 43.05%, P b 0.0001), but environmental factors may have influenced the frequency and severity of ARTIs (Nenna et al., 2017) .
HRSV showed a sharp winter/early-spring seasonality. When compared to previous reports, in the Parma area the HRSV epidemic season was shorter (Medici et al., 2006) .
HADV circulated persistently throughout the study with higher rates in fall and spring of the first two years of surveillance, eventually indicating a reduction of circulation due to the accumulation of immunity in the population.
The IV infection seasonality was consistent with previous reports (Del Manso et al., 2015a Manso et al., , 2015b .
No definite seasonality was shown by HBOV, HCOV, HMPV, HPIV, and HEV.
A correct viral diagnosis allows highlighting virus seasonality, which could favor both preventive measures and reduction of overuse of antibiotics. Accordingly, in children with HRSV or IV, bacterial infections are less prevalent than in those without a viral infection (Purcell and Fergie, 2004; Smitherman et al., 2005) .
The virus detection rate was significantly higher in children aged 0 to ≤3 years (68.34%), when compared to the N3 to ≤6 years (18.76%, P b 0.05) and N 6 to 14 years (12.9%, P b 0.0001) groups, suggesting that immaturity of the immune system and absence of previous immunity could influence the severity of ARTIs (Hammond et al., 2007; Monto, 2004) .
HRSV prevailed (67.98%) in younger infants (0 to ≤6 months), while HADV was prevalent in all other age groups studied. Accordingly, HRSV causes severe diseases in young children (Moe et al., 2017; Nguyen et al., 2017; Shi et al., 2017) .
For all the remaining respiratory viruses the highest detection rate was among children aged N6 months to ≤3 years. The attribution of an etiological role in ARTIs to HBOV has been extensively debated (Debiaggi et al., 2012; Schildgen et al., 2008) . In this study, HBOV was most frequently found in single infections (55.9%) than in coinfections, strengthening the idea that it can be considered a pathogen.
In conclusion, although this study has certain limitations, such as the lack of rhinovirus detection, it does provide relevant information on respiratory virus circulation in children with ARTIs in an area of Northern Italy with a temperate climate. The simultaneous use of different diagnostic approaches made it possible to carry out in-depth analysis of a large number of cases spanning over 3 years.
These findings contribute to estimate the disease burden associated with respiratory viruses, reinforcing the need for timely virologic diagnosis and continuous surveillance to optimize the prediction and control of ARTIs in children.
Supplementary data to this article can be found online at https://doi. org/10.1016/j.diagmicrobio.2019.01.008.
Funding This study was funded by grants from the University of Parma (Fondi di Ateneo FIL 2014 -FIL2014_DECONTO_MCS).
This article does not describe any studies with human participants or animals performed by any of the authors.
Respiratory virus detection was performed according to the medical order; hence there was no need to obtain informed consent for the epidemiological analysis of the related data.
|
